srs field guide

WSRC-MS-2000-00606

CAROLINA GEOLOGICAL SOCIETY2000 FIELD TRIP GUIDEBOOK

Savannah River SiteEnvironmental Remediation Systems In Unconsolidated

Upper Coastal Plain Sediments - Stratigraphic andStructural Considerations

Edited by:D. E. Wyatt and M. K. Harris

WSRC-MS-2000-00606

Carolina Geological Society

PresidentW. A. (Bill) Ransom

Department of Earth and EnvironmentalScience

Furman UniversityGreenville, SC 29613

Vice PresidentKevin Stewart

Department of Geological Sciences, CB #3315UNC-Chapel Hill

Chapel Hill, NC 27559

Secretary – TreasurerDuncan Heron

Division of Earth and Ocean Sciences, Box90230

Duke UniversityDurham, NC 27708-0230

Board Members

Jim FurrKubal-Furr & Associates

P. O. Box 80247Simpsonville, SC 29680

Dennis Lapoint105 Turnage Road

Chapel Hill, NC 27514

Joe McMurray109 Hillside DriveShelby, NC 28150

Dave Willis62 Williams RoadTrenton, SC 29847

Contributing Sponsors(at press time)

RCS Corporation, Aiken, SC

Science Applications International Corporation (SAIC), Augusta, GA

Exploration Resources, (EXR), Augusta, GA

Kubal-Furr & Associates, Simpsonville, SC

Blue Ridge Environmental, Shelby, NC

CSRA Geological Society, Aiken, SC

napLogic, Princeton, NJ

Arcadis Geraghty and Miller, Aiken, SC

Bunnell-Lammons Engineering, Inc., Greenville, SC

Carolina Geological Society, 2000 Annual Field Trip Guidebook WSRC-MS-2000-00606

iii

CAROLINA GEOLOGICAL SOCIETY2000 FIELD TRIP GUIDEBOOK

Aiken, South CarolinaNovember 4th, 2000

Savannah River SiteEnvironmental Remediation Systems

In Unconsolidated Upper Coastal Plain SedimentsStratigraphic and Structural Considerations

Edited by:Douglas E. Wyatt

andMary K. Harris

Co-sponsored by:

This document was prepared in conjunction with work accomplished under Contract No.DE-AC09-96SR18500 with the U.S. Department of Energy.

DISCLAIMER

This report was prepared as an account of work sponsored by an agency of the United States Government.Neither the United States Government nor any agency thereof, nor any of their employees, makes anywarranty, express or implied, or assumes any legal liability or responsibility for the accuracy,completeness, or usefulness of any information, apparatus, product or process disclosed, or represents thatits use would not infringe privately owned rights. Reference herein to any specific commercial product,process or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute orimply its endorsement, recommendation, or favoring by the United States Government or any agencythereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of theUnited States Government or any agency thereof.

This report has been reproduced directly from the best available copy.

Available for sale to the public, in paper, from: U.S. Department of Commerce, National TechnicalInformation Service, 5285 Port Royal Road, Springfield, VA 22161, phone: (800)553-6847, fax: (703) 605-6900, email: [email protected] online ordering:http://www.ntis.gov/ordering.htm

Available electronically at http://www.doe.gov/bridge

Available for a processing fee to U.S. Department of Energy and its contractors, in paper, from: U.S.Department of Energy, Office of Scientific and Technical Information, P.O. Box 62, Oak Ridge, TN37831-0062, phone: (865 ) 576-8401, fax: (865) 576-5728, email: [email protected]


v

Table of ContentsA-1. Description Of Field Stops ..........................................................................................................Carolina Geological Society

A-2. Savannah River Site Overview and General Site Information................................................................... edited by D. Wyatt

A-3. Overview of the Savannah River SiteStratigraphy, Hydrostratigraphy and Structure...............................................edited by D. Wyatt, R. K. Aadland, F. H. Syms

B-1. Establishing a Hydrostratigraphic Framework usingPalynology: An Example from the Savannah River Site,South Carolina, U.S.A........................................................................................................R. S. Van Pelt, D. W. Engelhardt,

R. A. Christopher, J. Lucas-Clark

C-1. Methodology and Interpretation of the PiezoconePenetrometer Test Sounding for EstimatingSoil Character and Stratigraphy at the Savannah River Site.........................................F. H. Syms, D. E. Wyatt, G. P. Flach

D-1. Contaminant Transport and Hydrologic Properties ofFault-Propagation Folds in Poorly Consolidated CoastalPlain Sediments .....................................................................................................................R. J. Cumbest and D. E. Wyatt

E-1. Outline of The Geology of Appalachian Basement Rocks Underlyingthe Savannah River Site, Aiken, SC........................................................................................................................A. Dennis

F-1. Significance of Soft Zone Sediments at theSavannah River Site..........................................................................................R. K. Aadland, W. H. Parker.,R. J. Cumbest

G-1. Carbonate Sediments in the General SeparationsArea, Savannah River Site, South Carolina ........................................................ R. K. Aadland, P. A. Thayer, W. H. Parker,

J. M. Rine, D. W. Engelhardt, R. J. Cumbest

H-1. The Savannah River Site E-Area Vadose ZoneMonitoring System........................................................................................................H. H. Burns, D. E. Wyatt, T. Butcher,

J. Cook, B. Looney, J. Rossaabi, M. Young

I-1. Modeling Aquifer Heterogeneity at the Savannah River Siteusing Cone Penetrometer Data (CPT) and StochasticUpscaling Methods ..................................................................................... M. K. Harris, G.P. Flach, A.D. Smits, F.H. Syms

J-1. Soil Vapor Extraction Remediation of TrichloroethyleneContamination at the Savannah River Site’sC-Area Burning Rubble Pit................................................................................................C. Switzer, D. Kosson, J. Rossabi

K-1. DNAPL Penetration of Saturated Porous MediaUnder Conditions of Time-DependentSurface Chemistry.................................................................................................... D. M. Tuck, G. M. Iversen, W. A. Pirkle

L-1. Subsurface Correlation of Cenozoic Stratain the Updip Coastal Plain,Savannah River Site (SRS), South Carolina.........................................................................R. K. Aadland and P. A. Thayer


vi

Table of Contents (Continued)

M-1. Geological Interpretation of the Pre-EoceneStructure and Lithostratigraphy of the A/M Area,Savannah River Site, South Carolina...................................................................D. E. Wyatt, R. K. Aadland, R. J. Cumbest

N-1. Use of Seismic Reflection Amplitude Versus Offset (AVO)Techniques to Image Dense Nonaqueous Phase Liquids(Dnapl) at the M-Area Seepage Basin, Savannah River Site,South Carolina ............................................................................................. M. G. Waddell, W. J. Domoracki, T. J. Temples

O-1. Geology: Improving Environmental Cleanupof the A/M Area, Savannah River Site .......................M. K. Harris, D. G. Jackson, B. B. Looney, A. D. Smits, R. S. VanPelt

P-1. Short Note:Seismic Monitoring at SRS................................................................................................................................D. Stevenson

Q-1. Short Note:Formation Damage Caused by Excessive Borehole FluidPressures in Coastal Plain Sediments ................................................................................................................. D. E. Wyatt

R-1. Bibliography ............................................................................................................................................................. Selected


A-1

DESCRIPTION OF FIELD STOPS

Stop #1 - H-Area Seepage Basin Extraction,Treat and Re-Injection Project(Coffee, juice and doughnuts provided at thisstop. Please follow all instructions and watchfor barricade ropes and warning signs.)

Until 1988, unlined seepage basins were used todispose of wastewater from the Savannah RiverSite's (SRS) separations facilities. Six were placedin operation in 1955, and a seventh was constructedand became operational in 1962 to replace a basinthat had stopped seeping. The wastewater wastransported approximately 3,000 feet from eachprocessing area through underground vitrified claypipes to the basins. The wastewater was allowed toevaporate and to seep into the underlying soil.

Use of all the basins was discontinued on Nov. 7,1988, when the F and H Effluent TreatmentFacility replaced them. The F and H EffluentTreatment Facility continues to fulfill this functionthrough chemical/mechanical treatment at thesurface. Annually, an average of about 80 milliongallons of wastewater was disposed in the basins.The three F Area basins cover about 6.5 acres, andthe four H Area basins cover about 15.5acres.

A groundwater monitoring well network wasinstalled in the 1950s. The monitoring network hasbeen continually expanded. Approximately 234monitoring wells, covering approximately 275acres in these areas, are currently sampled for avariety of chemical and radioactive parameters.Groundwater monitoring results show tritium andnitrate contamination, with elevated levels of someheavy metals and radionuclides. Some of the heavymetals are classified as hazardous waste as definedunder the Resource Conservation and RecoveryAct (RCRA).

In 1986, it was determined that the units should beregulated under RCRA as mixed waste (wastecontaining hazardous and radioactive components)disposal facilities, and closure plans were initiated.Closure activities began in June 1989, immediatelyafter the South Carolina Department of Health andEnvironmental Control (SCDHEC) approved theclosure plans. Closure activities were completed in1991, with SCDHEC certifying the F and H basinclosures. The total cost of closing the basins was$17 million. The overall basin closure planincluded:

♦ draining each basin;♦ placing granite aggregate on the basin bottoms

to stabilize the sludge and prevent airbornereleases of radionuclides during construction;

♦ installing limestone and blast furnace slag onthe aggregate to chemically stabilizecontaminants in the bottom of the basin;

♦ backfilling the basin with compacted soil;♦ placing a 2-foot thick clay cap over the basins

to minimize precipitation reaching thecontaminants; and

♦ covering the clay cap with topsoil, thensodding and/or seeding.

Groundwater remediation activities to reduceheavy metals, nitrates and some radionuclides andhydraulically control tritium are on-going.

Major components of each remediation systeminclude:

♦ groundwater extraction and collection system;♦ modular groundwater and secondary waste

treatment systems;♦ secondary waste collection and packaging

system;♦ treated water distribution and injection system.

Additional activities currently underway to furtherenhance contaminant removal include:

♦ modeling, fate and transport, chemicalinteraction and estimates to achievegroundwater clean up goals.

♦ Waste Minimization with a pilot demon-stration of the Disc Tube Nanofiltration Unit.

♦ Source reduction through Hot Spot Wells.

Stop #2 - Radioactive Burial Grounds,E-Area Vadose Zone Modeling Project(No eating, drinking or smoking is allowedduring this stop. Please follow all instructionsand watch for barricade ropes and warningsigns.)

The Burial Ground Complex occupiesapproximately 194 acres in the central section ofthe Savannah River Site (SRS) between the F andH Separations Area. The complex consists ofseveral former or present disposal sites for


A-2

hazardous and radioactive wastes from SRSactivities. It is divided into two areas:

1. The southern area, making up 76 acres, iscalled the Old Radioactive Waste BurialGround. In 1952, it began receiving wastesuntil it reached its full storage capacity in1972. The area contains 22 underground tankswhere used radioactive solvents were oncestored. Wastes are buried in slit trenches.

2. The northern area, called the Low-LevelRadioactive Waste Disposal Facility, occupies119 acres. It includes underground tanks forradioactive solvent storage, transuranic wastestorage pads and engineered and slit trenches.The facility began receiving wastes in 1970,and disposal continues there today. The 58-acre Mixed Waste Management Facility(MWMF), which is one of three separate areasof the 119-acre Low-Level Radioactive WasteDisposal Facility (LLRWDF), was closed inaccordance with RCRA regulations in 1991.Five additional areas, totaling 24 acres, of theLLRWDF were closed in accordance withRCRA regulations in the 1996-99 time frame.The remaining 37 acres of the LLRWDF areactive operations areas. Most of the 76-acreOld Radioactive Waste Burial Ground is beingcovered by a low permeability soil in 1996-98,in accordance with CERCLA Interim ActionRecord of Decision.

The degree of soil contamination in the complex isunknown. However, potential contaminants includechlorinated and organic solvents and toluene andbenzene, both components of degreasing agentsused to clean metal parts. Metals of concern arelead, cadmium and mercury. Potential radioactivematerials in the soil include alpha, beta and gammaray emitting materials.

Preliminary groundwater analyses in the complexshow contamination from tritium, cadmium, lead,mercury and chlorinated organic compounds.Tritium plumes are tracked originating from thewestern end of the Old Radioactive Waste BurialGround and the northern areas of the MWMF.

Radioactive operational and construction wastesare currently being disposed of in engineered slittrenches within the operational portion of thefacility. As part of the modeling and performanceassessment for the trenches, a vadose zonecharacterization program was initiated. Numerousvadose zone soil borings with sediment sampling

for hydraulic parameters were obtained forinstrumnt calibration and to sablish baselineconditions. CPT’s, neutron probe access wells,vertical and angled lysimeters, Vapor wells, TimeDomain Reflectometers, and AdvancedTensiometers have been installed in selectedhorizons to monitor moisture amount and flux. Thelysimeters are sampled for potential contaminantconcentrations. Using the CPT and boring data, athree dimensional model of the vadose zone hasbeen completed.

Lunch Stop - The Proposed Areas for theDOE Plutonium Disposition Facilities(A quick brown bag lunch will includesandwich, chips, apple, cookie and drink. Arestroom facility is available.)

The DOE Plutonium Disposition mission includesthe construction of three facilities proposed forsurplus plutonium disposition. These facilitiesinvolve pit conversion, immobilization, andmixed oxide (MOX) fuel fabrication. Eachprocess would be carried out within its ownspecific facility. Pit conversion would be donein the Pit Disassembly and Conversion Facility(PDCF), immobilization would be conductedin the Plutonium Immobilization Project (PIP)facility, and the MOX Fuel Fabrication Facility(MFFF) would produce MOX fuel assembliesfor domestic commercial power reactors.Several subsurface investigations have beencompleted in support of site selection andpreliminary design. Currently, structurespecific investigations are underway.

Stop 3 - TNX, Geo-Siphon and VadoseZone Remediation Project, TNX Ground-water Operable Unit

The TNX Area is an industrial facility where pilot-scale testing and evaluation of chemical processesin support for activities such as the Defense WasteProcessing Facility (DWPF), Separations Area, andfuel and target manufacturing areas took place.Groundwater contamination at TNX resulted fromthe use of unlined seepage basins to dispose ofwastewater, leakage from a network of processsewers, and leachate from various activities in thearea.

The TNX Groundwater has been found to becontaminated with chlorinated volatile organiccompounds (VOCs), primarily trichloroethylene


A-3

(TCE), and to a lesser extent tetrachloroethylene(PCE), and carbon tetrachloride. The VOC plumeunderlies eight acres, and has a maximumthickness of 20 feet. Groundwater characterizationactivities indicate that contamination is limited tothe shallow water table aquifer, and thecontaminant plume is outcropping in the TNXswamp, before it reaches the Savannah River.

An Interim Action Record of Decision (IROD) forthe TNX Groundwater Operable Unit (GWOU)was authorized by EPA, SCDHEC, and DOE onNovember 16, 1994. The objectives of the interimaction are: 1) reduce potential risk to human healthand the environment, 2) maintain risk at acceptablelevels to the onsite worker at the seepline, 3) massremoval of VOC contamination in the groundwaternear the plume core, 4) plume stabilization byinhibiting migration of elevated levels of VOCs(500 ppb TCE) to the swamp and 5) prevent furtheraquifer degradation.

The selected remedy for accomplishing the InterimRemedial Action goals was designated the HybridGroundwater Corrective Action (HGCA) system.The system has two components: 1) traditionalpump and treat technology to treat and inhibitfurther migration of the 500 ppb dissolved VOCplume, and 2) an innovative in-situ technology,airlift re-circulation well, located at the heart of theplume to expedite remediation.

Pump and treat is proven technology forgroundwater remediation. The airlift re-circulationwell is an emerging technology that couldpotentially reduce the contamination in the aquiferreducing remediation operation time frames for thepump and treat system. Based on testing performedin late FY 1996, it was determined that the re-circulation well was not effective in removingcontaminants at this location due to site specificconditions. Furthermore, it was determined that thepump and treat system would adequately meet theremedial objectives of the Interim Record ofDecision. Consequently, it was decided todiscontinue further operation of the re-circulationwell at TNX.

Other technological approaches are beingconsidered and/or evaluated, in the event that theInterim Remedial Action does not suffice for afinal remedy. These include: the IntrinsicRemediation Investigation, the GeoSiphon CellPilot Study, and the Soil Vapor ExtractionInvestigation.

The Intrinsic Remediation Investigation is a UnitedStates Geological Survey (USGS) project to assessnatural attenuation as a remedial strategy for thechlorinated volatile organic compoundcontamination in the groundwater of the TNXArea. The potential intrinsic remediation pathwaysto be evaluated include bioremediation, sorption,and phytoremediation.

A field trial of a GeoSiphon Cell is being pursuedin the TNX Area flood plain. The GeoSiphon Cellis essentially a large diameter well containinggranular cast iron, which reduces the COVCs in thegroundwater to ethane, methane, and chloride ions.The flow of groundwater through the treatment cellwill be passively induced by the natural hydraulichead difference between the cell and the SavannahRiver, where the treated groundwater willultimately be discharged.

Soil vapor analysis at TNX show that CVOCs arepresent in the soil at concentrations up to 55 ppmv.The residual CVOCs will continue to be a sourceof groundwater contamination as long as they arepresent in the subsurface. A pilot-study has beenproposed to evaluate the applicability of soil vaporextraction at TNX utilizing existing infrastructure.Removal of residual contaminants from theunsaturated zone would significantly reduce theoperating time frame of the air stripping system.

Stop 4 - C-Area Burning Rubble Pit, CPTGeo-Modeling and Air Sparging Remedia-tion

The C Area Burning/Rubble Pit is west of C Areaat the Savannah River Site (SRS). It is a shallow,unlined, earthen pit SRS excavated in 1951 toreceive used organic solvents, waste oils, paper,plastics and rubber materials. The pit is 350 feetlong, 25 feet wide and 10 feet deep.

Although the practice was discontinued in 1973,SRS periodically burned the pit debris. From 1973on, the pit continued to receive construction anddebris and empty, non- returnable drums. When thepit was full, workers backfilled it with soil andsediments to ground level.

SRS took preliminary soil gas surveys in andaround the pit in 1985 and 1986. These showchlorinated solvents in the soil. Furtherinvestigations took place in 1989, when SRS tooksoil samples and used ground-penetrating radar todefine the pit boundaries and locate buried objects.


A-4

In addition to chlorinated solvents, metalcontaminants of concern include antimony, lead,arsenic, vanadium, chromium, cadmium, nickel,barium, copper, tin and zinc. Also of concern arecyanide, xylene and phenolic constituents.

Preliminary groundwater monitoring conductedfrom 1984 through 1990 shows groundwaterbeneath the pit contains chlorinated solvents. Thehistory of solvent disposal at the CBRP and thepersistence of high VOC concentrations 25 yearsafter operations ceased suggests the presence ofVOC in non-aqueous phase liquid (NAPL) withinthe vadose zone beneath the CBRP. A Soil VaporExtraction and Air-Sparging (SVE/AS)remediation system are currently in operation forcleaning up the solvents in the soil andgroundwater at this site.

Extensive CPTU data were acquired as part of theoverall subsurface characterization for the CBRP.These data were calibrated to borings and core datato create a three dimensional model of thesubsurface.

Stop 5 - A/M Area Horizontal Wells, Baro-Balls, Vadose Zone Extraction, Modeling,Groundwater Pump and Treat, DNAPL,High Resolution Seismic Imaging for En-vironmental Characterization, A/M AreaGroundwater Cleanup(Soft drinks and snacks will be available atthis stop.)

In 1983, the SRS initiated one of the largest andmost successful remediation programs aimed atcleaning up organic solvents from the soils andgroundwater. Located in the northern portion of theSRS, the 350 acre A/M Area contained facilitiesfor reactor fuel and target assemblies, supportfacilities, laboratories, and administrationbuildings. At several stages in the fabricationprocess, fuel assemblies were degreased usingindustrial solvents, primarily trichloroethylene(TCE) and tetrachloroethylene (PCE).

From the 1950s into the early 1980s, wastewaterfrom fuel and target manufacturing operations wasdisposed in a settling basin. The basin, used from1958 to 1985, was an 8-million-gallon earthenbasin that received fluids from the M Areamanufacturing facilities. Waste entered the M AreaSettling Basin through an underground process

sewer line from a manufacturing facilityapproximately 2,500 feet north of the basin. Thebasin periodically overflowed to a natural seepagearea and shallow depression, known as Lost Lake,a Carolina Bay, via a drainage ditch. Thiscombined area is the M Area Hazardous WasteManagement Facility (HWMF).

The effluent contained heavy metals andchlorinated solvents, primarily trichloroethyleneand tetrachloroethylene similar to the chemicalsused in the dry cleaning industry. Most of theheavy metals were effectively captured by the soil.About half of the solvents released evaporated.Monitoring wells installed in 1981, however,showed that the remainder had seeped into thewater table, contaminating the groundwater.

After discovering groundwater contaminationbelow the M Area Settling Basin in June 1981,SRS established an interim groundwatermonitoring program. Multiple groundwater zonemonitoring wells were installed surrounding thebasin, the principal source of groundwatercontamination in the A/M Areas.

Groundwater cleanup was instituted voluntarily inFebruary 1983, using an experimental air-strippingsystem. This pilot testing project included a singlegroundwater pumping well and a prototype 70-gallon-per-minute air stripper unit. Later, a full-scale pump-and-treat system was constructed toremediate contaminated groundwater in M Area.The system, comprised of 11 groundwater recoverywells and an air stripper column, treatscontaminated groundwater from the shallowaquifer at a rate of approximately 550 gallons perminute. This system has been operating almostcontinuously since 1985, the year SRS submittedthe M Area Part B Permit corrective action plan tothe South Carolina Department of Health andEnvironmental Control (SCDHEC). The plan, oneof the first designed at a DOE facility, wasapproved in 1987 and requires operation andmonitoring for 30 years.

Today an extensive monitoring well networkconsisting of more than 400 wells and extendsthroughout the A/M Areas. Groundwater qualityand hazardous constituents defined by regulatorsare analyzed regularly at the wells. These activitieshave been conducted in compliance with the MArea Resource Conservation and Recovery ActPart B Permit.


A-5

The M Area air stripper has removed more than350,000 pounds of solvent from over 2.3 billiongallons of groundwater. Combined with thecontaminated groundwater treated by experimentaland prototype remediation units and the 5 soilvapor extraction units, more than 500,000 poundsof solvents have been removed from thesubsurface. These groundwater remediationactivities have been conducted in compliance withapproved wastewater and air quality controlpermits.

Air stripping systems work by pumpingcontaminated groundwater to the top of an airstripping column. As groundwater cascadesdownward through the column, pumped air isforced upward from the bottom of the column.When the water mixes with air, solvents in thegroundwater move from a liquid phase into a vaporphase, and volatile contaminants are stripped, andtreated prior to release to the atmosphere. When thesystem first came into use, the inlet solventconcentration of untreated groundwater wasapproximately 40,000 parts per billion. Theconcentration has now dropped to approximately9,000 parts per billion. Solvent concentration afterthe water has been treated in the air stripper isalmost always less than one part per billion. Thecleaned water is discharged through a permittedoutfall to a nearby stream.

A program was initiated in 1992 to removesolvents from the vadose zone, the layer ofunsaturated soils above the groundwater.Remediation of the vadose zone will further reducethe potential for more groundwater contamination.Five remediation systems, that use vacuumextraction to remove the solvents, have beeninstalled and are operating at known contaminationareas throughout the A/M area. Each unit isequipped with an offgas treatment system. Thistechnique will reduce cost and time of remediationwhile increasing remediation efficiency and publicand regulatory acceptability.

SRS researchers are continuing to develop newstrategies for treating solvent contamination. Thiswork includes geological modeling andgeochemical studies. In addition, a processutilizing horizontal wells, called in-situ airstripping, has been developed at SRS to removecontaminants in the groundwater and the vadosezone. During pilot program operation of the in-situair stripping system, 16,000 pounds of solventswere removed. This method is shown to be fivetimes as effective as traditional pump-and-treat

technology. In-situ bioremediation also tested withsuccess. SRS has proposed using in-situbioremediation at the Non Radiological WasteDisposal Facility.

Concentrated Contaminants, DNAPLs

DNAPLs are contaminants in an oil-like state thatare derived from chlorinated solvents, creosote andcoal tar used in industries ranging from drycleaners to electronic instrument makers. DNAPLsat SRS contain the contaminantstetrachloroethylene and trichloroethylene, and aredifficult to remediate.

SRS is looking at innovative ways to address theDNAPLs found under A/M Area. Although there isyet no known technology to clean up DNAPLs,scientists are researching several possiblealternatives. One technique would include injectioninto the groundwater of solutions that woulddestroy the DNAPLs underground allowing themto breakdown into harmless components.

Vadose Zone Remediation Units

Contamination to the vadose zone occurs ascontaminants seep downward from sources likewastewater basins, pipe-line breaks or surfacespills of chemicals. During the production andprocessing years at SRS, chemical solvents used bythe industry were released to wastewater dischargebasins or to surface drainage facilities. This was atypical and accepted practice by industry in thepast that is not allowed today without strictregulatory controls.

The Vadose Zone Remediation Process

SRS has recently added five vadose zone units tothe environmental cleanup program in A/M Area.These units remove solvent vapors at the rate ofabout 200 pounds per day from the vadose zone.Addition of vadose zone treatment increases thecleanup rate at the A/M Area by 500 percent. Hereis how it works:

♦ Vacuum units, connected to either horizontalor vertical wells placed into contaminatedsources like basins or spill areas, withdrawcontaminant vapors from the subsurface.

♦ The contaminated vapors pulled from thesubsurface flow to a thermal treatment unit,mounted near the vapor extraction well, which


A-6

destroys the contaminants via catalyticoxidation.

Extraction of these contaminants from the vadosezone prevents their downward movement into thewater table.

Cost Effectiveness

The soil vapor extraction technology is proven tobe cost effective, compared to other methods ofclean-up, including the removal of contaminatedsoil. The total project cost for the vadose zoneremediation project is approximately $4 million,compared to approximately $14 million foralternative methods of soil cleanup.

CORE AND GIS WORKSHOPS

Biology and Geology Department, USC-Aiken(Parking is available in the lot behind, west, ofthe science building. Please park in the white-lined spaces. The workshops will be on the 2nd.

floor of the Science Bldg., down the steps)

8:30 - 10:00 ........ Core Workshop 18:30 - 10:00 ........ GIS Geology Database

Presentation 110:00 ................... Refreshments10:30 - Noon ....... Core Workshop 210:30 - Noon ....... GIS Geology Database

Presentation 2

Core Workshop

The core workshop will consist of a hands-on viewing of representative sedimentsfrom the SRS. A presentation with an opendiscussion will be the format.

GIS Geology Database Presentation

The GIS presentation will demonstrate theutilization of GIS and geological data toaid in characterization and remediation.Data utilization, availability, andaccessibility will be discussed.


A-7

Stop #4 - TNX GeosiphonRemediation

Stop #1 - H Area SeepageBasin Pump & Treat Systems

Stop #2 - E Area Burial GroundsVadose Zone Modeling Project

Stop #5 - A/M Area RemediationProjects and Demonstrations

Stop #3 - C AreaSoil Vapor ExtractionRemediation

8 0 8 16 24 Kilometers

Geology (after Prowell)Quaternary alluvium 1Quaternary alluvium 2DunesDry Branch FormationHuber and Congaree FormationMcBean FormationTobacco Road Sandupland unit

SRS AreasMain Roads

N

CGS Field Trip Stops

Figure 1. SRS tour stop locations. Geology is from Prowell (1994).


A-8

Holley Inn235 Richland Ave.803-648-4265

USCA Science BuildingBiology and Geology

Second Floor

#

Richland Ave.

#

University Parkway

(/ 1

1 0 1 2 Miles

N

Sunday Workshops

Figure 2. Route from the Holley Inn to USCA.


A-9

Savannah River Site Overview and General Site Information

Compiled from various sources, edited by Doug Wyatt

INTRODUCTION

The Savannah River Site (SRS) is a keyDepartment of Energy industrial complexdedicated to the stewardship of theenvironment, the enduring nuclear weaponsstockpile and nuclear materials. Morespecifically, the SRS processes and storesnuclear materials in support of the nationaldefense and U.S. nuclear non-proliferationefforts. The site also develops and deploystechnologies to improve the environment andtreat nuclear and hazardous wastes left fromthe Cold War.

The SRS complex covers 198,344 acres, or310 square miles encompassing parts ofAiken, Barnwell and Allendale counties inSouth Carolina, bordering the Savannah River.

The site is owned by DOE and operated by anintegrated team led by WestinghouseSavannah River Company (WSRC). Underthe contract that went into effect Oct. 1, 1996,WSRC is responsible for the site's nuclearfacility operations; Savannah RiverTechnology Center; environment, safety,health and quality assurance; and all of thesite's administrative functions. The team alsoincludes Bechtel Savannah River Inc. (parentcompany: Bechtel National Inc.), which isresponsible for environmental restoration,project management, engineering andconstruction activities; BWXT SavannahRiver Company (parent company: BWXTechnologies), which is responsible forfacility decontamination anddecommissioning; and British Nuclear Fuels,Limited (BNFL) Savannah River Corporation(parent company: BNFL Inc.), which isresponsible for the site's solid waste program.

SAVANNAH RIVER SITE FOCUS

The Savannah River Site is committed to ourpeople, missions and the future. SRS has a

long track record of being the safest site in theDOE complex and one of the safest majorindustrial sites in the world. Protectingworkers, the public, the environment, andnational security interests is our highest goal.SRS will continue to maintain neededfacilities and infrastructure while training andretaining a skilled and motivated workforce toinsure our technical capability andperformance. The SRS team has madecommitments to its regulatory organizations,to the two states of the Central SavannahRiver Area, and to the community.Recognizing the imperative of opencommunication and trust, SRS will strive toaccomplish regulatory milestones andcommunity-driven obligations among thesite’s various neighbors and stakeholders. Wealso focus on cost effectiveness in contract andproject management and a cross-cuttingcorporate perspective that will best serve SRS,other Department of Energy sites and nationallabs, and the U.S. Government.

SRS BUSINESS APPROACH

SRS is positioned for continued success by ourcommitment to a business approach focusedon four cornerstones of success: technical andsafety excellence, a DOE mission-supportiveinfrastructure, cost effectiveness andcommunity support.

While the changing world has caused adownsizing of the site’s original defensemission, the future of SRS lies in severalareas: reducing the nuclear danger,transferring applied environmental technologyto government and non-government entities,cleaning up the site and managing the wasteSRS has produced, and forming economic andindustrial alliances.


A-10

HISTORY

During the early 1950s SRS began to producematerials used in nuclear weapons, primarilytritium and plutonium-239. Five reactors werebuilt to produce nuclear materials. Also builtwere support facilities including two chemicalseparations plants, a heavy water extractionplant, a nuclear fuel and target fabricationfacility, a tritium extraction facility and wastemanagement facilities.

Irradiated materials were moved from thereactors to one of the two chemical separationsplants. In these facilities, known as“canyons,” the irradiated fuel and targetassemblies were chemically processed toseparate useful products from waste. Afterrefinement, nuclear materials were shipped toother DOE sites for final application. SRSproduced about 36 metric tons of plutoniumfrom 1953 to 1988.

NEW MISSIONS

SRS is one of the primary DOE sites withmissions to address issues of national securityand non-proliferation, including legacymaterial disposition.

SRS has been designated to continue asDOE’s center for the supply of tritium to theenduring nuclear weapons stockpile. DOE hasannounced that its primary new source oftritium will be an existing commercial reactorin the Tennessee Valley Authority system.Tritium extraction from targets and loadinginto containers for shipment to the DefenseDepartment will continue to be a site mission.

SRS has been selected to “blend down” highlyenriched uranium from retired weaponscomponents and reactor fuel to low-enricheduranium that can be converted to commercialreactor fuel in a privatized venture.

Plutonium stabilization now being conductedat SRS will be expanded to include materialsfrom dismantled weapons and surpluses fromother DOE sites. In early January 2000, theSecretary of Energy announced that SRS was

to be the location for the Department’splutonium pit disassembly and conversion,mixed oxide fuel fabrication and plutoniumimmobilization facilities. These missionsestablish SRS’s vital role in plutoniummanagement for DOE.

ON-GOING MISSIONS

Tritium

Tritium, with a half-life of 12.5 years, must bereplenished, and SRS is the nation's onlyfacility for recycling and reloading tritiumfrom nuclear weapons reservoirs returnedfrom service. Recycling tritium allows theUnited States to stretch its tritium supplies.

All tritium unloading, mixing and loading isperformed in a facility that went into operationin 1994. It replaces older facilities thatprocessed the nation's tritium for 35 years. Anew tritium extraction facility is to be built inthe next few years to extract tritium created inthe Tennessee Valley Authority’s light-waterreactors.

Spent Fuel

Spent nuclear fuel currently stored at SRS isfrom the site’s production reactors, and fromdomestic and foreign research reactorprograms. All of this fuel is stored in water-filled concrete storage basins, which wereintended originally for interim storage whilespent fuel awaited processing in a chemicalseparations facility.

Until 1988, it was routine for foreignresearchers to return U.S.-origin spent fuel tothis country. At the urging of the U.S.Department of State and the InternationalAtomic Energy Agency, DOE renewed thatpolicy in 1996. The first shipment of foreign-research-reactor, spent nuclear fuel under therenewed policy arrived at SRS that September.

For three decades, the SRS Receiving Basinfor Offsite Fuels (RBOF) has provided safereceipt and interim storage of this fuel.Planning is under way to deinventory RBOF,


A-11

transferring spent fuel to SRS’s L AreaDisassembly Basin, a much larger, water-filled, reinforced-concrete facility. The basinwas modified and received its first shipment offoreign spent fuel in January 1997. Inaddition, studies are underway to findalternative technologies, such as dry caskstorage, for the spent fuel.

Canyon Operations

SRS has its two primary separations facilities— called canyons — located in F and H areas.F Canyon and H Canyon — together with theFB Line and HB Line, which are located atopthe canyons — are where nuclear materialshistorically have been chemically recoveredand purified.

HB Line has produced plutonium-238 forNASA. In 1995, SRS completed a five-yearcampaign to supply plutonium-238 forNASA's Cassini mission, an unmannedexpedition to the planet Saturn, which waslaunched October 13, 1997.

Currently, both canyons continue to stabilizeand manage most of the remaining inventoryof plutonium-bearing materials at SRS. FCanyon is scheduled to operate until about2002 to stabilize SRS materials. In addition,in July 1996, DOE determined that H Canyon,scheduled to operate until 2006, should beused to convert a large quantity of weapons-usable HEU to low-enriched material. Nolonger weapons-usable, the material will besuitable as fuel in commercial power reactors.

Waste Management

Weapons material production producedunusable byproducts, such as radioactivewaste. About 34 million gallons of high-levelradioactive liquid waste are stored in tanks.The Defense Waste Processing Facility(DWPF) is processing the highly radioactivewaste, bonding radioactive elements inborosilicate glass, a stable storage form.DWPF began operations in March 1996.Much of the volume in the tanks can beseparated as relatively low-level radioactive

salt solution, which is mixed with cement, ash,and furnace slag and poured into permanentconcrete monoliths for disposal at a facilitycalled Saltstone.

In addition to high-level waste, otherradioactive wastes at the site are: low-levelsolid and liquid waste; and transuranic waste,which contains alpha-emitting heavy isotopesthat have decay rates and concentrationsexceeding specified levels. Other wastesinclude hazardous waste, which is any toxic,corrosive, reactive or ignitable material thatcould affect human health or the environment;mixed waste, which contains both hazardousand radioactive components; and sanitarywaste, which, like ordinary municipal waste, isneither radioactive nor hazardous.

The site's solid, low-level radioactive wasteincludes items such as protective clothing,tools and equipment that have becomecontaminated with small amounts ofradioactive material. In October 1994, SRSopened engineered, concrete vaults forpermanent disposal of solid low-level waste.The nation’s first state-of-the-art waste vaults,they provide significantly better isolation fromthe environment than previous in-grounddisposal methods.

Waste that contains transuranic (TRU)nuclides (radioactive elements with an atomicnumber greater than uranium 92) is storedtemporarily at SRS. The site is developingcharacterization and treatment capabilities,pending eventual shipment to the WasteIsolation Pilot Plant in New Mexico.Hazardous wastes and mixed wastes are beingstored on site in Resource Conservation andRecovery Act-permitted facilities until theappropriate treatment facilities are operational.

Environmental Restoration

In 1981, SRS began inventorying waste units.There are now 515 inactive waste andgroundwater units included in theenvironmental restoration program. Wastesites range in size from a few feet to tens ofacres and include basins, pits, piles, burial


A-12

grounds, landfills, tanks and groundwatercontamination. Remediation of the waste sitesis regulated under the Resource Conservationand Recovery Act and the ComprehensiveEnvironmental Response, Compensation, andLiability Act.

The Resource Conservation and Recovery Act(RCRA) establishes a system for tracking andmanaging hazardous wastes from generationto disposal. This act also requires correctiveaction for releases of hazardous waste at activeor inactive waste units and treatment, storage,or disposal facilities.

The Comprehensive Environmental Response,Compensation, and Liability Act (CERCLA -also known as Superfund) addresses theprotection and cleanup of the environment.This act establishes a National Priorities Listof sites targeted for assessment and, ifnecessary, restoration. SRS was placed on thislist December 21, 1989.

In addition, the DOE has entered the FederalFacility Agreement (FFA) with the U.S.Environmental Protection Agency (EPA)Region IV and the South Carolina Departmentof Health and Environmental Control(SCDHEC). The FFA, effective August 16,1993, specifies how SRS will addresscontamination or potential contamination atwaste units to meet RCRA and CERCLArequirements. This agreement is requiredunder CERCLA.

SRS works closely with state and federalregulators to determine which waste unitsrequire cleanup. If preliminary evaluationsshow that a waste unit may be a candidate forcleanup, it undergoes investigation andcharacterization. The investigation phasebegins with looking at existing unit data, thendeveloping a work plan that prescribes how tocharacterize the unit.

This investigation may find that a waste unitdoes not pose a risk to human health or theenvironment. If EPA and SCDHEC agree withthis finding, with acceptance by the public, nofurther action is needed on that waste unit.

To date, 340 acres of land have beenremediated. Also, almost four billion gallonsof groundwater have been treated, with over925,000 pounds of solvents removed. Eventhough the site has had success, this cleanupprocess is expected to take decades.

SRS seeks public participation in prioritizingthe Environmental Management program.One way this is accomplished is through theSRS Citizens Advisory Board (CAB), formedin February 1994. This group of 25individuals with diverse viewpoints providesadvice to DOE, the U.S. EnvironmentalProtection Agency and the South CarolinaDepartment of Health and EnvironmentalControl.

Research and Development

The Savannah River Technology Center – thesite’s applied research and developmentlaboratory – creates, tests and deployssolutions to the site’s technologicalchallenges. SRTC researchers have madesignificant advances in glass technology,hydrogen technology, nonproliferationtechnology, environmental characterizationand cleanup, sensors and probes, and otherfields.

The laboratory’s 750-person staff includesseveral internationally recognized experts;one-fourth of the research staff holds Ph.D’s.SRTC’s unique facilities includebiotechnology laboratories, laboratories for thesafe study and handling of radioactivematerials, a field demonstration site for testingand evaluating environmental cleanuptechnologies and laboratories for ultra-sensitive measurement and analysis ofradioactive materials.

Today, while the laboratory continues to solvethe site’s technological challenges, half of itswork now comes from non-SRS customers,including DOE-Headquarters, other DOE sitesand other federal agencies. The laboratory’slargest work-for-others contract to date is a$31 million, multi-year contract to


A-13

demonstrate and evaluate the processes thatwill be used at the Hanford Site to treat anddispose of the waste in Hanford’s waste tanks.

Economic Development

Because of the increased emphasis on sharingthe site’s expertise with the nation that, formore than four decades, has invested in itswork, SRTC now forms strategic partnershipswith private industry, academia and othergovernment agencies to apply the laboratory’sunique expertise to challenges of mutualinterest. For example, SRTC, working with abroad-based consortium, applied its extensivehydrogen expertise to the development of ahydrogen-fueled bus that became part of theCity of Augusta’s public transit fleet beforebeing shipped to another DOE site for furtherdevelopment.

The laboratory also shares its expertise bylicensing private companies to manufactureand/or market technologies created at SRTC, amove that helps American businesses sharpentheir competitive edge and provides taxpayersa second return on their investment.

Environment

Originally farmland, SRS now encompasses atimber and forestry research center managedby the U.S. Forest Service. The site alsohouses the Savannah River EcologyLaboratory, an environmental research centeroperated for DOE by the University ofGeorgia.

In 1972, DOE's predecessor agency, theAtomic Energy Commission, designated SRSas the first National Environmental ResearchPark. The site is home to the bald eagle andthe red-cockaded woodpecker, an endangeredspecies. Other endangered species, includingthe shortnose sturgeon, peregrine falcon andwood stork, visit the site from time to time.Other wildlife commonly found on the siteincludes alligators, whitetailed deer, wildturkeys and otters.

Employment

Today, about 13,900 people are employed atSRS, making it one of the largest employers inSouth Carolina. About 89 percent areemployees of WSRC and its majorsubcontractors. DOE employees representabout 3.6 percent of the SRS population. Therest are other Westinghouse subcontractors andDOE contractors; the site's security contractor,Wackenhut Services Inc.; Savannah RiverEcology Laboratory; and U.S. Forest Service.

Economic Impact

The site's economic impact ripples across atwo-state area. Currently, the site's overallbudget is about $1.5 billion. Of that, morethan $900 million is payroll. In Fiscal Year1999, which ended Sept. 30, 1999, the sitepurchased about $215 million in goods andservices in South Carolina and Georgiacombined. Of that, about $110 million wasspent in the local area.


A-14

Overview of the Savannah River Site Stratigraphy, Hydrostratigraphy andStructure

Compiled from various sources, edited by D. E. Wyatt, R. K. Aadland and F. H. Syms

INTRODUCTION

This paper provides a brief overview of thestratigraphy, hydrostratigraphy and structureassociated with the Savannah River Site. Theinformation contained in this summary wasabstracted from a variety of authors andsources, the majority of whom are cited in thevarious papers in this field guide. Additional(and many of the same) references areincluded as a bibliography in this guidebook.

REGIONAL PHYSIOGRAPHY

The site region, defined as the area within a320-km (200-mile) radius of the center ofSRS, includes parts of the Atlantic CoastalPlain, Piedmont and Blue Ridge physiographicprovinces. SRS is located on the upperAtlantic Coastal Plain, about 50 km (30 miles)southeast of the Fall Line (Figure 1).

COASTAL PLAIN STRATIGRAPHY

The information in this section is based largelyon the work of Aadland et al., (1995) and hisreferences.

The sediments of the Atlantic Coastal Plain inSouth Carolina are stratified sand, clay,limestone, and gravel that dip gently seawardand range in age from Late Cretaceous toRecent. The sedimentary sequence thickensfrom essentially zero at the Fall Line to morethan 1,219 meters (4,000 feet) at the coast.Regional dip is to the southeast, although bedsdip and thicken locally in other directionsbecause of locally variable depositionalregimes and differential subsidence ofbasement features such as the Cape Fear Archand the South Georgia Embayment. A cross-section depicting these strata is presented inFigure 2.

.The Coastal Plain sedimentary sequence nearthe center of the region (i.e., SRS) consists ofabout 213 meters (700 feet) of LateCretaceous quartz sand, pebbly sand, andkaolinitic clay, overlain by about 18 meters(60 feet) of Paleocene clayey and silty quartzsand, glauconitic sand, and silt. ThePaleocene beds are in turn overlain by about107 meters (350 feet) of Eocene quartz sand,glauconitic quartz sand, clay, and limestonegrading into calcareous sand, silt, and clay.The calcareous strata are common in the upperpart of the Eocene section in downdip parts ofthe study area. In places, especially at higherelevations, the sequence is capped by depositsof pebbly, clayey sand, conglomerate, and clayof Miocene or Oligocene age. Lateral andvertical facies changes are characteristic ofmost of the Coastal Plain sequence, and thelithologic descriptions below are thereforegeneralized. A surface geologic map for SRSis presented in the “Descriptions of FieldStops” Figure 1 in this guide book. Thestratigraphic section (see Figure 2), whichdelineates the coastal plain lithology (seeFigure 3) is divided into several formationsand groups based principally on age andlithology. A reference geophysical log isshown on Figure 4.

Geology of the Coastal Plain Sediments -General

The following sections describe regionalstratigraphy and lithologies, with emphasis onvariations near the SRS. The data presentedare based upon direct observations of surfaceoutcrops; geologic core obtained duringdrilling of bore holes; microfossil age dating;and borehole geophysical logs. Several keyboring locations within the SRS boundariesand in the adjacent regions (presented inFigure 5) are referenced throughout thefollowing discussions.


A-15

Rocks of Paleozoic and Triassic ages havebeen leveled by erosion and areunconformably overlain by unconsolidated topoorly consolidated Coastal Plain. Thiserosional surface dips approximately 7 m/km(37 ft/mile) toward the southeast. TheAtlantic Coastal Plain sediments in SouthCarolina are stratified sand, clay, limestone,and gravel that dip gently seaward and rangein age from Late Cretaceous to Recent. Nearthe coast, the wedge is approximately 1,219meters (4,000 feet) thick.

Upper Cretaceous Sediments

Upper Cretaceous sediments overlie Paleozoiccrystalline rocks or lower Mesozoicsedimentary rocks throughout most of thestudy area. The Upper Cretaceous sequenceincludes the basal Cape Fear Formation andthe overlying Lumbee Group, which is dividedinto three formations. The sediments in thisregion consist predominantly of poorlyconsolidated, clay-rich, fine- tomedium-grained, micaceous sand, sandy clay,and gravel, and is about 213 meters (700 feet)thick near the center of the study area. Thinclay layers are common. In parts of thesection, clay beds and lenses up to 21 meters(70 feet) thick are present. Depositionalenvironments were fluvial to prodeltaic.

Cape Fear Formation

The Cape Fear Formation rests directly on athin veneer of saprolitic bedrock and is thebasal unit of the Coastal Plain stratigraphicsection at SRS. The saprolite ranges from lessthan 3 meters (10 feet) to more than 12 meters(40 feet) in thickness and defines the surfaceof the crystalline basement rocks andsedimentary rocks of the Newark Supergroup(Middle to Upper Triassic age). The thicknessof the saprolite reflects the degree ofweathering of the basement prior to depositionof the Cape Fear Formation. The Cape Fear isencountered at about 61 meters (200 feet) msljust south of well C-3 in the north and at about366 meters (1,200 feet) msl at well C-10 in thesouth. The Cape Fear does not crop out in the

study area, and its northern limit is north ofthe C-1 and P-16 wells and south of wells C-2and C-3. The unit thickens to more than 70meters (230 feet) at well C-10 and has amaximum known thickness of about 213meters (700 feet) in Georgia. The top of theCape Fear Formation dips approximately 5m/km (30 ft/mile) to the southeast across thestudy area.

The Cape Fear Formation consists of firm toindurated, variably colored, poorly sorted,silty, clayey sand and sandy silt and clay.Bedding thickness of the sand, silt and clayranges from about 1.5 meters (5 feet) to 6meters (20 feet), with the sand beds generallythicker than the clay beds. The sand grains aretypically coarse-grained with common granuleand pebble. The sand is arkosic with rockfragments common in the pebbly zones.

The Cape Fear Formation is more induratedthan other Cretaceous units because of theabundance of cristobalite cement in the matrix.The degree of induration decreases from northto south across the area. In the northern partof the area, the formation is represented ongeophysical logs as a zone of low resistivity.In the southern part of the study area, the unitis more sandy, and is noted on geophysicallogs by increased electrical resistivity (wellsALL-324 and C-10). The transition from themore indurated clayey sand in the north to thepoorly consolidated cleaner sand in the southmay be due to deeper fluvial incisement anderosion of the Cape Fear section to the north.This may bring the deeper, morecristobalite-rich part of the section intoproximity with the overlying unconformitythat caps the formation. Clark et al. attributethe differences between updip and downdiplithologies to changes in source materialduring deposition or to the southern limit ofthe cristobalite cementation process.

The lithologic characteristics and the paucityof marine fossils are indicative of ahigh-energy environment close to a sedimentsource area. Thus, these sediments mayrepresent deposition in fluvial-deltaicenvironments on the upper parts of a delta


A-16

plain, grading downdip to marginal marine.The Cape Fear Formation was erosionallytruncated prior to deposition of the overlyingMiddendorf Formation, resulting in adisconformity between the two formations.

Lumbee Group

Three formations of the Late CretaceousLumbee Group are present in the study area.These are, from oldest to youngest, theMiddendorf, Black Creek, and Steel CreekFormations.

The Lumbee Group consists of fluvial anddeltaic quartz sand, pebbly sand, and clay inthe study area. The sedimentary sequence ismore clayey and fine-grained downdip fromthe study area, reflecting shallow to deepmarine shelf sedimentary environments.Thickness ranges from about 122 meters (400feet) at well C-3 in the north, to about 238meters (780 feet) near well C-10 in the south.At least part of the group crops out in thenorthern part of the study area but it is difficultto distinguish the individual formations.Consequently, the Lumbee Group was mappedas undifferentiated Upper Cretaceous byNystrom and Willoughby. The dip of theupper surface of the Lumbee Group is to thesoutheast at approximately 4 m/km (20ft/mile) across the study area.

The Middendorf Formation (now in question)unconformably overlies the Cape FearFormation with a distinct contact. The contactis marked by an abrupt change from themoderately indurated clay and clayey sand ofthe underlying Cape Fear to the slightlyindurated sand and lesser clayey sand of theMiddendorf. The basal zone is often pebbly.The contact is unconformable and is markedby a sudden increase in electrical resistivity ongeophysical logs. Thickness of the formationranges from approximately 37 meters (120feet) in well C-2 in the north, to 73 meters(240 feet) in well C-10 in the south. It has amaximum known thickness of about 158meters (520 feet) in Georgia. The top of theformation dips to the southeast at about 4.9m/km (26 ft/mile) across the study area.

Fossil data for the Middendorf are sparse andthe formation is not well dated in the studyarea.

The sand of the Middendorf Formation ismedium to very coarse grained, typicallyangular, slightly silty, tan, light gray, andyellow in color. It is much cleaner and lessindurated than the underlying Cape Fearsediments. Sorting is generally moderate topoor. Pebble and granule zones are commonin updip parts of the study area, whereas claylayers up to 3 meters (10 feet) thick are morecommon downdip. Clay clasts are abundant inplaces. Some parts of the unit are feldspathicand micaceous, but not as micaceous as in theoverlying Black Creek Formation. Ligniticzones are also common.

Over much of the study area, a zone ofinterbedded sand and variegated clay up to 18meters (60 feet) thick is present at or near thetop of the Middendorf Formation. Theinterbedded sand is upward fining in places.This lithology and the marine microfaunafound in core samples indicate that the unitwas deposited in lower delta plain and deltafront environments under some marineinfluence. In the northern part of the studyarea, the formation is variably colored,composed of tan, red, and purple sand. Here,the sediments have the characteristics offluvial and upper delta plain deposits.

The Black Creek Formation is penetrated atvirtually all well-cluster sites in the study area.The unit ranges in thickness fromapproximately 46 meters (150 feet) at well C-2in the north to 91 meters (300 feet) near thecenter of the study area in well PBF-3 and to113 meters (370 feet) at well C-10 in thesouth. The unit dips approximately 4 m/km(22 ft/mile) to the southeast.

The Black Creek is distinguished from theoverlying and underlying Cretaceous units byits better sorted sand, fine-grained texture, andrelatively high clay content. It is generallydarker, more lignitic, and more micaceous,especially in the updip part of the section, thanthe other Cretaceous units. In much of the


A-17

study area, the lower one-third of theformation is mostly sand that is separated fromthe upper two-thirds of the unit by clay beds.These beds are 6 meters (20 feet) to 12 meters(40 feet) thick in the northern part of theregion and more than 46 meters (150 feet) atwell C-10 in the south. In general, the top ofthe Black Creek Formation is picked at the topof a clay bed that ranges from 3 meters (10feet) to 8 meters (25 feet) in thickness. Theclay bed is exceptionally thick but not laterallyextensive. For example, it is essentially absentin wells P-21, CPC-1, P-26, and P-29. Thissuggests lagoonal back barrier bay depositionassociated with nearby shorelines. Often thethick clay beds flank the areas where shoalingis suggested owing to uplift along the PenBranch and Steel Creek Faults, which wascontemporaneous with deposition. Overall,the Black Creek consists of two thick,fining-upward sequences, each capped bythick clay beds. The lower sequence ispredominantly silty, micaceous sand in thearea of SRS, while the upper sequence ismostly clay and silt.

Where the Black Creek Formation is presentnorth of SRS, it consists of clayey, micaceous,poorly to moderately well sorted, fine tomedium-grained, subangular to subroundedquartz sand beds and silty clay beds. Pebblybeds are present throughout the unit. Thissandy lithology is indicative of fluvial to upperdelta-plain environments; the clay beds thatcap the upward-fining sandy sequences aretypical of lower delta plain depositionalenvironments. Near Millet, SC, the basal bedsof the Black Creek consist of sand and siltyclay and are similar to underlying Middendorfsediments. Here, deposition occurred on alower delta plain. Fossils recovered from theunit suggest marine influences duringdeposition of the sediments, especially theclay.

In the central and downdip part of the studyarea (wells P-22, ALL-324, C-6, C-10), theunit grades into gray-green clayey silt, micriticclay, and fine- to medium-grained, upwardfining sand that is moderately well sorted,micaceous, carbonaceous, and locally

glauconitic. The sequence suggests depositionin a delta front or shallow shelf environment,as indicated by the lithology and an abundanceof marine macrofauna and microfauna. Thetransition from fine-grained, prodelta or deltafront deposits in the southern part of the studyarea to coarser-grained, more landward deltaicdeposits in the northern part of the area isreflected in the general increase in electricalresistivity noted on geophysical logs in thewells in the north, especially in the upper partof the Black Creek section.

The Peedee Formation was previouslyconsidered by some investigators to be absentin the study area; however, recentpaleontological evidence provides dates ofPeedee age from sediment samples in thesouthern part of SRS. Because there is aconsiderable difference in lithology betweenthe type Peedee and the sediments in the SRSregion, Peedee-equivalent sediments in thevicinity of SRS were referred to as the "SteelCreek Member” of the Peedee Formation (Ref.115). Raising the Steel Creek Member toformational status was recommended byAadland et al., and it is so used in thisdocument. The type well for the Steel CreekFormation is P-21, located near Steel Creek.The top of the Steel Creek is picked at the topof a massive clay bed that ranges from 1 meter(3 feet) to more than 9 meters (30 feet) inthickness. The formation dips approximately4 m/km (20 ft/mile) to the southeast.

The unit ranges in thickness fromapproximately 18 meters (60 feet) at well P-30to 53 meters (175 feet) at well C-10 in thesouth. It has a maximum known thickness of116 meters (380 feet) in Georgia. The SteelCreek section thins dramatically between theALL-324 and the P-22 wells due to truncationby erosion at the Cretaceous-Tertiaryunconformity. The Steel Creek Formationoverlies the Black Creek Formation and isdistinguished from it by a higher percentage ofsand, which is represented on geophysical logsby a generally higher electrical resistivity andlower natural gamma radiation count.


A-18

The formation consists of yellow, tan, andgray, medium to coarse, moderately sortedsand interbedded with variegated clay. Thelower part of the unit consists of medium- tocoarse-grained, poorly to well-sorted, quartzsand, silty sand, and off-white to buff clay thatcontains thin beds of micaceous andcarbonaceous clay. Pebbly zones arecommon, as are layers with clay clasts.Fining-upward sand is interbedded with theclay and silty clay beds in some areas. It isdifficult to differentiate the Steel Creek fromthe underlying Black Creek in thenorthwestern part of the study area. The unitappears to have been deposited in fluvialenvironments in updip areas and upper tolower delta plain environments in the south.The massive clay that caps the unit suggestslower delta plain to shallow shelf depositionalenvironments. The presence of certainmicrofossils indicates some marine influencein parts of the Steel Creek. A pebble-richzone at the base of the unit suggests a basalunconformity.

Tertiary Sediments

Tertiary sediments range in age from EarlyPaleocene to Miocene and were deposited influvial to marine shelf environments. TheTertiary sequence of sand, silt, and claygenerally grades into highly permeableplatform carbonates in the southern part of thestudy area and these continue southward to thecoast. The Tertiary sequence is divided intothree groups, the Black Mingo Group,Orangeburg Group, and Barnwell Group,which are further subdivided into formationsand members. These groups are overlain bythe ubiquitous Upland unit.

The Tertiary sedimentary sequence depositedin west-central South Carolina has beenpunctuated by numerous sea level low standsand/or affected by subsidence in the sourceareas (which reduced or eliminated sedimentavailability) resulting in a series of regionalunconformities. Four such regionallysignificant unconformities are defined in theTertiary stratigraphic section and include the“Cretaceous-Tertiary” unconformity, the

“Lang Syne/Sawdust Landing” unconformity,the “Santee” unconformity and the “Upland”unconformity. Based on these unconformities,four sequence stratigraphic units(unconformity bounded sedimentary units)have been delineated. Work is currentlyunderway to place the units in the globalsequence stratigraphic framework.

Black Mingo Group

The Black Mingo Group consists of quartzsand, silty clay, and clay that suggest upperand lower delta plain environments ofdeposition generally under marine influences.In the southern part of the study area, massiveclay beds, often more than 50 feet (15 meters)thick, predominate. Downdip from the studyarea, thin red to brown sandy clay beds, grayto black clay beds and laminated shaledominate the Black Mingo Group and suggestdeposition in clastic shelf environments. Atthe South Carolina coast, carbonate platformfacies-equivalents of the updip Black Mingoclastic sediments first appear. The carbonateunits are all referred to as "unnamedlimestones" by Colquhoun et al. These areequivalent to the thick beds of anhydrite anddolomite of the Paleocene Cedar KeysFormation and the lower Eocene glauconiticlimestone and dolomite of the OldsmarFormation. Both carbonate units aredelineated and mapped in coastal Georgia andnortheastern Florida.

Basal Black Mingo sediments were depositedon the regional “Cretaceous-Tertiary”unconformity of Aadland that defines the baseof Sequence Stratigraphic unit I. There is noapparent structural control of thisunconformity. Above the unconformity, theclay and clayey sand beds of the Black MingoGroup thin and often pinch out along thetraces of the Pen Branch and CrackerneckFaults. This suggests that coarser-grainedmaterials were deposited preferentially alongthe fault traces, perhaps due to shoaling of thedepositional surface. This, in turn, suggestsmovement (reactivation) along the faults. Thisreactivation would have occurred during Black


A-19

Mingo deposition, that is, in Paleocene andlower Eocene time.

The upper surface of the Black Mingo Groupdips to the southeast at 3 m/km (16 ft/mi.), andthe group thickens from 18 meters (60 feet) atwell C-2 in the north, to about 52 meters (170feet) near well C-10 in the south. The group isabout 213 meters (700 feet) thick at the SouthCarolina coast. Throughout the downdip partof the South Carolina Coastal Plain, the BlackMingo Group consists of the RhemsFormation and the overlying WilliamsburgFormation.

The Rhems Formation contains four members,each representing a depositional facies. Theyare the Sawdust Landing Member, an upperdelta plain fluvial deposit whichunconformably overlies the Cretaceous PeedeeFormation; the Lang Syne Member, a lowerdelta-plain deposit of estuarine and littoralorigin; the Perkins Bluff Member, a shallowshelf deposit; and the Browns Ferry Member,a deep-water shelf deposit. Additionally, anunnamed unit represents the carbonate-shelffacies.

In the updip part of the South Carolina CoastalPlain, the Black Mingo Group consists of theSawdust Landing and Lang Syne Formations,which are equivalent to the EllentonFormation of Siple; the Snapp Formation,which is the updip equivalent of theWilliamsburg Formation of Colquhoun et al.;and the Fourmile Formation, which is theupdip equivalent of the Fishburne Formationof Gohn et al.

Lang Syne/Sawdust Landing Formations.Siple proposed the name Ellenton Formationfor a subsurface lithologic unit in the SRS areaconsisting of beds of dark, lignitic clay andcoarse sand, which are equivalent to theSawdust Landing and Lang Syne Members ofthe Rhems Formation of Colquhoun et al.Fallaw and Price suggested that the SawdustLanding Member and the overlying Lang SyneMember of the Rhems Formation be raised toformational status and replace the termEllenton in the study area

In the absence of detailed paleontologicalcontrol, the Sawdust Landing Formation andthe overlying Lang Syne Formation could notbe systematically separated for mapping inthis region. Thus, they are treated as a singleunit; the Lang Syne/Sawdust Landingundifferentiated, on all sections and maps.This is consistent with the approach taken byFallaw and Price. The sediments of the unitgenerally consist of two fining-upwardsand-to-clay sequences, which range fromabout 12 meters (40 feet) in thickness at thenorthwestern boundary of SRS to about 30meters (100 feet), near the southeasternboundary. The unit is mostly dark gray toblack, moderately to poorly sorted, fine tocoarse-grained, micaceous, lignitic, silty andclayey quartz sand interbedded with dark grayclay and clayey silt. Pebbly zones, muscovite,feldspar, and iron sulfide are common.Individual clay beds up to 6 meters (20 feet)thick are present in the unit. Clay and silt bedsmake up approximately one-third of the unit inthe study area. The dark, fine-grainedsediments represent lower delta plain,bay-dominated environments. Tan, light gray,yellow, brown, purple, and orange sand,pebbly sand, and clay represent upper deltaplain, channel-dominated environments.

In the southern part of the study area, dark,poorly sorted, micaceous, lignitic sand andsilty sand containing a diverse assemblage ofpollen and microfauna of early and middlePaleocene (Midwayan) age are present. Thisis the Perkins Bluff Member of the RhemsFormation, which was deposited in lower deltaplain or shallow marine shelf environments.

Toward the coast, the Rhems Formationincludes shallow to increasingly deeper waterclastic shelf facies sediments (Browns FerryMember) that ultimately pass into a shallowcarbonate platform facies at the SouthCarolina coastline. Colquhoun et al., referredto the carbonate platform facies equivalent as“unnamed limestone." The carbonate platformsequence is correlative with the anhydrite- andgypsum-bearing dolomitized limestone andfinely crystalline dolomite of the lower part of


A-20

the Cedar Keys Formation that is mapped incoastal Georgia and northeastern Florida. Thecarbonate sequence is about 76 meters (250feet) thick at the South Carolina coastline.The Cedar Keys Formation has a maximumthickness of 130 meters (425 feet) in coastalareas of Georgia. The carbonate platformsediments of the Cedar Keys Formation aregenerally impermeable, and the unit acts as theunderlying confining unit of the FloridanAquifer System in the coastal areas of SouthCarolina and Georgia.

Snapp Formation (Williamsburg Formation).Sediments in the study area that are timeequivalent to the Williamsburg Formationdiffer from the type Williamsburg and havebeen designated the "Snapp Member of theWilliamsburg Formation”. Fallaw and Pricehave suggested that the "Snapp Member" ofthe Williamsburg be raised to formationalstatus. The Snapp Formation is used in thisreport. The unit is encountered in well P-22in the southeastern part of SRS near SnappStation. The basal contact with the underlyingLang Syne/Sawdust Landing undifferentiatedis probably unconformable. The SnappFormation appears to pinch out in thenorthwestern part of SRS and thickens toabout 15 meters (50 feet) near the southeasternboundary of the site.

The Snapp Formation (WilliamsburgFormation) crops out in Calhoun County. Thesediments in the upper part of the unit consistof low-density, fissile, dark-gray to blacksiltstone and thin layers of black clayinterbedded with sand in the lower part. Theseand similar sediments in Aiken andOrangeburg Counties were probably depositedin lagoonal or estuarine environments. Withinand near SRS, the Snapp sediments typicallyare silty, medium- to coarse-grained quartzsand interbedded with clay. Dark, micaceous,lignitic sand also occurs, and all are suggestiveof lower delta plain environments. In Georgia,the unit consists of thinly laminated, silty claylocally containing layers of medium- todark-gray carbonaceous clay. This lithology isindicative of marginal marine (lagoonal toshallow shelf) depositional environments.

Clayey parts of the unit are characterized ongeophysical logs as zones of low electricalresistivity and a relatively high-gamma rayresponse. In the southernmost part of thestudy area, the Snapp consists of gray-green,fine to medium, well-rounded, calcareousquartz sand and interbedded micritic limestoneand limey clay that is highly fossiliferous andglauconitic. This lithology suggestsdeposition in shallow shelf environmentssomewhat removed from clastic sedimentsources

The upper surface of the WilliamsburgFormation is defined by the “LangSyne/Sawdust Landing” unconformity.

Fourmile Formation. Early Eocene ages,derived from paleontological assemblages,indicate that the sand immediately overlyingthe Snapp Formation in the study area isequivalent to the Fishburne. These sedimentswere deposited on the “Lang Syne/SawdustLanding” unconformity and constitute thebasel unit of Sequence Stratigraphic Unit II.The Fishburne is a calcareous unit that occursdowndip near the coast. The sand was initiallydesignated the Fourmile Member of theFishburne Formation. Owing to the distinctivedifference in lithology between the type,Fishburne Formation and the time-equivalentsediments observed in the study area, Fallawand Price have recommended that theFourmile Member of the Fishburne be raisedto formational rank. The term FourmileFormation is used in this report.

The Fourmile Formation averages 9 meters(30 feet) in thickness, is mostly tan,yellow-orange, brown, and white, moderatelyto well-sorted sand, with clay beds a few feetthick near the middle and at the top of the unit.The sand is very coarse to fine grained, withpebbly zones common, especially near thebase. Glauconite, up to about 5%, is presentin places, as is weathered feldspar. In thecenter and southeastern parts of SRS, the unitcan be distinguished from the underlyingPaleocene strata by its lighter color and lowercontent of silt and clay. Glauconite and


A-21

microfossil assemblages indicate that theFourmile is a shallow marine deposit.

Overlying the Fourmile Formation in the studyarea is 9 meters (30 feet) or less of sandsimilar to the Fourmile. This sand is bettersorted, contains fewer pebbly zones, lessmuscovite and glauconite, and in many wellsis lighter in color. Microfossil assemblagesindicate that the sand is correlative with theearly middle Eocene Congaree Formation. Insome wells a thin clay occurs at the top of theFourmile, separating the two units; however,the difficulty in distinguishing the FourmileFormation from the overlying CongareeFormation has led many workers at SRS toinclude the entire 293 meters (960 feet)section in the Congaree Formation.

Orangeburg Group

The Orangeburg Group consists of the lowermiddle Eocene Congaree Formation(Tallahatta equivalent) and the upper middleEocene Warley Hill Formation and SanteeLimestone (Lisbon equivalent). Over most ofthe study area, these post-Paleocene units aremore marine in character than the underlyingCretaceous and Paleocene units; they consistof alternating layers of sand, limestone, marl,and clay.

The group crops out at lower elevations inmany places within and near SRS. Thesediments thicken from about 26 meters (85feet) at well P-30 near the northwestern SRSboundary to 61 meters (200 feet) at well C-10in the south. Dip of the upper surface is 2m/km (12 ft/mile) to the southeast. Downdipat the coast, the Orangeburg Group is about 99meters (325 feet) thick and is composed ofshallow carbonate platform deposits of theSantee Limestone.

In the extreme northern part of the study area,the entire middle Eocene Orangeburg Group ismapped as the Huber Formation. Themicaceous, poorly sorted sand, abundantchannel fill deposits and cross bedding, andcarbonaceous kaolin clay in the Huber is

indicative of fluvial, upper delta plainenvironments.

In the central part of the study area the groupincludes, in ascending order, the Congaree,Warley Hill, and Tinker/Santee Formations.The units consist of alternating layers of sand,limestone, marl, and clay that are indicative ofdeposition in shoreline to shallow shelfenvironments. From the base upward, theOrangeburg Group passes from cleanshoreline sand characteristic of the CongareeFormation to shelf marl, clay, sand, andlimestone typical of the Warley Hill andSantee Limestone. Near the center of thestudy area, the Santee sediments consist of upto 30 vol% carbonate. The sequence istransgressive, with the middle Eocene Seareaching its most northerly position duringTinker/Santee deposition.

Toward the south, near wells P-21, ALL-324,and C-10, the carbonate content of all threeformations increases dramatically. Theshoreline sand of the Congaree undergoes afacies change to interbedded glauconitic sandand shale, grading to glauconitic argillaceous,fossiliferous, sandy limestone. Downdip, thefine-grained, glauconitic sand, and clay of theWarley Hill become increasingly calcareousand grades imperceptibly into carbonate-richfacies comparable to both the overlying andunderlying units. Carbonate content in theglauconitic marl, calcareous sand, and sandylimestone of the Santee increases towards thesouth. Carbonate sediments constitute the vastmajority of the Santee from well P-21southward.

Congaree Formation. The early middle EoceneCongaree Formation has been traced from theCongaree valley in east central South Carolinainto the study area. It has beenpaleontologically correlated with the early andmiddle Eocene Tallahatta Formation inneighboring southeastern Georgia by Fallaw etal.

The Congaree is about 9 meters (30 feet) thicknear the center of the study area and consistsof yellow, orange, tan, gray, green, and


A-22

greenish gray, well-sorted, fine to coarsequartz sand, with granule and small pebblezones common. Thin clay laminae occurthroughout the section. The quartz grains tendto be better rounded than those in the rest ofthe stratigraphic column are. The sand isglauconitic in places suggesting deposition inshoreline or shallow shelf environments. Tothe south, near well ALL-324, the CongareeFormation consists of interbedded glauconiticsand and shale, grading to glauconitic,argillaceous, fossiliferous sandy limestonesuggestive of shallow to deeper shelfenvironments of deposition. Farther south,beyond well C-10, the Congaree grades intoplatform carbonate facies of the lower SanteeLimestone.

The equivalent of the Congaree northwest ofSRS has been mapped as the HuberFormation. At these locations it becomesmore micaceous and poorly sorted, indicatingdeposition in fluvial and upper delta plainenvironments. On geophysical logs, theCongaree has a distinctive low gamma raycount and high electrical resistivity.

Warley Hill Formation. Unconformablyoverlying the Congaree Formation are 3meters (10 feet) to 6 meters (20 feet) offine-grained, often glauconitic sand and greenclay beds that have been referred torespectively as the Warley Hill and Caw CawMembers of the Santee Limestone. The greensand and clay beds are referred to informallyas the "green clay" in previous SRS reports.Both the glauconitic sand and the clay at thetop of the Congaree are assigned to the WarleyHill Formation. In the updip parts of the studyarea, the Warley Hill apparently is missing orvery thin, and the overlying Tinker/SanteeFormation rests unconformably on theCongaree Formation.

The Warley Hill sediments indicate shallow todeeper clastic shelf environments ofdeposition in the study area, representingdeeper water than the underlying CongareeFormation. This suggests a continuation of atransgressive pulse during upper middleEocene time. To the south, beyond well P-21,

the green silty sand, and clay of the WarleyHill undergo a facies change to the clayeymicritic limestone and limey clay typical ofthe overlying Santee Limestone. The WarleyHill blends imperceptibly into a thick clayeymicritic limestone that divides the FloridanAquifer System south of the study area. TheWarley Hill is correlative with the lower partof the Avon Park Limestone in southernGeorgia and the lower part of the LisbonFormation in western Georgia.

In the study area, the thickness of the WarleyHill Formation is generally less than 6 meters(20 feet). In a part of Bamberg County, SouthCarolina, the Congaree Formation is notpresent, and the Warley Hill rests directly onthe Williamsburg Formation.

Tinker/Santee Formation. The late middleEocene deposits overlying the Warley HillFormation consist of moderately sorted yellowand tan sand, calcareous sand and clay,limestone, and marl. Calcareous sedimentsdominate downdip, are sporadic in the middleof the study area, and are missing in thenorthwest. The limestone represents thefarthest advance to the northwest of thetransgressing carbonate platform firstdeveloped in early Paleocene time near theSouth Carolina and Georgia coasts.

Fallaw et al., divided the Santee into threemembers in the study area: the McBean, BlueBluff, and Tims Branch Members. TheMcBean Member consists of tan to white,calcilutite, calcarenite, shelly limestone, andcalcareous sand and clay. It dominates theSantee in the central part of the study area andrepresents the transitional lithologies betweenclastics in the north and northwest (TimsBranch Member), and fine-grained carbonatesin the south (Blue Bluff Member).

The carbonates and carbonate-rich clastics arerestricted essentially to three horizons in thecentral part of, the Griffins Landing Memberof the Dry Branch Formation, the McBeanMember of the Tinker/Santee Formation andthe Utley Limestone member of theClinchfield Formation. The uppermost horizon


A-23

includes the carbonates of the GriffinsLanding Member of the Dry BranchFormation found below the “tan clay” intervalthat occurs near the middle of the Dry Branch.The isolated carbonate patches of the GriffinsLanding are the oyster banks that formed inthe back barrier marsh zone behind the barrierisland system. Underlying the Dry Branch,directly below the regionally significantSantee Unconformity, is the Utley LimestoneMember of the Clinch Field Formation.Without the benefit of detailed petrographicand paleontological analysis, the Utleycarbonates cannot be systematicallydistinguished from the carbonates of theunderlying Tinker/Santee Formation. Thus thecarbonate-rich sediments between the SanteeUnconformity, and the Warley Hill Formationare referred to as the Tinker/Santee (Utley)sequence in this report.

The Blue Bluff Member consists of gray togreen, laminated micritic limestone. The unitincludes gray, fissile, calcareous clay andclayey micritic limestone and very thinlylayered to laminated, clayey, calcareous, silty,fine sand, with shells and hard, calcareousnodules, lenses, and layers. Cores of BlueBluff sediments are glauconitic, up to 30% inplaces. The Blue Bluff lithology suggestsdeposition in offshore shelf environments.Blue Bluff sediments tend to dominate theformation in the southern part of the studyarea and constitute the major part of the“middle confining unit" that separates theUpper and Lower Floridan aquifers south ofthe study area.

Fallaw et al., described the Tims BranchMember of the Santee as the siliciclastic partof the unit, consisting of fine- andmedium-grained, tan, orange, and yellow,poorly to well sorted, and slightly tomoderately indurated sand. The clasticlithologies of the Tims Branch Memberdominate the Santee in the northern part of thestudy area. Because the clastic lithologiesdiffer so markedly from the type Santee,Fallaw and Price raised the Tims BranchMember of the Santee to formational rank,namely the Tinker Formation. Because the

clastic and carbonate lithologies that constitutethe Tinker/Santee sequence in the upper andmiddle parts of the study area arehydrologically undifferentiated, the units arenot systematically separated, and they aredesignated Tinker/Santee Formation on mapsand sections. The thickness of theTinker/Santee Formation is variable due inpart to displacement of the sediments, butmore commonly to dissolution of thecarbonate resulting in consolidation of theinterval and slumping of the overlyingsediments of the Tobacco Road and DryBranch Formations into the resulting lows.

The Tinker/Santee (Utley) interval is about 21meters (70 feet) thick near the center of SRS,and the sediments indicate deposition inshallow marine environments. The top of theunit is picked on geophysical logs whereTinker/Santee (Utley) sediments with lowerelectrical resistivity are overlain by the moreresistive sediments of the Dry BranchFormation. In general, the gamma-ray countis higher than in surrounding stratigraphicunits.

Often found within the Tinker/Santee (Utley)sediments, particularly in the upper third ofthe interval, are weak zones interspersed instronger carbonate-rich matrix materials. Theweak zones, which vary in apparent thicknessand lateral extent, were noted where rod dropsand/or lost circulation occurred duringdrilling, low blow counts occurred during SPTpushes, etc. The weak zones have variouslybeen termed as “soft zones”, the “criticallayer”, “underconsolidated zones”, “badground”, and “void”. For this report, thepreferred term used to describe these zoneswill be “soft zones.”

Barnwell Group

Upper Eocene sediments of the BarnwellGroup represent the Upper Coastal Plain ofwestern South Carolina and eastern Georgia.Sediments of the Barnwell Group arechronostratigraphically equivalent to the lowerCooper Group (late Eocene) of Colquhoun etal. The Cooper Group includes sediments of


A-24

both late Eocene and early Oligocene age andappears downdip in the Lower Coastal Plain ofeastern South Carolina.

Sediments of the Barnwell Group overlie theTinker/Santee Formation and consist mostly ofshallow marine quartz sand containingsporadic clay layers. Huddleston and Hetrickrecently revised the upper Eocene stratigraphyof the Georgia Coastal Plain, and theirapproach has been extended into SouthCarolina by Nystrom and Nystrom andWilloughby. These authors elevated theEocene “Barnwell Formation” to the“Barnwell Group.” In Burke County, Georgia,the group includes (from oldest to youngest)the Clinchfield Formation, and Dry BranchFormation, and the Tobacco Road Formation.The group is about 21 meters (70 feet) thicknear the northwestern boundary of SRS and 52meters (170 feet) near its southeasternboundary. The regionally significant SanteeUnconformity that defines of boundarybetween Sequence Stratigtraphic units II andIII (Figure 1.4-18) separates the ClinchfieldFormation from the overlying Dry BranchFormation. The Santee Unconformity is apronounced erosional surface observablethroughout the SRS region.

In the northern part of the study area, theBarnwell Group consists of red or brown, fineto coarse-grained, well-sorted, massive sandyclay and clayey sand, calcareous sand andclay, as well as scattered thin layers ofsilicified fossiliferous limestone. All aresuggestive of lower delta plain and/or shallowshelf environments. Downdip, the Barnwellundergoes a facies change to the phosphaticclayey limestone that constitutes the lowerCooper Group. The lower Cooper Grouplimestone beds indicate deeper shelfenvironments.

Clinchfield Formation. The basal late EoceneClinchfield Formation consists of light coloredquartz sand and glauconitic, biomoldiclimestone, calcareous sand, and clay. Sandbeds of the formation constitute the RigginsMill Member of the Clinchfield Formation andare composed of medium to coarse, poorly to

well sorted, loose and slightly indurated, tan,clay, and green quartz. The sand is difficult toidentify unless it occurs between the overlyingcarbonate layers of the Griffins LandingMember and the underlying carbonate layersof the Santee Limestone. The Clinchfield isabout 8 meters (25 feet) thick in thesoutheastern part of SRS and pinches out orbecomes unrecognizable at the center of thesite.

The carbonate sequence of the ClinchfieldFormation is designated the Utley LimestoneMember. It is composed of sandy, glauconiticlimestone and calcareous sand, with anindurated, biomoldic facies developed inplaces. In cores, the sediments are tan andwhite and slightly to well indurated. Withoutthe benefit of detailed petrographic andpaleontological analysis, the Utley carbonatescannot be systematically distinguished fromthe carbonates of the underlying Tinker/SanteeFormation. Thus the carbonate-rich sedimentsbetween the Santee Unconformity, and theWarley Hill Formation are referred to as theTinker/Santee (Utley) sequence in this report.

Dry Branch Formation. The late Eocene DryBranch Formation is divided into the IrwintonSand Member, the Twiggs Clay Member, andthe Griffins Landing Member. The unit isabout 18 meters (60 feet) thick near the centerof the study area. The top of the Dry Branchis picked on geophysical logs where a lowgamma-ray count in the relatively clean DryBranch sand increases sharply in the moreargillaceous sediments of the overlyingTobacco Road Sand.

The Dry Branch sediments overlying theTinker/Santee (Utley) interval in the centralportion of SRS were deposited inshoreline/lagoonal/tidal marsh environments.The shoreline retreated from its position innorthern SRS during Tinker/Santee (Utley)time to the central part of SRS in Dry Branchtime. Progradation of the shorelineenvironments to the south resulted in the sandsand muddy sands of the Dry Branch beingdeposited over the shelf carbonates andclastics of the Tinker/Santee (Utley) sequence.


A-25

The Twiggs Clay Member does not seem to bemappable in the study area. Lithologicallysimilar clay is present at various stratigraphiclevels in the Dry Branch Formation. The tan,light-gray, and brown clay is as thick as 4meters (12 feet) in SRS wells but is notcontinuous over long distances. This has beenreferred to in the past as the “tan clay” in SRSreports. The Twiggs Clay Member, thatpredominates west of the Ocmulgee River inGeorgia, is not observed as a separate unit inthe study area.

The Griffins Landing Member is composedmostly of tan or green, slightly to wellindurated, quartzose calcareous micrite andsparite, calcareous quartz sand and slightlycalcareous clay. Oyster beds are common inthe sparry carbonate facies. The unit seems tobe widespread in the southeastern part of SRS,where it is about 15 meters (50 feet) thick, butbecomes sporadic in the center and pinchesout. Carbonate content is highly variable. Inplaces, the unit lies unconformably on theUtley Limestone Member, which containsmuch more indurated, moldic limestone. Inother areas, it lies on the noncalcareous quartzsand of the Clinchfield. Updip, the underlyingClinchfield is difficult to identify or ismissing, and the unit may lie unconformablyon the sand and clay facies of theTinker/Santee Formation. The GriffinsLanding Member appears to have formed inlagoonal/marsh environments.

The Irwinton Sand Member is composed oftan, yellow and orange, moderately sortedquartz sand, with interlaminated andinterbedded clay abundant in places. Pebblylayers are present, as are clay clast-rich zones(Twiggs Clay lithology). Clay beds, which arenot continuous over long distances, are tan,light gray, and brown in color, and can beseveral feet thick in places. These are the “tanclay” beds of various SRS reports. IrwintonSand beds have the characteristics of shorelineto shallow marine sediments. The IrwintonSand crops out in SRS. Thickness is variable,but is about 12 meters (40 feet) near the

northwestern site boundary and 21 meters (70feet) near the southeastern boundary.

Tobacco Road Formation. The Late EoceneTobacco Road Formation consists ofmoderately to poorly sorted, red, brown, tan,purple, and orange, fine to coarse, clayeyquartz sand. Pebble layers are common, as areclay laminae and beds. Ophiomorpha burrowsare abundant in parts of the formation.Sediments have the characteristics of lowerDelta plain to shallow marine deposits. Thetop of the Tobacco Road is characterized bythe change from a comparatively well-sortedsand to the more poorly sorted sand, pebblysand, and clay of the “Upland unit.” Contactbetween the units constitutes the “Upland”unconformity. The unconformity is veryirregular due to fluvial incision thataccompanied deposition of the overlying"Upland unit" and later erosion. As statedpreviously, the lower part of the CooperGroup (upper Eocene) is the probabledowndip equivalent of the Tobacco RoadFormation.

“Upland Unit”/Hawthorn/Chandler BridgeFormations. Deposits of poorly sorted silty,clayey sand, pebbly sand, and conglomerate ofthe “Upland unit” cap many of the hills athigher elevations over much of the study area.Weathered feldspar is abundant in places. Thecolor is variable, and facies changes areabrupt. Siple assigned these sediments to theHawthorn Formation. Nystrom et al.., whomapped it as the "Upland unit", discussevidence for a Miocene age. The unit is up to18 meters (60 feet) thick. The environment ofdeposition appears to be fluvial, and thethickness changes abruptly owing tochanneling of the underlying Tobacco RoadFormation during “Upland” deposition andsubsequent erosion of the “Upland” unit itself.This erosion formed the “Upland”unconformity. The unit is up to 18 meters (60feet) thick.

Lithologic types comparable to the “Upland”unit but assigned to the Hawthorn Formationoverlie the Barnwell Group and the CooperGroup in the southern part of the study area.


A-26

In this area, the Hawthorn Formation consistsof very poorly sorted, sandy clay, and clayeysand, with lenses of gravel and thin beds ofsand very similar to the "Upland unit".Farther downdip, the Hawthorn overlies theequivalent of the Suwanee Limestone and actsas the confining layer overlying the FloridanAquifer System. It consists of phosphatic,sandy clay and phosphatic, clayey sand andsandy, dolomitic limestone interbedded withlayers of hard, brittle clay resembling stratifiedfuller's earth.

Colquhoun et al., suggest that the “Uplandunit”, Tobacco Road Formation, and DryBranch Formation are similar in granularityand composition, indicating that they might besimilar genetically, that is that they are part ofthe same transgressive/regressive depositionalcycle. The "Upland unit" represents the mostcontinental end member (lithofacies) and theDry Branch Formation represents the mostmarine end member. Thus, the “Upland unit”is the result of a major regressive pulse thatclosed out deposition of the BarnwellGroup/Cooper Group depositional cycle.Colquhoun et al., suggested that the "Uplandunit" is correlative with the Chandler BridgeFormation downdip toward the coast. Thishypothesis is significant because it impliesthat there was no major hiatus between the"Upland unit" and the underlying TobaccoRoad and Dry Branch Formations. Theexistence of a hiatus between the units hasbeen reported by numerous studies of theSouth Carolina Coastal Plain.

Quaternary Surfaces and Deposits

SRS lies within the interfluve area between theSavannah and the Salkahatchie Rivers. Thedrainage systems within the site consistentirely of streams that are tributary to theSavannah River. A series of nested fluvialterraces are preserved along the river andmajor tributaries. Fluvial terraces are theprimary geomorphic surface that can be usedto evaluate Quaternary deformation withinSRS. However, there is limited data availablefor the estimation of ages of river terraces inboth the Atlantic and Gulf Coastal Plains.

Major stream terraces form by sequentialerosional and depositional events in responseto tectonism, isostasy, and climate variation.Streams respond to uplift by cutting down intothe underlying substrate in order to achieve asmooth longitudinal profile that grades to theregional base level. Aggradation or depositionoccurs when down-cutting is reversed by a risein base level. The stream channel is elevatedand isolated from the underlying marine strataby layers of newly deposited fluvialsediments. Down-cutting may resume and theaggraded surface is abandoned. The result is alandform referred to as a fill terrace.

At the SRS there are two prominent terracesabove the modern floodplain (Qal). Thesedesignations are based on morphology andrelative height above local base level. Localbase level is the present elevation of theSavannah River channel. In addition, there areother minor terraces: one lower and severalhigher, older terrace remnants.

Carbonate and Soft Zones

Often found within the Tinker/Santee (Utley)sediments, particularly in the upper third ofthis section, are weak zones interspersed instronger carbonate-rich matrix materials.These weak zones, which vary in apparentthickness and lateral extent, were recordedwhere rod drops and/or lost circulationoccurred during drilling, low blow countsoccurred during SPT pushes, etc. They havevariously been termed as “soft zones”, “thecritical layer”, “underconsolidated zones”,“bad ground”, and “void”. The preferred termused to describe these zones is “soft zones”.Aadland et al. discuss these features in theirpaper in this guide book.

Deeper Sediments

Beneath the Coastal Plain sedimentarysequence and below a pre-Cretaceousunconformity are two geologic terranes: (1)the Dunbarton basin, a Triassic-Jurassic Riftbasin, filled with lithified terrigenous and


A-27

lacustrine sediments with possible minoramounts of mafic volcanic and intrusive rock;and (2) a crystalline terrane of metamorphosedsedimentary and igneous rock that may rangein age from Precambrian to late Paleozoic(reference Figure 6). The Paleozoic rocks andthe Triassic sediments were leveled byerosion, forming the base for Coastal Plainsediment deposition.

The lithologies and structures in the crystallineterrane are basically similar to that seen in theeastern Piedmont province as exposed in otherparts of the southeastern United States. Thecrystalline rocks form a volcanic – intrusivesequence of calc-alkaline composition,portions of which record both ductile andbrittle deformational events. Theserelationships indicate that these rocks are themetamorphosed and deformed remnants of anancient volcanic arc that are interpreted to beCarolina Terrane equivalents.

Ten core borings into the basin penetratedsedimentary rocks. Fanglomerate, sandstone,siltstone, and mudstone are the dominantlithologies. These rocks are similar to theclastic facies in other Newark Supergroupbasins

HYDROSTRATIGRAPHYINTRODUCTION

Two hydrogeologic provinces are recognizedin the subsurface beneath the SRS region(reference Figures 6 and7). The uppermostprovince, which consists of the wedge ofunconsolidated Coastal Plain sediments ofLate Cretaceous and Tertiary ages, is referredto as the Southeastern Coastal Plainhydrogeologic province. It is furthersubdivided into aquifer or confining systems,units, and zones. The underlying province,referred to as the Piedmont hydrogeologicprovince, includes Paleozoic metamorphic andigneous basement rocks and Upper Triassiclithified mudstone, sandstone, andconglomerate in the Dunbarton basin.

The following hydrogeological characteristicsare of particular interest:

• The layered structure of the coastalplain sediments effectively controlsmigration of contaminants in thesubsurface, limiting vertical migrationto deeper aquifers.

• Between the ground surface and theprimary drinking water aquifer(s) areseveral low permeability zones whichrestrict vertical migration from a givenpoint source.

• The abundance of clay size materialand clay minerals in the aquifer andaquitard zones affects groundwatercomposition and vertical migration.The concentration of some potentialcontaminants, especially metals andradionuclides, may be attenuated byexchange and fixation of dissolvedconstituents on clay surfaces.

• The recharge area(s) for the deeperdrinking water aquifers used are updipof SRS, near the fall line. Somerecharge areas are located at thenorthern-most fringe of the site.

• Recharge for the water-table aquifers,namely the Upper Three Runs andGordon Aquifers, is primarily fromlocal precipitation.

• Discharge of groundwater from theUpper Three Runs and Gordonaquifers is typically to the localstreams on the SRS.

• Groundwater at the SRS is typically oflow ionic, low dissolved solids andmoderate pH (typically ranging from4.4 to 6.0). Other constituents such asdissolved oxygen and alkalinity aremore closely associated with rechargeand aquifer material. Dissolvedoxygen is typically higher in the updipand near-surface recharge areas, andalkalinity, pH and dissolved solids aretypically higher in those portions ofthe aquifers regions containingsignificant carbonate materials.


A-28

• The presence of an upward verticalgradient or “head reversal” betweenthe Upper Three Runs and Gordonaquifers and the Crouch Branchaquifer is significant in that it preventsdownward vertical migration ofcontaminants into deeper aquifersover much of central SRS.

Hydrogeology

The method for establishing a nomenclaturefor the hydrogeologic units in the followingdiscussion is based on Aadland et al., andgenerally follows the guidelines set forth bythe South Carolina HydrostratigraphicSubcommittee. Hydrogeologic units are shownon Figure 2.

A hydrogeologic unit is defined by itshydraulic properties (hydraulic conductivity,hydraulic head relationships, porosity,leakance coefficients, vertical flow velocityand transmissivity) relative to those propertiesmeasured in the overlying and underlyingunits. The properties are measured at a typewell or type well cluster location. Aquifer andconfining units are mapped on the basis of thehydrogeologic continuity, potentiometricconditions, and leakance-coefficient estimatesfor the units. These properties are largelydependent on the thickness, areal distribution,and continuity of the lithology of the particularunit. However, a hydrogeologic unit maytraverse lithologic unit boundaries if there isnot a significant change in hydrogeologicproperties corresponding to the change inlithology.

Upper Three Runs Aquifer Unit

The Upper Three Runs Aquifer Unit occursbetween the water-table and the GordonConfining Unit and includes all strata abovethe Warley Hill Formation (in updip areas)and the Blue Bluff Member of the SanteeLimestone. It includes the sandy andsometimes calcareous sediments of theTinker/Santee Formation and all theheterogeneous sediments in the overlying

Barnwell Group. The Upper Three Runsaquifer is the updip clastic facies equivalent ofthe Upper Floridan aquifer in the carbonatephase of the Floridan Aquifer System.

The Upper Three Runs aquifer is defined bythe hydrogeologic properties of the sedimentspenetrated in well P-27 located near UpperThree Runs Creek in the center of SRS. Here,the aquifer is 40 meters (132 feet) thick andconsists mainly of quartz sand and clayey sandof the Tinker/Santee Formation; sand withinterbedded tan to gray clay of the Dry BranchFormation; and sand, pebbly sand, and minorclay beds of the Tobacco Road Formation.Calcareous sand, clay, and limestone, althoughnot observed in the P-27 well, are present inthe Tinker/Santee Formation throughout theGeneral Separations Area near well P-27.Downdip, at the C-10 reference well, theUpper Three Runs aquifer is 116 meters (380feet) thick and consists of clayey sand andsand of the upper Cooper Group; sandy, shellylimestone, and calcareous sand of the lowerCooper Group/Barnwell Group; and sandy,shelly, limestone and micritic limestone of theSantee Limestone.

Water-level data are sparse for the UpperThree Runs aquifer unit except within SRS.The hydraulic-head distribution of the aquiferis controlled by the location and depth ofincisement of creeks that dissect the area. Theincisement of these streams and theirtributaries has divided the interstream areas ofthe water-table aquifer into "groundwaterislands." Each "groundwater island" behavesas an independent hydrogeologic subset of thewater-table aquifer with unique recharge anddischarge areas. The stream acts as thegroundwater discharge boundary for theinterstream area. The head distribution patternin these groundwater islands tends to followtopography and is characterized by higherheads in the interstream area with graduallydeclining heads toward the bounding streams.Groundwater divides are present near thecenter of the interstream areas. Water-tableelevations reach a maximum of 76 meters (250feet) msl in the northwest corner of the study


A-29

area and decline to approximately 30 meters(100 feet) msl near the Savannah River.

Porosity and permeability of the Upper ThreeRuns aquifer are variable across the studyarea. In the northern and central regions, theaquifer yields only small quantities of water,owing to the presence of interstitial silt andclay and poorly sorted sediments that combineto significantly reduce permeability. Locallenses of relatively clean, permeable sandhowever, may, yield sufficient quantities fordomestic use. Such high-permeability zoneshave been observed in the GeneralSeparations Area near the center of the studyarea and may locally influence the movementof groundwater .

Porosity and permeability were determined forUpper Three Runs aquifer sand samplescontaining less than 25 percent mud, using theBeard and Weyl method. Porosity averages35.3 percent; the distribution is approximatelynormal, but skewed slightly toward highervalues. Geometric mean permeability is 31.5Darcies (23 m/day [76.7 ft/day]) with about 60percent of the values between 16 and 64Darcies (12 and 48 m/day [39 and 156ft/day]).

Recharge to the Upper Three Runs aquiferoccurs at the water-table by infiltrationdownward from the land surface. In the“upper" aquifer zone, part of this groundwatermoves laterally toward the bounding streamswhile part moves vertically downward. Thegenerally low vertical hydraulic conductivitiesof the “upper" aquifer zone and theintermittent occurrence of the “tan clay"confining zone retard the downward flow ofwater, producing vertical hydraulic-headgradients in the "upper" aquifer zone andacross the "tan clay” confining zone.

Downward hydraulic-head differences in the“upper” aquifer zone vary from 1.4 to 1.64meters (4.5 to 5.4 feet), and differences acrossthe “tan clay” are as much as 6.5 meters (15.8feet) in H Area. At other locations in theGeneral Separations Area, the head differenceacross the “tan clay” confining zone is only

0.3 to 1 meter (0.1 to 3.2 feet), essentiallywhat might be expected due simply to lowvertical flow in a clayey sand aquifer.Therefore, the ability of the "tan clay”confining zone to impede water flow variesgreatly over the General Separations Area.

Groundwater leaking downward across the"tan clay” confining zone recharges the"lower" aquifer zone of the Upper Three Runsaquifer. Most of this water moves laterallytoward the bounding streams; the remainderflows vertically downward across the Gordonconfining unit into the Gordon aquifer. Allgroundwater moving toward Upper ThreeRuns Creek leaks through the Gordonconfining unit or enters small streams.Vertical hydraulic-head differences in the"lower" aquifer zone range from 0.5 to 1 meter(1.5 to 3.2 feet) in H Area and indicate somevertical resistance to flow.

Gordon Confining Unit

Clayey sand and clay of the Warley HillFormation and clayey, micritic limestone ofthe Blue Bluff Member of the SanteeLimestone constitute the Gordon confiningunit. The Gordon confining unit separates theGordon aquifer from the overlying UpperThree Runs aquifer. The unit has beeninformally termed the "green clay" in previousSRS reports.

In the study area, the thickness of the Gordonconfining unit ranges from about 1.5 to 26meters (5 to 85 feet). The unit thickens to thesoutheast. From Upper Three Runs Creek tothe vicinity of L Lake and Par Pond, theconfining unit generally consists of one ormore thin clay beds, sandy mud beds, andsandy clay beds intercalated with subordinatelayers and lenses of quartz sand, gravelly sand,gravelly muddy sand, and calcareous mud.Southward from L Lake and Par Pond,however, the unit undergoes a stratigraphicfacies change to clayey micritic limestone andlimey clay typical of the Blue Bluff Member.The fine-grained carbonates andcarbonate-rich muds constitute the farthestupdip extent of the "middle confining unit" of


A-30

the Floridan Aquifer System (thehydrostratigraphic equivalent of the Gordonconfining unit), which dominates in coastalareas of South Carolina and Georgia.

North of the updip limit of the Gordonconfining unit, the fine-grained clastics of theWarley Hill Formation are thin, intermittent,and no longer effective in regionallyseparating groundwater flow. Here, the SteedPond aquifer is defined. Although thin andintermittent, the clay, sandy clay, and clayeysand beds of the Warley Hill Formation can besignificant at the site-specific level and oftendivide the Steed Pond aquifer into aquiferzones.

Gordon Aquifer Unit

The Gordon aquifer consists of all thesaturated strata that occur between the Gordonconfining unit and the Crouch Branchconfining unit in both the Floridan - MidvilleAquifer System and the Meyers Branchconfining system. The aquifer issemiconfined, with a downward potential fromthe overlying Upper Three Runs aquiferobserved in interfluvial areas, and an upwardpotential observed along the tributaries of theSavannah River where the Upper Three Runsaquifer is incised. The thickness of theGordon aquifer ranges from 12 meters (38feet) at well P-4A to 56 meters (185 feet) atwell C-6 and generally thickens to the east andsoutheast. Thickness variations in theconfining lithologies near the Pen BranchFault suggest depositional effects owing tomovements on the fault in early Eocene time.The Gordon aquifer is partially eroded nearthe Savannah River and Upper Three RunsCreek. The regional potentiometric map ofthe Gordon aquifer indicates that majordeviations in the flow direction are presentwhere the aquifer is deeply incised by streamsthat drain water from the aquifers.

The Gordon aquifer is characterized by thehydraulic properties of the sedimentspenetrated in reference well P-27 located nearthe center of SRS. The unit is 23 meters (75.5feet) thick in well P-27 and occurs from 38 to

15 meters (125 to 48 feet) msl. The aquiferconsists of the sandy parts of the SnappFormation and the overlying Fourmile andCongaree Formations. Clay beds and stringersare present in the aquifer, but they are too thinand discontinuous to be more than localconfining beds. The aquifer in wells P-21 andP-22 includes a clay bed that separates theCongaree and Fourmile Formations. The claybed appears sufficiently thick and continuousto justify splitting the Gordon aquifer intozones in the southeastern quadrant of SRS.

Downdip, the quartz sand of the Gordonaquifer grades into quartz-rich, fossiliferouslime grainstone, packstone, and wackestone,which contain considerably more glauconitethan the updip equivalents. Porosity of thelimestone as measured in thin-section rangesfrom 5 to 30 percent and is mostly moldic andvuggy.

South of SRS, near well ALL-324, the Gordonaquifer consists of interbedded glauconiticsand and shale, grading to sandy limestone.Farther south, beyond well C-10, the aquifergrades into platform limestone of the LowerFloridan aquifer of the carbonate phase of theFloridan Aquifer System.

The Gordon aquifer is recharged directly byprecipitation in the outcrop area and ininterstream drainage divides in and near theoutcrop area. South of the outcrop area, theGordon is recharged by leakage fromoverlying and underlying aquifers. Becausestreams such as the Savannah River and UpperThree Runs Creek cut through the aquifers ofthe Floridan Aquifer System, they representno-flow boundaries. As such, wateravailability or flow patterns on one side of theboundary (stream) will not change appreciablydue to water on the other side. In the centralpart of SRS, where the Gordon confining unitis breached by faulting, recharge to theGordon aquifer is locally increased.

Most of the Gordon aquifer is under confinedconditions, except along the fringes of UpperThree Runs Creek (i.e., near the updip limit ofthe Gordon Confining Unit) and the Savannah


A-31

River. The potentiometric-surface map of theaquifer shows that the natural discharge areasof the Gordon aquifer at SRS are the swampsand marshes along Upper Three Runs Creekand the Savannah River. These streamsdissect the Floridan Aquifer System, resultingin unconfined conditions in the stream valleysand probably in semiconfined (leaky)conditions near the valley walls. Reducedhead near Upper Three Runs Creek inducesupward flow from the Crouch Branch aquiferand develops the "head reversal" that is animportant aspect of the SRS hydrogeologicalsystem. The northeast-southwest orientedhydraulic gradient across SRS is consistentand averages 0.9 m/km (4.8 ft/mi). Thenortheastward deflection of the contours alongthe Upper Three Runs Creek indicatesincisement of the sediments that constitute theaquifer by the creek.

Hydraulic characteristics of the Gordon areless variable than those noted in the UpperThree Runs aquifer. Hydraulic conductivitydecreases downdip near well C-10 owing topoor sorting, finer grain size, and an increasein clay content.

Steed Pond Aquifer Unit

North of Upper Three Runs Creek where theFloridan - Midville Aquifer System is defined,the permeable beds that correspond to theGordon and Upper Three Runs aquifers of theFloridan Aquifer System are only locallyseparated, owing to the thin and intermittentcharacter of the intervening clay beds of theGordon confining unit (Warley HillFormation) and to erosion by the local streamsystems that dissect the interval. Here, theaquifers coalesce to form the Steed Pondaquifer of the Floridan-Midville AquiferSystem.

The Steed Pond aquifer is defined byhydrogeologic characteristics of sedimentspenetrated in well MSB-42 located inA/M Area in the northwest corner of SRS.The aquifer is 29.6 meters (97 feet) thick.Permeable beds consist mainly of subangular,coarse- and medium-grained, slightly gravelly,

submature quartz sand and clayey sand.Locally, the Steed Pond aquifer can be dividedinto zones. In A/M Area three zones aredelineated, the "Lost Lake" zone, and theoverlying "M Area" aquifer zones, separatedby clay and clayey sand beds of the "greenclay" confining zone.

In A/M Area, water enters the subsurfacethrough precipitation, and recharge into the"M-Area" aquifer zone occurs at thewater-table by infiltration downward from theland surface. A groundwater divide exists inthe A/M Area in which lateral groundwaterflow is to the southeast towards Tims Branchand southwest towards Upper Three RunsCreek and the Savannah River floodplain.Groundwater also migrates downward andleaks through the "green clay" confining zoneinto the "Lost Lake" aquifer. The "green clay"confining zone that underlies the "M-Area"aquifer zone is correlative with the Gordonconfining unit south of Upper Three RunsCreek.

Meyers Branch Confining System. TheMeyers Branch confining system separates theFloridan Aquifer System from the underlyingDublin and Dublin-Midville Aquifer Systems.North of the updip limit of the confiningsystem, the Floridan and Dublin-MidvilleAquifer Systems are in hydrauliccommunication and the aquifer systemscoalesce to form the Floridan-MidvilleAquifer System.

Sediments of the Meyers Branch confiningsystem correspond to clay and interbeddedsand of the uppermost Steel Creek Formation,and to clay and laminated shale of the SawdustLanding/Lang Syne and Snapp Formations. Inthe northwestern part of the study area, thesediments that form the Meyers Branchconfining system are better sorted and lesssilty, with thinner clay interbeds. This is theupdip limit of the Meyers Branch confiningsystem .


A-32

Crouch Branch Confining Unit

In the SRS area, the Meyers Branch confiningsystem consists of a single hydrostratigraphicunit, the Crouch Branch confining unit, whichincludes several thick and relativelycontinuous (over several kilometers) claybeds. The Crouch Branch confining unitextends north of the updip limit of the MeyersBranch confining system where the clay thinsand is locally absent and faulting observed inthe region locally breaches the unit. Here, theCrouch Branch confining unit separates theSteed Pond aquifer unit from the underlyingCrouch Branch aquifer unit. Downdip,generally south of the study area, the MeyersBranch confining system could be furthersubdivided into aquifer and confining units ifthis should prove useful for hydrogeologiccharacterization.

As indicated earlier, a hydraulic-headdifference persists across the Crouch Branchconfining unit near SRS. Owing to deepincisement by the Savannah River and UpperThree Runs Creek into the sediments of theoverlying Gordon aquifer, an upwardhydraulic gradient (vertical-head reversal)persists across the Crouch Branch confiningunit over a large area adjacent to the SavannahRiver floodplain and the Upper Three RunsCreek drainage system. This "head reversal"is an important aspect of the groundwater flowsystem near SRS and provides a natural meansof protection from contamination of the loweraquifers.

The total thickness of the Crouch Branchconfining unit where it constitutes the MeyersBranch confining system ranges from 17.4 to56.1 meters (57 to 184 feet). Updip, thethickness of the Crouch Branch confining unitranges from < 1 to 31.7 meters (3.3 to 104feet). The confining unit dips approximately 5m/km (16 ft/mi) to the southeast. Theconfining unit is comprised of the “upper” and“lower” confining zones, which are separatedby a “middle sand” zone.

In general, the Crouch Branch confining unitcontains two to seven clay to sandy clay beds

separated by clayey sand and sand beds thatare relatively continuous over distances ofseveral kilometers. The clay beds in theconfining unit are anomalously thin and fewerin number along a line that parallels thesouthwest-northeast trend of the Pen Branchand Steel Creek Faults and the northeastsouthwest trending Crackerneck Fault. Thereduced clay content near the faults suggestsshoaling due to uplift along the faults duringdeposition of the Paleocene Black MingoGroup sediments.

In A/M Area, the Crouch Branch confiningunit can often be divided into three zones: an“upper clay” confining zone is separated fromthe underlying “lower clay” confining zone bythe “middle sand” aquifer zone. The "middlesand" aquifer zone consists of very poorlysorted sand and clayey silt of the LangSyne/Sawdust Landing Formations. The"middle sand" aquifer zone has a flowdirection that is predominantlysouth/southwest toward Upper Three RunsCreek.

In places, especially in the northern part ofA/M Area, "upper clay" confining zone is verythin or absent. Here, only the “lower clay”confining zone is capable of acting as aconfining unit and the "middle sand" zone isconsidered part of the Steed Pond aquifer.Similarly, when the clay beds of the "lowerclay" confining zone are very thin or absent,the "middle sand" aquifer zone is consideredpart of the Crouch Branch aquifer unit. This isthe case in the far northeastern part of thestudy area.

The "lower clay" confining zone has beenreferred to as the lower Ellenton clay, theEllenton clay, the Peedee clay, and theEllenton/Peedee clay in previous SRS reports.It consists of the massive clay bed that capsthe Steel Creek Formation. The zone isvariable in total thickness and, based on 31wells that penetrate the unit, ranges from 1.5to 19 meters (5 to 62 feet) and averages 7.3meters (24 feet) thick.


A-33

Dublin Aquifer System. The Dublin AquiferSystem is present in the southeastern half ofSRS and consists of one aquifer, the CrouchBranch aquifer. It is underlain by theAllendale confining system and overlain bythe Meyers Branch confining system. Theupdip limit of the Dublin Aquifer System inthe study area corresponds to the updip limitof the Allendale confining system. North ofthis line, the Dublin-Midville Aquifer Systemis defined.

The thickness of the Dublin Aquifer Systemgenerally increases toward the south andranges from approximately 53 to 88 meters(175 to 290 feet). The top of the unit dips 3.79m/km (20 ft/mi) to the southeast. The unitthins to the east toward the Salkehatchie Riverand to the west toward Georgia. Near theupdip limit of the system, thicknesses arevariable and probably reflect the effects ofmovement along the Pen Branch Fault duringdeposition of the middle Black Creek clay.

The Dublin Aquifer System was defined andnamed by Clarke et al., for sedimentspenetrated by well 21-U4 drilled near the townof Dublin in Laurens County, Georgia. Theupper part of the Dublin Aquifer Systemconsists of fine to coarse sand and limestoneof the lower Huber-Ellenton unit. Comparablestratigraphic units serve as confining beds inthe SRS area and are considered part of theoverlying Meyers Branch confining system.Clarke et al.., noted that to the east near theSavannah River, clay in the upper part of thelower Huber-Ellenton unit forms a confiningunit that separates an upper aquifer ofPaleocene age from a lower aquifer of LateCretaceous age. The upper aquifer of Clarkeet al., is the Gordon aquifer as defined in thestudy, and their confining unit constitutes theMeyers Branch confining system of the SRSregion. The lower part of the Dublin AquiferSystem consists of alternating layers of clayeysand and clay of the Peedee-Providence unit.

Sediments typical of the Dublin AquiferSystem are penetrated in the reference wellP-22. The system consists of the well-sortedsand and clayey sand of the Black Creek

Formation and the moderately sorted sand andinterbedded sand and clay of the Steel CreekFormation. The aquifer is overlain by the claybeds that cap the Steel Creek Formation.These clay beds constitute the base of theMeyers Branch confining system.

The Dublin Aquifer System is 65 meters (213ft) thick in well P-22; the top is at an elevationof -68 meters (-223 feet) msl and the bottom at-133 meters (-436 feet) msl. The Dublinincludes five clay beds in this well.

In the southern part of the study area andfarther south and east, the Dublin shows muchlower values for hydraulic conductivity andtransmissivity, probably due to the increase offine-grained sediments toward the coast.

Dublin - Midville Aquifer System. TheDublin-Midville Aquifer System underlies thecentral part of SRS. The system includes allthe sediments in the Cretaceous LumbeeGroup from the Middendorf Formation up tothe sand beds in the lower part of the SteelCreek Formation. The system is overlain bythe Meyers Branch confining system andunderlain by the indurated clayey silty sandand silty clay of the Appleton confiningsystem. The updip limit of the system isestablished at the updip pinchout of theoverlying Meyers Branch confining system(see Figure 1.4-22). The downdip limit of theDublin-Midville is where the Allendalebecomes an effective confining system (seeFigure 1.4-22). The Dublin-Midville and theupdip Floridan-Midville Aquifer Systemswere referred to as the Tuscaloosa aquifer bySiple.

The thickness of the Dublin-Midville AquiferSystem ranges from approximately 76 to 168meters (250 to 550 feet). The dip of the uppersurface of the system is about 3.8 m/km (20ft/mi) to the southeast. Near the downdip limitof the system, thicknesses are variable andprobably reflect the effects of movement alongthe Pen Branch Fault. Shoaling along the faulttrace resulted in a relative increase in the


A-34

thickness of the aquifers at the expense of theintervening confining unit.

The Dublin-Midville Aquifer System includestwo aquifer units, the McQueen Branchaquifer, and the Crouch Branch aquifer,separated by the McQueen Branch confiningunit. The two aquifers can be tracednorthward, where they continue to be anintegral part of the Floridan-Midville AquiferSystem and southward where they constitutethe aquifer units of the Midville and DublinAquifer Systems, respectively.

The Dublin-Midville Aquifer System isdefined at the type well P-27. Here, thesystem is 153 meters (505 feet) thick andoccurs from -25 meters (-82 feet) msl to -179meters (-587 feet) msl. It consists of medium-to very coarse-grained, silty sand of theMiddendorf Formation and clayey, fine tomedium sand and silty clay beds of the BlackCreek Formation (Ref. 118). The systemincludes a thick clay bed, occurring from -100meters (-329 feet) msl to -117 meters (-384feet) msl, which constitutes the McQueenBranch confining unit.

A regional potentiometric surface mapprepared by Siple for his "Tuscaloosa aquifer,"indicates that the Savannah River hasbreached the Cretaceous sediments and is aregional discharge area for theFloridan-Midville Aquifer System, theDublin-Midville Aquifer System, and theupdip part of both the Dublin and MidvilleAquifer Systems. The Savannah River,therefore, represents a no-flow boundarypreventing the groundwater in these aquifersystems from flowing southward into Georgia.

Crouch Branch Aquifer Unit

The Crouch Branch aquifer constitutes theDublin Aquifer System in the southern part ofthe study area. Farther south, the Dublin canbe subdivided into several aquifers andconfining units. In the central part of thestudy area, the Crouch Branch aquifer is theuppermost of the two aquifers that constitutethe Dublin-Midville Aquifer System. Farther

north in the northwestern part of SRS andnorth of the site, the Crouch Branch aquifer isthe middle aquifer of the three aquifers thatconstitute the Floridan-Midville AquiferSystem.

The Crouch Branch aquifer is overlain by theCrouch Branch confining unit and is underlainby the McQueen Branch confining unit. Itpersists throughout the northern part of thestudy area, but near the updip limit of theCoastal Plain sedimentary clastic wedge, theCrouch Branch confining unit ceases to beeffective and the Crouch Branch aquifercoalesces with the Steed Pond aquifer.

The Crouch Branch aquifer ranges inthickness from about 30 to 107 meters (100 to350 feet). Thickness of the unit is variablenear the updip limit of the Dublin AquiferSystem where sedimentation was affected bymovement along the Pen Branch Fault. Thereduced clay content in this vicinity suggestsshoaling due to uplift along the fault duringLate Cretaceous and Paleocene time, resultingin the deposition of increased quantities ofshallow-water, coarse-grained clastics alongthe crest of the fault trace. The sandy beds acthydrogeologically as part of the CrouchBranch aquifer, resulting in fewer and thinner,less persistent clay beds in the overlying andunderlying confining units.

The Crouch Branch aquifer thins dramaticallyin the eastern part of the study area at the samegeneral location where the underlyingMcQueen Branch confining unit and theoverlying Crouch Branch confining unitthicken at the expense of Crouch Branch sand.Clay beds in the Crouch Branch aquifergenerally thicken in the same area andconstitute as much as 33 percent of the unit atthe well C-6.

Sediments of the Crouch Branch aquifer arechiefly sand, muddy sand, and slightlygravelly sand intercalated with thin,discontinuous layers of sandy clay and sandymud. Hydraulic conductivity of the CrouchBranch aquifer, determined from elevenpumping tests by Siple and from analyses


A-35

made by GeoTrans, ranges from 8.5 to 69m/day (28 to 227 ft/day). Comparatively highhydraulic conductivity occurs in anortheast-southwest trending regionconnecting D Area, Central Shops, and R Areaand defines a "high permeability" zone in theaquifer. Here, hydraulic conductivities rangefrom 36 to 69 m/day (117 to 227 ft/day). The"high permeability" zone parallels the trace ofthe Pen Branch Fault, and reflects changingdepositional environments in response tomovement along the fault as described above.South of the trace of the Pen Branch Fault,hydraulic conductivity for the aquifer reflectsthe return to a deeper water shelf/deltaicdepositional regime.

Allendale Confining System. The Allendaleconfining system is present in the southeasternhalf of the study area and separates theMidville Aquifer System from the overlyingDublin Aquifer System. In the study area, theAllendale confining system consists of asingle unit, the McQueen Branch confiningunit. The confining system is correlative withthe unnamed confining unit that separates theMiddendorf and Black Creek aquifers ofAucott et al., and with the BlackCreek-Cusseta confining unit of Clark et al.The system dips approximately 6.7 m/km (27ft/mi) to the southeast and thickens uniformlyfrom about 15.2 meters (50 feet) at the updiplimit to about 61 meters near the easternboundary of the study area. The rate ofthickening is greater in the east than in thewest. The updip limit of this confining systemis established where pronounced thinningoccurs parallel to the Pen Branch Fault.

Sediments of the Allendale confining systemare fine grained and consist of clayey, siltysand, clay, and silty clay and micritic claybeds that constitute the middle third of theBlack Creek Formation. North of the updiplimit of the confining system, where theMcQueen Branch confining unit is part of theDublin-Midville Aquifer System, the sectionconsists of coarser-grained, clayey, silty sandand clay beds.

McQueen Branch Confining Unit

The McQueen Branch confining unit isdefined by the hydrogeologic properties of thesediments penetrated in well P-27. At itstype-well location, the McQueen Branchconfining unit is 17 meters (55 feet) thick, andis present from -100 to -117 meters (-329 to –384 feet) msl. Total clay thickness is 14meters (45 feet) in three beds, which is 82% ofthe total thickness of the unit, with a leakancecoefficient of 3.14 x 10

-6 m/day (1.03x10-5

ft/day). The confining unit in well P-27consists of the interbedded, silty, often sandyclay and sand beds that constitute the middlethird of the Black Creek Formation.

The clay beds tend to be anomalously thinalong a line that parallels thesouthwest-northeast trend of the Pen BranchFault and the north-south trend of the AttaFault. The reduced clay content in these areassuggests shoaling due to uplift along the faultsduring Upper Black Creek-Steel Creek time.

Midville Aquifer System. The MidvilleAquifer System is present in the southern halfof the study area; it overlies the Appletonconfining system and is succeeded by theAllendale confining system. In the study area,the Midville Aquifer System consists of oneaquifer, the McQueen Branch aquifer unit.South of well C-10 (see Figure 1.4-21), thesystem may warrant further subdivision intoseveral aquifers and confining units.Thickness of the unit ranges from 71 meters(232 feet) at well P-21 to 103 meters (339feet) at well C-10. Variation in the thicknessof the unit, as well as the updip limit of thesystem, results from variation in the thicknessand persistence of clay beds in the overlyingAllendale confining system. Near the PenBranch Fault, contemporaneous movement onthe fault may have resulted in shoaling in thedepositional environment, which is manifestedin a thickening of the sands associated withthe Midville Aquifer System. The uppersurface of the aquifer system dipsapproximately 4.73 m/km (25 ft/mi) to thesoutheast across the study area.


A-36

The Midville Aquifer System was defined andnamed by Clarke et al. (Ref. 121) for thehydrogeologic properties of the sedimentspenetrated in well 28-X1, near the town ofMidville in Burke County, Georgia. Here, theupper part of the aquifer system consists offine to medium sand of the lower part of theBlack Creek-Cusseta unit. The Midville iscomparable to the lower portion of theChattahoochee River aquifer of Miller andRenken (Ref. 104) and correlative with theMiddendorf aquifer of Aucott and Sperian.

McQueen Branch Aquifer Unit

The McQueen Branch aquifer unit occursbeneath the entire study area. It thickens fromthe northwest to the southeast and ranges from36 meters (118 feet) at well AIK-858 to 103meters (339 feet) at well C-10 to the south.Locally, thicknesses are greater along the traceof the Pen Branch Fault because of theabsence and/or thinning of clay beds thatcompose the overlying McQueen Branchconfining unit. The upper surface of theMcQueen Branch dips approximately 4.7m/km (25 ft/mi) to the southeast.

The McQueen Branch aquifer unit is definedfor the hydrogeologic properties of sedimentspenetrated by well P-27 near the center of thestudy area. Here, it is 62 meters (203 feet)thick and occurs from -117 to -180 meters(-384 to –587 feet) msl. It contains 56 meters(183 feet) of sand in four beds, (which is 90%of the total thickness of the unit). The aquiferconsists of silty sand of the MiddendorfFormation and clayey sand and silty clay ofthe lower one-third of the Black CreekFormation. Typically, a clay bed or severalclay beds that cap the Middendorf Formationare present in the aquifer. These clay bedslocally divide the aquifer into two aquiferzones.

Appleton Confining System

The Appleton confining system is thelowermost confining system of theSoutheastern Coastal Plain hydrogeologicprovince and separates the province from the

underlying Piedmont hydrogeologic province.It is equivalent to the Black Warrior Riveraquifer of Miller and Renken and to the basalunnamed confining unit of Aucott et al. (Ref.122). The confining system is essentiallysaprolite of the Paleozoic and Mesozoicbasement rocks and indurated, silty and sandyclay beds, silty clayey sand and sand beds ofthe Cretaceous Cape Fear Formation.Thickness of saprolite ranges from 2 to 14meters (6 to 47 feet), reflecting the degree ofweathering on the basement unconformityprior to deposition of the Cape Fearterrigenous clastics. Thickness of saprolitedetermined from the Deep Rock Borings study(DRB wells) ranges from 9 to 30 meters (30 to97 feet) and averages 12 meters (40 feet) inwells DRB-1 to DRB-7. In the northern partof the study area, the Cape Fear Formationpinches out and the Appleton consists solelyof saprolite.

Some variability in thickness is noted alongthe trace of the Pen Branch Fault. It dips atabout 5.9 m/km (31 ft/mi) to the southeast andthickens from 4.6 meters (15 feet) in well C-2near the north end of the study area to 22meters (72.2 feet) in well C-10 in the south.Sediments of the confining system do not cropout in the study area. Thinning of theAppleton confining system in well PBF-2 isprobably a result of truncation of the sectionby the Pen Branch Fault.

The confining system consists of a singleconfining unit throughout the study area.Toward the coast, however, the Appletonconfining system thickens considerably andincludes several aquifers. The aquifersincluded in the confining system in thedowndip region are poorly defined becausefew wells penetrate them. They arepotentially water producing but the depth andgenerally poor quality of water in the aquifersprobably precludes their utilization in theforeseeable future. The Appleton confiningsystem includes no aquifer units or zones inthe northern and central parts of the studyarea.


A-37

Fine- to coarse-grained sand beds, often verysilty and clayey, occur in the upper part of theCape Fear Formation in the southern part ofthe study area. The sand appears to be incommunication with sand of the overlyingMcQueen Branch Aquifer System and isincluded with that unit.

Hydrogeology of the Piedmont Province

The basement complex, designated thePiedmont hydrogeologic province in thisreport, consists of Paleozoic crystalline rocks,and consolidated to semiconsolidated UpperTriassic sedimentary rocks of the Dunbartonbasin. All have low permeability. Thehydrogeology of the province was studiedintensively at SRS to assess the safety andfeasibility of storing radioactive waste in theserocks. The upper surface of the province dipsapproximately 11 m/km (36 ft/mi) to thesoutheast. Origins of the crystalline andsedimentary basement rocks are different, buttheir hydraulic properties are similar. Therocks are massive, dense, and practicallyimpermeable except where fracture openingsare encountered. Water quality in these unitsis also similar. Both contain water with highalkalinity and high levels of calcium, sodium,sulfate, and chloride. The low aquiferpermeability and poor water quality in thePaleozoic and Triassic rocks render themundesirable for water supply in the study area.

TECTONIC STRUCTURES: FAULTING,FOLDING, AND RIFT BASINS

Paleozoic Basement Beneath SRS

Information concerning structural features inthe basement beneath Savannah River Site ismainly derived from analysis of structuralfabrics recorded in core samples from deepborings and at larger scales from geophysicaltechniques such as gravity and magneticsurveys and seismic reflection profiles (seeDennis, this field guide).

An integrated analysis of the structural fabricin the basement core in addition to thegeophysical data concluded that at least two

regional scale ductile faults are present in thebasement beneath Savannah River Site andvicinity, the Upper Three Runs fault and theTinker Creek Fault. These faults areexpressed in the aeromagnetic data aslineaments and are interpreted to be associatedwith a thrust duplex that emplaces the rocks ofthe PBF Formation (Tinker Creek Nappe) overthe DRB formation . The age of the faulting isconstrained by a radiometric age on biotitethat dates the movement at about 300 Ma,which would indicate that these faults are partof the Paleozoic Eastern Piedmont FaultSystem.

In order to resolve faulting that deformCoastal Plain sediments, the topography of thebasement surface was mapped utilizing thedata listed above along with more recentlyacquired seismic reflection profiles. The mapof basement topography indicates that offsetsof the basement surface that range fromapproximately 30 meters (100 ft) in magnitudedown to the resolution limits of the data arepresent on the basement surface. Howevermost of these offsets are of relatively smallmagnitude and have limited lateral extents..The Crackerneck and Pen Branch Faults arerelatively well constrained with borings. Theother faults are projected from geophysicaldata only and their parameters are less wellknown. Of these faults the Pen Branch faulthas been extensively studied.

Post-Rift and Cenozoic Structures. Thesestructures include offshore sedimentary basins,such as the Carolina trough and the BlakePlateau basin; transverse arches andembayments, such as the Cape Fear arch andthe Southeast Georgia Embayment; CoastalPlain faulting; and paleoliquefaction featuresthat provide information on the recurrence ofthe Charleston earthquake.

Folding and Faulting

The most definitive evidence of crustaldeformation in the Late Cretaceous throughCenozoic is the reverse sense faulting found inthe Coastal Plain section of the eastern U.S.Under the auspices of the Reactor Hazards


A-38

Program of the late 1970s and early 1980s,USGS conducted a field mapping effort toidentify and compile data on all youngtectonic faults in the Atlantic Coastal Plain.Consequently, many large, previouslyunrecognized Cretaceous and Cenozoic faultzones were found (Prowell, 1983). Of 131fault localities cited, 26 were within North andSouth Carolina. The identification ofCretaceous and younger faults in the easternUnited States is greatly affected bydistribution of geologic units of that age.Many of the faults reported by Prowell arelocated in proximity to the Coastal Plain onlapover the crystalline basement. This may bedue to the ease of identifying basementlithologies in fault contact with Coastalsediments.

Prowell and Obermeier characterized thefaults as mostly northeast trending reverse slipfault zones with up to 100 km (62 miles)lateral extent and up to 76 meters (250 feet)vertical displacement in the Cretaceous. Thefaults dip 40° to 85°. Offsets were observed tobe progressively smaller in youngersediments. This may be due to an extendedmovement history from Cretaceous throughCenozoic .

Offset of Coastal Plain sediments at SRSincludes all four Tertiary unconformities.Following deposition of the Snapp Formationsome evidence indicates oblique-slipmovement on the existing faults. The offsetsinvolve the entire Cretaceous to Paleocenesedimentary section. In A/M Area, thisfaulting formed a serious of horsts and grabensbounded by subparallel faults that truncate atthe fault intersections. The strike orientationsof the individual fault segments vary from N11°E to N 42°E, averaging about N 30°E.Apparent vertical offset varies from 4.5 to 18meters (15 to 60 feet), but throws of 9 to 12meters (30 to 40 feet) are most common.

This faulting was followed by erosion andtruncation of the Paleocene section at the LangSyne/Sawdust Landing unconformity.Subsequent sediments were normal faultedfollowing deposition of the Santee Formation.

Typically, the offset is truncated at the Santeeunconformity, and the overlying TobaccoRoad/Dry Branch formations are not offset.Locally, however, offset of the overlyingsection indicates renewed movement on newor existing faults after deposition of TobaccoRoad/Dry Branch sediments.

In conjunction with these observations ofCoastal Plain faults, modern stressmeasurements provide an indication of thelikelihood of Holocene movement. Moos andZoback report a consistentnortheast-southwest direction of maximumhorizontal compressive stress (N 55-70°E) inthe southeast U.S. Their determination isbased on direct in situ stress measurements,focal mechanisms of recent earthquakes (seeStevenson this guide book), and younggeologic indicators. Shallow seismicity in thearea, within crystalline terranes, ispredominantly reverse character. Moos andZoback conclude that the northeast directedstress would not induce damaging reverse andstrike-slip faulting earthquakes on the PenBranch fault, a northeast striking Tertiary faultin the area. These same conclusions may beimplied for the other northeast trending faultsmapped by Prowell.

In A/M Area at SRS, faulting appears to havebeen episodic and to have varied in styleduring the Tertiary. Oblique-slip faultingdominated the Cretaceous/Paleocene events,with a local north-south stress orientation.Subsequently, left-lateral shear on thepre-existing faulting and normal faultingoccurred, with a corresponding shift in thedirection of maximum compressional stressoriented N 20°E to N 30°E.

Pen Branch Fault

The Pen Branch fault has been regarded as theprimary structural feature at SRS that has thecharacteristics necessary to pose a potentialseismic risk. As stated below, studies haveindicated that, despite this potential, the faultis not capable.


A-39

The Pen Branch fault is an upwardpropagation of the northern boundary fault ofthe Triassic Dunbarton basin that wasreactivated in Cretaceous/Tertiary time. Thefault dips steeply to the southeast. In thecrystalline basement, slip was originally downto the southeast, resulting in the formation ofthe Dunbarton rift basin. However, movementduring Cretaceous into Tertiary time wasreverse movement, that is, up to the southeast.There could also be a component of strike-slipmovement.

Joints

Joints are common on the SRS and vicinity.Though their mechanism of formation is notwell understood, their age was determined tobe constrained by interpretation that cutansoften developed along pre-existing joints. Thejoints, therefore, pre-dated the pedogenicprocesses that formed the cutans. Highly

variable orientations of cutans suggested thatthe orientation of joints on the SRS was alsohighly variable. A gradual and consistentchange in orientation of cutans over 30 to 60meters (100 to 200 feet) at some outcropssuggested the orientation of joints also locallychanged gradually and consistently. A lack ofconsistent preferred orientations of jointsacross the SRS did not favor a tectonic originfor these features. Furthermore, no clearrelationship existed between thejoint-controlled cutans and the localtopography. The joints, therefore, wereprobably not related to slope mass wasting. Alocal, gradual change in orientation overseveral hundred feet, and the commonoccurrence of closed depressions on the SRS,are consistent with differential settling fromsubsurface dissolution.


A-40

Figure 1. Generalized phyisiography for the SRS.


A-41

Figure 2. Regional stratigraphy.


A-42

Figure 3. Foot by foot sediment samples from well GCB-5. White strip indicates missing data.


A-43

Figure 3. continued.


A-44

Figure 3. continued.


A-45

Figure 4. Geophysical log signature with stratigraphic markers.


A-46

$T

$T

$T

$T

$T

$T$T

$T

$T

$T$T

$T

$T$T

$T$T$T

$T

#S#S

#S

#S

#S

#S

#S

#S#S

#S#S

#S

#S

#S

#S

#S

#S

#S

#S#S

#S

#S

#S #S

#S

#S

#S

#S

#S

#S

#S#S

#S

#S

#S

#S

#S

#S

#S

#S

#S

#S#S

#S

#S

#S

#S

#S

#S

#S

#S

#S

#S

#S

#S

#S

#S

#S

$T

$T

$T

$T

C1

C2

C5

C10

E

N

B F H

Z

D

A

R

L

S

P

K

C

M

RR

TNX

USFS

C7

P30

P4R

P5R

P6R

P7RP8R

P9R

CPC1

CPC2

HPC1

P11R

P12R

PBF1

PBF2

PBF3

PBF4

PBF5

PBF6

PBF8

PPC1

ALL27

BPC-1

CFD-1

DRB-1

DRB-2

DRB-3

DRB-4

DRB-7

DRB-9

GCB-1 GCB-2

GCB-3

GCB-8

NPR-5

SSW-1

SSW-2

SSW-3

SSW-5

AIK830 AIK831

AIK858

ALL324

CFD-18

DRB-10

DRB-11

GIRARD

MMP-3B

MMP-4B

GCB-5.1

GCB-7-L1

PD.

PW53A

GCB-4SSW-4

DRB-5DRB-6

DRB-8

P13A

P14C

P15A

P16A

P17A

P18A

P19A

P20B

P21A

P22A

P23A

P24A

P25A

P26A

P27B

P28AP29A

P30A

MMP-2A

80

220

220

160

240

260

80

100

120

140

180

200

200

22024

026028

0

300

320

100120

140160

180200

180160140

Borings PenetratingBasement

South Carolina

Georgia

SRS Roads

SRS Areas

#S

Offsite Cluster Wells

SRS Boundary

$T Onsite Cluster Wells

$

36^22 '

PL ANTNORTH

TRU ENORTH

3 0 3 6 9Kilometers

2 0 2 4 6Miles

Water Table (20 ft Interval)

Legend

Figure 5. Location of type and reference wells for hydrostratigraphic units at SRS.


A-47

Figure 6. Comparison of chronostratigraphic, lithostratigraphic, and hydrostratigraphc units in the SRSregion.


A-48

Figure 7. Hydrogeologic nomenclature for the SRS region.


B-1

Establishing a Hydrostratigraphic Framework using Palynology: AnExample from the Savannah River Site, South Carolina, U.S.A.

Robert S. Van Pelt, Bechtel Savannah River Inc., Savannah River Site, Aiken South CarolinaDonald W. Engelhardt (deceased), Earth Sciences Resources Institute, University of South Carolina, Columbia,South CarolinaRaymond A. Christopher, Dept. of Geological Sciences, Clemson University, Clemson, South CarolinaJoyce Lucas-Clark, Clark Geological Services, 1023 Old Canyon Road, Fremont, California

INTRODUCTION

The primary mission of the Savannah RiverSite (SRS) is to produce nuclear materials fornational defense. In the 1980’s concern aboutenvironmental issues precipitated studies ofthe hydrogeology of the SRS. Since 1990, themission has expanded to includeenvironmental restoration following more than40 years of waste-producing activities.Numerous waste-sites at SRS are currentlyundergoing environmental assessments and/orremediation under state and federalregulations.

An essential part of the environmentalassessments is the characterization of thesubsurface hydrogeology. Hydrogeologicalcharacterization involves establishing thehydrologic and geologic conditions (includingthe hydrostratigraphy), and incorporating thisinformation into groundwater flow andcontaminant transport models. Results fromthe hydrogeology and groundwater models arethen applied in the selection of remediationstrategies for a waste site. At the SRS,palynology has played a critical role inestablishing the hydrogeologic framework,including the identification andcharacterization of hydrogeologic units andtheir relationships.

Differentiating among the sedimentary unitsthat comprise the hydraulic system at SRS canbe difficult when based solely on lithology.Most of the units consist of sediments of nearshore and deltaic origin, regardless of age,such that similar depositional origins masklithostratigraphic distinctions. Many

sedimentary units are lenticular, compriseseveral depositional facies, and transgresstime. Palynology has been applied to thevarious sedimentary units for agedetermination and paleoenvironmentalinterpretation.

Palynology also plays an important role in thedetection of groundwater preferentialpathways, such as facies changes andstructural elements (e.g., faults). Prediction ofsuch pathways is essential for developingaccurate models of ground water contaminantflow and transport. The subsurfaceinvestigation at the SRS represents one of thefrontier applications of palynology tohydrogeology and environmental restoration.

PREVIOUS PALYNOLOGICAL WORKAT SRS

As early as the middle of the 1970's, the needfor accurate interpretation of thehydrostratigraphy beneath the SRS becameapparent to those concerned with potentialcontamination problems (Cahil, 1982; Fayeand Prowell, 1982; Marine and Root, 1975).Palynology has been used from the middle ofthe 1960's to the present as a basis for revisingthe formal stratigraphy (Siple 1967; Prowell etal., 1985a, 1985b; Gohn and Campbell, 1992;Nystrom et al., 1991; Fallaw and Price, 1992,1995). For example, the Sawdust LandingFormation (hydrostratigraphically equivalentto the middle part of the Crouch BranchConfining Unit in the northern part of theSRS; see Text-Figure 2) was determined to bePaleocene rather than Cretaceous in age bydinoflagellate and pollen analysis (Lucas-


B-2

Clark, 1992a). Subsequently, this confiningunit was separated into three distinctunconformity-bound units based onpalynological age interpretations (Aadland etal., 1992). In general, age interpretationsbased on palynology allow for more accuratecorrelation with other stratigraphic units in thesoutheastern Coastal Plain, in what has cometo be considered the Floridan Aquifer System.The revised lithostratigraphy and study ofgroundwater transport led to a completeformal hydrostratigrapic nomenclature for theSRS and surrounding vicinity (Aadland, 1990;Aadland and Bledsoe, 1992, Aadland et al.,1992; 1995) (Text-Figure 2).

Studies in the middle of the 1980's focused onTertiary carbonate units which wereconsidered as conduits for contaminants, andwhich were difficult to correlate due to theirdiscontinuous, lenticular, reef-like nature.Although palynomorphs are usually sparse inthe carbonates themselves, they are abundantin the interbedded and encapsulating clays andmud. Determining the age and identifyingreliable palynological datums in theseotherwise unfossiliferous fine-grained detritalunits was determined to be more critical to thecorrelation of the carbonates than were thefossils in the carbonates themselves.

More recent emphasis on environmentalrestoration has required compliance with stateand federal regulations, (e.g. ResourceConservation and Recovery Act (RCRA)/Comprehensive Environmental ResponseCompensation and Liability Act (CERCLA)),which require a complete subsurfacehydrogeologic/geologic characterization ofwaste sites, such as landfills, settling basins,storage tank areas, etc. (Van Pelt et. al., 1993;Lucas-Clark and Van Pelt, 1993; Van Pelt andLucas-Clark, 1996). In nearly all cases,palynology has been utilized to aid inidentifying hydrostratigraphic units andimproving and confirming well logcorrelation.

Certain hydrostratigraphic units, specificallyconfining units, have characteristic

dinoflagellate and pollen assemblages that canprovide age and paleoenvironmentinterpretations. The interpretations allow foraccurate correlation as well as prediction ofpinch outs and leaks in the confining units.Some confining units, such as the lower partof the Crouch Branch Confining Unit, weredeposited under marine conditions andrepresent a major transgressive sequence.Others units, such as the late Paleocene upperpart of the Crouch Branch Confining Unit, arecharacterized by dinoflagellates that areabundant in deltaic and possibly brackishwater environments. Overall, shelfaldinoflagellate assemblages appear to dominateconfining units that are more reliable ascontinuous barriers to hydrologic flow exceptwhere breached by unconformities; shallowwater (e.g., estuarine, lagoonal) dinoflagellateassemblages generally comprise confiningunits that are not reliable barriers to verticalgroundwater flow. This relationship isillustrated by the upper part of the CrouchBranch Confining Unit, which grades intocoarser grained non-marine sediments on thesoutheastern side of the Savannah River, andis absent on the northwest side of theSavannah River. This conclusion wasparticularly significant to the Trans-RiverFlow study conducted in the middle of the1990’s by the United States GeologicalSurvey and the Georgia State GeologicalSurvey. The focus of the Trans-River Flowstudy was to determine the possibility ofcontaminated groundwater migrating from theSavannah River Site under the SavannahRiver and into the aquifers in Georgia (Clarkeet al.,1994; Clarke et al., 1996; Leeth et al.,1996; Falls et al., 1997a and b; Huddlestunand Summerour, 1996; and Clarke and West,1997).

SAVANNAH RIVER SITE HYDRO-GEOLOGY AND PALYNOLOGICALSTRATIGRAPHY

The hydrology beneath the SRS andsurrounding area is controlled by the pre-Cretaceous Paleozoic metamorphic andTriassic clastic rocks and overlying Coastal


B-3

Plain sediments. Poor water quality and lowpermeability prevent usage of ground waterfrom within the pre-Cretaceous units.Overlying these units, however, the CoastalPlain sediments comprise a multi-layerhydrologic system in which retarding clay andmarl beds are inter-layered with sands andlimestone beds that transmit water morereadily (Aadland and Bledsoe, 1992). Inascending stratigraphic order, the hydrologicsystem is divided into four aquifers (McQueenBranch, Crouch Branch, Gordon, and UpperThree Runs Aquifers), each separated byintervening confining units (Text-Figure 3).The Gordon and Upper Three Runs Aquiferscoalesce in the northern part of the area toform the Steed Pond Aquifer (Lewis andAadland, 1992).

The relation between the palynology and thehydrogeology is presented below. Selectedpalynomorphs that are useful inhydrostratigraphic correlation are presented inText-Figures 4 through 8. The hydrogeologyis presented in ascending order.

Appleton Confining System

Separating the aquifers of the Coastal Plainsequence from basement is the AppletonConfining System, which consists of saproliteoverlying crystalline basement rock, and thesedimentary unit referred to as the Cape FearFormation. The Cape Fear Formation ismainly non-marine and dated by pollen andspores as middle Turonian to Santonian (LateCretaceous) in age (Lucas-Clark, 1992a). TheAppleton Confining System/Cape FearFormation is characterized by the presence ofa distinctive pollen type that has been referredto as aff. Porocolpopollenites spp. by Doyleand Robbins (1977) (Text-Figure 6A). Inaddition, the unit contains diverse andabundant representatives of the genusComplexiopollis, including C. abditusTschudy 1973 (Text-Figure 6B), as well asother described and undescribed forms. Allspecies of Complexiopollis that occur in theunit are of the “newer” variety ascharacterized by Christopher (1979).

Representatives of the genusPseudoplicapollis (Text-Figure 6C) also occurthroughout the unit, but they are not soabundant as are the Complexiopollis forms.Other genera that are present in assemblagesfrom the Appleton Confining System/CapeFear Formation are Praecursipollis (Text-Figure 6D), Osculapollis (Text-Figure 6E),Trudopollis (Text-Figure 6F), Plicapollis(Text-Figure 6G), Labrapollis (Text-Figure6H), and numerous undescribed forms.

McQueen Branch Aquifer

The McQueen Branch Aquifer overlies theAppleton Confining System and is the lowerof two Cretaceous aquifers at SRS. Thisaquifer represents the principal source fordomestic use at SRS. Domestic wells in thesesands commonly yield more than 1000 gallonsper minute of high quality water.

The McQueen Branch Aquifer corresponds tothe sands that are referred to as theMiddendorf Formation as well as these in thelower part of the Black Creek Formation atSRS. The Middendorf Formation isconsidered non-marine to shallow marine inorigin, and correlates with the lower (but notbasal) part of the Tar Heel Formation in NorthCarolina. Dinoflagellates and pollen in theshallow marine facies have yielded mainlyearly Campanian (Late Cretaceous) ages. Inmost marine samples from the MiddendorfFormation, dinoflagellates are sparse andassemblages consist mainly of spiny cysts anda few peridiniacean genera such asSubtilisphaera Jain & Millepied, SpinidiniumCookson & Eisenack, and PalaeocystodiniumAlberti. Pollen assemblages from unitsmapped as the Middendorf Formation at SRSare distinctly different from those in theunderlying Appleton Confining System/CapeFear Formation, which reflects theunconformity that separates these units.Occurring in the Middendorf assemblages arePlicapollis usitatus Tschudy 1975 (Text-Figure 6I), representatives of the genusProteacidites (Text-Figure 6J), twoundescribed species with monocolpate


B-4

apertures and echinate surface ornamentation,and a variety of undescribed taxa. In addition,the lowest stratigraphic occurrence ofHolkopollenites chemardensis Fairchild inStover et al., 1966 (Text-Figure 6M, N) occurswithin this unit (Text-Figure 4). The pollendata suggest an early Campanian age for theunit.

The lower part of the Black Creek Formationis marine to non marine in origin, and containspollen assemblages similar to those in theunderlying Middendorf Formation.Dinoflagellates in the shallow marine facieshave yielded mainly Campanian ages. In mostof the marine samples, dinoflagellates areabundant, and assemblages consist ofabundant small peridiniacean species ofSubtilisphaera, Spinidinium,Palaeohystrichophora Deflandre, and largerperidiniacean species of Andalusiella (Riegel)Riegel & Sarjeant, Xenascus Cookson &Eisenack, Palaeocystodinium Alberti,Cerodinium Vozzennikova, andTrithyrodinium Drugg. Gonyaulacaceandinoflagellates other than ceratioid types aresparse (Lucas-Clark, 1992a). Age diagnosticdinoflagellate species in the McQueen BranchAquifer/Black Creek Formation includeAndalusiella spicata (May) Lentin andWilliams, Xenascus sarjeantii (Corradini)Stover & Evitt, and Palaeohystrichophorainfusorioides Deflandre.


The McQueen Branch Confining Unitseparates the two Cretaceous (McQueenBranch and Crouch Branch) aquifers. Theunit thins and pinches out in the vicinity of thewell cluster P 19 (Text-Figure 1) near thecenter of the SRS (where the two Cretaceousaquifers are in communication) but it persiststhroughout the remainder of the site region.The unit corresponds to the clays of themiddle part of the Black Creek Formation.

Dinoflagellates from this unit have yieldedmainly late Campanian ages. Characteristicsof the dinoflagellate assemblages vary. Some

assemblages are dominated by spiny cysts;others by Areoligera coronata (Wetzel)Lejeune-Carpenter/senonensis; others bysmall peridiniacean cysts; and others byXenascus ceratioides (Deflandre) Lentin &Williams/sarjeantii complex. Useful speciesof dinoflagellates include Dinogymniumeuclaense Cookson & Eisenack, D. digitus(Deflandre) Evitt et al., andPalaeohystrichophora infusorioides var. Aand B of Aurisano (1989). Resedimentation ofearly Cretaceous dinoflagellates is notuncommon (Lucas-Clark, 1992a).

Pollen assemblages from the McQueenBranch Confining Unit include a variety oftriporate forms that can be assigned to thegenera Momipites (Text-Figure 7Q, Z),Caryapollenites (Text-Figure 7R), andBetulaceoipollenites (Text-Figure 7S) Inaddition, representatives of the genusHolkopollenites (Text-Figure 7V) are diverseand, at times, common elements in theseassemblages. A late Campanian age issuggested by the pollen data (Text-Figure 4).

Crouch Branch Aquifer

The Crouch Branch Aquifer represents theupper of the two Cretaceous aquifers at SRS.It consists of the sand layers at the uppermostpart of the Black Creek Formation and themajority of the overlying Steel CreekFormation. Dinoflagellates from the fine-grained marine portions of this aquifer suggestages of late Campanian and Maastrichtian.Unfortunately, the clean sands and highlyaltered kaolinitic clays comprising the upperpart of this aquifer often do not contain well-preserved fossil material. However, genericlevel identification of dinoflagellates such asDinogymnium spp. and species ofRugubivesiculites (Text-Figure 7BB) pollenare useful in distinguishing Cretaceous fromoverlying Tertiary units (Lucas-Clark and VanPelt, 1993). Diagnostic dinoflagellate speciesinclude: Dinogymnium acuminatum Evitt et al,Palaeocystodinium benjaminii Drugg, andCerodinium striatum (Drugg) Lentin &Williams.


B-5

Pollen data from the Crouch Branch Aquifersuggest the presence of an unconformitywithin the unit; this unconformity correspondsto the contact between the Black Creek andSteel Creek Formations, and represents theuppermost part of the Campanian and most ofthe lower Maastrichtian. Osculapollisaequalis Tschudy 1975 (Text-Figure 7W) isrestricted to the units below the unconformity,as are several morphotypes of the genusHolkopollenites (Text-Figure 7O,V).Restricted to the units above the unconformityare Libopollis jarzenii (Text-Figure 7X) and athin-walled species of Baculostephanocolpites(MPH-1 of Wolfe, 1976) (Text-Figure 7Y). Inthe SRS and surrounding vicinity, the SteelCreek Formation is primarily non-marine tomarginal marine, whereas the underlyingBlack Creek Formation is near shore to openmarine in origin.


The confining system that overlies the CrouchBranch Aquifer represents the principalconfining unit for the SRS and vicinity. Thisconfining unit prevents migration ofcontaminated water from the Steed Pond andGordon Aquifers of Tertiary age into theunderlying Cretaceous Crouch BranchAquifer. However, in some areas within theSRS, the confining unit is discontinuousresulting in local communication of theTertiary and Cretaceous aquifers. It is thuscritical to identify the confining unit where itis present, and important to understand itsrelation to other units.

The Crouch Branch Confining Unitcorresponds to lower and upper Paleocene andlower Eocene lithologic units, whichcorrespond to the Lang Syne, SawdustLanding, the Snapp, and the Fourmile BranchFormations. The base of the Crouch BranchConfining Unit/Sawdust Landing Formation ischaracterized by the first appearance datum(FAD) of several Danian (Paleocene)dinoflagellate species: Carpatella cornutaGrigorovich (Text-Figure 8A),

Damassadinium californicum (Drugg)Fensome et al. (Text-Figure 8R) Spiniferitesseptatus (Cookson & Eisenack) McLean,Senegalinium ?dilwynense (Cookson &Eisenack) Stover & Evitt (Text-Figure 8B),and Spinidinium pulchrum (Bensen) Lentin &Williams.

The upper part of the Crouch BranchConfining Unit (equivalent to the upper partof the Lang Syne, Snapp and Fourmile BranchFormations) is characterized by dinoflagellateassemblages that are dominated by one or atmost, a few species. These assemblages maycomprise small peridiniacean species alongwith Areoligera Lejeune-Carpenter spp.and/or Cordosphaeridium Eisenack spp.Apectodinium homomorphum (Deflandre &Cookson) Lentin & Williams (Text-Figure8H), A. quinquelatum (Williams & Downie)Lentin & Williams, and Eocladopyxispeniculata Morgenroth make their firststratigraphic appearance in the upper part ofthe confining unit (Edwards, 1990; Lucas-Clark, 1992b).

There are several pollen species that havetheir bases or first appearances ( FAD) in theCrouch Branch Confining Unit/ SawdustLanding Formation (Text-Figure 5A,B).These include Choanopollenites conspicuus(Groot & Groot 1962) Tschudy 1973,Choanopollenites discipulus Tschudy 1973(Text-Figure7G), Favitricolporitesbaculoferus ( Pflug in Thompson and Pflug1953) Srivastrava 1972 (Text-Figure 7E),Nudopollis terminalis ( Pflug & Thompson inThompson & Pflug, 1953) Pflug 1953 (Text-Figure 7D), Nudopollis thiergarti ( Thompson& Pflug 1953) Pflug 1953 (Text-Figure 7I),Momipites dilatus Fairchild in Stover et al.(1966), Pseudoplicapollis limitataFrederiksen 1978, Tricolpites asperFrederiksen 1978, and Trudopollis plenusTschudy 1975 (Text-Figure 7J). All of thesetaxa have tops or last appearances (LADs)higher in the stratigraphic section but theassemblage is characteristic of earlyPaleocene age. The range top ofPseudoplicapollis serenus Tschudy 1975


B-6

(Text-Figure 7O,P) that is within the SawdustLanding Formation is considered byFrederiksen (1991) to mark the top of theDanian or early Paleocene.In the upper part of the Crouch BranchConfining Unit which includes the upper LangSyne and Snapp Formations, there are severalpollen and spore species that have their lastappearances. These include taxa which hadtheir first appearances in the lower part of theCrouch Branch Confining Unit:Choanopollenites conspicuus,Favitricolporites baculoferus (Text-Figure7E), Interpollis paleocenicus (Elsik 1968)Frederiksen 1980, Lusatisporis indistinctaFrederiksen 1979 (Text-Figure 7H),Momipites dilatus (Text-Figure 6Z),Nudopollis thiergartii (Text-Figure 7I),Pseudoplicapollis limitata (Text-Figure 7b),Tricolpites asper (Text-Figure 7K), andTrudopollis plenus (Text-Figure 7J).Holkopollenites chemardensis Fairchild inStover et al. (1966) Text-Figure 6M,N), whichfirst appears in the Upper Cretaceous has a topin this interval. Corsinipollenites? verrucatusFrederiksen 1988 has a first and lastappearance within the Snapp Formation.Frederiksen (1998) reports this species fromthe lower Eocene. Taxa such asSubtriporopollenites nanus (Pflug &Thompson in Thompson and Pflug 1953)Frederiksen 1980 (Text-Figure 7F),Pseudolaesopollis ventosus (Potonie’ 1931)Frederiksen 1979, and Milfordia hungarica (Kedves 1965) Krutzsch & Vanhoorne 1977(Text-Figure 7R) have their first appearancesor bases within the Snapp Formation orequivalents in the southeastern United States (Frederiksen 1980a and b, 1988 and 1998).

Gordon/Steed Pond Aquifer and GordonConfining Unit

The middle to upper Eocene aquifer (SteedPond/Gordon) and confining unit (Gordon)that overlies the Crouch Branch ConfiningUnit is a complex set of strata consisting, inpart, of carbonates (limestone) that arelenticular and discontinuous, as well asalternating clay and sand layers that act as

confining units. Understanding the potentialcomplexity of groundwater flow requiresconsiderable accuracy in the identification andcorrelation of the lithologic units.Lithologic units that correspond to the SteedPond and equivalent aquifers include theCongaree Formation, Warley Hill Formation,Santee Formation, Clinchfield Formation,Irwinton Sand and Griffins Landing Membersof the Dry Branch Formation, and theTobacco Road Formation. Sedimentscomprising the Tobacco Road Formation areconsistently barren of organic-walledmicrofossils. The other lithologic units haveyielded dinoflagellate and pollen assemblages,usually from the clay- and silt-rich intervals.

Dinoflagellates useful in correlating thecomplex of middle to upper Eocene unitsinclude: Pentadinium favatum Edwards, P.goniferum Edwards (Text-Figure 8I),Glaphrocysta Stover & Evitt spp. (Text-Figure 8F), Wetzelliella articulata Eisenackgroup (Text-Figure 8E), Membranophoridiumaspinatum Gerlach, M. bilobatum Michouxand Charlesdowniea variabilis (Bujak) Lentin& Vozzhennikova (Text-Figure 8Q).

Throughout most of the SRS, confiningconditions (Gordon Confining Unit) exist toseparate the Steed Pond Aquifer into twodistinct aquifer zones, the Upper Three Runsand underlying Gordon Aquifer zones. TheGordon Confining Unit consists of sandyclays to clayey sands which represent parts ofthe Warley Hill and Tinker Formations, andthe Blue Bluff Unit. Dinoflagellates firstappearing within this confining unit includePentadinium laticinctum Gerlach (Text-Figure8N), Cordosphaeridium cantharellus(Brosius) Gocht, and Charlesdownieavariabilis (Text-Figure 8Q) Pentadiniumgoniferum (Text-Figure 8I) last appears withinthe upper part of this confining unit.

There are not as many extinctions or tops forpollen taxa in the middle and late EoceneGordon Confining Unit as in the CrouchBranch Confining Unit. The most distinctivespecies are Bombacacidites aff. B. reticulatus


B-7

Krutzsch 1961, Subtriporopollenites nanus(Text-Figure 7F), Malvacipollis tschudyi(Frederiksen 1973) Frederiksen 1980 (Text-Figure 7W), Brosipollis striata Frederiksen1988 (Text-Figure 7M), Milfordia hungarica(Text-Figure 7R), Lanagiopollis hadrodictyaFrederiksen 1988, Plicatopollis triradiata(Nichols 1973) Frederiksen and Christopher1978 (Text-Figure 7T), Ulmipolleniteskrempii (Anderson 1960) Frederiksen 1979and Retibrevitricolpites simplex Frederiksen1988 (Text-Figure 7X). Dicolpopollis spp.and Proxapertites spp. are reported byTschudy (1973a and b) and Frederiksen(1988) to have their last appearances in themiddle and late Eocene.

In addition to the tops observed in the GordonConfining Unit several species have firstoccurrences within this unit.Retibrevitricolpites simplex (Text-Figure 7X)first appears in the Congaree Formation.Other species that have bases in the Congareeand Warley Hill and Santee formations areJuglans nigripites Wodehouse 1933,Ulmipollenites undulosus Wolff 1934,Lymingtonia cf. L. rhetor Erdtman 1960(Text-Figure 7Q) in Frederiksen (1988),Symplocos? jacksoniana Traverse 1955,Gothanipollis cockfieldensis Engelhardt 1964,and Boehlensipollis hohlii Krutzsch 1962(Text-Figure 7A). Graminidites spp. isreported by Tschudy (1973) and Frederiksen(1988) to have a first appearance in the lateEocene.

Many of the first appearance datums or bases( FAD) and last appearance datums (LADs) ortops are shown in Text-Figures 5A and B.These figures summarize the results ofobservations of palynomorph taxa in the coresfrom the Savannah River Site and publisheddata. The principal publications on pollen andspores that were utilized in the study include:Elsik, 1974, Elsik and Dilcher, 1974,Frederiksen, 1978, 1979,1980a, 1980b, 1988,1991,1998, in press, Frederiksen andChristopher, 1978, and Tschudy 1973a, 1975.

CONCLUSIONS

Dinoflagellate and pollen biostratigraphy hasbeen applied successfully as a geologic toolfor helping establish the subsurfacehydrogeologic framework beneath the SRS.Because of the rapid vertical and lateralchanges in updip coastal plain setting,dinoflagellate, and pollen and sporebiostratigraphy, and to some extentdinoflagellate paleoenvironmentalinterpretations, have been critical in thecorrelation of aquifers and confining units andin the identification of facies changes andstructural elements (e.g., Pen Branch Fault,Aadland et al., 1995). Recognition of thesefeatures is important as they serve aspreferential pathways for groundwatermovement and contaminant transport.

REFERENCES

Aadland, R.K., 1990, HydrogeologicCharacterization of the Cretaceous-Tertiary Coastal Plain Sequence at theSavannah River Site, South Carolina. In:Zullo, V.A., Harris, W.B., Price, V., (eds.)1990, Savannah River Region: Transitionbetween the Gulf and Atlantic CoastalPlains, Proceedings of the Second BaldHead Island Conference on CoastalPlains Geology p. 57-60.

Aadland, R.K. And Bledsoe, H.W., 1992,Hydrogeologic Classification of theCretaceous-Tertiary Coastal PlainSequence at the Savannah River Site.Geological Society of America Abstractswith Programs, Vol. 24, No.2, p.1.

Aadland, R.K., Thayer, P.A., And Smits, A.D.,1992, The Hydrogeological Atlas of West-Central South Carolina. WestinghouseSavannah River Company, WestinghouseSavannah River Company TechnicalReport 90-987.

Aadland, R.K., Gellici, J.A., And Thayer,P.A.,1995, Hydrogeologic Framework ofWest-Central South Carolina. SouthCarolina Department of NaturalResources, Water Resources DivisionReport No. 5, 200 p.

Aurisano, R.W., 1989, Upper Cretaceousdinoflagellate biostratigraphy of the


B-8

subsurface Atlantic Coastal Plain of NewJersey and Delaware, U.S.A. Palynology,v.13, p.143-180.

Berggren, W.A., Kent, D.V., Swisher, C.C. Iii, andAubry, M.-P., 1995, Arevised Cenozoicgeochronology and chronostratigraphy: inBerggren, W. A., Kent, D.V., Aubry, M.-P., Hardenbol, Jan (eds.), Geochronology,Time Scales andGlobal StratigraphicCorrelation: Society of EconomicPaleontologists and Mineralogists,Special Publication No. 54, P. 129-212.

Cahil, J.M., 1982, Hydrology of the low-levelradioactive solid-waste burial site andvicinity near Barnwell, South Carolina.U.S. Geological Survey Open File Report82-863, 101 p.

Christopher, R.A., 1979, Normapolles and triporatepollen assemblages from the Raritan andMagothy Formations (Upper Cretceous) ofNew Jersey: Palynology, v.3, p. 73-121,15 text-figs., 1 table, 9 pls.

Clarke, J.S., Falls, W.F., Edwards, L.E.,Frederiksen, N.O., Bybell, L.M., Gibson,T.G., And Litwin, R.J., 1994, Geologic,Hydrologic, and Water Quality Data for aMulti-Aquifer System in Coastal PlainSediments Near Millers Pond, BurkeCounty, Georgia, 1992-93, GeorgiaDepartment of Natural ResourcesInformation Circular 96, 34p.

Clarke, J.S., And West, C.T., 1997, Ground-WaterLevels, Predevelopment GroundwaterFlow, and Stream-Aquifer Relations in theVicinity of the Savannah River Site,Georgia and South Carolina, United StatesGeological Survey, Water-ResourcesInvestigations Report 97-4197, 120p.

Doyle, J.A., And Robbins, E.I., 1977, Angiospermpollen zonation of the continentalCretaceous of the Atlantic Coastal Plainand its application to deep wells in theSalisbury embayment: Palynology, v. 1, p.43-78.

Edwards, L. E., 1990, Dinocysts from the LowerTertiary Units in the Savannah River area,South Carolina and Georgia. In: Zullo,V.A., Harris, W.B., Price, V., (eds.) 1990,Savannah River Region: Transitionbetween the Gulf and Atlantic CoastalPlains, Proceedings of the Second BaldHead Island Conference on CoastalPlains Geology, p.89-90.

Elsik, W.C., 1974, Characteristic Eocenepalynomorphs in the Gulf Coast, U.S.A.:

Palaeontographica, Abt. B, v. 149, p.90-111.

Elsik, W.C., and Dilcher, D.L., 1974, Palynologyand age of clays exposed in Lawrence claypit, Henry County, Tennessee:Palaeontographica, Abt. B, v. 146, p.65-87.

Engelhardt, D.W., 1964, A new species ofGothanipollis Krutzsch from the CockfieldFormation (middle Eocene) ofMississippi: Pollen et Spores, v.6 no. 2,p.597-600.

Erdtman, G., 1960, On three new genera from theLower Headon Beds, Berkshire:Botaniska Notiser, v. 113, p. 46-48.

Fallaw, W.C. And Price, V., 1992, Outline ofstratigraphy at the Savannah River Site.In: Fallaw, W.C., and Price, V., (eds.),1992, Carolina Geological Society FieldTrip Guide Book, Nov. 13-15, 1992,Geological Investigations of the centralSavannah River area, South Carolina andGeorgia: U.S. Department of Energy andS.C. Geological Survey, p. CGS-92-B-11-1-33.

Fallaw, W.C. And Price, V., 1995, Stratigraphy ofthe Savannah River Site and Vicinity.Southeastern Geology, V.35, No.1, p.21-58.

Falla, W.F., Baum, J.S., And Prowell, D.C., 1997,Physical Stratigraphy andHydrostratigraphy of Upper Cretaceousand Paleocene sediments, Burke andScreven Counties, Georgia, SoutheasternGeology, v.36, no. 4, p. 153-176.

Faye, R.E. And Prowell, D.C., 1982, Effects ofLate Cretaceous and Cenozoic faulting onthe geology and hydrology of the CoastalPlain near the Savannah River, Georgia,and South Carolina. U.S. GeologicalSurvey Open File Report 82-156.

Frederiksen, N.O., 1978, New Paleogene pollenspecies from the Gulf Coast and AtlanticCoastal Plains: U.S. Geol. Survey Journ.Research, 6: 691-696.1979, Paleogene sporomorphbiostratigraphy, northeastern Virginia:Palynology, v. 3, p. 129-167.1980a, Sporomorphs from the JacksonGroup (upper Eocene) and adjacent strataof Mississippi and western Alabama: U.S.Geol. Survey Professional Paper 1084,75p.1980b, Paleogene sporomorphs fromSouth Carolina and quantitative


B-9

correlations with the Gulf Coast:Palynology, v. 4, 125-179.1988, Sporomorph biostratigraphy, floralchanges, and paleoclimatology, Eoceneand earliest Oligocene of the eastern GulfCoast: U.S. Geol. Survey ProfessionalPaper 1448, 68p.1991, Midwayan (Paleocene) pollencorrelations in the eastern United States:Micropaleontology, v. 37,no. 2, p. 101-123.1998, Upper Paleocene and lowermostEocene angiosperm pollen biostratigraphyof the eastern Gulf Coast and Virginia:Micropaleontology, v. 44, no.1, p. 45-68.In Press, Lower Tertiary pollenbiostratigraphy of testholes in Georgianear the Savannah River Site: U.S. Geol.Survey Professional Paper.

Frederiksen, N.O. And Christopher, R.A., 1978,Taxonomy and biostratigraphy of LateCretaceous and Paleogene triatriate pollenfrom South Carolina: Palynology, v. 2, p.113-145.

Gohn, G.S. And Campbell, B.G., 1992, Recentrevisions to the stratigraphy of subsurfaceCretaceous sediments in the Charleston,South Carolina area. South CarolinaGeology, v.34, p.25-38.

Huddlestun, P.F. And Summerour, J.H., 1996, TheLithostratigraphic framework of theuppermost Cretaceous and lower Tertiaryof Eastern Burke County, Georgia,Georgia Geologic Survey Bulletin 127,94p.

Krutzsch, W., 1960, Ueber Thomsonipollismagnificus (Th. and Pf. 1953) n.f. gen. n.comb. und Bemerkungen zur regionalenVerbreitung einiger Pollengruppen imaelteren Palaeogen: FreibergerForschungshefte, v. C86, p. 54-65.1961, Beitrag zur Sporenpalaeontologieder praaoberoligozaenen kontinentalenund marinen TertiaerablagerungenBrandenburgs: Berichte der GeologischenGesellchaft in der DeutschenDemokratischen Republik, v. 5, p.290-343.1962, Stratigraphisch bzw. botanischwichtige neue Sporen -und Pollenformenaus dem deutschen Tertiaer: Geologie, v.11, no. 3, p.265-307.

Krutzsch, W. And Vanhoorne, R., 1977, DiePollenflora von Epinois und Loksbergen

in Belgien: Palaeontographica, Abt. B, v.163,p. 1-110.

Lewis, S. And Aadland, R.K.,1992, HydrogeologicSetting of the A/M Area: Framework forGroundwater Transport. WestinghouseSavannah River Company TechnicalReport 92-355, 80 p.

Leeth, D.C. And Nagle, D.D., 1996, Shallowsubsurface geology of part of theSavannah River alluvial valley in theupper Coastal Plain of Georgia and SouthCarolina, Southeastern Geology, v.36,no.1, p. 1-14.

Lucas-Clark, J., 1992a, Problems in Cretaceouspalynostratigraphy (dinoflagellates andpollen) of the Savannah River Site Area.In: Zullo, V.A., Harris, W.B., Price, V.,(eds.) 1992, Savannah River Region:Transition between the Gulf and AtlanticCoastal Plains, Proceedings of the SecondBald Head Island Conference on CoastalPlains Geology, p.74.

Lucas-Clark, J., 1992b, Some Characteristic fossilDinoflagellate cysts of Eocene Strata,Savannah River Site, South Carolina. In:Fallaw, W. and Price, V. 1992, GeologicalInvestigations of the Central SavannahRiver Area, South Carolina and Georgia.Carolina Geological Society Field TripGuidebook 1992, p. 58-60.

Lucas-Clark, J. And Van Pelt, R.S., 1993,Application of DinoflagellateBiostratigraphy to EnvironmentalRestoration Activities at the SavannahRiver Site, South Carolina, U.S.A., (abst.)Abstracts with Programs: FifthInternational Conference on Modern andFossil Dinoflagellates, Zeist, TheNetherlands, April 1993.

Marine, I.W. And Root, K.R., 1975, A groundwater model of the Tuscaloosa aquifer atthe Savannah River Plant. In: Crawford,T.V. (ed.) Savannah River LaboratoryEnvironmental Transport and EffectsResearch Annual Report, USERDAReport DP-1374, E.I. duPont de Nemours& Co., Savannah River Laboratory, Aiken,South Carolina.

Martini, E., 1971, Standard tertiary and QuaternaryCalcareous Nannoplankton Zonation,Proceedings of the II PlanktonicConference, Rome, 1970, p.739-785.

Nystrom, P.G., Jr., Willoughby, R.H. And Price,L.K., 1991, The Cretaceous and Tertiarystratigraphy of the South Carolina upper


B-10

Coastal Plain. In: Horton, J.W., Jr., andZullo, V.A., (eds.), Geology of theCarolinas. University of Tennessee Press,Knoxville, Tennessee, p.221-240.

Prowell, D.C., Christopher, R.A., Edwards, L.E.,Bybell, L.M., And Gill, H.E.,1985a,Geologic section of the updip CoastalPlain from central Georgia to westernSouth Carolina. U.S. Geological SurveyMiscellaneous Field Studies, Map MF-1737.

Prowell, D.C., Edwards, L.E., And Frederiksen,N.O., 1985b, The Ellenton Formation inSouth Carolina-A revised age designationfrom Cretaceous to Paleocene. U.S.Geological Survey Bulletin, 1605-A,p.A63-A69.

Siple, G.E., 1967, Geology and ground water of theSavannah River Plant and vicinity, SouthCarolina. U.S. Geological Survey Water-Supply Paper 1841, 113 p.

Sissingh, W.,1977, Biostratigraphy of CretaceousCalcareous Nannoplankton, Geologic enMijnbow, v.56, no.1, p.37-65.

Srivastava, S.K.,1972, Some spores and pollenfrom the Paleocene Oak Hill Member ofthe Naheola Formation, Alabama(U.S.A.): Rev. of Palaebot. and Palyn.v.14, p. 217-285.

Stover, L.E., Elsik, W.C. And Fairchild,W.W.,1966, New genera and species ofEarly Tertiary palynomorphs from theGulf Coast: Kansas Univ.Paleontological Contrib. Paper 5, p.10.

Thomson, P.W. And Pflug, Hans,1953, Pollen andsporen des mitteleuropaEischen TertiaErs:Palaeontographica, Abt. B, v. 94, p.1-138.

Traverse, A.F.,Jr.,1955, Pollen analysis of theBrandon lignite of Vermont: U.S. Bur.Mines Rept. Inv. 5151, p. 107.

Tschudy, R.H.,1973a, Stratigraphic distribution ofsignificant Eocene palynomorphs of the

Mississippi embayment: U.S. Geol.Survey Professional Paper 743-B, p. 24.1973b Complexiopollis pollen lineage inMississippi Embayment rocks: U.S.Geological Survey Professional Paper743-C, p. C1-C15.1975 Normapolles pollen from theMississippi Embayment: U.S. GeologicalSurvey Professional Paper 865, 42 p.

Van Pelt, R.S., Waanders, G.L., Lucas-Clark, J.,And Harris, M.K., 1993, The Applicationof Palynology to Waste-SiteCharacterization: An Example from theSavannah River Site, South Carolina,(Abst.) American Association ofStratigraphic Palynologists 1993 AnnualMeeting Programs and Abstracts, p.82.

Van Pelt, R.S., And Lucas-Clark, J., 1996, TheApplication of Palynology toEnvironmental Restoration Activities: AnExample from the Savannah River Site,South Carolina, (Abst.) NinthInternational Palynological CongressPrograms and Abstracts, June 1996, p.82.

Wodehouse, R.P., 1933, Tertiary pollen. II The oilshales of the Eocene Green Riverformation: Torrey Bot. Club Bull., v.60,p.479-524.

Wolfe, J.A., 1976, Stratigraphic distribution ofsome pollen types from the Campanianand lower Maastrichtian rocks (UpperCretaceous) of the Middle Alantic States:U.S. Geological Survey ProfessionalPaper 977, p. 1-18.

Wolff, H., 1934, Mikrofossilien des pliocaenenHumodils der Grube Freigericht beiDettingen a.M. und Vergleich mit alterenSchichten des Tertiaers sowieposttertiaeren Ablagerungen: Preuss.Geol. Landesanstalt, Inst. Palaeobotanikund Petrographie Brennsteine Arb., v.5,p.55-86.

Carolina G

eological Society, 2000 Annual F

ield Trip G

uidebookW

SRC

-MS-2000-00606

B-11

Figure 1. Location of the study area, Savannah River Site (SRS). Palynological data have been obtained for all or part of thewells shown on the map (modified from Fallaw and Price, 1995).

Carolina G


ield Trip G

uidebookW

SRC

-MS-2000-00606

B-12

Figure 2. Comparison of chronostratigraphic, lithostratigraphic, and hydrostratigrahpic units in the SRS region: IRM=IrwintonSand member; CF=Clinchfield Formation; RMM=Riggins Mill member; BBU=Blue Bluff unt; UM=Utley member; GLM=GriffinsLanding member; ODB=Orangeburg District bed.


B-13

Figure 3. Cross section of the Uper Cretaceous and Tertiary hydrostratigraphic sequencebeneath the SRS (modified from Aadland et al., 1995).


B-14

Figure 4. Selected Cretaceous palynomorph datums. The relatonship of formationalboundaries to stage boundaries is modified following Berggren et al., 1996.


B-15

Figure 5A. Selected Tertiary palynomorph (first appearance) datums. The relationshp offormational boundaries to stage boundaries is modified following Berggren et al., 1996:IRM=Irwinton Sand Member; CF=Clinchfield Formation; RMM=Riggins Mill Member;BBU=Blue Bluff Unit; UM=Utley Member; GLM=Griffins landing Member;ODB=Orangeburg District Beds.


B-16

Figure 5B. Selected Tertiary palynomorph (last appearance) datums. The relationshp offormational boundaries to stage boundaries is modified following Berggren et al., 1996:IRM=Irwinton Sand Member; CF=Clinchfield Formation; RMM=Riggins Mill Member;BBU=Blue Bluff Unit; UM=Utley Member; GLM=Griffins landing Member;ODB=Orangeburg District Beds.


B-17

Figure 6. Selected Cretaceous pollen used in hydrostratigraphic correlation. All figures are 800 x, except BB, which is 400 x(reduced 25% for reproduction): (a) Porocolpopollenites sp.(sensu Doyle, 1969); (b) Complexiopollis abditus Tschudy 1973; (c)Pseudoplicapollis longion- nulata Christopher 1979; (d) Praecursipollis plebus Tschudy 1975; (e) Osculapollis Sp.; (f) Trudopollissp.; (g) Plicapollis sp.; (h) Labrapollis sp.; (i) Plicapollis usitatus Tschudy 1975; (j) Proteacidites sp.; (k) Echimonocolpites sp. 45(Cl-45); (1) Echimonocolpites sp. 46 (Cl-46); (m,n) Holkopollenites chemardensis Fairchild in Stover, Elsik and Fairchild 1966; (o)Holkopol- lenites sp. A (=CP3d-1 of Wolfe, 1976); (p) Santalacites minor Christopher 1979; (q) Momipites sp.; (r) Caryapollenitessp.; (s) Betulaceoipollenites sp. (=NO-3 of Wolfe, 1976); (t) ?Holkopol- lenites sp. (CP3E-1 of Wolfe, 1976); (u) Pseudoplicapollisendocuspis Tschudy 1975; (v) Holkopol- lenites sp. B (=CP3D-2 of Wolfe, 1976); (w) Osculapollis aequalis Tschudy 1975; (x)Libopollis jarzenii Farabee, Daghlian, Canright and Oftedabi 1984; (y) Baculostephanocolpites sp. B (= MPH-2 of Wolfe, 1976); (z)Mornipites dilatus group (=NP-2 of Wolfe, 1976); (aa) Plicatopollis cretacea Frederiksen and Christopher 1979; (bb)Rugubivesiculites sp.


B-18

Figure 7. Selected Tertiary pollen used in hydrostratigraphic corrlation. All figures are 625 x, except fotr (p), which is 1500 x(reduced 25% for reproduction: (a) Boehlensipollis holii Krutzsch 1962; (b) Pseudoplicapollis limitata Froderiksen 1978; (e)Thomsonipollis magnificus (Pflug in Thomson and Pflug) Krutzsch 1960; (d) Nudopollis terminals (Thomson and Pflug in Thomsonand Pflug, 1953); (e) Favitticolpozites baculoferous (Thomson and Pflug) Pflug 1953; (f) Sub- tziporopollenites nanus (Thomson andPflug in Thomson and Pflug, 1953); (g) Choanopollenites discipulus Tschudy 1973; (h) Lusatispozis indistincta Froderikson 1979; (i)Nudopollis thiergartii (Thomson and Pflug) Pflug 1953; (j) Trudopollis plenus Tschudy 1975; (k) Tilcolpites asper Froderiksen 1978;(1) Piolencipollis endocuspoides Froderikson 1979; (m) Bz-osipollis striata Froderiksen 1988; (n) Tricolpites crossus Froderiksen1979, (o) Pseudoplicapollis serenus Tschudy 1975; (p) Pseudoplicapollis serenus Tschudy 1975 1500 x; (q) Lymingtonia cf. L. rhetorErdtman 1960 in Froderiksen 1988; (r) Milfordia hungarica (Kedves, 1965) Krutzsch and Vanhoome 1977; (s)Intratziporopollenites stavensis Froderiksen 1980; (t) Plicatopollis tdradiata (Nichols 1973) Frederiksen and Christopher 1978; (u)Bombacacidites reticulatus Krutzsch 1961; (v) Quadmpol- lenites vagus Stover in Stover et a]., 1966; (w) Malvacipollis tschud@di(Frederikson 1973) Frederiksen 1980; (x) RetibzvWtricolpites simplex Froderikson 1988.


B-19

Figure 8. Selected Tertiary dinoflagellates used in hydrostratigraphic correlation. All figures are 500 x. (reduced 25% forreproduction): (a) Carpatelia cornuta Grigorovich 1969; (b) Senegalinium ?dil"nense (Cookson and Eisenack) Stover and Evitt1978; (c) Senegaliniurn bicavaturn Jain and Millepied 1973; (d) Adnatosphaeridium multispinosum Willianis and Downie 1966; (e)Wet- zelialia articulate Eisenack 1938; (f) Glaphyrocysta ?vicina (Eaton) Stover and Evitt 1978; (g) Dracodinium sp.; (h)Apectodinium homomorphum (Deflandre and Cookson) Harland 1979; (i) Pentadinium goniferum Edwards 1982; (i) Samlandiachlamydophora Eisenack 1954; (k) Palaeotetradinium minusculum (Alberti) Stover and Evitt 1978; (1) Hafniasphaera septata(Cook- son and Eisenack) Hansen 1977; (m) Wetzelielia cf. W. aiticulata Eisenack 1938; (n) Pentadinium laticinctum Gerlach 1961;(c)) Wetzelielia sp.; (p) Lentinia sp.; (q) Chariesdowniea variabilus (Bujak) Lentin and Vozzhennikova 1989; (r) Damassadiniumcalifornicurn (Drugg) Fensome et al., 1993.


B-20

Notes


C-1

Methodology and Interpretation of the Piezocone Penetrometer TestSounding for Estimating Soil Character and Stratigraphy at the SavannahRiver Site

Frank H. Syms, Bechtel Savannah River, Inc.Douglas E. Wyatt, Westinghouse Savannah River CompanyGreg P. Flach, Westinghouse Savannah River Company

PREFACE

Experience with the Piezocone Penetration Test (CPT) technology, and site-specific correlationsdeveloped at the Savannah River Site (SRS), have resulted in a methodology that makes the CPT aviable tool for geologic mapping. This methodology has evolved through the collective interest andcontributions of numerous investigators. This paper will review the test equipment as well as presentongoing research for interpretation of upper Coastal Plain sediments at the SRS using the CPT.

INTRODUCTIONThe piezocone penetration test (CPT) hasbecome a common tool for subsurfacecharacterization at the Savannah River Site(SRS) due to the cost and quality benefits ofthe technology. In the last 8 years, over 800CPT soundings have been pushed for variousgeotechnical, environmental and geologicinvestigations (Figure 1). Many of thesesoundings reach depths greater than 60 metersproviding a vast amount of data for the middleand upper Eocene (and younger) sediments.Use of the CPT for geologic mapping hasevolved from gained experience, provenreliability and correlation capabilities of thetechnique. Because the CPT was originallydeveloped for engineering applications, thetechnology was not broadly exposed togeologists until its use became more commonon environmental projects in the late 1980’s.The increased use and popularity of the CPTcan be attributed to several qualities including:• Standard equipment and test procedure• Depth control on the scale of centimeters• Multiple measurement parameters• Highly repeatable, nearly continuous,

electronic data• Depth of penetration• Cost effectivenessDirect correlation of CPT measurements tostratigraphic units within the upper 60 metersat the SRS has been demonstrated onnumerous investigations (Ref WSRC, 1994,

WSRC, 1995, Wyatt, et al, 1997). Severalinvestigations at the SRS beginning in theearly 1990’s, used closely spaced CPTsoundings and soil borings for localstratigraphic correlation. From this data, itbecame evident that CPT measurements canbe used not only for determining local

$$$

$$$$$$$$$$

$$$$$$

$$$$$$$$$$ $$$$$$$$

$$$$$$$$$$$$$$$

$$$$$$$$$$

$$$$$

$$

$

$$$$$

$$

$

$$$$$$$$$$$$$$$$ $$$$$$$$$$$

$$ $$$$$$$$$

$

$

$

$ $

$$$$$$$$$$$$$$$$$$$$$$

$$$$$ $$$

$$$$$$$$

$$$$

$$$$$$$ $$$

$$$

$

$$

$

$$$$$

$

$$ $ $$$

$

$$$$$$$$$$

$

$ $$$$$$$$$$$

$$$$$$

$$$$$$$

$$$$$$

$

$$$

$$$

$$$

$$$

$

$$$$

$$$

$

$$$$$$

$$$$$

$$$$$$$ $$

$$$$$$$

$$

$$$$$$$$$

$$$$ $$$$$$$$$$$$$$$$$$

$$$$$

$$$

$$$$

$$$$$$$$$$$$$$

$$

$$$$$$$$$$$$$$$$$$$$$$$$$$

$

$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$

$$

$

$$$$$$$$$$$$$$$$$$$$

$$

$$$$$$$$$$

$$

$

$

$$ $$$$$$$$$$$$$$$$$$$$

$$$$$$$$ $$

$$$$$$$$$

$

$$$$

$

$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$

$

$$

$$$$$$$$$$$$$$$$$$$$$$$

$$$$$$$$$$

$$$$

$$$$$$

$$$$$$$$$

$$$$ $

$

$

$

$ $

$$

$$$$ $$

$ $$

$$

$$$$$$$$$$

4 0 4 8 Kilometers

$$$

$$$$$$$$$$$$

$ $$$$$

$

$

$$

$$

$ $ $

$ $$$$$$

$

$

$$$$$$$$

$$$$$$$

$$$$$$$

$$$$

$ $$

$

$$

$$$$$$

$$$$

$$$$$$

$ $$

$

$$

$

$

$$

$

$$$$$

$

$$

$

$

$$

$

$$$$$$$$$$$$

$$$$$$$$$$

$$$$$$

$

$$

$$

$$

$$$$

$$

$

$$$

$$

$

$$

$

$

$$

$

$$

$$

$

$$

$

$$

$$

$$

$ $$

$$

$$$

$

$$

$ $$

$

$

$ $$

$ $

$ $

$$$$

$$$$$$$$$

$$$

$$$

$$$

$$$$$

$

$$$$$

$$$$$$$$$$$$$$$$$$$

$$$$

$$$$$$$$$$

$

$ $

$

$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$

$

$$

$$$$$$$

$$$

$

$

$$

$$$$ $$

$$$

$

$

$

$ $

$

$

$$

$$$

$

$$$

$

$

$$$$$$$$$$

$$$

$$$$$$$$$$

$$$

$

$$

$

$

$$$

$$$

$

$$$

$$$

$$

$

$$

$$$$

$

$$

$

$$ $

$

$

$

$

$$

$$

$

$

$

$

$

$$$$$$$$$ $$$

$$

$

$$$$$$$$

$$

$

$$

$

$

$

$$

$

$

$$

$$

$

$

$$

$

$$$$$$$$$$$$$$

$

$ $

$

$

$$$$

$$$

$$

$

$

$$$

$$

$

$$$

$$$$

Figure 1

Site layout map o f the SRS showing clusters o f CPT loca tions. Note these CPT's are l imited to sound ings greater than 30 meters.

N

Figure 1. Distribution of CPT soundingsaround the SRS.

subsurface layering, but also to describe othergeologic characteristics of the sediments suchas vertical sedimentation sequences and faciesdistribution. This paper discusses the standard


C-2

CPT equipment in use at the SRS as well asproposed methods of geologic interpretation.

TEST METHOD AND EQUIPMENT

The first generation CPT’s were developed inthe 1930’s for investigating the bearingstrength of fairly shallow, loose, sensitive soilsnative to Holland. These early mechanicaltools eventually evolved into the electro-mechanical devices in use today. The testequipment and procedures in use at the SRSare controlled by American Standard TestMethod (ASTM) D5778, Standard TestMethod for Performing Electronic FrictionCone and Piezocone Penetration Testing ofSoils.

The CPT discussed here is a quasi-static,electro-mechanical probe that measures threeseparate parameters with depth (see Figure 2).These include tip resistance (qc), sleeveresistance (fs) and pore pressure (u). A fourthparameter, friction ratio (Rf) is computed bydividing the sleeve friction by the tipresistance and multiplying by 100.Measurements are acquired at roughly 1second intervals. At a penetration rate ofabout 2cm per second, a vertical resolution ofabout 2cm is achieved. Forces are sensed byload cells and a pressure transducer. Data aretransmitted from the probe assembly via acable running through the push rods to thedata acquisition system at the surface. Theanalog data are digitized and recorded bycomputer in the penetrometer truck. Thetypical CPT test rig used at the SRS consistsof a hydraulic load frame mounted on a heavytruck with the weight of the truck providingthe necessary reaction mass (see Figure 3).

INTERPRETATION OF CPT DATA

Interpretation of CPT data is similar tomethods used for down-hole geophysical logsi.e., measurements can be related to somecharacteristic of the sediment. For the SRS,the relative amount of fine grained soils (finescontent) is the most useful for engineering andhydrogeologic mapping. The vertical

resolution of CPT provides well definedboundaries where changes in lithology may beinferred. Discrete sedimentation sequencessuch as fining or coarsening sequences mayalso be defined.

Umbilical Cable

Friction Sleeve

Strain Gauge

Load Cell

Pore Pressure Transducer

Conical TipPorous Filter(Piezo Element)

Type 2 Shoulder Cone

Figure 2. Type 2 Shoulder Cone CPTprobe.

Basic CPT Response Characteristics

Each sensor on the CPT probe provides aseparate measurement of the soil properties asthe probe is being advanced. In general, tipresistances are relatively higher in sands anddecrease as the fines content increases. Thetip resistance is considered to be one of themost reliable CPT parameters. Inversely,sleeve resistances are relatively lower in sandsand increase as the fines content increases.Sleeve resistances are acquired over largersurface areas behind the tip and are consideredto be less reliable and of lower resolution.Pore pressures are lower in sands and higherin finer grained soils due the highertransmissivity of sands. The pore pressureelement measures an induced pressure as theprobe is advanced and is referred to as adynamic pore pressure. Therefore, in finergrained sediment, the induced pressures willtend to be higher and dissipate slower than a


C-3

more porous (sandy) sediment. Theserelationships have shown that as a generalrule, friction ratio increases as fines contentand plasticity increase (Lunne, et al, 1997).As shown in Figure 4, fines content resultsfrom sieve analysis plotted against frictionratio values support this general trend,however there is significant scatter in the data.These CPT characteristics are general butprovide a fundamental basis for datainterpretation although numerous exceptionscan be demonstrated for SRS soils.

Several published CPT soil classificationcharts have been developed based ontheoretical relationships as well as empiricaldata. However, many factors influence CPTmeasurements resulting in inconsistentperformance of these charts for SRS soils. Anunderstanding of the influences that can affectthe measurements is required for properinterpretation.

Factors Affecting CPT Measurements

Several factors can affect CPT measurementsand should be considered before interpretingthe data. Several of these

Figure 3. Typical CPT rig used at the SRS.

factors have been mitigated at SRS bystandardizing the type of CPT probe used,calibration requirements, and data acquisitionand processing parameters. However, otheraffects from geologic conditions such as layertransition, aging, cementation, weathering,consolidation and in-situ state of stress, mayresult in variations in the CPT measurementsthat if not corrected for, will result inmisleading

Figure 4. Friction ratio versus measuredpercent fines.

interpretations. Experience and familiaritywith the CPT, as well as understanding thesubsurface conditions of the site, is paramountto assessing data quality. One selectedreference which discusses these affects as wellas methods of interpretation is Lunne, et al,(1997).

Applicability of Soil Classification Charts

The most common classification chartparameters use the relationship of tipresistance (qt) and friction ratio (Rf) to predicta soil type. The most recent charts use theseparameters normalized for stress effects (Qt

and Fr) as well as normalized pore pressure(Bq).

An evaluation of 11 published CPT soilclassification charts using high quality CPTsoundings paired with adjacent soil boringsdemonstrated inconsistent predictions of soiltype for SRS sediments. Figures 5 and 6 showthe distribution of sieve test results correlatedto adjacent non-normalized and normalizedCPT measurements, respectively. Indicated onthese charts, however, is the tendency of quitedifferent fines contents to co-populate areas onthe charts. This demonstrates that the

0

1

2

3

4

5

6

7

8

9

10

0 10 20 30 40 50 60 70 80 90 100

Fines Content

Fri

ctio

n R

atio

(R

f)


C-4

relationship of these CPT parameters forpredicting a soil type is not unique. This islikely due to various subsurface conditions asdiscussed previously. The degree of this issuevaries for all classification charts but is mostprominent for soils that have moderate tosignificant ratios of fines to sand-sizedconstituents.

Soil type alone does not provide adequateresolution of sediment gradation to be usefulfor interpreting sedimentation sequences andsequence boundaries. Therefore, a method topredict gradation (sand versus fines content)directly from CPT measurements has beendeveloped from site-specific correlations at theSRS.

Figure 5. Fines content distribution plottedagainst correlating non-normalized CPTmeasurements.

RESULTS

Proposed Method for Estimating GradationChanges from the CPT

To interpret sedimentation sequences fromCPT measurements requires a method to relatethe CPT measurements to changes ingradation. Closely spaced (nominal 10 feet)CPT soundings combined with soil boringshave been used to develop a site-specificrelationship between fines content (defined asthe combined percent by weight of silt andclay passing a No. 200 mesh sieve) and CPTmeasurements (Qt and Fr). This allowsresolution vertical gradation trends for SRSsediments.

Figure 6. Fines content distribution plott-ed against correlating normalized CPTmeasurements.

The basis for this approach follows amethodology suggested by Robertson and Fear(1995). Two baseline equations, presentedbelow, for estimating a Classification Index(Ic) and ultimately fines content (FC)(Jefferies and Davies, 1993 and Robertson andFear, 1995, respectively) were modified usingdata acquired from the CPT and soil boringpairs.

Ic=[(3.47-logQt) 2+ (logFr+1.22)2]0.5

From Ic:

FC%=1.75Ic3-3

Where: Qt=(qt-σvo)/σ’vo

Fr=[fs/(qt-σvo)]X100%σvo - Total overburden stressσ’vo - Effective stress

Significant improvement was achieved bydeveloping stratigraphic specific relationships.Currently, three modified equations areproposed for each layer and are discussedbelow. However, these equations arecurrently being re-evaluated based onadditional data and analytical techniques.

1

1 0

1 0 0

1 0 0 0

0 2 4 6 8 1 0

Fr ic t io n R a t io ( R f )

Tip

Res

ista

nce

(qt) 0 - 1 0 % f in e s

1 1 - 2 0 % f in e s

2 1 - 3 0 % f in e s

3 1 - 4 0 % f in e s

4 1 - 5 0 % f in e s

> 5 0 % f in e s

1

10

100

1000

0.10 1.00 10.00

Friction Ratio (Fr)

Tip

Res

ista

nce

(Q

t) 0-10 % fines

11-20 % fines

21-30 % fines

31-40 % fines

41-50 % fines

>50 % fines


C-5

Stratigraphic Interpretation and Predictionof Gradation Trends

The high vertical resolution, repeatability andmulti-parameter measurements of the CPTmake it an ideal stratigraphic mapping tool forunconsolidated sediments. One of the primaryuses and advantages of the CPT is definingsubsurface layering of the investigation site(Schmertmann, 1978 and Robertson andCampanella, 1988). Figure 7 shows a cross-section developed from CPT soundings. Thehigh vertical resolution of the CPT is obviousbut more importantly the continuity ofmeasurements (repeatability) betweenpenetrations is what makes the techniqueuseful for subsurface mapping.

Figure 8 represents a CPT sounding andadjacent soil boring. Also shown on thisfigure is a fines content profile produced fromthe stratigraphic Ic and FC equations.Measured fines content values determinedfrom sieve test results from the adjacent soilboring are plotted for comparison.

SRS Stratigraphy

Altamaha

As shown on Figure 8, the upper portion of theprofile represents the Altamaha Formation.This unit consists of massive clayey sandswith varying amounts of fine and coarse-grained sediments, but typically contains amoderate to high amount of fines (greater than30 percent).

The CPT response can be described asmoderate to high tip (50-300 tsf) and sleeve(2-5 tsf) resistances and a resulting highfriction ratio (3 percent or greater). High porepressures are noted as well even though thegroundwater table is much deeper. These CPTresponses are typical for overconsolidatedsoils, a known condition of the Altamaha (RefWSRC, 1995). The overconsolidated statetends to yield higher sleeve friction valuesthan would be expected for correspondinghigh tip resistances. This can also account forthe high pore pressures as well.

The proposed equations for the Altamaha areas follows:

IcA=[(3-logQt) 2+ (logFr+1.5)2]0.5

FCA=(1.75(Ic)3.5)-6

Tobacco Road/Dry Branch

As shown on Figure 8, the middle portion ofthe profile corresponds to the Tobacco Roadand Dry Branch. These units consist ofnormally to slightly overconsolidatedclayey/silty sands with most of the clay boundas matrix material. Fines content of theTobacco Road typically range from low tomoderate (5-25 percent). The Dry Branchfines content is generally less than theoverlying Tobacco Road Formation yet theoccurrence of clayey beds are more massiveand can form confining units locally.

This unit can be distinguished from theoverlying Altamaha by the reduced sleeveresistance and pore pressure measurements.CPT measurements can be described asmoderate tip resistances (100-300 tsf) whereincreasing resistance with depth can be noted.Sleeve resistance is generally low to moderate(1-2 tsf). Friction ratios are uniform andrelatively low except where higher valuesresult from decreases in tip resistances. Porepressure response is generally moderate (at, orslightly above, hydrostatic pressure) withhigher measurements noted in zones withhigher friction ratios.

The proposed equations for the TobaccoRoad/Dry Branch are as follows:

IcTD=[(2-logQt) 2+ (logFr+1.7)2]0.5

FCTD=(1.75(Ic)3.5)

Santee/Tinker

The lower portion of the profile consists of theSantee/Tinker as shown on Figure 8. Thisportion of the section represents one of themost lithologically complex units in theshallow subsurface of the SRS. Thesesediments were deposited under near shore


C-6

marine environments resulting in varyingamounts of carbonate bearing sediments.Cementation and/or dissolution have occurredirregularly throughout this unit. The Santeecan be described as silty to clayey fine sands.

CPT measurements can be described as low tohigh tip resistances (20-300 tsf) and moderateto high sleeve resistances (2- 4 tsf). Frictionratios are moderate to high (2-4 percent) inlow tip resistance intervals and low (1-1.5percent) in high tip resistance intervals. Porepressure measurements are generally erraticbut are high in the high friction ratio intervalsand around or slightly below hydrostatic in thehigh tip resistance intervals.

Proposed equations for the Santee/Tinker areas follows:

IcS=[(1.5-logQt) 2+ (logFr+1.7)2]0.5

FCS=(1.75(Ic)3.5)-1

Recommended Approach to EstimatingFines Content

1. Care must be taken to assure the CPTmeasurements are consistent for a givensite. This requires a subjective review ofthe data to assess the reported qt, fs, Rf andu measurements.

2. The CPT data must consist of piezoconemeasurements with proper input for thegroundwater table. The data must benormalized as specified in the Ic equations.

3. Specific stratigraphic layers (Altamaha,Tobacco Road, Dry Branch, Santee) mustbe defined where CPT measurementsshow similar characteristics (refer to thediscussion of shallow stratigraphy andCPT measurements). Differentrelationships should be expected indifferent parts of the SRS.

4. The application of the proposed Icequations are for guidance and must beadjusted based on site specific sieve data.

5. For a given site, the continuity ofstratigraphic CPT layers must be assessedand used as guidance to determinereasonable sample intervals and locationsto verify the Ic equations.

SUMMARY AND APPLICATION

The CPT provides many advantages forsubsurface mapping. However, interpreta-tionof the data requires an initial understanding ofthe subsurface conditions and testingparameters. Correlation with adjacent soilborings has provided a means to estimategradational trends directly from CPT data.This is a significant advancement forengineering and hydrogeologic applica-tion.

The ability to better predict the amount offines can be used to quantitatively describesubsurface units. In the case of hydrogeologicmapping, units can be mapped definitively asmore fine grained (less transmissive) or moresandy (more transmissive). For engineeringapplication, more compressible (fine grained)units can be distinguished from more lesscompressible (sandy) units. Also, the abilityto estimate a fines content allows a UnifiedSoil Classification System (USCS) soil typebe assigned.

ADVANCEMENTS AND FUTURETRENDS

One recent advancement in CPT technologyused at the SRS is the CPT sampler. In softintervals where traditional drilling techniqueshave generally failed in recovery of such softmaterials, the CPT sampler has provedrepeatedly to have a nearly perfect recoveryrate. Another advancement being tested forapplication is down-push video. Severalsoundings on site have been pushed usingcameras with various focal lengths providingmagnified images with on-screen scales.Down-push geophysical logs (resistivity andnatural gamma) coupled with the standardCPT measurements are becoming standard testsuites. Environmental probes are numerousand include water and gas samplers as well as


C-7

in-situ characterization/detection systems. Useof the CPT will continue to grow not only forenvironmental and engineeringcharacterization and will become an additionaltool for geologic mapping in unconsolidatedsediments.

REFERENCES AND SELECTEDBIBLIOGRAPHY

Jefferies, M.G., Davies, M.P., 1993, “Use of theCPTu to Estimate Equivalent SPT N60”,Geotechnical Testing Journal, GTJODJ, Vol. 16,No. 4, December, pp 458-468.

Lunne, T., Robertson, P.K., and Powell,J.J.M.,(1997) “Cone Penetration Testing inGeotechnical Practice”, Blackie Academic andProfessional, U.K.

Robertson, P.K., Fear, C.E., 1995, “Liquefaction ofSands and it’s Evaluation”, IS TOKYO 95, FirstInternational Conference on EarthquakeGeotechnical Engineering, Keynote Lecture,November.

Robertson, P.K., and Campanella, R.G., (1983a)“Interpretation of the Cone Penetrometer Test; PartI: Sand.” Canadian Geotechnical Journal, 20(4).718-733.

Robertson, P.K., and Campanella, R.G., (1983b)“Interpretation of the Cone Penetrometer Test; PartII: Clay.” Canadian Geotechnical Journal, 20(4).734-745.

Robertson, P.K., and Campanella, R.G., (1988)“Guidelines for Geotechnical Design using theCPT and CPTU.” University of British Columbia,Vancouver, Dept. of Civil Engineering, SoilMechanics Series 120.

Schmertmann and Crapps, Inc. (1988) “Guidelinesfor using the CPT, CPTU and Marchetti DMT forGeotechnical Design. Volume II – Using CPT andCPTU Data.” PB88-211644.

U.S. Dept. of Transportation, (1992) “The ConePenetrometer Test”, Pub. No. FHWA-SA-91-043.

WSRC (1994), “Geological Characterization in theVicinity of the In-Tank Precipitation Facility,”WSRC-TR-94-0373, Rev. A, October.

WSRC (1995) “In-Tank Precipitation (ITP) and H-Tank Farm (HTF) Geotechnical Report,” WSRC-TR-95-0057, Rev. 0, September.

Wyatt, D.E., Cumbest, R.J., Aadland, R.K., Syms,F.H., Stephenson, J.C., and Sherrill, J.C., 1997,Investigation on the Combined Use of GroundPenetrating Radar, Cone Penetrometer and HighResolution Seismic Data for Near Surface andVadose Zone Characterization in the A/M Area ofthe Savannah River Site, WSRC-RP-97-0184.


C-8

110

120

130

140

160

170

180

190

210

220

230

240

260

270

280

290

100

150

200

250

300

0

200

400

Tip (COR)(tsf)

0 50 100 150 200 250 300 350Distance (feet)

0246

Sleeve(tsf)

Ele

vatio

n (fe

et)

0

200

400

Tip (COR)(tsf)

0246

Sleeve(tsf)

0 10 20 30 40 50

% Fines

0

200

400

Tip (COR)(tsf)

02468

Sleeve(tsf)

0 10 20 30 40 50

% Fines

0 20 40 60 80 100

N-Value

APSF SectionSouth North

d:/gelinas/cpt/fpair/sect-2j.grf 09-29-98uses apsfsect.xls sheet section2_whitepaper

CPT-36, FB-6 CPT-1 CPT-14 CPT-10; FB-3FB-10

TR

UC

K D

OC

K

MAA

0

200

400

Tip (COR)(tsf)

02468

Sleeve(tsf)

% Fines

Approximate Water Table

Scale for % Fines

Scale for CPT Tip (COR)Stress, Resistance in TSF

Scale for CPT SleeveStress, Resistance in TSF

KEY:

AL

TR

DB

TCI

DB

ST

WH

CG

AP

SF

Geo

logi

cU

nits 110

120

130

140

160

170

180

190

210

220

230

240

260

270

280

290

100

150

200

250

300

0 50% Fines

Figure 7. Cross-section constructed from CPT soundings. Note the continuity and repeatabilityof measurements. The plotted parameters include CPT tip resistances to the right of the axiswith sleeve resistances plotted inversely to the left. Also, fines content values determined fromsieve tests from adjacent borings are plotted along the tip resistances. The stratigraphic columnon the right denotes the Altamaha (AL), Tobacco Road (TR), Dry Branch (DB) including theTan Clay Interval (TCI) and Santee (ST).


C-9

Figure 8. CPT sounding profile with adjacent boring data. Note the sharp breaks atstratigraphic boundaries. Fines content values determined from sieve tests are plotted as reddots. A calculated fines content curve was generated from applying the proposed Ic and FCequations.


C-10

Notes


D-1

Contaminant Transport and Hydrologic Properties of Fault-PropagationFolds in Poorly Consolidated Coastal Plain Sediments

R. J. Cumbest and D. E. Wyatt, Westinghouse Savannah River Co., Aiken, SC 29808

INTRODUCTION

The Atlantic margin Coastal Plain at regionalscales is commonly characterized as a thin,seaward thickening wedge of poorlyconsolidated sediments that overly rigidcrystalline, and indurated sedimentary rocks.Primary structure in the Coastal Plainsedimentary cover is characterized by a gentle(a few degrees) seaward dip. In many casesand certainly at regional scales this is anaccurate and useful approximation. However,it has previously been recognized that locallythe sediments of the Coastal Plain recordsecondary structure due to tectonicdeformations that are primarily shallowexpressions of high angle faulting in theunderlying rigid basement rocks (Howell andZupan, 1974; Prowell and O’Connor, 1978;Prowell, 1983, York and Oliver, 1976; Cramerand Arden, 1978; Bramlett and others, 1982,Mixon and Newell, 1982; Prowell, 1988,Domoracki, 1995). Because of the way thebasement deformation is partitioned in thepoorly consolidated Coastal Plain sedimentarysequence, recognition of this deformation inthe near surface is difficult and in some casesmay be so subtle as to require high resolutiontechniques that image most of the coversequence and possibly the basement surfaceitself. This is because the high angle faultingin the rigid basement rocks is propagated as adiscrete offset only in the lower parts of thesedimentary package. Above this thestructural deformation is accommodated byfault-propagation folding, which by the time itreaches the surface may result in dipanomalies of only a few degrees. Theserelatively minor deformations in the shallowsection would be extremely difficult torecognize in near surface geologiccharacterizations such as those commonlyperformed in most environmentally related

studies. However, these relatively smalldeformations may have significant effects onboth hydrologic behavior and contaminantmigration properties of Coastal Plainsediments. In consequence, recognition andcareful characterization of these near surfaceexpressions of Coastal Plain faulting may beextremely useful in designing effective andefficient remediation efforts in settings wherethese types of structures are present.

Fault Propagation Folding in LooselyConsolidated Cover above High AngleBasement Reverse Faults

Fault-propagation folding has been analyzedand discussed by Mitra, 1993, Narr andSuppe, 1994, and McConnell, 1994.McConnell, 1994, described several generalcharacteristics (Fig. 1) of fault-propagationfolds based on extensive field observations:(1) Fold axial surfaces intersect faults at or

near the basement cover contact anddiverge up-section.

(2) Fold hinges are typically narrow andangular but may become more open up-section.

(3) Heterogeneous thickness changes resultin thickening in syncline hinge zonesand thinning of forelimbs adjacent toanticline axial planes.

(4) Foot wall synclines are typicallypreserved so that faults cut steep foldforelimbs or are located near anticlineaxial planes.

McConnell (1994) also developed a kinematicmodel for the geometry and evolution of thesestructures that gives quantitative relationshipsfor the major geometric elements. Inparticular, McConnell (1994) gives the


D-2

following relationships, which are importantto this discussion:

γs=tan-1[1/(2cot α - cot γa)] (1),

relates the exterior angle of the foot wallsyncline axial surface (γs) to the exterior angleof the hanging wall anticline axial surface (γa)and the dip of the fault (α).

β =tan-1[H/(u cot γs - u cot γa - H cot α)] (2),

gives the angle of inclination of any un-faulted bed as a function of the structuralrelief on the basement surface (H), thedistance above the basement surface (u), andthe exterior angles of the axial surfaces andthe fault.

Th/To = sin γa’/sin γa (3)

Tf/To = sin γs’/sin γs (4)

relate the relative thickness changes of faultedunits in the hanging wall (Th/To) and foot wall(Tf/To) to the interior and exterior inter-limbaxial angles (γa), (γa’), (γs), and (γs’). Theseequations can be powerful tools for predictingspecific geometric elements of fault-propagation folds based on their inter-limbangles or other easily measured parameters. Inaddition, McConnell (1994) gives thefollowing relationship between the length todisplacement ratio for the fault, defined asrelative stretch (εr), and the inter-limb angles:

εr = (sin γa’ sin γs)/(sin γa sin γs’)(5).

Small values of relative stretch indicate thatthe fault tip has migrated a relatively smalldistance up-section relative to the totalamount of basement offset and that most ofthe total strain is being accommodated byfolding ahead of the fault tip. This type ofbehavior would be expected from loosely orunconsolidated material. In contrast, largevalues of relative stretch would indicatesignificantly larger fault tip migration relative

to total offset and would be indicative ofstronger materials that could notaccommodate significant strain by foldingbefore failure by faulting.

Examination of the above equation (5) showsthat values of relative stretch are minimized,as the divergence of the axial surfaces ismaximized, that is as (γa’) and (γs) are madesmaller as (γa) and (γs’) are forced to 90degrees. This in effect forces open the anglebetween the hanging wall anticline and footwall syncline axial surfaces and allowsdeformation to be spread over largerhorizontal distances with consequent smallerstrain gradients and minimization of energy.However, this effect is mitigated by thethickening of the fold limbs. Note fromequation (4) that minimizing (γs) causes severethickening in the foot wall syncline. Thismeans that the angle of divergence of the axialsurfaces will be a competition betweenminimization of the relative stretch andresistance to thickness changes of the foldlimbs. In effect the angle between the hangingwall anticline and foot wall syncline axialsurfaces are indicative of the materialstrength. Stronger materials will resistthickness changes of the fold limbs and showlittle axial surface divergence. In contrast,weak materials, such as loosely orunconsolidated sediments, will allowthickness changes easily and will seek tominimize energy by exhibiting relatively openaxial surfaces.

Note that equation (1) may be rearranged togive:

α = tan-1[2/(cot γs + cot γa)] (6)

which relates the dip of the basement fault tothe exterior angles of the axial surfaces. Forthe special case of a vertical basement fault(γ=90 degrees) the exterior angles of the axialsurfaces are constrained to be complementaryangles resulting in a symmetrical structure.This indicates that for relatively steepbasement faults (γa) may become greater than


D-3

90 degrees particularly in weak materials suchas unconsolidated sediments.Seismic reflection imaging of Coastal Plainsediments located above steeply dippingreverse basement faults (Fig.2) often showcharacteristics that closely approximate thosediscussed above (see Fig. 1). Distinct offset istypically observed at the basement surface butdies out up section. Above the fault tipdeformations in the shallower parts of thesection are accommodated, at least at largerscales, by fault propagation folding. Thisfolding can be characterized as a monoclinewith steeply dipping limbs directly above thefault tip that progressively show less dip upsection. The axial surfaces of the hanging wallanticline and foot wall syncline divergeupward and are relatively open resulting indeformation occurring over a large horizontalarea. In addition the exterior angles observedfor the hanging wall axial surfaces of thesestructures are typically near vertical or greaterthan 90 degrees with very open folding. Thesecharacteristics are all indicative of fault-propagation folding in unconsolidated toloosely consolidated sediments.

Potential Hydrologic and ContaminantTransport Properties of Fault-PropagationFolds

The addition of regional dip to the relativelysimple geometry exhibited by a typical fault-propagation fold (Fig. 4) has potentiallysignificant consequences for the hydrologicand contaminant transport properties exhibitedby the local stratigraphy. These consequencesare critically dependant on the way that thefaulting and folding is accommodated at smallscales. Below the fault tip in the lower part ofthe section where significant offset ofstratigraphic units may occur potentialhydrogeologic effects include, dammingresulting from juxtaposition of impermeableconfining layers against more permeable units(Fig. 4) or connection of previously isolatedaquifers by moving one opposite the other.The potential for both of these situations isenhanced by the heterogeneous strains

experienced by the hydrogeologic units in theforelimbs of the foot wall syncline andhanging wall anticline as expressed inequations (3) and (4).

In the shallower section, above the fault tippotential hydrogeologic effects are morecomplicated and dependent on the failureproperties of the confining units (i.e. clays). Ifthe confining lithologies respond to thefolding in a completely plastic mode then it isfeasible that the confining units will retaintheir competency and no connection willoccur between the aquifers. However, if theconfining units exhibit some component ofbrittle failure and develop open fractures to asignificant degree then it is likely that theirconfining properties will be compromised.

The interaction between the elevation changesin confining units resulting from fault-propagation folding and potentiometricsurfaces may also result in subtle localizedhydrologic effects. Areas where thepotentiometric surface is below a particularconfining unit will exhibit “downward head”relative to the confining unit. However, in thevicinity of the foot wall syncline, where theconfining unit is lower in elevation due tofolding, the confining unit will drop below thepotentiometric surface and locally an “upwardhead” condition will be observed. In theshallow parts of the stratigraphic sectionwhere inter-limb fold angles of only a fewdegrees are present this local “head reversal”condition could be very enigmatic if the fault-propagation folding were not recognized.

The localized dip anomalies resulting fromfault-propagation folding may also haveprofound effects on the migration behavior ofdense non-aqueous phase liquids (DNAPL,Fig. 5) since the migration of these materialsis to a significant degree gravity controlled. Ifthe DNAPL source is up regional dip thesematerials will have a tendency to migratedown dip and pool in the trough formed by thefoot wall syncline. Again, however, thesubsequent migration behavior will depend inlarge part on the failure mode exhibited by the


D-4

material comprising the confining units. If thefolding is accommodated completely byplastic deformation the confining unitsprobably will fully retain their confiningproperties and the DNAPL will remain abovethe confining unit. In this case as long as theDNAPL does not fill the synclinal trough andspill over the lip formed by the hanging wallanticline it will remain localized or at leastchanneled down the trough. However, if thefolding of the confining units results in opentransmissive fractures, particularly in thehinge zone of the synclinal trough then it islikely that the DNAPL will migrate throughthe confining unit below and into the loweraquifers. In this scenario since the strainsexperienced by the confining units in thelower parts of the section due to the fault-propagation folding are larger it is likely thatthe DNAPL will migrate down sectionthrough the hinge zones of the foot wallsyncline to the first competent unit. , probablythe basement. Note however that at this pointthe dam formed by the high angle reverseoffset will effectively stop any continueddown dip movement.

More complicated scenarios may beenvisioned when the hydrologic effects andmigration properties of fault-propagationstructures are combined. It may be entirelyfeasible in hanging wall synclines in whichopen fracturing has compromised confiningunits to have DNAPL contaminants migratingdown section while at the same timegroundwater migrates up-section due to thelocalized “head reversal” effect. This couldlead to a plume of dissolved componentsignificantly down hydrologic gradient fromthe structure in spite of the fact that the freephase is effectively localized by the fault-propagation fold structure.

DISCUSSION AND CONCLUSIONS

Deformation of poorly to unconsolidatedCoastal Plain sediments resulting from non-vertical high angle reverse faulting in rigidbasement rocks typically is expressed as fault-

propagation folding in the shallower part ofthe stratigraphic section. The local dipanomalies associated with this type ofstructure and the behavior of confining unitsmay have significant effects on localhydrologic behavior and contaminanttransport properties depending upon thedetails on the failure mode that accommodatesthe folding in the confining lithologies. In thedeeper section below the fault tip wheresignificant offsets are recorded bygeohydrologic units, damming may occurwhere less permeable units are juxtaposedagainst more permeable units. Or connectionbetween aquifers may occur due to the offset.These phenomena will be amplified due to thethickening and thinning of geohydrologicunits in the forelimb of the structure.

In the shallower section above the fault tipwhere deformation of the section isaccommodated by fault-propagation foldingmore subtle effects may be observed. If thefolding is accompanied by open fracturing ofthe confining units then connection betweenaquifers may occur. However, even if theconfining units fail completely in a plasticmode and retain their integrity localized “headreversals” may occur due to the interaction oflocal dip anomalies and potentiometricsurfaces.

Localized dip anomalies will also affect themigration of contaminants such as DNAPL,which move primarily down gravitationalgradients. For a scenario in which confiningunits retain their integrity DNAPL willmigrate down regional dip and pool in thetrough formed by the footwall syncline. Inscenarios in which the confining units arecompromised by open fracturing DNAPL willlikely migrate down section along the axialsurface of the footwall syncline to the lowestuncompromised confining unit which in themost extreme case will be the rigid basement.However at this location migration will bearrested by the dam formed by the up-thrownside of the high angle reverse fault. In bothcases migration pathways and pooling areasmay be predicted based on the locations of the


D-5

axial surfaces of the fault-propagation fold.Since these pooling areas will be areas ofDNAPL concentration knowledge of thisbehavior in detail may lead to extractionpoints that result in highly effective andefficient remediation system design. Also,since groundwater could migrate up-sectionwhile at the same time DNAPL is flowingdown section plumes of dissolved componentmay be present at large distances from the freephase in spite of the fact that the free phase islocalized by the structure.

It should be noted that for typical basementoffsets recorded for the Atlantic Coastal Plain(i.e. < 100ft.) shallow section dip anomalieswill be very subtle (on the order of a fewdegrees) and difficult to recognize especiallyby conventional borehole techniques. Forinstance shallow boreholes that sample onlythe upper part of the fault-propagation foldwill show the enigmatic effect that the closerthe boreholes are together the less apparentoffset that they record. The shallow section isalso very difficult to image with highresolution geophysical techniques such asreflection seismic profiling. Howeverreflection profiling in most cases can imagethe intermediate and deeper parts of thesection in addition to the basement surfaceoffset to indicate the presence of faultpropagation folding. If the geometry of thestructure can be determined in the deepersection the geometries of shallow stratigraphicunits may be effectively calculated usingpreviously determined kinematic models forthese structures. This interpretation techniquemay be used to explain unusual “headreversals” or used to design efficientcontaminant remediation systems.

REFERENCES

Bramlett, K.W., Secor, D.T. Jr., and Prowell, D.C.(1982) The Belair fault: A Cenozoic reactivationstructure in the eastern Piedmont. GeologicalSociety of America Bulletin, v.93, p.1117.

Cramer, H.R., Arden, D.D Jr. (1978) Faults inOligocene rocks of the Georgia Coastal Plain.

Geological Society of America Abstracts withPrograms, v.10 p.166.

Domoracki, W.J. (1995) A geophysicalinvestigation of geologic structure and regionaltectonic setting at the Savannah River Site, SouthCarolina. Ph.D. Dissertation, Virginia PolytechnicInstitute and State University, Blacksburg, Va.236pp.

Howell, B. B., Zupan, A.W. (1974) Evidence forpost-Cretaceous tectonic activity in the WestfieldCreek area, North of Cheraw, South Carolina.Geologic Notes, South Carolina State DevelopmentBoard, v.18, p.98-105.

McConnell, D. A., (1994) Fixed-hinge, basementinvolved fault-propagation folds, Wyoming.Geological Society of America Bulletin, v.106(12), pp. 1583-1593.

Mitra, S. (1993) Geometry and kinematic evolutionof inversion structures. The American Associationof Petroleum Geologists Bulletin, v.77 (7), p.1159-1191.

Mixon, R.B., Newell, W.L. (1982) Mesozoic andCenozoic compressional faulting along the AtlanticCoastal Plain Margin, Virginia. In CentralAppalachian Geology NE - SE Geological Societyof America Field Trip Guide Books, P.T. Lyttle(ed) American Geological Institute.

Narr, W., Suppe, J. (1994) Kinematics ofbasement-involved compressive structures.American Journal of Science, v.294, p802-860.

Prowell, D.C., O’Connor, B.J. (1978) Belair faultzone: Evidence of Tertiary fault displacement ineastern Georgia. Geology, v.6, p.681-684.

Prowell, D.C. (1983) Index of faults of Cretaceousand Cenozoic age in the eastern U.S. U.S.Geological Survey Miscellaneous Field StudiesMap, MF-1269.

Prowell, D. C. (1988) Cretaceous and Cenozoictectonism on the Atlantic Coastal Margin. in TheGeology of North America, v. I-2 The AtlanticContinental Margin: US. Geological Society ofAmerica.

York, J.E., Oliver, J.E., (1976) Cretaceous andCenozoic faulting in Eastern North America.Geological Society of America Bulletin, v.87,p.1105-1114.


D-6

Figure 1. Geometric elements of fault-propagation fold kinematic model (after McConnell,1994). (a & β dip of fault surface; (γa + γa’): anticline inter-limb angles; (γs + γs’): syncline inter-limb angles; (U): dip of unfaulted bed; (H): structural relief on basement surface; (u):elevation above basement surface; (To): thickness of undeformed beds; (Tf): thickness offaulted bed in syncline forelimb; (Th) thickness of faulted bed in anticline forelimb.


D-7

Figure 2. Seismic reflection profiles across the Crackerneck fault (a) in the Coastal Plain ofSouth Carolina (Domoracki, 1995). Events corresponding to basement surface and othergeologic units in cover sequence annotated in addition to the axial surfaces of the foot wallsyncline and hanging wall anticline.


D-8

Figure 3. Fault-propagation fold with regional dip. Potential hydrogeologic effects illustrated.A local condition of “upward head” may occur relative to the confining unit above the crosshatched area.


D-9

Figure 4. Fault-propagation fold with regional dip illustrating the probable behavior ofDNAPL when confining unit integrity is not compromised.


D-10

Figure 5. Fault-propagation fold with regional dip illustrating probable behavior of DNAPL ifconfining units are compromised.


E-1

Outline of the Geology of Appalachian Basement Rocks Underlying theSavannah River Site, Aiken, SC

Allen J. Dennis, Biology and Geology, University of South Carolina Aiken, Aiken SC 29801-6309John W. Shervais, Department of Geology, Utah State University, Logan, UT 84322 andHarmon D. Maher, Jr., Geology and Geography, University of Nebraska Omaha, Omaha, NE 68182-0199

INTRODUCTION

Basement core collected over a forty yearcampaign at the Savannah River Site andsurrounding environs affords an unparalleledopportunity to observe Appalachian basementrock and its alteration during Mesozoic riftingand subsequent Tertiary tectonism. Thestructures that are preserved would surely notsurvive long at the earth�s surface, and havethe potential to change how we think about thehistory of Piedmont rock exposures we use fordetailed bedrock geologic mapping.Remarkably no less than 5000m of coresample fault zones that have beenintermittently active throughout thePhanerozoic. These fault zones acted asconduits for fluids throughout the Mesozoicand Cenozoic. Additionally, evidencepresented here and elsewhere in this volumeindicates that syn- and post-depositionalfaulting of the Tertiary section isfundamentally controlled by long-livedAppalachian structures.

The following short paper draws on our inpress or submitted manuscripts for GeologicalSociety of America Bulletin, AmericanJournal of Science, and Geological Society ofLondon treating the geochemistry andpetrology of metamorphosed Neoproterozoicarc rocks underlying SRS, the Paleozoicstructure of those rocks, and evidence forMesozoic and younger fluid flow andpseudotachylyte generation. This work wassupported by South Carolina UniversitiesResearch and Education Foundation contract170.

We are grateful to Randy Cumbest, SharonLewis, Van Price, Doug Wyatt (all at SiteGeotechnical Services for all or some fractionof the time this work was done), Tom Temples

(at DOE during the time this work was done)for their assistance and support during thestudy. Josh Mauldin, Hampton Uzzelle,Lynda Bolton, John Harper, and Tom Creechassisted in this work.

Lithology

Greenschist to granulite faciesmetamorphosed volcanic arc rocks thatunderlie the Savannah River Site are divided,from northwest to southeast, by metamorphicgrade, into the Crackerneck MetavolcanicComplex, Deep Rock MetaigneousComplex, and Pen Branch MetaigneousComplex (Fig. 1; Dennis and others, 2000,Shervais and others, 2000). The CrackerneckMetavolcanic Complex are the lowest graderocks and comprise very weaklymetamorphosed mafic and felsic tuffs, withrelict lapilli. These are best observed in wellsP30 and P6R (Fig. 2). Deep RockMetaigneous Complex is made up of the DeepRock Metavolcanic Suite and the DRBMetaplutonic Suite. Rocks of the Deep RockMetavolcanic Suite are generally dark grey-green mafic metavolcanic rocks, �well-exposed� in DRB wells 2-7 at depths from ca.1100� to ca. 1900�. These metavolcanic rocksare locally garnet-bearing, indicatingamphibolite facies conditions. These rocksare cut by metamorphosed mafic and felsicdikes that locally preserve good primarytextures. The DRB Metaplutonic Suitecomprises metadiorite and (locally mylonitic)quartz diorite gneiss. In the DRB-1 where thisrock is observed, a serrated contact betweenthe diorite and well-foliated maficmetavolcanic rock is observed. Furthermore,deformation in the diorite is veryheterogeneous and may be best described asthin, late stage anastomosing shear zones.The Pen Branch Metaignous Complex is


E-2

likewise made up of The Pen BranchMetavolcanic Suite and PBF MetaplutonicSuite. The Pen Branch Metavolcanic Suite ismade up of distinctive layered (transposed)intermediate to mafic orthogneisses. Theserocks are observed in PBF-7 between 2500-3500 where they are screens within the PBFMetaplutonic Suite, and between 3600-3648�(TD) in that well. Pen Branch MetavolcanicSuite is also observed in the SeismicAttenuation (SA) Core in spot cores between1959� and 4002�. The PBF MetaplutonicSuite is best observed in PBF-7 and is agarnet-bearing granodiorite to granodioritegneiss, locally with hornblende megacrysts.In that well it is heterogeneously deformed,with a locally intense K-metasomaticoverprint. The variety of deformation andmetamorphic textures in this unit, as well asthe variability of the metasomatic overprint isdisplayed in PBF-7 2550�-3600�.

A schematic interpretation of the temporal orstratigraphic relations between these units ispresented in Figure 3.

Dennis and others (1997) report U-Pb zirconages and Sr and Nd isotopic data for the DRBMetaplutonic Suite and PBF MetaplutonicSuite. Dennis and others (1997) report U-Pbzircon crystallization ages of 625.1± 1.4 Ma(weighted mean 207Pb/206Pb dates of foursize fractions) and 619.2 ± 3.4 Ma (average of207Pb/206Pb dates from only two sizefractions analyzed) for the PBF-7 granodioriteand DRB-1 diorite respectively. Neithersample showed evidence of inheritance. �Ndvalues for PBF-7 and DRB-1 rocks are 2.0 and3.5 respectively (Dennis and others, 1997).DRB-1 tonalite yields a 87Sr/86Sr ratio of0.701939. Sr isotopic systematics of PBF-7granodiorite are significantly complicated bythe addition of Rb during the pinking eventdiscussed below, but a primary initial ratio of0.703 may be estimated.

Based on the differences between Nd isotopesin PBF Metaplutonic Suite and DRBmetaplutonic suite and similarities in REEchemistry (unpublished data, Dennis andothers, 2000, Shervais and others, 2000)

between Pen Branch Metaigneous Complex,Deep Rock Metavolcanic Suite and differencesin REE patterns between those two “units”and DRB Metaplutonic Suite we interpret thatThe DRB Metaplutonic Suite may be faultedinto a Pen Branch Metaigneous Complex-Deep Rock Metavolcanic Suite intrusive-extrusive complex. We interpret that thisfaulting may have occurred in Neoproterozoictime, and may have served to initially localizethe Tinker Creek nappe.

Alleghanian (Pennsylvanian) Structure

A body of evidence supports a faultedoverturned nappe limb as the likely contactbetween the Pen Branch MetaigneousComplex and Deep Rock MetaigneousComplex. The overturned nappe limb is calledthe Tinker Creek nappe, and the shear zonethat �decapitates� the Tinker Creek nappe iscalled the Four Mile Branch fault. The nameFour Mile Branch fault has been usedpreviously in the SRS region to describeoffsets of Tertiary strata by Wyatt and others(1996) and Harris and others (1997). Weinterpret the Tinker Creek nappe and FourMile Branch fault to be Carboniferous in agebased on the similarity of structural style withthat observed in the exposed eastern Piedmontof South Carolina (e.g., Maher , 1987, Secorand others, 1986), supplemented by a single40Ar/39Ar age (305±5 Ma) of a hornblendeneoblast from these rocks reported by Rodenand others (1996).

The evidence supporting the interpretationthat the structure underlying the DRB (1-7)well array is an overturned nappe limbincludes parasitic folds, chloritic slip surfacesand shear bands, and evidence of extensivetransposition (Fig. 4). These data have beensupplemented by P-T calculations based ongarnet-biotite thermometry that supportinferences based on map-scale petrology, andthe seismic reflection processing ofDomoracki (1995). These structures (TinkerCreek nappe - Four Mile Branch fault)controlled the location of the TriassicDunbarton basin border fault. An


E-3

Andersonian (1951) analysis supports thisinterpretation. Tinker Creek nappemesoscopic structures such as the chloriticslip surfaces were later reactivated as conduitsfor fluid flow (pinking event), gouge zones,and ultimately pseudotachylyte.

Mesoscopic structures that indicate nappeemplacement from southeast to northwest areabundant in DRB formation and includeparasitic folds, chloritic slip surfaces andshear bands (Fig. 5). These structures are bestrepresented in DRB well series, particularly inDRB-6. Mesoscopic folds interpreted to beparasitic have a down dip vergence thatindicates an overturned limb geometry. Thetrailing limb of these mesoscopic folds has adip of 50-60° typically (assuming that the coreis vertical), and an overturned limb that isnearly horizontal. Interlimb angles are usuallyin the range of 30-60°. It can be shown thatthe axes of these folds are nearly horizontal.Most foliation dips are in the range of 40-60°in DRB formation; rarely are dipssubhorizontal. There are also �floating�isoclinal fold hinges, intact isoclinal folds, andevidence for transposition within the DRBformation. Probably most of this isoclinal-transposed fabric records nappe deformationbut some may record a Late Precambrianorogenic event.

Chloritic slip surfaces cross-cut parasitic,mesoscopic folds and transposed layering, andshow hanging wall up offset. Chloritic slipsurfaces have the same strike as foliation andalmost always dip in the direction of foliation,but with a shallower dip, ca. 30-50° (Fig. 5g).A small fraction have the opposite dip. Thesurfaces range in thickness from less than 1mm thick to 4 mm thick. Sense of offsetacross these features is consistentlyhangingwall up. The amount of offset can beup to several cm. Chloritic slip surfaces areseparated by 2-10 cm. In DRB-6 arespectacular examples of down-dip vergent(overturned limb) parasitic folds cross cut byhangingwall up chloritic slip surfaces. Theprecise origins of the geometry of thesesurfaces and the chlorite along them areunclear, but are of considerable interest.

Mesocopic shear bands are present in bothvertical and horizontal sections of core, andindicate hangingwall up, and dextral motionrespectively. Horizontal sections of a fewsamples showed sinistral shear sense. It wasnot possible to discern overprinting relationsof dip-slip and strike-slip shear bands. It maybe simplest to intepret these structures ashaving formed during dextral tranpression thathas not been partitioned spatially as it seemsto be the case of the Irmo shear zone andAugusta fault (Secor and others, 1986; Maherand others, 1991).

Norian (Late Triassic) Metasomatism –“The Big Pink”

The Dunbarton Triassic basin (Marine, 1974;Marine and Siple, 1974) is one of a wellstudied array of basins described by Olsen andothers (1991) related to opening of the presentday Atlantic. The border fault rocks andfanglomerates of the northwestern margin ofthe basin are well exposed in coreholes PBF-7and 8. Mesozoic and younger veining andalteration paragenesis are generally bestexpressed in these wells.

Pinking Event

The effects of a hydrothermal alteration(�pinking�) event are widespread andheterogeneous in Pen Branch MetaigneousComplex and Deep Rock MetaigneousComplex. Hot waters with abundantdissolved silica, K and Al were responsiblefor metasomatic alteration along fracturesurfaces and foliation planes. Hydrolysis ofgroundmass biotite in the PBF-7 granodiorite

is interpreted to be the source of K+; thisphase is consistently altered to chlorite in theunpinked granodiorite. In DRB rocks, pinkingis observed as zones typically less than 1 cmthick that reactivate chloritic fractures orquartz veining.

Pinking of the PBF granodiorite occurs aslimited alteration proceeding along foliationplanes with grain size reduction andnucleation of quartz and potassium feldspar


E-4

grains along the faces of potassium feldsparmegacrysts parallel and adjacent to foliationplanes. It is also observed as veins ofalteration cross-cutting foliation. In pinkedrocks, sericitization of plagioclase feldspars isobserved, and these crystals appear to have adusty coating in thin section. Microprobeanalysis indicates that the An content ofplagioclase is ≤An5.

Pinking may occur along fractures in as aband less than 5mm thick, or as broad zonesthat affect ten�s of meters of core (Fig. 6).The most spectacular location that this effectmay be observed covers the range 3496� (box167) - 3602� (box 173) in PBF-7, over whichthe rocks are completely pinked with less than5% mafic minerals. The presence of garnetand altered hornblende megacrysts (altered tomagnetite and pyroxene) in a felsic gneissaffirms that the protolith of these rocks wasthe PBF-7 granodiorite. These rocks are veryhard and recrystallized as befits thesilicification accompanying their alteration.Few later fractures crosscut this zone. Belowthis zone, there is no evidence of pinking, andthe rock is a garnet amphibolite. Biotite andamphibole have not been chloritized.

An isocon diagram (Fig. 7; O�Hara, 1988,1990; O'Hara and Blackburn, 1989; Grant,1986, after Gresens, 1964) was prepared forthe PBF-7 granodiorite. The construction ofthis diagram requires the ratio of major andtrace element analyses for altered rock(pinked) and protolith (unpinked) becalculated, and scaled such that all ratios maybe observed on a single plot. If ametamorphic or hydrothermal event isisochemical the ratios of analysed elements oroxides in the altered rock and protolith willequal 1 and lie along a line with a slope of 1.Deviations from a slope of 1 may suggestrelative enrichment (>1) or leaching (<1).However if certain trace elements areimmobile (e.g., Ni, Sc, Zr, Cr, Ti, V), then theline that passes through those elements� ratioson the isocon diagram may demark theboundary between leaching and enrichment inother elements. Using this logic, it isconcluded that MnO, MgO, and to a lesser

extent CaO are relatively depleted in thepinked rocks and K2O, Na2O, SiO2, Rb, andto a lesser extent Al2O3 are enriched in thepinked rocks.

Pinking is observed to cross-cut differentdeformation-metamorphic facies of the PBFMetaplutonic Suite granodiorite and thus mustpostdate the event responsible for fabricdifferences. Kish (1992) reports a 2 point Rb-Sr isochron of 220±5 Ma (Norian) for agranite recovered from the C-10 borehole.Petrography and chemistry of this sampleconfirm that it has been pinked, and that it issimilar to the PBF-7 granodiorite. Theisochron is interpreted to date the pinkingevent in this area. That the pinking is mostintensive in the PBF-7 well, and that the

biotite (interpreted to be source of the K+) isubiquitously altered to chlorite in the PBF-7granodiorite suggests that the DunbartonBasin border fault was the conduit for thepinking event hydrothermal fluids. It isestimated that the temperature of the waterswas low enough to allow chlorite andoligoclase to form.

Greening Event

A "greening" event postdates the pinkingevent. In crystalline rocks (particularly of thePBF-7 core) this event is recognized byepidote (and rare pyrite) veining that maycause local brecciation as the vein is injected.Epidote veins are generally 1 mm- 1 cm thickand fine-grained. The veins typically have thesame strike as foliation, and dip steeply in thesame direction (concordant), or in the oppositedirection (discordant). The effects of thegreening are, however, most spectacularlydeveloped in Triassic fanglomerates of PBF-7(Fig. 8; boxes 10-21, footage 1150-1330).The matrix of these conglomerates is nolonger brick red but instead is a pale green,and an irregular white and grey rind (5mm to1 cm thick) appears on oxidized gneiss clastsgreater than 2-3 cm in diameter (Fig. 8). Thealternating white and grey bands correspondto gneissosity in the oxidized clast, andinterpreted to represent partially resorbededges and the penetration of reducing fluids


E-5

along gneissosity. In box 10 it is apparent thatshallowly dipping fractures (≤ 25° dip) controlgreening, and greening extends 1-10 cm awayfrom these fractures. In boxes 13-18 thisgreening of conglomerate matrix is complete.Microprobe analyses of fibrous intergranularmaterial (cement?) in the (PBF-7-16) matrixindicate that this is green muscovite. Locallythis material contains fine disseminated pyrite.

Similar relations are observed in GCB-2-648'(Fig. 8). Here saprolitized Triassicfanglomerate is observed. The matrixcomprises coarser grained white to creamygreen material, while the clasts are finergrained (clayey) and brick red. The marginsof these clasts are discernible as reducedghosts in the matrix, while the clasts' reddishinteriors have very highly irregular shapedthat are again interpreted as the delimiting theextent of penetration of reducing fluids alongfoliation. In poorly sorted Cretaceous (?)sediments 30' above (GCB-2-615') the reducedfanglomerate, a clot of euhedral chalcopyritegrains 2.5 cm x 5 cm (long axis) parallel tobedding plane is recognized. It is unlikely that

this (≥ 10 cm3) euhedral sulfide clot is adetrital clast (in sand), and it is interpreted torecord the greening event in Coastal Plainrocks.

The "greening" event indicates a flushing ofreducing waters along Triassic basin borderfaults, that may be related to migration ofhydrocarbons derived from Triassicsedimentary rocks (e.g., D. Richers, pers.comm.) The presence of chalcopyrite in rockstentatively identified as Cape Fear,approximately 10 m above reduced clasts andmatrix of a thin zone of preserved Triassicfanglomerate in GCB-2-648, suggests that thisevent postdates Cape Fear deposition(Coniacian). Differences in oxidation statebetween Cape Fear rocks and overlying redand yellow Middendorf Formation (?) sandsindicate that the greening event occurred priorto Middendorf deposition in the Santonian(86.6-83.0 Ma). Furthermore the presence ofreduced fanglomerate in GCB-2 suggests thata border fault of another Triassic sub-basinmay lie in the the vicinity of the GCB-2 well.

Quartz Veining and Vuggy Quartz

Quartz veining, locally quite vuggy, cutsepidote veins in PBF-7 core. These veins aresubparallel to epidote veins when crosscuttingrelations are observed. The veins generallyhave the same strike as foliation, and usuallydip parallel to foliation or less commonly inthe opposite direction. The veins can beirregular and not simple planar features. Theyare typically not more than 2 cm thick, butthick veins may present open vugs up to 7 mmlined with well terminated crystals. Thesevuggy veins are most notable below PBF-7-(box) 111-2662, and occur with irregularincreasing frequency through box 163-3452',and then less frequently in completely pinkedmylonite through 173-3602'. Thick veins mayalso contain cm-scale clasts of the variablypinked granodiorite gneiss they cut. Boththese observations strongly suggest high fluidpressures in the vicinity of the DunbartonBasin border fault during Late Mesozoicthrough Paleogene time.

Zeolite ± calcite

Zeolite ± calcite, locally euhedral, veins cutand locally fill vuggy quartz veins throughoutDRB and PBF formations. These veins aresubparallel to foliation as the epidote andquartz veining events, but may crosscutfoliation. The zeolite is typically pinkishorange or less commonly white. Verycommonly it forms radiating sprays, less than2 cm in diameter, of euhedral crystals, lessthan 1 cm in length, on the fracture surface.Calcite is typically colorless and rarely formseuhedral blocky crystals.

Strike normal zeolite

Nearly vertical (75-90°), nearly "strike-normal" zeolite-filled fractures are verycommon in DRB and PBF formation rocks.Map view of individual core segmentsindicates that often the strike of veins is 15-20° clockwise of foliation dip direction ratherthan precisely strike-normal. The veins aretypically less than 1 cm thick, and are filledwith pinkish orange to white crystals. Some


E-6

veins show evidence of (considerable)indeterminate amounts of offset, but thisoffset must predate filling of fractures by largewell-formed crystals.

Approximately strike-normal joint sets havebeen recognized in the eastern Piedmont ofSouth Carolina during the course of detailedbedrock mapping by the authors (unpublisheddata; Bartholomew and others, 1997, Brodieand Bartholomew, 1997, Heath andBartholomew, 1997). Assuming a regionalstrike of 065 it is estimated that the orientationof the "strike-normal" zeolite filled fractureset is 350-355°, and probably indicates anepisode of approximately E-W extension,accompanied by minor strike-slip motion.

Understanding the timing of this motion isimportant for an understanding of Teritaryregional kinematics; Dennis and others (2000)report an attempt was made to date thesemineralized veins. These efforts wereencouraged by the report of Secor et al. (1982,p. 6952) of the successful K-Ar dating of astrike-parallel (060) set of zeolites in theCarboniferous Winnsboro (SC) granite nearthe Virgil Summer nuclear station. The agereported is 45±5 Ma (South Carolina Electric& Gas Company 1977). This set is post-datedby a second zeolite array oriented 332(approximately "strike normal" to regionalstrike, Secor et al. 1982, their fig 5c.) thatshows several cm of oblique slip. Both setsare reported to be steeply dipping. Based onmicroprobe analysis, PBF-7 strike-normalzeolite crystals are laumontite-chabazite andhave K in the range of 0.2 weight percent. Asingle 3 gram sample (PBF-7-139-3089) ofhandpicked (2mmx7mm) zeolite grainsyielded a conventional K-Ar age of 22.9±1.5Ma. This is considered to be a minimum age,because radiogenic Ar may have been lostfrom the system continuously or episodicallysince the crystals formed. Other strike normalzeolites from the SRS basement should bedated in an effort to push this minimum ageback.

Pseudotachylyte

Over two dozen pseudotachylytes have beenrecognized in core (Fig. 9). The commoncharacterstics of these features are 1)formation in a fine-grained intermediate tomafic protolith; 2) nucleation along pre-existing, thin chloritic slip surfaces; 3)evidence that suggests K-metasomatism alongchloritic slip surfaces influencedpseudotachlyte generation, perhaps byreducing the melting temperature of chloriticgouge; 4) evidence for multiple events withina single pseudotachylyte vein; different eventsare recognized based on discrete banding ofhydrated or devitrified vs. glassy or simplyaphanitic matrix; 5) preservation of angularclasts at all scales; 6) mm scale teepee orinjection structures that may penetrate theadjacent rock from 3 mm to 2 cm. It hasgenerally not been possible to recognize themagnitude or sense of offset across thepseudotachylytes. Locally it has beenrecognized that presence of felsic layeringwithin several cm of the chloritic slip surfacesmay influence pseudotachylyte development;perhaps through a very local increase indifferential shear stress.

Pen Branch fault

The Pen Branch fault has been described bySnipes and others (1993) and Stieve andStephenson (1995, based on Domoracki,1995). These authors recognized the PenBranch fault as a surface across which thebasement-coastal plain unconformity has beenoffset as much as 50 m (southeast side up)with offset of middle Eocene units decreasingto as little as 3-4 meters. The extent of controlof the Pen Branch fault on outcrop patterns ofthe Upland unit, or surface drainages and therelative importance of strike slip motion alongthis fault system is a matter of somecontroversy (R.J. Cumbest, T.J. Temples,pers. comms.). Domoracki (his fig. 23, 1995)plots two recent hypocenters (6/1985 and8/1988) on SRS line 1 on the basement faultabove the Tinker Creek nappe, in the vicinityof the Pen Branch fault, at depths of .96 and2.86 km respectively.


E-7

In our observations the Pen Branch fault cutscrystalline basement in PBF-7 between boxes25 (1375') and 78 (2183'). Over this intervalat least two intrusive units of granodioriticcomposition are hydrated and retrogressed sothat they appear as chlorite schists.Deformation is highly heterogeneous, andseveral less completely retrogressed slices orlenses are preserved. The largest slice occursbetween boxes 38-48, with a much smallerexample at about 2078'. None of thepreviously described vein sets appear tocrosscut the sheared, hydrated schists, while atleast some are observed in the protolith.Evidence of the orthogneiss protolith isdemonstrated by preservation of pink graniticvein arrays that maintain orientation in theretrogressed rock, and the presence of relictfeldspar grains (porphyroclasts), up to 1 cmlong. within the chloritic matrix. Gouge,breccia zones and chloritic fractures typicallyoccur with 15-20° dips, parallel to foliation(ca. 45°) and steeply dipping (ca. 75°).Breccia zones, up to 10's of cm thick, andgouge zones, up to 2 cm thick, appear tonucleate on chloritic fractures Within 15 m ofthe fault zone, chloritic fracture surfaces areoxidized to a black hematite. At somelocations these oxidized coatings are striated.

Pseudotachylyte and the Pen Branch fault areboth interpreted to largely postdate strike-normal zeolite though both may extend backthrough the Mesozoic. That pseudotachylytepostdates the pinking event (220±4 Ma, Kish,1990; Norian) is evinced by the observation ofpseudotachylyte crosscutting rocks that havebeen visibly pinked. Microprobe analyses ofpseudotachylyte glass and devitrifiedglass/chlorite grains show that in DRBformation mafic metavolcanic rocks with abulk composition that is typically well below4 weight % K2O (Shervais and others, 2000),glass and devitrified glass compositions aretypically in the range 8-12 weight % K2O.This indicates the likelihood of pinking fluidsaltering chloritic vein material and adjacentmafic metavolcanic and subsequentreactivation of these chloritic slip surfaces assites of pseudotachylyte generation.

In over two dozen observations there are onlya couple instances of any crosscuttingmineralized veins cutting pseudotachylyte. Inone very thin (<1 mm) calcite vein cutspseudotachylyte at an oblique angle (DRB-7-9-1419). In the second, it is clear thatpseudotachylyte predates at least one zeolitemineralized fracture event. Thepseudotachlyte (Fig. 9; PBF-7-89-3089) is anirregular brick red, oxidized, devitrified zone,greater than 1 cm thick that is subparallel toand crosscuts epidote and calcite + zeoliteveins and includes clasts of mineralized veinswithin it. These clasts are less than 1 cm andand are visible at all microscope scales. Thebrick red, cherty appearance of the featurestrongly suggests oxidizing fluids passedthrough a cryptocrystalline or glassy material.The relative timing of this event can only besaid to be post-zeolite + calcite veining. Theoxidizing fluids appear to predate the cross-cutting zeolite-filled fracture that issubparallel to the margin of thepseudotachylyte and "decapitates" (oxidized)teepees.

In the case of basement expression of the PenBranch fault (250 m (800') over PBF-7- 25through 78) it can be shown that this featurecross-cuts all mineralized veins includingstrike-normal zeolite. Intact mineralizedfractures are not preserved in the intervalPBF-7-25 through 78. Given the relativedecrease in offset of Tertiary units in theAtlantic Coastal Plain cover (Snipes andothers, 1993) above the expression of PenBranch fault motion in the crystalline rocks,the retrogression of fault rock protolith anddisruption of mineralized fractures and veinsin bedrock is interpreted to have occurredfrom the Late Cretaceous through theNeogene.

CONCLUSIONS

Repeated episodic Phanerozoic reactivationsof basement fault zones are documented inmore than 5000m of basement core recoveredfrom beneath the updip coastal underlying theUS Department of Energy Savannah River


E-8

Site, near Aiken, South Carolina.Neoproterozoic (ca. 619 and 624 Ma)metadiorite and metagranodiorite rockscontain a heterogeneously developed,anastomosing mylonitic fabric and intrudefoliated mafic metavolcanic rocks. Atapproximately 305 Ma, granulite faciesorthogneisses were thrust over amphibolitefacies metaigneous rocks on (northwest-vergent? and dextral) structures called theTinker Creek nappe and Four Mile Branchfault. The overturned limb of the TinkerCreek nappe and Four Mile Branch faultcontrol the location the border fault of theDunbarton Triassic basin. The Dunbartonbasin border fault region acted as a conduit forfluids in Mesozoic and Cenozoic time. Ca.220±5 Ma, a potash and silica metasomaticevent ("pinking" event) affected crystallinerocks throughout the SRS basement region. A"greening" event flushed highly reducingfluids through crystalline rocks, Triassicfanglomerates and affected rocks interpretedto be as young as Santonian (Cape Fear Fm.).Based on the occurrence of saprolite offanglomerates affected by greening in core,the remains of a subbasin of the Dunbartonbasin were identified in the northwest part ofthe site (GCB-2 well). A consistent veinparagenesis interpreted to to be Cretaceousand younger overprints the previous eventsand includes 1) locally vuggy quartz veinswith the same strike as foliation, 2) zeolite ±calcite with the same strike as foliation, and 3)a near vertical strike-normal zeolite set. Aconventional K-Ar age of a separate ofeuhedral, 2mm x 7mm zeolite grains from thelast set yields an age of 22.9 ± 1.5 Ma,interpreted to be a minimium age. More than30 pseudotachylytes are found throughout themetavolcanic terrane and are preferentiallylocalized along Alleghanian age chloriticfractures dipping less steeply than foliationthat have been �pinked.� In virtually everycase, pseudotachylyte can be shown to post-date mineralized fractures. Thus,pseudotachylytes beneath SRS are likely noolder than Mesozoic, and many are probablyTertiary in age. The Pen Branch fault is"exposed" in approximately 250 m ofbasement core between PBF-7-(box) 25-

(footage) 1375 and PBF-7-78-2168. Mostrecent motion on this fault must postdatestrike-normal zeolites, because the zone cross-cuts all mineralized fractures.

REFERENCES

ANDERSON, E.M. 1951. The dynamics offaulting. Oliver & Boyd, Edinburgh, 206p.BARTHOLOMEW, M.J., HEATH, R.D., BRODIE,

B.M., & LEWIS, S.E. 1997. Post-Alleghaniandeformation of Alleghanian granites(Appalachian Piedmont) and the AtlanticCoastal Plain: Geologic Society of AmericaAbstracts with Programs, 29/3, 4.

BRODIE, B.M. & BARTHOLOMEW, M.J. 1997.Late Cretaceous-Paleogene phase ofdeformation in the Upper Atlantic CoastalPlain. Geologic Society of America Abstractswith Programs, 29/3, 7.

CUMBEST, R.J., PRICE, V. & ANDERSON, E.E.1992. Gravity and magnetic modelling of theDunbarton Triassic basin, South Carolina.Southeastern Geology, 33, 37-51.

DANIELS, D.L. 1974. Geologic interpretation ofgeophysical maps, central Savannah Riverarea, South Carolina and Georgia. USGeological Survey, Geophysical InvestigationsMap, GP-893.

DENNIS, A.J. , WRIGHT, J.E. , MAHER, H.D.,MAULDIN, J.C. & SHERVAIS, J.W. 1997.Repeated Phanerozoic reactivation of aSouthern Appalachian fault zone beneath theup-dip coastal plain of South Carolina.Geological Society of America Abstracts withPrograms, 29/6, 223.

DENNIS, A.J., MAHER, H.D, SHERVAIS,J.W.,MAULDIN, J.C, & UZZELLE, G.H.2000a. The role of fluid flow in basin-bounding faults and metasomatictransformation of basement: An example fromthe Mesozoic and Cenozoic deformation of aNeoproterozoic volcanic arc beneath theAtlantic Coastal Plain, Savannah River Site,South Carolina, USA, In: HOLDSWORTH,R.E. (ed) Nature and tectonic significance offault zone weakening, Geological Society ofLondon Special Volume, in press.

DENNIS, A.J. , WRIGHT, J.E. , MAHER, H.D., &SHERVAIS, J.W. 2000b. Neoproterozoic andLate Paleozoic deformation and metamorphismof a volcanic arc beneath the Atlantic CoastalPlain, Savannah River Site, South Carolina.ms.


E-9

DOMORACKI, W. J. 1995. GeophysicalInvestigation of geologic structure andregional tectonic setting at the Savannah RiverSite. Ph.D. thesis, Virginia Polytechnic andState University.

GRANT, J.A. 1986. The isocon diagram - A simplesolution to Gresens' equation for metasomaticalteration. Economic Geology, 81, 1976-1982.

GRESENS, R.L. 1967. Compostion-volumerelationships of metasomatism. ChemicalGeology, 2, 47-55.

HARRIS, M.K., THAYER, P.A., & AMIDON, M.B.,1997, Sedimentology and depositionalenvironments of middle Eocene terrigenous-carbonate strata, southeastern Atlantic CoastalPlain, USA. Sedimentary Geology, 108: 141-161.

HEATH, R.D. & BARTHOLOMEW, M.J. 1997.Mesozoic phase of post-Alleghaniandeformation in the Appalachian Piedmont.Geologic Society of America Abstracts withPrograms, 29/3, 23.

KISH, S.A. 1992. An initial geochemical andisotopic study of granite from core C-10. In:FALLAW, W.C. & PRICE, V. (eds) GeologicalInvestigations of the central Savannah riverArea, South Carolina and Georgia: CarolinaGeological Society Guidebook. South CarolinaGeological Survey, Columbia, B-IV-1 - B-IV-4.

MAHER, H.D. 1987a. Kinematic history ofmylonitic rocks from the Augusta fault zone,South Carolina and Georgia. AmericanJournal of Science, 287, 795-816.

MARINE, I.W. 1974. Geohydrology of buriedTriassic basin at Savannah River Plant, SouthCarolina. American Association of PetroleumGeologists Bulletin, 58, 1825-1837.

MARINE, I.W. & SIPLE, G.E. 1974. BuriedTriassic basin in the Central Savannah RiverArea, South Carolina and Georgia. GeologicalSociety of America Bulletin, 85, 311-320.

O'HARA, K. 1988. Fluid flow and volume lossduring mylonitization: An origin for phyllonitein an overthrust setting, North Carolina,U.S.A.. Tectonophysics, 156, 21-36.

O'HARA, K. 1990. State of strain in mylonites formthe western Blue Ridge Province, southernAppalachians: The role of volume loss.Journal of Structural Geology, 12, 419-430.

O'HARA, K. & BLACKBURN, W.H. 1989.Volume-loss for trace-element enrichments inmylonites. Geology, 17, 524-527.

OLSEN, P.E., FROELICH, A.J., DANIELS, D.L.,

SMOOT, J.P. & GORE, P.J.W. 1991. Riftbasins of early Mesozoic age. In: HORTON,J.W., & ZULLO, V.A. (eds) The Geology ofthe Carolinas. University of Tennessee Press,Knoxville, 142-170.

PETTY, A.J., PETRAFESO, F.A. & MOORE, F.C.1965. Aeromagnetic map of Savannah RiverPlant area, South Carolina and Georgia. USGeological Survey, Geophysical InvestigationsMap, GP-489.

RODEN, M.F., LATOUR, T.E., CAPPS, C., &WHITNEY, J.A. 1997. Incompletemetamorphic reequilibration in a metadioritefrom the crystalline basement of SavannahRiver Site. Geological Society of AmericaAbstracts with Programs, 29/3, 65.

SECOR, D.T., JR. , PECK, L.S., PITCHER, D.M.,PROWELL, D.C., SIMPSON, D.H., SMITH,W.A. & SNOKE, A.W. 1982. Geology of thearea of induced seismic activity at MonticelloReservoir, South Carolina. Journal ofGeophysical Research, 87, 6945-6957.

SECOR, D.T., JR., SNOKE, A.W., BRAMLETT,K.W., COSTELLO, O.P. & KIMBRELL, O.P.1986. Character of the Alleghanian orogeny inthe southern Appalachians. Part I. Alleghaniandeformation in the eastern Piedmont of SouthCarolina . Geological Society of AmericaBulletin, 97, 1314-1328.

SHERVAIS, J.W., DENNIS, A.J. & MAULDIN, J.C.2000. Petrology and geochemistry of avolcanic arc beneath the Atlantic Coastal Plain,Savannah River Site, South Carolina. ms.

SHERVAIS, J.W., SHELLEY, S.A., & SECOR, D.T.1996. Geochemistry of volcanic rocks of theCarolina and Augusta terranes in central SouthCarolina. In: NANCE, R.D. & THOMPSON,M.D. (eds) Avalonian and related peri-Gondwanan terranes of the circum-NorthAtlantic. Geological Society of America,Boulder, Special Paper, 304, 219-236.

SNIPES, D.S., FALLAW, W.C., PRICE, V., JR. &CUMBEST, R.J. 1993a, The Pen Branch fault:Documentation of Late Cretaceous-Tertiaryfaulting in the Coastal Plain of South Carolina.Southeastern Geology, 33, 195-218.

SOUTH CAROLINA ELECTRIC & GAS COMPANY1977. Virgil C. Summer Nuclear Station FinalSafety Analysis Report, 2. South CarolinaElectric & Gas Company, Columbia.

SPEER, J.A. 1982. Description of granitoid rocksassociated with two gravity minima in Aikenand Barnwell counties, South Carolina. SouthCarolina Geology, 26, 15-24.


E-10

STIEVE, A. & STEPHENSON, D. 1995.Geophysical evidence for post-Late Cretaceousreactivation of basement structures in thecentral Savannah River area. SoutheasternGeology, 35, 1-20.

WYATT, D.E., WADDELL, M.G., and SEXTON,G.B. 1996. Geophysics and shallow faults inunconsolidated sediments. Groundwater, 34,236-334.


E-11

84

83

82

81

34

33

32

A t l a n t i c

C o a s t a l P l a i n

Columbia

Atlanta

Augusta

GEORGIA

SOUTH CAROLINA

F

C-GS

H

B

CPS

GHMZ��

��

��

��

��

�

��

��

N

��

aeromag - E P F S

M O D O CZ O N E

Clark Hill Lake

AFZ

Belair fault zone

subsurface Belair belt

��

�

��

��

*

*

Tr-J mafic igneouscomplex from aeromag

33¡00'

33¡15'

81¡15'33¡30'

81¡15'33¡15'

81¡45'

33¡00'

81¡30'

82¡00'

33¡30'

82¡00'

Martin f.: southern border of Dunbarton basin

��

Carolina slate belt

aeromag lineament: Tinker Cr. nappe limb & EPFS dextral shear zone

Graniteville pluton

Aiken

Kiokee belt

Springfield pluton

C-7

C-5

SAL-1

C-2

DRB wellcluster

33¡15'

81¡45'

C-1

edge of associatedgravity anomaly

33¡30'

81¡45'

10 km

��33¡30'

81¡30'C-3

* subsurface Tr border faultreactivated as Pen Branch Fault

C-10

��

��

��

��

C-1

no s

trat

ord

erim

plie

d

Pen Branch MIC

Deep Rock MIC

Crackerneck MIC

Post meta plutons

Tr sedimentary basin

Tr mafic igneous rocks

fault offsetting Tertiary strata

aeromag linerosional window

through ACP

offsite welllocation

��Carb.Dev.

Four Mile Branch fault

Figure 1. Eastern Piedmont of Georg-olina. Compiled from our work and interpreted fromPetty et al. 1965, Daniels 1974, Speer 1982, Cumbest et al. 1992, and Snipes, pers. comm. Off-site C-wells and SAL-10 located on this map.


E-12

C-7

C-5C-1

C-3

C-2

DRB-1DRB-3

DRB-4

GCB-1

SSW-3

SSW-1

SSW-2

DRB-2

GCB-5.1

8

51

LINE 7

LINE 3

21

623

8

LINE 4

LINE 4

LINE 7

GCB-2

P6R

MMP-4P30

GCB-3

DRB-5

DRB-7DRB-6

GCB-4PBF-7PBF-8

6

N

0 6 km

ab

GCB-8

Figure 2. Index map showing location of cores used in this study, as well as location ofDomoracki's (1995) seismic lines 3, 4, 7, and epicenters (a, 6/9/1985; b, 8/4/1988) plotted byDomoracki (his p. 23). Locations where other seismic lines cross Lines 3, 4, 7 are indicated: 1,21, 8 cross line 7; 5 crosses line 4; and 6 and 8 cross line 3).


E-13

��

��

��

��

��

��

Deep Rock Metaigneous Complex

Pen Branch Metaigneous Complex

?

619±3.4 Ma (U-Pb z)

626.1 ± 4.4 Ma (U-Pb z)

PBF-7 extrusive equivalent?

Post-620 ? Persimmon Fork-Uwharrie equivalent?

Deep Rock Metavolcanic Suite

Crackerneck Metavolcanic Complex

DRB-1 metadiorite

Pen Branch Metavolcanic Suite

PBF-7 Metaplutonic Suite


pre 626 Ma volcanic pile

original nature of this contact unclear:fault or unconformity?

Figure 3. Schematic column of Appalachian litho-“stratigraphic” units beneath the US DOESavannah River Site.


E-14

Figure 4. Overturned limb structure of the Tinker Creek nappe from structural analysis ofover 3000m of basement core in the DRB well cluster.


E-15

Dunbarton basin border fault

Pen B

ranch fault

Deep Rock Metaigneous Complex


Pen Branch Metaigneous Complex

TR

lower grade (?)

Deep RockMetaigneousComplex ?

T i n k e rC r e e k n a p p e

Kiokee belt - Savannah River terrane of Maher and others (1991)interpreted to be more extensively Alleghanian remobilized

ca. 620 Ma basement

ca. 8 km

land surfacePiedmont-Coastal Plain unconf.

Figure 5. Schematic cross-section of Appalachian structure underlying the middle third of theSavannah River Site.


E-16

Figure 6. Comparison of the relative amounts of pinking or K-metasomatic alteration of PBFMetaplutonic Suite observed in PBF-7. Second and third numbers refer to box # and footage(depth from surface). XRF analysis of whole rock powders from this set was used to constructthe isocon diagram (Fig 8).


E-17

Pin

ked

Rb

Y

K O2

SiO2

Na O2

MnO

ZrV Sr

Ni Cu

Ba Fe O2 3

TiO2

Al O2 3

P O2 5

isochemical ?Sc

elements/ major element oxides enriched during pinking

elements/oxidesleached/removed during pinking

Unpinked (protolith)MgO

most enriched in K during pinking

"isochemical"slope of 1

CaO

Cr S

Figure 7. Isocon diagram showing changes in ratios of major and trace elements as a result ofpinking in PBF-7 gneisses of PBF Metaplutonic Suite. Discussed in text. Standard deviationsare shown. Slope of “1” represents isochemical conditions in theory; ratio of elements/majoroxides before/after pinking equals 1. Perhaps a line that drives through Ni, Zr, Sc, V, TiO2

supposedly immobile elements really represents isochemical conditions – ratios that have notbeen changed by metasomatism. Elements/major oxides that appear below this line have beenleached or depleted during the pinking event, elements/major oxides that appear above the linerepresent increasing additions of certain elements during pinking, in particular K2O, Y, Rb,SiO2, Na2O and (to a somewhat lesser extent ) Al2O3. 15 unpinked protolith (boxes 103, 107,108, 139, 143, 145, 149, 157, 159, 161, 165, 171, 173, 175, 177) and 8 pinked (101, 115, 115t, 167,171, 171-3568, 175, 176-3541) samples. Compare Figure 6.


E-18

Figure 8. PBF-7-16-1249' (and PBF-7-13) and GCB-2-648' illustrate the “greening” event. Abrick-red arkose (DRB-11-2741’) from within the Dunbarton basin is included to comparecolor. The matrix of the PBF-7 sample on a wet surface is grayish yellow green (5GY 7/2) onthe Munsell rock color chart. Microprobe analyses of a fibrous mineral that defines the matrixindicates that it is muscovite. Irregular margins of oxidized fanglomerate clasts have beenreduced and now appear cream colored. Reducing fluids penetrated along gneissosity. TheGCB-2 sample is saprolite, but shows a very similar texture. Coarser grained matrix iscompletely reduced. The ragged irregular margins of (brick-red) fine-grained clayey clastsshow incomplete penetration of reducing fluids parallel to foliation.


E-19

Figure 9. Pseudotachylyte hand specimens: a) DRB 3-13-1038, b) DRB 7-6-1394, c) DRB-7-1419’. Inspection of pseudotachylyte samples commonly shows that these features reactivatethin chlorite shear surfaces that dip less steeply than foliation, and show hangingwall up senseof displacement. It can be shown that most pseudotachylytes must post-date Alleghanian timebecause the rock they cut has been pinked, and pinking is a Triassic event. Somepseudotachylytes seem to form near felsic mafic contacts, and may reflect local shear stressaccumulations at these sites. PBF-7-89-2344: pseudotachylyte clearly postdates pinking, the“greening” as observed in crystalline rocks , and zeolite ±calcite veining (pale pinkish orange).


E-20

Notes


F-1

Significance of Soft Zone Sediments at the Savannah River Site

R. K. Aadland, Westinghouse Savannah River Co., Aiken, SC 29808W. H. Parker, Science Applications International Corp., Augusta, GA 30901R. J. Cumbest, Westinghouse Savannah River Co., Aiken, SC 29808

ABSTRACT

Rod drops and lost circulation during drilling,commonly occur in the middle Eocene marinecarbonates and clastics of the SanteeFormation, the Clinchfield Formation and to alesser extent the Dry Branch Formation.Historically the rod drop (including low blowcounts in SPT borings) was assumed toindicate the presence of “soft zones.” Untilrecently, sediment from soft zones as definedby SPT N-values, rod drops and lostcirculation was never recovered. Conventionalcoring methods was unable to recoversediment from the intervals. Therefore, softzone sediment, and certainly the precursorsediment to the final soft zone sediment wasunknown. Nearby sediments to the soft zones,whether they are carbonate-bearing tocarbonate-rich clastics or limestones havebeen observed to have undergone minor toextensive carbonate dissolution. Silicificationwas commonly noted where the precipitationof silica was minor to extensive, replacing andcementing much of the original sediment. Ithas generally been assumed that the carbonatein the sediment in the soft zones hasundergone extensive if not complete carbonatedissolution leaving behind vugular, moldyclastic debris. Thus a soft zone occurs in the“end member” sediment where the carbonatedissolution noted in varying degrees in thenearby non-soft zone sediments was complete.

Core recovered from a soft zone in theSantee/Clinchfield section in F-Area (GSA)using CPT technology, indicates that thesediment had not undergone extensivecarbonate dissolution. Instead, the precursorcarbonate sediment underwent extensive opal-CT (amorphous silica) replacement much like

the silica replacement/cementation thatcommonly occurs in overlying and nearbycompetent sediments.The silica wasprecipitated as 2-5 µm opal-CT lepispheresthat replaced carbonate mud (micrite) matrix,and precipitated as linings within molds andvugs and in micropores in the lime mudmatrix. Thus, opal-CT replacement of theoriginal carbonate sediment in the soft zoneinterval, is the crucial event in the formationof the soft zone. Where carbonate has beenreplaced by silica the soil properties of theresulting sediment is such that it has neithercohesion nor very much compressive strength.Here the tip resistance to the CPT push wouldbe low, rod drops would be encounteredduring conventional drilling and thesurrounding relatively unaltered carbonatesediment supports the overburden.

The extent and distribution of soft zones in theGSA was mapped. Soft zones are generallythicker and more concentrated wherecarbonate is thickest and most concentrated(has the highest average percentages).

1Information developed under Contract DE-AC09-89SR19015 with the U.S. Dept. ofEnergy

INTRODUCTION

The purpose of this paper is to provideinformation on the origin and extent of “softzones” encountered in the carbonate bearingstrata at the Savannah River Site (SRS). (seeFigure 1, in Aadland and others, 2000, thisvolume). Near the center of the SRS, middleEocene marine carbonates and clastics of theMcBean member of the Santee Formation, theUtley limestone member of the Clinchfield


F-2

Formation and to a lesser extent the GriffinsLanding member of the Dry Branch Formation(Fallaw and Price, 1995 and Aadland andothers, 1995), occur at depths between 100and 250 feet below the ground surface (seeFigure 2, in Aadland and others, 2000, thisvolume). Often found within the carbonaterich to carbonate bearing sediments,particularly in the upper third of this section,are “weak zones” interspersed in strongermatrix materials. These weak zones, whichvary in apparent thickness and lateral extent,have been variously termed “soft zones,” “thecritical layer,” “underconsolidated zones,”“bad ground,” and “void.” The preferred termused here is “soft zones”.

Rod drops, during drilling, commonly occur inthe carbonate-rich sediments in theSantee/Clinchfield and Dry Branch section(see Figures 2 and 4, in Aadland and others,2000, this volume). Historically the rod drop(including low blow counts in SPT borings)was assumed to indicate the presence of a“soft zone.” Due to the depth at which softzones occur (typically greater than 100 ftbelow the ground surface) there has been noopportunity for direct visual observations oftheir nature and geometry. Conclusions thathave been drawn on soft zone size are basedon the limited penetrations ofgeotechnical/geophysical borings and ConePenetrometer Tests (CPT).

The prevailing assumption of the causalmechanism for the rod drops in the soft zoneshas been dissolution of the carbonate-richsediments in the zone, resulting in vugularporosity where the drill rod meets little or noresistance to penetration. An alternativehypothesis for this phenomenon is that thedrill rod was pushed into uncemented sandswhere the overburden is supported by dense orsemi-cemented beds that overly theuncemented sands. A third alternativehypothesis from data accumulated in thisstudy (Parker, 1999 and 2000) is that softzones form where carbonate has been largely

replaced by opal-CT (amorphous silica). Theuncertainty of the origin and extent of softzones has lead to very conservativeengineering analyses of this subsurfacecondition.

Historically, it was assumed the soft zonesobserved in neighboring borings wereinterconnected to varying degrees, thus theywere grouted as an expedient way of resolvingany potential foundation stability issues(Corps of Engineers, 1952). This methodcontinued through the restart of the K-reactor(see Figure 1, in Aadland and others, 2000,this volume), where the project chose to grout“soft zones” in the Santee/Clinchfieldcarbonate interval beneath the cooling waterlines to resolve potential foundation stabilityissues (WSRC, 1992a, b; GeotechnicalEngineers Inc., 1991). The results of thateffort were carefully studied and it was foundthat the grout was not having the desiredeffect on the subsurface soft zones as waspreviously thought. The results showed thatsoft zones in adjacent wells were poorlyinterconnected. The grout traveled in thinsheets along preferential pathways oftenbypassing “soft zones” in adjacent wells. Softzones that existed prior to grouting stillexisted after grouting was completed. Thegrouting provided only limited benefit inreducing the potential settlement from the softzones.

A grout assessment study was conducted atthe In Tank Precipitation Facility (ITP) in HArea (Aadland and others, 1994; Syms, 1995and WSRC, 1995) (see Figure 1, in Aadlandand others, 2000, this volume). CPT and SPTborings identified locations and intervals inthe Santee/Clinchfield section where softzones were identified. Grout was injected intothe soft zone intervals. Wells were drilled andcored at distances from 5 to 40 feet from thegrout injection wells and the sedimentanalyzed to determine where the grouttraveled. It was observed that the grout residedin and mixed with the clastic sediments of the


F-3

overlying Dry Branch Formation. Indeed thegrout not only did not reside in nearbyinterconnected soft zones, it did not evenremain within the Santee/Clinchfield interval.

Soft Zone Identification Criteria

1. SPT N-values of 5 and less (includingweight of rods and weight of hammer) areconsidered to be primary indicators of softzones. Also, when these criteria are metduring SPT sampling, continuous SPTsampling is required until the soft zoneinterval is passed.

2. The CPT has none of the limitationsassociated with sampler mass or fluidinjection into the formation. Because ofits high data sampling frequency and arrayof test parameters and reliability of themeasurements, it is the primary tool fordetecting and locating soft zone intervals.The primary criterion is a tip resistance of15 tsf or less over a continuous 2 feetthick interval. Further, the sleeveresistance, pore pressure and shear wavevelocity measurements should be acquiredand evaluated.

3. The loss of drilling fluids is not areasonable indicator that soft materialshave been intercepted by the borehole.Also, it is not a reliable indicator of softzone size or extent. Fluid losses are aroutine occurrence in sediments below100 feet depth, and likely indicatehydrofracturing has occurred. In addition,identifying the location of leakage into theformation is notoriously difficult.Typically the only definitive informationthat can be obtained from fluid loss is thedepth to its first occurrence and the depthof any marked increases in loss (or totalloss of circulation).

4. During grouting programs it has beenfound that while significant quantities ofgrout were injected into the sediments, a

substantial quantity of the grout residesprobably in thin seams. Thus, the volumeof grout-take in a boring is not a reliableindicator of the location or extent of a softzone.

5. Rod drops recorded from historicaldrilling accounts require some knowledgeof the drilling rig, type of bit and pumppressure. Therefore, rod drops duringdrilling operations are not consideredprimary indicators of soft zones

6. Depending on the type of investigation,field conditions and other circumstances,other techniques including cross-holeseismic tomography or other geophysicaltechniques may be used. Thesetechniques must, however, be calibrated tomore standard soft zone identificationcriteria such as the CPT and SPTmeasurements.

Origin and Delineation of Soft Zones

Until recently, sediment from soft zones asdefined by SPT N-values, rod drops and lostcirculation was never recovered. Conventionalcoring methods were unable to recoversediment from the intervals. Therefore, softzone sediment, and certainly the precursorsediment to the final soft zone sediment wasunknown.

Nearby sediments (often fossiliferous) to thesoft zones, whether they are carbonate-bearingto carbonate-rich clastics or limestones havebeen observed to have undergone minor toextensive carbonate dissolution (see Figure12, in Aadland and others, 2000, this volume)(Thayer and others, 1994; Rine andEngelhardt, 1999). Silicification wascommonly noted where the precipitation ofsilica was minor to extensive (replacing andcementing much of the original sediment) (seeFigures 11, 15, 16 and 17, in Aadland andothers, 2000, this volume), (Thayer andothers, 1994; Rine and Engelhardt, 1999 and


F-4

Parker, 1999 and 2000). Only minor carbonatecementation has been noted in near-bysediments (see Figure 14, in Aadland andothers, 2000, this volume). It has generallybeen assumed that the carbonate in thesediment in the soft zones has undergoneextensive if not complete carbonatedissolution leaving behind vugular, moldyclastic debris. Thus a soft zone occurs in the“end member” sediment where the carbonatedissolution noted in varying degrees in thenearby non-soft zone sediments was complete.

Soft Zone Study at APSF

Evidence from the soft zone study at theAPSF suggests that the sediment in the roddrop/no recovery zone had not undergoneextensive carbonate dissolution. Instead, theprecursor carbonate sediment underwentextensive opal-CT (amorphous silica)replacement much like the silicareplacement/cementation that commonlyoccurs in overlying and nearby competentsediments (Figure 1; see Figure 11, inAadland and others, 2000, this volume).

A series of closely spaced CPT’s were pushedat the APSF Site (Figure 2) (WSRC, 1998) inthe northern portion of F Area (see Figure 1,in Aadland and others, 2000, this volume).One purpose of the investigation program wasto aid in establishing the stratigraphic andlateral distribution of soft zones in the area.Using the defining criteria of tip stress valuesless than 15 tsf to define the presence of softzones, it was determined that soft zones vs.non-soft zones occurred at comparablestratigraphic horizons in CPT pushes spacedless than 10 feet apart (Figure 2). An 8 feetthick soft zone occurred in theSantee/Clinchfield section immediatelybeneath the Santee unconformity. The softzone was not observed in the surroundingCPT’s that were pushed only a few feet away.The soft zone was continuously cored (usingthe newly available CPT coring tool) and

three lithologies were delineated (Parker,1999); sandy, biomoldic chert (see Figures 11and 15, in Aadland and others, 2000, thisvolume); siliceous sandy mud (see Figure 16,in Aadland and others, 2000, this volume) andterrigenous sand (Figure 17, in Aadland andothers, 2000, this volume).

The sandy, biomoldic chert and the siliceous,sandy mudstone were originally sandy limemud (micrite) supported fossiliferouswackestone and sandy, lime mudstones(Parker, 1999). The carbonate mud wasreplaced molecule for molecule withauthigenic opal-CT lepispheres andchalcedony without any loss in sedimentvolume (Figure 1; see Figure 11, in Aadlandand others, 2000, this volume). The originalsandy fossiliferous wackestone suggestsdeposition in clear, open-marine water ofnormal salinity on the inner to middle shelf(Parker, 1999). The opal-CT matrix wasoriginally lime mud (micrite) matrix, and itsabundance indicates deposition in quiet waterbelow normal marine wave base. In short theprecursor soft zone sediment was typical ofthe Santee/Clinchfield carbonate sedimentdeposited in the GSA.

The sediment in the soft zone is not a distinctmicrofacies separate and distinct from thesediments that comprise the surroundingcarbonate, but is diagenetically altered wherethe initial carbonate matrix was completelyreplaced with opal-CT (Figures 1 and 3; seeFigure 11, in Aadland and others, 2000, thisvolume). The silica was precipitated as 2-5µm opal-CT lepispheres that replacedcarbonate mud (micrite) matrix, andprecipitated as linings within molds and vugsand in micropores in the lime mud matrix. Theopal-CT formation postdates solution of thefossil shells because the opal-CT andchalcedony line the interiors of some molds.

Since the original carbonate sediment in thesoft zone interval underwent complete opal-CT replacement, the working hypothesis is


F-5

that opal-CT silica replacement of carbonatesediments, not carbonate dissolution is thecrucial event in the formation of the softzones. Where carbonate has been replaced bysilica the soil properties of the resultingdiagenetically altered sediment is such that ithas neither cohesion nor very muchcompressive strength. Here the tip resistanceto the CPT push would be low (less than 15tsf), rod drops would be encountered duringconventional drilling and the surroundingrelatively unaltered carbonate sedimentsupports the overburden.

Soft Zone Study R Area

Sediment from a 6 foot thick soft zoneencountered in the Santee/Clinchfield sectionin R Area (see Figure 1, in Aadland andothers, 2000, this volume) was continuouslysampled using CPT technology andpetrographically analyzed to corroborate thefindings at the APSF site in the GSA, (Parker,2000). The lithologies described were sandy,biomoldic chert and siliceous sandy mud verysimilar to the silica replaced sediments notedat the APSF site. Here, as at the APSF site,precursor calcareous wackestone andmudstone (micrite) was replaced by opal-CT,and supports the assertion that opal-CTreplacement of carbonate is the primarymechanism for soft zone development.

If replacement of carbonate by amorphoussilica is the controlling factor in thedevelopment of soft zones, then thecontrolling mechanisms for soft zonedevelopment would include (Folk and Pitman,1971; Meliva and Siever, 1988):

• Adequate supply of amorphous silicain the form of sponge spicules anddiatoms for the continuedprecipitation of the opal-CT.

• Adequate permeability to flush thesediment with the volumes of silicasaturated pore waters needed to

maintain the precipitation of thesilica.

• Chemistry and mineralogy of the hostsediment.

• Bulk pore-water ph.• Presence of organic matter.• Concentration of silica and certain

other ions, such as sulfate.

Assuming the above mentioned controls areavailable for silica replacement/cementationto occur; the spatial geometry of the silicified“areas” would be very irregular (Figure 2).The silicification of the enclosing sedimentwould follow and spread along beddingplanes, along microfractures of variedorientations, along corridors of locallyenhanced permeability etc. The resulting “softzone” could be in the form of irregularisolated pods, extended thin ribbons orstacked thin ribbons separated by interveningunsilicified parent sediment. Soft zonesencountered in one CPT could be absent in theneighboring CPT only a few feet away. Onlywhere silicification has spread far enoughaway from the bedding planes and/or fracturesalong which the silica replacement has takenplace, where all the intervening sediment isreplaced, would the soft zones be largeenough and coherent enough to pose aquestion for the siting of new facilities. In alllikelihood this would be a most uncommonevent.

Delineation of the Extent of CarbonateZones/Soft Zones

CPT logs were the primary data sourcequeried to determine the presence, geographicdistribution and stratigraphic position of thesoft zones in the GSA. Where the tip stresswas less than 15 tsf, a soft zone was defined.In addition, a search of the data from the fieldlogs from the SPT boring reports wasconducted for indications of "soft zones"encountered during drilling in the GSA. Thesedata included blow count N values, where Nvalues less than 5 are considered evidence of


F-6

soft zones. In addition, rod drop intervals, lostcirculation zones were considered to furtherdelineate the areal extent of the soft zones.These last two, rod drop intervals and lostcirculation zones proved of little value indelineating soft zones and were not includedin the final data base queries. The blow countdata from the SPT boring logs were queried asbackup to the primary data source, namely thetip stress measurements from the CPT pushes.

Three D profiles were created in EARTHVISION to illustrate the extent (stratigraphicand geographic distribution) of soft zones andcarbonate zones in the GSA. The geographicdistribution, and concentration of carbonatesediment in the Santee/Clinchfield section theGSA was analyzed by querying the footage ofcarbonate greater than or equal to 5% found inthe cored wells and borings (see Figure 20, inAadland and others, 2000, in this volume).Three areas stand out where several coredwells indicated appreciable thickness ofcarbonate. From east to west they include theITP area (#1 on Figure 1, in Aadland andothers, 2000, in this volume), the “H” Basinarea (#2 on Figure 1, in Aadland and others,2000, in this volume) and the Burial Groundsarea (#3 on Figure 1, in Aadland and others,2000, in this volume). Variable density anduneven distribution of the core and CPTboring data precludes defining these locationsas the only areas of carbonate concentration inthe GSA since the areal extent of each area isnot well constrained nor are other areas wheredata is sparse. The areas outlined are however,considered typical of the carbonate-richterrain in the GSA.

Query Results

Three D profiles were generated at the H AreaSeepage Basin (Figures 4 and 5) and theBurial Grounds (Figures 6 and 7) illustratingthe thickness and percentage of carbonate, andthe thickness and distribution of soft zones. Ingeneral where the carbonate is thickest, the

average carbonate percentage is greatest.Comparing the extent and lateral distributionof the soft zone “hits” with the distribution ofcarbonate indicates that soft zones aregenerally thicker and more concentratedwhere carbonate is thickest and mostconcentrated (highest average percentages).Indeed, throughout the GSA it is observed thatthe carbonate thickness and concentration isdirectly related to the isopach thickness of theSantee/Clinchfield interval. Where theSantee/Clinchfield interval is thick, carbonateis more concentrated, however, where theinterval is thin, carbonate thickness andconcentration is reduced. Since the thicknessand distribution of soft zones is closely linkedto the thickness and distribution of carbonate;those areas where the Santee/Clinchfieldsection is thin and where a great deal ofcarbonate has been removed would be areaswhere soft zones are less likely to be present.


F-7

REFERENCES

Aadland, R. K., Gellici, J. A. and P. A. Thayer,1995, Hydrogeologic Framework ofWest-Central South Carolina. State ofSouth Carolina Department of NaturalResources, Water Resources DivisionReport 5, 200p.

Aadland, R. K., Cumbest, R. J., Howe, K., Lewis,S. E., Price, v. Jr., Richers, D. E.,Wyatt, D. E. and K. S. Sargent, 1994,Geological Characterization in theVicinity of the In-Tank PrecipitationFacility (U). Westinghouse SavannahRiver Co., Site Geotechnical ServicesDept., U. S. Department Of EnergyReport, WSRC-TR-94-0373, Rev. 0,88p.

Geotechnical Engineers, Inc. 1991, “K-ReactorArea, Geotechnical Investigation forSeismic Issues, Savannah River Site(U), Volume 1: Seismic StructuralEngineering,” WSRC-TR-91-47, March1991.

Fallaw, W. C., and Price, V., 1995, Stratigraphy ofthe Savannah River Site and vicinity:Southeastern Geology, v. 35, p. 21-58.

Folk, R. L., and J. S. Pittman, 1971, Length-slowchalcedony: a new testament forvanished evaporates: Journal ofSedimentary Petrology, v. 41, p. 1045-1058.

Maliva, R. G., and R. Siever, 1988, Mechanismsand controls of silicification of fossils inlimestones: Journal of Geology, v. 96,p. 387-398.

Parker, W. H., 1999, Petrographic Analysis andMicrofacies Delineation for a Soft ZoneInterval Beneath F-Area, SavannahRiver Site, Aiken, South Carolina.Report prepared by SAIC undersubcontract AB6029N to DOE. 24p.

Parker, W. H., 2000, Petrographic Analysis andMicrofacies Delineation for a Soft zoneInterval Beneath R-Area, SavannahRiver Site, Aiken, South Carolina.Report prepared by SAIC undersubcontract AB6029N to DOE. 12p.

Rine, J. M. and D. W. Engelhardt, 1999, GeologicCharacterization of Core from BoringsHBOR-34 and HBOR-50, In TankProcessing (ITP) Area, Savannah Riversite, South Carolina. Report preparedfor DOE by the Earth Sciences andResources Institute, University of SouthCarolina, Columbia South Carolina.ESRI-USC Technical Report 99-3-F143. 16p.

Syms, F. H., 1995, In Tank Precipitation Facility(ITP) Grout assessment Final Report(U), Report prepared for DOE by SiteGeotechnical Services (SGS),Westinghouse Savannah River Co.,Savannah River Site, Aiken, SouthCarolina. Report No. WSRC-TR-950127, Rev.0

Thayer, P.A., Smits, A. D., Parker, W. H., Conner,K. R., Harris, M. K., and M. B.Amidon, 1994, Petrographic Analysis ofMixed Carbonate-ClasticHydrostratigraphic Units in the GeneralSeparations Area (GSA); SavannahRiver Site (SRS), Aiken, SouthCarolina, Westinghouse Savannah RiverCompany, U. S. Department of EnergyReport, WSRC-RP-94-54, 83p.

Westinghouse Savannah River Company, 1992a,“Integrated Geologic Analysis of the K-Reactor Area (U),” Reactor EngineeringDepartment, Document No. WSRC-TR-92-42-004, January, 1992.

Westinghouse Savannah River Company 1992b,“K-Area Soil Stabilization Program(KASS)," Summary Overview and


F-8

Volume I, Systems EngineeringDepartment, Document No. WSRC-TR-92-299, Rev. 0, July, 1992.

Westinghouse Savannah River Company, 1995,“In-Tank Precipitation Facility (ITP)and H-Tank Farm (HTF) GeotechnicalReport”, Site Geotechnical ServicesDepartment, Document WSRC-TR-95-0057, Rev. 0, September, 1995).

Westinghouse Savannah River Company, 1998,“APSF Packaging and Storage Facility

Soft Zone Settlement Analysis (U)”,Site Geotechnical Services Department,Calculation No. K-CLC-F-00034

US Army Corps of Engineers (COE), CharlestonDistrict, 1952, "Geologic EngineeringInvestigations, Savannah River Plant",Volumes I and 2, WaterwaysExperiment Station, Vicksburg, MS.


F-9

Figure 1. Diagrammatic representation of the interpreted diagenetic history of the siliceoussandy mudstone microfacies.


F-10

Figure 2. Photomicrograph of siliceous sandy mud microfacies, Santee/Clinchfield sectionfrom Core CPT-157S3 @ 108.35 ft. below land surface, APSF Site, F Area, GeneralSeparations Area, Savannah River Site. Plane-polarized light. Amorphous (Opal—CT) silicareplaced calcite micrite (mud) matrix.


F-11

Figure 3. This diagram is illustrating the stratigraphic and lateral distribution of soft zonesdue to silica replacement of carbonate in the GSA. Replacement/precipitation of silica occursalong bedding planes, microfracture systems, and zones of enhanced permeability resulting inhighly irregular pods, stringers and sheets of silica replaced carbonate (i.e., soft zones).

Carolina G


ield Trip G

uidebookW

SRC

-MS-2000-00606

F-1

2

Figure 4. This figure is a 3D profile illustrating the relative thickness and percentage of carbonate from core and SPT boringsat the Burial Grounds. The location of the Burial Grounds is outlined in Figure 1, in Aadland and others, 2000, in this volume.The upper bounding surface is the top Santee Unconformity surface and the lower bounding surface is the top Warley Hillsurface. The tan colored “sticks” indicate carbonate “hits” occurring in the SPT borings without regard to carbonate contentother than its presence at that particular depth in the boring. The percentage carbonate in the cored wells is color coded asfollows: Purple >65%; Blue 45-65%; Yellow 25-45%; Green 5-25%.

Carolina G


ield Trip G

uidebookW

SRC

-MS-2000-00606

F-1

3

.

Figure 5. This figure is a 3D profile illustrating soft zones at the Burial Grounds. The location of the Burial Grounds is outlined inFigure 1, in Aadland and others, 2000, in this volume. The upper bounding surface of the query is the top Santee Unconformitysurface and the lower bounding surface is the top Warley Hill surface. CPT logs were the primary source queried to determine thepresence, geographic distribution and stratigraphic position of the soft zones in the GSA. Where the tip stress was <15 tsf, a softzone was defined. Blow count N values from SPT logs were queried. N values <5 were considered evidence of soft zones. The blowcount data was queried as back up to the primary data source, namely the tip stress measurements from the CPT pushes.

Carolina G


ield Trip G

uidebookW

SRC

-MS-2000-00606

F-1

4

Figure 6. This figure is a 3D profile illustrating the relative thickness and percentage of carbonate from core and SPT borings atthe H Area Seepage Basin. The location of the Seepage Basin is outlined in Figure 1, in Aadland and others, 2000, in this volume.The upper bounding surface is the top Santee Unconformity surface and the lower bounding surface is the top Warley Hill surface.The tan colored “sticks” indicate carbonate “hits” occurring in the SPT borings without regard to carbonate content other than itspresence at that particular depth in the boring. The percentage carbonate in the cored wells is color coded as follows: Purple>65%; Blue 45-65%; Yellow 25-45%; Green 5-25%.

Carolina G


ield Trip G

uidebookW

SRC

-MS-2000-00606

F-1

5

Figure 7. This figure is a 3D profile illustrating soft zones at the H Area Seepage Basin. The location of the seepage Basin isoutlined in Figure 1, in Aadland and others, 2000, in this volume. The upper bounding surface of the query is the top SanteeUnconformity surface and the lower bounding surface is the top Warley Hill surface. CPT logs were the primary source queried todetermine the presence, geographic distribution and stratigraphic position of the soft zones in the GSA. Where the tip stress was<15 tsf, a soft zone was defined. Blow count N values from SPT logs were queried. N values <5 were considered evidence of softzones. The blow count data was queried as back up to the primary data source, namely the tip stress measurements from the CPTpushes.


F-16

Notes


G-1

Carbonate Sediments in the General Separations Area, Savannah River Site,South Carolina

R. K. Aadland, Westinghouse Savannah River Co., Aiken, SC 29808P. A. Thayer, University of North Carolina at Wilmington, Wilmington, NC 28403-3297W. H. Parker, Science Applications International Corp., Augusta, GA 30901J. M. Rine, Earth Sciences and Resources Institute, University of South Carolina, Columbia, SC 29208D. W. Engelhardt, Earth Sciences and Resources Institute, University of South Carolina, Columbia, SC 29208R. J. Cumbest, Westinghouse Savannah River Co., Aiken, S.C. 29808

ABSTRACT

Carbonates and carbonate-rich clastics in theGeneral Separations Area (GSA) are restrictedto the middle Eocene Warley Hill Formation,Santee and Clinchfield Formations and GriffinsLanding member of the Dry Branch Formation.The preponderance of carbonate occurs in theSantee/Clinchfield section. The calcareoussediment consists of calcareous sand, calcareousmud, limestone, sandy limestone, muddylimestone, and sandy muddy limestone. Theywere deposited in shallow generally subtidalmarine environments with the exception of theDry Branch oyster beds that were deposited in aback barrier marsh/lagoonal environment.

The carbonates and limey sands of theSantee/Clinchfield sequence underwent acomplex combination of deposition of diverselithologies, post-depositional dissolution-precipitation-replacement of calcite, silica andother minerals. The result is sediment often farremoved from its initial depositional fabric andmineralogy. Local dissolution is sometimes sopervasive that the precursor sediment is nolonger identifiable.

The thickness and concentration of carbonate inthe GSA is directly related to the overallthickness of the Santee/Clinchfield interval. Ingeneral where the Santee/Clinchfield interval isthickest the footage of carbonate is greatest andthe average carbonate content (percentage) inthe cores is greatest. Where the interval is thinthe thickness and percentage of carbonate in thecore is reduced.

Where the Santee/Clinchfield section is thickand where carbonate is concentrated, theoverlying “upland unit” and the TobaccoRoad/Dry Branch section is generallystructurally high. Where the section is thin andthe carbonate content is reduced or absent, theoverlying “upland unit” and Tobacco Road/DryBranch section is structurally low. Since the DryBranch/Tobacco Road/”upland unit” section isrelatively uniform in thickness over both thestructurally high and low areas and not drapedover the features, the removal of carbonate andthe thinning (collapse) of the Santee/Clinchfieldinterval occurred in post Tobacco Road/”uplandunit” time.

1Information developed under Contract DE-AC09-89SR19015 with the U.S. Dept. ofEnergy

Geologic Setting

Sediment underlying the Savannah River Site(SRS) (Figure 1) forms a wedge of strata thatthickens from about 700 ft in the northwest toalmost 1400 ft at the southeastern boundary ofthe site (see Figure 2, in Introduction to thisvolume, Wyatt, ed.). measuring on average 900ft in thickness in the GSA (Figure 2). Regionaldip is to the southeast and decreases upwardfrom about 48 ft/mi at the base of theCretaceous-aged strata to about 15 ft/mi at thetop of middle Eocene-aged strata.

Late Cretaceous sediments of possibleConiacian/Turonian (?) to Santonian throughMaastrichtian age (91 to 66 Ma) (Fallow andPrice, 1995; Aadland and others, 1995) average


G-2

550 feet in thickness in the GSA (Figure 2). Thesequence consists of quartz sand; pebbly sandand sandy clay generally deposited in lower toupper delta plain environments (Figure 3).Following a period of erosion of the Cretaceoussection, Tertiary sediments were deposited whenseas transgressed onto the Cretaceous-Tertiaryunconformity surface. Tertiary sedimentsaverage about 350 feet in thickness in the GSA(Figure 2), range in age from Early Paleocene toMiocene (?)/Oligocene (?) and were depositedin fluvial to marine shelf environments (Figure3).

Four regionally significant unconformities aredefined in the Tertiary section (Figure 4). Frombase upward, they include the “Cretaceous-Tertiary” unconformity, the “LangSyne/Sawdust Landing” unconformity, the“Santee” unconformity, and the “Upland”unconformity). Four sequence stratigraphic units(labeled Sequence I, II, III and IV) that arebounded by the unconformities have beendelineated (Figures 2 and 4).

The “calcareous zone” in the GSA includescalcareous strata belonging to the Warley HillFormation, Santee and Clinchfield Formationsand Griffins Landing member of the Dry BranchFormation (Figure 4). The preponderance ofcarbonate occurs in the Santee/Clinchfieldsection (Thayer and others, 1994), and consistsof calcareous sand, calcareous mud, limestone,sandy limestone, muddy limestone, and sandymuddy limestone.

Chronology of Tertiary Geologic Events in theGSA

The chronology of events and the depositionalsetting that characterized deposition of theTertiary sequence in the GSA is presented here.

Oldest

• The Snapp Formation (Figures 2 and 4)and older Paleocene sediments(Sequence Stratigraphic unit I) were

deposited in shallow clastic shelf anddelta plain environments (Figure 3).Shallow shelf (platform) carbonatedeposition was restricted to coastalareas of South Carolina at this time.

• Sequence I deposition was followed byerosion of the section (LangSyne/Sawdust Landing unconformity)(Figure 4).

• A rise in sea level initiated deposition ofthe Congaree/Fourmile sands (SequenceStratigraphic unit II) in shoreline toshallow shelf depositional environments(Figure 3). Shallow shelf (platform)carbonate deposition rapidly expandedlandward and reached the southernboundary of SRS, but remained to thesouth and east of the GSA (see Figure 2,in Introduction to this volume, Wyatted.).

• Sea level continued to rise resulting indeposition of the Warley Hill clays(Sequence Stratigraphic unit II) indeeper shelf depositional environments(Figures 3 and 4).

• A lowering of sea level and/or anincreased rate of sediment supplyresulted in deposition of the Santee andClinchfield carbonates and clastics inshallow shelf depositional environmentsin the GSA (Figure 5). This completeddeposition of the Sequence Stratigraphicunit II sedimentary package (Figures 2and 4).

• A retreat of the sea resulted insubstantial erosion of theSantee/Clinchfield (Sequence II)sediments (Santee unconformity).

• Deposition of the Dry BranchFormation (Sequence III), including theGriffins Landing limestone member,was deposited over theSantee/Clinchfield section in shorelineto lagoonal/marsh depositionalenvironments (Figure 5). The shorelineretreated from its position in thenorthern part of SRS, where it waslocated during Santee/Clinchfield time,


G-3

to the central part of SRS in Dry Branchtime. Progradation of the shorelineenvironments to the south resulted inthe sands and muddy sands and locallythe oyster banks of the Griffins Landingmember of the Dry Branch beingdeposited over the shelf carbonates andclastics of the Santee/Clinchfieldsequence (Figure 5). The shallow shelf(platform) carbonates that had reachedtheir most northwesterly updip positionin Santee/Clinchfield time, reaching theGSA, retreated far to the south and eastof SRS in Dry Branch time (see Figure2, in Introduction to this volume, Wyatted.).

• Deposition of the Dry Branch clasticswas followed by deposition of theTobacco Road Formation (Sequence III)probably in moderate to high-energylagoonal/open bay environments(Figures 4 and 5).

• A dramatic increase in sediment supplyand a retreat of the shoreline towardsthe southeast, resulted in deposition ofthe "upland unit” sand and gravel(Sequence IV) in fluvial to upper deltaplain environments (Figure 3).

Youngest

In conclusion, the deposition of shallow shelfcarbonates generally remained to the south andeast of the Savannah River Site during theTertiary (see Figure 2, in Introduction to thisvolume, Wyatt ed.). Only during middle EoceneSantee/Clinchfield time did deposition of thecarbonates extended as far inland as the GSA(Figures 4 and 5).

Stratigraphy of the Carbonate Sediments inthe GSA

The carbonates and carbonate-rich clastics areessentially restricted to two intervals in theTertiary section in the GSA; the Dry BranchFormation and the underlying Santee andClinchfield formations. The upper most interval

includes the carbonates of the Griffins Landingmember of the Dry Branch Formation foundbelow the “tan clay” beds that occur near themiddle of the Dry Branch (Figures 2 and 4). Theisolated carbonate patches of the GriffinsLanding are the oyster banks that typically formin the back barrier marsh zone behind a barrierisland system (Figures 4 and 5). Underlying theDry Branch, directly below the regionallysignificant Santee unconformity, is the UtleyLimestone member of the ClinchfieldFormation. Without the benefit of detailedpetrographic and paleontological analysis, theUtley carbonates cannot be systematicallydistinguished from the carbonates of theMcBean member of the underlying SanteeFormation. Thus the carbonate-rich sedimentsdeposited in shallow shelf environments foundbetween the Santee unconformity (Figure 4),and the Warley Hill Formation are referred to asthe Santee/Clinchfield sequence in this report.The preponderance of the carbonate observed inthe GSA occurs in the Santee/Clinchfieldinterval. Upon occasion carbonate mud andwackestone deposited in deeper shelfenvironments is observed in the underlyingWarley Hill Formation (Figure 4).

Distribution and Depositional Environmentsof Carbonates at SRS

The regional distribution of theSantee/Clinchfield carbonate-rich sequence isillustrated in Figure 6. Carbonate content in thesequence is minimal in northwestern SRS andpredominates near the southeast boundary of thesite. The GSA is in that part of the mixedclastics/carbonate zone where the clasticsediments generally constitute a greaterpercentage of the section than the carbonates. Innorthern SRS the depositional equivalent of theSantee/Clinchfield sediments are mostly sandsand muddy sands deposited in shoreline to lesserlagoonal and tidal marsh environments (Figure5). The sands and muddy sands constitute theTinker Formation (Fallaw and Price, 1995). Inthe central SRS the mixed carbonates and


G-4

clastics of the McBean member of the SanteeFormation and the Utley member of theClinchfield Formation (Figure 2) were depositedin middle marine shelf environments. The resultis a varied mix of lithologies ranging fromcarbonate-rich sands and mud to sandy andmuddy limestone (mostly micrites andwackestones, and to a lesser extent packstonesand grainstones) (Thayer and others, 1994)(Figure 7). In southern SRS the Blue Bluffmember of the Santee/Clinchfield sequence wasdeposited further offshore, further removedfrom riverine clastic input into the shelfenvironment resulting in deposition of carbonatemud (micrite) (Figure 5).

Approximately 40-50% of the wells that drilledthrough the Santee/Clinchfield interval in theGSA penetrated quantities of carbonate rangingfrom 5-78% of the sediment sampled. Thayerand others, 1994, conducted a binocularmicroscope and thin-section examination of coresamples from the Santee/Clinchfield section anddelineated five carbonate, two terrigenous, andone siliceous microfacies. The eight microfaciesinclude: 1) lime mud (micrite); 2) skeletalwackestone; 3) skeletal packstone 4) skeletalgrainstone; 5) microsparite; 6) terrigenous mud;7) terrigenous sand and sandstone; and 8)siliceous mudstone. Skeletal wackestone is thedominant microfacies in the stratigraphicinterval, and is found throughout the "calcareouszone" beneath the GSA, forming 55 percent ofthe sampled population. The carbonate-richclastics and limestone in the Santee/Clinchfield-Dry Branch stratigraphic interval vary widely inthickness. Calcareous sands range from 2 feet to33 feet in thickness, the sandy and muddylimestones range in thickness from 3 feet to 30feet. This complex sedimentary packageunderlies most of the facilities in the GSAincluding the burial grounds, the tank farms, andthe seepage basins (Figure 1).

The presence of glauconite along with a normalmarine fauna including foraminifers, molluscs,bryozoans, and echinoderms, indicates that theSantee/Clinchfield limestones and limy

sandstones were deposited in clear, open-marinewater of normal salinity on the inner to middleshelf (Figures 3 and 5) (Thayer and others,1994; Parker, 1999 and 2000). The abundanceof carbonate mud (micrite) in the limestonessuggests deposition in quiet water below normalmarine wave base. The presence of abraded andwell-worn skeletal grains indicates that bottomtransport by currents or storm-generated wavesalternated with quiet-water conditions in whichthe sediments accumulated.

The presence of varying percentages of fine,subangular quartz sand and terrigenous mud inthe carbonate lithologies indicates the varyingbut ever presence proximity of the carbonates toriverine input (Figures 3 and 4). Viewing theSantee/Clinchfield sedimentary package parallelto the shoreline (Figure 5), the carbonate-richsediments would be concentrated in the areasfurthest removed from the tidal inlets at theshoreface where clastic sediments supplied byriverine input is concentrated. The clastic-richsediments on the other hand would concentrateopposite the tidal inlet areas where clasticsediment is more readily available. The lateralfacies transition of the sediments in the subtidalshelf environment from carbonate-rich toclastic-rich lithologies is therefore gradual andmeasures in the thousands of feet. Shiftinglocations of the tidal inlets at the shoreline hasresulted in a complex sedimentary package(Figure 7) where facies gradually transitionfrom one lithology to another both laterally andvertically. Therefore both vertical and laterallithologic variability in the Santee/Clinchfieldsequence is the result.

The Dry Branch Formation (Sequence III),including the Griffins Landing limestonemember and the Tobacco Road Formation, weredeposited over the Santee/Clinchfield section inshoreline to lagoonal/marsh depositionalenvironments (Figure 5). The DryBranch/Tobacco road sedimentary package inthe GSA is interpreted to be upward fining lowsinuosity distributary channel and over bankdeposits that retain there depositional fabric and


G-5

thickness over several miles as the distributarychannels swept across the lagoonal/marshenvironment. Thus the primary depositionalfabric (several upward fining sequences fromcoarse channel sand to overbank silty mud) andthe uniformity in the thickness of the sequencecan be traced for miles in the central part of theSRS.

Locally however, the contact between carbonatesediments and laterally comparable clasticsediments is sharply drawn, occurring overdistances of only a few feet. Here the carbonateis interpreted to be the remains of morecontinuously deposited beds (Thayer and others,1994: Parra and others, 1998) where postdepositional dissolution and removal ofcarbonate left isolated remnants over time. Theoverlying Dry Branch/Tobacco Road (”Upland”unit ?) clastics being uniform in depositionalfabric and thickness over extensive lateraldistances, are interpreted to have slumpeddownward compensating for the removal of thecarbonate (Figure 8). The slumping into the lowsleft by the removal of the Santee/Clinchfieldcarbonate did not result in the draping andthinning of the overlying Dry Branch/TobaccoRoad sediments over the carbonate rich structuralhighs. That would have occurred if thedissolution and differential collapse of theSantee/Clinchfield section predated DryBranch/Tobacco Road deposition. On thecontrary, the Dry Branch/Tobacco Road sectionis uniform in thickness over both the underlyingstructural highs and lows. Thus the consolidation(collapse) of the Santee/Clinchfield section dueto the removal of the carbonate, post-datesdeposition of the Dry Branch and Tobacco Roadsands and possibly the “upland unit” sands andgravel as well.

Diagenesis of the Carbonate Sediments andFormation of Soft Zones

The carbonate-bearing and carbonate-richSantee/Clinchfield sediments underwentextensive diagenesis due to dissolution and/orprecipitation of calcite, silica and other

minerals. Post-depositional events affecting thelimestones and limey sandstones occurred inboth the marine and freshwater phreatic(saturated) environments.

The Santee/Clinchfield sediments weredeposited in the subtidal “normal” marineenvironment where marine phreatic watersbathed the newly deposited sediment.Diagenetic events occurring in the marinephreatic environment included: 1) boring andmicritization of skeletal grains to form micriteenvelopes (Figures 9 and 10), mainlypelecypods, and 2) precipitation of subseacements, including pyrite and glauconite.Glauconite precipitates under slightly reducingconditions in water depths greater than 10 m; itspresence indicates slow sedimentation rates(Nystrom and others, 1989, 1991) in marineshelf environments. The pervasive presence ofglauconite in the Santee/Clinchfield sedimentscorroborates the interpretation that thesediments were deposited in subtidal marineshelf environments based on fossil content andsediment fabric analysis. Pyrite precipitates andis associated with organic matter within thecarbonate mud (micrite) matrix of thelimestones indicating that it formed underlocalized reducing conditions in areascontaining putrifying tissue where sulfate-reducing bacteria were abundant (Scoffin,1987).

The limestones commonly underwent minorpost-depositional compaction within the firstfew hundred feet of burial as shown by thepresence of broken pelecypod and other shells(Figure 11).

The Santee/Clinchfield sediments underwent atleast two episodes of fresh water flushing. Theinitial episode of dissolution/erosion of themiddle Eocene Santee/Clinchfield sedimentsoccurred soon after burial following the drop insea level at the Santee unconformity (Figure 4).The diagenetic changes that occurred thatmaterially effected the mineralogy and/or the


G-6

engineering properties of the sediments includethe following:

• Skeletal grains, mostly pelecypods,dissolved to create molds (Figures 9, 10and 11). The molds underwent solution-enlargement to form vugs (Choquette andPray, 1970) and channels (Figure 12).This was an important porosity-creatingprocess in the limestones and limeysandstones because pelecypods formed upto 20 percent of the originalunconsolidated sediment (Thayer andothers, 1994). Locally the molds and vugscollapsed due to compaction following thedissolution of the fossil debris (Figure13).

• Some of the carbonate derived fromdissolution of the fossil shell debrisdescribed above was precipitated nearbyas pore-reducing or pore-filling calcitecement (Figure 14). Rine and Engelhardt,1999, noted that at least two generationsof spar cements appear to be present in thecarbonates from core at the ITP site in HArea (Figure 1), both related to separateepisodes of dissolution. Carbonateprecipitation however was a minorprocess in the limestones and limeysandstones. Most of the carbonatematerial released from dissolution offossil shells was exported out of thesystem as indicated by the small numberof pores that are lined or filled with sparrycalcite cement. Where it occurredhowever, the carbonate was often lithified(cemented) into hard limestone.

• Biogenic opal-A in the form of spongespicules and diatom valves dissolved tocreate silica-rich pore water. Some of thesilica was precipitated as 2-5 µm sizedopal-CT lepispheres that replacecarbonate mud (micrite) matrix (Figure11), precipitated within molds and vugsand in micropores in the lime mud matrix.Opal-CT formation generally postdatessolution of skeletal grains because thesilica lines the interiors of some molds.

The amount and selectivity of the Opal-CT andchalcedony precipitation and replacement ofcarbonate is variable in the Santee/Clinchfieldsediments (Parker, 1999). Precipitation of theopal-CT silica varies from rare to pervasivewhere the entire precursor carbonate sedimentwas 100 percent replaced by silica (Figures 15and 16). The main limiting factor for silicaprecipitation and replacement of carbonatedepends on the quantity and availability ofamorphous biogenic opal-A (sponge spiculesand diatom valves) that dissolved to createsilica-rich pore water for mobilization andprecipitation of the silica into the nearbysediment (Maliva and Seiver, 1988). Thesilicification process ceased when the silicasupply (diatom and sponge-spicule biogenicopal-A) was exhausted.

Fibrous chalcedony precipitates locally as rimcements on detrital quartz grains (Figures 15and 17). Cement stratigraphy indicates thatchalcedony formed after precipitation of opal-CT.

Following the episode of fresh water flushingand exposure of the Santee/Clinchfieldsediments at the Santee unconformity, TobaccoRoad and Dry Branch sediments were depositedin shoreline environments in the GSA. Now theunderlying Santee/Clinchfield section would bealternately flushed by the meteoric phreatic zone(fresh water)/marine phreatic zone (salt water)mixing zone (Figure 18). Here furtherdissolution of the Santee/Clinchfield carbonatesoccurred albeit at a reduced rate, and siliceousreplacement and cementation of the clastics andcarbonates progressed until the supply of silicawas exhausted. Eventually, as Dry Branch andespecially Tobacco Road sedimentationcontinued, the Santee/Clinchfield sedimentswere buried more deeply and the mixing zonewaters could no longer reach the section (Figure18). Here the sequence remained in the marinephreatic environment for extended lengths oftime and diagenetic alteration of the sedimentswas reduced to a minimum or essentially ceased.


G-7

A dramatic increase in sediment supply and aretreat of the shoreline seaward to the southeast,resulted in deposition of the "upland unit” sandand gravel (Sequence IV) in fluvial to upperdelta plain environments (Figure 3). This is thetime frame of the further dissolution andremovable of the Santee/Clinchfield carbonatedue to renewed fresh water flushing of thesequence. Here the overlying DryBranch/Tobacco Road (”Upland” unit ?) clasticsare interpreted to have slumped downwardcompensating for the removal of the carbonate(Figure 8). Slumping into the lows left by theremoval of the Santee/Clinchfield carbonatecorroborates the assertion that the consolidationof the interval post-dates deposition of the DryBranch and Tobacco Road sands and possiblythe “upland unit” sand and gravel.

Rine and Engelhardt, 1999, noted vugs filledwith sand and silt sized limestone fragments andquartz grains present in thin sections from theITP site in H Area (Figures 1 and 19) andconcluded that the detritus was probably derivedfrom leaching and collapse of rock adjoining thevug. The voids not containing the detritus areprobably molds of single shells and are notconnected to other large pores. These limestoneand quartz debris filled vugular limestonesformed during the “collapse” and consolidationof the Santee/Clinchfield carbonates in “uplandand/or post upland unit” time.

In conclusion, two episodes of freshwaterflushing of the Santee/Clinchfield sectionresulting in two (or more) stages of fossil shelldissolution are hypothesized based on themultiple episodes of erosion/dissolution of thesection. The first occurred at the time of theSantee unconformity, the second at the time ofdeposition of the “upland unit” following theUpland unconformity. During the interim periodbetween the two primary episodes of fresh waterflushing of the Santee/Clinchfield section,dissolution of carbonate and precipitation andreplacement of carbonate by silica continuedalbeit at a slower rate during deposition of theDry Branch and Tobacco Road sands.

The carbonates and limey sands of theSantee/Clinchfield sequence underwent acomplex combination of deposition of diverselithologies, post-depositionaldissolution/precipitation/replacement of calcite,silica and other minerals. The result is sedimentoften far removed from its initial depositionalfabric and mineralogy. Local dissolution is sopervasive that the precursor sediment is nolonger identifiable.

Extent of Carbonate Zones

A database was established that cataloged alldata retrieved from core, geophysical logs, CPTlogs and geotechnical (SPT) borings related tothe stratigraphy and the presence of carbonatesin the Dry Branch-Santee/Clinchfield interval.The primary sources of data for mapping theextent (stratigraphic and geographic distributionand “concentration”) of the carbonate zoneswere binocular and petrographic examination ofcore retrieved from wells drilled in the GSA,SPT boring logs and borehole geophysical logs.The percentage of carbonate was recorded on afoot by foot basis from core described in thecore lab. The thickness, distribution andpercentage of carbonate were mapped from thecore data. Mapping the stratigraphic and lateralextent of the carbonate was augmented withlithologic data interpreted from the geophysicaland SPT boring logs.

A file was assembled with stratigraphicinformation from core and geophysical log datathat includes picks for the top of the Dry BranchFormation, top of the Santee/Clinchfield section(i.e., top surface of the Santee unconformity)and top of the Warley Hill Formation for theGSA area. The top surfaces were mapped andused to constrain the data query for the presenceof calcareous sediments to the stratigraphicinterval between the top of the Santeeunconformity and the top of the Warley HillFormation where the zones are located.


G-8

RESULTS

The geographic distribution, and concentrationof carbonate sediment in the Santee/Clinchfieldsection in the GSA was analyzed by queryingthe footage of carbonate greater than or equal to5% found in the cored wells and SPT borings.Figure 20 is a 3D profile illustrating the relativethickness and percentage of carbonate from coreand SPT borings throughout the area.

Three areas stand out where several cored wellsindicated an appreciable thickness of carbonate.From east to west they include the H Area ITPsite ( #1 on Figure 1), the H Area Seepage Basin(#2) and the Burial Grounds area (#3). In “F”Area at the APSF site (#4) and the F Basin area(#5) extensive SPT boring data is availableindicating concentrations of carbonate. Neitherarea has concentrated core data that is needed toverify the SPT findings. In other areas carbonateis encountered in isolated cored wells. Here theextent of the carbonate is indeterminate. Theselarge areas of sparse data precludes making theassumption, as was done in the past (Aadlandand others, 1991; Richardson 1994; Thayer andothers, 1994 and Aadland and others, 1994), thatthe carbonate is concentrated only in the areasindicated on the profile.

The average percentage of carbonate in eachcore was mapped. In general where thecarbonate is thickest, the average carbonatepercentage is greatest. The thickness of theSantee/Clinchfield section (interval between theSantee unconformity and the Warley HillFormation) was mapped in the GSA. Carbonatethickness and concentration is directly related tothe isopach thickness of the Santee/Clinchfieldinterval. Where the Santee/Clinchfield intervalis thick carbonate is more concentrated, wherethe interval is thin carbonate thickness andconcentration is reduced. It is further observedthat where carbonate is concentrated in theSantee/Clinchfield section the overlying “uplandunit”, Tobacco Road/Dry Branch section isgenerally structurally high, and where thecarbonate content is reduced or absent the

overlying “upland unit”, Tobacco Road/DryBranch section is generally structurally low.This fact, and the uniformity in thickness of theDry Branch/Tobacco Road section in the areamapped, indicates that the removal of carbonateand the thinning of the Santee/Clinchfieldinterval occurred in post Tobacco Road(“upland unit”) time.

Significant Findings and Observations

• The carbonates and limey sands of theSantee/Clinchfield sequence underwenta complex combination of deposition ofdiverse lithologies, post-depositionaldissolution/precipitation/replacement ofcalcite, silica and other minerals. Theresult is sediment often far removedfrom its initial depositional fabric andmineralogy. Local dissolution issometimes so pervasive that theprecursor sediment is no longeridentifiable.

• The thickness of the Santee/Clinchfieldinterval in the GSA is variable. Thethickness and concentration ofcarbonate is directly related to thethickness of the Santee/Clinchfieldinterval. In general where theSantee/Clinchfield interval is thickestthe footage of carbonate is greatest andthe average carbonate content(percentage) in the cores is greatest.Where the interval is thin the thicknessand percentage of carbonate in the coreis reduced.

• It is further observed that where theSante/Clinchfield section in thick andwhere carbonate is concentrated, theoverlying “upland unit” and TobaccoRoad/Dry Branch section is generallystructurally high (Figure 8). Where thesection is thin and the carbonate contentis reduced or absent, the overlying“upland unit” and Tobacco Road/DryBranch section is structurally low butuniform in thickness with the section inthe structurally high areas. This


G-9

indicates that the removal of carbonateand the thinning of theSantee/Clinchfield interval occurred inpost Tobacco Road/”upland unit” time.

• All degrees of carbonate dissolutionfrom minor to complete are observed inthe Santee/Clinchfield sediments.

• Two primary episodes of freshwaterflushing of the Santee/Clinchfieldsection resulting in two (or more) stagesof fossil shell dissolution arehypothesized based on the multipleepisodes of erosion/dissolution of thesection. The first occurred at the time ofthe Santee unconformity, the second atthe time of deposition of the “uplandunit” following the uplandunconformity. During the interim periodbetween the two primary episodes offresh water flushing of theSantee/Clinchfield section, dissolutionof carbonate and precipitation andreplacement of carbonate by silicacontinued albeit at a slower rate duringdeposition of the Dry Branch andTobacco Road sands.

• The second episode of fresh waterflushing in “upland unit” time resultedin the further dissolution and removableof the Santee/Clinchfield carbonate.Here the overlying Dry Branch/TobaccoRoad (”Upland unit” ?) clastics areinterpreted to have slumped downwardcompensating for the removal of thecarbonate (Figure 8). Slumping into thelows left by the removal of theSantee/Clinchfield carbonatecorroborates the assertion that theconsolidation of the interval post-datesdeposition of the Dry Branch andTobacco Road sands and possibly the“upland unit” sand and gravel.

REFERENCES

Aadland, R. K., Gellici, J. A. and P. A. Thayer,1995, Hydrogeologic Framework ofWest-Central South Carolina. State ofSouth Carolina Department of NaturalResources, Water Resources DivisionReport 5, 200p.

Aadland, R. K., Cumbest, R. J., Howe, K.,Lewis, S. E., Price, v. Jr., Richers, D.E., Wyatt, D. E. and K. S. Sargent,1994, Geological Characterization inthe Vicinity of the In-TankPrecipitation Facility (U).Westinghouse Savannah River Co.,Site Geotechnical Services Dept., U.S. Department Of Energy Report,WSRC-TR-94-0373, Rev. 0, 88p.

Aadland, R. K., Harris, M. K., Lewis, C. M.,Gaughan, T. F. and T. M. Westbrook,1991, Hydrostratigraphy of theGeneral Separations Area, SavannahRiver Site (SRS), South Carolina,USDOE Report WSRC-RP-91-13,Westinghouse Savannah RiverCompany, Aiken, SC., 29808, 114 p.

Choquette, P. W. and L. C. Pray, 1970,Geological Nomenclature andClassification of Porosity inSedimentary Carbonates. AmericanAssociation of Petroleum GeologistsBull 54, p. 207-250.

Fallaw, W. C., and Price, V., 1995, Stratigraphyof the Savannah River Site andvicinity: Southeastern Geology, v. 35,p. 21-58.

Maliva, R. G., and R. Siever, 1988, Mechanismsand controls of silicification of fossils


G-10

in limestones: Journal of Geology, v.96, p. 387-398.

Parra, J.O., Price, V., Addington, C., Zook, B.J.,and Cumbest, R.J., 1998, Interwellseismic imaging at the SavannahRiver Site, South Carolina:Geophysics, 63: 1858-1865.

Parker, W. H., 1999, Petrographic Analysis andMicrofacies Delineation for a SoftZone Interval Beneath F-Area,Savannah River Site, Aiken, SouthCarolina. Report prepared by SAICunder subcontract AB6029N to DOE.24p.

Parker, W. H., 2000, Petrographic Analysis andMicrofacies Delineation for a Softzone Interval Beneath R-Area,Savannah River Site, Aiken, SouthCarolina. Report prepared by SAICunder subcontract AB6029N to DOE.10p.

Richardson, G. A., 1994, Petrographic andHydrogeologic Characteristics ofEocene Calcareous Strata of the

General Separations Area, SavannahRiver Site, South Carolina; MastersThesis, University of North Carolina,Wilmington, 59p.

Rine, J. M. and D. W. Engelhardt, 1999,Geologic Characterization of Corefrom Borings HBOR-34 and HBOR-50, In Tank Processing (ITP) Area,Savannah River site, South Carolina.Report prepared for DOE by the EarthSciences and Resources Institute,University of South Carolina,Columbia South Carolina. ESRI-USCTechnical Report 99-3-F143. 16p.

Thayer, P.A., Smits, A. D., Parker, W. H.,Conner, K. R., Harris, M. K., and M.B. Amidon, 1994, PetrographicAnalysis of Mixed Carbonate-ClasticHydrostratigraphic Units in theGeneral Separations Area (GSA).Westinghouse Savannah RiverCompany, Savannah River Site (SRS),Aiken, South Carolina, Reportprepared for the U. S. Department ofEnergy, WSRC-RP-94-54, 83p.


G-11

Figure 1. Location map of the General Separations area (GSA) at the Savannah River Site. TheBurial Grounds, H Area Seepage Basin and the H Area ITP Site used in the Carbonate and softzone query are outlined in red.


G-12

Figure 2. Lithostratigraphic and Sequence Stratigraphic Units, at the Savannah River Site.


G-13

Figure 3. Spatial relationships of depositional environments typical of the Tertiary sediments atSRS.


G-14

Figure 4. Lithologic data retrieved from core, and the summary of insoluble residue andcarbonate WT % data, from Well HSB-TB, H Area, General Separations Area, Savannah Riversite.

Carolina Geological Society, 2000 Annual Field Trip Guidebook W

G-15

Figure 5. Spatial relationships of depositional environments typical of the Dry Branch andSantee/Clinchfield sediments at SRS. Progradation seaward puts the younger tidalflat/marsh/shoreline (inner shelf) sediments of the Dry Branch Formation over the older middleshelf sediments of the Santee/Clinchfield Formations in the General separations Area at SRS.


G-16

Figure 6. Regional distribution of carbonate in the Dry Branch-Santee/Clinchfield sequence atSRS.


G-17

Figure 7. Shifting locations of tidal inlets at the shore face would result in shifting positions of theclastic-rich/carbonate-rich sediments deposited out on the subtidal shelf. The result is a complexmixture of sediment types that gradually transition from one lithology to another.


G-18

Figure 8. Slump faulting at the H Area ITP Site. Dissolution of carbonate in the Santee/Clinchfieldsediments especially the Clinchfield, resulted in slumping and collapse of the section with DryBranch, Tobacco Road and ‘upland unit” sediments slumping into the resulting low areas.


G-19

Figure 9. Diagrammatic summary showing steps in solution and precipitation of calciumcarbonate by fresh water.


G-20

Figure 10. Photomicrograph of quartz-rich skeletal packstone from the Burial Grounds, GSA,SRS that has been diagenetically altered to sandy, molluscan-mold microsparite. From theSantee/Clinchfield section in Well BGO-14A, 122 ft. Crossed nicols. Blue area is void space filledwith blue epoxy glue.


G-21

Figure 11. Diagrammatic representation of the interpreted diagenetic history of the sandybiomoldic chert microfacies.


G-22

Figure 12. Extensive dissolution of fossil fragments in a skeletal grainstone (?) resulting in vugularporosity. Only micrite envelopes of the shell fragments remain. From the Santee/Clinchfieldsection, Well HBOR-50 @ 165 ft. H-Area ITP Site, General Separations Area, Savannah RiverSite. Void space filled with blue epoxy glue.


G-23

Figure 13. Collapsed and fractured shell fragments in a skeletal Grainstone (?). Well HBOR-50 @179 ft. in the Santee/Clinchfield section. Crossed Nicols. H-Area ITP Site, General SeparationsArea, Savannah River Site. Void space filled with blue epoxy glue.


G-24

Figure 14. Sandy, pelecypod, pellet packstone. Santee/Clinchfield section, Burial Grounds,General Separations Area, Savannah River Site. Crossed nicols. Void space filled with blue epoxyglue.


G-25

Figure 15. Photomicrographs of sandy biomoldic chert, Santee/Clinchfield section, Core CPT-157S3 @ 109.85 ft. F-Area APSF Site, General Separations Area, Savannah River Site. Bottompicture plane-polarized light. Amorphous silica (Opal-CT) replacement of calcite micrite (mud)matrix.


G-26

Figure 16. Photomicrograph of sandy siliceous mudstone, Santee/Clinchfield section, BurialGrounds, General Separations Area, Savannah River Site. Well BGX-2B, @ 151.2 ft. Crossednicols. Calcite micrite (mud) matrix was replaced by amorphous (Opal-CT) silica.


G-27

Figure 17. Chalcedony-cemented, quartz arenite, Terrigenous sand and sandstone microfacies.Santee/Clinchfield section, H Area, General Separations Area, Savannah river Site. Well YSC02A@ 109 ft. Crossed nicols.


G-28

Figure 18a. Model for development of facies and soft zones in the General Separation Area,Savannah River Site.

Figure 18b. Model for development of facies and soft zones in the General Separation Area,Savannah River Site.


G-29

Figure 19. Quartz-rich, skeletal grainstone, with carbonate/quartz detritus present in vuggy pores.Skeletal Grainstone Microfacies, Santee/Clinchfield section, H Area ITP Site, General SeparationsArea, Savannah river Site. Well HBOR-50 @ 166 ft. Crossed nicols. Void space filled with blueepoxy glue.

Carolina G


ield Trip G

uidebookW

SRC

-MS-2000-00606

G-3

0

Figure 20. This figure is a 3D profile illustrating the relative thickness and percentage of carbonate from core and SPTborings throughout the GSA. The upper bounding surface is the top Santee Unconformity surface and the lowerbounding surface is the top Warley Hill surface. The tan colored “sticks” indicate carbonate “hits” occurring in the SPTborings without regard to carbonate content other than its presence at that particular depth in the boring. Thepercentage carbonate in the cored wells is color coded as follows: Purple >65%; Blue 45-65%; Yellow 25-45%; Green 5-25%.


H-1

The Savannah River Site E-Area Vadose Zone Monitoring System

Heather Holmes Burns, Bechtel Savannah River Company, Aiken, SC 29808Doug Wyatt, Tom Butcher, Jim Cook, Brian Looney, Joe Rossabi, WSRC, Aiken, SC 29808Michael Young, Georgia Institute of Technology, Atlanta, GA 30332

ABSTRACT

Shallow disposal trenches located in theSavannah River Site’s E-Area contain low-level radioactive waste. To help protect thegroundwater below these trenches, acomprehensive vadose zone monitoringsystem was developed and deployed. Avadose zone-based monitoring approach wasutilized since groundwater contamination fromnearby facilities would limit the sensitivityand increase the cost of a traditionalgroundwater-based monitoring system. The E-Area Vadose Zone Monitoring System(VZMS) monitors the three soil parameters(i.e., water content, water potential or tension,and contaminant concentration) that determinehow fast and in what concentrationcontaminants are traveling to the groundwater.Forty instruments were installed in threevertical boreholes and four angled boreholesdrilled underneath the centerline of the trench.The various monitoring instruments weremounted on assemblies that were installed ineach borehole. The equipment at each depthtypically included: 1) a porous-cup vacuumlysimeter (i.e., solution sampler) for collectingwater from the vadose zone for analysis, 2) anadvanced tensiometer (AT) for measuring soil-water tension, and 3) a water contentreflectometer (WCR) for measuring soil-watercontent. The electronic instruments wereconnected to data loggers with cellulartelephone links to allow direct downloading ofdata and calibration information to both onsiteand offsite computers. Thin-walled stainlesssteel tubes were installed near the instrumentholes to allow more detailed water contentlogging using neutron probes. The initial dataindicate that many of the instruments haveequilibrated with the moisture in thesurrounding backfill and environment. Thedata are being closely monitored to track the

wetting fronts and analyze the potentialmovement of contaminants.

Several innovative methods forcharacterization, drilling and installation wereutilized during implementation of the EMOP-Phase IA. Extensive characterization of thegeology using a combination of conepenetrometer lithology data and conventionaldrilled soil core data from around the trenchesenabled development of 3-D models of thelithology allowing for precise instrumentplacement. An innovative approach forinstalling the WCRs was developed to ensureinstrument contact with “undisturbed” soils.The team also developed a new method forinstallation of the angled lysimeters that wouldallow removal and replacement, whenrequired.

The number of vadose zone wells installed thisfiscal year was deliberately limited to allowsystem evaluation. By evaluating thestrengths and weaknesses of the system, futuredeployment of vadose zone monitoringsystems could be improved. The strengthsand weaknesses of the system are described inthis document with respect to installation,maintenance, and operation. The results arealso discussed. Preliminary results indicatethat the VZMS will provide cost effective,high quality performance monitoring and asensitive early warning of any releases fromthe shallow disposal trench operation.

INTRODUCTION

E-Area, located in the central portion of SRS,is the principal disposal facility for low-levelradioactive wastes. Some of these wastes aredisposed of in unlined trenches. Vaults arealso used for disposal if additional isolation isrequired due to higher radionuclide content.


H-2

The trenches vary in length and width but aretypically constructed within the upper eightmeters of the local sediments. The vadosezone underlying the E-Area study extends to adepth of approximately 21 meters. Waterfrom the surface must pass through the vadosezone to reach the saturated water table, andtherefore, potential contaminants from thetrenches must also pass through the vadosezone.

The approach prior to VZMS installationprimarily relied on groundwater monitoring.This monitoring program could not meet DOErequirements because of the several tritiumplumes that emanate from the Burial GroundComplex (BGC) disposal facilities locatedupgradient of the shallow trenches and vaults.Because of the proximity of these facilities tothe E-Area Disposal Facility (EADF) units,the previous tritium contamination masks anypotential contributions from the EADF.Therefore, monitoring the vadose zone was thebest alternative to demonstrate compliancewith DOE Order 435.1 which requires both 1)a comprehensive monitoring program thatvalidates the performance assessment (PA)and 2) an assessment of impacts togroundwater. At SRS, this has beeninterpreted to mean that concentrations ofradionuclides in groundwater not exceeddrinking water standards (DWS).

The significant accomplishments of thisprogram are summarized in Table 1.0. Thedata supplied by the monitoring programprovides information about the integrity of thedisposal unit, and whether or not elements ofthe disposal site are functioning as intendedand predicted. Initial data obtained from bothvadose zone characterization and monitoringdemonstrates that the E-Area disposal trenchesare meeting the requirements established inDOE Order 435.1. The E-Area VZMS isdesigned to detect and quantify the rate andvolume of both water and contaminantmovement released from the trenches to theundisturbed environment. The VZMS will beused to show that facilities are operating safelyduring waste emplacement, and also, serve asan early-warning indication of the magnitude

of any releases from the shallow disposaltrench operation. The early-warning measurewill allow corrective actions to be conductedto prevent further degradation of thegroundwater resources.

Brief Description of Monitoring Strategy

Instruments were selected that had beendemonstrated in industry or by research anddevelopment efforts ready for deployment.An array of VZMS instruments installedaround the perimeter of the E-Area disposaltrenches was selected to provide data fordetermining the quantity, rate, and direction ofsoil-water and contaminants in the vadosezone. Three types of vadose zone monitoringdevices were selected to comprise the coretechnology for this system:

• Advanced Tensiometers measure soil-water tension or potential. This parameterrelates to how well the soil retainsmoisture.

• Water Content Reflectometers sense thedielectric constant of the soil to determinemoisture content.

• Porous-cup lysimeters or solutionsamplers collect soil-water samples thatwill be analyzed for tritium and othercontaminants.

Tritium was the contaminant of interest sincethe PA predicts tritium to be the contaminantfirst reaching the groundwater. The VZMSwas to define tritium transport in two ways tomeet the two primary objectives stated in thissection. The first was to provide soil-watersamples from the vadose zone using porous-cup lysimeters for determining theconcentration of tritium. The second was toprovide the flux velocity of tritium carried tothe groundwater using data from the AT andWCR.

The instruments are installed in boreholestraditionally used for monitoring contaminantsat depth. Data collection using tensiometersand WCRs is automated, so that site personnel


H-3

are not needed for data collection. Othermonitoring points will be taken with a neutronprobe which is designed to provide acontinuous water content profile for the entirethickness of the vadose zone. Probes arelowered into the borehole, readings are taken,and the probe is removed. The neutron probeaccess tubes are located adjacent to eachborehole containing the WCRs to allowredundant measurements of water content.

Brief Description of Installation Strategy

The installation strategy was dictated by themonitoring strategy as described below:

• Lysimeters were placed under the center-line of the trench using angled boreholesto monitor vertical contaminant transport.

• Instruments were placed in vertical wellsat the north end of the trench to monitorthe greatest tritium activity and to monitorlateral contaminant transport.

• Vadose zone monitoring wells wereinstalled in a triangular pattern to providedata that could be used to better determinethe flow direction or gradient.

The E-Area VZMS was installed on twoadjacent sides of the outermost completeddisposal trench. Seven boreholes wereinstalled, including three vertical boreholesand four angled boreholes. The angledboreholes were cased with PVC pipe. Verticalboreholes were not cased but backfilled withalternating layers of silica flour and nativebackfill. The vertical boreholes containedeach of the three instruments (i.e., ATs,WCRs, and solution samplers) emplaced atfour depths, whereas, the angled boreholescontain one solution sampler that is retrievableif failure occurs.

CHARACTERIZATION OF THE SRSVADOSE ZONE

Field characterization to support the VZMSdeployment was completed in January 1999.The vadose zone in E-Area was characterized

by three methods: 1) split spoon sampling, 2)Shelby Tube samples and subsequentlaboratory analysis on sediment samples, and3) PiezoCone Penetrometer Testing (CPTU) todetermine soil characteristics or thesand/silt/clay horizons and dielectric constantand pore pressure to determine moisturecontent. Geological models were developedfrom the characterization data. The 3-Dmodels were utilized in determining theoptimum depths for locating the instruments inthe subsurface strata with the highest moisturecontent. Laboratory analysis of field samplescollected during characterization allowedcomparison of actual field data to thehydraulic parameters and the determination ofcoefficients required as input to the modelsused in the E-Area RPA. Soil hydraulicproperty data was also used to calculate thesoil-water flux through the vadose zone thatcould be compared to PA mathematical modeloutput.

The 3-D models of the vadose zone werecorrelated with data obtained from the ShelbyTube soil samples. The models depicted thesand/silt/clay horizons (See Figure 1.0) andsoil moisture. The 3-D model indicates thereare three major horizons underneath the E-Area disposal trenches that differ in ways thatare important for placement of vadose zoneinstruments. The A-horizon dominates theupper 7 meters of the vadose zone and is apredominantly clay layer. The B-horizonbegins at about the 7 meter depth and extendsto 18 meters and is characterized by highersand content than the A or C horizons. The C-horizon is at the 20 meters depth and includesthe water table overlaying a clay-confiningzone. Within the area of the localized model,there are apparent channels interspersed inthese horizons that cause variations in the clayand sand content underlying the trenches.Because of this, there are lateral variations inthe overall soil moisture throughout thevadose zone.

The approximate depths selected forplacement of instruments is presented inFigure 2.0. The uppermost tensiometer wasplaced just below the clay surface. The


H-4

second tensiometer was placed above the clayinterbedded in the sandy zone at about 14meters depth. The next two tensiometers wereinstalled slightly above and in the capillaryfringe zone.

DEPLOYMENT OF THE SRS VADOSEZONE MONITORING SYSTEM

The goal of deploying a vadose zonemonitoring system in E-Area was toinstrument the disposal trenches with a systemthat would meet the two primary objectives: 1)validation/calibration of the performanceassessment mathematical model and 2)protection of groundwater resources. Thestrategy used to meet the stated objectives wasto monitor the soil-water parameters thatdirectly relate to water movement andcontaminant migration from the trenchdisposal area. Therefore, the VZMS wasdesigned to monitor soil-water content,tension, and contaminant concentration whichare the soil-water parameters that directlyrelate to water movement and contaminantmigration from the trench disposal area.These three primary soil parameters could beutilized to determine the contaminant flux, orhow fast a given contaminant is traveling tothe groundwater through the vadose zone.

SRS is the first site to utilize the AT incombination with other instruments to monitorsoil parameters at up to 18 meters depth in thevadose zone at a humid site. The AT haspaved the path for a permanently installedtensiometer allowing it to be utilized in long-term monitoring programs. The AT has theseadvantages over other tensiometers: (a) it canoperate at depths exceeding 30 meters, (b) itcan be checked for proper operation in thefield, (c) the pressure sensor can be calibratedin the field, and (d) the pressure sensor can bereplaced, if required. All of these advantagesresult in reduced maintenance and servicingneeds for the AT.

The AT and the WCR instruments haveelectrical leads that are connected to dataloggers. Three data loggers that are suppliedwith power utilizing a solar panel have been

installed adjacent to each of the three verticalwells. The data loggers are equipped withcellular telephone links to allow directdownloading of data and calibrationinformation to both offsite and onsite personalcomputers. The soil-water samples collectedby the vacuum lysimeters are manuallycollected.

Methodology

Seven vadose zone wells were installedaround two sides of the outermost completeddisposal trench (i.e., slit trench #1). Thelocation of each of the vadose zone wells wasselected based on the followingconsiderations:

1) to allow monitoring of the entirelength of the outermost trench,2) to determine the flow of contaminantsin vertical and lateral flow patterns,3) to facilitate the earliest detection oftritium by being placed next to the highesttritium containing waste and underneath thecenterline of the trench, and4) to monitor the water flow through thetrench.

Four of the wells were angled and placedunder the centerline of the trench to enablemonitoring of vertical contaminant transport.Each of the angled wells were equipped withone lysimeter and were spaced approximately30 meters apart to enable monitoring the entirelength of the trench. Three vertical wells wereplaced at the NE and NW corners of the trenchand on the eastern side of the trench to form atriangular pattern. This arrangement wouldallow the soil-water flux to be measured basedon the gradient between the instruments andwould allow monitoring of lateral contaminanttransport. Instrument clusters containing oneAT, one porous-cup lysimeter, and one WCRwere placed at four vertical depths inside eachvertical well. Neutron probe access ports wereplaced adjacent to each of the vertical wells.The E-Area VZMS vertical wells contain ATsat four different depths so that the rate atwhich a wetting-front is moving through thevadose zone can be used to determine the soil-


H-5

water velocity, the direction of watermovement, and the soil-water flux. Thecontaminant flux that is directly comparable tothat estimated in a performance assessment isobtained from both the soil-water flux and thecontaminant concentration.

Innovative Installation Methods

Two instruments required innovativeinstallation methodologies to ensure properinstrument performance and sensitivity: 1) thewater content reflectometer and 2) the neutronprobe access port.

1) Water Content Reflectometer

The WCR requires monitoring of undisturbedsoil and can be installed via two methods: 1)insertion into the side-wall or 2) placementagainst the side-wall of the borehole. TheWCRs were installed in boreholes by forcingthe assembly against the borehole wall using ashort lever-arm attached to the WCR as shownin Figure 3.0. To provide redundantmeasurement of water content, a neutronprobe access port was installed adjacent toeach vertical well to provide continuousmeasurement of soil-water content.

2) Neutron Probes

The neutron probe enables a continuous soil-water content measurement along the length ofthe vadose zone. Periodic monitoring withthese probes will allow the tracking ofmoisture fronts traveling through the E-Areavadose zone. It has long been recognized thatthe nature of the access borehole cansignificantly affect neutron probe sensitivity(i.e., performance). Stainless steel wasselected as the material of construction for theneutron access pipe. Steel tubing waspreferred over PVC pipe since it does notcontain chlorine to effect neutron attenuationand is commonly available. The steel tubingused in this application consisted of flushthreaded well casing. The tubing wasschedule 5 with a maximum thickness ofschedule 40 near each threaded coupling. Thestainless steel material of the access tube was

selected not only because of its minimal effecton the neutron flux, but also because of itsdurability and its ability to be pushed in thesoil by a hydraulic cone penetrometer test(CPT) rig. The access tube was installed by aCPT rig to eliminate the generation of anannular space created by conventional drilling,and therefore, to eliminate the need forbackfill. The most commonly used backfillmaterials have different moisture contents thanthe actual formation. Therefore, the moisturecontent of the backfill, being the closestmaterial to the neutron source, will cause thethermalization of a large fraction of the high-energy neutron flux resulting in significantattenuation of the neutron flux before it entersthe formation.

PERFORMANCE OF INSTRUMENTS

In this section, the performance of all threevadose zone instruments/samplers (i.e., theAT, WCR, and solution sampler) and theanalyzed results are discussed. Two soilparameters are required to calculate thecontaminant flux to the groundwater: 1) thesoil-water flux determined from the AT andWCR data, and 2) the contaminantconcentration obtained via the solutionsampler. Tritium was selected as the targetconstituent because: 1) it is present in thewaste disposed of in the trenches, and 2) theSRS PA model indicates that it would be thefastest traveling radionuclide to thegroundwater.

Water Content Reflectometer SensorPerformance

WCR data was collected and analyzed severalmonths after installation, allowing forequilibrium of the sensor with the host soil.The data are very consistent with timeshowing little change in water content in thevadose zone at the depths installed. Averagewater contents after equilibrium with the hostsoil are values of 0.284, 0.181 and 0.266m3/m3 for the three vertical wells. Thesevalues are consistent in time and reasonablegiven the soil texture.


H-6

Advanced Tensiometer Performance

The water potentials as measured by the ATsreflected expected conditions in thesubsurface. The lowest water potentialsreflecting dryer conditions were located in theshallower zones while the deeper zones nearthe wetter capillary fringe had higherpotential. Figure 4.0 shows the soil-watertension data obtained from the AT sensors invertical Boring #5 at all four depths in thevadose zone (from 7 to 17-meters depth).Data from the AT sensors in boring #5 wereselected for illustration since this boreholecontains the most operative solution samplers.The solution samplers, when placed undervacuum, collect soil moisture until the vacuumequilibrates with the surrounding soil tension.At that time, water from the soil is no longerdrawn into the lysimeter collection chamber.In Figure 4.0, the sharp decreases in tension(days #134, #141, #182, #225, and #230)occurred on the days in which the solutionsamplers were placed under vacuum. TheATs performed as expected since the soil-water tension decreased as the surroundingsoils become dryer when soil-water was beingcollected into the solution sampler. The stepchange seen in the instrument tension at thedeepest depth (17-meters) was at the time of aheavy rainfall resulting in the water tablerising. The instrument AT-5-55 indicates thatsaturation occurred at that time and was to beexpected since boring #5 soils were wetter at ashallower depth than that of either boring #6and boring #7).

All of the data from the ATs in Boring #5(AT-5), Boring #6 (AT-6), and Boring #7(AT-7) indicate that the soil tension isrelatively constant ranging between -100 cm(wetter) to -200 cm (dryer) and is consistentwith expected tensions for SRS soils. Waterpotential appeared to be unaffected by daily oryearly infiltration events at these depths. Thetotal variation in water potential was less than100 cm in all tensiometers over the 3.5-monthperiod.

Solution Samplers

Sixteen vacuum solution samplers were placedin the seven boreholes around the sides of theE-Area disposal trenches. Table 2.0 lists thesolution samplers and the sample volumescollected during FY99. Thirteen out ofsixteen solution samplers have collectedsamples. Each of the solution samplers in thefour angled boreholes placed under thecenterline of the trench have been successfullycollecting samples. Nine out of the twelvesolution samplers placed in the three verticalboreholes have been able to collect samples.A sample as small as 5-10 milliliters issufficient for obtaining the concentration oftritium, which is the primary contaminant ofinterest. The three solution samplers that havenot collected samples on either of the twocollection dates, have not yet been declaredinoperative. Other methods to improve thecollection performance are planned whichinclude placing solution samplers under aconstant vacuum. The first set of wateranalyses, as expected, shows that tritiumabove background levels is moving throughthe vadose zone beneath the trench area wheredebris with the highest tritium content hasbeen buried.

DATA ANALYSIS

The SRS PA model estimates the peakgroundwater concentrations based on 1) anestimated groundwater flux and 2) a maximumcurie concentration in the disposal area thatwill not exceed the DWS. Comparison of themeasured tritium concentrations with PAmodel results indicates that groundwaterconcentrations will meet the DWS asdiscussed in this section.

The first round of sampling from the vacuumlysimeters produced tritium concentrationsranging from about 1.0 pCi/ml to about 84pCi/ml as shown in Table 2.0. A simplecalculation was made to compare the highestconcentration of 84 pCi/ml with the results ofthe vadose zone model used in theperformance assessment. The comparison


H-7

indicated that a tritium concentration of about400 pCi/ml migrating through the vadose zonewas required to produce a tritiumconcentration equal to the DWS of 20 pCi/mlat a hypothetical 100-meter well. Themeasured vadose zone concentration of 84pCi/ml is about 20% of the 400 pCi/ml.Therefore, this maximum tritiumconcentration measured by the SRS VadoseZone Monitoring Program should result in agroundwater concentration of about 4.0pCi/ml, or 20% of the DWS at the 100-meterwell.

OVERALL ASSESSMENT: LESSONSLEARNED

This section describes the lessons learnedfrom implementation of the first phase of thevadose zone monitoring system deploymentprogram at SRS. The number of vadose zonewells and instruments installed during the firstphase was deliberately limited to allow systemevaluation. The next phase of monitoring willbe designed to incorporate the “lessonslearned” from the installation and operation ofthe first VZMS at SRS. The three largestimprovement areas in subsequent monitoringdesigns are (1) redundancy in monitoringcritical soil parameters, (2) maintenance andlong-term monitoring, and (3) statisticalvalidation of data.

Redundancy in Monitoring Critical SoilParametersRedundant monitoring of critical soilparameters using different instruments shouldimprove the confidence that changes in thesoil-water conditions are real and not affectedby the monitoring systems themselves. Thefirst phase of the SRS VZMS deploymentincluded redundant monitoring of soil watercontent by neutron logging and by WCRs.

Maintenance and Long-Term Monitoring

The SRS VZMS deployment program utilizedthe borehole monitoring strategy forinstallation of instruments because of its use intraditional well installations and because of itsprimary advantage that allows placement of

instruments at lower depths. However, theborehole monitoring strategy suffers from asignificant disadvantage for long-termmonitoring: replacement of instruments. Lossof fixed instruments in the borehole results ina loss of monitoring points. The loss of a fewinstruments can leave potentially large gaps ofthe subsurface unmonitored. Boreholeinstrument replacement requires re-drillingwhich is not feasible.

Subsequent monitoring programs shouldmaximize access ports for multiple instrumentuse and easy replacement upon failure. Theexistence of access tubes and ports willfacilitate the use of new technologies. Long-term monitoring programs (e.g., greater than10 years) can benefit from the installation ofaccess boreholes. Future developments indownhole instrumentation could improve theaccuracy in monitoring subsurface conditions,and the presence of access tubes couldfacilitate the testing and use of theseinstruments.

Statistical Validation of Data

Data collection from the SRS VZMS has beenon-going since installation in May 1999. Datacollection intervals are important aspects ofstatistical analyses. In addition, up to threemonths is required for the data to equilibratewith the surrounding environment. Samplingover a longer period of time permits thequantification of seasonal variability and othertime-dependent behavior. Insufficient dataand inadequate methodology for representingsignificance and uncertainty in data can lead tofaulty conclusions. Without sufficientamounts of data it is not statistically possibleto conclude whether baseline conditions havechanged.


H-8

Table 1. EAVZ accomplishments and lessons learned.

EVALUATIONOF:

SUMMARY OF ACCOMPLISHMENTS

Vadose ZoneCharacterization

- Developed 3-D Models of lithology allowing for precise instrument placement.

BoreholeDrilling

- Successfully utilized a mud rotary drilling rig to drill angled holes.- Vertical boreholes remained open allowing proper instrument placement.- Obtained split-spoon samples during installation to verify placement of

instruments in optimum moisture zones.

InstrumentInstallation

- Developed innovative method for placing WCRs against the side-wall of theborehole.

- Developed innovative approach to eliminate annular space around the neutronprobe access port. A cone penetrometer rig pushed the stainless steel accessport directly into the ground.

InstrumentPerformance

- Deep lysimeters (i.e., > 12-meters) are able to collect large samples for labanalyses.

- ATs and WCRs have equilibrated with surrounding environment.- ATs and WCRs indicate consistent soil-water tension and content that is

independent of influx events.

PA Validation - Characterization data obtained from analyses of soil samples are comparable toperformance assessment input data.

DOE OrderCompliance /PA Validation

- Lysimeter data indicates tritium concentration underneath trenches will meet theDrinking Water Standard at the compliance point (i.e., 100-meter well).

Summary ofLessonsLearned

- Increase redundant monitoring of critical soil parameters using differentinstruments to improve confidence in data.

- Increase usage of access ports to facilitate the use of multiple technologies.- Increase usage of cased boreholes to allow instrument replacement in case of

failure.- Design system to include all phases of monitoring: pre-operational, operational,

and post-operational.- Establish baseline to definitively distinguish between background radiation and

that contributed by the disposal area.- Quantify uncertainty in measurements and observations. Utilize statistical

analyses to validate data.


H-9

Figure 1. 3-D model depicting sand/clay horizons using CPT Friction Ratio. Brighter colors (dark reds)indicate more clay-rich sediments, and lighter colors (grays) indicate more sand-rich sediments.


H-10

Figure 2. Instrument depths as determined from the 3-D models.

0

1020

3040

5060

70

10’

20’

30’

40’

50’

60’

70’

0’ 5’ 10’ 15’ 20’

Base Upland Uni t /Top Tobacc o Road Fm.

Surface GL

Trench 1 Trench 2 Trench 3(Active)

Contact “Tan C layConfining Zone”

Aproximate ContactDry Branch Fm.

View from northern end of trenches f acing southw ard.(relative scale is correct)

~20’

~65o

EAVZCPT2-FR

~10’

~50’

Porous Cup Lysimeter

Advanced Tensiometer

Advanced Tensiometer

Neutron ProbeAcces s Tube


H-11

Figure 3. Conceptual diagram of the installation method for the Advanced Tensiometer and WaterContent Relectometer instrument cluster.

Installation steps for Advanced Tensiometer (AT) Assembly: 1) lower to depth whileTDR probe is retracted, 2) release TDR and pull back to engage, 3) test TDR proberesponse, 4) while holding AT system, emplace silica flour

TDR MoistureSensor Deployment

System

PVC Lever Arm

Rigid PVC Backing

polyethylene foam

Sensor (Campbell505 or similar)

retraction string*

* for use during deployment

standard porous cup lysimeter with dual tubeconnection for vacuum collection andpressure lift delivery to surface (SoilMoisture Equipment 1920 or similar)

Silica flour (<200 mesh) backfill aroundlysimeter section

Bentonite seal

Bentonite seal


H-12

Figure 4. Advanced Tensiometer Data for Boring #5 (AT-5).

AT-5 Water Potentials

-1000

-800

-600

-400

-200

0

100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250

Time into Year (d)

ial

PT_1_22

PT_2_32

PT_3_41PT_4_55


H-13

Table 2. Volumes of soil-water samples collected by the lysimeters.

Solution Sampler # Sample Volume (mls) AverageTritiumConc.

(pCi/ml)July 1 Sept 13

BORING #1 (AL-1) 800 600 4.3BORING #2 (AL-2) 800 600 1.0BORING #3 (AL-3) 800 700 2.5BORING #4 (AL-4) 800 600 3.3

BORING #5AT-5-23 900 0* 84.0AT-5-33 825 0* 29.7AT-5-42 0 10 14.1AT-5-57 775 700 21.7

BORING #6AT-6-24 0 0 --AT-6-34 0 0 --AT-6-43 400 350 13.3AT-6-55 150 0 12.4

BORING #7AT-7-12 0 0 --AT-7-23 0 5 2.2AT-7-42 0 200 7.4AT-7-54 5 – 10 0 --

* Tubing had leaks and could not sustain a vacuum. Tubing has been temporarily repaired as parts arebeing ordered for permanent replacement.


H-14

Notes


I-1

Modeling Aquifer Heterogeneity at the Savannah River Site Using ConePenetrometer Data (CPT) and Stochastic Upscaling Methods

M. K. Harris, G.P. Flach, WSRC, Savannah River Technology Center (SRTC), Aiken, SC 29808A.D. Smits, Science Applications International Corporation (SAIC), Augusta, GA 30901F.H. Syms, BSRI, Site Geotechnical Services, Aiken, SC 29808

ABSTRACT

Cone penetrometer tests (CPT) have becomean increasingly popular characterizationmethod for Tertiary shallow subsurfaceinvestigations (less than 200 feet) in theAtlantic Coastal Plain of South Carolina.These sediments consist primarily ofinterbedded and interfingering fluvial, deltaic,and shallow marine sediments. CPT providesrelatively inexpensive subsurface data withoutdrilling fluid or cuttings. For environmentalrestoration applications at the Savannah RiverSite (SRS), CPT is typically used to obtaindepth-discrete groundwater samples, small-diameter permeability samples, and to definehydrostratigraphic horizons. A recentexample is the C-reactor area at the SRS, thefocus of this study. At this site, 140 CPTlocations were used to define contaminantplumes and hydrostratigraphy over a 3 mi2

area, rather than conventional boreholetechniques (e.g. monitoring wells, cores,electric well logs, slug and pumping tests). Inthis investigation, the CPT lithologic data arefurther used to predict hydraulic conductivityvariations within hydrostratigraphic zones ofthe uppermost aquifer unit. The methodinvolves correlating tip resistance, sleeveresistance and pore pressure measurements topercent fines (mud, silt, and clay) content andhydraulic conductivity. Predicted percent finescontent at the scale of the CPT measurements(0.1 ft) were categorized into high, mediumand low conductivity and upscaled to the flowmodel resolution using geostatistical methods.The resulting model conductivity fieldprovided a realistic representation of aquiferheterogeneity in coastal plain sediments. Themethod significantly improved groundwatermodeling flow path predictions. The resulting

tritium transport simulations compared wellwith field data.

PURPOSE OF STUDY

A geospatial mapping and groundwatermodeling case study was completed at the C-Reactor Area at the Savannah River Site(SRS) based primarily on cone penetrationtesting (CPT) (Figures 1 and 2). For adetailed discussion on CPT methods see Symsand others (2000). CPT was utilized for thedefinition of hydrostratigraphic unitboundaries and hydraulic conductivitiesinstead of conventional hydrogeologicalmethods (e.g. monitoring wells, cores, electricwell logs, slug and pumping tests).Approximately 140 CPT tests were used tomap the hydrostratigraphic unit boundariesand define hydraulic conductivities of thehydrostratigraphic units. In conjunction withthe CPT tests approximately 1000 depth-discrete water samples were taken tocharacterize the tritium and volatile organiccompound (VOC) contaminant plumes in thearea. A three-dimensional geospatial model ofthe hydrostratigraphy was constructed basedon correlation of the CPT data, while three-dimensional groundwater plume maps wereconstructed based on the water samples. Theresulting hydrogeological framework wascoupled with a groundwater flow model forparticle track analysis and subsequenttransport modeling. This paper concentrateson two main areas of the case study. The firstillustrates CPT data for detailedhydrogeological mapping and integration ofall characterization data (geological,hydrological, and contaminant) to understandthe relations of groundwater and plume


I-2

movement. Secondly this study presents amethod for transforming CPT data intohydraulic conductivity data for definition ofhydrogeologic properties in groundwatermodeling.

SITE DESCRIPTION AND HISTORY

The study area is located in the Upper AtlanticCoastal Plain physiographic province insouthwestern South Carolina at the U.S.Department of Energy’s Savannah River Site(SRS). The SRS occupies about 300 mi2 andwas set aside in 1950 as a controlled area forthe production of nuclear material for nationaldefense and specialized nuclear materials.This study focuses on a 3 mi2 area in thecentral portion of the SRS near the C ReactorArea which is undergoing environmentalcharacterization and remediation for RCRAand CERCLA regulatory investigations. TheC-Reactor Seepage Basins (CRSBs) and theC-Area Burning/Rubble Pit (CBRP) (Figures1 and 2) are the two principal reactor wasteunits associated with C-Area.

The CRSBs consist of three unlined (earthen)basins constructed in 1957 to containradioactive water discharged from the reactorprocess buildings. The seepage basins wereused from 1959 to 1970 to dispose of low-level radioactive process purge water from thereactor disassembly basin. In 1963, the purgewater was deionized and filtered and toremove solids and sludge prior to discharge tothe basins. The seepage basins were not usedfrom 1970 to 1978. During this period, purgewater was mixed with large volumes ofcooling water from the heat exchanger anddischarged to area streams. Discharge to theCRSBs resumed in 1978 after improvementswere made to the processing of purge waterfrom the disassembly basin. The C-Reactorwas shutdown for repairs in 1985, and theCRSBs have not received wastewater since1987. The CRSBs are currently undergoingfeasibility studies to determine the regulatorypath forward. Waste disposal records indicatethat the main basin received aqueousradioactive waste. Radionuclides in the

wastewater from the disassembly basin,sumps, tanks and drums included tritium,chromium-51, cobalt-60, cesium-134, cesium-137, and other beta-gamma fission products(WSRC, 1993). Tritium releases to CRSBs aredocumented in Environmental ProtectionDepartment (EPD) annual monitoring reports.

The CBRP was constructed in 1951 and usedto dispose of and burn organic solvents, wasteoils, paper, plastics, and rubber through 1973.After 1973, the pit was filled with inert rubbleand soil (WSRC, 1994). The CBRP was ashallow, unlined excavation (approximately25 feet wide and 350 feet long) with depths ofapproximately 8 to 12 feet. The history ofsolvent disposal at the CBRP and thepersistence of high VOC concentrations 25years after operations ceased suggests thepresence of VOC in non-aqueous phase liquid(NAPL) within the vadose zone beneath theCBRP. A Soil Vapor Extraction and Air-Sparging (SVE/AS) remediation system arecurrently in operation for cleaning up thesolvents in the soil and groundwater at thissite.

Primary groundwater contaminants in C-areainclude tritium for the CRSBs and VOC forthe CBRP. This study will only discuss theCBRP VOC groundwater contaminationdistribution in detail.

CHARACTERIZATION DATA

Geologic descriptions from three cores andCPT lithologic data (tip stress, sleeve stress,and pore pressure) from 139 locations wereused to identify and delineatehydrostratigraphic boundaries beneath thestudy area (Figure 2). The CPT data werealso used to infer conductivity distributionusing stochastic methods. Thehydrostratigraphic boundaries were compiledinto a hydrogeologic model usingEarthVision geospatial analysis (DynamicGraphics Inc., Alameda CA, www.dgi.com).Contaminant plume mapping was based onover 1000 depth-discrete water samples takenwith the cone-penetrometer to characterize


I-3

tritium and VOC plumes in the area.Analytical data was compiled into three-dimensional maps utilizing EarthVision™software.

HYDROGEOLOGY

The SRS lies within the Upper AtlanticCoastal Plain, a southeast-dipping wedge ofunconsolidated and semi-consolidatedsediment that extends from its contact with thePiedmont Province at the Fall Line to the edgeof the continental shelf. The sediment rangesfrom Late Cretaceous to Miocene in age andcomprises layers of sand, muddy sand, andmud with minor amounts of calcareoussediment (Fallaw and Price, 1995). TheCoastal Plain sediment rests unconformablyon Triassic-aged sedimentary rock of theDunbarton Basin and Paleozoic-agedcrystalline rock.

This study involves the Tertiary-age sediment,principally the Eocene to Miocene sequence.The up-dip portion of the Floridan aquifersystem includes the Gordon aquifer, theGordon confining unit and the Upper ThreeRuns aquifer (Aadland and others, 1995), andis considered in geospatial and groundwateranalysis. The Upper Three Runs aquifer wasanalyzed in more detail in relation to the CPTdata and will be discussed in more detail thanthe underlying hydrostratigraphic units.Figure 3 correlates the hydrostratigraphicnomenclature with the local lithostratigraphyas defined by Fallaw and Price (1995).

Gordon Aquifer Unit (GAU)

The Gordon aquifer constitutes the basal unitof the Floridan aquifer system at the SRS andis the lowermost unit characterized in thisstudy. The aquifer includes loose sand andclayey sand of the Congaree Formation and,where present, the sandy parts of theunderlying Fourmile Branch and SnappFormations (Figure 3), (Harris and others,1990; Aadland and others, 1991 and 1995).The sand within the Gordon aquifer isyellowish to grayish orange and is sub- to well

rounded, moderately to poorly sorted, andmedium- to coarse-grained. Pebbly layers andzones of sand cemented with iron and silicaare common. The Gordon aquifer includesrare interbeds of light tan to gray clay thatrange up to three feet in thickness. Lenses ofclay less than 6 inches in thickness arecommon near the base of this unit. TheGordon aquifer contains a small amount ofsporadically distributed calcareous sediment.The thickness of this unit is variable, rangingfrom approximately 60 feet to 160 feet acrossthe SRS (Flach and others, 1999). Multi-wellpumping tests performed on Gordon aquiferwells at the SRS and vicinity produceconductivity values that average 35 ft/day(Aadland and others, 1995). Ten single-wellpumping tests within the Gordon aquifer givepermeability values that range from 0.82 to143 ft/day (Flach and others, 1999).

Gordon Confining Unit (GCU)

The Gordon confining unit separates theGordon aquifer from the Upper Three Runsaquifer. This unit is commonly referred to asthe “green clay” in previous SRS literatureand includes sediment of the Warley HillFormation (Figure 3). The unit comprisesinterbedded silty and clayey sand, sandy clayand clay. The clay is stiff to hard and iscommonly fissile. Glauconite is a commonconstituent and imparts a distinctive greenishcast to the sediment, hence the informal nameof “green clay” given to this unit. Within thestudy area the GCU includes some calcareoussediment and limestone, primarilycalcarenaceous sand and clayey sand withsubordinate calcarenaceous clay, micritic clay,and sandy micrite and limestone.

The GCU dips toward the south-southeast,increasing in thickness downdip across theSRS from approximately 10 feet to 80 feet.Laboratory tests of undisturbed samples takenfrom the GCU at locations across the SRSindicate vertical permeability ranges from1.14E-06 to 4.27E-01 ft/day and horizontalpermeability ranges from 5.40E-06 to 1.22E-


I-4

01 ft/day within this unit (Flach and others,1999).

Upper Three Runs Aquifer (UTRA) Unit

The Upper Three Runs aquifer includes theinformally named “upland” unit, TobaccoRoad Sand, Dry Branch Formation,Clinchfield Formation, and Santee Limestone.At SRS, the UTRA is divided into informal“lower” and “upper” aquifer zones separatedby the “tan clay” confining zone (Figure 3).

“Lower” Aquifer Zone (LAZ). The "lower"aquifer zone of the UTRA consists ofsiliciclastic and calcareous sediments of theSantee Formation and parts of the Dry BranchFormation beneath the “tan clay” confiningzone (TCCZ) (Figure 3). The zone is highlyvariable in thickness ranging from around 40 –80 feet thick in the study area. The LAZrepresents shallow, moderate to high-energynearshore fluvial and marine environments(Harris and others, 1997). Core descriptionswithin and adjacent to the study area indicatethe LAZ represents a dual facies, one that isdominated by siliciclastic components and theother dominated by carbonate componentsresulting in highly permeable sedimentsdirectly in contact with sediments of lowerpermeability. Typically, the sands and clayeysands are unconsolidated fine- to mediumgrained, moderate to well sorted and havegood to excellent measured porosities (>30%)with permeabilities ranging from 6.8 to 40.8ft/day (Harris and others, 1997). In contrastthe carbonate sediments are unconsolidatedand consolidated with excellent measuredporosities (>35%) and moderate to lowpermeabilities ranging from .03 to 4.0 ft/day(Harris and others, 1997). The carbonatesediments are generally mud supported(wackestones and packstones). The lithifiedcarbonates are sometimes partially silicifiedwith biomoldic characteristics. Bothlithologies exhibit good porosity but lowpermeability because the pores are notinterconnected or are filled with micrite mudthat is highly porous but not very permeable.

Based on the heterogeneity within the LAZand for the purposes of this study, the LAZwas sub-divided into five informalhydrostratigraphic intervals. Characteristiclog signatures of tip, sleeve, and pore pressuredata from the CPT data were utilized todelineate permeable (dominantly siliciclastic)and non-permeable (dominantly calcareous)intervals. This sub-division was necessary tomap the calcareous sub-zones which functionas localized confining zones and stronglyinfluence groundwater flow direction. Fromthe bottom up, these intervals include thelower siliciclastic interval (lLAZ), calcareousinterval 2 (CC2), middle siliciclastic interval(mLAZ), upper calcareous interval (CC1), andupper siliciclastic interval (uLAZ) of the LAZ.

“Tan Clay” Confining Zone (TCCZ). The“tan clay” confining zone is lithologicallyequivalent to the Twiggs Clay and IrwintonSand Members of the Dry Branch Formation(Figure 3). The zone contains light-yellowishtan to orange clay and sandy clay interbeddedwith clayey sand and sand. Clay layers aredispersed vertically and horizontallythroughout the zone and are probably notlaterally continuous over distances greaterthan 100 to 200 feet (Harris and others, 1990;Aadland and others, 1991). The TCCZconsists of two sequences of interbedded mudand sand with some calcareous sediment inthe study area. The lower sequence containslight green to tan, slightly fissile clay and siltysand that is commonly interbedded with sand,silty sand, and coarse-grained, carbonategravel. The carbonate gravel consists ofoysters and other shell debris that is mixedwith mud, sand, and gravel. The lowersequence commonly includes interbeds ofrelatively dense, well-indurated layers of amatrix-supported, shelly, sandy carbonatemudstone. This material consists ofsiliciclastic sand, mud, and gravel, shelldebris, and other carbonate fragmentscontained in a micrite (mud) matrix.

The upper sequence within the TCCZ consistsof interbedded siliciclastic sand and mud andrepresents an interbedded dual facies with low


I-5

permeability sediments closely interbeddedwith high permeability sediments . The mudis waxy and commonly fissile, with a veryhigh clay content. The interbedded sand isgenerally well-sorted and medium-grained,with good interstitial porosity. Thicknesses ofthe TCCZ range from approximately 2 feet to40 feet. The TCCZ becomes very thin in thevicinity of Caster Creek and is deeply incisedat Fourmile Branch. Laboratory tests ofundisturbed samples taken from the TCCZ incores across the SRS indicate verticalpermeability ranges from 3.70E-08 to 2.39E-01 ft/day and horizontal permeability rangesfrom 1.45E-05 to 2.04E-01 ft/day within thisunit (Flach and others, 1999).

“Upper”Aquifer Zone (UAZ). The “upper”aquifer zone includes the “upland” unit,Tobacco Road Sand, and part of the DryBranch Formation (Figure 3). Massive beds ofsand and clayey sand with minor interbeds ofclay characterize the UAZ. The sedimentwithin the “upland” unit is commonly verydense and clayey and often contains gravelysand.

The top of the UAZ is defined by the present-day topographic surface. The UAZ may besubdivided into four informalhydrostratigraphic intervals that are delineatedby characteristic log signatures of tip, sleeve,and pore pressure data from the CPT tool.From the bottom up, these intervals includethe "transmissive" zone (TZ), “AA” interval,“A” interval, and an undifferentiated soilsinterval (“uu” interval) (Figure 3). All theunits vary greatly in thickness due to thefluvial nature of deposition. The UAZaverages around 75 feet in thickness in thestudy area.

Hydrostratigraphic Mapping

Geologic descriptions from three cores andCPT lithologic data from 139 locations(Figure 2) were used to identify and delineatehydrostratigraphic boundaries within the nearsurface sediments beneath the study area.Figure 4 is a composite log illustrating

correlation between CPT, core, andgeophysical data which form the basis for thecorrelation of the hydrostratigraphic unitswithin the Upper Three Runs aquifer. Thehydrostratigraphic boundaries were compiledinto a 3-dimensional hydrogeologic modelusing EarthVision geospatial analysissoftware. EarthVision processes sets ofspatial and property data by calculatingminimum-tension grids to contour a “best fit”of the data. The grids can contour data in 3dimensions (x,y,z), such as the top of ageologic unit, as two-dimensional grids, orcontour data in 4-dimensions: x,y,z, and a“property.” Figure 5 is a 3-dimensionalvisualization of the hydrostratigraphic modelillustrating the relations of the units to streamand valley incisions.

Hydrogeological Conceptual Model

Based on the hydrogeology and knowledge ofgroundwater recharge and discharge zones, ahydrogeological conceptual model for the areawas developed. Figure 6 presents aschematic depiction of the hydrogeologicconceptual model for groundwater flow withinthe study area. The figure illustrates the watertable and the hydrostratigraphy previouslydiscussed. Groundwater flow in the UpperThree Runs aquifer is driven by recharge, withFourmile Branch and tributaries interceptingflow from higher elevation. The underlyingGordon aquifer is influenced at a regionalscale by the Savannah River and Upper ThreeRuns. Within C-Area, the Gordon aquifer isrecharged by the overlying Upper Three Runsaquifer. Except for process water outfalls,surface water bodies gain from groundwaterdischarge. Groundwater flow in the Gordonaquifer appears to be influenced significantlyby recharge from the overlying UTR aquifer,and lateral flow within the study areaboundaries. Solute groundwater contaminationoriginating in the C-Area is expected to beconfined to the Upper Three Runs and Gordonaquifers. Most surface recharge groundwaterdischarges to the nearest stream (e.g. FourmileBranch), with the balance entering the Gordonaquifer. In C-Area, the flow between the


I-6

deeper coastal plain aquifers and the overlyingGordon aquifer is upward; thereforecontamination is not expected to be migratedownward.

Hydrogeologic Property Mapping UsingCone Penetration Testing

CPT data were also used to define hydraulicconductivity variation within thehydrostratigraphic zones previously discussed.The method discussed herein involvescorrelating CPT tip resistance, sleeve frictionand pore pressure measurements to finegrained (mud, silt, and clay) sedimentpercentages and hydraulic conductivity. Theresultant small-scale conductivity estimateswere upscaled and interpolated onto thegroundwater model grid using stochasticmethods.

Fines Correlation to CPT

Soil classification charts in the geotechnicalliterature typically correlate CPT tipresistance, sleeve resistance, and porepressure to soil categories (e.g. Lunne andothers, 1997), but do not predict fine-grainedsediments percentages within categories.Fine-grained sediments, or fines, are definedas particle sizes passing a #200 sieve (e.g.mud, silt, clay). Techniques for prediction offines content have been demonstrated byJefferies and Davies (1993) and Robertsonand Fear (1995). These studies were used asguidelines for the prediction of the fine-grained sediment percentages.

Data specific to the Savannah River Site(SRS) were used to develop the correlations.Sieve data from F- and H-areas of the SRSwere paired with nearby CPT pushes. A totalof 516 data pairs were assembled. The paireddata were next segregated into 4 categories offines content based on #200 sieve results: 0-15%, 15-30%, 30-50% and 50%-100%. Thedata pairs are plotted in Figures 7 and 8,color-coded by percentage fines content(%FC) category: 0-15% (group 1, blue), 15-

30% (group 2, green), 30-50% (group 3, red)and 50%-100% (group 4, yellow/black). Thedata are plotted in terms of normalized tip(Qtn), sleeve (Fsn) and pore pressure (Bq)parameters that, in part, account for deptheffects. Normalized tip resistance, tnQ , isdefined by

σ

σ−= t

tnq

Q (1)

where qt is the corrected tip resistance, and σand σ are the total and effective overburdens.Normalized tip resistance can be interpretedas the net force required to advance the conetip (deform the medium) divided by the grain-to-grain contact stress. Normalized sleevefriction (a type of friction ratio) is defined by

%100q

fF

t

ssn ×

σ−= (2)

and represents sleeve friction relative to thenet force required to advance the cone tip.Normalized pore pressure is similarly definedas

σ−

−=t

q qpu

B (3)

The numerator is net pore pressure in excessof hydrostatic conditions.

Figures 7 and 8 illustrate that although the%fines groups significantly overlap, they areseparated in an average sense. "Center ofmass" points are depicted in each figure, andrepresent the average positions of each %finesgroup. The first three groups (0-50% fines)follow a definite, nearly linear, trend in Figure7. Specifically, %fines increases with frictionratio (Fsn) and pore pressure (Bq) on average.The sieve data show little dependence onnormalized tip resistance (Qtn) in Figure 8.However, the 50-100% fines group, whichrepresents low conductivity sediments,deviates from the trend established by the 0-


I-7

50% fines data. Specifically, friction ratiodecreases on average when the fines contentexceeds 50%.

Because the objective of this study is todistinguish between high and low conductivitysediments, accurate prediction of fines >50%is not critical and these data were omitted inregression analysis in order to achieve a bettercorrelation in the 0-50% range. The resultingcorrelation is

tnq

sn

QBFFC

10

10

log*869.6*14.24

log*98.15009.5%

+++=

(4)

The correlation coefficient (R2) for theregression is 0.45, meaning that the equation(4) explains about 45% of the variation in%FC. While there is large uncertainty in thepredictions, the correlation is neverthelessuseful given the absence of other more precisecharacterization data in C-area.

Hydraulic Conductivity Correlation

Hydraulic conductivity has been directlycorrelated to CPT measurements on SRSsediments by Parsons and ARA (1997) andCeleste (1998). Also, conductivity has beencorrelated to %fines (e.g. Kegley, 1993;Parsons and ARA, 1997; Celeste, 1998; Flachand others, 1998; Flach and Harris, 1999).These correlations could be used inconjunction with equation (4) to predictconductivity from CPT data. In this study, ahybrid approach was preferred, wherebyconductivity is related to CPT measurementsindirectly through predicted %fines fromequation (4), and directly through thenormalized pore pressure parameter, Bq.

The %FC ranges of 0-15%, 15-30% and >30%are considered to correspond to "high","medium" and "low" conductivity. Based onconductivity data taken at various scales,previous flow models, and groundwater flowmodel calibration, reasonable values for theseconductivity categories appear to be 20, 2 and0.01 ft/d. Since Bq is good predictor of low

conductivity intervals, we also use Bq > 0.1 todefine an "extra low" conductivity category.The "extra low" conductivity zone is assigneda value of 0.0001 ft/d, irrespective ofpredicted %fines. The chosen hydraulicconductivity relationship is summarized inTable 1.

The conductivity settings listed in Table 1 arebased partially on qualitative knowledge ofhow conductivity varies with lithology andsupport scale. As with all groundwater flowmodels, the specific values of conductivityassigned to a grid are ultimately based onmodel calibration. The settings in Table 1 aretypical values for small-scale conductivitymeasurements on sediments ranging fromclean sand (high K) to clay (extra low K) (e.g.Kegley, 1993; Parsons and ARA, 1997;Celeste, 1998; Flach and others, 1998; Flachand Harris, 1999). Example small-scale testsinclude the laboratory falling headpermeameter and mini-permeameter. Afterthe correlation is applied to the CPT data, thesmall-scale conductivity estimates are scaledup to the flow model grid (field scale) througha process described in the next section. Theupscaled estimates for horizontal conductivitycompare favorably with average field-scalemeasurement data from slug and pumpingtests, and previous calibrated flow models.The predicted conductivity fields based onTable 1 are viewed as a starting point for finalflow model calibration to additional targetssuch as water level and plume data.

Upscaling to the Groundwater Flow ModelGrid

CPT measurements have a vertical resolutionof 0.1 ft, and the radius of influence of thecone in the horizontal plane is similarly small.Therefore, a method for upscaling CPTmeasurements to the coarser resolution of thegroundwater flow model mesh is required.Upscaling refers to the process of replacing aheterogeneous conductivity field within aparticular finite volume with a single,"equivalent", conductivity value. The readeris referred to Sanchez-Vila and others (1995),


I-8

Wen and Gomez-Hernandez (1996), andRenard and de Marsily (1997) for a review ofupscaling and related topics in the stochastichydrology.

The stochastic upscaling approach in thisstudy is based on combining the work ofGelhar and Axness (1983), Ababou and Wood(1990), and Desbarats (1992). Gelhar andAxness (1983) derived analytical expressionsfor the effective conductivity tensor of aninfinite, ergodic, anisotropically-correlatedmedium subjected to a uniform mean flow.The three-dimensional anisotropy of theheterogeneous medium is defined in terms ofan exponential autocovariance function withdistinct correlation scales for each coordinatedirection, �1, �2 and �3. Ababou and Wood(1990) pointed out that expressions of the typederived by Gelhar and Axness (1983) canalternatively be written in terms of a "power-average" defined by

( )p/1

pp/1

i

pip KK

N1

K ��

��=

��

�

�≡ � (5)

where the exponent (power) p is applicationspecific. Arithmetic, geometric and harmonicaveraging are special cases of equation (5)gotten by choosing p = 1, 0 and -1. Hence,power-averaging is a generalization ofcommon approaches to averaging. Throughnumerical studies, Desbarat (1992)demonstrated that the power-averages are aneffective approximation to upscaledconductivity, provided an appropriateexponent is chosen. The work of Gelhar andAxness (1983) provides guidance for selectingexponents appropriate for upscaling horizontaland vertical conductivity in C-area.

The conductivity estimates inferred from CPTmeasurements have a vertical resolution of 0.1ft and are considered to be "point"measurements. Atlantic Coastal Plainsediments are clearly stratified and implyanisotropic correlation scales in the horizontaland vertical directions, �h and �v. Their ratio

cannot be derived from the CPT data becausenone of the 139 locations are close enough inthe horizontal plane (i.e. within inches).However, judgement based on knowledge ofthe depositional environment and visualinspection of outcrops suggests a reasonableratio is approximately �h / �v = 10.Furthermore, the vertical and horizontalcorrelation scales within stratigraphic zonesare probably on the order of a few inches andseveral feet, respectively. Considering thatthe typical model block will be a few feetthick and span on the order of 100 ft in thehorizontal plane, the "equivalent" blockconductivity should be close to the "effective"conductivity (Kitanidis, 1997), in theterminology of Sanchez-Vila and others(1995). Therefore, power averagingparameters derived from Gelhar and Axness(1983) can reasonably be applied to upscalingof C-area CPT data. For �h / �v = 10, thepower-averaging exponents are ph = +0.86 andpv = -0.72 for horizontal and verticalconductivity, respectively. For horizontalconductivity the averaging is nearlyarithmetic, and for vertical conductivityaveraging is similar to harmonic.

To avoid introducing bias into the predictedconductivity fields, the following upscalingprocess is chosen for Kh and Kv:

1) Transform the point data K to pK , wherep = +0.86 for horizontal conductivity andp = -0.72 for vertical conductivity

2) For each CPT push, arithmetically averagepK over the thickness of each model

layer

3) Interpolate pK within each layer usingtwo-dimensional block kriging

4) Back transform pK block estimates by

computing p/1p )K( , where p takes onthe value used in step 1)


I-9

Vertical averaging in step 2) reduces thethree-dimensional CPT data to a sequence oftwo-dimensional data sets, one per modellayer. Model layers conform to stratigraphicunits, zones or sub-zones. Kriging is chosenin step 3) because it is an exact interpolatorfor point estimation and can be optimized forthe characteristics of individual formationsthrough the variogram model.

For kriging, an isotropic exponentialsemivariogram model was chosen (e.g. deMarsily, 1986). The functional form is

��

��

��

�

�−−=ahch 3

exp1)(γ (6)

where

γ(h) ≡ semivariogram;

[ ]��

�� −+

2pp )x(K)hx(KE21 �

�

�

h ≡lag distance, h�

c ≡contribution

a ≡effective range

A variogram analysis of the synthetic pKdata for each model layer in the UTRA wasperformed to estimate the effective range, a.As an example, the experimental variogramsfor horizontal and vertical conductivity formodel layer #13 in the Transmissive Zone arepresented in Figure 9. The variograms arebased on the 139 lithologic CPT pushes inC-area, among a larger total number thatincluded groundwater sampling. Theestimated variogram ranges (a) vary from 300to 1000 ft, confirming that the vertically-upscaled CPT estimates are relevant at fieldscale. Occasionally a nugget effect isobserved. For convenience and consistency, amodel variogram with a range of 750 ft and nonugget effect was subsequently used in theblock kriging performed for each unit. Themodel variograms for model layer #13 are also

shown in Figure 9. The contribution, c, doesnot affect the block conductivity estimates soits specification can be arbitrary. Thecontribution does affect the uncertainty of theconductivity estimates, and an estimate wouldbe needed if an uncertainty analysis weredesired.

Predicted Conductivity Fields for C-Area

Figures 10 through 13 illustrate predictedupscaled hydraulic conductivity for key modellayers which represent in descending order,the TZ, the TCCZ, the uLAZ, and CC1(Figure 3). The spatial trends observed in thepredicted conductivity fields are supportive ofgeological interpretations based on CPT data,groundwater flow directions and plumemovement. Figure 10 illustrates the predictedhorizontal conductivity for the TZ of theUAZ. This zone is importanthydrogeologically because it represents asandy saturated zone overlying the TCCZ andis a controlling factor for the water table flowdirections. The predicted conductivities overa large portion of the study area exceed 10feet/day, which is consistent with field testconductivities for this aquifer zone from otherareas at SRS. Figure 11 illustrates verticalconductivity for the TCCZ underlying the TZ.In the northern portion of the study area thepredicted conductivities of roughly 10-3 ft/dayare consistent with known TCCZconductivities at SRS and are indicative of agood confining zone. In contrast, the southernportion of the study area near Caster Creekexhibits much higher conductivities (roughly3×10-2) that are more typical of aquifers atSRS. Interpretation of CPT data suggest thattoward the south the zone becomes a coarsegrained sandy material while the clayeysediments become very thin anddiscontinuous. A possible explanation isfluvial channeling which would producecoarser grained sandy sediments with higherconductivities and aquifer characteristics.Figure 12 illustrates the predictedconductivities for the uLAZ which liesdirectly beneath the TCCZ. Good aquiferconductivities are observed in the northern


I-10

portion of the study area and excellent aquiferconductivities are prevalent in the southernregion. Figure 12 illustrates the predictedconductivities for CC1. Geologically thiszone is interpreted to represent a calcareousconfining sub-zone within the LAZ. Incontrast to the TCCZ conductivities decreasefrom the northern to southern portions of thestudy area. Predicted values indicate that CC1is a very competent confining zone havingvalues of 10-4 ft/day or lower. Thecombination of the range of aquifer andconfining zone conductivities of these fourhydrostratigraphic intervals control shallowgroundwater flow and contaminant migration.

Tritium Contaminant Migration Pathways

Prior to the detailed conductivity fieldpredictions just presented, preliminarygroundwater modeling flow predictions ofgroundwater flow from the CRSB indicated awesterly migration of tritium towards TwinLakes and Fourmile Branch (Figure 13), alonga path parallel to the VOC contamination fromCBRP. However, this interpretationcontradicted depth discrete tritium data fromCPT. Tritium concentrations above 400pCi/ml are shown in Figure 13. These dataindicate that the tritium plume is movingprimarily south toward Caster Creek, withsmaller concentrations discharging toFourmile Branch well south of Twin Lakes.

Based on initial hydrostratigraphic CPTinterpretations, the initial groundwater modelassumed that the LAZ functioned as arelatively homogenous aquifer zone, and theTCCZ as a competent confining zone. Carefulre-examination of the CPT data, coupled withnearby core data, identified two lowpermeability calcareous intervals in the LAZ.Based on the heterogeneity within the LAZand for the purposes of this study, the LAZwas subsequently sub-divided into fiveinformal hydrostratigraphic intervals.Additionally, the TCCZ was recognized asbecoming extremely sandy in the southernregion of the study area with the clays verythin and discontinuous. Within this

framework, the observed CRSB plumemovement can be explained by thecorresponding conductivity variationspredicted in Figures 10 through 13. In thesouthern portion of the study area, higherhorizontal conductivities in the TZ and uLAZ,coupled with the water table being supportedby the deeper CC1 rather than the TCCZ,preferentially draw groundwater towardCaster Creek.

Figure 15 illustrates simulated tritiummigration from CRSB based on the revisedhydrostratigraphic intervals and conductivityvariations. In agreement with the CPT depthdiscrete contaminant data the bulk of thetritium plume discharges to Caster Creek witha low concentration fringe discharging toFourmile Branch.

CONCLUSIONS

CPT was successfully utilized in the absenceof conventional borehole data for geospatialand contaminant plume mapping. Thevariations of hydraulic conductivity within thehydrostratigraphic zones of the UTRA werepredicted by coupling CPT data andgeostatistical methods. Incorporating thehydraulic conductivity variations intogroundwater modeling significantly improvedflow path predictions. The model conductivityfield provided a realistic representation ofaquifer heterogeneity in coastal plainsediments. The resulting tritium transportsimulations compared well with field data.The CPT geospatial mapping is a viablemethod for characterization ofhydrostratigraphy and sediment behavior incoastal plain depositional environments.


I-11

REFERENCES

Aadland, R. K., Gellici, J. A., and Thayer, P. A.,1995. Hydrogeologic Framework of West-CentralSouth Carolina. Report 5, Water ResourcesDivision, South Carolina Department of NaturalResources, Columbia, SC.

Aadland, R. K., Harris, M. K., Lewis, C. M.,Gaughan, T. F., and Westbrook, T. M., 1991.Hydrostratigraphy of the General Separations Area,Savannah River Site (SRS), South Carolina.WSRC-RP-91-13, Westinghouse Savannah RiverCompany, Aiken, SC 29808.

Ababou, R. and E. F. Wood, 1990, Comment on"Effective groundwater model parameter values:influence of spatial variability of hydraulicconductivity, leakance, and recharge" by J. J.Gomez-Hernandez and S. M. Gorelick, WaterResources Research, v26, n8, 1843-1846.

Celeste, 1998, Permeability of R-area sediments,Savannah River Site, South Carolina:Minipermeameter vs. cone penetrometertechnology, M. S. Thesis, Department of EarthSciences, University of North Carolina atWilmington.

de Marsily, G., 1986, Quantitative hydrogeology;Groundwater hydrology for engineers, AcademicPress.

Desbarats, A. J., 1992, Spatial averaging ofhydraulic conductivity in three-dimensionalheterogeneous porous media, MathematicalGeology, v24, n3, 249-267.

Fallaw, W. C. and V. Price, 1995, Stratigraphy ofthe Savannah River Site and vicinity, SoutheasternGeology, v35, 21-58.

Flach, G. P., L. L. Hamm, M. K. Harris, P. A.Thayer, J. S. Haselow and A. D. Smits, 1998, Amethod for characterizing hydrogeologicheterogeneity using lithologic data, in G. S. Fraserand J. M. Davis, Hydrogeologic models ofsedimentary aquifers, SEPM Concepts inHydrogeology and Environmental Geology No. 1.

Flach, G. P. and M. K. Harris, 1999, Integratedhydrogeological modeling of the GeneralSeparations Area, Volume 2, Groundwater flowmodel, WSRC-TR-96-0399, Rev. 1.

Flach, G. P., M. K. Harris, R. A. Hiergesell, A. D.Smits and K. L. Hawkins, 1999, Regionalgroundwater flow model for C, K, L, and P areas,Savannah River Site, Aiken, South Carolina (U),WSRC-TR-99-00248, Rev. 0.

Harris, M. K., Aadland, R. K., and Westbrook, T.M., 1990. Lithological and HydrologicalCharacteristics of the Tertiary HydrostratigraphicSystems of the General Separations Area, SavannahRiver Site, South Carolina. Savannah RiverRegion: Transition Between the Gulf and AtlanticCoastal Plains, Proceedings of the Second BaldHead Island Conference on Coastal PlainsGeology, University of North Carolina atWilmington, pp 68-73.

Harris, M.K., Thayer, P.A. and Amidon, M.B.,1997, Sedimentology and depositionalenvironments of middle Eocene terrigenous-carbonate strata, southeastern Atlantic CoastalPlain, USA, v108, nos. 1-4, February, 163-180.

Gelhar, L. W. and C. L. Axness, 1983, Three-dimensional stochastic analysis of macrodispersionin aquifers, Water Resources Research, v19, n1,161-180.

Jefferies, M. G. and M. P. Davies, 1993, Use of theCPT to estimate equivalent SPT N60, GeotechnicalTesting Journal, GTJODJ, v16, n4, December, 458-468.


I-12

Kegley, W. P., 1993, Distribution of permeabilityat the MWD well field, Savannah River Site,Aiken, South Carolina, M. S. Thesis, ClemsonUniversity.

Kitanidis, P. K., 1997, Groundwater flow inheterogeneous formations, in G. Dagan and S. P.Neuman, eds., Subsurface flow and transport: astochastic approach, Cambridge University Press.

Lunne, T., P. K. Robertson and J. J. M. Powell,1997, Cone Penetration Testing in GeotechnicalPractice, Blackie Academic and Professional.

Parsons Engineering Science, Inc. and AppliedResearch Associates, Inc., 1997, Evaluation ofcone penetrometer data for permeability correlationat the Savannah River Site, WSRC-RP-97-63.

Renard, Ph. and G. de Marsily, 1997, Calculatingequivalent permeability: a review, Advances inWater Resources, v20, n 5-6, 253-278.

Robertson, P. K. and C. E. Fear, 1995,Liquefaction of sands and its evaluation, ISTOKYO 95, First International Conference onEarthquake Geotechnical Engineering, KeynoteLecture, November.

Sanchez-Vila, X., J. P. Giradi and J. Carrera, 1995,A synthesis of approaches to upscaling of hydraulicconductivities, Water Resources Research, v31, n4,867-882.

Syms, F.H., Wyatt, D.E. and Flach G.P., 2000,Methodology and Interpretation of the PiezoconePenetrometer Test Sounding (CPT) for EstimatingSoil Character and Stratigraphy at the SavannahRiver Site, this volume.

Wen, X-H. and J. J. Gomez-Hernandez, 1996,Upscaling hydraulic conductivities inheterogeneous media: An overview, Journal ofHydrology, 183, ix-xxxii.

WSRC, 1993. Site Evaluation Report for the C-Area Reactor Seepage Basin (904-68G)(U),WSRC-RP-93-360, Westinghouse Savannah RiverCompany, Aiken, SC, 29808.

WSRC, 1994, Phase II, Revision 2, RCRA facilityinvestigation/remedial investigation plan for the C-Area Burning/Rubble Pit (131-C) (U), WSRC-RP-91-1122, Rev. 2.

Table 1. Hydraulic conductivity correlation.

Conductivity category Definition Value (ft/d)

High 0-15% fines predicted from equation (4) 20

Medium 15-30% fines predicted from equation (4) 2

Low >30% fines predicted from equation (4) 0.01

Extra Low Bq > 0.1 irrespective of predicted %fines 0.0001


I-13

Figure 1. Location of study area.


I-14

Figure 2. Locations of CPT and borehole data within C-area.


I-15

Figure 3. Lithostratigraphic and hydrostratigraphic nomenclature.


I-16

Figure 4. Composite log illustrating correlation between CPT, core and geophysicaldata.


I-17

Figure 5. Three-dimensional map of hydrostratigraphic model.


I-18

Figure 6. Conceptual model of hydrogeology and groundwater flow.


I-19

Fsn

Bq

10-1 100 101-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

10-15% fines (center of mass)15-30% fines (center of mass)30-50% fines (center of mass)50-100% fines (center of mass)0-15% fines (sieve data)15-30% fines (sieve data)30-50% fines (sieve data)50-100% fines (sieve data)

Figure 7. Normalized pore pressure (Bq) and friction ratio (Fsn) CPT data, color-codedby %fines content (group 1 / blue = 0 to 15%; group 2/green = 15 to 30%; group 3 / red= 30 to 50%; group 4 / yellow & black = 50 to 100%).


I-20

Fsn

Qtn

10-1 100 10110-1

100

101

102

0-15% fines (center of mass)15-30% files (center of mass)30-50% fines (center of mass)50-100% fines (center of mass)0-15% fines (sieve data)15-30% fines (sieve data)30-50% fines (sieve data)50-100% fines (sieve data)

Figure 8. Normalized tip resistance (Qtn) and friction ratio (Fsn) CPT data, color-codedby %fines content (group 1 / blue = 0 to 15%; group 2/green = 15 to 30%; group 3 / red= 30 to 50%; group 4 / yellow & black = 50 to 100%).


I-21

Lag, ft

Var

iogr

amof

(Kh)

p ,(ft/

d)2p

Var

iogr

amof

(Kv)

p ,(ft/

d)2p

0 500 1000 1500 20000

5

10

15

20

25

0

1000

2000

3000

4000

5000

6000

7000

8000

Experimental variogram - (Kh)p

Model exponential variogram - (Kh)p

Experimental variogram - (Kv)p

Model exponential variogram - (Kv)p

Figure 9. Variogram analysis for Kp in model layer #13 in "transmissive" zone (TZ).


I-22

1.51.31.10.90.70.50.30.1

-0.1-0.3-0.5

log10 Kh(ft/d)

Fourmile Branch

Caster Creek

CBRP

CRSB

N

1000 ft

Figure 10. Block kriging estimates of Kh in model layer #12 (Transmissive Zone).


I-23

0-0.5-1-1.5-2-2.5-3-3.5-4

log10 Kv(ft/d)

Fourmile Branch

Caster Creek

CBRP

CRSB

N

1000 ft

Figure 11. Block kriging estimates of Kv in model layer #11 (Tan Clay Confining Zone).


I-24

1.51.31.10.90.70.50.30.1

-0.1-0.3-0.5

log10 Kh(ft/d)

Fourmile Branch

Caster Creek

CBRP

CRSB

N

1000 ft

Figure 12. Block kriging estimates of Kh in model layer #9 (upper LAZ).


I-25

0-0.5-1-1.5-2-2.5-3-3.5-4

log10 Kv(ft/d)

Fourmile Branch

Caster Creek

CBRP

CRSB

N

1000 ft

Figure 13. Block kriging estimates of Kv in model layer #8 (CC1).


I-26

Fourmile

Branch

Caster Creek

C ReactorSeepage Basins

C Burning/Rubble Pit

VOC plume

Tritiumplume

N

1000 ft

Figure 14. Plan view of CPT tritium concentration data exceeding 400 pCi/ml.


I-27

20000100005000200010005002001005020

pCi/ml

N

1000 ft

Figure 15. Simulated tritium plume originating from C-Reactor Seepage Basins.


I-28

Notes


J-1

Soil Vapor Extraction Remediation of Trichloroethylene Contamination atthe Savannah River Site’s C-Area Burning Rubble Pit

Christine Switzer and David Kosson, Department of Civil and Environmental Engineering, Vanderbilt University,Box 1831, Station B, Nashville, TN 37235Joseph Rossabi, Savannah River Technology Center, Aiken, SC, 29808

ABSTRACT

Soil vapor extraction (SVE) has become theEnvironmental Protection Agency’s (EPA)presumptive remedy for cleanup of the vadosezone when contaminated with volatile organiccompounds. Air sparging (AS) has beenemployed to remediate shallow groundwatercontamination, normally in conjunction withSVE. A SVE/AS pilot system was installed at asmall waste area on the Savannah River Site toremediate trichloroethylene (TCE)contamination and evaluate the technology forapplication to future projects. The system hasbeen operating since October 1999.Concentrations of TCE measured in vadose zoneextraction wells and monitoring points betweenextraction wells have decreased since operationbegan. After periods of shutdown, subsurfacesoil gas concentrations of TCE have reboundedto high levels indicating remainingcontamination, possibly in the form of dispersednon-aqueous phase liquid (NAPL) or a masstransfer limited phase in the subsurface.

INTRODUCTION

In the past ten years, soil vapor extraction (SVE)has become the presumptive remedy for theremediation of volatile organic compound(VOC) contamination in soils and sediments.Because of the unique aspects of individualwaste sites and contaminant configurations, nostandard approach to SVE design or modelinghas been adopted, although several guidancedocuments have become available (USEPA,1994).

In typical implementations of SVE, wells arescreened in the vadose zone according to thegeology and the contaminant distribution.Vacuum is applied to the configuration and thecontaminant is drawn out of the subsurface andeither treated or discharged. Treatment methodsinclude collection on activated carbon,

collection of the VOC by condensation, andthermal oxidation techniques.

During SVE operation, three phases of masstransport have been observed (Hiller andGuidermann, 1989; Johnson et al., 1990). At thebeginning of SVE operation, concentrationsremain approximately constant as the availablegas phase contaminant is removed from thesubsurface pore spaces that are in the directlyaccessible flow path of the vapor extractionsystem. The length of this period depends on thetype of soil, the vapor flow achieved through thesubsurface, and the extent of contamination.Operation then enters a transition period whereconcentrations decline drastically as the VOC inthe accessible flow path is depleted. Diffusion ofVOC from less accessible pores and masstransfer from the liquid or aqueous phase to theaccessible pores limit VOC recovery.Eventually, a point is reached where massremoval becomes diffusion-limited. In thediffusion-limited regime, contaminant massremoval rates may decline further, even if totalmass flow rates are held constant, because VOCdiffusive flux may have been diluted byadvective flow from clean areas. When thesystem is shutdown, the accessible pores in theflow path had time to re-equilibrate with the lessaccessible sources of contaminant. In thisregime, the rebound regime, subsurfacecontaminant concentrations in the soil gasincrease as more contaminant became availablefor removal.

Remediation by SVE/AS can be greatly affectedby the presence of non-aqueous phase liquid(NAPL) in the subsurface. Modeling approacheshave varied with respect to inclusion of a NAPLcomponent. Some researchers neglected theNAPL phase entirely for reasons that includedthe age of the spill and the type of soil in whichthe spill occurred (Ng and Mei, 1996, Kalerisand Croise, 1999, Ng, 1999). The presence ofpartially saturated, low permeability material inthe subsurface increases the likelihood of NAPL,


J-2

as capillary action retains the NAPL in the porespace (Massmann et al., 2000). Partiallysaturated conditions also diminish the effectivediffusivity of the contaminant in the gas phaseby the cube (approximately) of the gas phasesaturation (Rossabi, 1999). Smaller effectivediffusivity limits the volatilization and removalof the NAPL.

SITE BACKGROUND

The C-Area Burning Rubble Pit (CBRP) at theDepartment of Energy’s Savannah River Site(SRS) was used for trichloroethylene (TCE)disposal during the site’s Cold War activities.The pit was active from the 1950’s to the1980’s. The amount of TCE disposed in the pitis not accurately known because disposalrecords were not kept. The site was selected as apilot site for evaluation of soil vapor extractionand air sparging (SVE/AS) technologies.

Prior to startup of SVE/AS operation, sitecharacterization was carried out to aid in systemdesign. In January 1999, cone penetrometerpushes were carried out at six locations todetermine subsurface stratigraphy. Thirty-nineSVE wells and seventeen AS wells wereinstalled in the western portion of the pit inFebruary and March (Figure 1). Some wellswere installed in a clustered configuration ofthree depths per location. Since these threedepths differed in each cluster, the deepest hasbeen designated as the “A” well, the middledesignated the “B” well and the most shallowdesignated the “C” well. Locations 1 through 12and 15 through 17 have a single well installed atthe equivalent “A” depth. Locations 13, 14 and18 through 23 have the clustered configuration.The vadose zone wells were not developedbefore operation. The wells are manifoldedtogether with underground piping leading to thesoil vapor extraction and catalytic oxidationtreatment unit. The unit was designed to operatewith a vacuum of up to 24 kilopascals at eachwell and a flow capacity of more than 470 cm3/s(600 scfm).

Three well locations (SVE 18, 19 and 22) wereselected for extensive characterization. All threeof these locations were clustered configurations.In May 1999, twenty-three soil cores were takenfrom locations along the triangle connectingSVE 18, 19 and 22. In June, EPA installed

twenty-six concentration and pressuremonitoring implants along the same triangle(Figure 1).

Operation began at CBRP in October 1999.Operation was limited to SVE 18, 19 and 22 atthe B and C depths. Samples were collected atthe wells and the implants at regular intervals.Operation of additional wells began inDecember.

In January 2000, the system was shutdown for afew days and rebound was observed based on asingle sampling event. In late January, thesystem was shutdown for two weeks to measurethe rebound effect with more detail. The systemwas turned on again for a few days in mid-February before it was shutdown for repairs.Many of the wells and some of the undergroundpipes in the manifold system had becomeclogged with sediments and required cleaningbefore operation could resume.

Operation resumed in late April using a subsetof available wells. The wells that continuedoperating were 12, 13AB, 17, 18AC, 19AC,21A, 22AC and 23C.

RESULTS AND DISCUSSION

Initially, data were collected daily at alloperating SVE points. Monitoring transitionedgradually to semiweekly, weekly, biweekly andthen monthly. Between each transition, VOCconcentration and operating pressure weremeasured at the implants. After each long-termshutdown event, the sampling schedule wasrevised to capture the expected system behavior.

Operational data showed the characteristicbehavior of a SVE system (Figure 2).Concentrations decreased over time andeventually reached a point where continuedoperation produced little mass removal. Whenthe system was shutdown, concentrationsrebounded significantly at most locations. Sinceoperation began, TCE concentrations havedecreased at the wells and implants from theirinitial values. The observed high rebound hasindicated that residual TCE contamination mustcontinue to be addressed.


J-3

The highly heterogeneous vadose zone withmany low permeability layers (Figure 3)increased the probability of NAPL in the vadosezone, despite the age of the pit. The observedhigh soil gas concentration rebound at wells 18,19 and 22 during the two monitored shutdownperiods also suggested the possibility of NAPL,because TCE concentrations rebounded to levelsat or near those before SVE operation began(Figure 2). Concentrations of TCE at the Cseries wells were as high as 5% of the TCEsaturated partial pressure during the reboundevents. Pressure and concentration data recordedat the EPA implants have helped provide a betterunderstanding of contaminant distribution andbehavior during operation and rebound periods(Figure 4). Implant data suggested that, althoughhigh concentrations have persisted at the wells,SVE operation has reduced concentrations ofTCE in the subsurface at distances away fromthe wells.

Many of the wells experienced low vapor flowor a loss of flow due to clogging by sedimentsand water. The screened zones were selected tobe in permeable materials (sandy) that wereadjacent to fine grain units (silt and clay) whereresidual contamination was likely to reside.Because of the heterogeneity of the formation,however, screened zones often crossed somefine grain units as well as the more permeablezones. Since the wells were not developedbefore operation began, the initial operatingperiod served as a well development cycle,producing an accumulation of fines and water inthe wells. The wells were cleaned by bailing andbrushing. Limited operation of a subset of thewells was conducted so that the remaining wellscould be cleaned and fitted with filters to limitfines traveling through the pipes to theextraction unit. The limited operation at the pitarea, which began in April, isolated severalproblem wells. Flow criteria were defined as“good flow” at wells producing 5 cfm or greater,“acceptable flow” as wells producing 3 cfm orgreater and “poor flow” at wells producing lessthan 3 cfm. Wells 18C and 22C had poor flowand high concentrations. In contrast, well 19Chad good flow and a much lower concentration.When the system was shutdown again the firstfew days of June, soil gas concentrations of TCEat all wells rebounded. Poor flow was observedin most of the operating wells in the limitedconfiguration (Figure 5).

After several weeks of operation, the removalrate of TCE observed at well 19C was more thantwice that observed at 18C and four times thatobserved at 22C (Figure 5). Possibleexplanations for the poor flow at 18C and 22Cincluded well placement and sedimentaccumulation. Because 18C and 22C werescreened mostly into the stiff clay layer, residualmoisture may have decreased the permeabilityof the sand layers above and below, where highpermeability would normally be expected.Advective flow was determined by the relativepermeability of the sediments, which dependedon the tortuosity of the subsurface material. Thisparameter also scaled with the cube of the gassaturation in subsurface pore space (Rossabi,1999). As the relative saturation of water in atwo fluid phase porous medium increased, theadvective flow decreased. The remedies to theflow problem included removing theaccumulated sediments from the wells,developing the wells, and injecting air in at bothlow and high pressures to create pathways forairflow to the wells. These repairs have beensuccessful for many but not all of the wells.Some of the poorly performing wells may havebeen screened in zones that are inaccessible topermeable sediments.

Early zone of influence testing at CBRP, whichused the EPA implants to measure pressurechanges in the subsurface when each well in thegroup of SVE 18, 19 and 22 was operatedindividually, provided interesting insight to theflow problems. During pressure testing, a singleSVE well in that group was operated andpressure was recorded at the EPA implants.Implant depth varied by location. For mostlocations, there were two implant depthsinstalled. For simplicity in discussion, thesewere grouped as “upper” (above 30ft) and“lower” (below 30ft). These groupingscorresponded with location of the stiff claylayer, which was present between 25 and 30 ft(Figure 3b). Comparison of vacuum at 19C andvacuum response at the implants, a well that thatproduced high flow rates, and 22C, a well thatproduced low flow rates, supported wellplacement as a significant contributing factor towell performance. Both of these wells wereincluded in the “upper” grouping becauseportions of their screens were above 30ft. While19C was operating, vacuum diminished in thecorresponding “upper” implants as distance from


J-4

the well increased. Vacuum recorded at the“lower” implants was close to or at zero. While22C was operating, vacuum at the “upper”implants was small, approaching zero whilevacuum at the “lower” implants was muchgreater, suggesting that 22C was inducing flowbelow the stiff clay layer (in the clayey sand)instead of above it (Figure 6).

This same zone of influence testing at 18C alsoprovided insight to the low flow there. When18C was operated, a significant vacuum wasmeasured at the nearest of the shallow implantsonly. Little to no vacuum was measured at all ofthe other implants, suggesting a small radius ofinfluence for 18C (Figure 7).

CONCLUSIONS

Soil gas concentrations of TCE in the subsurfacehave decreased significantly at many locationssince operation of the SVE/AS system began.The observed high rebounds in soil gas TCEconcentration at several locations indicated thatTCE contamination must continue to beaddressed. Limited operation at the pit area,which began in April, isolated several poorlyperforming wells. Low flow was observed inmost of the operating wells in the limitedconfiguration. Possible explanations for the poorflow included well placement and sedimentaccumulation. Some of the wells were placedacross lower permeability zones. The attemptedremedies to the poor flow problem have beenpartially successful. Despite these problems, theobserved decrease in soil gas TCEconcentrations at the implants indicatedsignificant removal of TCE from the subsurfaceby the SVE/AS system.

ACKNOWLEDGEMENTS

The authors gratefully acknowledge the financialsupport of the Consortium for Risk Evaluationwith Stakeholder Participation (a cooperativeagreement with the Department of Energy) andthe assistance of Jerry Nelsen, Mary Harris,Gregory Flach, Michael Morgenstern, JohnBradley, Keith Hyde and Brian Riha.

REFERENCES

USEPA (1994). EPA Soil Vapor Extraction (SVE)Treatment Technology Resource Guide. Office ofSolid Waste and Emergency Response TechnologyInnovation Office Washington, DC, 20460,EPA/542-B-94-007.

Hiller, D., and H. Guidemann (1989). “Analysis ofvapor extraction data from applications in Europe.”In Proceedings of Third International Conference onNew Frontiers for Hazardous Waste Management,EPA/600/8-89/072, Pittsburgh, PA, September 10-13.

Johnson, P. C., M. W, Kemblowski, J. D. Colthart(1990). “Quantitative analysis for the cleanup ofhydrocarbon-contaminated soils by in situ venting.”Ground Water, 28(3): 413-429.

Kaleris, V. and J. Croise (1999). “Estimation ofcleanup time in layered soils by vapor extraction.”Journal of Contaminant Hydrology 36: 105-129.

Massmann, J. W., S. Shock, et al. (2000).“Uncertainties in cleanup time for soil vaporextraction.” Water Resources Research 36(3): 679-692.

Ng, C.-O. (1999). “Macroscopic equations for vaportransport in a multi-layered unsaturated zone.”Advances in Water Resources 22(6): 611-622.

Ng, C.-O. and C. C. Mei (1996). “Aggregatediffusion model applied to soil vapor extraction inunidirectional and radial flows.” Water ResourcesResearch 32(5): 1289-1297.

Rossabi, J. (1999). “The influence of atmosphericpressure variations on subsurface soil gas and theimplications for environmental characterization andremediation.” Ph.D. thesis, Clemson University,University of Michigan Press.


J-5

Figure 1. Diagram of SVE and AS well locations with region of study indicated.

1.E-01

1.E+00

1.E+01

1.E+02

1.E+03

1.E+04

1.E+05

8/28/1999 11/6/1999 1/15/2000 3/25/2000 6/3/2000 8/12/2000

time

conc

entr

atio

n (p

pmv)

18C

18B

1.E-01

1.E+00

1.E+01

1.E+02

1.E+03

1.E+04

1.E+05

8/28/1999

11/6/1999

1/15/2000

3/25/2000

6/3/2000 8/12/2000

time

con

cen

trat

ion

(p

pm

v) 19C

19B

1.E-01

1.E+00

1.E+01

1.E+02

1.E+03

1.E+04

1.E+05

8/28/1999 11/6/1999 1/15/2000 3/25/2000 6/3/2000 8/12/2000

time

con

cen

trat

ion

(pp

mv)

22C

22B

Figure 2. Concentration of TCE recorded at a) SVE18, b) SVE 19 and c) SVE 22.

a) b)

c)


J-6

Figure 3. Detailed stratigraphy based on CPT information (courtesy of Greg Flach and MaryHarris, Savannah River Technology Center). SVE well locations are indicated by the larger blockswith screen (A, B or C) designated, and shelby tube sample locations are indicated by the smallerblocks.

a)

b)

Carolina G


ield Trip G

uidebookW

SRC

-MS-2000-00606

J-7

Figure 4. A two-dimensional perspective on the concentration data recorded at CBRP. Plots are two-dimensional in space, taken at three discrete time points : before operation, after three months of operation,and after two weeks of rebound. Area of the bubble is proportional to concentrations.


J-8

0

5

10

15

20

25

30

4/4/2000 4/24/2000 5/14/2000 6/3/2000 6/23/2000

time

flo

w (

ft^3

/min

)

18c

19c

22c

0

1

2

3

4

5

6

7

8

4/4/2000 4/24/2000 5/14/2000 6/3/2000 6/23/2000

time

vacu

um

pre

ssu

re (

in H

g)

18c

19c

22c

0

500

1000

1500

2000

2500

3000

3500

4/4/2000 4/24/2000

5/14/2000

6/3/2000 6/23/2000

time

con

cen

trat

ion

(p

pm

v)

18c

19c

22c

0.0E+00

5.0E-04

1.0E-03

1.5E-03

2.0E-03

2.5E-03

4/4/2000 4/24/2000 5/14/2000 6/3/2000 6/23/2000

time T

CE

rem

ova

l rat

e (s

cfm

)

18c

19c

22c

Figure 5. a) Total flow, b) vacuum pressure, c) concentration and d) removal rate of TCE recordedat SVE 18C, 19C and 22C after the system was down for one month of repairs.

0

5

10

15

20

25

0 6 12 18 24 30

distance from operating well (ft)

vacu

um

pre

ssu

re (

in H

2O)

22c on - upper (<30ft)22c on - low er (>30ft)19c on - upper (<30ft)19c on - low er (>30ft)

0

1

2

3

0 6 12 18 24 30distance from operating well (ft)

vacu

um

pre

ssu

re (

in H

2O)

22c on - upper (<30ft)22c on - low er (>30ft)19c on - upper (<30ft)19c on - low er (>30ft)

Figure 6. Vacuum pressure measured during preliminary pressure testing at the EPA implantsplotted with respect to distance from the SVE well that was operating, shown at a) large scale (0-25kPa) and b) small scale (0-3kPa). Implant depth varied by position so locations were categorizedas upper (above 30ft) and lower (below 30ft).

a) b)

c) d)

a) b)


J-9

0

5

10

15

20

25


vacu

um

pre

ssu

re (

kPa)

upper (<30ft)middle (30-49ft)low er (50-53ft)upper (<30ft)low er (>30ft)

0

1

2

3


vacu

um

pre

ssu

re (

kPa)

upper (<30ft)middle (30-49ft)low er (50-53ft)upper (<30ft)low er (>30ft)

Figure 7. Vacuum pressure measured during preliminary pressure testing at the EPA implantswhile SVE 18C was operating, shown at a) large scale (0-25kPa) and b) small scale (0-3kPa). Datawere recorded at two sets of implants near 18C.

a) b)


J-10

Notes


K-1

DNAPL Penetration of Saturated Porous Media Under Conditions ofTime-Dependent Surface Chemistry

David M. Tuck, Westinghouse Savannah River Company, Aiken, SC 298081,Gary M. Iversen, WSRC, Aiken, SC, 29808William A. Pirkle, University of South Carolina, Aiken, SC 29801

1 Current address NAPLogic, Inc. P.O. Box 72, Princeton, NJ 08542-00722; email: [email protected]

ABSTRACT

Two long-term experiments were run toobserve the behavior of a DNAPL containinga nonvolatile, surface active dye in water-saturated porous media. The DNAPL wastetrachloroethylene (PCE), and the dye wasSudan IV, the most frequently used dye usedto stain organic phases in multiphase flowvisualization studies. The DNAPL penetratedthe top two millimeters of the porous mediumin which a PCE-dye solution with 0.508 g/L ofdye was used. The results provide the firstqualitative experimental evidence in support ofa pore-scale model which explicitlyincorporates the time-dependent surfacechemistry exhibited in the water-PCE-SudanIV dye system.

INTRODUCTION

The presence of dense, non-aqueous phaseliquids (DNAPLs) in the subsurface is a majorenvironmental problem facing the SavannahRiver Site, as well as other commercial andindustrial sites around the country. It isimportant to understand and attempt to predicttheir behavior during any enhancedremediation effort to avoid exacerbating theexisting problem. Previous research (Pirkle etal. 1997a) indicated that entry pressures areconsiderably different for PCE/Sudan IV dyesolutions than for pure PCE, due to the surfaceactive nature of the dye. This is importantsince most sites contaminated with PCE orother DNAPLs involve impure solutionscontaining potentially surface active

substances. The purpose of the currentresearch was to experimentally examine thesubsurface behavior of perchloroethylene(PCE), which is the major component of theSRS M-Area DNAPL, under conditions oftime-dependent surface chemistry.

BACKGROUND AND THEORY

Capillary forces play a major role in controllingwhere and how DNAPLs migrate in thesubsurface. DNAPL fluid saturation in aninitially water-saturated porous medium is afunction of the capillary pressure, i.e., thedifference between the wetting and non-wetting fluid pressures. This capillarypressure-saturation relationship is importantinput data required for modeling multiphasefluid flow in porous media.

In previously reported research (Pirkle et.al.,1996, 1997a), the Hobson Equation, amodification of the Young-Laplace equation(HOBSON, 1954, 1975), was changed toaccount for the fact that, under theexperimental conditions used, there was nobouyant force supporting the PCE column priorto its entry into the medium. For ourexperimental set-up (Figure 1) the Young-Laplace equation was expressed as:

t

owovc r

cos2gLP

θσ=ρ=

(1)


K-2

Where Lv is the height of the PCE or SUDANIV dye/PCE solution column prior topenetrating the medium (cm), σow is theinterfacial tension between SUDAN IV dyed-PCE solution and water (dynes/cm), θ is thecontact angle between the PCE-waterinterface and a glass surface, rt is the“effective” pore throat radius (cm), ρo is thedensity of PCE (g/cm3), and g is theacceleration of gravity (980 cm/ sec2 ).

SUDAN IV dye dissolved in PCE reducesinterfacial tension between the PCE and waterand changes the wetting relationship betweenPCE, water and glass. These changes occurrelatively slowly in a roughly log-linear fashionover time spans of minutes to days (Tuck andRulison 1998). The extent to which interfacialtension reduction and wetting changes takeplace depends upon the concentration andactivity of the dye (Tuck and Rulison 1998). Inaddition to the above considerations, theaddition of a dye may also affect otherparameters such as the fluid density andviscosity. The effects of the SUDAN IV dyeare analogous to the surface chemistry andphysical changes one might expect of a weak,oil-soluble surfactant. The extent to whichSudan IV dye effects the system surfacechemistry (contact angle and interfacialtension) depends on the concentration of thedye (see Figure 2 a and b).

The Hobson equation treats the surfacechemistry as constant, i.e., as equilibriumvalues. Recent theoretical work (Tuckaccepted, Tuck 1999, Tuck et al. 1998b)suggests that equilibrium values do notnecessarily yield the best understanding and,therefore most accurate modeling ofmultiphase flow in porous media. Surfacechemical kinetics, i.e., the time-dependentnature of surface chemical properties (surfaceor interfacial tension and contact angle), may,under certain circumstances, dominate thepore-scale processes which govern multiphase

flow in porous media. The time-dependencemay be represented as σ = σ(t) and θ = θ(t),where σ is the interfacial tension, θ is thecontact angle, and t is time.

The conditions under which surface chemicalkinetics (i.e., the time rate of change of thesurface chemical properties interfacial tensionand contact angle) become important are whenthe capillary pressure is close to the criticalcapillary pressure required for immiscibledisplacement to occur.

Tuck (accepted) postulated the age of aninterface gets re-set each time it becomesunstable and advances through the subsequentpore-body during a "Haines jump." Hesuggested this results in a time-dependentcritical capillary pressure for any given pore-throat because the surface chemistrycontrolling the stability of the interface is time-dependent. Thus for a given pore-throat ofradius rt and constant fluid pressure difference,Pc, a critical age (tcrit) at which an interfacewill become unstable is given when thefollowing equation is satisfied

0r

)t(cos)t(2P

t

critcritc =

θσ− (2)

Using this model, he showed that immiscibledisplacement can occur at lower hydrostaticfluid pressure differences than those normallymeasured in experimental work. His modelwas designed to approximate a fine sand, andthe surface chemical kinetics were those forthe water-glass-PCE-Sudan IV dye systemwith a dye concentration of 0.508 g/L. Themodeled PCE penetration rate under theseconditions decreased in log-linear fashion asthe hydrostatic fluid pressure differencedecreased. Penetration of the model mediumcontinued at hydrostatic fluid pressuredifferences well below the entry pressuredifference for pure PCE. The model results


K-3

thus predict that Sudan IV-dyed PCE shouldpenetrate a water saturated porous media evenat pressures below the normally defined "entrypressure", but the rate of penetration should bequite slow.

The objective of this research was to test theconceptual theoretical model developed byTuck (accepted). Under the relatively slowsurface chemical kinetics associated with theSudan IV dye, PCE should penetrate a watersaturated medium at hydrostatic fluid pressuredifferences less than measured in previousexperimental work (Pirkle et al. 1997a), but therate of penetration is expected to be very slow.

EXPERIMENTAL PROCEDURE

The experimental apparatus is illustrated inFigure 1. Two identical experimental set-upswith "uniform" diameter glass bead porousmedia were used. The glass beads wereobtained from Potter Industries Inc. Theywere washed and re-sieved at the Universityof South Carolina Aiken. The beads were drysieved to reduce the amount of material forfurther processing. Dry-sieved beads weresoaked overnight in ethanol, and thenthoroughly washed with ethanol to remove anyhydrocarbon residues. The beads werethoroughly rinsed with deionized (DI) water toremove the ethanol. Washed beads were thenwet classified on stainless steel sieves to obtainthe desired size range (US 170 mesh sieve,106µ > db > 90µ). Finally, the beads werepermitted to air dry, were homogenized, andstored in a dessicator prior to use. Theexperiments were set up with the same,narrow-size-range of glass beads used inpreviously reported work (Pirkle et al. 1997a).Both fluid reservoirs (water reservoir buretteand PCE standpipe) were Schellbach burettes,which allowed fluid volume changes of 0.01mL to be clearly detected. This corresponds toapproximately 33,000 unit voids (pores) ofdisplacement ocurring in the glass bead

medium assuming orthorhombic packing ofuniform spheres with a grain diameter midwaybetween the sieve limits (Graton and Fraser1935).

A known mass of the glass beads was loadedinto each apparatus. A vacuum of 21 inches ofmercury was applied. This vacuum was lessthan that which would cause water to boil atthe prevailing range of laboratory temperatures(approximately 22o ± 2 o C). The medium ofglass beads was saturated from the bottomunder the vacuum conditions. Several porevolumes of DI/DA water were draineddownward through the media as a final form ofpacking and stabilization of the media. Falling-head permeability measurements were thenmade on each medium (Pirkle et al. 1997a,1997b).

For the current experiments, two solutionsused in prior research were utilized except thatthey were equilibrated with deionized waterprior to use. Table 1 contains the dyeconcentration data. These concentrationscorrespond with concentrations Tuck andRulison (1998) used to characterize thesurface chemical kinetics of the water-glass-PCE-Sudan IV-dye system (see Figure 2).Solutions C2 and C4 were chosen for thecurrent research and were equilibrated withdeionized water. Originally the secondexperiment was to be run with pure PCE.However it was desirable to have some dye inthe solution for visualization purposes. Thesurface chemistry illustrated in Figure 2 andthe entry pressures reported in Table 1 indicatethat the C4 dye concentration solution behavesvirtually identically as pure PCE. The twolong-term experiments were designated "C2170F" and "C4 170H".

The media were flushed with ten pore volumesof pre-equilibrated water following thepermeability determinations for eachexperiment. This water had been pre-equilibrated with the SUDAN IV dye/PCEsolution. The water reservoir burette was


K-4

adjusted so the water pressure was justsufficient to fully saturate the medium. Thusthe top of the bead media was convenientlyselected as the reference height for all fluidhead measurements.

The levels in the water reservoir burette andthe media were allowed to equilibrate for 15minutes. A Schellbach burette, which had beencut off and fused to a glass bulb with an O-ring, was used as the standpipe for the PCEcolumn. The standpipe assembly was insertedto 2mm above the media surface. Finally, acolumn of the PCE-Sudan IV dye solution wasadded to the Schellbach standpipe to a heightof 9 cm above the top of the water saturatedglass bead porous media. This height was wellbelow the expected entry pressure for PCEhaving the 0.508 g/L dye Sudan IV dyeconcentration.

In prior work (Pirkle et al. 1997a). the meanentry pressure for PCE with this dyeconcentration was 15.45 cm of PCE with astandard deviation of 0.66 cm (n=4) (Table 1).This translates to an expected capillary entrypressure of 24,500 baryes (g/cm-sec2). The95% confidence interval for the entry pressureof PCE with this dye concentration in thismedium was thus 13.35 to 17.55 cm of PCE(21,200 to 27,900 baryes). The run times forthe previous experiments were less than eight(8) hours. Thus the surface chemical time-dependence became the primary variablegoverning penetration of the glass beadmedium in the experiments described heresince the target fluid pressure difference wasless than 60% of the expected entry pressureand less than 70% of the lower 95%confidence limit for entry pressure.

Evaporation of the fluids was anticipated to besignificant in these experiments since theywere to be run for many weeks. Thereforeseparate evaporation studies were conductedin order to assess the extent of evaporation.Initial evaporation studies were conducted inordinary Schellbach burettes. Later, the

burettes were cut off and fused at the base toeliminate fluid losses through the stop-cocks.

RESULTS

Evaporation Studies

The results of the evaporation rate studieswere divided into three essentially linearevaporation rate groups for both water andPCE. The groups correspond to the followingevaporation burette conditions: early, unsealedburette; later, unsealed burette; and sealedburette. High rates of evaporative loss wereinterpreted to be the combined result ofevaporation loss upward through the buretteabove the fluid surface and evaporative lossthrough the burette stopcock. The stopcockswere tightened in an effort to reduce thatsource of fluid loss. Stopcocks were later cutoff the Schellback burettes, and the open endswere fused shut. Evaporation studies wereinitiated in the sealed burettes when the losseswere still noted to be high. Thus evaporationwas only possible upward above the top of thefluid surfaces in these sealed burettes, allowingan estimate to be made of the separateevaporative loss components. The loss ratethrough a stopcock was estimated assumingthe early, unsealed rate represents a sum ofstopcock loss plus evaporative loss out the topof the burette. The stopcock loss rates wereestimated to be 0.0057 mL/day for water,0.0561 mL/day for C2-PCE, and 0.0348mL/day for C4-PCE.

Experiment C2 170F

The rates of fluid loss from the water reservoirburette in experiment C2 170F were noted tobe high. Periodic additions of water weremade to the burette to prevent the waterpressure from decreasing to the point wherethe capillary entry pressure for the dyed-PCEsolution would be reached. The water loss wasessentially linear over the time intervalsbetween additions, but the rates of fluid loss


K-5

decreased by an order of magnitude inapproximately exponential fashion over thecourse of the experiment (Table 3). Thecumulative fluid volume loss measured in thewater reservoir burette over the nearly 70 daysof the experiment was 9.31 mL. Thecumulative fluid loss measured in the PCEstandpipe was 0.34 mL.

The capillary pressure history for experimentC2 170F is shown in Figure 3. The capillarypressure followed a saw-tooth pattern withtime since water was periodically added tokeep the reservoir pressure differencerelatively constant. It is important to note thespike in capillary pressure that occurred due tofluid loss from the water reservoir burette overthe first four days of the experiment. Thereservoir pressure difference, at its peak,exceeded the lower 95% confidence limit ofthe maximum stable pressure difference forthis medium and this dye concentration.

Tiny fiber-like structures, presumablyconsisting of water sheathed with a PCE-filmcontaining a significant surface excess of thedye, were noted emanating from the top of thewater saturated porous medium on the secondday of the experiment. PCE was visuallyobserved to be entering the top of the mediumon day four with "fingers" extending to depthsof up to 1 mm. At the same time, the tinyfibrous structures appeared to be continuing to"grow" out of the medium. Examination of thefibrous structures with a magnifying glassrevealed an appearance resembling a "string ofpearls" with small bubbles of water connectedvia the fibers. The PCE fingers extended asmuch as 1.8 to 2 mm below the porousmedium surface by day five. In addition, manyof the tiny water balloons had begun toagglomerate into larger drops having anexterior texture resembling a tight cluster ofgrapes. PCE continued to accumulate in thetop 2 mm of the porous medium leading to anoticeable darkening of that portion of themedium by day 30. There was never anyvisual indication that PCE had penetrated to a

depth greater then the top 2 mm. While waterwas clearly escaping from the top of themedium, there was never any significantaccumulation of water in the PCE standpipe.

Figures 4 and 5 are photographs showing earlyand ending conditions in experiment C2 170F.The photo in Figure 4 was taken approximately1.7 hours after starting to add the dyed-PCEsolution to the standpipe in experiment C2170F. Figure 5 illustrates the penetration ofthe dye to a maximum depth of 2mm after 69days.

The glass beads in the top 2 mm of the columnwere found to be stained red by Sudan IV dyeupon disassembling the column. Excavation ofthe medium confirmed a) that penetration hadoccurred relatively uniformly over the top ofthe medium, and b) that penetration had notoccurred to any detectable degree beyondapproximately 2 mm below the top of themedium.

Experiment C4 170H - ControlExperiment

As in experiment C2 170F, capillary pressurefollowed a saw-tooth pattern with time (Figure6). The early rate of fluid loss measured inexperiment C4 170H was high (0.217 mL/day),significantly exceeding the measuredevaporation loss rate. This high rate lastedapproximately 7 days after which it declined toa rate of 0.068 mL/day, approximating the lossrate in an unsealed burette (i.e., including astopcock). This second rate declined furtherbeginning on day 18 when it decreased to0.016 mL/day. The spike in fluid reservoirpressure difference occurred when a "major"leak occurred in the apparatus. This spike(approximately 25,500 baryes), however,remained well below the expected entrypressure for the medium of 27.71 cm of PCE(44,050 baryes) (see Table 1) and the lower95% confidence limit for the entry pressure(21.64 cm of PCE or 34,400 baryes).


K-6

There was no indication of PCE penetrationinto the glass bead medium throughout thecourse of experiment C4 170H (43 days). Thecontact between the water-saturated glassbeads and the free dyed-PCE phase remainedvisually sharp throughout the course of theexperiment. Figure 7 is a photo of the PCE-glass bead contact at the start of experimentC4 170H (3.5 hours after starting to add thedyed-PCE solution to the standpipe), andFigure 8 shows the contact at the end of theexperiment (43 days). Excavation of the glassbead medium indicated there was no stainingof the beads nor any penetration in the interiorportion of the column away from the glasswalls.

DISCUSSION

PCE Evaporation

The order of magnitude difference betweenC2 and C4 PCE solution evaporation rates isnotable. It suggests that minor (<<1%),nonvolatile components dissolved in volatilesolvents can greatly reduce solvent volatility.The dye concentration in C2 was about 300ppm by weight; in C4 it was about 3 ppm. Amajor implication is that one of the importantmechanisms for solvent depletion in the vadosezone, volatilization, may be significantlycurtailed by comparable minor components in afield DNAPL. The concentration of nonvolatilecomponents would increase as evaporativeremediation progresses. This process couldplace a limit on the fraction of solventcontamination in the vadose zone that can berecovered by vapor extraction. At the least, itcould be expected to significantly reduce therate of solvent recovery. More detailed studyis needed to confirm these results.

Time-Dependent Penetration of aCapillary Barrier

The results of experiment C2 170F areinteresting, but do not as yet provide solidexperimental support for the hypothesis that

the time-dependent chemistry can govern thepenetration of a porous medium capillarybarrier. The spike in capillary pressure early inthe experiment exceeds the lower 95%confidence limit for entry pressure for thismedium and this dyed-PCE solution. It onlyexceeded that limit for less than a day,however, and likely for only a matter of hours.The correspondence of PCE penetration of themedium beginning on day 4 raises questionsregarding whether the entry pressure for thismedium, as defined by previous work (Pirkle etal. 1997a) was exceeded. On the other hand,the slow and limited penetration of the mediumis consistent with the time-dependent pore-scale behavior modeled by Tuck (accepted). Inaddition, Pirkle et al. (1997a) found that if theentry pressure had been exceeded, the PCEpenetrated the whole length of the medium andeven drained into the open space below thescreen any time an experiment was leftovernight. The broad, relatively uniformpenetration over the whole upper portion of themedium is also consistent with Tuck's modelingresults for capillary pressures significantlybelow the entry pressure for pure PCE (i.e.,zero dye content). An experimental apparatusspecifically designed for maintaining constantcapillary pressure during longer-termexperiments is needed to provide a better testof Tuck's model of pore-scale behaviorincorporating time-dependent surfacechemistry.

There is additional anecdotal evidencesupporting the role of the time-dependentsurface chemistry in the behavior observed inthe experiment, including the fiber-likestructures emanating from the medium and thestaining of the glass beads. Formation of thefiber structures was not observed until daytwo; at least 10 hours passed before they wereobserved. The best explanation for thesestructures is that partitioning of the dye into thewater-PCE interface lowers interfacial tensionsufficiently to allow tubes of water to formwhich rise due to the unstable densityconfiguration. Adsorption of the dye onto the


K-7

glass bead surfaces was indicated by the factof their having become stained. Dye adsorptionexplains the change in PCE-glass-watercontact angles observed by Tuck and Rulison(1998) and illustrated in Figure 2.

Incorporating the role of time-dependentsurface chemistry into our understanding ofmultiphase flow in porous media is important,presuming it is a real phenomenon and behavescomparably to the pore-scale model developedby Tuck. In the case of NAPL contamination,any separate phase penetration of finer-grained layers will be significantly harder toremediate. Relatively little penetration, such asobserved in experiment C2 170F, can result ina significant amount of NAPL penetration intofiner layers. DNAPL contamination at theSavannah River Site A/M-Area has beenfound to be distributed in multiple, relativelythin zones in the M-Area aquifer (Tuck et al.1998a). The magnitude of this issue can beestimated by making a few simpleassumptions. A total bulk volume of 40,000 Lof DNAPL-containing, finer-grained layersresults if we assume ten such zones each of1000 m2 in area and bounded above and belowby finer-grained layers each of which has beenpenetrated to a depth of 2 mm by DNAPL.The total pore space available to contain fluidsis 1,520 L assuming an average porosity of0.38 (Eddy-Dilek, et al. 1994). If we furtherassume a range of DNAPL saturationbetween 0.10 and 0.50, the total volume ofDNAPL caught in those bounding zones due tothe time-dependent penetration would be 150to 760 L. This volume of DNAPL would beexpected to be more difficult to remediate.

Another important issue arises from how wemodel and understand DNAPL flow anddistribution following its initial release into thesubsurface. The result of time-dependentsurface chemistry during flow of a NAPL is toincrease the effective interfacial tensiongoverning that flow (Tuck et al. 1998b). Theresult is that the faster a NAPL-front isadvancing, the higher will be the average

interfacial tension governing the capillaryforces acting to resist the NAPL movement.This arises from the lower average time thatNAPL-water interfaces spend in pore-throats.The influence of average interface age on theeffective interfacial tension can be seen inFigure 2b. The net result in terms of modelingmultiphase flow is that there will be anadditional iteration loop required sinceinterfacial tension plays an important role in thecapillary pressure-saturation relationship,which in turn will effect the average rate ofadvance of NAPL-water interfaces throughthe medium.

Further research is needed to confirm ornegate the validity of incorporating time-dependent surface chemistry in pore-scalemodels of multiphase flow as developed byTuck (accepted) given its potential importancefor better understanding of NAPL behavior inporous media. Additional research shouldinclude experimental testing under moreclosely maintained capillary pressure conditionsand the development of a more comprehensivepore-scale model. The means for scaling-upthe time-dependent pore-scale behavior intocontinuum scale models is also an importantarea requiring research and development.

ACKNOWLEGEMENTS

We are grateful to the Department of Energyat the Savannah River Site for providingfunding for the experimental work reportedhere. The research was performed at USCAiken under South Carolina UniversitiesResearch and Education Foundation contractAA00900T. The authors also wish to thankWestinghouse Savannah River Company andits management and colleagues for support andencouragement. In particular, we wish tothank Dr. Brian B. Looney, Dr. RichardStrom, Joette Sonnenberg, and Deb Moore-Shedrow. The authors also wish to expresstheir appreciation to the USC Aikenadministration, especially Dr. Harold Ornes,former Chair, and Dr. Allen Dennis, current


K-8

Chair, of USC Aiken's Department of Biologyand Geology for their support andencouragement of the project.

REFERENCES

Eddy-Dilek, C.A., T. R. Jarosch, M. A. Keenan, W.H. Parker, S. P. Poppy, J. S. Simmons. 1994.Characterization of the Geology and ContaminantDistribution at the Six Phase HeatingDemonstration Site at the Savannah River Site (U).Westinghouse Savannah River Company, Aiken,SC, WSRC-TR-93-678. .

Graton, L. C. and H. J. Fraser. 1935. “SystematicPacking of Spheres with Particular Relation toPorosity and Permeability,” J. Geology, v.43, p.785-909.

Hobson, G. D. 1954. Some Fundamentals ofPetroleum Geology, Oxford University Press, NY,NY.

Hobson, G. D. and E. N. Tiratsoo. 1975.Introduction to Petroleum Geology, Scientific Press,Ltd., Beaconsfield, England.

Pirkle W. A., G. M. Iversen, and D. M. Tuck. 1996.DNAPL Entry Pressure Into Water SaturatedPorous Media, Final Report Submitted inCompletion of SCUREF Task 185. University ofSouth Carolina Aiken.

Pirkle W. A., G. M. Iversen, and D. M. Tuck. 1997a.The Effect of an Organic Dye on DNAPL EntryPressures into Water Saturated Porous Media, FinalReport Submitted in Completion of SCUREF Task185, Phase II, Vols. I&II. University of SouthCarolina Aiken.

Pirkle W. A., G. M. Iversen, and D. M. Tuck. 1997b.DNAPL penetration of capillary barriers due totime-dependent wetting relationships. University ofSouth Carolina-Aiken. Final Report Submitted inCompletion of SCUREF Task 232. University ofSouth Carolina Aiken.

Tuck, D. M., and C. Rulison. 1998. Interfacial effectsof a hydrophobic dye in the tetrachloroethylene-water-glass system (U). Westinghouse SavannahRiver Company, WSRC-TR-97-0038. Aiken, SC, 23p.

Tuck, D. M., G. G. Exner, B. B. Looney, B. Riha, J.Rossabi. 1998a. Phase I Field Test Results of anInnovative DNAPL Remediation Technology: TheHydrophobic Lance(U). Savannah RiverTechnology Center, WSRC-TR-98-00308. Aiken,SC, 24 pp .

Tuck, D. M., G. M. Iversen, W. A. Pirkle, C.Rulison. 1998b. Time-dependent interfacial propertyeffects on DNAPL flow and distribution. In:Wickramanayake, G.B., Hinchee, R.E. (Eds.),Nonaqueous-Phase Liquids: Remediation ofChlorinated and Recalcitrant Compounds. BattellePress, Columbus, OH, 73-78.

Tuck, D. M. 1999. Time for a new dimension in Pc-Sspace. in: Seventh Annual Clemson UniversityHydrogeological Symposium, April 9, 1999,Clemson, SC, Clemson University, p. 13.

Tuck, D. M. accepted. Primary drainage of NAPLgoverned by time-dependent interfacial properties.Chemical Geology.


K-9

Table 1. Sudan IV dye concentrations in PCE used in this and previous research and resulting entry pressuresfor dyed PCE in glass-bead media of the same type used here (Tuck and Rulison year, Pirkle, et al. 1997a).________________________________________________________________________________________

Entry Pressures from Prior Research in glass-bead media

Solution Sudan IV dye Number of Mean Entry Pressure Standard Deviation Mean EntryPressureLabel Conc. (g/L) Experiments (cm of PCE) (cm of PCE) (baryes, g/cm-sec2)_________________________________________________________________________________________

C0 0.00000 12 27.14 2.48 43,150

C4 0.00508 7 27.71 2.48 44,050

C2 0.508 4 15.45 0.66 24,560_________________________________________________________________________________________

Table 2. Rates of evaporative loss for the different fluids. The difference between sealed burette and unsealedburette loss rates provides an estimate of evaporative loss through a stopcock.

_________________________________________________________Water C2 PCE C4 PCE

_________________________________________________________Early, Unsealed Burette rate (mL/day) 0.0121 0.0598 r2 0.9215 0.9925 n 17 36Later, Unsealed Burette rate (mL/day) 0.0086 0.0554 0.0676 r2 0.9919 0.9929 0.9818 n 29 14 30Sealed Burette rate (mL/day) 0.0064 0.0037 0.0328 r2 0.9933 0.9846 0.9939 n 40 33 32_________________________________________________________

Table 3. Fluid loss rates from the water reservoir burette in experiment C2 170F______________________________________________________________________Time Interval Loss Rate Number of AverageMidpiont (days) (mL/day) R^2 Data Points, n Temperature (C)______________________________________________________________________

1.97 0.6114 0.9931 25 23.16 4.41 0.8739 0.9951 5 22.93 5.91 0.3753 0.9985 7 23.44 8.43 0.2748 0.9854 7 23.1711.40 0.1947 0.9994 6 22.7214.58 0.1777 0.9881 5 22.2021.66 0.1112 0.9933 14 22.8435.58 0.0693 0.9966 18 22.73


K-10

51.23 0.0516 0.9946 17 22.4664.27 0.0460 0.9842 11 22.60______________________________________________________________________


K-11

Water SaturatedPorous Medium

WaterReservoirBurette

PermeameterDrainline

PermeameterDrainline Valve

Water ReservoirValve

d

D

L

Stand-pipe

Water ReservoirLevel Same as Topof Porous Medium

Three-WayStop-cock

High Vacuum Valves

CapillaryPressure,Pc = ρPCEgLv

Lv

Figure 1. Experimental set-up used for PCE entry pressure in the water-glass-PCE-Sudan IV dye system.

4321020

30

40

50

60

log(Interface Age[sec])

Con

tact

Ang

le (d

egre

es)

a.

Undyed PCE0.005 g/L

0.508 g/L1.27 g/L

1.69 g/L4.0 g/L

DNAPL-DIWDNAPL-GW

Legend: System Descriptions

100001000100101.120

30

40

50

Average Interface Age (sec)

Inte

rfac

ial T

ensi

on (m

N/m

)

b.

Figure 2. Time-dependence of interfacial tension and contact angle in the water-glass-PCE-Sudan IV dyesystem.


K-12

Figure 3. Capillary pressure history for experiment C2-170F. Sudan IV dye concentration in PCE was 0.5 g/L.

Capillary Pressure - Exp. C2 170F

1300014000150001600017000180001900020000210002200023000

0 10 20 30 40 50 60 70

Time (days)

Cap

illar

y Pr

essu

re (b

arye

s, g

/cm

-se

c^2)

Capillary Pressure


K-13

Figure 4. Beginning of experiment C2 170F. Photo was taken approximately 1.7 hours (0.069 days) after thefirst addition of PCE to the PCE-Standpipe.


K-14

Figure 5. End of experiment C2 170F. Photo was taken on the 69th day from the start of the experiment.


K-15

Figure 6. Capillary pressure history for experiment C4 170H. Sudan IV dye concentration was 0.005 g/L, toolow to significantly change the entry pressure for PCE.

Fluid Reservoir Pressure Difference, Pc - Exp. C4 170H

11000

16000

21000

26000

0 5 10 15 20 25 30 35 40 45

Time (days)

Pre

ssur

e D

iffe

renc

e (b

arye

s, g

/cm

-se

c^2)

Series1


K-16

Figure 7. Beginning of experiment C4 170H. Photo was taken approximately 3.5 hours (0.15 days) after thefirst addition of PCE to the PCE-Standpipe.


K-17

Figure 8. End of experiment C4 170H. Photo was taken on the 43rd day from the start of the experiment.


L-1

Subsurface Correlation Of Cenozoic Strata In The Updip Coastal Plain,Savannah River Site (SRS), South Carolina

R. K. Aadland, Westinghouse Savannah River Co. Aiken, S.C. 29808P. A. Thayer, University of North Carolina at Wilmington, Wilmington N.C. 28403-3297

ABSTRACT

Difficulties with correlation and the paleo-geographic reconstruction of fluvial andmarginal-marine terrigenous strata near theFall Line result from: 1) lithologic similaritythroughout the section, typically coarse,poorly sorted unconsolidated sand and mud; 2)few marker beds; 3) paucity of age-diagnosticfossils due to original depositional environ-ment and post-depositional leaching; 4) facieschanges; 5) hard to recognize unconformities;and 6) structural complexity. Down-holegeophysical logs and core descriptions from245 soil borings and monitoring wells wereused to delineate the stratigraphy and structureof the Cenozoic section in the 15 square mileA/M Area at SRS. The density of subsurfacedata made possible the detailed correlation ofstratigraphic units and the understanding offacies relationships. In addition, the impor-tance of faulting, which has largely goneunrecognized in previous studies in theAtlantic Coastal Plain, has been demonstrated.The juxtaposition of differing lithologies incoastal plain strata has commonly beeninterpreted to result from facies change. Thisstudy however, indicates that fault displace-ment of stratigraphic units often accounts forthe perceived ”facies changes.” In order todevelop correct depositional models incorpo-rating the relative contribution of eustacy,sediment supply and tectonics, the density ofsubsurface data must be sufficient to distin-guish stratigraphic variations resulting fromsedimentological factors versus post-depositional structural events.

1Information developed under Contract DE-AC09-89SR19015 with the U.S. Dept. of Energy

INTRODUCTION

The Savannah River Site (SRS) is a 300-

square-mile facility located near the Fall Linein the Upper Atlantic Coastal Plain of SouthCarolina (Figure 1). The A/M Area, whichcovers approximately 15 square miles, issituated in the northwestern part of the Site(Figure 1). The sedimentary section underly-ing A/M Area is 700 to 800 feet thick andconsists of coarse to fine, often gravely sandsand interbedded clays of Late Cretaceous toHolocene age.

The purpose of this study is to describe andevaluate the stratigraphy and structure ofunconsolidated coastal plain strata in the A/MArea. The study will aid in delineatinghydraulic characteristics of the sediments formodeling groundwater movement and con-taminant transport.

METHODS

Detailed analysis of down-hole geophysicaldata and core from more than 245 monitoringwells and soil borings were utilized in thestudy. Down-hole geophysical logs, includinggamma ray, gamma ray density, neutrondensity, resistivity surveys, and detaileddescriptions of drill core were used to deline-ate lithologic types and facies relationships.

Eight cross-sections were prepared and used tocorrelate lithologic units. Structure contourand isopachous maps were constructed toshow the overall geometry, thickness, andcontinuity of the stratigraphic units in thestudy area.

STRATIGRAPHY

The SRS is underlain by Paleozoic igneousand metamorphic rocks, Mesozoic terrigenousclastics of the Dunbarton rift basin (Marineand Siple, 1974), and Late Cretaceous andTertiary sediments of the Atlantic CoastalPlain.


L-2

Paleozoic basement and Triassic/Jurassic (?)sedimentary rocks of the Dunbarton Basinhave been beveled by erosion and are uncon-formably overlain by unconsolidated to poorlyconsolidated Late Cretaceous and Tertiarycoastal plain sediments (Cooke, 1936; Siple,1967; Colquhoun and Johnson, 1968;).Coastal plain strata form a mostly clasticwedge that thickens and dips to the southeast(Colquhoun and Johnson, 1968). The sedi-mentary wedge is about 700 feet thick in theA/M Area, and accumulated in environmentsranging from fluvial to shallow marine shelf.Fluctuating conditions of deposition accountfor the complex variations in sedimentlithology throughout the study area.

Paleontological data is sparse in coastal plainstrata of A/M Area. Further south, in thesouthern and central parts of SRS, coastalplain sediments were deposited in lower deltaplain and shallow marine environments. It ishere that age-diagnostic fossils are moreplentiful; hence, age and stratigraphic relation-ships can be resolved with greater certainty.The stratigraphy of coastal plain sediments inA/M Area was determined by correlation totype wells located less than 10 miles to thesoutheast, which were established by Fallawand Price (1995).

Cretaceous Strata

In A/M Area Upper Cretaceous strata overliePaleozoic basement. The Upper Cretaceoussequence includes the basal Cape FearFormation and overlying Lumbee Group. TheLumbee Group includes from the base upward(Faye and Prowell, 1982), Middendorf, BlackCreek, and the Steel Creek Formations (Figure2). The Cretaceous sequence is about 400 feetthick in A/M Area, and consists of quartzsand, pebbly sand and sandy clay, whichaccumulated in lower to upper delta plainenvironments.

Tertiary Strata

Tertiary sediments are approximately 300 feetthick in A/M Area and range from Paleoceneto Miocene/Oligocene (?) in age (Figure 2).

They are interpreted to have been deposited influvial, marginal marine, and shallow marineshelf environments. Tertiary strata consist ofmoderately to very poorly sorted quartz sandand pebbly quartz sand with varying amountsof mud matrix. Interbedded mud units consistmainly of quartz silt and kaolinitic clay andare commonly sandy.

Four regionally significant unconformities arerecognized within the Tertiary section of A/MArea. From the base upward, they include theCretaceous-Tertiary unconformity, LangSyne/Sawdust Landing unconformity, Santeeunconformity, and ”Upland” unconformity(Figures 2, 3 and 4). Four sequence strati-graphic units, here labeled Sequence I, II, IIIand IV, which are bounded by the unconfor-mities, are delineated (Figure 2). Work iscurrently underway to place these sequencesinto a global sequence stratigraphic frame-work.

STRUCTURE

Post Sequence I Faulting

Figure 4 is a cross-section showing faulting inA/M Area and Figures 5, 6 and 7 illustratefault distribution in plan view. The oldestpost-Sequence I oblique-slip reverse faults arelabeled alphabetically, and the youngest post-Sequence II and III/IV series of normal faultsare labeled numerically.

Several episodes of oblique-slip reversefaulting or renewed faulting occurred in A/MArea following the Triassic/Jurassic riftingevent. The first episode began at the close ofCape Fear time and continued intermittentlythrough Paleocene time (Aadland, 1997). Thefinal episode of oblique-slip reverse offsetinvolved the upper Paleocene Lang SyneFormation, and was followed by erosion andtruncation of the unit at the LangSyne/Sawdust Landing unconformity. Theunconformity is the defining event in A/MArea, as it is regionally, separating PaleoceneSequence I strata and associated tectonicevents from Sequence II and youngerstrata/tectonic events.


L-3

The pre-Sequence II oblique-slip reverse faults(Figures 4, 5 and 6) form a series of closely-spaced (< 2000 ft), high-angle reverse faults,any one of which can be traced for approxi-mately 1 to 4 miles. The faults overlapresulting in a fault zone where several sub-parallel faults are noted. The faults form aseries of horsts and grabens, some of whichtruncate at fault intersections. The strike ofindividual fault segments varies from N 17o Eto N 42o E, averaging about N 30o E. Apparentvertical offset on the faults varies from 15 to60 feet (Figures 3, 4, 5 and 6), with throws of30 to 40 feet most common in the Cretaceous-Paleocene section.

Post Sequence II and III/IV Faulting

Faulting that offsets post-Sequence II andIII/IV strata form an en echelon series ofnormal faults that strike from N 21o W to N 6o

E, averaging N 11o W (Figures 6 and 7). Theen echelon normal faults strike at an acuteangle (about 41o ) to the average orientation ofthe oblique-slip reverse faulting that offsetsthe underlying late Cretaceous-Paleocenesediments of Sequence I (Figure 6). Theapparent throw on the normal faults usuallyvaries from 10 to 20 feet (Figure 4).

DISCUSSION

Active tension in a direction perpendicular tothe strike of faults in the post-Sequence II andIII/IV section would account for the observednormal faulting. However, regional tensionacting in this direction cannot account for theen echelon fault pattern. Left-lateral shearalong one of the pre-existing post-Sequence Ioblique-slip reverse faults beneath the passiveSequence II and III/IV strata (Figure 8), wherethe northwestern block moves to the southwestrelative to a southeastern block, could accountfor the en echelon pattern of normal faultsobserved in the Sequence II and III/IV section.Alternately, a couple (Figure 9), caused by anortherly block moving towards the southwestrelative to a southerly block moving towardsthe northeast, with a block wedged betweenthe two moving blocks, could produce the typeof en echelon normal faults observed in the

overlying Sequence II and III/IV passivesediments.

SUMMARY AND CONCLUSIONS

1. The density of subsurface data has madepossible the detailed correlation of strati-graphic units and understanding of faciesrelationships in the A/M Area.

2. The importance of faulting, which has gonelargely unrecognized in Atlantic Coastal Plainstudies, is demonstrated.

3. The juxtaposition of differing lithologies incoastal plain strata has commonly beeninterpreted to be the result of facies change.This study however, indicates that faultdisplacement of stratigraphic units oftenaccounts for the perceived ”facies changes.”

4. The relative contribution of tectonics,versus eustacy and sediment supply, to thedepositional model developed in the updipcoastal plain setting of A/M Area is critical toremediation efforts of contaminants insubsurface aquifers. The density of subsurfacedata must be sufficient to distinguish strati-graphic variations resulting from sedimen-tological factors versus post-depositionalstructural events in determining the hydrauliccharacteristics of sediments for modelinggroundwater movement and contaminanttransport.

REFERENCES

Aadland, R. K., 1997, Tertiary geology of A/MArea, Savannah River Site, S.C. (U): U.S. Depart-ment of Energy, WSRC-RP-0505, 94 p. + 17plates.

Billings, M.P., 1972. Structural Geology. 3d ed.Englewood Cliffs, New Jersey: Prentice-Hall, Inc.

Cooke, C. W., 1936, Geology of the Coastal Plainof South Carolina: U.S. Geological Survey Bulletin867, 196 p.

Colquhoun, D. J., and Johnson, H. S., Jr., 1968,Tertiary sea-level fluctuation in South Carolina:


L-4

Palaeogeography, Palaeoclimatology, Palaeoecol-ogy, v. 5, p. 105-126.

Fallaw, W. C., and Price, V., 1995, Stratigraphy ofthe Savannah River Site and vicinity: SoutheasternGeology, v. 35, p. 21-58.

Faye, R. E., and Prowell, D. C., 1982, Effects ofLate Cretaceous and Cenozoic faulting on thegeology and hydrology of the coastal plain near theSavannah River, Georgia, and South Carolina: U.S.Geological Survey Open-File Report 82-156, 73 p.

Marine, I. W., and Siple, G. E., 1974, BuriedTriassic basin in the Central Savannah River Area,South Carolina and Georgia: Geological Society ofAmerica Bulletin, v. 85, p. 311-320.

Pascal, R. and R.W. Krantz, 1991. “Experimentson Fault Reactivation in Strike-Slip Mode,”Tectonophysics. 188:p. 117-131.

Siple, G. E., 1967, Geology and groundwater of theSavannah River Plant and vicinity, South Carolina:U.S. Geological Survey Water-Supply Paper 1841,113 p.


L-5

Figure 1. Location of the Savannah River Site and the A/M Area.

Savannah River

5 MILES

Aiken C

o.

Barnw

ell C

o.

Allendale Co.

A/MArea

SCGA

N

LowerAtla

nticCoas

tal P

lain

UpperAtla

nticCoas

tal P

lain

Atlantic

Ocean

Augusta

Aiken

Line

FallColumbia

Savannah

River

SavannahRiver Site

GASC


L-6

Figure 2. Stratigraphic units in A/M area at Well MSB-69. The location ofMSB-69 is shown in Figure 5.

G.R.ELEV. 382'

ABOVE MSLLITHOSTRATIGRAPHYAGE

UPLAND UNCONF.

“UPLAND” UNITMIOCENE /OLIGOCENE?

LATEEOCENE

MIDDLEEOCENE

EARLYEOCENE

PALEOCENE

LATECRETACEOUS

DRY BRANCHFORMATION

SANTEEFORMATION

WARLEY HILLFORMATION

CONGAREE &FOURMILE BRANCH

FORMATIONS

LANG SYNE &SAWDUST LANDING

FORMATIONS

STEEL CREEKFORMATION

(LUMBEE GROUP)

“C” SAND HORIZON

SEQ

UEN

CE

IIIS

EQU

ENC

E I

LANG SYNE / SAWDUSTLANDING UNCONF.

K-T UNCONF.

SEQ

. IV

TOBACCO ROADFORMATION

SANTEE UNCONF.S

EQU

ENC

E II

RES 16N 64N

10

50

100

150

200

250

300

350

?

MUDSAND & MUDDY SANDNO RECOVERY


L-7

Figure 3. Stratigraphic Units in Wells MCB-CH2 and MCB-CH3. Missing Section in BothWells is due to Greater Erosional Truncation on the Upthrown Fault Blocks.

5 x 103

100

150

200

100

250

300

50

00 200100

0 200100

102 103

GR (API)GR (API) RESISTIVITY

RFOC

ILD

MCB-CH2MCB-CH3

ILM

“Upland”Unit

Tobacco RoadFm.

Dry BranchFm.

SEQ

UEN

CE

IVSE

QU

ENC

E III

SEQ

UEN

CE

IISE

QU

ENC

E I

SanteeFm.

Steel CreekFm.

Lang Syne /Sawdust Ldg.

Fm.

Congaree /Fourmile Br.

Fm.

Warley Hill Fm.

“C” Sand Top

K

T

Carolina G


ield Trip G

uidebookW

SRC

-MS-2000-00606

L-8

Figure 4. West-east lithostratographic cross-section A/M Area, Savannah River Site.


L-9

Figure 5. Middle Eocene time “C” Sand structure map. Note location of wells MSB-69,MCB-CH2 and MCB-CH3 (see Figs. 2 and 3). “C” Sand mapped from the LangSyne/Sawdust Landing unconformity surface datum. The unconformity surface wasarbitrarily assigned an elevation of 200 feet from which the “C” Sand surface was mapped.The offset noted on faults A through L predate Congaree/Fourmile deposition.


L-10

Figure 6. “C” Sand horizon structure map. Stratigraphic position of “C” Sand shown inFigures 2 and 3. “C” Sand horizon mapped from mean sea level. Contours in feet abovemean sea level.


L-11

Figure 7. Top of Warley Hill structure Map Contours in feet above mean sea level.


L-12

Figure 8. En echelon pattern of normal faults in the sediments formed above a 90°°°° base-ment fault reactivated in a pure strike-slip mode. (Modified from Billings, 1972 and Pascaland Kranz, 1991).


L-13

Figure 9. En echelon pattern of normal faults in the sediments formed above a 90°°°° base-ment fault couple reactivated in a pure strike-slip mode, where a northerly block movedtoward the west relative to a southerly block, resulting in the en echelon normal faulting inthe passive sediments of the intervening block. (Modified from Billings, 1972 and Pascaland Krantz, 1991).


L-14

Notes


M-1

Geological Interpretation of the Pre-Eocene Structure andLithostratigraphy of the A/M Area, Savannah River Site, South Carolina

D. E. Wyatt, R. K. Aadland and R. J. Cumbest, WSRC, Aiken, SC 29808

ABSTRACT

The geological interpretation of the structureand stratigraphy of the A/M Area wasundertaken in order to evaluate the effects ofdeeper Cretaceous aged geological strata andstructure on shallower Tertiary horizons. Theshallow horizons are often involved inenvironmental issues. Additionally, this studyattempted to evaluate the nature, extent,affects and timing of regional faults that haveaffected the area.

This interpretation was completed bycombining 87 existing cone penetrometertests, 10 new cone penetrometer tests, 234existing boring and geophysical log locationsand 21 new geophysical log locations with 22km of new high resolution seismic data, 27 kmof existing data and a 275 m by 350 m 3-Dseismic grid. Horizon information and faultintervals were interpreted from the seismicdata and combined into a database withstratigraphic picks from the geophysical logsand cone penetrometers. Regional deep boringcorrelations were made using the geophysicallogs to constrain the seismic horizonidentification. Seismic horizon to stratigraphicinterval correlations were made from sonicinformation derived from geophysical logging.

This study interpreted four major, crystallinebasement involved faults. The apparent oldestregional fault trends NNE and is identified asthe Crackerneck Fault. Three parallel(younger?) faults trend north. Thewesternmost fault is identified as theMWESTA Fault and the easternmost fault asthe Steed Pond Fault. The central faultintersects the Crackerneck Fault and isidentified as the Lost Lake Fault. These faultsare not thought to be single breaks in all areasbut linear zones of disruption. These faults are

thought to be predominantly reverse faultswith possible dip-slip, or minor strike slipmotion.

At least three episodes (and possibly more) offaulting are observed in the data, 1) Mesozoic,probably Jurassic to Triassic in age, faultingin the crystalline basement rocks, 2) post-Cretaceous aged faulting that disrupted theentire Cretaceous sediment column and 3)Paleocene or later faulting that disruptedyounger sediments. These younger structurestrend to the NW and occur in an en-echelonpattern. This pattern may affect strata to nearsurface or surface horizons.

The stratigraphic information for the deeperhorizons is limited due to the limited numberof deep geophysical logs tied to the seismicinformation. However, a comparison ofregional deep logs and seismic data confirmthe near shore, fluvial to deltaic, transitionalto near shore marine, depositional character ofthe deeper A/M Area sediments.

INTRODUCTION

The A/M Combined Geological andGeophysical Program was organized toprovide a comprehensive correlationcapability between existing geologicalinformation, newly acquired geological core,advanced borehole geophysical data, surfacehigh resolution reflection seismic informationand ground penetrating radar data within thegreater A and M Areas of the Savannah RiverSite. Specifically, the locations for all borings,seismic, cone penetrometer and groundpenetrating radar data were chosen tomaximize correlation to existing or newlyacquired data and to compliment existing orongoing environmental programs. This reportpresents an integration of new and advanced


M-2

data with existing information combined intoa comprehensive picture of the deepsubsurface geology of the A/M Area. The areaof investigation is defined on Figure 1. Thelocation of the seismic lines and deep boringsis shown on Figure 2.

The A/M Advanced Geological Study was ajoint effort supported by several organizationsand agencies. Funding for the program wasprovided through the WSRC EnvironmentalRestoration Department and administered byWSRC Site Geotechnical Services. Drillingand field oversight was provided by the USGeological Survey through an inter-agencyagreement administered through DOE-SRSEnvironmental Division. All field work wascoordinated through WSRC Site GeotechnicalServices. Additional subcontract services wereprovided as follows: drilling by EM&TC, Inc.,core descriptions by SAIC, field support andcoordination by Bechtel Savannah River, Inc.,and Raytheon. Core geochemical samplingwas provided by Microseeps, Inc. Core fromthe upper 150 feet of each boring was splitwith half sent to Rutgers University fordetailed soils and transport analysis as part ofthe DOE-HQ FERC program. Boreholegeophysical data were acquired bySchlumberger Wireline Services, Inc, andWestern-Atlas Wireline, Inc. Additional welllogging was provided by Graves LoggingServices. Downhole acoustic suspensionlogging was acquired in GCB-1 and GCB-2 byAgbabian and Associates. Surface seismicinformation was obtained and processed bythe Earth Sciences and Resources Institute ofthe University of South Carolina.Palynological data and additional coreanalysis were provided by the Department ofGeological Science of Clemson University.Ground Penetrating Radar data for the GCBprogram were acquired and processed byMicroseeps, Inc. All data analysis,interpretation and correlation are provided byWSRC, Site Geotechnical Services and theauthors. This report utilizes seismic and deep boreholegeophysical data to characterize the pre-

Tertiary strata. Tertiary maps are made as acomparison to Aadland’s (1997) work and asa tie to his interpretations. Aadland’s (1997)work was completed using a manualinterpretation and mapping approach. APPROACH TO INTERPRETATION The general process for geological correlationwithin the A/M Area follows a basement tosurface “bottoms up” depositional andlithostratigraphic approach. For deeper strata,the interpretation of geophysical log well tiesare made to seismic reflection events and areused to establish base correlation horizons.For shallower strata, older geophysical logsand core are correlated with advancedgeophysical logs and core from the GCB-1,2,3borings to establish type lithostratigraphiclocations. Activities defining the generalapproach are described in order: • Install borings, obtain and analyze core,

and geophysically log GCB-1, GCB-2 andGCB-3. Obtain spectral gamma logs inadditional A/M deep cased wells P8R,P6R, P9R and from deep boring MMP-4SB.

• Acquire and process A/M seismic data. • Correlate GCB’s to existing basement

boreholes using geophysical logs, coredata, field description of core, Type Wellpicks and Aadland (1997) picks.

• Establish synthetic seismic data from

GCB wells correlated to lithologic picksfrom core and geophysical log horizons.

• Correlate seismic horizons to GCB

lithologic picks. Establish fault andstructure trends through the area based onseismic offsets in the basement. Establishprincipal lithostratigraphic horizons onseismic and geophysical logs.


M-3

• Correlate existing A/M well data picks tolithostratigraphic picks in the GCBborings.

• Incorporate CPT data into well grid.

Correlate shallowest interpretable horizonto well picks.

• Incorporate geophysical data from new

shallow environmental borings to fill datagaps

• Generate combined correlation maps

based on interpretable seismic horizonsand geophysical logs. Establish basementstructural trends based on seismic offsets.

• Combine maps with structure into a three

dimensional A/M data volume. • Develop and present cross-sections

through the data volume. • Interpret data.

SOURCES OF INFORMATION

Nine high resolution reflection seismic lineswere obtained to support deep mapping. Thelocation of these lines are shown on Figure 2.In addition to these lines, three existingregional seismic lines were also utilized.These location of these lines are also shownon Figure 2.

Field data were acquired using an OYO DAS-1 signal enhancement seismograph operatingin 96 channel mode. 40 Hz OYO geophoneswere used in geophone strings having 3phones per string. The energy sourceconsisted of an IVI MiniVIBE® remotelytriggered vibrator sweeping from 50 to 200Hz. In general, an asymmetric split spreadacquisition geometry (24-gap-72 spread) wasutilized, with shot and geophone spacings of1.5 and 3.1 meters depending on the seismicline.

The following processing sequence wasapplied to the data: reformat SEG-2 field filesto SEG-Y format, apply vibroseis whitening,F-K filter, predictive deconvolution, bandpassfilter, trace edit, crooked line sort, statics, bulkvelocity shift, velocity analysis, residualstatics, apply normal moveout, stack,bandpass filter, trace amplitude balance, applyAGC.

The dominant frequency from the highresolution survey is approximately 120 Hertz.The highest frequencies observed wereapproximately 200 Hertz. Based on thefrequencies available, it was possible todiscern horizons approximately 4 meters inthickness. Features that were thin, or sporadic,may have been imaged in some cases, butwere not interpreted in this study.

Velocity logs from GCB-1 and 2, and fromP8R and P9R, were used to construct syntheticseismic sections tied to stratigraphic horizons.The synthetic seismic section from GCB-1tied to seismic line A-5 is shown in figure 3.Using the synthetic seismic ties, it waspossible to identify specific reflectionhorizons on the seismic lines as specificdefined stratigraphic horizons. The crystallinebasement reflector, Top of Cape FearFormation, Top of Middendorf equivalentzone, and in some cases, the McQueen BranchConfining Unit were interpretable. Thesesurfaces represent regional unconformities oreustasy related sequence boundaries andgenerally represent an acoustic impedancezone detectable by the seismic technique. Inonly a few cases were shallower horizonsinterpretable.

BACKGROUND GEOLOGY

Figure 3 is the generalized stratigraphiccolumn for the Savannah River Site region.The depositional environments and sedimentcharacter for GCB-1, located in the center ofthe study area and thought to berepresentative, are shown on Figure 4a. Figure4b shows the tie between key borngs GCB-1,2, and 3. A detailed discussion of the


M-4

stratigraphy and deposition is found in Fallowand Price (1995) and Aadland et al., (1995).Parts 1 and 2 of the Advanced A/M GeologyStudy (Wyatt et al., 1997 a and b) discuss thedata derived from the ground penetratingradar, cone penetrometers and the geophysicalcharacterization of borings GCB-1, 2 and 3.

A conceptual model of the anticipated geologyin the region of the A/M Area provides anintellectual framework to interpret theborehole geophysical and seismic data. Thebasic model should be consistent with knownregional geology as described in establishedscientific literature. As data are evaluatedagainst the model, an iterative thought processwill allow features to be interpreted within theframework, or will suggest that modificationsneed to be made in the model. If there arehighly unusual features that violate thecharacter of the anticipated geology within themodel, then a data “bust” may be indicated, ornew information may be present to advancethe known geology within the area.

The conceptual model of the anticipateddeposition environment within the A/M Areais a near shore to delta plain environment.Therefore, sediments associated with a nearshore deltaic, fluvial and near shore marineare anticipated. Variations in faciesarchitecture associated with depositionalprocesses in these environments are alsoanticipated. Facies architecture variationsinclude “fining upwards” or “coarseningupwards” sedimentary sequences in atransgressive-regressive system. Sedimentsassociated with near shore, littoral zone andbeach environments may be highly variableboth laterally and vertically. The modelassumes that depositional sequences withinthe area are consistent with those published byFallaw and Price (1995) and anticipates theregional unconformities present in thesesequences to also be present in the area.

Structural features within the area areassumed to be consistent with the known postdepositional regional stress fields. Thepresence of the Crackerneck Fault as defined

in Stephenson and Stieve (1992) is alsoassumed. Any subsequent faulting within thearea is assumed to be consistent with theCrackerneck Fault or associated with the postdepositional stress fields. The conceptualmodel assumes no preferred age restrictionson structural features. The presence of faults,fractures and folds are anticipated.

The description of the sediments encounteredin GCB-1 and described on Figure 5 verifiesthe basic conceptual model.

STRATIGRAPHIC INTERPRETATION

Geophysical logs and core data from shallowborings and wells are the dominant source ofdata within the A/M Area. Few boringspenetrate depths greater than 100m (330 feet)and most logs terminate within the LangSyne/Sawdust Landing Formation (near thePaleocene-Eocene contact) and are tooshallow for direct correlation with the seismicinformation. To allow for direct correlationwith the higher quality information from thedeep borings, the geophysical logs and coredata from the shallow wells were tied to thestratigraphy determined from the deeperborings. The correlation of the shallowborings is discussed in Aadland (1997) andelsewhere in this guidebook.

Cone penetrometer soundings generallyextend to depths of approximately 50m (150feet) or approximately to the top of theWarley Hill Formation and occasionally reachthe “green clay” bed within the formation. TheFriction Ratio’s from the cone penetrometerdata were correlated with the geophysical logsfrom the shallow borings. See Syms et al., inthis guidebook for a further discussion of theCPT.

INTERPRETATION OF DEEP COR-RELATION BORINGS

Lithostratigraphy

The interpretation of the GCB borings wasguided by the need to establish relationships


M-5

between core derived and geophysicallyderived geology. Additionally, the correlationbetween the geophysical log interpretation andsurface derived seismic data is important tounderstanding the overall A/M, and SRS,geological setting. The GCB-1,2,3interpretation will be presented as a unit byunit correlation comparing core geology andlog derived depositional signatures.Hydrostratigraphic nomenclature follows thatof Aadland et al. (1995). Stratigraphicnomenclature and environmental comparisonsare from Fallaw and Price (1995) and Reineckand Singh (1980). Log interpretations followsuggestions from Schlumberger (1989) andSerra (1985).

Saprolite and Basement (Pzm/Sap andAcoustic Basement)

Crystalline basement lithologies are similaracross the GCB-1,2,3 area. In GCB-1, an oliveto gray/green weathered chlorite schistoverlies a highly fractured bluish gray gneiss.The thickness of the schist is approximately47 feet (14.3 m) and a distinct contact ispresent between the schist and gneiss. InGCB-2 and 3, only the chlorite schist intervalwas sampled which had a similar appearanceto the schist of GCB-1. Fractures within theschist of GCB-1 have an approximatenorthwest strike orientation with a dip varyingbetween 13 and 27 degrees. Fractures andquartz veins within the gneiss appear to have anortheast strike orientation and gentle dips,however, only the upper few feet of the gneisswere sampled for dip information. No orienteddip or strike information is available for GCB-2 and 3. All of the crystalline basementlithologies encountered in GCB-1,2 and 3were expected as referenced in Cumbest andPrice (1989).

The top of basement contact is the firstcontact with the highly weathered andfractured saprolite. Generally, the saproliteincreases in thickness from GCB-2 and 3 toGCB-1 from approximately 13 feet (4 m) inGCB-3 and 15 feet in GCB-2 toapproximately 32 feet (9.8 m) in GCB-1. The

character of the saprolite varies in eachboring. In GCB-1 the saprolite contact isdefined by a sharp change between a wellindurated sandy/silty clay to gray-green claywith numerous iron oxide stains. The colorvaries from gray-green to brown-red. In GCB-2 the saprolite is a brown to red mottled claygrading into a dense, massively foliated olivegreen clay while in GCB-3 the saprolite is abluish gray to off white, well indurated clay.

Cape Fear Formation (Top CF)

The Cape Fear Formation generally thickenstowards the southeast from GCB-3 to GCB-1.The overall lithology of the formation consistsof interbedded gravel, sands and clays ofvarying colors and sizes. Rock fragments andangular sands suggest that the Cape FearFormation was deposited near its source.There are generally three member sequences,alternating coarsening and fining upwards,that define the formation and that are mostdeveloped in GCB-1. The overall geophysicallog pattern of the sequences, which matchesthe field core descriptions, suggests that eachunit within the formation is a lag channel,shoaling deposit fill with pebbles and heavysands grading into finer sands withinterspersed clays. This interpretation isconsistent with the description in Aadland etal., (1995) and Fallaw and Price, (1995).

Within the Cape Fear and at the basalMiddendorf, rock resistivity values and deepversus shallow curve separation (an indicatorof relative permeability) decrease rapidly. Theoverall neutron porosity drops from anaverage of approximately 35% toapproximately 20%. The low relativepermeability and porosity suggests tighterformations. The lower resistivity values areassociated with higher concentrations ofheavy minerals and interstitial waters are moremineralized than shallower aquifers. Overall,the Cape Fear and saprolite are indurated withlower permeability, forming the AppletonConfining Unit.


M-6

Middendorf Zone (Top Mid)

The Middendorf Zone (formerly Formation)gradually thickens towards the southeast fromGCB-3 to GCB-1. There are correlatablepackages of sediments across the formationbut it is difficult to correlate specific sand orclay units across the area defined by GCB-1,2 and 3. Channel and sheet type sands areinterbedded with indurated clays suggesting alower delta plain depositional environment.Many of the sand units in GCB-1,2,3 appearto alternate between fining and coarseningupwards suggesting that they are marineinfluenced channels or bars. Spectral gammadata, when combined with field coredescriptions and lab core analysis, suggeststhat there are few clays of significantthickness within the formation. Generally, theclays are thin and less than one meter thick.Numerous heavy minerals and micas, as wellas fine grained sands with clay balls, generatea spiked appearance of the gamma curvewhich may be interpreted as numerousinterspersed clays. Focused or InductionResistivity data suggests that there arenumerous high permeability zones separatedby numerous ‘tight’ streaks of clay or silt tosilty-sand. These zones also suggest that thereare non-correlatable sand bodies within theformation with variable permeability,therefore, tracking groundwater movementwithin the Middendorf within the A/M Areawould be extremely difficult.

Black Creek Formation (Top BC)

The Black Creek Formation in borings GCB-1,2,3 may be described in terms of an upperand lower member, possibly two sequences,separated by the McQueen Branch ConfiningUnit (MqBCU) regional aquitard. Both upperand lower zones have geophysical log andcore characteristics that define them asprimary regional aquifers.

The lower zone lies unconformably over theMiddendorf and consists of interbedded thinclays and sands occurring, more or less,randomly distributed and uniformly spaced

throughout the interval. Spectral gamma peakto peak correlation, indicating regional clayunits, are possible to track across the borings.These clays are probably localized maximumflooding surfaces, while the basal clays,similar across the area, were deposited in anear shore marine environment. Theinterbedded sands and clays, often withinterspersed iron staining and heavy minerals,is suggestive of a fluvial to upper deltadepositional environment.

The MqBCU is a sequence of three principalclays and silty clays occurring in a 40 foot(12.2 m) interval separated by clayey to siltysands. Each of the three units is approximately10 to 15 feet (3.0 to 4.6 m) thick with thelower unit having the most consistent claycontent and thickness. Many of the sandswithin the MqBCU are moderately sorted andsub-angular with a consistent presence ofmicas suggesting that these sands are fluvialand deposited near their source. The clays alsoconsistently have micas presence, suggestingthat the clays and the sands are both fluvialdeltaic to near shore sub-tidal.

The upper zone of the Black Creek is variableacross the region of the GCB borings. InGCB-2, the interval consists principally ofinterbedded clays and sands in beds averaging5 to 7 feet (1.7 to 2.1 m) thick. This ‘serrated’depositional style is common along themargins of bar-finger sands (Busch and Link,1985). In GCB-1 and 3, the dominantlithology in the upper zone is sand with thininterspersed silt and clay intervals. The sandsare dominantly moderately to well sorted, tanto cream in color, and sub-angular to sub-rounded. Micas are present and sometimesabundant with rare heavy minerals. The sandis 75 feet (22.9 m) thick in GCB-3 and 37 feet(11.4 m) thick in GCB-1. These sands are notmassive but consist of stacked bar finger, bar-axis and barrier sequences with dips (in GCB-1) of 2o to 12o to the northwest in the upperunits, to 14o to 21o to the south-southwest inthe middle sands followed by 15o to 29o dipsto the northeast in the lower sand units. Theaverage dip direction is to the south. The


M-7

alternating dip directions are indicative ofnear shore influence and the measured dipsare probably alternating barrier or duneforeset beds.

Samples for palynological analysis wereobtained from the GCB-1 boring in a dark,organic rich clay from a depth of 439.5 feet(134 m, Figure 4a) at the base of the uppersand unit and top of the MqBCU.Palynomorphs were present indicating a lateCampanian to early Maestrichtian age. Thespores, pollen, plant cuticle, wood fibers andinertinite, plus a lack of marine algae,indicates that this interval is non-marine(Christopher, personnel communication,1996).

Steel Creek Formation (Base of CBCU toTop BC)

The Steel Creek Formation is defined from thebase of the Crouch Branch Confining Unit(CBCU) to the top of the Black CreekFormation. Dominant lithologies consist ofmicaceous clays interbedded with sands, sandsand silty sands interbedded with thin clays andpebble zones. Micas, including angular grains,are present throughout the interval and ironstaining is common. In the basal 50 feet (15.2m) of each boring, clay content increases andalternates with fine grained silty sands. Theoverall dip direction is generally random.Sands appear to be more prevalent in theupper portions of the Steel Creek and exhibitmorphologies similar to distributary barrierand bar sequences. In GCB-1, these sandsgenerally fine upwards, and exhibit varyingdip angles from up to 20o northwest to 12o

southeast and 18o northeast. The ‘C” sand ofAadland et al., (1995) is present in all wellsand gradually varies from a coarseningupwards distributary bar finger and delta lobesands (Busch and Link, 1985) in GCB-1 and 2to a more marginal bar finger sand or deltalobe in GCB-3. The remaining sands in theupper portion of the Steel Creek aredepositionally similar to the ‘C’ sand but tendto be more tidal in GCB-3. In general, thesands of the Steel Creek appear to be more

landward in GCB-3 and more seaward inGCB-1. It is possible that theCretaceous/Tertiary boundary (K/T), andcontact with the Paleocene, is present at thebasal contact of the clay unit immediatelyoverlying the ‘C’ sand. The presence ofpebble lag layers and heavy minerals maydefine this time sequence marker.

Crouch Branch Confining Unit (Top CBCUto Base CBCU)

The Crouch Branch Confining Unit varies inthickness and clay content from approximately17 feet (5.2 m) of thin and sparse clays inGCB-3 to 44 feet (13.4 m) of thick silts andclays in GCB-2 to 50 feet (15.2 m) of siltyclays and clays in GCB-1. The clays tend tobe a white to light gray, plastic kaolinite withthinly laminated purple mineral horizons. TheCBCU in GCB-1 is predominantly kaoliniticclay with only minor interspersed laminae ofsilty clay. In GCB-2, thin lamina of heavyminerals and silts are interspersed within theclays. Thin sands within the unit often haveheavy minerals and abundant micas. Thinzones of cemented iron laminae are alsopresent. In GCB-3, only thin clay units arepresent interspersed with silty clays, sandysilts and sands. Numerous lamina with ironstaining and heavy minerals are present. Thethin sands tend to grade into pebbles and clayballs. The thin clays, clay balls and shoalingsands units in GCB-3 suggest that GCB-3CBCU sediments were deposited in a fairlyhigh energy, near shore, possibly intertidalzone, transitioning into a subtidal to deltadistributary depositional environment in GCB-1.

This unit, which consists of parts of variousformations, forms the principal confining unitacross the SRS; therefore is discussed in detailas it controls downward contaminantmovement. Permeability variations in theCBCU, determined from magnetic resonancegeophysical logs, are observable in GCB-1.The interval from 240 feet (73.2 m) to 250feet (76.2 m) is a well indurated and mottledclay. Permeability average less than 200


M-8

millidarcies. From 240 feet (73.2 m) to 265feet (80.8 m), subsurface, the measuredpermeabilites increase to greater than 1 darcy.All clay within this interval is well induratedand without other noticeable features. Variouscolor changes, suggesting changes in thegeochemical makeup of the clays, formnumerous lamina throughout the section. Themagnetic resonance tool is probably imagingthe horizontal permeability associated withthese lamina. In the base of the CBCU, fromapproximately 265 feet (80.8 m) to the top ofthe Steel Creek Formation at 290 feet (88.4 m)subsurface, the permeability decreases fromapproximately 500 millidarcies to zero. Thisinterval is also a well indurated clay withfewer lamina, grading into a silty, micaceousclay. At 290 feet (88.4 m) subsurface, themagnetic resonance permeability increasedramatically in the sands of the Steel CreekFormation.

Lang Syne/Sawdust Landing Members (TopLS/SL to Top CBCU)

The Lang Syne and Sawdust LandingFormations (refer to Aadland et al., 1995 andFallaw and Price, 1995 for nomenclaturediscussions) are not lithologicallydifferentiated within the A/M Area of theSavannah River Site. Within the area of GCB-1,2,3, these formations vary in thicknessbetween 28 feet (8.5 m) in GCB-3 to 46 feet(14 m) in GCB-2 and 12 feet (3.6 m) in GCB-1. Two distinct, lithologically different andcorrelatable, members are seen in each boring.The basal member of the Lang Syne/Sawdustlanding is a lower bar to delta sheet sand thatthins downdip to the southeast. The uppermember is a silty clay to clay coarseningupwards into a delta front sheet sand that hasthe greatest thickness in GCB-2. The silty-sand interval in the upper member in GCB-1appears to be the distal edge of the delta front.Overall, the general dip of the LS/SLsediments is south-southeast. The sandinterval in the upper member appears to be abar sand that is probably regressive, intertidaland defines the Lang Syne/Sawdust Landingregional unconformity.

The lower 10 feet (3 m) (230 feet (70.1 m) to240 feet (73.2 m) subsurface) of the LS/SLhas an average permeability that exceeds 1darcy. This interval is a well indurated claywith iron oxide staining and containing somepyrite. The measured permeability in thisinterval is probably due to microfracturing andthe horizontal character of the iron oxidelamina. This interval is capped by a lowpermeability (less than 10 millidarcies) thinlybedded clay and silty clay. Overlying thisclay, is a zone that transitions into a highpermeability (greater than 1 darcy) thatprobably corresponds with the LangSyne/Sawdust Landing unconformity. InGCB-1, the combination of the underlyingCBCU and the high permeability associatedwith the LS/SL unconformity create the zonewith the highest TCE concentrations. Theability to track and map the LangSyne/Sawdust Landing Formations and theunconformity zone suggests a way to possiblytrack halogenated solvents.

The following stratigraphic units, althoughEocene in age, are discussed for continuitywith the work of Aadland in this guidebook.

Congaree Formation (Top Congaree)

Unconformably overlying the LangSyne/Sawdust Landing Formations are theassemblages of Early Eocene sediments thatcomprise the Fourmile Formation. It islithologically difficult to distinguish thesespecific formations from the overlying lowermiddle Eocene Congaree Formation and theentire interval is generally referred to as theCongaree Formation (Aadland et al., 1995).Generally, the lower sands within the interval,when present, are considered equivalent to theFourmile Formation while the upper sands,when present, are a portion of the CongareeFormation.

Generally, the depositional character of theCongaree in GCB-1,2,3 is similar. There is aslight increase in formation thickness from


M-9

GCB-3 towards GCB-1 and 2 suggesting aseaward dip. The lithology of GCB-2 and 3consists of sands, medium to coarse grained,interspersed with thin indurated clays. GCB-2is dominantly sands with thin silts and a fewclay lamina and GCB-3 is dominantly sandwith thin clay and iron stained lamina. InGCB-1, there are lithified sand intervals 1.2 to1.6 inches (30 to 40 mm) thick and siliceoussandstones up to 2 inches (50 mm) thick. Wellindurated clays are present as a central unitwithin the formation and are approximately 10to 15 feet (3 to 4.6 m) thick associated withsilty clays and clayey silts.

Permeability zones are variable within theCongaree. In GCB-1, generally, within theformation, cleaner sand zones have higherpermeability, approaching 500 millidarcies,and silty to clayey sands have permeabilityless that 5 millidarcies. Traces of PCE andTCE are generally seen in the thin sands thatare immediately above the well induratedclays.

A shallow shelf to shoreline environments ofdeposition for the Congaree is evidenced bythe dip directions in GCB-1. Basal Congareesands (possible Fourmile or FishburneFormation equivalents) generally rotate from alow angle south-southwest dip through awestward to north-northwest dip directionwith an angle of 6o to 8o to a low anglenortheasterly dip. The upper sand unit(probable Congaree Formation equivalent) hasvery high, up to 24o dips generally striking tothe northeast. The upper sand interval is mostindicative of a possible dune sand with thehigh angle dips suggesting foreset bedding.The basal sands are most indicative of deltaicto near shore barriers and bars.

Warley Hill Formation (Top WH)

The Warley Hill Formation is present in GCB-1, 2 and 3. In GCB-3, the formation isgenerally geophysically indistinguishablefrom the over and underlying units, however,a subdued gamma pattern is similar to thepatterns from GCB-1 and 2. In GCB-1, the

Warley Hill consists of a sand withinterspersed iron oxide modules overlying asilty sand to thinly laminated silty clay. Thesilty clay, and to some extent the silty sand,form the Gordon Confining Unit withmeasured log permeability of less than 50millidarcies. This clay interval is regionallyextent and is thought to represent a maximumflooding surface. Sands within the upperportion of the formation generally havepermeability on the order of 100 to 200 md. InGCB-1, the dip directions within the WarleyHill are varied.

In GCB-2, the Warley Hill sediment characteris similar to GCB-1 but the formation islocally exposed on the surface and has beenaffected by near surface processes. In GCB-3,the formation is generally sandier than GCB-1, and lacks the thin clay lamina and siltysands that constitute the confining unit.

Correlation With Other Deep Borings

The deep correlation borings used to calibratethe seismic data are GCB-1, 2, 3 and P8R,P9R, MMP-4-SB and MBC08ASB. Thelocations for these boring are shown on Figure2. The stratigraphic horizons from GCB-1, 2,3 were used to correlate the horizons fromP8R, P9R, MMP-4-SB and MBC08ASB. Thegeophysical data from all of these borings(except MCB08ASB) are calibrated tointernational standards and were used tocorrelate and verify all geophysical logswithin the A/M Area. Data were correlatedusing LandMark® and TerraSciences®

geological software. Figure 6 is a structuralcorrelation panel showing the stratigraphichorizons for these deep correlation wells. Thispanel is flattened on an elevation of 97m (300feet) relative to mean sea level. Thestratigraphic horizons are those utilized tomap specific intervals within the study area. Ingeneral, only the Late Cretaceous horizons arevisible on the seismic data.

The figure 6 correlation panel follows noparticular strike or trend except that P8R,GCB-1 and P9R are in the central portion of


M-10

the A/M Area with P8R being more westerly,GCB-2 is in the southeastern area, MMP-4SBand GCB-3 are in the eastern area andMCB08ASB is in the northern area.Therefore, the correlation panel proceeds fromthe west central portion of the study areasoutheastward, westward and then to thenorth. These logs are used only to evaluate therelationship of the stratigraphic units acrossthe study area and, along with their velocityand other data, provide a framework forinterpreting the seismic data. Also,“flattening” on stratigraphic horizons allowsfor the relative interpretation of depositionalcharacter and relative fault motions to bemade.

This panel demonstrates several interestingfeatures. All horizons, except the McQueenBranch Confining Unit, from MCB08ASB inthe north to P8R in the southwest, show thesoutherly basinal dip. There are offsets incrystalline basement, particularly associatedwith GCB-3,P9R and GCB-1, that arecontinuous through to the Lang Syne/SawdustLanding unconformity. In GCB-3 andMCB08ASB the regional structure does nottrack as well as in the wells to the south andwest and it is possible that there is anothererosion event beveling the structure prior tothe Lang Syne/Sawdust Landing event, andprobably occurring at the Cretaceous/Tertiaryboundary. In well P8R, the structure reversesfrom a general downward sense of motion toan upward sense of motion in the lower BlackCreek Formation. It is possible that well P8Rcuts a reverse fault in the Middendorf/BlackCreek intervals. This is supported fromhistoric anecdotal evidence that the lowerintervals of well P8R are deviated along afault plane.

The Middendorf, Black Creek (includingMcQueen Branch Confining Unit) and SteelCreek Formations (generally, the entireCretaceous section) are fairly uniform inthickness across the area. The Middendorfthins (or the Cape Fear may be missing) inP9R suggesting that the upwards motion notedin basement was present and either erosion or

non-deposition occurred during Middendorftime. The upper Black Creek also thinssuggesting reactivation of the upward sense ofmotion during Black Creek time.

The Steel Creek thins in GCB-3 suggestingerosion or non-deposition at this locationduring or in post Steel Creek time. The overallthickness of the Steel Creek to LangSyne/Sawdust Landing interval including theclays of the early Paleocene, may be describedas a transitional strata. The LangSyne/Sawdust Landing shows only minimaleffects from the underlying structure (on theorder of 3 to 4 meters). The basement relatedstructures show approximately 30 to 40 metersof relief.

The Congaree and Warley Hill Formations arefairly uniform in thickness. The sag seen inthe Warley Hill in GCB-3 is contrary toregional dip and is probably associated withthe basement structure. The sag seen in GCB-1 may also be associated with a basementrelated structure. The interval from the Santeeunconformity to the Warley Hill unconformityis generally uniform but thickens in GCB-1.The Santee surface is also a cut and fillsurface and suggests that the motiondocumented in GCB-1 was not active in post-Santee time.

Interpretation of Seismic Data

The seismic data were interpreted using theLandMark® software system according to thefollowing approach. Each processed seismicline was plotted on the regional map withassociated deep correlation borings. Availableborehole sonic data from the deep borings,(from GCB-1, 2, 3, P8R, P9R and MMP-4-SB), were utilized to generate a syntheticseismic image, using the GMA® seismicsoftware, for the area adjacent to the borehole.The synthetic seismogram ties well derivedstratigraphic horizons to specific seismichorizons and allows for the interpretation andtracking of stratigraphy and structure across aseismic profile. Figure 5 is the syntheticseismogram from GCB-1.


M-11

The synthetic seismogram for each well wascorrelated with the adjacent seismic profiles.Horizons on the seismic profile that werecorrelative with the synthetic well tie weretraced across the profile. In general thecrystalline basement reflector (rock), the CapeFear/Middendorf contact (top Cape Fear), thebasal Black Creek/Middendorf contact (topMiddendorf), and McQueen Branch ConfiningUnit horizons were interpretable stratigraphicevents. The Lang Syne/Sawdust Landingintervals were not consistently discernibleacross the seismic data. In a few areas, theWarley Hill interval appeared to beinterpretable, however this occurred toosporadically to map. In general, no shallowerhorizons were obviously interpretable on theseismic data. Figures 7 and 8 showrepresentative seismic data with stratigraphichorizons and faults interpreted.

Stratigraphic horizons were traced from welltie points following dominant seismicamplitude events. In areas where the horizonswere difficult to track, no interpretations weremade. In many cases, one stratigraphichorizon may be interpretable while otherswere not. The two way travel time values foreach seismic shotpoint with an interpretedhorizon were tracked, converted to depthbased on the synthetic seismograms, andtransferred to EarthVision 4.0® for mapping.

Faulted intervals were determined using thecriteria of Sheriff (1982) and Campbell (1965)and include: 1) abrupt reflector terminations,2) direct fault plane reflections, 3) presence ofdiffractions especially at a termination, 4)visible drag and rollover, 5) correlations ofreflectors across a fault plane and possibly 6)loss of coherency beneath a fault plane ordistorted dips seen through the fault plane(fault shadow). Using these criteria, faultswere defined on the seismic profiles inLandMark® and transferred into EarthVision®

for mapping.

It is probable that channels and otherdepositional features exist within the A/MArea and are imaged on the seismic data.

However, the channels or other feautures arenot highlighted on the seismic data and are notdiscussed in detail in this document becausethey will require a more intense interpretiveeffort using seismic signal attribute data (suchas instantaneous frequency and phase). Ifobvious channels are observed in the data,then they are mentioned. The criteria used todefine a channel are discussed in Reineck andSingh (1980) and include: 1) an obvious,incised erosional surface, 2) the presence ofcross-bedding, 3) presence of point barsequences, 4) an apparent thalweg, and 5)alternating sand/clay/gravel bars.

Individual 2-D Seismic Line Interpretations

Representative seismic data are shown onfigures 7 and 8 and the overall pattern anddistribution of seismic coverage is shown onfigure 9. Figure 9 also shows the relativeseismic depths in time and the trend of themajor faults in the area.

All seismic data are interpreted for faultingand stratigraphic horizon markers, asdetermined from the synthetic seismic datashown on figure 5. For all lines, the yellowhorizon represents the top of “rock” or thesurface of the unweathered Paleozoiccrystalline basement. The dark red horizonmarker is the contact of the Cape FearFormation (Coniacian/Turonian, 90 mybp)with the overlying Middendorf Formation.The pink horizon is top of the MiddendorfFormation (Campanian/Santonian, 82 mybp)or contact with the overlying Black CreekFormation, a regional unconformity. Theorange horizon is the top of the confining unitor clay package that has been designated asthe McQueen Branch Confining Unit(Aadland et al., 1995). This horizon lieswithin the Black Creek Formation and mayrepresent the Campanian/Maestrichtiancontact, another major unconformity. Thesefour Cretaceous aged horizons appearconsistently on all of the data and are used asthe principal subsurface horizons for seismicmapping.


M-12

On all seismic lines, only five possible faultsare shown (reference Figure 9). Faults, andprojection lines, designated in red refer to theCrackerneck Fault (CNF). Faults in greenrefer to the MWESTA Fault and faults in pinkrefer to the Lost Lake Fault. Fault segments inorange refer to the Steed Pond Fault. Yellowfault segments are faults with no assignedorientation or name. These faults may havelimited extent, may be associated with anamed fault, or their extent may not bedetermined from our seismic coverage. Theangle of the fault segments posted on eachseismic line is somewhat arbitrary since ourdata were not designed to image structurebelow the crystalline surface.

Line SRS-1 is shown on figure 7 and isdiscussed as a representation of the remainingseismic data.. This approximate north-southline is the definitive data for the existence ofthe CNF, shown as a red line. Regional dipfrom this data is southerly at approximately1.5%. The vertical offset in the CNF isapproximately 12 ms two way time (TWT), or30 to 35 meters, on the basement horizon. Thehorizontal extent of the offset is across twoshotpoints or approximately 35 meters.Compensating faults, or more recent faults,associated with the CNF, are shown in yellow.These faults may accommodate the stressesassociated with the CNF. The northwesternfault (in yellow) is associated with the greendashed line which is the projection of theAMWESTA fault, therefore, these two faultsare probably the same. The fault noted in pinkto the southeast of the CNF is the Lost LakeFault and the dashed pink lines are theprojections of this fault through the area fromother seismic data. The basement reflector, inyellow is consistently tracked by the dark redCape Fear marker horizon except immediatelynorthwest of the CNF where there is anapparent increase in thickness of the CapeFear, or where there may be remnant Triassictrough sediments. The pink horizon is theMiddendorf Formation contact with the BlackCreek Formation. Over the faulted intervals, itis difficult to track a consistent horizon anddisturbed or eroded sediments are seen. The

orange McQueen Branch horizondemonstrates a fault propagation foldgeometry. Folded and flexed sediments areapparent from the McQueen Branch toapproximately 100 ms, although these eventsare not traceable across the data. The overallsediment wedge thickness appears to increaseto the southeast as expected. Changes inamplitude laterally along horizons to thesoutheast, particularly in the lower BlackCreek, may indicate the stratigraphic or facieschanges associated with shallow channeling,other near shore or fluvial processes. This lineintersects five of the A/M lines CNF-1, A-3,A-1A, A-2 and SRS-7. In all cases, theinterpreted horizons match those from theother lines and the data are consistently tied.

3-Dimensional Seismic Data

A three dimensional seismic survey wasobtained immediately southeast of line MCB-1. Results are shown in figure 10. Therectangular cube was obtained over a surfacearea of approximately 275 meters by 352meters using 120 cross-lines and 154 inlineswith a 2.2 meter spacing (a 2.2 meter binsize). An IVI MiniVibe vibratory source wasused. In general, best results were obtainedfrom 200 ms and deeper. Little to no datawere interpretable above 200 ms.

Figure 10 shows time slices, in relativeamplitude, through the data cube. Each timeslice has the time and stratigraphic intervalnoted. The projection and strike of the LostLake Fault, with relative sense of motion,which is projected through the area from the2-D seismic lines, is included for reference.The 325 ms time slice corresponds to theapproximate top of “rock” or crystallinebasement. A pronounced negative amplitude(purple) feature is seen in the southeasternportion of the cube, surrounded by a higheramplitude event. This low amplitude feature isthe same low amplitude feature seensurrounding the heart shaped high amplitudefeature seen in the 320 ms time slice(approximately 20 meters shallower).Therefore, the low amplitude event is higher


M-13

by 5 ms and has the same approximate senseof motion as the Lost Lake Fault predicts. At340 ms, approximately 40 meters below thetop of “rock”, the trend of the Lost Lake Faultbisects a phase change (of about 180o) eventpresumed to be the same for the shallowerhorizons. It is not evident that this relationshipexists upwards through the cube, althoughthere are notable changes across the projectedstrike of the Lost Lake Fault. The features thattrend northwest to southeast in the 260 ad 270ms time slices may be faults at oblique anglesto the main basement trend. This is supportedin the time slices at 256 and 260 ms. The NW-SE trending high amplitude reflector in theeastern central portion of the cube in slice 260ms is the same high, but weak, reflector in thecentral west portion of the cube, betweeninlines 40 and 60 of time slice 256 ms. Thisimplies a general up to the west sense ofmotion, however, the trend of the features isoblique to the trend of the Lost Lake Fault. Itis interesting to note that the trend of thesefeatures parallel the NW to SE trending sidesof the high amplitude basement feature seen inthe 320 and 325 ms time slices. This suggeststhat the basement trend of the Lost Lake Faultzone is NE to SW but individual faults withinthe zone may be parallel orsynthetic/antithetic to the basement trend.This also suggests that individual basementfault blocks may be small in size along thefault zone (approximately 150 by 150 metersin the case of the high amplitude feature in thebasement). It is not evident how far the faultextends into basement on the 500 ms timeslice. These data are basically random and outof the range of the seismic investigation.

It should also be noted that there are linear,amplitude coherency events in the 3-D data.These features are often associated withstructural features such as fault cuts, and arealso associated with sedimentary features suchas channel boundaries. Within the useable 3-Ddata, there are no obvious channel features orunique facies changes, although it is possiblethat there are facies changes associated withthe slopes of the faulted blocks. The trend ofthe low amplitude events in time slice 256 ms,

for example, may be interpreted as a channelmeander, however, inline evaluation suggeststhat the reflection event is structurallycontrolled. The lack of channels and definitivestructures is probably a function of the smallsize of the 3-D cube and the representativesubsurface sampled.

DISCUSSION

Mapping

Seismic horizon picks, converted from time todepth using the synthetic information fromGCB-1, GCB-3, P8R and P9R were combinedwith the stratigraphic picks from thegeophysical logs and CPT’s (whereapplicable) to create the database for mapping.Mappable horizons from the seismic datainclude the acoustic crystalline basement(“rock”), Cape Fear, Middendorf andMcQueen Branch surfaces. In a few cases theBlack Creek horizon was observable but thesewere too sporadic to adequately map. Datawere forced, using a static shift andvelocity/depth corrections, to match actualdepths from the deep correlation borings. Thedata from each horizon were then mapped. Forthe seismically derived horizons, the projectedmajor fault planes (MWESTA, CNF, LostLake and Steed Pond) were included in themapping database and used to constraincontours adjacent to the faults. Each mappedsurface is discussed.

Stratigraphic intervals from shallowerhorizons are discussed in Aadland (thisguidebook). The previous intervals from thework of Aadland (1997) were reviewed andmodified according to the regional deepboring correlations. Stratigraphic picks fromavailable cone penetrometer data werecombined with new and existing information.These stratigraphic “picks” were incorporatedinto Earth Vision® 4.0 for mapping. Shallowmapped horizons include the top of BlackCreek Fm., top of Steel Creek Fm., LangSyne/Sawdust Landing unconformity, and


M-14

Warley Hill unconformity. A surface elevationmap is included for discussion.

Top of Crystalline Basement (“rock”)

Figure 11 is the surface of the crystallinebasement. The dominant faults projectedthrough the area and seen on figure 9 areshown. It is obvious that the area in the centerof the map is complexly disturbed. The areabetween the Steed Pond Fault (most easternfault) and the Crackerneck Fault is structurallyhigher than areas east and west. Deeperbasement is to the southwest and westward ofthe CNF and MWESTA faults. The overallmaximum structural elevation change acrossthe area is approximately 70 meters. Regionaldip is to the southeast. Structure and dip onthe northwest, northern, northeast andsoutheast, and possibly the southwest edges ofthe map are influenced by map edge griddingaffects and are extrapolations of data.

Top of Cape Fear Formation

The surface of the Cape Fear Formation isshown on Figure 12. Although the same mapedge affects are present as in the basementmap, the general trend of the data are similar.The deeper basin is to the west southwest inthese data and the structural high through thecenter of the study area associated with theCNF and Lost Lake Fault is present. The CapeFear is basically draped over the basementstructure and has smoothed many of the minorstructural perturbations found on the basementmap by infilling. However, the formation isoffset by faulting. The structural changeacross the area on the Cape Fear is alsoapproximately 70 meters, suggesting thatmuch of the structure is post Cape Fear in age.Basinward dip is to the west southwest. Thedepositional character is not obvious on thismap, although sediments primarily associatedwith structurally lower areas east of thecentral high are present. It is probable that thisarea has been lowered structurally, but it isalso possible that there was a trough or flowpassage through the area as well.

Top of Middendorf Inteval

The structural top of the Middendorf intervalis shown on figure 13. The central structure ispresent with an overall elevation changeacross the area of approximately 50 meters.Generally the formation dips to the south. Thestructural affects of the Crackerneck andSteed Pond Faults are immediately obvious.Map edge effects are great, particularly thesteep structural change to the north betweenthe Crackerneck and Steed Pond Faults. Aramp like structure may be present from thecentral high eastward to the Steed Pond Fault.Structure is more complex west of theCrackerneck/Lost Lake Faults.

Top of McQueen Branch Unit

The top of the McQueen Branch unit is shownon figure 14. This map is different than theunderlying Middendorf in that the regional dipappears to be dominantly to the southeast.Structural elevation change across the area isapproximately 20 meters and is represented bya monocline or dip reversal flexure across theregional faults. Minimal effects are seen fromthe MWESTA and Steed Pond Faults. Moreconvoluted structure is seen associated withthe CNF and Lost Lake Faults, particularly inthe area of the greatest seismic data density inthe center of the map. The gridding edgeeffects are great in the northern portion of thismap.

Top of Steel Creek Formation

The top of the Steel Creek Formation is shownon figure 15 and is generated from boreholegeophysical data. These values are the same asthose used in the manually contoured maps ofAadland (1997). The shallower faulting on thesurface of the Cretaceous from Aadland(1997) is shown on the maps for reference andmay be affecting the depositional pattern. Fewborings penetrated the Steel Creek horizonwithin the area and most are centrally located,therefore the edge effects on this map are


M-15

large. There are few regional faulting effectsseen in the map except for the convolutedstructure adjacent to the CNF and Lost LakeFaults. Between the time of deposition of theMcQueen Branch unit and Steel CreekFormation, there was a large near shoremarine depositional influence across the area.Thick marine sands were deposited in theupper Black Creek Formation, and a regionalunconformity or disconformity at the top ofthe Black Creek-contact with basal SteelCreek has smoothed the surface. A regionalunconformity, the possibleCretaceous/Tertiary boundary global erosionevent has further planed the Steel Creeksurface.

Surface of the Lang Syne/ Sawdust LandingUnconformity

The top of the Lang Syne/Sawdust LandingFormation is shown on figure 16. This mapwas also created using the same data asAadland (1997). The northwest to southeasttrending structures may be associated withshear reactivation along the CNF and LostLake Faults (shown on this and subsequentmaps for trend information). There arethickness variation in the Lang Syne/SawdustLanding sediments across the area that appearto be related to older structure on the surfaceof the Steel Creek Formation. The trend of thestructural features is oblique to the strike ofthe regional faulting. Deformation extent maybe restricted by the bounding AMWESTA andSteed Pond Faults. The initial timing of thisoblique faulting/folding is observed in thestructural map of the Steel Creek Formation,therefore it is assumed that the timing of thereactivation is post-Ypresian. The change instructural elevation across the features isapproximately 12 meters as compared withapproximately 14 to 16 meters at Steel Creektime. It is not known whether the structure isaccomodated by folding or faulting but theabrupt elevation change in some areas maysuggest faulting while the less abrupt changesmay suggest folding. The en-echelon, periodiccharacter of the structural events (high, low,high. low, high) occurring approximately

every 1000 meters, is strong evidence forstructural control. A current dominated nearshore sand depositional character is notthought be the cause of the en-echelon patternbecause the features are along dip into thebasin and not along margin strike.

Faulting and Structure

The four major faults, interpreted from theseismic data, are accompanied by severalsmaller, unassigned faults. It is possible thatmany of the unassigned faults may becorrelated together and form additional faultsystems, possibly those of Aadland (1997). Itis also probable that many of the smaller faultare basement splays from the regional faultsand that the larger faults are the primaryoffsets in a complex fault system. However, itis not possible, without additional data, toconfirm these alignments. Many of theunassigned faults have significant offsets, orhave complex structures, and may beimportant to shallower horizons, however,additional data are required to examine thispossibility in detail.

The oldest fault thought to affect the A/MArea is the CNF. This fault trendsapproximately N35E across the study area andmay extend an indeterminate distance to theSSW and NNE. This fault shows reversemotion and is generally up on the SE anddown on the NW. The amount of throw oncrystalline basement varies fromapproximately 30 to 35 meters maximum toapproximately 10 meters minimum within thestudy area. The angle of the fault is generallygreater than 60o from horizontal. The CNFfollows a propagation fold model (Cumbestand Wyatt, 1997) and disturbs sedimentsthrough the Warley Hill Formation. It ispossible that horizons shallower than theWarley Hill are influenced but this is difficultto determine from the seismic data or boreholemapping. The correlation of CPT, GPR andhigh resolution seismic data from Wyatt, etal., (1997) suggests that there are near surfacestructural influences, however, it is not clear


M-16

that the shallow effects are from the CNF.They may possibly be from later faulting.

On seismic lines SRS-1 and A-5 (figures 7 and8 respectively), which cross the faultperpendicularly, a synclinal trough is noted inthe footwall block. An additional sediment fillis seen in the trough suggesting that eithersyndeposition was occurring during pre-CapeFear compressional faulting(Coniacian/Turonian?, 90 mybp) or that theCNF is a re-activated Mesozoic normal faultwith a depositional remnant in the downwarped interval associated with theextensional event (Triassic/Jurassic/earlyCretaceous, > 90 mybp?).

A schematic of the fault propagation modelsimilar to the CNF is shown on Figure 17along with an actual seismic schematic of theCNF from line SRS-1. The comparisonbetween these two figures is obvious althoughthe seismic schematic does not demonstrateany fore limb thickening of sediments exceptpossibly in the Lang Syne or Warley Hillintervals. The lack of sediment thickeningmay be from the high angle of the CNF andunconsolidated nature of the sediments abovethe axis.

The Lost Lake Fault is thought to be youngerthan the CNF because it apparently intersectsand disturbs the CNF near the northern end ofseismic line A-3 (refer to figure 9). Because ofthe conjugate nature of the Steed Pond andMWESTA faults to the Lost Lake Fault, theyare thought to also be younger than the CNF,and approximately the same age as the LostLake. These faults trend approximately N10E,and within the study area, only the Lost LakeFault intersects the CNF. To the southwest,and beyond the study area, the MWESTAFault may also intersect the CNF.

The MWESTA, Lost Lake and Steed PondFaults were grouped because they generallyexhibited the same relative sense of motion,along linear trends, throughout the study area.However, the relative sense of motion mayvary along these faults because of shear

displacement. On many of the mappedhorizons, the sense of offset varies along thesefaults and between faults. At the McQueenBranch Confining Unit (Figure 14) forexample, the Steed Pond and MWESTAFaults appear to have a left lateral offset whilethe Lost Lake may have more of a right lateralmotion. On this figure, it is possible that offsetexists in the shallow horizons (up on the NE)and that the fault splays into a positive flowerstructure, indicative of a shear motion.

On figure 16 the Lang Syne/Sawdust LandingNW/SE trending highs appear to truncate atthe MWESTA and Steed Pond Faults. Thesehighs may be flexural folds associated withthe differential shear motion between theMWESTA and Steed Pond Faults. The overallsense of motions for the fault zone boundedby the MWESTA and Steed Pond faultswould be right lateral. The axial crests of thestructural highs appear to be offset byapproximately 3000 meters from crest axis tocrest axis, which may be a predictive tool forlocation of additional highs.

Of the MWESTA, Steed Pond and Lost LakeFaults, only the Lost Lake appears toinfluence shallow sediments. In general, thestructural influence from these faults, asderived from the seismic data, appears todiminish above the Middendorf horizon.,however, the resolution of the data do notpreclude the possibility of shallowdeformation.

Assuming that structural influences haveaffected the topography of variousstratigraphic horizons through time, then itshould be possible to estimate timing and totaldisplacement of sediments. The effects oferosion are difficult to estimate. Figure 18 is agraph of approximate maximum elevationrelief, average, minimum and maximuminterval thickness for the stratigraphichorizons. Data were extracted from mapfigures 11 through 16 and values from thestratigraphic picks


M-17

Several observations may be made from thisgraph. The amount of structural relief on thesurface of the crystalline basement is thegreatest observed. The structural relief on theCape Fear and Middendorf exceeds theaverage and maximum sediment thickness,suggesting post or syndepositional structuredominated during the Mesozoic throughSantonian time. The minimum (and average)sediment thickness from the McQueen BranchConfining Unit to the Middendorf (lowerBlack Creek), and Steel Creek to McQueenBranch (Steel Creek and upper Black Creek),exceeds the overall amount of structural reliefon these horizons. This suggests thatdepositional events predominated overstructural events during the Campanianthrough Maestrichtian time. For the LangSyne/Sawdust Landing to Steel Creek interval,the structural relief is greater than theminimum sediment thickness and less than theaverage thickness. This suggests that acombination of structural and depositionalevents have occurred during the Paleocene.From the Warley Hill to Lang Syne/SawdustLanding (Lutetian to Ypresian), the sedimentthickness exceeds the overall amount ofstructural relief available suggesting adominantly depositional environment. Fromthe present day land surface to the Warley Hill(Bartonian and younger), the averagesediment thickness is exceeded by thestructural relief but the maximum sedimentthickness exceeds the relief. Therefore, thissuggests a combination of depositional andstructural events.

Based on the data from this graph, it ispossible to establish the following series ofevents: 1) Paleozoic to Coniacian/Turonian:structural events dominant (assumed), 2)Coniacian/Turonian to Santonian: structuralevents dominant, 3) Campanian toMaestrichtian: depositional events dominant,4) Paleocene: depositional and structuralevents, 5) Lutetian to Ypresian depositionalevents dominant, 6) Bartonian and younger,depositional and structural events.

Stratigraphy

A variety of depositional environments andstructural influences have affected the stratawithin the A/M Area. In general, all sedimentswithin the area may be characterized as nearshore marine to fluvial punctuated bysubmarine or subaerial erosional events.Within formations, the depositional charactermay change across the area as theenvironment of deposition changed alongslope. Adjacent to faults, the sedimentcharacter may vary due to tectonic influences.The regional stratigraphy is discussed inFallaw and Price (1995), Aadland et al.,(1995) and Aadland (1997).

The deep borings and seismic data may becombined to define the lateral variability andcharacter of sediments within a specificformation. The log signatures of thegeophysical curves may be used to interpretthe depositional sequence or environmentrelative to formation and depth, particularlyfor the sand sequences (Busch and Link, 1985,and Serra, 1985). This, in turn, may beassociated with an area along a seismic lineand further associated over a larger area. Eachdeep Cretaceous stratigraphic horizon will bediscussed. Tertiary strata are discussed inAadland (1997) and this guidebook.

Cape Fear Formation

The Cape Fear Formation varies in thicknessacross the A/M Area from approximately 6meters eastward to 12 meters westward. Thegeophysical log signature is similar in allpenetrating borings, therefore the depositionalenvironments are thought to be similar,although no definitive environment may beinterpreted from the geophysical data. Ingeneral, and referencing the core descriptionsin Wyatt et al., (1997), the formation appearsto be a fluvial or shoaling lag deposit withinterspersed sands, cobbles, pebbles, silts andheavy minerals. Aadland (1997) contends thatthe Cape Fear may be missing in well P9R.


M-18

The geophysical logs from GCB-1, GCB-3,P8R and MMP-4-SB suggest that theformation may consist of at least two andpossibly three shoaling events. Threedistinctive lower energy intervals, punctuatedby sandy/silty higher energy events are notedin the GCB-1 log. In MMP-4-SB, three eventsare also noted but have a different character,suggesting a low energy or shoaling (possibleheavy mineral) lag, then a higher energyenvironment, followed by another low energyenvironment. Since structural relief exceedsdepositional thickness, it is possible that thestratigraphic and depositional character of theCape Fear is dominated by episodic structuralevents causing periodic shoaling.

The seismic data provide little information onthe depositional or stratigraphic character ofthe Cape Fear. The seismic data does supporta thickening of the interval westward but ingeneral, the thickness of the Cape Fear is lessthan the resolution of the data. Overall, thenear source, shoaling environmentsinterpreted for the Cape Fear are consistentwith the initial rising sea levels during theLate Cretaceous and suggests that the A/MArea was near shore during this eustatic event.

Middendorf Interval

The Middendorf interval varies in thicknessacross the area from approximately 25 metersalong a central axis to 35 meters eastward andwestward in the area. It is possible to trackconsistent members through the formation,therefore the thinned areas are apparentlycaused by erosion. Individual members orbeds within the formation alternate betweenthinly bedded, high frequency, dominantlysands with interspersed clays to thinly bedded,high frequency, dominantly clays withinterspersed sands. The varying pattern ofsand units from “spikey” to “bell”, “funnel”and “serrate” shapes all suggest a bar, bar-finger, delta front sheet sand, depositionalenvironment. The best developed sand unitsare in GCB-3, PW-113, MMP-4-SB andMBC08ASB, suggesting that the north and

west portions of the study area wereshoreward during the Middendorf deposition.Neither the sand or clays are well developedin GCB-1, P8R or GCB-2 suggesting anenvironment with more of a marine influence.

The seismic data suggest that the Middendorfdevelops a more variable character to thesouth-southeast. Line SRS-7, an approximatedip line for Middendorf depositional time,shows a more variable character seaward,while having a relatively stable “railroadtracks” character landward or up dip. Seismicline CNF-2, an updip strike line forMiddendorf deposition also displays a“railroad track” appearance while line CNF-1,further down dip, has a more variablecharacter.

Black Creek Formation (including theMcQueen Branch Confining Unit)

The Black Creek Formation’s depositionalcharacter and environment is similar to theMiddendorf Formation except that an increasein silts and clays throughout the formation issuggestive of a lagoonal to back bay shallowwater and lower delta plain environments.Distinctive sand and clay units are mappableacross the area and are generally welldeveloped to the north and east and lessdeveloped to the south and southwest.

The McQueen Branch Confining Unit(MqBCU) is a member of the Black CreekFormation, consisting of thick claysintercalated with thin sand/silt layers. TheMqBCU approximately divides the BlackCreek into upper and lower units. In GCB-1, 2and 3, the MqBCU contains rich organic claysinterspersed with heavy minerals and ironstaining. A greater abundance of heavyminerals and staining is present in GCB-3suggesting deposition further up slope. Acrossthe structural high associated with GCB-1 andP9R, the MqBCU thins suggesting a structuralinfluence affecting deposition. In general, theMqBCU appears to be distributed in a nearshore, back barrier bay environment and mayrepresent a sequence time line or surface


M-19

erosional interval between depositional cyclesof the over and underlying Black Creek.

The lower Black Creek zone is generally moreclay dominated than the upper zone. Several(generally three) clay units may be trackedacross the lower portion of the Black Creekbut are missing or thin above the GCB-1 andP9R high. The sand units within the lowerzone have log characteristics of bar-fingersands or delta sheet sands combined with barsands. Since the sand signatures are difficultto track across the data, they are probablymore associated with discontinuous barsrather than delta type sheet sands.

The upper Black Creek zone is sanddominated and was probably deposited in ahigher energy environment. Individual sandunits are difficult to map across the area andthe overall log signatures suggest bar-fingersands and bar pinchouts. In some logs,MBC08ASB for example, there may be atidal inlet fill log signature.

Across the A/M Area, the overall characterand variation of the lower Black Creek andMqBCU remains fairly constant. Furtherdown dip to the south and southeast (southand east of the Crackerneck Fault), thedepositional character appears to change and amore varied depositional pattern is seen. Thismay be the transition from upper delta, tidallyinfluenced deposition to a lower delta marineinfluenced environment.

The seismic character of the upper BlackCreek is less distinct than the lower BlackCreek and the overall depositional pattern ismore difficult to discern. In lines CNF-2 andSRS-7, the character is more consistent in thedown dip areas to the south and southeast.This may indicate a more stable, marineinfluenced environment, while a higherenergy, more discontinuous environmentexisted to the north and northwest.

Steel Creek Formation

The stratigraphic character of the Steel CreekFormation is not discernible from the seismicdata. From the deep borings, the formation isfairly uniform in thickness across most of thestudy area but thins in GCB-3. The GCB-3 logsuggests that the upper portions of the SteelCreek have been eroded. There are distinctivesand units within the formation that may belocally tracked, however, they are nottraceable through all of the deep borings. Theoverall log character of the sand units suggestsa near shore bar-finger environment.Figure 11 shows the suggested depositionalenvironments by formation.

SUMMARY

This study identified four principal crystallinebasement involved faults. The oldest regionalfault trends NNE and is identified as theCrackerneck Fault. Three younger faults trendnorth. The westernmost fault is identified asthe MWESTA Fault and the easternmost faultas the Steed Pond Fault. The central faultintersects the Crackerneck Fault and isidentified as the Lost Lake Fault. These faultsare not thought to be single breaks in all areasbut linear zones of disruption. Theintersection of the Lost Lake and CrackerneckFaults is a highly complex area. These faultsare thought to be predominantly reverse faultswith possible dip-slip, or minor strike slipmotion. The overall pattern of theCrackerneck Fault and possibly the Lost lakeFault follows a fault propagation fold model.The Crackerneck may be a re-activatedMesozoic normal fault. The overall subsurfacestructural relief within the A/M Area isthought to be controlled by these faults forstrata older than the Middendorf and by acombination of sedimentation and structurefor more recent sediments.

At least three ages of faulting are observed inthe data. The oldest faulting is probably pre-Cretaceous and affected deposition of theCape Fear Formation. There may be sediments


M-20

pre-Cape Fear in the synclinal troughsassociated with these older faults. Themajority of faults are post-Cretaceous in ageand disrupt the entire Cretaceous sedimentcolumn. At least two pulses of post-Cretaceous movement are noted. A morerecent structural event is seen in the Paleoceneand younger sediments. This event may be are-activation of the deeper faulting in a strikeslip mode. These younger structures trend tothe NW and occur in an en-echelon pattern.This pattern may affect strata to near surfaceor surface horizons.

The stratigraphic information for the deeperhorizons is limited due to the limited numberof deep geophysical logs tied to the seismicinformation. However, a comparison ofregional deep logs allows for a discussion ofpossible deep stratigraphy. In general, allsediments were deposited in a near shore,marine to fluvial environment and vary from atidally dominated to a wave dominateddepositional character. For all Cretaceousstrata, there is a noticeable change indepositional character south and southeast ofthe Crackerneck Fault from a near shore tomore distal environment. Overall, there is ageneral thickening of Cretaceous strata to thesoutheast and south. Figure 19 provides adepositional relationship of the variousstratigraphic units underlying the A/M Area toa deltaic, near shore model.

REFERENCES

Aadland, R. K., 1997. Tertiary Geology of A/MArea Savannah River Site, South Carolina (U),WSRC-RP-97-0505, Rev 1., 62 p.

Aadland, R.K., J. A. Gellici, and P.A. Thayer,1995. Hydrogeologic Framework of West-CentralSouth Carolina. State of South CarolinaDepartment of Natural Resources, Water ResourcesDivision. Report 5.

Busch, D. A. and D. A. Link, 1985. ExplorationMethods for Sandstone Reservoirs, OGCIPublications, Tulsa, 327 p.

Campbell, F. F., 1965. Fault criteria. Geophysics,Vol. 30, No. 6, pp. 976-997.Cumbest, R. J. and D. E. Wyatt, 1997. Fault-bendfolding in unconsolidated Coastal Plain sedimentsand possible effects on hydrologic and contaminanttransport properties. Geological Society of AmericaAbstracts with Programs, Southeastern Section, V.29(3), p. 44.

Fallaw, W.C., and V. Price, 1995. Stratigraphy ofthe Savannah River Site and Vicinity, SoutheasternGeology. Vol. 35, No. 1, p. 21-58.

Faye, R. E., and D. C. Prowell, 1982. Effects ofLate Cretaceous and Cenozoic Faulting on theGeology and Hydrology of the Coastal Plain Nearthe Savannah River, Georgia, and South Carolina.U.S. Geological Survey. Open File Report 82-156.

Pascal, R., and R.W. Krantz, 1991. Experiments onFault Reactivation in Strike-Slip Mode,Tectonophysics. 188:117-131.

Prowell, D. C, 1994. Preliminary geologic map ofthe Savannah River Site, Aiken, Allendale, andBarnwell counties, South Carolina. USGS OFR 94-181. 11 p.

Reineck, H. E. and I. B. Singh, 1980. Depositionalsedimentary environments, 2nd ed. Springer-Verlag, New York, 549 p.

Serra, O., 1985. Sedimentary Environments fromWireline Logs, Schlumberger Educational Services,Houston, 211 p.

Sheriff, R. E., 1982. Structural interpretation ofseismic data, Education Course Series #23, AAPG,Tulsa, 73 p.Stephenson, D. and A. Stieve, 1992. Structuralmodel of the basement in the Central SavannahRiver Area, South Carolina and Georgia (U),WSRC-TR-92-120, 43 p.

Wyatt, D. E., R. J. Cumbest, R. K. Aadland, F. H.Syms, D. E. Stephenson, J. C. Sherrill, 1997.Investigation on the Combined Use of GroundPenetrating Radar, Cone Penetrometer and HighResolution Seismic Data for Near Surface andVadose Zone Characterization in the A/M Area ofthe Savannah River Site, WSRC-RP-97-0184


M-21

Wyatt, D. E., R. J. Cumbest, R. K. Aadland, F. H.Syms, D. E. Stephenson, J. C. Sherrill, 1997.Interpretation of Geological Correlation Borings 1,2, 3 in the A/M Area of the Savannah River SiteWSRC-RP-97-0185


M-22

Tu

Tu

Ttr

Ttr

Tdb

Tdb

Tm

Tm

Qal2

Qal1

3685000

3690000

4300

00

4350

00

N

Figure 1. General study area. Colors represent the surface exposure of various geological formations.‘Tu’ is the Upland unit, ‘Ttr’ is the Tobacco Road Formation. ‘Tdb’ is the Dry Branch Formation. ‘Tm’is the McBean Formation. ‘Qal1’ and ‘Qal2’ represent Quaternary alluvial fill material, with thenumbers representing relative ages of deposition. Grid coordinates are in UTM. Roads are noted as thinblack lines. The heavy black line defines the area of data used for this study. Geology is from Prowell(1994).


M-23

MBC-8

GCB-3

MMP-4B (P-30)

GCB-2GCB-1

P9R

P8R

Figure 2. Location of seismic lines and deep correlation wells and borings used in the study. Red lines arethe seismic lines and the well or boring locations are labeled. Data are superimposed on surface geology(from Figure 1).


M-24

0

100

200

300

400

500

600

700

800

900

MiddendorfFormation

Black CreekFormation

Lum

bee

Gro

up

Cape FearFormation

BlackMingoGroup

Groups Formations

"Upland Unit"

Dry BranchFormation

OrangeburgGroup

BarnwellGroup

SanteeFormation

Paleozoic Crystalline Basementor Triassic Newark Supergroup

Hydrostratigraphic Units

Congaree Formation

Northern SRS Southern SRS


Gordon Confining Unit

Piedmont Hydrogeologic Province

Sou

thea

ster

n C

oast

al P

lain

Hyd

roge

olog

ic P

rovi

nce

App

roxi

mat

e D

epth

in F

eet

McBean Mbr.Blue Bluff Mbr.

Tims Branch Mbr.

Warley Hill Formation

ClinchfieldFormation

Utley Mbr.Riggins Mill Mbr.

Period/EpochAge Unknown

Eocene

Paleocene

Tert

iary

Late

Cre

tace

ous

Lithostratigraphic Units(Fallaw and Price, 1995)

Irwinton Sand Mbr.

Twiggs Clay Mbr.Griffins Landing Mbr.

Fourmile Formation

Tobacco Road Formation

Snapp Formation

Lang Syne/Sawdust Landing

Formation

Steel CreekFormation

Nomenclature

Upper Three Runs Aquifer

Gordon Aquifer

Steed PondAquifer

MeyersBranch

ConfiningSystem

Crouch BranchConfining Unit


AllendaleConfiningSystem


UndifferentiatedAppletonConfiningSystem

Flor

idan

-Mid

ville

Aqu

ifer

Sys

tem

GO

001H

R4,

Rev

. 4

DublinAquiferSystem

MidvilleAquiferSystem

FloridanAquiferSystem

Figure 3. Stratigraphic chart representing the sediments within the A/M Study Area. Data are fromAadland et al., (1995).


M-25

Crouch B

ranch Aquifer

McQ

ueen Branch A

quifer

GCB-1

Gamma

SGR

CGR

0

0

200

200

Spectral

Thor.Uran.

Pot.

00

0

2010

1

Depth (ft)

Core Data

%Mud

%Sand%Gravel

000

100100100

Dens ./NPor.Acoustic

NPhi

Resistivity

DtCo60 30 0

10 0 01.5 2 .5RhoD ILD

ILM

0.20.2

20002000

Geochem./Perm.

0

0

PCE

TCE

1

3

MPrm0 1000AHT60

AHT20

44

105084 0

0

50

100

150

200

250

300

350

400

450

500

550

600

650

700

750

800

Environmentsof Deposition

clays and silty clays,thin sand laminaeshallow marine shelf,near shore to intertidal to delta distributary

distributary barrier and bar sequences to marginal bar finger sands and delta lobes

Interbedded sandsand clays, lower delta front to marine influenced channels and bars

interbedded sands & gravels, lag channel, shoal deposits, near source

chlorite schist,metavolcanics,fractured

weathered andfractured

interbedded thin clays and sands with a nearshore influenceupper delta plain

fluvial to nearsource clays, sands and silty sands, fluvialdeltaic to sub-tidal

bar finger and bar axis,upper delta, non-marine

Vadose Zone

Lost lake Aquifer Z

one

Green Clay Confining Zone

delta & interidal sands/clays

shallow marine shelf,near shore to intertidal to subtidal and shoreline, delta distributary clays

shallow marine shelf,near shore to intertidal to subtidal and shoreline, upper deltaplain

Top Congaree

Top Lang Syne/Sawdust Landing (LS/SL)

Top

Base

Top Black Creek (BC)

Top Middendorf (Mid)

Top Cape Fear (CF)

Pzm/Sap.

Top Warley Hill (WH)

Acoustic Basement

Interpreted Contacts

Base of Uplandnot inferred

Top Steel Creek (SC)

439.5’late Campanian to early Maestrichtianterrestrial,non-marine

McQueen BranchConfining Unit (MqBCU)

Base

Top

Age

Priabonian

Bartonian

Lutetian

YpresianThanetianDanian

Maestrichtian

Santonian

Coniacian?Turonian?

Paleozoic

Late Cretaceous

Paleocene

Eocene

90

82

74

62

54

43

40

<570

Numbers are ages in MYBP

Crouch BranchConfiningUnit

Appleton ConfiningSystem

“M” Area Aquifer ZoneS

teed Pond A

quifer

Cam

panian

Piedmont ProvinceAquifer System

Figure 4a. Combined interpretation of GCB-1. Stratigraphic picks are from the correlation of GCB-1,2and 3 to regional horizons. Depositional environments are interpreted from the geophysical logsignatures, field core descriptions and core laboratory data. Hydrostratigraphy is from Aadland et al.,(1995). Stratigraphy and chronostratigraphy are from Aadland et al. (1995), Fallaw and Price, (1995)and DNAG, 1983. Palynology is from Christopher (personal communication).


M-26

GCB-1 GCB-2 GCB-3

330’ msl

0

100

200

300

400

500

600

700

0

100

200

300

400

500

600

700

0

100

200

300

400

500

600

700

800

Surface

Top CongareeSteed Pond Aquifer

Top LS/SL

CBCU

TR/DB

Top BC

MqBCU


Top Mid

Top CFPzm/SapAcoustic Basement


Top WHSantee

GCCZ

“C” sand

Figure 4b. GCB-1, 2 and 3 correlation panel. Datum is set to 330’ (110 m) msl. Depths are in feet. Colorof correlation lines correspond to color of horizon name. “Top WH” is the top of the Warley HillFormation. “GCCZ” is the Green Clay Confining Zone. “Top LS/SL” if the top of the LangSyne/Sawdust Landing Formations. “CBCU” is the Crouch Branch Confining Unit. “Top BC” is the topof the Black Creek Formation. “MqBCU” is the McQueen Branch Confining Unit. “Top Mid” is the topof the Middendorf Formation. “Top CF” is the top of the Cape Fear Formation. “Pzm/Sap” is the contactwith the weathered basement rocks or saprolite. Acoustic Basement is the velocity crossover from coastalplain seismic velocities to those more typical of crystalline rocks.


M-27

Figure 5. Synthetic seismogram and tie to seismic data. Sonic and density data from GCB-1 were used tocalculate the synthetic seismogram. Traces from seismic lines A-5 and CNF-1, adjacent to GCB-1, were used forcorrelation. The synthetic data have been phase shifted by 300o. Key stratigraphic horizons, identifiable on theseismic data, have been defined.


M-28

Figure 6. Deep borehole correlation panel flattened on a 30 foot datum relative to mean sea level. Keystratigraphic horizons are labeled. Natural gamma, spectral gamma and/or calculated gamma ray curvesare shown.


M-29

Figure 7. Synthetic seismogram from GCB-1 plotted against seismic line SRS-1 and core data adjacent tothe Crackerneck Fault.


M-30

Figure 8. A portion of seismic line A-5, shown as representative of the seismic datadistribution shown in Figure 9.


M-31

Figure 9. Distribution of seismic data interpreted for this paper. Seismic lines SRS-1, SRS-7, and SRS-12are regional lines acquired in 1987. Lines CNF-1 and CNF-2 were obtained to help map the CrackerneckFault system for regional implications and are high resolution data. Lines A-1 through A-5 were obtainedto aid in mapping the subsurface associated with the M-Area plume remediation efforts and are also highresolution data. Line MCB-1 was acquired to evaluate the shallow structure and stratigraphy associatedwith the Miscellaneous Chemical Basin waste unit and to evaluate the use of very high resolutionacquisition with a vibroseis source. The shaded colors correspond with calculated depth to crystallinebasement, darker colors (blues ) are shallower and lighter colors (reds) are deeper. The pink linescrossing the seismic data are the projections of the faults.


M-32

1 20 40 60 80 100 120 140

20

40

60

80

100

120

UD

270 ms. (middle Middendorf)

1 20 40 60 80 100 120 140

20

40

60

80

100

120

UD

320 ms. (approximate top of saprolite)1 20 40 60 80 100 120 140

20

40

60

80

100

120

UD

1 20 40 60 80 100 120 140

20

40

60

80

100

120

UD

1 20 40 60 80 100 120 140

20

40

60

80

100

120

UD

1 20 40 60 80 100 120 140

20

40

60

80

100

120

UD

325 ms. (approximate top of crystalline rock)300 ms. (approximate top of Cape Fear)

310 ms. (basal Cape Fear)260 ms. (upper Middendorf)

N

Figure 10. Three dimensional seismic data. Time slices are shown as labeled. Red colors are area ofhigher amplitude, and blues are lower amplitude. Amplitude changes across the figure suggest eitherstructural or stratigraphic variations.


M-33

Figure 11. Subsurface map on top of the crystalline basement “rock” as determined from the seismicdata and the deep borings. Elevation are in meters relative to mean sea level. The grid is in UTMcoordinates. SRS roads are shown for location. The pattern of seismic data is shown by thick black lines(numbered with shotpoints where data has been interpreted). The Crackerneck Fault is shown as a northtrending solid thin black line, the MWESTA, Lost Lake and Steed Pond Faults are shown as NNEtrending dashed lines. Faults interpreted from the deepest stratigraphic investigation of Aadland (1997)are shown as dotted lines.


M-34

Figure 12. Subsurface map on top of the Cape Fear Formation as determined from the seismic data andthe deep borings.


M-35

Figure 13. Subsurface map on top of the Middendorf horizon as determined from the seismic data andthe deep borings.


M-36

Figure 14. Subsurface map on top of the McQueen Branch Confining Unit as determined from theseismic data and unit penetrating borings.


M-37

Figure 15. Subsurface map on top of the Steel Creek Formation as determined from the seismic data andunit penetrating borings.


M-38

Figure 16. Subsurface map on top of the Lang Syne/Sawdust Landing unconformity as determined fromthe seismic data and unit penetrating borings.


M-39

Foot Wall Syncline Hanging Wall Anticline

CNF

SENW

WH

LS/SDL

MqBCU

Midd.

CF

rock

400 m.

Figure 17. Top figure is fault propagation folding over a reverse fault. (Modified after Cumbest andWyatt, 1997). The lower figure is the actual seismic horizon schematic from line SRS-1 over the CNF. Thehorizons are labeled as; “rock” is crystalline basement, “CF” is the Cape Fear Fm., “Midd.” is theMiddendorf Fm., “MqBCU” is the McQueen Branch Confining Unit, “LS/SL” is the Lang Syne/SawdustLanding Fm., and “WH” is the Warley Hill Fm. The offset in basement rock is approximately 30-35 m.The distance from the anticline axis to the syncline axis is approximately 400 meters.


M-40

0

20

40

60

80

100

120

140

rock CF

Midd.

McQBCU SC

LS/S

LW

H

Surfac

e

Stratigraphic Horizon

met

ers

Maximum Relief

Average Sediment Thickness

Maximum Thickness

Minimum Thickness

Figure 18. Graph of structural relief, average, maximum and minimum sediment thickness for themapped stratigraphic horizons. “CF’ is Cape Fear, “Midd” is Middendorf, “McQBCU” is McQueenBranch Confining Unit, “SC” is Steel Creek, “LS/SL” is Lang Syne/Sawdust Landing, “WH” is WarleyHill.


M-41

Lagoon environment Continental slope

Abyssal plain(Deep marine environment)

Coastal Plain

Continental shelf(Shallow marine environment)

Prodelta

Delta slope

upperdeltaplain

Former shoreline

Fluvial environment

River

Distributary channels

Shoreline and Beach

lowerdeltaplain

Rev. 1GO1507

Extent of floodplain

UD

Ebb Tidal Delta

Middendorf

Black Creek

Cape Fear

Steel Creek

LS/SL

Congaree

Figure 19. Environments of deposition anticipated within the A/M Area. Figure is modified fromAadland (1997). Stratigraphic units, with the anticipated environments of deposition within the A/MArea are shown in red.


M-42

Notes


N-1

Use of Seismic Reflection Amplitude Versus Offset (AVO) Techniques toImage Dense Nonaqueous Phase Liquids (DNAPL) at the M-Area SeepageBasin, Savannah River Site, South Carolina

M. G. Waddell, W. J. Domoracki, T. J. Temples, Earth Sciences and Resources Institute, University of SouthCarolina, Columbia, SC 29208

ABSTRACT

One of the most difficult problems in designinga remediation plan for cleaning up DNAPLcontamination is locating the pools of free-phaseDNAPL. The seismic modeling of a DNAPLsaturated sand versus a water saturated sandsuggests that with frequencies of 120 Hz orhigher there would be an amplitude versus offset(AVO) anomaly. Seismic survey acquisitionparameters for imaging the DNAPL werederived from an AVO model, which was basedupon previously acquired seismic and well data. The seismic line was located in such a mannerthat it started in an area where theunconsolidated sand was water saturated andcrossed a known pool of free-phase DNAPLsaturated sand and continued back into a watersaturated sand. Using a weighted stackprocessing technique (Smith and Gidlow, 1987)a �fluid factor� stack indicated an anomaly at thedepth and location of the known free-phaseDNAPL plume. The initial results suggest thatunder certain conditions free-phase DNAPL canbe imaged using high-resolution seismicreflection techniques.

INTRODUCTION

Imperative to any DNAPL remediation effort isthe ability to locate high concentrations ofcontaminants. Traditional techniques such asborehole sampling run the risk of cross-contamination of aquifers. In addition, highconcentrations of DNAPL can be missedbecause of inadequate spatial sampling. Seismicreflection profiling is a geophysical techniquethat can be conducted to yield horizontalmeasurements every foot and verticalmeasurements every 3-5 feet depending on thesurvey design. The principal data collected froma seismic reflection survey are the arrival time

and amplitude of acoustic waves reflected fromsubsurface. The section can reveal faults, burialchannels, and subsurface lows that mightinfluence the migration of DNAPL in thesubsurface.

The amplitude and arrival time of seismic wavesis dependent on the elastic parameters of thesubsurface including bulk modulus, shearmodulus, density, Poisson ration, and angle ofthe incidence of the impinging energy. The lasttwo parameters can, in certain instances, be usedto infer the fluid content within the pore spaces.In the Petroleum industry seismic amplitudeversus offset (AVO) methods are used to directlydetect hydrocarbon bearing strata. In this studywe adapt some of these techniques to directlydetect the presence of DNAPLs at the UnitedStates of Department of Energy Savannah RiverSite (Waddell and Domoracki, 1997). If thesetechniques can be found to be applicable to othersites, remediation of DNAPLs can be facilitatedand be achieved at lower cost.

Savannah River Site

The Savannah River Site (SRS) is located inSouth Carolina on the South Carolina- Georgiaborder. During the Cold War SRS was a majorproduction facility of nuclear materials fordefense purposes. Since the late 1980�s anemphasis on environmental restoration has led tothe development of programs and strategies toremediate DNAPL spill at the site. One area ofSRS targeted for remediation is the M-Areaseepage basin.

The M-Area seepage basin was constructed in1958 to contain uranium wastes and residualsolvents produced from reactor fuel and targetdegreasing operations (Fig. 1). An estimatedtwo million pounds of residual solvents were


N-2

released into the eight million gallon unlinedsurface impoundment over a period of nearlythirty years. In 1988 the basin was closed,backfilled, and covered with an impermeablecap. Chlorinated solvents, including free-phaseconstituents, have been detected in thegroundwater near the seepage basin since 1981.The majority of the DNAPL found in thesubsurface is composed of Trichloroethylene(TCE), Tetrachloroethylene (PCE), andTrichloroethane (TCA) (Looney, 1992). Environmental remediation strategies haveincluded groundwater pump and treat, soil vaporextraction, in situ air stripping, and in situbiomediation.

The near-surface geology at the M-Area seepagebasin is comprised of Eocene age Upper CoastalPlain sedimentary units. On the surface is the�Upland unit� which consists of cobbles andcoarse sand ranging in thickness from zero tofifty feet. Underlying the �Upland unit� are theTobacco Road and Dry Branch formations whichare composed of fine to medium grain sand,clayey sand, and discontinuous layers of clay. The amount of clay in the formations decreaseswith depth. Directly below the Dry BranchFormation is the Warley Hill Formation rangesin thickness from 0 to 10 feet and occurs at adepth of approximately 155 feet below thesurface. This clay, when present, is consideredthe confining unit which separates the overlyingsurficial aquifer from the semi- to confinedaquifer below. The DNAPL in the M-Areapools on top of the �green clay�.

OBJECTIVES

The primary objective of this study was to testthe feasibility of using high-resolution seismictechniques and direct hydrocarbon indicatoranalyses to image free-phase and dissolved phaseDNAPLs at the M-Area seepage basin. Otherobjectives were to use the seismic data to mapthe subsurface geology and to determine thegeologic controls on the distribution of theDNAPL.

METHODOLOGY

The approach taken was thee fold consisting of1) evaluation of existing geological andgeophysical data concerning the amount anddistribution of DNAPL, 2) seismic modeling todetermine whether or not an AVO anomalywould be expected from DNAPL saturatedsediments, 3) acquisition and processing of aseismic line designed specially to image theDNAPL.

Figure 1. Location map of all the 2-D seismiclines acquired for AVO analysis. Well MSB-3is adjacent to well MSB-22.

Seismic Amplitude Techniques

In the 1960's petroleum companies recognizedthat in young sediments (Tertiary age) largeseismic amplitudes were associated with gassaturated sands. However, it was soon realizedthat not all large seismic amplitudes representedhydrocarbon saturated sands. The normalincidence (NI) reflectivity (bright spot)techniques involved three different scenariosbased upon a water saturated state and ahydrocarbon saturated state (for this discussiona sand/shale or sand/clay interface). Thescenarios are classified by changes in NIreflectivity from a water saturated sand to a gasor hydrocarbon bearing sand.

The three scenarios are:

• Dim-spot- a large positive amplitude that


N-3

is reduced to a smaller positive amplitude,

• Phase reversal- a small positive amplitudethat changes to small negative amplitude and

• Bright spot- a negative amplitude increasingto a large negative amplitude.

The dim-spot is generally associated with alarge acoustic impedance contrast and is atechnique for inferring lithology. Bright spotanomalies are generally used for interpretinglithology and estimating sand thickness. Phasereversal reflections are not generally reliablebecause geologic features (e.g. faults) can causethe reflections to appear to reverse phase. Thus,a reflection phase reversal does not necessarilyindicate a change in lithology (Verm andHilterman, 1995). The bright spot technique wasthe first direct hydrocarbon indicator.

In 1984 Ostrander published a article entitled"Plane-wave reflection coefficients for gas sandsat non-normal angles of incidence.� Ostranderobserved that the P wave reflection coefficient atthe interface between two media varies with theangle of incidence of the impinging energy.Ostrander also investigated the phenomenon ofcompressional wave reflection amplitudevariation with angle of incidence and changes inPoisson�s ratio. Poisson�s ratio is defined as theratio of transverse strain to longitudinal strainand can be expressed in terms of P wave and Swave velocity (Sheriff, 1991):

��

�

�

��

�

�−�

�

�

�

−=

12

2

2

2

s

p

s

p

VV

VV

σ (1)

σ = Poisson�s ratioVp= compressional wave velocityVs= shear wave velocity

Much of Ostrander�s work was based uponKoefoed�s 1955 work on determining thereflection coefficients of plane longitudinal

waves reflected at a boundary between twoelastic media. Koefoed observed that if therewere two elastic media with the top mediumhaving a smaller Poisson�s ratio than theunderlying medium there would be an increasein the reflection coefficient with increasing angleof incidence. He also observed that if Poisson�sratio of the lower medium was lower than theoverlying medium, the opposite would occurwith a decrease in the reflection coefficient withan increase in the angle of incidence. Anotherobservation showed that the relative change inreflection coefficient increases as the velocitycontrast between two media decreases. Koefoedalso observed that if Poisson�s ratio for twomedia was increased, but kept equal, thereflection coefficient at the larger angles ofincidence would also increase. Ostrander (1984)observed that changes in Poisson�s ratio causedby the presence of hydrocarbons in the porespace has dramatic effect on the P wavereflection coefficients and that these effects canbe related to seismic amplitude anomalies.

In order to understand amplitude versus offset(AVO) analysis, one has to understand offsetdependent reflectivity. Offset dependentreflectivity is defined as the variation of thereflection and transmission coefficients withchanging of the incident angle (Castagna andBackus, 1993). The offset refers to source toreceiver separation. Increasing source toreceiver separation results in increasing angle ofincidence for raypaths as measured from thenormal to a horizontal interface. Coincidentsource and receiver locations results in what istermed normal incidence, i.e. verticaltransmission and reflection from a horizontalinterface. The amplitude of the reflected andtransmitted waves are described by the reflectionand transmission coefficients. The followingdiscussion describes the equations and theoryused to determine amplitude for both thereflected and transmitted waves at an acousticboundary for normal incident energy and non-normal incident energy, i.e. offset dependentreflectivity. Note that this discussion is for asimplistic model of a single interface. Themajority of reflections observed on a seismicprofile are the superposition of events from


N-4

multiple layers and has a more complex AVOresponse.

As implemented in the Petroleum industry, AVOanalysis involves comparing modeled AVOresponses with field data to find a deviation froman expected background response. The expectedbackground response is usually taken to be awater saturated reservoir; hence, the deviationfrom the expected response is an indication ofhydrocarbon presence but does not indicatequantity. DNAPLs have the same acousticcharacteristics as liquid hydrocarbons; therefore,if DNAPL is present in free-phase and in largeenough quantities, similar types of analyzes asthose performed for the petroleum industry canbe applied to directly detect the presence ofDNAPL.

An understanding of reflection AVO techniquescan be obtained by a review of elastic wavepropagation. A P wave incident on a boundarybetween two linear elastic homogeneousisotropic (LEHI) media generates four types ofwaves: 1) transmitted P wave, 2) reflected Pwave, 3) reflected S wave, 4) a transmitted Swave (Figure 3). The amplitudes of the reflectedand transmitted waves at the boundary dependupon the P wave and S wave velocities. Thedensity of the two media and the angles ofincidence and refraction, are determined fromSnell�s Law.

Figure 2. Elastic waves generated at aboundary. A P wave incident at an angle θθθθon a boundary between two liner elastichomogenous isotopic (LEHI) materials

generates four wave types: reflected P,reflected S, transmitted P, and transmitted S.Snell’s law from optics governs the angles ofreflection and refraction. The P wavevelocity, density, and Poisson’s ratio describethe material properties of the media. The Swave velocity can be found from the P wavevelocity and Poisson’s ratio.

Snell�s Law,

V =

V =

V =

V = p

S

2

S

121

2121

φφ sinsinsinsin

ΡΡ

ΘΘ (2)

Vp1 = P wave velocity in medium 1,Vp2 = P wave velocity in medium 2,Vs1 = S wave velocity in medium 1,Vs2 = S wave velocity in medium 2,Θ1 = incident P wave angle,Θ2 = transmitted P wave angle,φ1 = reflected S wave angle,φ2 = transmitted S wave angle, andp is the ray parameter.

The reflection coefficient of the P wave as afunction of the incident angle, Rp (Θ), is definedas the ratio of the amplitude of the reflected Pwave to that of the incident P wave (Castagnaand Backus, 1993). The P wave transmissioncoefficient, Tp (Θ), is the ratio of the amplitudeof the transmitted P wave to that of the incidentP wave (Castagna and Backus, 1993). The Pwave reflection coefficient, Rp ,at normalincidence is given by the following equation:

II* *

21

II*

21 =

I + II - I = R

p

p

pa

p

pp

ppp

1

2

12

12 ln≈∆ (3)

Ip2 = acoustic impedance of medium 2 = ρ2Vp2ρ2 = density of medium 2Ip1 = acoustic impedance of medium 1 = ρ1Vp1ρ1 = density of medium 1Ipa = average acoustic impedance across the interface= (Ip2+Ip1)/2, and∆Ip = Ip2-Ip1.

The P-wave transmission coefficient at normalincidence Tp is given by:

Tp = 1-Rp. (4)The variation of the reflection and transmission


N-5

coefficients with incident angle and source toreceiver offset is referred to as offset dependentreflectivity (Castagna and Backus, 1993). Thevalues of the reflection and transmissioncoefficients for non-normal angles of incidenceare given by the Zoeppritz (1919) equations. Thecomplexity of the Zoeppritz equations has led tonumerous approximations to simplify thecalculations. Some of the approximations areBortfield (1961), Aki and Richards (1980), andShuey 1985. The Aki and Richards (1980)approximation to the Zoeppritz equations isgiven below:

)VV( c+)( b+)

VV( )_aR(

S

S

P

P ∆∆∆ρρθ (5)

θ = arcsin [ (VP2 /VP1 ) sinθincident]a = ½ + tan2 θb= 0.5-[(2VS

2 / VP2 ) sin2θ]

c = -( 4VS2 / VP

2 ) sin2θVP = (Vp1 + Vp2 ) / 2; Vp is P wave velocityVS= (Vs1 + Vs2 ) / 2; Vs is S wave velocityρ = (ρ1 + ρ1 ) / 2; ρ is bulk density∆VP = VP2 - VP1∆VS = VS2 - VS1∆ρ = ρ2 - ρ1

In this equation the reflection amplitude isexpressed in three terms containing P wavevelocity, S wave velocity, and bulk density.

The Aki and Richards approximation (Equation5) was rearranged by Shuey (1995) to obtain amore convenient form. The Shueyapproximation of the Zoeppritz equationsstresses the importance of Poisson�s ratio as theprimary determinant of the AVO response of areflection (Allen and Peddy, 1993). In this studywe used the Shuey approximation, whichexpresses angle dependent reflectivity in termsof P wave velocity, bulk density, and Poisson�sratio to compute the AVO models.

The Shuey (1985) formula for the reflectioncoefficient (amplitude) is:

θθθσσθ sintansin 22

p

p22OO0 -*

VV*

21+)*

-1+R*A(+R)R(

∆∆≈(6)

∆Vp = (Vp2 - Vp1 ) θ = (θ2 + θ1) / 2

Vp= ( Vp2 + Vp1 ) / 2 ∆σ = (σ2 + σ1)

∆Vs = ( Vs2 - Vs1 ) σ = (σ2 + σ1) / 2

Vs = ( Vs2 + Vs1) / 2

∆ρ = ( ρ2 - ρ1) ρ = ( ρ2 + ρ1 ) / 2

)+VV(

21

Rp

po ρ

ρ∆∆≈ R

))-(1

1(+A=A0

20σ

σ∆

)-12-1( B)+(1 2-B=A0 σσ

ρρ∆∆

∆

+ VV

VV

=B

p

p

p

P

Vp1= P wave velocity of layer oneVp2= P wave velocity of layer twoVs1= S wave velocity of layer oneVs2= S wave velocity of layer twoρ1 = density of layer oneρ2 = density of layer twoσ1 = Poisson ratio of layer oneσ2 = Poisson ratio of layer twoθ1 = incidence and transmission angle layer one.θ2 = incidence and transmission angle layer two.R0 = Normal incidence (NI), i.e. reflection

coefficient for zero-offset.B = is the AVO gradient.A0= is the normal incidence amplitude.

The approximation above can be simplifiedfurther. If the average Poisson�s ratio is assumedto be 1/3 and higher order terms that areinsignificant above 30� are dropped, the formulafor the angle dependent reflection coefficientbecomes:

θσθ sin2NI)-49(+NI)R( ∆≈ (7)

θsin2B+NI≈ (8)

θθθ sin22 PR+NIcos)R( ≈ (9)NI and PR are defined as


N-6

)z+z(zSUB1)-z(=

)V(+)V()V(-)V(

=NI12

2

1p2p

1p2p

12

12

ρρρρ z is the acoustic impedance: (10)

)-(1=PR 2

avgσσ∆ (11)

(Verm and Hilterman, 1995). Reflection AVOcan be thought of a combination of normalincidence reflectivity and a far offset reflectivity,or �Poisson reflectivity,� that arises primarily asa result of changes in the Poisson�s ratio betweenmedia.

To implement AVO analysis, the NI and PRcoefficients are extracted from a common depthpoint (CDP) gather. Initially the CDP gather istransformed from a function of offset to incidentangle. This transformation requires knowledgeof the root mean square (RMS) velocity, whichis ideally obtained from borehole sonic logs. Inthe absence of sonic logs a rough estimate of theRMS velocity can be obtained from the normalmoveout (NMO) stacking velocities. Next, theNI and PR coefficients are found by fitting eitherEquation (8) or (9) to the amplitudes.

Modeling

The most important aspect to this study is theAVO modeling, which was used to design thefield acquisition parameters for seismic profilesM-1, M-2, and M-3. The first type of model wasgenerated using the Aki-Richards (1980) andSmith and Gidlow (1987) approximations to theZoeppritz equations. The model is a sand wedgesaturated with either water or TCE overlyingeither a clay layer or a water saturated sand

(Table 1).

Using the parameters in Table 1, the results ofthe modeling suggested that there would be anAVO effect caused by the presence of DNAPL(Figures 3,4, 5). Furthermore, these resultsindicated that changes in the reflectioncoefficient would begin to occur atapproximately 220 angle of incidence. Using thisinformation the seismic lines were designed suchthat half the receivers would be under thisincident angle and half would be over.

The second type of model was generated usingthe P wave, S wave, and density from an existingwell. For this type of modeling a synthetic CDPgather was generating using the velocity anddensity data from the well is shown in Figure 6(left). To investigate the AVO response causedby the presence of DNAPL a ten foot interval ofthe logs corresponding to the depth where theDNAPL is accumulating was replaced with thesame velocity and density parameters for TCEused to generate the models shown in Figure 3,4, and 5. Also, the model was convolved with a120 Hz Ricker wavelet to simulate typical high-resolution seismic data. Figure 6 (right) is themodeled CDP gather for a TCE contaminatedsand layer.

Comparison between the synthetic CDP gathermodels demonstrates that, in this case, an AVOeffect can be expected due to the presence ofTCE. This AVO effect is manifested as a largeincrease in amplitude with increasing shot toreceiver offset. The increase in amplitude ismuch larger than if no TCE were present

TABLE 1. Parameters Used to Generate AVO Models

Lithology Vp Vs Densityft/s ft/s g/cc

WedgeWater Sand 5800 1450 1.9TCE Sand 4968 1634 2.07SubstrateClay 5600 1300 1.85Water Sand 5600 1460 1.89


N-7

Figure 3. Graph of reflection coefficient versus angle of incidence using the Aki-Richards(solid line) and Smith and Gidlow (dashed line) approximations to the Zoeppritz equationsfor water saturated sand overlying the “green clay”. At the sand/clay interface thereflection coefficient is slightly negative and becomes more negative with increasing angleof incidence. At large angles of incidence (greater than 50º) this effect is pronounced.

Figure 4. Graph of reflection coefficient versus angle of incidence using the Aki-Richards (solidline) and Smith and Gidlow (dashed line) approximations to the Zoeppritz equations for TCEsaturated sand overlying the “green clay”. In this scenario, the amplitude is slightly negative, butat 22º the reflection coefficient goes from slightly negative to positive and increases with angle ofincidence.


N-8

Figure 5. Graph of reflection coefficient versus angle of incidence using the Aki-Richards (solidline) and Smith and Gidlow (dashed line) approximations to the Zoeppritz equations for watersaturated sand overlying TCE saturated sand. The reflection coefficient at 0 angle of incidence isnear 0º, or is slightly negative, and at 22º becomes increasingly negative.

Figure 6. Synthetic CDP gather generated using P wave and S wave velocity logs from wellMHM-21 and a density log from nearby well MHT-1. Notice that the reflection at 120 ms has aslight increase in amplitude with increasing offset. The synthetic seismogram on the right is thesame as on the left, but the P wave, S wave, and density values have been replaced using theparameters as in Figures 4-5 to simulate a 10 foot section of TCE saturated sediment. Notice thatthe reflection at 120 ms has a much greater increase in amplitude with offset as a consequence ofthe presence of TCE.


N-9

Seismic Data Acquisition

The original project consisted of recording oneseismic profile, line M-1, across a knownDNAPL plume that has a free-phase component(Fig. 1). However, after processing M-1 andpreliminary AVO analysis (fluid factor stack)was preformed, an AVO anomaly was detectedat the location and depth of the known DNAPLplume. As a result, it was decided to acquire twoadditional lines M-2, and M-3 and a verticalseismic profile (VSP) at well MSB-9A (Fig. 1).

All of the data were collected with a 24 bit OYODAS-1 seismograph recording either 48 channels(M-1) or 96 channels (M-2 and M-3). The M-Area is a notorious low signal-to-noise seismicdata area. Two test lines were shot prior to theacquisition of the M-series lines using mini-vibrator, downhole Seisguntm, and EWG-1

weight drop sources. None of these sources wereable to combine the requirements of adequatesignal penetration, high frequency content, andminimal source generated noise. Lastly, a 10 lb.Sledgehammer was tested. The sledgehammerwas found to be a repeatable high frequencysource that generated a relatively small surfacewave.

The acquisition parameters for seismic line M-1appear in Table 2. The parameters such as groupspacing , near offset, and far offset, were basedupon the modeling results. The target intervalwas the �green clay� aquitard at approximately155 feet depth. The recording offsets werechosen in such a way as to have at least 15 to 20geophone groups under 220 angle of incidence sothat the AVO analysis described above could bepreformed on the seismic lines.

TABLE 2. Acquisition Parameters for Seismic Line M-1

Number of Channels 48Group interval 2ft.Shot interval 2 ft.Near offset 20 ft.Far offset 114 ft.Nominal CDP fold 24Geophone frequency 40 HzEnergy source Hammer /8 hitsSample rate 0.25msRecord length 500 ms

After line M-1 was processed, modificationswere made to the basic recording parameters foracquisition of lines M-2 and M-3. For thoselines, a 96 channel recording system was usedwhich allowed a greater range of offsets to berecorded. For lines M-2 and M-3 the minimumand maximum offsets were 29 and 219 feetrespectively.

Seismic Data Processing

The seismic data were processed with standardCDP data processing sequence (Yilmaz, 1987)that included frequency-wave number filtering(F-K or pie slice filtering) to eliminate linearnoise trains, spiking deconvolution, and iterativevelocity analysis and residual statics application. For display, the data were filtered to a 90-275Hz passband and five point running mean was


N-10

applied to enhance the lateral continuity ofreflections. To analyze AVO variations, nearand far offset stacks were generated as well asSmith and Gidlow (1997) fluid factor stack.

DIRECT DETECTION OF DNAPL

At the M-Area seepage basin two primary typesof DNAPL are present below the water table,trichloroethylene (TCE) and tetrachloroethylene(PCE). The latter being the predominate type ofsolvent. Water samples taken from well MSB-3D (Fig. 1) consisted of a separate phase liquidcomposed of PCE with a subordinate amount ofTCE (Looney, 1992).

In this project two methods were used toinvestigate any AVO effects caused by thepresence of DNAPL. As previously mentioned,the recording geometry was such that 15 to 20geophone groups were under 220 angle ofincidence. If the models were correct, therewould not be any significant change in amplitudeunder 220 . Method one was ranged limitedstacking. In this method data were gathered andstacked using subsets of the range of offsets toproduce a near offset section (Fig. 7 top) and faroffset section (Fig. 7 middle). AVO anomaliesproduce by the presence of DNAPL shouldshow as high amplitudes present on the far offsetsection , but not on the near offset section. Thesecond AVO analysis technique used was theSmith and Gidlow �fluid factor� stack (Smithand Gidlow, 1987) (Fig. 7 bottom).

The �fluid factor� stack is derived from the Akiand Richards (1980) approximation of Zoeppritzequations and the Castagna, Batzle andEastwood (1985) �mudrock line� for 100%water saturated clastic silicate rocks. In thismethod a series of time and space variantweighting factors are applied to the CDP gatherafter NMO (normal moveout) corrections. If themodel is valid, the CDP stack will be zero for100% water saturated clastic sediments. Any

residual reflections should denote sedimentssaturated with fluids other than water. In thisproject, DNAPL would be the only fluid otherthan water to saturate or partially saturate thesediments.

Profile M-1 Offset Range Limited Stack

In Figure 7 the upper profile is a near offsetstack and the middle profile is a far offset stack. If there are any AVO anomalies, they should bepresent on the far offset stack and absent on thenear offset stack. At shot point (SP) 79 at about89 ms is an anomaly that was drilled ( wellMOX-6) and TCE was found to be present. Onthe near offset stack the anomaly is absent. Another anomaly is shot point 297 at about 109ms. This anomaly is adjacent to well MSB-3which had free phase DNAPL. Anotheranomaly is at shot point 430 at about 110 mswhich is believed to be DNAPL, but is as yetuntested.

Profile M-1 Smith-Gidlow (Fluid Factor)Stack

Contained in Figure 7 (bottom) is a fluid factorstack based upon the Smith and Gidlow (1987)weighted stack technique. The amplitudeanomalies observed on the far offset stack arepresent. The anomaly observed on the far offsetrange limited stack at shot point 79 at 89 ms ispresent as is the other anomalies at shot points297 at 109 ms and shot point 430 at 110 ms. However, it must be pointed out that anyanomaly above 100 ms must be treated withcaution, because the fluid factor analysisassumes 100% water saturation. The water tableat the M-Area seepage basin occurs atapproximately 100 ms; therefore, any anomalyabove 100 ms in not at 100% water saturationunless it is perch water.


N-11

Figure 7. Offset range limited stacks and fluid factor stack for profile M-1. Near offset section(top), far offset (middle), and fluid factor (bottom). The middle section was generated by stackingoffsets greater than 58 feet. High amplitudes that occur only at far offsets should denote presenceof DNAPL. Every second trace is plotted. CDP spacing is one foot. The bottom section is a fluidfactor stack of M-1. The water table occurs at a depth corresponding to approximately 100 mstime. The amplitude envelope (magnitude of Hilbert transform) is displayed.

CONCLUSIONS

At the M-Area seepage basin, Savannah RiverSite it appears that DNAPLs can be imaged inthe subsurface using high-resolution seismicdata. Two wells were drilled on anomaliesrecognized by the seismic data and DNAPLswere found at the predicted depth in parts permillion (head space data).

Caution must be exercised in applying thistechnique to other areas. Before any seismic

data are acquired some basic modeling has to bedone. The modeling will determine what is theminimum amount of DNAPL that can be imagedgiven the geologic conditions for a particular siteand provide the necessary data for designingseismic acquisition parameters.

ACKNOWLEDGMENTS

Funding for this project was provided by U.S.Department of Energy, Savannah River Site. Wewould also like to thank Mike Graul at Texseis,Inc. of providing the initial modeling data. We


N-12

also thank Landmark Graphics Corporation fordonating the ProMAX software, Hampson-Russel Software Services Ltd. for donating theirAVO software, and Seismic Micro Technologyfor donating their KINGDOM software.

REFERENCES

Aki, K. and Richards, P.G., 1980, QuantitativeSeismology: W.H. Freeman and Co., SanFrancisco, 932 p.

Bortfield, R., 1961, Approximation to the reflectionand transmission coefficients of planelongitudinal and transverse waves:Geophysical Prospecting, v. 9, p. 485-502.

Castagna, J.P. and Backus, M. M., eds., 1993,Offset-dependent reflectivity - Theory andpractice: SEG Investigations inGeophysics No. 8, Society of ExplorationGeophysicists, Tulsa, OK, 345 p.

Castagna, J.P., Batzle, M.L., and Eastwood, R.L.,1985, Relationships betweencompressional-wave and shear-wavevelocities in clastic silicate rocks:Geophysics, v. 50, p. 571-581.

Gardner, G.H.F., Gardner, L.W., and Gregory,A.R., 1974, Formation velocity anddensity-the diagnostic basics forstratigraphic traps: Geophysics, v. 39, p.770-780.

Koefoed, O., 1955, On the effect of Poisson�sratios of rock strata on the reflectioncoefficients of plane waves: GeophysicalProspecting, p. 381-387.

Looney, B.B., 1992 �Assessing DNAPLContamination in A/M Area, SRS: Phase1,� WSRC-RP-92_1302, Prepared for theU.S. Department of Energy underContract No. DE-AC09-89SR18035, p 1-85.

Mavko, G., Mukerji, T. and Dvorkin, J., 1998: TheRock Physics Handbook: CambridgeUniversity Press, p 1-220.

Ostrander, W.J., 1984, Plane-wave reflectioncoefficients for gas sands at non-normalangles of incidence: Geophysics, v. 49, p.1637-1648.

Rutherford, S.S. and Williams, R.H., 1989,Amplitude-versus-offset variations in gassands: Geophysics, v. 54, p. 680-688.

Sheriff, R, 1991, Encyclopedic Dictionary ofExploration Geophysics: SEGGeophysical Reference Series No. 1,Society of Exploration Geophysicists,Tulsa, OK, 376 p.

Shuey, R.T., 1985, A simplification of theZoeppritz equations: Geophysics, v. 50, p.609-614.

Smith, G.C. and Gidlow, P.M., 1987, Weightedstacking for rock property estimation anddetection of gas: Geophysical Prospecting,v. 35, p. 993-1014.

Verm, R. and Hilterman, F., 1995, Lithology color-coded seismic sections: The calibration ofAVO crossplotting to rock properties: TheLeading Edge, v. 14, n. 8, p. 847-853.

Waddell, M.G, Temples, T.J., and Domoracki,W.J., 1997, Using high-resolutionreflection seismic to image free-phaseDNAPLs at the M-area, Savannah RiverSite (Abstract): Ann. Mtg. Am. Assoc.Petroleum Geologists, Dallas, TX.

Waters, K.H., 1981, Reflection seismology: A toolfor energy resource exploration: JohnWiley and Sons, New York, 453 p.

Zoeppritz, K., 1919, Uber reflexion und durchgangseismischer wellen durchUnstetigkerlsflaschen: Berlin, UberErdbebenwellen VII B, Nachrichten derKoniglichen Gesellschaft derWissenschaften zu Gottingen, math-phys.K1., p. 57-84.


O-1

Geology: Improving Environmental Cleanup of the A/M Area, SavannahRiver Site

Mary K. Harris1, D. G. Jackson1, Brian B. Looney1, Andrew D. Smits2, and R. S. VanPelt3

1Savannah River Technology Center, Aiken SC2Science Applications International Corporation, Augusta GA3Environmental Restoration Division, Savannah River Site, Aiken SC

Introduction

Successful environmental cleanup programsrely on an interdisciplinary team of scientists– geologists, engineers, chemists,mathematicians, and others. The solutionsdeveloped by the team are based on focusedenvironmental characterization followed byselecting and deploying clean-up technologiesthat are matched to the problem. Eachtechnological advance is grounded in a clearlystated conceptual model that is continuallydeveloped and refined using the scientificmethod. Successful technologies always obeynatural laws, and often rely on naturalprocesses or capabilities. In the work,geology plays a particularly important role.Stratigraphy and hydrostratigraphy controlboth the migration of contaminant plumes andthe performance of cleanup technologies.Understanding depositional processes andpost-depositional alterations – erosionalfeatures and the like – are critical to properselection and optimization and ultimatesuccess of cleanup. Cleanup of the A/M Areaof the Savannah River Site exemplifies boththe interdisciplinary nature of the work, aswell as the pivotal role of geology.

Objective

The primary goal of geological support in theA/M Area is to assist in safe and effectiveoperations and in environmental stewardship.This overall objective requires comprehensivegeological activities including both basic andapplied studies. These include a wide varietyof invasive, noninvasive and datainterpretation tools. In the sections, below,we introduce the general philosophy for the

A/M Area environmental clean-up program,present a brief background of A/M Area,provide a summary of the local geologicconditions, and provide detail from a fewspecific A/M area geology studies. In theconclusions, we relate the various geologicalfindings to specific clean-up technology andimplementation decisions within the program.

Anatomy of a Contaminated Site

Figure 1 depicts a conceptual diagram of acontaminated site that has impacted itssurroundings – in this case, the underlying soiland groundwater. The three ovals – thesource zone, the primary contaminant plume,and the dilute fringe – represent differentportions of the impacted environment thateach have a different character. The sourcezone contains significant contamination inconcentrated and hazardous forms. Thesource zone can contain materials such asundissolved organic liquids (oils, fuels orsolvent), strong acids or bases, high levels ofradiation, and/or toxic chemicals or elements.The second oval, the primary contaminantplume, is comprised of contaminatedgroundwater or vapor that carries pollutants atlower levels, but levels that still represent apotentially significant present or futurehazard. The third oval, the dilute fringe,contains contamination at relatively lowconcentrations, but in large volumes of water.

Efficient and effective environmental clean uprequires matching the character of the clean-up and stabilization methods to the characterof the target zone of contamination. Thus,aggressive and relatively expensive methodsare often appropriate for the source zone,


O-2

classical pump-and-treat methods are oftengood for the primary contamination zone, andvarious methods based on natural processesare often best for the dilute fringe. Figure 1identifies several example technologies thatare appropriate for each of the zones.

Figure 2 extends this conceptual model byidentifying the cost basis for the typical cleanup technologies. In the source zone,stabilization and removal methods arenormally priced in terms of volume of soil oramount of contaminant in the treatment zone($ per cubic yard treated, $ per pound and thelike). The reference source zone technologiesrequire aggressive access and subsequent useof targeted energy or chemical reagents. It isclear that in the source zone it is important tocharacterize the site in such a way that theprecise location of the source zone isdelineated as carefully as possible as increasesin treatment volume result in additionalremediation costs. Detailed sourcecharacterization reduces costs by focusingenergy or reagent to areas where they areneeded. Equally important, however, is adesire to minimize any undesired negativeimpacts (wasting energy, harmingmicrobiological populations, etc.) associatedwith using aggressive remedies on regionswithout source level contamination.

In the primary contaminant plume, treatmenttechnologies are normally priced in terms ofthe amount of water (or vapor) treated ($ pergallon and the like). Thus, the goal ofcharacterization in this portion of the plume isto define the flow directions and generalplume structure to allow the most contaminantto be treated in the fewest “gallons”. Figure 3illustrates an important-final extension to oursimplified conceptual model. This diagram ofthe primary contaminant plume in A/M Areashows that contamination moves in responseto many factors – contaminant release locationand type, geology, sources and discharges ofwater, and others. The resulting contaminatedsoil and groundwater zone occupies acomplicated three-dimensional shape ratherthan the simple ovals that we began with.

This complexity must be recognized whendeveloping and implementing technologies forboth characterization and clean up of theprimary contaminant plume.

The dilute fringe contains low concentrationsof contamination in large volumes of water.Thus, the best technologies for this zone arethose that are priced in terms of time ($ / yearand the like). To be successful, thesetechnologies must rely on natural-sustainable-measurable processes. This class oftechnology has gained recent regulatorysupport under the terminology “monitorednatural attenuation”. For the dilute fringe,technology selection is biased towardunderstanding the contaminant destructionand stabilization capabilities of native speciesand natural populations. A second step isidentifying engineering interventions, ifneeded, to maximize the performance and toassure that the attenuation process will operatefor extended periods. As a consequence,logical and cost-effective monitoringstrategies are critical to the success of “naturalattenuation” technologies.

The three zones depicted in Figure 1 arepresent at contaminated sites of all sizes. At a“mom-and-pop” gas station, the entirecontaminated zone – all three ovals – mightoccupy a portion of a city block. At a largeindustrial facility like the A/M Area at SRS,the contaminated zone can extend over a fewsquare miles. The size of a problem impactshow distinct the actions to address thedifferent zones need to be. Time is also afactor. Concentrations change, as cleanupprogresses, so that dilute fringe technologiesbecome appropriate for polishing areas thatwere formerly at higher concentrations.

General Background on A/M Area

The Savannah River Site (SRS) is a U.S.Department of Energy (DOE) facility that wasset aside in 1950 as a controlled area forproduction of nuclear materials for nationaldefense. This report focuses on the A/M Areaat the north boundary of the SRS (Figure 4).


O-3

This area includes the facilities which wereused for fabrication of reactor fuel and targetassemblies (M Area), laboratory facilities(SRTC), and administrative support facilities(A Area). Operations within the A/M Arearesulted in the release of chlorinated solventsto the subsurface (Marine and Bledsoe, 1984).Released solvents include trichlorethylene(TCE), tetrachloroethylene (PCE) and 1,1,1-trichloroethane (TCA) which havecontaminated the soil and groundwater withinthe area. Since the detection of these volatileorganic compounds (VOCs), the SRS haspursued a program of soil and groundwaterremediation.

The SRS remediation program integratestechnologies to address various portions of thecontaminant plume conceptually presentedearlier. Within the immediate source zone aseries of soil-vapor extraction units arecombined with enhanced sourceremoval/destruction methods, and anextensive network of groundwater recoverywells. This combination provides for theremoval of contaminants in the vadose zoneand reduces additional migration in thegroundwater. The extensive groundwaterpump & treat system also provides for theextraction and treatment of contaminatedgroundwater in the primary contaminant zone.A series of vertical recirculation wells treatsgroundwater before it enters the dilute fringe.SRS maintains an on-site research anddemonstration program for developinginnovative remediation techniques andmaximizing the efficiency and effectiveness ofthe SRS remediation strategy. The SRSremediation strategy includes on-goingcharacterization of subsurface features andconditions that may have a significantinfluence on groundwater flow andcontaminant transport.

Recent characterization efforts emphasizelocating dense non-aqueous phase liquid(DNAPL) in the subsurface beneath A/M Area(Looney and others, 1992). Upon release atthe surface, DNAPL migrates down throughthe vadose zone into the saturated zone. The

relatively high density of the DNAPL causesit to move in an overall vertical direction,following a “path of least resistance” throughzones of highly permeable gravel and coarse-grained sand. The DNAPLs tend to slow andspread laterally upon encountering stratacomposed of less-permeable silty sand andclay (Jackson and others, 1996). Potentialmigration paths and distribution of DNAPLare therefore related to the distribution andconfiguration of less-permeable, fine-grainedsediment.

Once emplaced within the saturated zone,DNAPLs serve as a source for the release ofdissolved VOCs to the groundwater. Thegroundwater carries dissolved VOCs asplumes, which emanate from the DNAPLsource and extend in a directioncorresponding to the local hydraulic gradient.Movement of VOC-contaminatedgroundwater is largely controlled by thegeometry and distribution of more permeablestrata, the nature and extent of the DNAPLsource, and the overall dynamics of waterrecharge and discharge from the subsurfacesystem.

The distribution of fine-grained sedimentlayers and variations in the quantity of clayand silt within the sediment is extremelyimportant in order to fully characterize theextent of DNAPL layers beneath the A/MArea and determine the geometry of thecontaminant plumes emanating from them.Indirect assessment techniques have beenapplied to historical groundwaterconcentrations in the A/M Area and indicatethat the distribution of DNAPL beneath A/MArea may be laterally extensive (Jackson andothers, 1996). The delineation of layers ofless-permeable sediment is therefore critical toaid remediation efforts. Detailed mapping ofthese geologic features can aid in locatingareas where contamination is most likely tohave migrated into the saturated zone. Thesedata are valuable aids for accurate modelingof groundwater flow and contaminanttransport. In addition, this information can beused to refine the current remediation systems


O-4

or assist in designing new ones. The results ofa recent geospatial mapping study arepresented below to exemplify the work inA/M Area.

A/M Area Geological Setting

The SRS comprises approximately 300 squaremiles within Aiken, Barnwell, and Allendalecounties in southwestern South Carolina. Thecenter of the SRS is 22.5 miles southeast ofAugusta, Georgia, approximately 100 milesfrom the Atlantic Coast within the UpperAtlantic Coastal Plain Physiographic Province(Figure 4). The Savannah River forms thesouthwest boundary of the SRS. The SRS lieson the Aiken Plateau of the Atlantic CoastalPlain at an average elevation of 300 feet abovemean sea level (ft msl). The Aiken Plateau iswell drained, although many poorly drainedsinks and depressions exist, especially inupland areas. Overall, the Aiken Plateaudisplays highly dissected topography,characterized by broad inter-fluvial areasseparated by narrow, steep-sided valleys.Local relief can attain 280 feet (Siple, 1967).The A/M Area comprises approximately eightsquare miles near the northern boundary of theSRS (Figures 4 & 5).

LithostratigraphyThe SRS is underlain by sediment of theAtlantic Coastal Plain. The Atlantic CoastalPlain consists of a southeast-dipping wedge ofunconsolidated and semi-consolidatedsediment that extends from its contact with thePiedmont Province at the Fall Line to the edgeof the continental shelf. These deposits consistof sediment that is deltaic and near-shoremarine in origin, with evidence ofconsiderable fluvial influence (Fallaw andPrice, 1995). The sediment consists ofinterbedded sand, muddy sand, and mud (clayand silt), with a subordinate amount ofcalcareous sediment. Strata range from LateCretaceous to Miocene in age and restunconformably on crystalline and sedimentarybasement rock (Figure 6). The A/M Area lieswithin the up-dip area of the coastal plaindeposits, approximately 20 miles from the Fall

Line. Beneath the A/M Area, Cretaceous-aged strata rest on crystalline basement rock ata depth of 500 to 700 feet below the landsurface (ft bls). For a detailed discussion ofthe lithostratigraphy the reader is referred toWyatt, 2000 (this volume). This paper willconcentrate on the hydrostratigraphy of theA/M Area.

HydrostratigraphyThe hydrostratigraphy of the SRS has been thesubject of several different systems ofhydrostratigraphic classification. This reportincorporates the hydrostratigraphicnomenclature currently established for theSRS region and A/M Area by Aadland andothers (1995a, 1995b). This study focuses onthe up-dip part of the Floridan-Midvilleaquifer system as defined for A/M Area byAadland and others (1995b). Strata withinunits of the Floridan-Midville aquifer systemthat exhibit internally consistent hydrauliccharacteristics and which, on a local scale,behave as distinct hydrostratigraphic units, aredelineated as informal aquifer and confiningzones (Aadland and others, 1995b). Figure 6correlates the nomenclature of Aadland andothers (1995b) with the lithostratigraphy ofFallaw and Price (1995).

The hydrostratigraphy of the A/M Areaconsists of three aquifers within the Floridan-Midville aquifer system divided by oneconfining unit and one confining system. Inthe vicinity of SRS, this aquifer systemincludes, the McQueen Branch aquifer,overlain by the Crouch Branch aquifer, andthe Steed Pond aquifer. The Crouch Branchaquifer is the principal water-producingaquifer at SRS and is the deepest unit that isroutinely sampled by the current monitoringwell system. Since the principle componentsof large scale contaminant migration in theA/M Area are dissolution of source (DNAPL)solvents followed by dissolved transportwithin the shallow aquifers overlying theCrouch Branch aquifer, the informationselected for presentation below addressescritical features of the Steed Pond aquifer and


O-5

the nature of the aquitard between the SteedPond aquifer and Crouch Branch aquifer.

Within the A/M Area, the flow of shallowgroundwater is controlled by numerous localand regional parameters. These parametersinclude the influence of the Savannah River tothe west of the A/M Area, the facies changeswithin the Crouch Branch confining zone inthe northeast portion of the study area, thetopographic relief that creates a strongdownward gradient through out the entireA/M Area, and the incision of locallyimportant confining zones due to erosion inthe southern portion of the study area.

The majority of the groundwatercontamination in the A/M Area is located inthe Steed Pond aquifer and is the target ofcurrent remediation activities. This aquifercontains the water-table unit. In the CentralA/M Area the Steed Pond aquifer is dividedinto the M-Area and the Lost Lake aquiferzones, which are separated by the Green Clayconfining zone. Towards the north, in thevicinity of the SRTC complex, these units arenot well distinguished and the unit is treatedas one aquifer unit. The elevated topography,location of drainage features, and thenumerous interbedded layers of sands, siltsand clays combine to create a predominatelydownward flow direction within the aquifer.In the southern region of the study area the M-Area and Lost Lake aquifer zones becomedistinct hydrologic units due to increasedthickness of clays within the Green Clayconfining zone. The groundwater flow isstrongly downward in the M-Area aquiferzone in this region. Upon entering the LostLake aquifer zone the flow transitions to thelateral direction and is controlled by regionaldischarges to surface streams towards both thesouth and the southeast. At the extremesoutheastern portion of the study area, erosionof the Green Clay confining unit has resultedin the outcropping of the contaminant plumecontained in the Lost Lake aquifer zone.

Lithofacies Mud Mapping of CriticalIntervals in A/M Area

Geologic drill core descriptions andgeophysical logs (caliper, gamma-ray andresistivity) for 42 cores were utilized in thisstudy to delineate hydrostratigraphicboundaries beneath the study area (Figure 5).The drill core descriptions are detailed andprovide a foot-by-foot microscopic descriptionwhich includes percent mud, gravel, and sandalong with estimated porosity and sortingcharacteristics. Altitude contour maps,isopach maps, and lithofacies maps wereproduced for each hydrostratigraphic unitusing EarthVision® software (DynamicGraphics Inc., Alameda CA, www.dgi.com).Lithofacies maps were created to depict thelateral distribution of mud and the internalvariation of the mud content within the unitusing the geometric mean and standarddeviation of the parameter. The maps of thegeometric mean are interpreted as an averageexpression the overall mud content within theunit while the standard deviation provides anindication of the degree of vertical lithologicvariation, such as interbedding of sedimenttypes within the unit. Both of theselithofacies parameters are functions of thethickness of the unit. Isopach maps andgeometric mean mud content maps forselected hydrostratigraphic units are presentedin this paper. The Crouch Branch confiningunit and the Green Clay confining zone willbe discussed in detail. For discussion of theother hydrostratigraphic units the reader isreferred to Smits and others, 1998, and Parkerand others, 1999.


The Crouch Branch confining unit separatesthe Crouch Branch aquifer from the SteedPond aquifer (Figure 6). The Crouch Branchconfining unit may be divided into threehydrogeologic zones beneath the A/M Area(Aadland and others, 1995b). These zones, inascending order, are the “lower clay”confining zone, the “middle sand” aquiferzone, and the “upper clay” confining zone. As


O-6

described below, the “upper clay” confiningzone thins and disappears toward the northernedge of A/M Area. In this region, the “middlesand” aquifer zone coalesces with theoverlying “Lost Lake” aquifer zone of theSteed Pond aquifer. In the western part of theA/M Area, these zones are not delineated dueto the absence of the sand beds referred to asthe “middle sand” aquifer zone. In thenortheastern portion of the area the clay bedsof the “upper clay” and “lower clay” zone ofthe unit are thin or absent. It is in this regionwhere groundwater flow occurs between theoverlying Steed Pond aquifer and theunderlying Crouch Branch aquifer (Aadlandand others, 1995b). The interpretationoutlined below is supported by a recentgeophysical study that creatively combinedthe results of time domain electromagnetic(TDEM) soundings and shallow seismicreflection surveys to characterize the CrouchBranch confining unit (Eddy-Dilek and others,1997). The geophysical study wasparticularly successful in noninvasivelycorroborating both the position andlithological character of the Crouch Branchconfining unit (e.g., areas where the CrouchBranch confining unit is dominated by sand).

“Lower” Interval of the Crouch Branchconfining unitThe surface of the “lower” interval of theCrouch Branch confining unit ranges from119 to 75 ft msl and dips to the south-southeast. The isopach map of this intervalindicates a variation from 17 to 60 ft inthickness (Figure 7). The “lower” interval ofthe Crouch Branch confining unit is generallythicker in areas that correspond with channel-like features on the top of the Crouch Branchaquifer. These variations are attributed todepositional fluvial facies changes such aschannel and point bar deposits.

Figure 8 illustrates the geometric mean of themud percentage within the “lower” interval ofthe Crouch Branch confining unit. Thegeometric mean of the mud percentage isgenerally high which corresponds to thickintervals, ranging from 75% in the center of

the study area to 90-95% in the southern part.In contrast, this interval of the Crouch Branchconfining unit contains relatively lowquantities of mud beneath the northeast cornerof the study area, corresponding to thechannel-like feature beneath A Area. Thisinterpretation is further supported by smallstandard deviations in this area whichrepresent clean well-sorted sands indicative offluvial deposition. In contrast, the areas oflow mud percentages with higher standarddeviations are interpreted to representinterbedded sands and clays of fluvio-deltaicorigin.

“Middle” Interval of the Crouch Branchconfining unitThe surface of the “middle” interval of theCrouch Branch confining unit ranges from105 to 140 ft msl and varies in thickness from46 ft to 3 ft (Figure 9). The isopach mapindicates that thick parts of this intervalgenerally correspond with channel-likefeatures on the top of the “lower” interval.The mean mud fraction varies from less than5% to greater than 60%. The larger mudpercentages generally correlate with thethinner parts of the interval, especially in thesouthern sector of the A/M Area and in thewestern part of the study area (Figures 9 and10). The standard deviation of the mudfraction within the “middle” interval is higherin areas where the mean mud fraction ishigher. These fine-grained muddy sedimentsof the "middle” interval of the Crouch Branchconfining unit are interpreted to representinterbedded layers of sand and mud that arenot laterally extensive.

“Upper” Interval of the Crouch Branchconfining unitThe surface of the “upper” interval rangesfrom 117 to 158 ft msl and varies in thicknessfrom 5 to 46 ft (Figure 11). Two channel-likefeatures were recognized in this interval, onethat trends northwest to southeast beneath Aand M Areas and the other in the westernportion of the study area. The mean mudfraction ranges from less than 5% to greaterthan 80%. Areas with high mean mud values


O-7

correspond with thin parts of the interval(Figures 11 and 12). Standard deviation ofthis interval indicates high internal variabilityof the mud content. Where significantamounts of clay do exist, they are highlyinterbedded with sand. As with the “lower”and “middle” interval the northeastern portionof the study area has a sporadic distribution ofmud. The areas of mean mud values greaterthan 50% within the “upper” interval probablydefine the areal extent of the “upper clay”confining zone of the Crouch Branchconfining unit (Figure 12). The absence ofsignificant quantities of mud from all threeintervals in the northern sector of the A/MArea indicates that the Steed Pond aquifer isprobably in hydraulic communication with theCrouch Branch aquifer in this part of the studyarea.

Steed Pond Aquifer

The Steed Pond aquifer is composed primarilyof sand and clayey sands interbeds (Aadlandand others, 1995b). The Steed Pond aquifer isdivided into the M Area aquifer zone, GreenClay confining zone, and Lost Lake aquiferzone (Figure 6). The Lost Lake aquifer zoneconsists of yellow, tan, orange, and brown ,loose to slightly indurated, fine to coarse,moderately to well-sorted, occasionally pebblysand and minor clayey sand. The Green Clayconfining zone overlies the Lost Lake aquiferzone. This zone is considered correlative withthe clay and silty clay beds of the Gordonconfining unit of the Floridan aquifer systemsouth of Upper Three Runs (Aadland andothers, 1995b). In the A/M Area the GreenClay confining zone consists of primarilyorange and yellow, fine to coarse, poorly towell-sorted, often pebbly sand and clayey sandinterbedded with gray, green, and tan clay tosilty clay beds (Fallaw and Price, 1992).Aadland and others (1995b) consider theseclay and silty clay beds the Green Clayconfining zone when they are sufficientlythick and continuous. The M Area aquiferzone extends from the water table to the topof the Green Clay confining zone (Aadlandand others, 1995b). In A/M Area the M Area

aquifer zone consists of orange to tan andyellow, fine to coarse, poorly to well-sortedsand of the Tobacco Road and Dry Branchformations (Fallaw and Price, 1992). Pebblylayers, interbedded clay laminae, and clayinterbeds (up to 8 ft thick) are common. Onlythe Green Clay confining zone will bediscussed in detail. The Green Clay confiningzone was theorized to be a principle controlon the accumulation and migration of DNAPLbelow the water table in A/M Area (Looneyand others, 1992). This hypothesis wasfurther supported using indirect DNAPLcharacterization techniques along withheuristic DNAPL modeling (Jackson andothers, 1996).

Green Clay confining zoneThe surface of the Green Clay confining zonevaries from 197 to 223 ft msl with thicknessvarying from 9 to 32 ft (Figure 13). TheGreen Clay confining zone is generally thickerin the southern part of the study area. Thethicker areas are related to the channel-likefeatures in the Lost Lake aquifer zone. TheGreen Clay confining zone is interpreted torepresent a fluvio-deltaic environment withvarying amounts of nearshore marineinfluence.

The mean mud content within the Green Clayconfining zone varies from less than 5% togreater than 50% and is significantly higher inthe southeast corner and southern edge of thestudy area (Figure 14). The isolated thick areanortheast of A Area is characterized byrelatively small mud percentages. The GreenClay confining zone contains very little mudat the western end of the study area. Thestandard deviation of the mean mudpercentages is generally greater in the areaswith larger mud fractions. The parts of theGreen Clay confining zone that lack mud tendto be thicker, and show smaller standarddeviation values. This indicates that theGreen Clay confining zone consists primarilyof sand, with moderate amounts ofinterbedded clay present in the southeasternpart of the study area where the unit isrelatively thin. The maps of mud distribution


O-8

indicate that the zone is probably a competentconfining zone southeast of the M Area Basin.The low mud percentages and low standarddeviation values west of the M Area Basinsuggest that the Green Clay confining zone isrelatively permeable in this area.

Summary and Conclusions

The distribution of mud layers and variationsin the percentage of fine-grained sediments inthe subsurface is essential in characterizingand understanding the DNAPL accumulation,movement, and geometry of the resultingcontaminant plumes beneath the A/M Area.In general, characterizing the nature anddistribution of confining zones in CoastalPlain sediments is challenging. Thesesediments were deposited by fluvial, deltaic,and/or coastal processes and are often alteredby post depositional erosion and weathering.As a result, they exhibit significant lateral andvertical variations typical of ephemeraldepositional settings. Variations in lithologyand stratigraphy can occur abruptly orgradually. Abrupt changes can occur due tolocal depositional environment (e.g.,meanders), structural modification, (e.g.,faults), and erosional events (e.g.,unconformities). Gradual changes are oftenmore difficult to evaluate and are oftenmanifested in coarsening or fining ofsediments vertically or along the dip of a bed.Precision mapping of critical sediment layerscan aid in locating areas where contaminationis most likely to have migrated. Additionally,this information can be used to refineremediation systems or assist in designingnew systems.

The mud distribution for the “lower”,“middle”, and “upper” intervals of the CrouchBranch confining unit indicates a series offacies changes beneath the study area. Thedistribution of mud in the northern sector ofthe study area indicates all three intervalsconsist primarily of sand. In the remainder ofthe study area, the “middle” and “upper”intervals both consist primarily of sand, with arelatively sporadic distribution of mud in

comparison to the “lower” interval. The“lower interval” consists primarily of thickclay layers with high mud percentages in thecentral and southern sectors of the study areaand represents the most competent andextensive portion of the Crouch Branchconfining unit. The absence of significantquantities of mud from all three intervals inthe northern sector of the A/M Area indicatesthat the overlying Steed Pond aquifer is inhydraulic communication with the CrouchBranch aquifer in this region. This facieschange is probably a significant contributor tothe migration of dissolved contamination fromthe Steed Pond aquifer into the Crouch Branchaquifer (Jackson and other, 1997).

The Green Clay confining zone is within theSteed Pond aquifer and has a mean mudcontent ranging from approximately 5% to50%. While this zone is relatively thin andsandy it contains enough fine-grainedsediment to control the accumulation andmigration of DNAPL below the water tablenear known sources (e.g., the M-Area SettlingBasin). This is illustrated in Figure 15 thatpresents the results of a DNAPL screeninganalysis where the pure-phase solubility wascompared to historically reportedgroundwater. Screening of historicalconcentrations against 1% of pure-phasesolubility identified 43 monitoring wells and10 recovery wells that likely containedDNAPL within the A/M Area.

The geological characterization has hadspecific impacts on the design and operationof the A/M Area environmental clean-up.Near the original DNAPL sources, delineationof confining layers above and below the watertable is the basis for design of samplingstrategies and for targeting aggressivetreatment technologies. Below the watertable, these structures impede downwardmigration and serve to control accumulationof solvent into thin-laterally-extensive layersof contaminant rich fluid within overlyingsand and ultimately into “pools” occupyingstructural lows. One such area, southwest ofthe M-Area Settling Basin, was identified


O-9

using geophysics and sampling and wastreated using injection of Fenton’s reagent(Jerome and others, 1997). Above the watertable, our data indicate that residual DNAPLsolvents are trapped within fine-grainedsediment layers. As a result, we performedgeological investigations of these shallowintervals (Parker and others, 1999) and haveidentified and implemented source treatmentsthat have specific abilities to overcome themass transfer limited clean-up rates. Thethree enhanced remediation technologiestested to date are heating technologies – steamheating (Dynamic Underground Stripping,DUS), radio frequency heating (Jarosch andothers, 1994), and six phase joule heating(Gauglitz and others, 1994). The differencesin technology selection for the differentsource scenarios is a direct result ofcontrolling geological conditions and theobserved interaction of contaminant with thesubsurface materials.

Once the contaminant dissolves from thesource zone, the resulting plume migrates withthe groundwater toward the discharge alongTim’s Branch and Upper Three Runs. Thefocus for geology contribution to remedialdesign shifts to understanding the precisetrajectory of the contaminant plume. Weoptimize the performance of “pump and treat”remediation using knowledge of the plumecenterline and by careful placement ofrecovery well screens. Geologicalinformation and depth discrete samplingperformed during recovery well installationallow maximum efficiency because design canbe customized to extract the maximum amountof contaminant in the minimum volume ofwater. Similarly, the in well vapor strippingsystems installed in A/M Area were designedwith the intake screen centered in a carefullyidentified discrete layer of contamination thatoccupied only a portion of the Lost Lakeaquifer – maximizing system efficiency.

A particularly significant example of theimportance of geology in the A/M Area cleanup is the change made to the groundwatercorrective action system based on knowledge

about the Crouch Branch confining unit. Theobserved facies changes – the thinning andmore sandy character of this majorhydrostratigraphic “confining unit” near AArea – was the basis of modification of thepump and treat operation. To minimize futuredownward migration and contamination of theCrouch Branch aquifer, increased watertreatment capacity was installed and thenumber of recovery wells and pumping rate inshallow groundwater in A Area wasaggressively expanded. These modificationsrepresent a significant improvement in longterm protection of the important CrouchBranch aquifer for a relatively low cost.

Finally, the geological and chemical dataindicated that the bulk of the plume ismigrating toward Tim’s Branch and UpperThree Runs and suggested likely areas offuture plume discharge. Recently, passivesamplers placed in active flow lines near theprojected groundwater-surface waterdischarge area have confirmed this conceptualmodel. Geological considerations,particularly knowledge of the area where thestream channel has eroded below the greenclay, were critical in successfully determiningthe location of the future contaminatedgroundwater discharges (since the plume is inthe middle of the Lost Lake aquifer in thesouthern portion of A/M Area). The flow pathknowledge is now available for design andimplementation of effective long termremediation concepts. Two examples of suchconcepts are phytoremediation and microbialdegradation in the wetland sediments adjacentto the stream.

Delineation of the hydrogeologic frameworkof a contaminated site is a critical componentof effective environmental characterizationand remediation. A/M Area demonstrates thatdevelopment and refinement of a conceptualmodel using a variety of approaches, includingdrilling, coring, geophysical logging, aquifertesting, surface geophysics, modeling and datainterpretation, enhances decision making andtechnical performance.


O-10

References

Aadland, R. K., Gellici, J. A, and Thayer, P.A., 1995a. Hydrogeologic Framework ofWest-Central South Carolina, Report 5, WaterResources Division, South CarolinaDepartment of Natural Resources, Columbia,SC.

Aadland, R. K., Lewis, S. E., and McAdams,T. D. 1995b. HydrogeologicalCharacterization Report for the A/M Area(U). WSRC-RP-95-0052, WestinghouseSavannah River Company, Aiken, SC 29808.

Eddy-Dilek, C. A., Looney, B. B., Hoekstra,P., Harthill, N., Blohm, M., and Phillips, D.R., 1997. “Definition of a Critical ConfiningZone Using Surface Geophysical Methods”,Ground Water, v. 35(3): 451-462.

Fallaw, W. C. and Price, V., 1992. “Outline ofStratigraphy at the Savannah River Site”,Geological Investigations of the CentralSavannah River Area, South Carolina andGeorgia, Fallaw, W. C., and Price, V., eds.,Carolina Geological Society Field Trip GuideBook, November 13-15, 1992, CGS-92-B-11-1-33, U.S. Department of Energy and S.C.Geological Survey.

Fallaw, W. C. and Price, V., 1995.“Stratigraphy of the Savannah River Site andVicinity”, Southeastern Geology, 35: 21-58.

Gauglitz, P. A., Bergsman, T. M., Caley, S.M., Heath, W. O., Miller, M. C., Moss, R. W.,Roberts, J. S., Schalla, R., Schlender, M. H.,Jarosch, T. R., Eddy-Dilek, C. A., andLooney,B. B., Six Phase Soil Heating for EnhanceRemoval of Contaminants: Volatile OrganicCompounds in Non-Arid Soils IntegratedDemonstration, Savannah River Site, PNL-10184, Pacific Northwest NationalLaboratory, Richland WA.

Jackson, D. G., Payne, T. H., Looney, B. B.,and Rossabi, J. R., 1996, Estimating theExtent and Thickness of DNAPL within theA/M Area of the Savannah River Site (U),

WSRC-RP-06-0574, Westinghouse SavannahRiver Company, Aiken, SC 29808.

Jackson, D. G., B. B. Looney, and H. W.Campbell. 1997, Assessment of ChlorinatedSolvent Contamination in the Crouch BranchAquifer of the A/M Area (U). WSRC-RP-97-0247: September 30, 1997. WestinghouseSavannah River Company, Aiken, SC 29808.

Jarosch, T. R., Beleski, R. J., and Faust, D.,1994, Final Report: In Situ Radio FrequencyHeating Demonstration, WSRC-TR-93-673,Westinghouse Savannah River Company,Aiken, SC 29808.

Jerome, K. M., Riha, B. D., and Looney, B.B., 1997, Final Report for Demonstration ofIn Situ Oxidation of DNAPL Using the Geo-Cleanse® Technology, WSRC-TR-97-00283,Westinghouse Savannah River Company,Aiken, SC 29808.

Looney, B. B., Rossabi, J. R., and Tuck, D.M., 1992. Assessing DNAPL contamination,A/M-Area, Savannah River Site: Phase IResults (U), WSRC-RP-92-1302,Westinghouse Savannah River Company,Aiken, SC 29808.

Marine, I. W. and Bledsoe, H. W. 1984.Supplemental Technical Summary M-AreaGroundwater Investigation, DPSTD-84-0112,E. I. du Pont de Nemours & Co., SavannahRiver Laboratory, Aiken, SC 29808.

Siple, G. E., 1967. Geology and GroundWater of the Savannah River Plant andVicinity, South Carolina, U.S. GeologicalSurvey Water Supply Paper 1841.

Smits, A. D., Harris, M. K., Jackson, D. G.,and Hawkins, K. L., 1998, Baseline MappingStudy of the Steed Pond Aquifer and CrouchBranch Confining Unit Beneath A/M Area,Savannah River Site, Aiken, South Carolina(U), WSRC-RP-98-00357, WestinghouseSavannah River Company, Aiken, SC 29808.


O-11

Parker, W. H., Smits, A. D., Harris, M. K.,Jackson, D.G., and Hawkins, K.L, 1999,Baseline Mapping Study of the Steed PondAquifer and Vadose Zone beneath A/M Area,Savannah River Site, Aiken, South Carolina(U), WSRC-RP-99-00295, WestinghouseSavannah River Company, Aiken, SC 29808.

Wyatt, D. E., Aadland, R. K., and Syms, F. H.,2000, Overview of the Savannah River SiteStratigraphy, Hydrostratigraphy, andStructure (This volume).


O-12

Figure 2: Conceptual Diagram of a Contaminated Site withcost considerations for clean-up technologies

Figure 1: Conceptual Diagram of a Contaminated Site

Anatomy of a Contaminated Site

Source Zone

Characteristics:High ConcentrationsSignificantly perturbedgeochemistry

Need:Aggressive technologiesto limit long term damage

Examples:destruction or stabilizationin place; heat/steam;chemical oxidation orreduction; immobilization.

Primary Groundwater /Vadose Zone Plume

Characteristics:Moderate to high aqueous/vaporphase concentrations

Need: Baseline methods ormoderately aggressive alternatives

Examples: pump (gas or water) andtreat;recirculation wells; enhancedbioremediation

Dilute Plume / Fringe

Characteristics:Low aqueous/vaporphase concentrations;Large water volume.Need: innovativetechnologies - sustainablelow energy conceptsExamples: Passive pumping(siphon, barometric, etc.);bioremediation;phytoremediation,geochemical stabilization

Wastesite

Diagnosing and Treating aContaminated Site

Source Zone

Costs:$/lb contaminant or $/cuyd. Removalexamples:< $50-$100/cu yd or< $100/lb for chlorinatedsolvents

hot spot characterizationreduces cleanup volume

Primary Groundwater/ VadoseZone Plume

Costs:$/treatment volume (gallon/cu ft)example:<$0.5-$10 / 1000 gallons

zone of capture characterizationneeded, optimize extraction toreduce treatment volume

Dilute Plume/Fringe

Costs:Operation andmaintenance costs $/time

mass transfer and fluxcharacterization needed

Wastesite


O-13

Figure 3. Cut-away diagram showing the 3Dstructure of a real groundwater plume


O-14

Figure 4: Location of the Savannah River Site and the A/M Area

Figure 5: Detailed map of A/M Area illustrating core locations


O-15

Figure 6: Comparison of Chronostratigraphic, Lithostratigraphic, andHydrostratigraphic Units

CHRONOSTRATIGRAPHIC LITHOSTRATIGRAPHIC UNITS HYDROSTRATIGRAPHIC UNITSUNITS (Modified from Fallaw and Price, 1995) (Modified from Aadland and others, 1995a)

ERA System Series Group Formation

Miocene(?) "upland" unit

Tobacco Road Sand

Upper Dry Branch

Barnwell Formation

GroupClinchfield

Eocene FormationSantee

Orangeburg FormationMiddle Group Warley Hill Formation

Congaree Formation

Lower Fourmile Branch Formation

Upper Black Snapp Formation

Paleocene Mingo Lang Syne FormationLower Group Sawdust Landing Formation

Steel Creek Formation

Black Creek

Upper CretaceousGroup

Middendorf Formation

Cape Fear Formation

ME

SOZ

OIC

Tri

assi

c

Newark Supergroup

Sedimentary Rock (Dunbarton Basin)

Piedmont Hydrogeologic

LA

TE

(?)

PR

OT

ER

OZ

OIC

Pre

-Cam

bria

n(?)

Crystalline Basement Rock

Province

McQueen Branch confining unit

Appleton confining system

Crouch Branchconfining unit

Steed Pond aquifer

Southeastern Coastal P

lain Hydrogeologic P

rovince

"green clay" confining zone

"Lost Lake" aquifer zone

Crouch Branch aquifer

CE

NO

ZO

IC

Ter

tiary

Cre

tace

ous

"M Area" aquifer zone

"middle sand" aquifer zone

Floridan-Midville aquifer system

McQueen Branch aquifer

vadose zone

water table

ground surface


O-16

Figures 7 & 8: Isopach and Geometric Mean Mud percentage in the “Lower”interval of the Crouch Branch Confining Unit


O-17

Figures 9 & 10: Isopach and Geometric Mean Mud percentage in the “Lower”interval of the Crouch Branch Confining Unit


O-18

Figures 11 & 12: Isopach and Geometric Mean Mud percentage in the“Lower” interval of the Crouch Branch Confining Unit


O-19

Figures 13 & 14: Isopach and Geometric Mean Mud percentage in the GreenClay Confining Zone


O-20

Figure 15: Monitoring Wells considered to have DNAPL contamination basedon screening analysis comparing the pure-phase solubility to historicallyreported groundwater concentrations (Jackson and others, 1996).


P-1

Short Note:Seismic Monitoring at SRS

D. Stevenson, WSRC, Aiken, SC 29808

INTRODUCTION

The Savannah River Site (SRS) has beenoperating and maintaining a continuousrecording seismic network on Site since themid-seventies. The network was developed tomonitor Site and regional seismic activity thatmay potentially impact the safety of existing orplanned structures and systems at SRS.Operation and maintenance of the network is animportant ongoing task contributing to Siteseismic hazards analyses. It provides SRS withearthquake data collection and analysiscapability to insure accurate reliableinformation on current earthquake activity thatmay occur within the immediate Site vicinity.Additionally, network seismic data directlyaddresses seismic safety considerations byproviding an historical background databasedocumenting levels of seismic activityimpacting the Site through the years.

Current station configuration includes nineshort-period seismic recording stations, six areconfined within the plant boundary with anadditional three off-Site (Figure 1). The off-site station locations are within a 10 to 50 kmradius from the center of the SRS. Currently,we have a mix of digital and analog stationsdepending on quality of communication linkbetween the field and the central recordinglaboratory. One station, Hawthorne Firetower,has a unique configuration consisting of both ashort-period and long period instrument housedin a borehole package. This instrumentationpackage is presently installed in a relativelyshallow borehole (100ft). However, through acooperative effort with the USGS, drilling of adeeper bedrock hole (~1000 ft) at the samelocation is planned for January 2001. Uponcompletion of a deeper hole the existing sensorpackage will be re-deployed. It is hoped that thestation will then become a part of the USGSNational Seismic Network.

In addition to local events, regional earthquakeactivity occurring outside the relatively smallarea covered by the SRS monitoring networkalso has the potential for significantlyimpacting seismic safety considerations of theSite. Good quality regional seismic data arerequired when defining the range ofseismogenic sources that may influenceestimates of Site vibratory ground motion aswell as identification of possible seismicallyactive tectonic features. Regional coverage iscurrently provided cooperatively with theUniversity of South Carolina through data fromthe SCSN (South Carolina Seismic Network).The University of South Carolina seismicnetwork stations serve to compliment the SRSnetwork contributing greatly to the regional andSite area seismic monitoring effort.

On-Site Earthquake Activity

Three earthquakes of MMI III or less haveoccurred with epicentral locations within theboundaries of SRS (Figure 2). On June 9,1985, an intensity III earthquake with a localduration magnitude of 2.6 occurred at SRS. Feltreports were more common at the western edgeof the central portion of the plant site. Figure 3shows the resulting isoseismal intensity map. Another event occurred at SRS August 5,1988,with an MMI I-II and a local durationmagnitude of 2.0. A survey of SRS personnelwho were at the plant during the 1988earthquake indicated that it was not felt at SRS. Neither of these earthquakes triggered theseismic alarms (set point 0.002g) at SRSfacilities. These earthquakes were of similarmagnitude and intensity as several recentevents within the region.

On the evening of May 17, 1997, at 23:38:38.6UTC (7:38 pm EDST) an MD ~ 2.3 (DurationMagnitude) earthquake occurred within theboundary of the Savannah River Site. It wasreported felt by workers in K-Area and by


P-2

Wackenhut guards at a nearby barricade. AnSMA (strong motion accelerograph) located 3miles southeast of the epicenter at GunSite 51was not triggered by the event. The SMAlocated approximately 10 miles (16 km) northof the event in the seismic lab building 735-11A was not triggered. The closest instrument to the epicenter (GunSite 51) is set at a triggerthreshold of 0.3% of full scale where full scaleis 2.0g (0.006g). The more distant lab SMA isset to trigger at a threshold of 0.1% of full scalewhere full scale is 1.0g (0.001g).

Other Seismic Monitoring Efforts at SRS

Monitoring of structural response to earthquakemotions is being done through deployment andoperation of accelerographs (strong groundmotion recorder) in ten selected mission-criticalstructures. Instrument placement ranges fromfoundation level, selected elevations and in thefree-field depending on the structure beingstudied. Free-field instrumentation data will beused to compare measured response to thedesign input motion for the structures and todetermine whether the OBE has been exceeded.The instruments located at the foundation leveland at elevation in the structures will be used tocompare measured response to the design inputmotion for equipment and piping, and will beused in long-term evaluations. In addition,foundation-level instrumentation will providedata on the actual seismic input to the missioncritical structures and will be used to quantifydifferences between the vibratory groundmotion at the free-field and at the foundationlevel. All instruments are Kinemetrics EtnaStrong Motion Accelerographs with dial-upmodem data download capability. All SMAinstrumentation is set to trigger at 2.0% fullscale with full scale being 1g (i.e. trigger set at0.02g).

b

b

b

b

b

b

bb

b

MBY

DXN

NPR

ELK

SRPD

SRPW

SRPNSRAV

HAWT

E

N

B F H

Z

D

A

R

L

S

P

K

C

M

RRTNX

USFS

Legend

b

SRS Areas

SRS Roads

SRS Boundary

Georgia

South Carolina

Short Period Recording Stations

36̂ 22 '

PLANTNORTH

TRU ENORTH

2 0 2 4 6Miles

3 0 3 6 9Kilometers

Figure 1. SRS short period recording stations.

25 0 25 50 Miles

$

$

$$

$

$

$

$

$$

$

$

$

$

$

$$

$

$

$$

$

$

$

$

#

#

##

#

#

#

#

#

#

#

#

#

#

##

#

#

##

#

#

1/16/7911/15/78

7/26/45

15

22

21

19

18

17

16

14

13

12

11

10

9

8

7

6

54

3

2

1

20

1/7/92

Legend

60 Mile Buffer

New locations#

Old locations$

SRS Boundary

Interstates

Georgia

South Carolina 36^22 '

PLANTNORTH

TRU ENORTH

9 0 9 18 27Miles

10 0 10 20 30 40Kilometers

Figure 2. Historical earthquake epicenterslocated on the SRS. Triangles with date arehistorically mis-located.


P-3

Figure 3. Isoseismal Intensity map from the June 9, 1985, intensity III earthquake.


P-4

Notes


Q-1

Short Note:Formation Damage Caused by Excessive Borehole Fluid Pressures inCoastal Plain Sediments

D. E. Wyatt, WSRC, Aiken, SC 29808

IntroductionDuring the initial characterization andengineering studies of the SRS, performed bythe Corps of Engineers, numerous zones oflost circulation, “grout take” and rod dropwere noted. Large foundation groutingprograms discovered that intervals of “micro-channeling” and fracturing occurred duringpumped grout programs. Ongoing drillingactivities often report zones of lost circulationand rod drop, generally thought to be “softzones” (see Aadland et al. in this guidebook).In many cases rod drops and lost circulationare due to soft zones, however an abundanceof cone penetrometer (CPT) data suggests thatmany coastal plain sediments may be subjectto hydraulic fracturing during drillingoperations. Hydraulic fracturing may be thecause of some of the lost circulation problemsand may cause problems if key aquitards aredamaged.

An Example of the ProblemState regulations and SRS Drilling Proceduresrequire that surface casing be set inmonitoring wells penetrating deeper aquifersin 1) areas where known contamination existsin overlying aquifers, and 2) areasdowngradient from facilities or areas withknown contamination. Typical casing setpoints are within the “Green Clay” interval(Gordon Confining Unit) of the Warley HillFormation and occasionally within the “TanClay” interval of the Dry Branch Formation(Tan Clay Confining Zone of the Upper ThreeRuns Aquifer). These two units are clayeyaquitards that are generally less than ten feetthick for the “Tan Clay” and less than six feetthick for the “Green Clay”. Neither of theseunits are massively bedded clays but generallyconsist of thin to thick (<1” to <2’) laminainterspersed within clayey silts and finegrained sands. However, few studies have

been completed on the potential formationdamage effects caused by hydraulic fracturingduring casing and cementing activities withinthese sediments. Exploration conducted toinvestigate the effects of foundation groutingprograms showed that large volumes ofinjected grout traveled radially, in thin layers,considerable distances from the borehole.Intuitive experience from the petroleumindustry suggests that formation damage islikely when the hydraulic pressure gradientfrom the drilling or grouting exceeds thefracture gradient of the formation. Formationdamage may lead to “leaking” of downwardmoving contaminants or upward movinggroundwater through fractures and micro-fractures adjacent to the borehole. Hydraulicover-pressuring may be caused by excessivemud weight, grout weight or pump over-pressuring.

ApproachSince few measurements of formation damagehave been made at the SRS, a theoreticalapproach is made based on data collected atthe Hydrogeological Field Test Site in wellMWD-12. This well is centrally located at theSRS and is representative of the majority ofmonitoring wells in the area. MWD-12 wascompleted to provide core information for acomparison study between core propertymeasurements and data obtained fromgeophysical tools, principally the magneticresonance and formation fluid/pressure tester(the MRIL and Formation Multi-Tester (FMT)tools). In addition to core and geophysicalmeasurements, Cone Penetrometer Test (CPT)information was also obtained immediatelyadjacent to the boring location. CPT data areavailable to a depth of 153’ and geophysicaldata to a depth of 342’.


Q-2

Formation damage is probable once the fluidpressures in a borehole exceed the fracturestress pressure of the formation. When thishappens, borehole fluids will tend to flowunder gravity and/or pumped pressure into theformation along preferentially oriented planesof weakness. In most sediments, these planesof weakness are lateral or horizontal (alongthe shear plane) rather than vertical (along thecompressive plane) (Adams and Charrier,1985). However, vertical fractures are knownin soft sediments in outcrop Formationdamage is caused by flowing borehole fluidsdisplacing formation fluids and sediments.

Two key values are required to estimatewhether formation damage is likely, thepressure of the borehole fluids at a givendepth, and the fracture stress gradient of thesediments. For this study, it is assumed thatsediments at the SRS are not over-pressured(for example, due to hydraulic charging fromoil or gas) and are in iso-static equilibrium. Ifthe density or composition of the boreholefluids is known, or the weight of the fluids perunit volume, then the hydrostatic pressure ofthe fluid per depth may be calculated, or readfrom prepared tables. The formation fracturegradient, by rule of thumb in Gulf CoastalPlain type sediments is approximately 1.0psi/ft (Adams and Charrier, 1985). However,in the shallow sediments of the SRS, whichare relatively less consolidated than deeperGulf Coastal Plain sediments, (although theyare of the same relative type and age) it ispossible to calculate and extrapolate a stressgradient based on data from the CPT and FMTinformation.

Figure 1 is a graph showing the calculatedpressure at depth for typical drilling mud andgrout as defined in the Halliburton CementingTables. Overburden pressures, from the CPTtool from MWD-12 and hydrostatic pressuresfrom the FMT tool are also shown. The FMTis a direct measure of total formation fluidpressure. CPT overburden pressure arecalculated as “total” which is a function ofdepth, while “effective” overburden pressuresconsider the effects of saturated sediments

below the water table. The CPT data(measured to a depth of 153’) are extrapolatedto approximately 300’.

If the pore pressure is known, then it ispossible to calculate the formation fracturestress gradient using the equation of Hubbertand Willis (in Adams and Charrier, 1985) asfollows:

P/Z (min, max) = (1/3 min, ½ max)(Sz/Z + 2p/Z)

Where:P = fracture pressure, psiZ = depth, feetSz = overburden at depth Z, psip = pore pressure, psi

Example:

for p = 43 psi, Sz = 68 psi, Z=100 ft.,then P = 78 psi minimum and 118 psimaximum

The maximum and minimum calculatedfracture pressures from this equation are alsoshown on Figure 1. It should be understoodthat this equation is generally used for muchdeeper strata, however, the match between theCPT derived Effective Overburden pressureand the minimum fracture pressure, isinteresting (the pore pressure, p, used in theHubbert and Willis equation is utilized fromthe CPT data).

The FMT tool measures formation fluidpressure directly at discrete depths. These dataare also shown on Figure 1. Compared to thenormal hydrostatic curve (assumingfreshwater) demonstrates that the measuredFMT pressures are higher and tend to parallelthe normal hydrostatic curve.

An assumption that the CPT EffectiveOverburden pressure, measured FMT MWD-12 hydrostatic pressure and the minimumFormation Fracture pressures are equivalent isvalid because the curves are similar on Figure1 (the CPT Effective Overburden pressure andthe minimum Fracture Pressure curves are


Q-3

measured to 153’ and extrapolated toapproximately 300’). This assumption isintuitive for saturated sediments if the totalpressure at a given depth is a sum of thelithologic and hydrostatic pressure at thatdepth. Since the sediments are not over-pressured, the measured formation pressurefrom the FMT and measured EffectiveOverburden Pressure from the CPT will beequivalent to the hydrostatic pressure of theformation at a given depth.

Results and ConclusionsA review of the graph demonstrates that:

♦ calculated overburden pressures from theCPT and the minimum fracture pressuresfrom the Hubbert and Willis equationgenerally agree, and are equivalent withthe FMT derived pressures,

♦ the CPT Total Overburden and EffectiveOverburden curves diverge at the watertable,

♦ the hydrostatic pressure from the 13.2lb./gal. grout exceeds the Total EffectiveOverburden curve the at approximately 90feet (Area A) suggesting the possibility ofhydraulic fracturing in sediments at thisdepth,

♦ the 10 lb./gal. mud weight exceeds theCPT Effective Overburden and minimumFracture Pressure a approximately 240feet, suggesting the possibility ofhydraulic fracturing at this depth (AreaB),

♦ the FMT pressures are essentiallyequivalent to the extrapolated minimumFracture Pressure and CPT Effective

Overburden, but demonstrate a divergencein Area C suggesting that measuredformation pressures and calculatedformation pressures also diverge at thesedepths.

♦ the maximum Formation and TotalOverburden Pressures are generallygreater than pressures anticipated in thedepth ranges evaluated,

♦ any pumping pressures from the surfacemust be added to the hydraulic pressuresto estimate subsurface pressure effects.

Based on these observation, it is possible fordrilling operation to hydraulically fracturetypical unconsolidated sediments when groutand mud weights exceed the CPT measuredEffective Overburden Pressure or thecalculated Minimum Formation FracturePressure. It may be worthwhile, in areas to beevaluated using monitoring wells requiringcasing, to obtain CPT derived data formationpressure data, and perform a hydraulicfracture potential analysis. Key formations,such as those mentioned earlier, may bedamaged by excessive mud and/or groutweight induced hydraulic fracturing.

ReferencesAdams, N. J. and Charrier, T., 1985, DrillingEngineering, PennWell Books, Tulsa, 960 p.Field Data Handbook, Dowell-Schlumberger,Houston, TX

Halliburton Cementing Tables, 1981, HaliburtonServices Co., Duncan, OK

Carolina G


ield Trip G

uidebookW

SRC

-MS-2000-00606

Q-4

Figure 1. This graph presents the calculated maximum and minimum fracture pressures, measured hydrostatic pressures from the FMT tool, total and effective overburdenpressures measured from the CPT tool, pressure weights of typical grout and drilling mud versus depth, from well MWD-12. A normal freshwater hydrostatic curve is alsoshown for comparison. The interpreted formation horizons are shown across the top of the graph associated with the depth. “Upland” is the Upland Unit, “TR/DB” is theTobacco Road Fm. – Dry Branch Fm. contact, “Santee” is the Tinker/Santee Fms. interval, “GC” is the Green Clay or Gordon Confining unit of the Warley Hill Fm.,“”Gordon”” is the Gordon Aquifer Unit and “CBCU” is the Crouch Branch Confining Unit. Several lines are extrapolated (dashed lines) to depths beyond the available CPTdata. The water table depth is shown as a blue triangle. These data do not include surface pump pressures the inclusion of which would shift the intersection areas “A” and “B”to shallower depths.


R-1

Bibliography

Selected references from the SRS Generic Safety Analysis Report

Aadland R. K., and Bledsoe, H. W., 1990b,"Hydrostratigraphy of the Coastal Plain sequence,Savannah River Site (SRS), South Carolina": GSAAbstracts with Programs, v. 22, no. 7.

Aadland, R. K., 1997, Tertiary Geology of A/MArea Savannah River Site, South Carolina, WSRC-RP-97-0505, Rev 1. Westinghouse Savannah RiverCompany, Aiken SC 298098, 60 p.

Aadland, R. K., and Bledsoe, H. W., 1990a,Classification of hydrostratigraphic units at theSavannah River Site, South Carolina: USDOEReport, WSRC-RP-90-987, WestinghouseSavannah River Co., Westinghouse SavannahRiver Laboratory, Aiken, SC 29808, 15 p.

Aadland, R. K., et al. Significance of Soft ZoneSediments at the Savannah River Site (U),Historical Review of Significant Investigations andCurrent Understanding of Soft Zone Origin, Extentand Stability. WSRC-TR-99-4083, Rev. 0,Westinghouse Savannah River Company,September 1999.

Aadland, R. K., Smits, A. D., and Thayer, P. A.,1992a, Geology and hydrostratigraphy of the A/MArea, Savannah River Site (SRS), South Carolina(U): USDOE Report WSRC-RP-92-440,Westinghouse Savannah River Company,Savannah River Site, Aiken, SC 29808, 104 p.

Aadland, R. K., WSRC-RP-97-0505, Rev. 1,Tertiary Geology of A/M Area Savannah RiverSite, South Carolina (U).

Aadland, R.K. Geology of A/M Area, SavannahRiver Site South Carolina. WSRC-RP-96-0505,Rev. 0, Westinghouse Savannah River Company,Aiken, SC, 1996.

Aadland, R.K., Gellici, J. A. and P. A. Thayer,1995, Hydrogeologic Framework of West-CentralSouth Carolina, South Carolina Department ofNatural Resources, Water Resources Division,Report 5, Columbia, South Carolina, 200 p.

Amick, D. and Gelinas, R. “The Search forEvidence of Large Prehistoric Earthquakes alongthe Atlantic Seaboard.” American Association forthe Advancement of Science, February 1991. Vol.251, 1991.

Amick, D.R., et.al. Paleoliquefaction FeaturesAlong the Atlantic Seaboard. NUREG/CR-5613,U.S. Nuclear Regulatory Commission,Washington, DC, 1990.

Anderson, E. E. “The Seismotectonics of theSavannah River Site: The Results of a DetailedGravity Survey”. M.Sc. Thesis, University ofSouth Carolina, Columbia, SC, 246p. 1990.

ASCE. Technical Engineering and Design GuidesAdapted from the US Army Corps of Engineers,No. 9, Settlement Analysis, American Society ofCivil Engineers Press, New York, New York,1994.

ASTM. Laboratory Compaction Characteristics ofSoil Using Modified Effort. ASTM D1557-91,American Society for Testing and Materials,Conshohocken, PA, 1999.

Aucott, W. R., and Speiran, G. K., 1985a, "Groundwater flow in the Coastal Plain aquifers of SouthCarolina": Journal of Ground Water, v. 23, no. 6,p. 736-745.

Aucott, W. R., Davis, M. E., and Speiran, G. K.,1987, Geohydrologic framework of the CoastalPlain aquifers of South Carolina: U.S. GeologicalSurvey Water-Resources Investigations Report 85-4271.

Barstow, N.L., et.al. An Approach to SeismicZonation for Siting Nuclear Electric PowerGeneration Facilities in the Eastern United States:NUREG/CR-1577. U.S. Nuclear RegulatoryCommission, 1981.

Bates, R. L., and Jackson, J. A., 1980, eds.Glossary of Geology: American GeologicalInstitute, Falls Church, Va., 857 p.

Baum, G.R., et.al. “Structural and StratigraphicFramework for the Coastal Plain of NorthCarolina.” Field Trip Guidebook for CarolinaGeological Society and Atlantic Coastal PlainGeological Society. North Carolina Department ofNatural Resources and Development, 1979.

Beard, D. C., and Weyl, P. K., 1973, "Influence oftexture on porosity and permeability ofunconsolidated sand": American Association ofPetroleum Geologists Bulletin, 57, p. 349-369.


R-2

Bechtel Corporation, 1982, Studies of postulatedMillett Fault: Report prepared for Georgia PowerCompany with Southern Company Services.

Benedict, M., Cooke, J. B., Deere, D. U., Garrels,R. M., Graham, R. M., and Wolman, A., 1969,Permanent storage of radioactive separationsprocess wastes in bedrock on the Savannah RiverPlant Site: E. I. du Pont de Nemours and Co.,Savannah River Laboratory, Aiken, SC 29801, 57p.

Berkman, E. High Resolution Seismic Survey, PenBranch Fault, Savannah River Site, South Carolina.WSRC-TR-91-38, Westinghouse Savannah RiverCompany, Aiken, SC, 1991.

Bledsoe, H.W., Aadland, R. K., and Sargent, K. A.,1990, Baseline Hydrogeologic Investigation -Summary Report. WSRC-RP-90-1010,Westinghouse Savannah River Company, Aiken,SC, 1990.

Bobyarchick, A.R. “The Eastern Piedmont FaultSystem and Its Relationship to AlleghanianTectonics in the Southern Appalachians.” Journalof Geology. Vol. 89, 1981.

Bollinger, G.A. “Reinterpretation of the IntensityData for the 1886 Charleston, South CarolinaEarthquake.” Studies Related to the Charleston,South Carolina, Earthquake of 1886--APreliminary Report. U.S. Geological SocietyProfessional Paper 1028, 1977.

Bollinger, G.A. “Seismicity of the SoutheasternUnited States." Bulletin of the SeismologicalSociety of America. Vol. 63, No. 5, 1973.

Bollinger, G.A. “Specification of Source Zones,Recurrence Rates, Focal Depths, and MaximumMagnitudes for Earthquakes Affecting theSavannah River Site in South Carolina.” U.S.Geological Survey Bulletin. No. 2017, 1992.

Bollinger, G.A. “Specification of Source Zones,Recurrence Rates, Focal Depths, and MaximumMagnitudes for Earthquakes Affecting theSavannah River Site in South Carolina.” U.S.Geological Survey Bulletin. No. 2017, 1992

Bollinger, G.A., et.al. “An Analysis of EarthquakeFocal Depths in the Southeastern United States."Geophysical Research Letters. Vol. 12, 1985.

Bollinger, G.A., et.al. “Seismicity of theSoutheastern United States: 1698-1986."Neotectonics of North America. GeologicalSociety of America DNAG Map Vol. 1, 1991

Bollinger, G.A., et.al. “Seismicity of theSoutheastern United States -- 1968–1986." Decadeof Geology. 1987.

Bramlett, K.W. Geology of theJohnston-Edgefield Area, South Carolina, and itsRegional Implications. (M.S. Thesis). Universityof South Carolina, Columbia, SC, 1989.

Bramlett, K.W., et.al. “The Belair Fault: ACenozoic Reactivation Structure in the EasternPiedmont." Geological Society of AmericaBulletin. Vol. 93, 1982.

Brown, P.M., et.al. Structural and StratigraphicFramework, and Spatial Distribution ofPermeability of the Atlantic Coastal Plain, NorthCarolina to New York. U.S. Geological SurveyProfessional Paper 796, 1972.

Chapelle, S. H. and Lovley, D. R., 1991,Competitive exclusion of sulfate reduction byF(III) – reducing bacteria: A mechanism forproducing discrete zones of high-iron groundwater,Groundwater, Vol. 30 No. 1, p. 29-36.

Chapman, W.L. and DiStefano, M.P. SavannahRiver Plant Seismic Survey, 1987-88. ResearchReport 1809-005-006-1-89, Conoco, Inc., SeismicAcquisition Section, 1989.

Chowns, T.M., Fritz, W.J., Latour, T.E., 1996,“Investigation of Triassic Basin Sediments –Savannah River Site. Final Report Prepared forWestinghouse Savannah River Co. Prime ContractNo. :DE-AC09-89SR18035 ERDA Task #52.

Christopher, R. A., 1978, "Quantitative palynologiccorrelation of three Campanian and Maestrichtiansections (Upper Cretaceous) from the AtlanticCoastal Plain": Palynology, v.2, p. 1-27.

Clarke, J. S., Brooks, R., and Faye, R. E., 1985,Hydrogeology of the Dublin and Midville aquifersystems of east-central Georgia: GeorgiaGeological Survey Information Circular 74.

Clarke, J. S., Hacke, C. M., and Peck, M. F., 1990,Geology and ground-water resources of the coastalarea of Georgia: Georgia Geological SurveyBulletin 113, 106 p.

Colquhoun, D. J., J. M. Rine, M. P. Segall, M. P.Katuna, A. D. Cohen, and D. L. Siron, 1994,Sedimentology and Stratigraphy of the UplandUnit, Final. Prepared under SCUREF cooperationAgreement with DOE, Task 119AA00900T.Columbia, South Carolina: Universityof South Carolina, Geological SciencesDepartment, Earth Science and Resources Institute.


R-3

Colquhoun, D. J., Oldham, R. W., Bishop, J. W.,and Howell, P. D., 1982, Up-dip delineation of theTertiary limestone aquifer: South CarolinaDepartment of Geology, University of SouthCarolina, Columbia, SC, 29201

Colquhoun, D.J. and Steele, K.B.Chronostratigraphy and Hydrostatigraphy of theNorthwestern South Carolina Coastal Plain.Annual Cooperative Grant Agreement No. 13040R-83-591, Project No. G868-05, Interim TechnicalReport to Water Resources Research Institute,Clemson University, Clemson, South Carolina,1985.

Colquhoun, D.J., et.al. “Surface and SubsurfaceStratigraphy, Structure and Aquifers of the SouthCarolina Coastal Plain.” SCDHEC Report ISBN0-9613154-0-7, 1983.

Cook, F.A., et.al. “COCORP Seismic Profiling ofthe Appalachian Orogen Beneath the Coastal Plainof Georgia.” Geological Society of AmericaBulletin. Vol. 92, No. 10, 1981.

Cook, F.A., et.al. “Thin-skinned Tectonics in theCrystalline Southern Appalachians: COCORPSeismic Reflection Profiling of the Blue Ridge andPiedmont.” Geology. Vol. 7, 1979.

Cooke, C.W. “Geology of the Coastal Plain ofSouth Carolina.” U.S. Geological Survey Bulletin.No. 867, 1936.

Coruh, C., et.al. “Results from Regional VibroseisProfiling: Appalachian Ultradeep Core Hole SiteStudy.” Geophysical Journal of the RoyalAstronomical Society. Vol. 89, 1987.

Coruh, C., et.al. “Seismogenic Structures in theCentral Virginia Seismic Zone.” Geology. Vol.16, 1988.

Cox, J. and Talwani, P. “Paleoseismic Studies inthe 1886 Charleston EarthquakeMeizoseismal Area.” Geological Society ofAmerica, Abstracts with Programs. No. 16, 1983.

Cumbest, R.J., et.al. “Gravity and MagneticModeling of the Dunbarton Triassic Basin, SouthCarolina.” Southeastern Geology. Vol. 33, No. 1,1992.

Cumbest, R.J., Stephenson, D.E., Wyatt, D.E.,Maryak, M. “Basement Surface Faulting forSavannah River Site and Vicinity” WSRC-TR-98-00346. 1998.

Dallmeyer, R.D., et.al. “Character of theAlleghanian Orogeny in the SouthernAppalachians: Part II, Geochronological

Constraints on the Tectonothermal Evolutions ofthe Eastern Piedmont in South Carolina."Geological Society of America Bulletin. Vol. 97,1986.

Daniels, D.L. “Geologic Interpretation ofGeophysical Maps, Central Savannah River Area,South Carolina and Georgia." GeophysicalInvestigation Map GP-893, U.S. GeologicalSurvey, 1974.

Davis, G.J. The Southwestern extension of theMiddleton-Lowndesville Cataclastic Zone in theGreensboro, Georgia, Area and its RegionalImplications (M. S. Thesis). University ofGeorgia, Athens, Georgia, 1980.

Dennehy, K.F., et.al. “Geohydrology of theDefense Waste Processing Facility and Vicinity,Savannah River Plant, South Carolina.” U.S.Geological Survey Water Resources Investigation,WRI 88-4221, 1988.

Dennis, A.J. “Is the Central Piedmont Suture aLow-Angle Normal Fault?” Geology. Vol. 19,1991

Dennis, A.J., Maher Jr., H. D., Mauldin, J.C.,Shervais, J.W., 1996, “Repeated PhanerozoicReactivation of a Southern Appalachian Fault ZoneBeneath the Up-Dip Coastal Plain of SouthCarolina. Report submitted for SCURF Task 170.

Dept. of the Navy. Soil Mechanics, NAVFACDM-7.01, Washington D. C., September 1986.

Dewy, J. W., “Relocation of instrumentallyrecorded pre-1974 earthquakes in the SouthCarolina region”. In Gohn, G.S. (ed.), StudiesRelated to the Charleston, South CarolinaEarthquake of 1886 – Techtonics and Seismicity,U.S. Geological Survey Professional Paper 1313,q1-q9, 1983.

Dillon, W.P. and Popenoe, P. “The Blake PlateauBasin and Carolina Trough." The Geology ofNorth America. V I-2, The Atlantic ContinentalMargin. DNAG Publication, Geological Society ofAmerica, Boulder, CO, 1988.

Dobry, R., et.al. Predication of Pore WaterPressure Buildup and Liquefaction of SandsDuring Earthquakes by the Cyclic Strain Method.National Bureau of Standards Building ScienceSeries 138, 1982.

Domoracki, W.J. A “Geophysical Investigation ofGeologic Structure and Regional Tectonic Settingat the Savannah River Site, South Carolina”. PhDDissertation, Virginia Polytechnic Institute,Blacksburg VA. 1995.


R-4

Domoracki, W.J., et.al. “Seismotectonic StructuresAlong the Savannah River Corridor, SouthCarolina, USA.” Journal of Geodynamics. 1997.

Dutton, C. E. The Charleston Earthquake ofAugust 31, 1886. U.S. Geological Survey, 1890.

Dutton, C.E. “The Charleston Earthquake ofAugust 31, 1886.” U.S. Geological Survey AnnualReport 1887-1888. Government Printing Office,Washington, DC, 1889.

Evans, E. K. and Parizek, R. R., 1991,Characterization of hydraulic conductivityheterogeneity in Tertiary sediments within theGeneral Separations Area, Savannah River Site,South Carolina: Department of Geosciences,Pennsylvania State University, Pa., 157 p.

Facility Safety. DOE Order 420.1, U.S.Department of Energy, Washington, DC, 1995.

Fallaw, W. C. and Price, V., 1992, Outline ofStratigraphy at the Savannah River Site: In Fallaw,W. C., and Price, Van, eds., 1992, CarolinaGeological Society Field Trip Guidebook, Nov. 13-15, 1992, Geological Investigation of the CentralSavannah River Area, South Carolina and Georgia:US Dept of Energy and SC Geological Survey, P.CGS-92-B-11-1-33.

Fallaw, W.C. and Price, Van. “Stratigraphy of theSavannah River Site and Vicinity.” SoutheasternGeology. Vol. 35, No. 1, 1995.

Fallaw, W.C., et.al. “Effects of Varying Degreesof Marine Influence on Tertiary Sediments inSouthwestern South Carolina.” Geological Societyof America Abstracts with Programs. Vol. 22, No7, 1990.

Farrar, S.S. “Tectonic Evolution of theEastern-Most Piedmont, North Carolina."Geological Society of America Bulletin. Vol. 96,1985.

Faye, R.E. and Prowell, D.C. Effects of LateCretaceous and Cenozoic Faulting on the Geologyand Hydrology of the Coastal Plain Near theSavannah River, Georgia and South Carolina. U.S.Geological Survey Open File Report 82-156, 1982.

Fetter, C. W., 1986, Applied Hydrogeology (2ndEdition) Merrill Publishing Company, Columbus,Oh.

Fullagar, P.D. and Bartholomew, M.J.“Rubidium-Strontium Ages of the Watauga RiverCranberry, and Crossing Knob Gneisses,Northwestern North Carolina." GeologicalInvestigations in the Blue Ridge of Northwestern

North Carolina: 1983 Guidebook for the Carolinas.Geological Society, North Carolina Division ofLand Resources, Article 11, 1983.

Fullagar, P.D. and Odom, A.L. “Geochronology ofPrecambrian Gneisses in the Blue Ridge Provinceof Northwestern North Carolina and Adjacent Partsof Virginia and Tennessee.” Geological Society ofAmerica Bulletin. Vol. 84, 1973.

Fullagar, P.D., et.al. “1200 m.y.-Old Gneisses inthe Blue Ridge Province of North and SouthCarolina.” Southeastern Geology. Vol. 20, 1979.

General Design Criteria. DOE Order 6430.1A,U.S. Department of Energy, Washington, DC,1989.

GeoTrans, Inc., 1988, Characterization of groundwater flow and transport in the GeneralSeparations Area, Savannah River Plant: Effects ofclosure options on groundwater flow at the F-Areaseepage basins.

Glover, L., III, et.al. “Diachronous PaleozoicMylonites and Structural Heredity ofTriassic-Jurassic Basins in Virginia." GeologicalSociety of America Abstracts with Programs. Vol.12, 1980.

Gohn, G.S. “Late Mesozoic and Early CretaceousGeology of the Atlantic Coastal Plain: NorthCarolina to Florida.” The Geology of NorthAmerica. Vol. I-2, The Atlantic ContinentalMargin, pp. 107-130, DNAG Publication,Geological Society of America, Boulder, CO.,1988

Gohn, G.S. Studies Related to the CharlestonEarthquake of 1886: Tectonics and Seismicity.U.S. Geological Survey Professional Paper 1313,1983.

Gohn, G.S., et.al. “Field Studies of EarthquakesInduced Liquefaction-Flowage Features in theCharleston, South Carolina, Area PreliminaryReport”. U.S. Geological Survey Open-File Report84-670, 1984.

Goldsmith, R., et.al. “Geologic Map of theCharlotte 1x2 Quadrangle, North Carolina andSouth Carolina.” U.S. Geological SurveyMiscellaneous Investigations Series Map I-1251-E,Scale 1:250,000, 1988.

Gore, P.J.W. “Depositional Framework of aTriassic Rift Basin: The Durham and SanfordSub-basins of the Deep River Basin, NorthCarolina." Society of Economic Paleontologistsand Mineralogists Field Guidebook. Third AnnualMidyear Meeting, Raleigh, NC, 1986.


R-5

Griffin, V.S., Jr. “Geology of the Abbeville East,Abbeville West, Latimer and LowndesvilleQuadrangles, South Carolina." South CarolinaGeological Survey MS-24. 1979.

Grow, J.A., et.al. “Structure and Evolution ofBaltimore Canyon Trough." The Geology of NorthAmerica. Vol. I-1, The Atlantic ContinentalMargin. DNAG Publication, Geological Society ofAmerica, Boulder, CO, 1988.

Grubb, H. F., 1986, Gulf Coast regional aquifer-system analysis - A Mississippi perspective: U. S.Geological Survey Water-Resources InvestigationsReport 86-4162, 22 p.

Guidelines for Use of Probabilistic Seismic HazardCurves at Department of Energy Sites.DOE-STD-1024, U.S. Department of Energy,Washington, DC, December 1992.

Gutierrez, B.J. (1999). Letter from the DOE to toL.A. Salomone and F. Loceff, on “RevisedEnvelope of the Site Specific PC-3 Surface GroundMotion”, September 9, 1999.

Hall, M.H. “Heavy Mineral Provinence of TriassicDunbarton Basin Deposits, South Carolina”. MSc.Thesis, University of North Carolina Wilmington,Wilmington, NC. 1997.

Hanson, K.L., et.al. “Applications of QuaternaryStratigraphic, Soil-Geomorphic, and QuantitativeGeomorphic Analyses to the Evaluation ofTectonic Activity and Landscape Evolution in theUpper Coastal Plain, South Carolina."Proceedings, 4th DOE Natural Phenomena HazardsMitigation Conference. Vol. 2, Atlanta, Georgia,1993.

Hasek, M. J. Settlement of H-Area Waste StorageTanks and Structures (U), K-ESR-H-00011, Rev.0.0, October 1999.

Hatcher, R.D., et.al. “The Smokies FoothillsDuplex and Possible Significance of the GuessCreek Fault, A Corollary to the Mapping of Kingand Neuman.” Geological Society of AmericaAbstracts with Programs. Vol. 18, 1986a.

Hatcher, R.D., et.al. Continent-Ocean Transect E5Cumberland Plateau (North American Craton) toBlake Plateau Basin. The Decade of NorthAmerican Geology Project, Geological Society ofAmerica, Boulder, CO, 1994

Hatcher, R.D., Jr. “Developmental Model for theSouthern Appalachians.” Geological Society ofAmerica Bulletin. Vol. 83, 1972.

Hatcher, R.D., Jr. “Tectonics of the WesternPiedmont and Blue Ridge, Southern Appalachians:Review and Speculation." American Journal ofScience. Vol. 278, 1978.

Hatcher, R.D., Jr. and Butler, J.R. Guidebook forSouthern Appalachian Field Trip in the Carolinas,Tennessee, and Northeastern Georgia.International Geologic Correlation Program Project27, University of North Carolina, Chapel Hill, NC,1979.

Hatcher, R.D., Jr. and Edelman, S.H.“Macro-Scale Partitioning in the Southern andCentral Appalachians: Thrusting and Strike-Slip asProducts of Alleghanian Collision.” GeologicalSociety of America Abstracts with Programs. Vol.19, 1987.

Hatcher, R.D., Jr. and Goldberg, S.A. “The BlueRidge Geologic Province." The Geology of theCarolinas. Carolina Geological Society 50thAnniversary Volume, 1991.

Hatcher, R.D., Jr., et.al. “Eastern Piedmont FaultSystem: Speculations on its Extent.” Geology.Vol. 5, 1977.

Hatcher, R.D., Jr., et.al. “Geometric and TimeRelations of Thrusts in the Crystalline SouthernAppalachians.” Geometry and Mechanisms ofAppalachian Thrusting, with Special Reference tothe Appalachians. Geological Society of AmericaSpecial Paper 222, 1988.

Hatcher, R.D., Jr., et.al. “Tectonic Map of the U.S.Appalachian, Plate 1, the Applachian-OuachitaOrogen in the United States." The Geology ofNorth America. Vol. F-2. Geological Society ofAmerica, Boulder, CO, 1990.

Heller, P.L., et.al. “Episodic Post-Rift Subsidenceof the United States Atlantic Margin.” GeologicalSociety of America Bulletin. Vol. 93, No. 5, 1982.

Herrick, S. M., and Vorhis, R. C., 1963, Subsurfacegeology of the Georgia Coastal Plain: GeorgiaDepartment of Natural Resources, Division ofMines, Mining, and Geology Information Circular25, 80 p.

Herrmann, R.B. “Surface-Wave Studies of SomeSouth Carolina Earthquakes." Bulletin of theSeismological Society of America. Vol. 76, No. 1,1986.

Heyn, T. “Geology of the Hinge Zone of theSauratown Mountains Anticlinorium, NorthCarolina: Structure of the Sauratown MountainsWindow, North Carolina." Carolina GeologicalSociety Guidebook. 1988.


R-6

Hooper, R.J. and Hatcher, R.D., Jr. “The PineMountain Terrane, a Complex Window in theGeorgia and Alabama Piedmont-Evidence from theEastern Termination.” Geology. Vol. 16, 1988.

Hopson, J.L., et.al. “Geology of the Eastern BlueRidge of Northeast Georgia and the AdjacentCarolinas." Georgia Geological SocietyGuidebook. Vol. 9, 1989.

Hopson, J.R. “Structure, Stratigraphy, andPetrogenesis of the Lake Burton Mafic-UltramaficComplex." Georgia Geological SocietyGuidebook. Vol. 9, 1989.

Horton, J.W., Jr. “Shear Zone Between the InnerPiedmont and Kings Mountain Belts in theCarolinas.” Geology. Vol. 9, 1981.

Horton, J.W., Jr. and McConnell, K.I. “TheWestern Piedmont.” The Geology of theCarolinas. Carolina Geological Society 50thAnniversary Volume, 1991.

Housner, G.W. Earthquake Criteria for theSavannah River Plant. DPE-2383, E.I. du Pont deNemours and Company, Savannah River Plant,Aiken, SC, 1968.

Huddlestun, P.F. and Hetrick, J.H. “Stratigraphyof the Tobacco Road Sand - A New Formation."Georgia Geologic Survey. Bulletin 93, 1978.

Hutchinson, D.R. and Klitgord, K.D. “Evolutionof Rift Basins on the Continental Margin OffSouthern New England: Triassic-Jurassic Rifting:North America and Africa.” American Associationof Petroleum Geologists Memoir, 1986.

Hydrogeologic Characterization of the MWMF(Mixed Waste Management Facility), SRS.WSRC-RP-92-837, Westinghouse Savannah RiverCompany, Aiken, SC, 1992

Kanter, L.R. Tectonic Interpretation of StableContinental Crust in: The Earthquakes of StableContinental Regions. Electric Power ResearchInstitute, Palo Alto, CA, 1994.

Keen, C.E. and Haworth, R.T. “DNAG TransectD-3: Rifted Continental Margin Off Nova Scotia:Offshore Eastern Canada.” Geological Society ofAmerica, Centennial Continent/Ocean Transect #4.Boulder, CO, 1985.

Kimball, J.K. “The Use of Site DependentSpectra.” Proceedings of the U.S. GeologicalSurvey Workshop on Site Specific Effects of Soiland Rock on Ground Motions and the Implicationsfor Earthquake-Resistant Design. U.S. GeologicalSurvey Open File Report 83-845, 1983.

King, P.B. “A Geologic Cross Section Across theSouthern Appalachians, An Outline of the Geologyin the Segment in Tennessee, North Carolina, andSouth Carolina." Guides to Southeastern Geology.Geological Society of America Annual Meeting,1955.

King, P.B. Systematic Pattern of Triassic Dikes inthe Appalachian Region, Second Report. U.S.Geological Survey Professional Paper 759-D, 1971

King, P.B., et.al. 1968, Geology of the GreatSmoky Mountains National Park, Tennessee andNorth Carolina. U.S. Geological SurveyProfessional Paper 587.

Klitgord, K.D. and Schouten, H. “PlateKinematics of the Central Atlantic.” The Geologyof North America. Vol. M, The Western NorthAtlantic Region. DNAG Publication, GeologicalSociety of America, Boulder, CO, 1986.

Klitgord, K.D., et.al. “US Atlantic ContinentalMargin: Structural and Tectonic Framework." TheGeology of North America. V I-2, The AtlanticContinental Margin. DNAG Publication,Geological Society of America, Boulder, CO,1988.

Koester, J.P. and Franklin, A.G. CurrentMethodologies for Assessing the Potential forEarthquake-Induced Liquefaction in Soils.NUREG/CR-430, U.S. Nuclear RegulatoryCommission, Washington, DC, 1985.

Krause, R. E. and Randolph, R. B., 1989,Hydrology of the Floridan aquifer system insoutheast Georgia and adjacent parts of Floridaand South Carolina: U. S. Geological SurveyProfessional Paper 1403-D, 64 p.

Krause, R. E., 1982, Digital model evaluation ofthe predevelopment flow system of the Tertiarylimestone aquifer, southeast Georgia, northeastFlorida, and southern South Carolina: U. S.Geological Survey Water-Resources InvestigationsReport 82-173, 27 p.

Lahr, J.C., “HYPOELLIPSE: a computer programfor determining local earthquake hypocentralparameters, magnitude, and first motion pattern.”U.S. Geological Survey Open-File report 84-519,35p., 1984.

Latour, T.E., Roden, M.F., Vanko, D.A., Whitney,J.A. “Crystalline Basement Core – Savannah RiverSite. Final Report”. ERDA Task 53, 303p, 1995.

Lee, R.C. Investigations of Nonlinear DynamicSoil Properties at the Savannah River Site.


R-7

WSRC-TR-96-0062, Rev. 0, WestinghouseSavannah River Company, Aiken, SC, March 1996

Leutgert et.al. “Crystal Structure Beneath theAtlantic Coast Plain of South Carolina.”Seismological Research Letters. Vol. 65, No. 2,1994.

Lewis, M. R., Arango, I., Kimball, J. K. and T. E.Ross. “Liquefaction Resistance of Old SandDeposits,” Proceedings of the XI PanamericanConference on Soil Mechanics and GeotechnicalEngineering. Foz do Iguassu, Brazil, August 1999.

Li, W. T. Settlement of Defense Waste ProcessingFacility (U), K-ESR-S-00003, Rev. 0, September1999.

Lindholm, R.C. “Triassic-Jurassic Faulting inEastern Northern America - A Model Based onPre-Triassic Structures." Geology. Vol. 6, 1978.

Logan, W. R. and Euler, G. M., 1989, Geology andground-water resources of Allendale, Bamberg,and Barnwell Counties and part of Aiken County,South Carolina: Water Resources CommissionReport 155, 113 p.

Lohman, S. W., et al, 1972, Definitions of selectedgroundwater terms - revisions and conceptualrefinements: U. S. Geological Survey Water-Supply Paper 1988, 21 p.

Long, L.T. “The Carolina Slate Belt-Evidence of aContinental Rift Zone.” Geology. Vol. 7, 1979.

Long, L.T. and Chapman, J.W., Jr. “BouguerGravity Map of the Summerville-Charleston, SouthCarolina Epicentral Zone and TectonicImplications.” Studies Related to the Charleston,South Carolina Earthquake of 1886--A PreliminaryReport. U.S. Geological Survey Professional Paper1028, 1977.

Luetgert, J.H., H.M. Benz, and S. Madabhushi,“Crustal Structure beneath the Atlantic coastalplain of South Carolina.” Seismological ResearchLetters, vol 30, no. 7, 180-191, 1994

Madabhushi, S. and Talwani, P. “Fault PlaneSolution and Relocations of Recent Earthquakes inMiddleton Place Summerville Seismic Zone NearCharleston South Carolina." Bulletin of theSeismological Society of America. Vol. 83, No. 5,1993.

Maher, H.D. “Stratigraphy and Structure of theBelair and Kiokee Belts near Augusta, Georgia,Geological Investigations of the Eastern Piedmont,Southern Appalachians.” Carolina Geological

Society Field Trip Guidebook. South CarolinaGeological Survey, 1978.

Maher, H.D., et.al. “40Ar/39Ar Constraints onChronology of Augusta Fault Zone Movement andLate Alleghanian Extension, Southern AppalachianPiedmont, South Carolina and Georgia.” AmericanJournal of Science. Vol. 294, 1994.

Maher, H.D., et.al. “The Eastern Piedmont ofSouth Carolina.” The Geology of the Carolinas.University of Tennessee Press, Knoxville, 1991.

Maher, H.D., Jr. “Kinematic History of MyloniticRocks from the Augusta Fault Zone, SouthCarolina and Georgia." American Journal ofScience. Vol. 287, 1987.

Maher, H.D., Jr. Stratigraphy, Metamorphism, andStructure of the Kiokee and Belair Belts NearAugusta, Georgia (MS Thesis). University ofSouth Carolina, Columbia, SC, 1979.

Maher, H.D., Jr., et.al. “Magmatic Softening in theOrogenic Hinterlands Southern AppalachianPiedmont, Georgia.” Geological Society ofAmerica Abstracts with Programs. Vol. 24, No. 2,1992.

Manspeizer, W., et.al. “Separation of Morroccoand Eastern North America: A Triassic-LiassicStratigraphic Record." Geological Society ofAmerica Bulletin. Vol. 89, 1978.

Marine, I. Geohydrology of the Buried TriassicBasin at the Savannah River Plant Groundwater.Vol. 2, 1974a.

Marine, I. W., 1966, "Hydraulic correlation offracture zones in buried crystalline rock at theSavannah River Plant near Aiken, South Carolina":Geological Survey Research 1966, Chapter D, U.S.Geological Survey Professional Paper 550-D, p.D223-D227.

Marine, I. W., 1967a, "The permeability offractured crystalline rock at the Savannah RiverPlant near Aiken, South Carolina": GeologicalSurvey Research 1967, Chapter B, U.S. GeologicalSurvey Professional Paper 575-B, p. B203-B211.

Marine, I. W.,1967b, "The use of tracer test toverify an estimate of groundwater velocity infractured crystalline rock at the Savannah RiverPlant near Aiken, South Carolina": in IsotopeTechniques in the Hydrologic Cycle: GeophysicalMonograph Series, v. 11, ed., G. E. Stout,Washington, DC, American Geophysical Union, p.171-178.


R-8

Marine, I.W. “Geohydrology of Buried TriassicBasin at Savannah River Plant, South Carolina.”American Association of Petroleum GeologistsBulletin. Vol. 58, 1974b.

Marine, I.W. and Siple, G.E. “Buried TriassicBasin in the Central Savannah River Area, SouthCarolina and Georgia." Geological Society ofAmerica Bulletin. Vol. 85, 1974.

Martin, G.R., et.al. “Fundamentals of LiquefactionUnder Cyclic Loading." Journal of theGeotechnical Engineering Division. ProceedingsPaper 11284, ASCE, Vol. 101, No. 5, May 1975.

Mauldin, J., Shervais, J.W., Dennis, A.J., 1996,“Petrology and Geochemistry of MetavolcanicRocks from a Neoproterozoic Volcanic Arc,Savannah River Site, South Carolina”. Reportsubmitted for SCURF Task 170.

McBride, J.H. “Constraints on the Structure andTectonic Development of the Early MesozoicSouth Georgia Rift, Southeastern United States;Seismic Reflection Data Processing andInterpretation." Tectonics. Vol. 10, No. 5, 1991.

McConnell, K.I. “Geology of the SauratownMountains Anticlinorium; Vienna and Pinnacle 7.5Minute Quadrangles: Structure of the SauratownMountains Window, North Carolina.” CarolinaGeological Society Guidebook. 1988.

Meisler, H., 1980, Plan of study for the northernAtlantic Coastal Plain regional aquifer systemanalysis: U.S. Geological Survey Water-ResourcesInvestigations Report 80-16, 27 p.

Miller, J. A, and Renken, R. A., 1988,Nomenclature of regional hydrogeologic units ofthe Southeastern Coastal Plain aquifer system: U.S. Geological Survey, Water ResourcesInvestigations Report 87-4202, 21 p.

Miller, J. A., 1986, Hydrogeologic framework ofthe Floridan aquifer system in Florida and inparts of Georgia, Alabama, and South Carolina:U.S. Geological Survey Professional Paper 1403-B, 91 p.

Mitchell, J. K. “Practical Problems fromSurprising Soil Behavior,” The 20th TerzaghiLecture, Journal of Geotechnical Engineering, Vol.112, No. 3, American Society of Civil Engineers,March 1986.

Mitchell, J. K. and Solymar, Z. V. “TimeDependent Strength Gain in Freshly Deposited orDensified Sand,” Journal of GeotehcnicalEngineering Division. Vol. 110, No. 11, AmericanSociety of Civil Engineers, November 1984.

Mitwede, S.K., et.al. “Major ChemicalCharacteristics of the Hammett GroveMeta-Igneous Suite, Northeastern South Carolina.”Southeastern Geology. Vol. 28, No. 1, 1987.

Moos, D. and Zoback, M.D. In Situ StressMeasurements in the NPR Hole, Savannah RiverSite, South Carolina: Final Report to WestinghouseSavannah River Co., Vol. 1, Results andInterpretations. Subcontract AA00925P, ScienceApplications International Corporation, Augusta,GA, 1992.

National Center for Earthquake EngineeringResearch. Proceeding of the NCEER Workshop onEvaluation of Liquefaction Resistance of Soils.Technical Reprot NCEER-97-0022, Buffalo, NY,December 31, 1997.

National Earthquake Hazards Reduction Program(NEHRP) Recommended Provisions for SeismicRegulations for New Buildings and OtherStructures. Building Seismic Safety Council, part1 Provisions (FEMA 302).

Natural Phenomena Hazards Assessment Criteria.DOE-STD-1023-95, U.S. Department of Energy,Washington DC, 1995.

Natural Phenomena Hazards Design andEvaluation Criteria for Department of EnergyFacilities. DOE-STD-1020-94, U.S. Department ofEnergy, Washington, DC, April 1994.

Natural Phenomena Hazards PerformanceCategorization Criteria for Structures, Systems, andComponents. DOE-STD-1021-93, U.S.Department of Energy Washington, DC, 1993.

NCEER, Youd, T. L. and I. M. Idriss. Proceedingof the NCEER Workshop on Evaluation ofLiquefaction Resistance of Soils. Technical ReprotNCEER-97-0022, Buffalo, NY, December 31,1997.

Nelson, A.E., et.al. “Generalized Tectonic Map ofthe Greenville 1x2 Quadrangle, Georgia, SouthCarolina, and North Carolina.” U.S. GeologicalSurvey Miscellaneous Field Studies MapMF-1898, Scale 1:250,000, 1987.

Nelson, K.D., et.al. “New COCORP Profiling inthe Southeastern United States: Part I: LatePaleozoic Suture and Mesozoic Rift Basin.”Geology. Vol. 13, 1985.

Neuman, R.B. and Nelson, W.H. Geology of theWestern Great Smoky Mountains, Tennessee. U.S.Geological Survey Professional Paper 349-D,1965.


R-9

Noel, J.R., et.al. “Paleomagnetism and 40 Ar/39Ar Ages from the Carolina Slate Belt, NorthCarolina: Implications for Terrane Amalgamation.”Geology. Vol. 16, 1988.

Nystrom, P. G., and Willoughby, R. H., 1982,"Geological investigations related to thestratigraphy in the kaolin mining district, AikenCounty, South Carolina": in, South CarolinaGeological Survey, 1982 Carolina GeologicalSociety Field Trip Guidebook, 183 p.

Nystrom, P. G., Jr., Willoughby, R. H. and Kite, L.E., 1986, "Cretaceous-Tertiary stratigraphy of theupper edge of the Coastal Plain between NorthAugusta and Lexington South Carolina": in,Carolina Geological Society Field Trip Guidebook,South Carolina Geological Survey.

Nystrom, P.G., et.al. “Claibornian Stratigraphy ofthe Savannah River Site and Surrounding Area,1990, Savannah River Region: Transition Betweenthe Gulf and Atlantic Coastal Plains.” Proceedingsof the Second Bald Head Island Conference onCoastal Plains Geology. Hilton Head Island,November 6-11, 1990.

O’Connor, B.J. and Prowell, D.C. “The Geologyof the Belair Fault Zone and Basement Rocks ofthe Augusta, Georgia Area.” Georgia GeologicalSociety Guidebook 16. 1976.

O’Connor, B.J. and Prowell, D.C. “The Geologyof the Belair Fault Zone and Basement Rocks ofthe Augusta, Georgia Area.” Georgia GeologicalSociety Guidebook 16. 1976.

O’Connor, et.al. “Recently Discovered Faults inthe Central Savannah River Area.” abstract,Georgia Academy of Science Bulletin. Vol. 32,1974.

Obermeier, S.F., et.al. “Geologic Evidence forRecurrent Moderate to Large Earthquakes nearCharleston, South Carolina.” Science. Vol. 227,1985.

Obermeier, S.F., et.al. Earthquake-InducedLiquefaction Features in the Coastal Setting ofSouth Carolina and in the Fluvial Setting of theNew Madrid Seismic Zone. U.S. GeologicalSurvey Professional Paper No. 1504, 1990.

Olsen, P.E. “On the Use of the Term Newark forTriassic and Early Jurassic Rocks of Eastern NorthAmerica.” Newsletters on Stratigraphy. Vol. 7,1978.

Olsen, P.E. and Schlische, R.W. “Unraveling theRules of Rift Basins.” Geological Society ofAmerica Abstracts with Programs. Vol. 20, 1988.

Olsen, P.E., et.al. “Rift Basins of Early MesozoicAge." The Geology of the Carolinas. Universityof Tennessee Press, Knoxville, 1991.

Ou, G.B. and Herrmann, R.B. “A Statistical Modelfor Ground Motion Produced by Earthquakes atLocal and Regional Distances." Bulletin of theSeismological Society of America. Vol. 80, No. 6,1990.

Parsons, et.al. Bedrock Waste Storage Project,Triassic Basin Fault Probing Program Report. E. I.du Pont de Nemours and Co., Savannah RiverPlant, Aiken, SC, 1973.

Pen Branch Fault Basement Drilling.Westinghouse Environmental and GeotechnicalServices, 1991.

Petersen, T.A., et.al. “Structure of the RiddlevilleBasin from COCORP Seismic Data andImplications for Reactivation Tectonics." Journalof Geology. Vol. 92. 1984.

Phinney, R.A. and Roy-Chowdhury, K. ReflectionSeismic Studies of Crustal Structure in the EasternUnited States: Geophysical Framework of theContinental United States. Geological Society ofAmerica Mem. 172, 1989.

Poag, C.W. and Valentine, P.C. “Mesozoic andCenozoic Stratigraphy of the US AtlanticContinental Shelf and Slope." The Geology ofNorth America. V I-3, The Atlantic ContinentalMargin. DNAG Publication, Geological Society ofAmerica, Boulder, CO, 1988.

Poland, J. F., Lofgren, B. E., and Riley, F. S., 1972,Glossary of selected terms useful in studies of themechanics of aquifer systems and land subsidencedue to fluid withdrawal: U. S. Geological SurveyWater Supply Paper 2025, 9 p.

Preliminary Report on Belair Exploratory TrenchNo. 10-76 Near Augusta, Georgia. Open FileReport No. 77-411, U.S. Geological Survey,Atlanta, GA, 1977.

Price, V., et.al. Pen Branch Fault InvestigationProgram Plan. ESS-SRL-89-395, 1989.

Prowell, D. C., Christopher, R. A., Edwards, L. E.,Bybell, L. M., and Gill, H. E., 1985a, "Geologicsection of the updip Coastal Plain from centralGeorgia to western South Carolina": USGSMiscellaneous Field Studies, Map MF-1737.

Prowell, D.C. “Cretaceous and CenozoicTectonism on the Atlantic Coastal Margin, theGeology of North America, the Atlantic


R-10

Continental Margin." Geological Society ofAmerica. Vol. I-2, 1988.

Prowell, D.C. Index of Faults of Cretaceous andCenozoic Age in the Eastern United States. U.S.Geological Survey Miscellaneous Field StudiesMap MF-1269, 1983.

Prowell, D.C. Preliminary Geologic Map of theBarnwell 30’ x 60’ Quadrangle, South Carolinaand Georgia. Open File Report 94-673, U.S.Geological Survey, Atlanta, GA, 1994.

Prowell, D.C. and O’Connor, B.J. “Belair FaultZone: Evidence of Tertiary Fault Displacement inEastern Georgia." Geology. Vol. 6, 1978.

Prowell, D.C. and Obermeier, S.F. “Evidence ofCenozoic Tectonism.” The Geology of theCarolinas. University of Tennessee Press,Knoxville, TN, 1991.

Prowell, D.C., et.al. “The Ellenton Formation inSouth Carolina: A Revised Age Designation fromCretaceous to Paleocene.” U.S. Geological SurveyBulletin. No. 1605-A, 1985.

Prowell, D.C., et.al. Preliminary Evidence forHolocene Movement Along the Belair Fault ZoneNear Augusta, Georgia. Open File Report 75-680,U.S. Geological Survey, Atlanta, GA, 1976.

Rast, N. and Kohles, K.M. “The Origin of theOcoee Supergroup.” American Journal of Science.V. 286, 1986.

Ratcliffe, N.E. “The Ramapo Fault System in NewYork and Adjacent Northern New Jersey: A Caseof Tectonic Heredity." Geological Society ofAmerica Bulletin. Vol. 82, 1971.

Reactor Site Criteria. 10 CFR 100, Washington,DC, November 15, 1983.

Reactor Site Criteria. 10 CFR 100, Washington,DC, November 15, 1983

Rozen, R.W. “The Middleton-LowndesvilleCataclastic Zone in the Elberton East Quadrangle,Georgia.” Geological Investigations of the KingsMountain Belt and Adjacent Areas in theCarolinas. Carolina Geological SocietyGuidebook, 1981.

Sacks, P.E. and Dennis, A.J. “The ModocZone-D2 (Early Alleghanian) in the EasternAppalachian Piedmont, South Carolina andGeorgia, Anatomy of the Alleghanian Orogeny asSeen from the Piedmont of South Carolina andGeorgia.” Carolina Geological Society Field Trip

Guidebook. South Carolina Geological Survey,1987

Samson, S., et.al. “Biogeographical Significanceof Cambrian Trilobites from the Carolina SlateBelt.” Geological Society of America Bulletin.Vol. 102, 1990.

Sawyer, D.S. “Thermal Evolution." The Geologyof North America. Vol. I-2, The AtlanticContinental Margin. DNAG Publication,Geological Society of America, Boulder, CO,1988.

Schlische, R.W. and Olsen, P.E. “QuantitativeFilling Model for Continental Extensional Basinswith Application to the Early Mesozoic Rifts ofEastern North America.” Journal of Geology.1992.

Schmertmann, J. H. “The Mechanical Aging ofSoils,” The 25th Terzaghi Lecture, Journal ofGeotechnical Engineering, Vol. 117, No. 9, ,American Society of Civil Engineers, September1991.

Schmertmann, J. H. “Update on the MechanicalAging of Soils,” for the symposium “SobreEnvejecimiento de Suelos,” The Mexican Societyof Soil Mechanics, Mexico City, August 1993.

Scholz, C.H. The Mechanics of Earthquakes andFaulting. Cambridge University Press, 1990.

Secor, D.T., Jr. “Regional Overview, Anatomy ofthe Alleghanian Orogeny as Seen from thePiedmont of South Carolina and Georgia.”Carolina Geological Society Field Trip Guidebook.South Carolina Geological Survey, 1987.

Secor, D.T., Jr. et.al. “Character of theAlleghanian Orogeny in the SouthernAppalachians, Part I: Alleghanian Deformation inthe Eastern Piedmont of South Carolina.”Geological Society of America Bulletin. Vol. 97,1986a.

Secor, D.T., Jr., et.al. “Character of theAlleghanian Orogeny in the SouthernAppalachians: Part III, Regional TectonicRelations.” Geological Society of AmericaBulletin. Vol. 97, 1986b.

Secor, D.T., Jr., et.al. “Confirmation of CarolinaSlate Belt as an Exotic Terrane.” Science.Vol. 221. 1983.

Seed, H.B. “Soil Liquefaction and Cyclic MobilityEvaluation for Level Ground During Earthquakes."Journal of Geotechnical Engineering Division.


R-11

Vol. 105, No. 2, American Society of CivilEngineers, February 1979.

Seed, R.B., Idriss, I. M. and I. Arango.“Evaluation of Liquefaction Potential Using FieldPerformance Data.” Journal of GeotechnicalEngineering Division. Volume 109, No. 3,American Society of Civil Engineers, 1983.

Seismograph Services Corporation. Gravity andMagnetic Survey, Savannah River Plant, SouthCarolina. Unpublished report prepared for E.I. duPont de Nemours and Co., Inc., EngineeringDepartment, Wilmington, DE, 1972.

Sen, A.K. and Coruh, C. Removing Near-SurfaceEffects in Seismic Data: Application forDetermination of Faults in the Coastal PlainSediments. WSRC-TR-92-169, WestinghouseSavannah River Company, Aiken, SC, 1992.

Sheridan, R.E. and Grow, J.A. “The Geology ofNorth America." The Atlantic Continental Margin.Vol. I-1 and I-2, DNAG Publication, GeologicalSociety of America, Boulder, CO, 1988.

Shervais, J.W., Mauldin, J., Dennis, A.J., 1996,“Petrology and Geochemistry of Neo- ProterzoicArc Plutons beneath the Atlantic Coastal Plain:Savannah River Site, South Carolina”. ReportSubmitted for SCURF Task 170.

Sibol, M.S. and Bollinger, G.A. EarthquakeCatalog for the Southeastern United States,1698-1989. Virginia Polytechnic Institute andState University Seismological ObservatoryComputer File, Blacksburg, VA, 1990.

Sibol, M.S. and Bollinger, G.A. EarthquakeCatalog for the Southeastern United States,1698-1989. Virginia Polytechnic Institute andState University Seismological ObservatoryComputer File, Blacksburg, VA, 1990.

Sibson, R.H. “Earthquakes and Rock Deformationin Crustal Fault Zones.” Annual Review of EarthPlanet Sciences. Vol. 14, 1986.

Sibson, R.H. “Frictional Constraints to Thrust,Wrench, and Normal Faults.” Nature. Vol. 249,1984.

Siple, G.E. “Geology and Ground Water of theSavannah River Plant and Vicinity, SouthCarolina." Paper No. 1841, United StatesGeological Survey Water Supply. 1967.

Skempton, A. W. “Standard Penetration TestProcedures and the Effects in Sands of OverburdenPressure, Relative Density, Partical Size, Aging

and Overconsolidation,” Geotechnique, Vol. 36,No. 3, 1986.

Smoot, J.P. “The Closed-Basin Hypothesis and ItsUse in Facies Analysis of the NewarkSupergroup.” Proceedings of the Second U.S.Geological Survey Workshop on the EarlyMesozoic Basins of the Eastern United States.U.S. Geological Survey Circular 946, 1985.

Snipes, D. S., et.al. “The Pen Branch Fault:Documentation of Late Cretaceous-TertiaryFaulting in the Coastal Plain of South Carolina.”Southeastern Geology. Vol. 33, No. 4, 1993.

Snoke, A.W. and Frost, B.R. “Exhumation of HighPressure Pelitic Schist, Lake Murray Spillway,South Carolina: Evidence for Crustal ExtensionDuring Alleghanian Strike-Slip Faulting.”American Journal of Science. Vol. 280, 1990.

Snoke, A.W., et.al. “Deformed Hercynian GraniticRocks from the Piedmont of South Carolina."American Journal of Science. Vol. 280, 1980.

Sohl, N. F., and Christopher, R. A., 1983, TheBlack Creek-Peedee formational contact (UpperCretaceous) in the Cape Fear River region ofNorth Carolina: U. S. Geological SurveyProfessional Paper 1285, 37 p.

Somerville, P.G., et.al. Comparison of SourceScaling Relation of Eastern and Western NorthAmerican Earthquakes. BSSA, 77, #2, 322-346.1987.

Somerville, P.G., et.al. Comparison of SourceScaling Relation of Eastern and Western NorthAmerican Earthquakes. BSSA, 77, #2, 322-346.1987.

Steckler, M.S., et.al. “Subsidence and BasinModeling at the U.S. Atlantic Passive Margin."The Geology of North America. Vol. I-2, TheAtlantic Continental Margin. DNAG Publication,Geological Society of America, Boulder, CO,1988.

Steltenpohl, M.G. “Kinematics of the Towaliga,Bartletts Ferry, and Goat Rock Fault Zones,Alabama: The Late Paleozoic Dextral ShearSystem in the Southernmost Appalachians.”Geology. Vol. 16, 1988.

Stephenson, D.E. Savannah River PlantEarthquake. DPST-88-841, E.I. du Pont deNemours and Co., Savannah River Laboratory,Aiken, SC, August 1988.

Stephenson, D.E. and Stieve, A. Structural Modelof the Basement Based on Geophysical Data in the


R-12

Central Savannah River Area, South Carolina andGeorgia. WSRC-TR-92-120, WestinghouseSavannah River Company, Aiken, SC, 1992.

Stephenson, D.E., et.al. “Savannah River SiteDisaggregated Seismic Spectra.” Fourth DOENatural Phenomena Hazards MitigationConference. U.S. Department of Energy,Washington, DC, 1993.

Stephenson, D.E., P. Talwani, J. Rawlins,Savannah River Plant Earthquake of June 1985.DPST-85-583, E.I. du Pont de Nemours & Co.,Savannah River Laboratory, Aiken, SC, 1985.

Stephenson, D.S. and Chapman, W.L. “StructureAssociated with the Buried Dunbarton Basin,South Carolina from Recent Seismic Data.”Geological Society of America, SoutheasternSection Abstracts with Programs. Vol. 20, 1988.

Stieve, A.L. Pen Branch Fault Program: InterimReport on the High Resolution, Shallow SeismicReflection Surveys. WSRC-TR-91-30,Westinghouse Savannah River Company, Aiken,SC, 1991b.

Stieve, A.L. and Stephenson, D.E. “GeophysicalEvidence for Post Late Cretaceous Reactivation ofBasement Structures in the Central Savannah RiverArea.” Southeastern Geology. Vol. 35, No. 1,1995.

Stieve, A.L. et.al. Confirmatory Drilling ProjectFinal Report. WSRC-RP-94-0136, WestinghouseSavannah River Company, Aiken, SC, 1994.

Stieve, A.L. et.al. Confirmatory Drilling ProjectFinal Report. WSRC-RP-94-0136, WestinghouseSavannah River Company, Aiken, SC, 1994.

Stieve, A.L., et.al. Pen Branch Fault Program:Consolidated Report on the Seismic ReflectionSurveys and the Shallow Drilling.WSRC-TR-91-87, Westinghouse Savannah RiverCompany, Aiken, SC, 1991a.

Stokoe, K.H., et.al. Correlation Study of NonlinearDynamic Soil Properties: Savannah River Site,Aiken, South Carolina. Rev. 0, File No.SRS-RF-CDP-95, University of Texas at Austin,Department Civil Engineering, September 13,1995.

Stringfield, V. T., 1966, Artesian water in Tertiarylimestone in the Southeastern States: U.S.Geological Survey Professional Paper 517, 226 p.

Strom, R. N., and Kaback, D. S., 1992, SRPbaseline hydrogeologic investigation: aquifercharacterization, groundwater geochemistry of the

Savannah River Site and vicinity: WSRC-RP-92-450, Westinghouse Savannah River Company,Aiken, SC 29808, 96 p.

Talwani, P. “Current Thoughts on the Cause of theCharleston, South Carolina Earthquakes.” SouthCarolina Geology. Vol. 29, 1985.

Talwani, P. “Seismotectonics of the CharlestonRegion.” Proceedings, National Conference onEarthquake Engineering, 3rd. EarthquakeEngineering Research Inst., Vol. 1, 1986.

Talwani, P. The Woodstock Fault is Alive andTicking Near Charleston, South Carolina.Submitted to the Bulletin of the SeismologicalSociety of America, 1982.

Talwani, P. and Cox, S. “An Internally ConsistentPattern of Seismicity Near Charleston, SouthCarolina.” Geology. Vol. 10, 1985.

Talwani, P. and Cox, S. “Paleoseismic Evidencefor Recurrence of Earthquakes near Charleston,South Carolina." American Association for theAdvancement of Science. Vol. 229, July 1985.

Talwani, P., “Crustal Structure of South Carolina.”Second Technical Report to the U.S. GeologicalSurvey, contract #14-08-0001-14553, 1975.

Talwani, P., B.K. Rastogi, and D.A. Stevenson,“Induced seismicity and earthquake predictionstudies in South Carolina.” Tenth TechnicalReport to the U.S. Geological Survey, contract#14-08-0001-17670, 1980.

Talwani, P., et.al. Induced Seismicity andEarthquake Prediction Studies in South Carolina.Tenth Technical Report to the USGS, Contract14-08-001-17670.1980.

Talwani, P., J. Rawlins, D. E. Stephenson. “TheSavannah River Plant, South Carolina, Earthquakeof June 9, 1985 and its Tectonic Setting.”Earthquake Notes. Vol. 56, No. 4, 1985.

Talwani, P.J., et.al. “The Savannah River Plant,South Carolina, Earthquake of June 9, 1985 and itsTectonic Setting.” Earthquake Notes. Vol. 56, No.4, 1985.

Tarr, A., P. Talwani, S. Rhea, D. Carver, and D.Amick. “Results of Recent South CarolinaSeismological Studies." Bulletin of theSeismological Society of America. Vol. 71, 1981.

Tarr, A.C., et.al. “Results of Recent SouthCarolina Seismological Studies." Bulletin of theSeismological Society of America. Vol. 71, 1981.


R-13

URS/John A. Blume and Associates, Engineers.Update of Seismic Criteria for the Savannah RiverPlant, Vol. 1 of 2, Geotechnical. USR/JAB8144,San Francisco, CA. Prepared for E.I. du Pont deNemours and Company as DPE-3699, SavannahRiver Plant, Aiken, SC, 1982.

US Army Corps of Engineers (COE), CharlestonDistrict. “Geologic Engineering Investigations,Savannah River Plant”, Volumes 1 and 2,Waterways Experimental Station, Vicksburg, MS,1952.

Vick, H.K., et.al. “Ordovician Docking of theCarolina Slate Belt.” Paleomagnetic Data:Tectonics. Vol. 6, 1987.

Visvanathan, T.R., “Earthquakes in SouthCarolina, 1698-1975.” South Carolina Geol. Surv.Bull. , vol. 40, 61p., 1980.

Wait, R. L., and Davis, M. E., 1986, Configurationand hydrology of the pre-Cretaceous rocksunderlying the Southeastern Coastal Plain aquifersystem: U.S. Geological Survey Water-ResourcesInvestigations Report 86-4010, 1 sheet.

Walton, W. C., 1970, Ground Water ResourceEvaluation, McGraw-Hill Book Co., New York, N.Y., 664 p.

Wang, W. Some Findings in Liquefaction. WaterConservancy and Hydroelectric Power ScientificResearch Institute, Beijing, China, August 1979.

Webster, D. S., Proctor, J. R., and Marine, I. W.,1970, Two-well tracer test in fractured crystallinerock: U.S. Geological Survey Water-Supply Paper1544-I.

Wehr, F. and Glover, L., III. “Stratigraphy andTectonics of the Virginia-North Carolina BlueRidge: Evolution of a Late Proterozoic-EarlyPaleozoic Hinge Zone." Geological Society ofAmerica Bulletin. Vol. 96, 1985.

Wentworth, C.M. and Mergener, K.M.“Regenerate Faults of the Southeastern UnitedStates.” Studies Related to the Charleston, SouthCarolina, Earthquake of 1886, Tectonics andSeismicity. U.S. Geological Survey ProfessionalPaper 1313, 1983.

Winkler, C.D. and Howard, J.D. “Correlation ofTectonically Deformed Shorelines on the SouthernAtlantic Coastal Plain.” Geology. Vol. 5, 1977.

Wyatt. D. E., Aadland, R. K., Cumbest, R. J.,Stephenson, D. E., Syms, F. H., 1997 A/M AreaAdvanced Geological Study, Part 3 of 3,Geological Interpretation of the Structure and

Stratigraphy of the A/M Area, Savannah RiverSite, South Carolina (U), WSRC-RP-97-0186, Rev.0, 46 p.

Youd, T.L. “Liquefaction, Flow and AssociatedGround Failure.” U.S. Geological Survey Circular.No. 688, 1973.

Youd, T.L. and Hoose, S.N. “LiquefactionSusceptibility and Geologic Setting.” Proceedings,6th World Conference on Earthquake Engineering.New Delhi, India, Vol. III, 1977.

Zoback, M.D. and Zoback, M.L. Tectonic StressField of North America and Relative Plate Motionsin Neotectonics of North America. Vol. I, DecadeMap, Geological Society of America, Boulder, CO,1991.

Zoback, M.L. and Zoback, M.D. “State of Stressin the Conterminous United States.” Journal ofGeophysical Research. Vol. 85, 1980.


R-14

Notes

srs field guide

Documents

sc naplogic

nc csra geological society

united states government

sc blue ridge environmental

department of energy

field trip guidebook

department of commerce

south carolina