rr3i9i3i - semspub.epa.gov

tDEPARTMENT OF HEALTH & HUMAN SERVICES Public Health SerJicP *l.

Agency for Toxic Substancesand Disease Registry

Atlanta GA 30333March 25, 1998

Mr. Gregg CrystallU.S. Environmental Protection AgencyRegion inHazardous Waste Management (3HWOO)841 Chestnut Building, 9th FloorPhiladelphia, PA 19107

Dear Mr. Crystall:

Please find enclosed a final copy of Health Consultation No. 2 and a public comment copy ofHealth Consultation No. 3 for the Drake Chemical Superfund Site Thermal Destruction Facility,Lock Haven, Pennsylvania, dated March 25, 1998.

ATSDR will receive comments regarding Health Consultation No. 3 for 30 days, endingMay 3,1998. That health consultation addresses concerns regarding whether the levels of stackemissions and fugitive emissions produced at the Drake site during the trial burn periodwere sufficient to cause any adverse public health effects.

ATSDR has also enclosed a fact sheet that summarizes the findings and recommendationsof Health Consultations No. 2 and No. 3.

Please address any comments or correspondence regarding these two documents to theChief, Program Evaluation, Records, and Information Services Branch, Division of HealthAssessment and Consultation, Agency for Toxic Substances and Disease Registry, ATTN:Drake Chemical Superfund Site, 1600 Clifton Road, NE (E56), Atlanta, Georgia 30333.

If you have any immediate questions, you may contact ATSDR at 1-800-447-1544. Thankyou for your interest.

Sincerely yours,

Max M. Howie, Jr.Chief, Program Evaluation, Records,and Information Services Branch

Division of Health Assessmentand Consultation

Enclosures

You May Contact ATSDR TOLL FREE at1-800-447-1544 or

Visit our Home Page at: http://atsdrl.atsdr.cdc.gov:8080/

RR3I9I3I

Petitioned Health Consultation No. 2

Thermal Destruction Facility

DRAKE CHEMICAL

LOCK HAVEN, CLINTON COUNTY, PENNSYLVANIA

CERCLIS NO. PAD003058047

MARCH 25, 1998

U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICESAgency for Toxic Substances and Disease RegistryDivision of Health Assessment and Consultation

Atlanta, Georgia

A R 3 I 9 I 3 2

Health Consultation: A Note of Explanation

An ATSDR health consultation is a verbal or written response from ATSDR to a specificrequest for information about health risks related to a specific site, a chemical release,or the presence of hazardous material. In order to prevent or mitigate exposures, aconsultation may lead to specific actions, such as restricting use of or replacing watersupplies; intensifying environmental sampling; restricting site access; or removing thecontaminated material.

In addition, consultations may recommend additional public health actions, such as conductinghealth surveillance activities to evaluate exposure or trends in adverse health outcomes;conducting biological indicators of exposure studies to assess exposure; and providing healtheducation for health care providers and community members.

This document has previously been released for a 30 day public comment period. Subsequentto the public comment period, ATSDR addressed all public comments and revised orappended the document as appropriate. The health consultation has now been reissued. Thisconcludes the health consultation process for this site, unless additional information isobtained by ATSDR which, in the Agency's opinion, indicates a need to revise or append theconclusions previously issued.

You May Contact ATSDR TOLL FREE at1-800-447-1544

orVisit our Home Page at: http://atsdrl.atsdr.cdc.gov:8080/

A R 3 I 9 I 3 3

PETITIONED HEALTH CONSULTATION NO. 2

Thermal Destruction Facility

DRAKE CHEMICAL



Prepared by:

Petition Response SectionExposure Investigation and Consultation BranchDivision of Health Assessment and ConsultationAgency for Toxic Substances and Disease Registry

&R3I9131*

Drake Chemical - Final Health Consultation #2

Foreword

This health consultation evaluates the incinerator operational plans, trial burn protocols, andemissions modeling. Health Consultation #3 will provide an evaluation of data from actualsampling and monitoring of the incinerator during the trial bums and risk burns. In HealthConsultation #3, the Agency for Toxic Substances and Disease Registry (ATSDR) will alsoevaluate the ambient air data for public health implications.

Background and Statement of Issues

This health consultation is the second in a series of consultations ATSDR is conducting to addresspublic health issues relating to the Drake Chemical Superfund site in Lock Haven, Pennsylvania.The consultation evaluates emissions and modeling information used in the trial burn riskassessment and the operational plans and protocols for the incinerator. The first healthconsultation evaluated the trial burn risk assessment (excluding incinerator emissions andmodeling information) and the ambient air sampling plans and activities [1]. ATSDR refersreaders to that health consultation for general background information on the site. In the future,ATSDR will evaluate the public health implications of air and stack emissions sampling conductedduring the trial burn process.

Since the incinerator trial burn is complete, and ATSDR did not become involved at the site untilafter the incinerator was operational, we have reviewed the documents to (1) evaluate the designand operational conditions that might affect public exposure and (2) look at the modeling and riskassessment approach. In this consultation, we comment on the risk assessment modeling approachand discuss changes we would like to have incorporated into the risk assessment for full operationbased on the trial burn stack emissions testing. We also comment on the planned operatingconditions, identified in the documents, that may affect public exposure. This consultation is not acritical review of the projected emissions for the trial burn period and the risk associated withthose projected emissions. ATSDR does not consider it efficient to review and comment onhypothetical projected emissions when the actual emissions data from the trial burn and risk burnwill soon be available. In another consultation, we will evaluate the emissions data and operationalconditions during the trial burn and comment on the potential health effects of full operation basedon the actual stack emissions. Comments provided in this consultation are intended to help theEnvironmental Protection Agency (EPA) develop the risk assessment for full operation.

Discussion

The comments in this consultation are based on ATSDR staff review of the following documentsand the comments provided by an ATSDR modeler [2]:

Drake Chemical Site Incinerator Trial Burn Risk Assessment, Volumes I - HI and Appendices[3]-

1

AR3I913S


Trial Burn Plan for the Drake Chemical Superfund Site's Mobile Hazardous Waste Incinerator,Volumes I and II [4].

Facility Operations Description

The Drake incinerator is a transportable rotary kiln incinerator system owned by OHMRemediation Service Corporation. The soils to be treated are stockpiled in a feed buildingmaintained under negative pressure. The soils are conveyed at a maximum feed rate of 60 tons perhour to the rotary kiln (primary combustion chamber). The organic contaminants are volatilizedand the combustion process begins in the rotary kiln which is operated at 1,200°F. The organicsand any partially combusted compounds complete the combustion process in the secondarycombustion chamber, which is operated at 1,800°F. A thermal relief valve (TRV) is at the top ofthe secondary combustion chamber. Cyclones between the kiln and the secondary combustionchamber remove larger particulate matter (entrained soil) and combine it with the treated soilfrom the kiln and the fly ash from the secondary combustion chamber. The combined treated soiland fly ash are sprayed with water to control fugitive emissions and are conveyed to a coveredstorage area, where they are stored until analysis is completed.

An evaporative cooler after the secondary combustion chamber reduces the gas temperaturebefore the gas goes to the baghouse for removal of smaller particulate matter. The gases arefurther cooled and scrubbed in a venturi quench before going to a caustic scrubber, whichremoves and neutralizes any acid gases before releasing the exhaust gases to the atmosphere fromthe 150-foot stack. Fly ash from the evaporative cooler and from the baghouse is transported inan enclosed conveyor and sprayed with water to control fugitive emissions before it is dischargedin the ash storage area. See Figures 1 and 2 [3].

According to Drake Chemical Site Incinerator Trial Burn Risk Assessment, Volume n, Appendix3B [3] and the Air Quality Equivalency Document issued by the Commonwealth of Pennsylvania[5], the following conditions initiate the automatic shutdown of all waste feed materials to thekiln:

• Kiln-firing hood pressure greater than -0.1 inch water column for ten seconds; or greaterthan zero inches water column instantaneously.

• Secondary combustion chamber discharge temperature less than 1,800 °F.• Carbon monoxide concentration in stack gas exhaust greater than 100 parts per million by

volume (ppmv) for one hour average or greater than 500 ppmv instantaneously or loss ofthe signal.

• Stack gas flow greater than 60 feet per second (or a modified value to be determined fromresults of the trial bums).

• Failure of both secondary combustion chamber burners.• Caustic/water scrubbing solution flow to either of the packed bed acid gas scrubbers at a

rate less than 450 gallons per minute.

A R 3 I 9 I 3 6


• Acid gas scrubbing solution pH less than 6.0 for one hour.• Differential pressure across the baghouse less than 1.0 inch water column.• Nitrogen oxide emissions greater than 300 parts per million by volume (daily average) or

loss of signal.• Kiln discharge gas temperature less than 1,000°F instantaneously or the hourly rolling

average less than 1,200°F.• Failure of both kiln burners.• Kiln dry ash conveyor stopped.

• • Kiln rotation less than 0.4 revolutions per minute.• TRV on secondary combustion chamber not closed.• Solids feed rate hourly average greater than the maximum rate demonstrated during the

trial burn.

Even though the waste feed will be discontinued when these conditions occur, the flames will bemaintained and the air pollution control equipment will continue to operate; therefore it is unlikelythat there would be any increase in stack emissions.

The following conditions will initiate an emergency system shutdown that stops the waste feed,fuel feeds, and induced draft fan and opens the TRV:

• Induced draft fan failure.• Evaporative cooler discharge gas temperature greater than 500°F.• Electric power failure.• . Acid gas scrubber inlet temperature greater than 250°F.

When the TRV opens, there will be an initial puff of emissions consisting of the gases in thesecondary combustion chamber and the rotary kiln. The organics released will likely consist offully combusted decomposition products, partially combusted organics, and organics justvaporized from the soil. Because the gases will bypass the ah" pollution control equipment, therewill be no removal of metals, particulate matter, or acid gases. As the temperature, pressure, andair flow decrease, the TRV stack emissions will decrease rapidly. The prediction is that they will.approach zero within 25 minutes of the TRV opening.

When the gas temperature at the exit of the secondary combustion chamber is greater than2,600°F, the system automatically shuts down all waste feed and the fuel to both combustionchambers. The hot gases will continue to flow through and be treated in the air pollution controlequipment for removal of the particulate and acid gases; however, there will be an increase inorganic emissions when the fuel is shut off, causing the flame to go out.

When the kiln discharge temperature exceeds 2,200°F, the system will automatically shut downall waste feed and the fuel to the kiln only; the flame will be maintained in the secondarycombustion chamber. Because combustion will continue to occur in the secondary combustion .

HR3I9I37


chamber and the gases will still be treated in the air pollution control equipment, stack emissionsare not expected to increase.

The Pennsylvania Air Quality Equivalency Document, Section 7, dictates continuous monitoringand recording of the key operating parameters. All of the conditions that will cause an automaticwaste feed shut off (AWFSO) or TRV opening are continuously monitored and recorded. Thedocument also dictates that the following stack emissions be continuously monitored and that databe recorded:

CO carbon monoxide concentration,THC total hydrocarbon concentration,02 oxygen concentration, andNOX nitrogen oxides concentration.

Continuous monitoring and recording of all key operating conditions and parameters allow thosewho oversee the incinerator operations to verify how the facility was operated and whatconditions triggered upset conditions at any point.

The trial bum operating conditions were designed to represent maximum input of soil taken fromthe contaminated areas and fed to the incinerator operating at minimum combustion temperatureand maximum combustion gas flow rate [4]. These conditions represent worst case operatingconditions. If the stack emissions measured during the trial burn and risk burn do not pose anunacceptable health risk to the community, then operation of the incinerator under normaloperating conditions should not pose a health risk.

The trial bum plan appeared to be well designed and specified standard or updated draftEnvironmental Protection Agency (EPA) sampling and analytical procedures and quality controlmeasures. If the trial burn and risk burn were conducted in accordance with the plan, the data •should be adequate for evaluating the public health implications of the incinerator.

General Comments

• The" approach to estimating emissions from the incinerator for the trial burn period wasconservative and very thorough.

• The term "hazard index" should be defined and explained on first use. The public may notbe familiar with the term.

Dispersion Modeling

The air quality dispersion modeling analysis performed in the trial burn risk assessment used theIndustrial Source Complex Version 3 (ISC3) model, CALMET/CALPUFF models, and the

A R 3 I 9 I 3 8


INPUFF model to predict the ambient air concentrations of contaminants expected to be releasedfrom the stack during the trial burn. Dispersion of emissions caused by a possible upset condition,which was defined as a TRV opening, were also modeled to determine the potential ambientimpacts. The modeling of stack emissions predicted the monthly average ambient concentrationsand ground-level deposition in the area extending 13 miles (20 kilometers) from the Drake site.Modeling predicted one hour and monthly average ambient concentrations of contaminants for theupset release condition. Modeling also provided one hour and eight hour average ambientconcentration predictions for carbon monoxide [3].

Because the modeling was conducted before the incinerator was constructed and operated on site,the designed physical parameters for the incinerator were used. The air quality modeling analysisreflected surface data from the Williamsport, Pennsylvania, National Weather Service station;upper air data from the Pittsburgh, Pennsylvania, National Weather Service station; andprecipitation data from Philipsburg, Pennsylvania. Although on-site meteorological data exist,EPA determined the data inadequate for use in the modeling analysis because of instrumentmalfunctions and improper sensor exposure, which together resulted in unrepresentative windspeed data for the Lock Haven area [3].

The following comments assume that the same general modeling protocol will be used for the riskassessment for full operation except (1) modeling will cover the full year; (2) the modeling willreflect local meteorology as much as possible; (3) stack emissions data from the trial burn will bemodeled; and (4) the trial burn worst case operational data, such as stack gas temperature andvelocity, will be used.

• Modeling for potential acute effects should reflect the maximum hourly and maximumdaily ambient concentrations.

• A comparison of the Williamsport wind data and the actual site-specific wind data for themonths that the data are available will provide additional scientific support to the use ofthe Williamsport meteorological data. A time histogram or auto-correlation may be usefulfor comparing the two data sets. . .

• The use of five years of meteorological data will help assure that likely worst case publicexposure concentrations are part of the evaluation.

• Are there other sources of local meteorological data, such as the Piper Memorial Airport,the state college, other industrial air sources, river and stream flow stations, etc.? Adefinitive statement as to the unavailability or unsuitability of any such data (and whatattempts were made to locate other local data sources) would strengthen the full operationrisk assessment.

AR3I9I39


• The uncertainty and sensitivity analyses do not present quantitative assessment of the airquality modeling results. One approach to quantifying the uncertainty and sensitivity issueswould be to exercise one of the models at a few locations, systematically varying thevarious input assumptions. This would yield an ensemble of model predictions that couldhelp health scientists assess the values used in the risk assessment.

• Volume m, Section 5.3, of the risk assessment assumes that the particulate releasedduring a TRV event would be primarily metals (not organics) and that they would berelatively large and thus would deposit on site. This assumption is contrary to theinformation in Table 2.4-2 of Volume I of the risk assessment and Volume n, Section3B.3.3. These sections say all the metals except barium, beryllium, and chromium will be100% volatilized during normal and upset conditions. Barium will be 50% volatilized, andberyllium and chromium will be 5% volatilized. It seems reasonable to assume that most ofthe metals will be in the vapor state when exiting the TRV and will be dispersed in theplume rather than being large particles that fall on site. •

• EPA should provide references to support the following assumptions: (1) the assumptionthat after an initial puff of one minute duration, the emissions will decrease exponentiallyto zero within 25 minutes and, (2) the assumption that during the first minute thedestruction and removal efficiency will decrease from 99.99% to 99%.

Note: This health consultation evaluates the incinerator operational plans, trial burnprotocols, and emissions modeling. Health Consultation #3 will provide an evaluation ofdata from actual sampling and monitoring of the incinerator during the trial burns andrisk burns. In Health Consultation #3, ATSDR will also evaluate the ambient air data forpublic health implications.

Conclusions

• Automatic waste feed shut-offs (AWFSOs) were set on the key operating conditions, andthe documents specified that the flame was to be maintained whenever it was safe to doso. The operating conditions and AWFSOs should ensure safe operation of the incineratorand prevent or minimize to the maximum extent possible exposure of the community.

• The continuous monitoring and recording of the key operating conditions and stackemissions will provide continuous documentation of the operation of the incinerator andthe causes of any releases that may occur.

• The circumstances specified when the thermal relief valve (TRV) will be opened are onlythose that constitute true emergency situations, when it would be more protective of

AR3I9U40


public health to vent the hot gases through the TRV than to allow catastrophic equipmentfailure that would release the hot gases at ground level.

Estimates of TRV emissions for full operation and modeling of those values would helpinvestigators predict the public health impacts of upset conditions.

New dispersion modeling would be useful for evaluating the stack emissions measuredduring the trial burn. Dispersion modeling to date is not adequate to project the ambientair concentrations in the communities around the Drake incinerator during full operation.

Recommendations

1. Conduct new modeling using five years of meteorological data and stack emissions ratesmeasured during the trial burn and risk burn. Agency for Toxic Substances and DiseaseRegistry (ATSDR) staff should review the draft modeling protocol before the modeling isconducted so that our input can be more pertinent and timely.

2. Reevaluate and provide references for the following assumptions used in projecting theemissions that will be released when operating conditions cause the thermal relief valve(TRV) to open: (1) the assumption that after an initial puff of one minute duration theemissions will decrease exponentially to. zero within 25 minutes and, (2) the assumptionthat during the first minute the destruction and removal efficiency will decrease from99.99% to 99%. •

Note: This health consultation evaluates the incinerator operational plans, trial burnprotocols, and emissions modeling. Health Consultation #3 will provide an evaluation ofdata from actual sampling and monitoring of the incinerator during the trial burns andrisk burns. In Health Consultation #3, ATSDR will also evaluate the ambient air data forpublic health implications.

A R 3 I 9 I U I


Preparers of Report

Betty C. Willis, MSEnvironmental Health ScientistProgram Evaluation, Records, and Information Services Branch

Gregory M. Zarus, MSEnvironmental Health ScientistExposure Investigation and Consultation Branch

8

AR3I9U2


References

1, Agency for Toxic Substances and Disease Registry. Health consultation for Drake ChemicalSuperfund site thermal destruction facility, Lock Haven, Pennsylvania. Atlanta: USDepartment of Health and Human Services, Public Health Service. 1997 July 24.

2. Dabberdt WF. Draft comments on Drake Chemical site incinerator trial burn risk assessment.1997 Jan 24 . ,

3. Roy F. Weston, Inc. Drake Chemical site incinerator trial bum risk assessment (volumes I, E,HI). Prepared for Environmental Protection Agency Region HI. 1996 June 21.

4. Midwest Research Institute. Trial burn plan for the Drake Chemical Superfund site's mobilehazardous waste incinerator, volumes I and n. Prepared for OHM Remediation ServicesCorp. 1994 Oct 24 (revised 1996 Sept 20).

/5. Commonwealth of Pennsylvania, Department of Environmental Protection, Bureau of Air .

Quality. Air quality equivalency document. 1995 Dec 8.

AR3I9U3


FIGURES

10

AR3I9IM*

EXHAUSTSTACK AH

POLLUTION/ x CONTROL

SYSTEMMETEOROLOGICALSTATION

FEEDPREPARATIONBULD1NG

ROTARYKILN .• < I

/

N3COOASHHANDLINGBUILDING

AMERICAN COLORAND

CHE.M CAL COMPANY

FEED ANDASH CONVEYORSYSTEH

ASHSTABILIZATIONBUILDING

DRAKECHEMICAL

SITEPROPANSSTORAGHAREA A WATER

TREATMENTFACILITY

DRUMMEDCAUSTICSTORAGE

— — — — Sits bojndaryRailroad tracksExisting 8' natural gas line

-J^ Locauon current meBoodogical

HAUUEHMIU.PAPER COMPANY

Note:Location of sits features s approximate only.This plan is not to be construed as a survey map

Drake Chemical S'rteLock Haven, PennsylvaniaGRID NORTH

SO 100 150 200

RGURE1SITE PLOT PLANSOURCE: SCALE INFECT"

Aeapted (ram DraXa Chemical Supertund S!» - Dwg. No. 220521-C-01 by OHMRemedanon Services Corp.. not dated - supcfied by EPA

11 AR319IU5

12 AR3I9U6


APPENDIX A

13

AR3I9U7


RESPONSE TO COMMENTS RECEIVED

Comments were received on this health consultation from a Lock Haven resident and from EPARegion IH EPA comments were responses or explanations for a number of issues raised in thisconsultation. Their comments are attached. If changes were made based on their comments,ATSDR has added a response noting those changes. We also added responses to several of theissues where further explanation of ATSDR's position was needed.

The resident submitted twenty-two newspaper articles and two reports from the GovernmentAccountability Project for our review and requested that we also consider the report issued byBob Martin, the EPA-Washington ombudsman. She also requested that ATSDR collect healthinformation from the local hospitals on cancers, lung problems, skin problems, and birth problemsincluding difficulties, defects, deaths, and miscarriages. She requested that this" data be used as abaseline for conducting a longitudinal health study of the Lock Haven area. The commenter alsoquestioned whether history would support that incineration is a safe technology or that it is adangerous practice that should be banned.

Occupational and community health studies have been conducted for the Lock Haven area for anumber of years to evaluate health outcomes related to the workers' and community's exposureto p-naphthylamine (BNA) resulting from the Drake Chemical plant's operation and productionof BNA. Further health studies do not seem warranted at this time. Because of limited resourcesand the expenses involved in conducting a longitudinal health study, ATSDR has developed

• criteria for when it is appropriate to conduct a health study. One of the key criteria is that thereshould be documented community exposure to contaminants at levels of public health concern. Inthe draft Health Consultation #3, ATSDR evaluates the potential public health effects from theincinerator. In that consultation, ATSDR concluded that the community is not likely to experienceadverse health outcomes due to exposure to contaminants from the Drake Chemical Superfundsite or emissions from the incinerator being used to remediate that site. If new information isprovided that indicates exposure of the public to contaminants at levels of health concern, theAgency will re-evaluate the need for conducting additional health studies in the Lock Havencommunity.

Incineration is a technology that has been around for hundreds of years. History so far has notindicated that incineration is a dangerous practice that should be banned. There has been at leastone documented instance where a commercial hazardous waste incineration facility which had noair pollution control equipment and operated under poor combustion practices did cause adversehealth effects in nearby residents (see Caldwell Systems, Inc. discussion in Appendix B).However, stack testing and/or ambient air monitoring at many other incinerators have notindicated releases at levels of health concern. In the last 15-20 years, hazardous wasteincineration practices and facilities in the USA have in general improved dramatically, so it isunlikely that history will find that the newer facilities are so dangerous that they should be banned.

14

A R 3 l 9 l l * 8


Appendix B contains a summary of the health studies related to incineration that ATSDR hasfunded.

The Preliminary Report Regarding Shakedown and Trial Burn Activities at the Drake ChemicalSuperfund Site in Lock Haven, Pennsylvania, dated October 18, 1996, by the GovernmentAccountability Project raised the following issues. Each issue is addressed here, no changes weremade in the text of the health consultation.

Issue 1: When the mini-risk bum sampling for polychlorinated dibenzo-p-dioxin andpolychlorinated dibenzofuran (PCDD/F) and manganese was conducted soil had only been burnedfor four hours before the sampling began. The authors were concerned that this was insufficienttime to establish a PCDD/F equilibrium due to the hysteresis effect that has been reported byseveral researchers. Therefore, that data might be biased low.

ATSDR Response: We did not use the mini-risk burn data when evaluating the public's exposureto PCDD/Fs and manganese in our Health Consultation #3. We used ambient air samplescollected at community monitors for two years and stack samples taken during the full risk burnconducted in January 1997. The full risk burn was conducted after several months of burningcontaminated soil, so the hystereses effect should not be an issue with that data.

Issue 2: On October 9, 1996, the waste feed shut-off did not occur automatically within one hourof the pH falling less than 6.0 like it should have. The issue was raised that the increase in HC1could have caused an increase in PCDD/F formation.

ATSDR Response: It is our understanding that strict measures were put into place because of thisevent to assure that operators assure compliance with the Pennsylvania specified operatingconditions and do not override automatic waste feed shut-offs (AWFSOs). To our knowledge nosuch events have occurred since then. Health Consultation #2 endorsed the AWFSOs establishedby the state of Pennsylvania, which includes the one-hour delay on low pH in the scrubber,because there is an immediate AWFSO if the inlet temperature to the scrubber exceeds 250 °F.Since this AWFSO will prevent the scrubber temperature from getting into the dioxin and furanformation temperature range, an increase in hydrogen chloride (HCl) caused by a pH less than sixshould not increase dioxin or furan emissions. A number of researchers have found that PCDD/Fformation does not increase with increased HCl concentrations; however it does correlate withchlorine (Clj) concentrations and the presence of copper and temperatures between 480 °F and660 °F l.

'Olie, K., Addink, R,, and Schoonenboom, M. "Metals as Catalysts during the Formationand Decomposition of Chlorinated Dioxin andFurans in Incineration Processes". Journal of theAir & Waste Management Association 46:101-105 (February 1998).

15

AR3I9U9


Issue 3: The issue was raised as to whether PCDD/F would increase during an event such as thethermal relief valve opening on October 11, 1996, if contaminated soil was being processed at thetime. The point was made that the gases would be exiting at elevated temperatures and cooled inthe environment through the dioxin formation temperature range.

ATSDR Response: Fundamental laws of chemistry state that the precursor chemicals (products ofincomplete combustion) must come in contact with each other for a sufficient period of time toreact and form other chemicals such as dioxins or furans. This has been shown to occur in postcombustion equipment where gases remain for several seconds in the 480T to 660°F dioxin andfuran formation temperature range. However, when the thermal relief valve opens and the hotgases are vented to the atmosphere, the gases are not confined and the products of incompletecombustion/reactants are not likely to be in close proximity to each other. Furthermore, they willquickly cool to ambient temperature so it is highly unlikely that the necessary reactants will be incontact with each other for several seconds in the ambient air in the necessary temperature range.Therefore, it is highly unlikely that an increase in dioxin and furari formation would occur during athermal relief valve opening.

Issue 4: The authors pointed out mechanical problems that have occurred.

ATSDR Response: All mechanical devices (even brand new cars) have problems. The fact thatproblems have occurred does not affect any of the conclusions or recommendations in HealthConsultation #3. The most important issue related to mechanical problems is did that event causepublic exposure to chemicals at levels of health concern? Health Consultation #3 evaluates thedata from the off-site community ambient air monitors and discusses the potential health impactsbased on that data.

Issue 5: Several regulatory issues were raised to EPA in this document. -

ATSDR Response: ATSDR is not a regulatory agency and will only address health issues relatedto the Drake site.

The Government Accountability Project Preliminary Report Number 2, Drake ChemicalSuperfund Site, Lock Haven, Pennsylvania dated January. 28, 1997, raised several additionalissues. Each issue is addressed here, no changes were made in the text of the health consultation.

Issue #1: The laboratory has analytical problems with the test method being used to measureambient concentrations of beta-naphthyl amine.

Santoleri, J. J. "Dioxin Emissions - Effects of Chlorine, Time, Temperature Relationship at300 °C". Proceedings of International Incineration Conference (1995).

16

AR3I9I50


ATSDR Response: ATSDR Health Consultation #1 identified these same problems andrecommended that EPA find a better method. Health Consultation #3 also addresses this issue.

Issue #2: Was burned ash mixed with the contaminated soil that was burned during the trial burnand risk burns in January 1997?

ATSDR Response: ATSDR inquired about the alleged occurrence of mixing burned ash with on-site contaminated soil. It is ATSDR's understanding that during the optimization period afterOctober 3, 1996, that the previously burned ash was placed on top of the contaminated soilstockpile due to space constraints on-site. The previously treated ash is visibly different in colorand texture from the contaminated soil, so it is possible to see if previously treated ash is broughtinto the feed building. The ash resulting from this testing is of a discernable color and texture thatcan be differentiated from the contaminated on-site soil. Ash was said to be removed from thestockpile, prior to the use of the contaminated soil for the trial bum tests. This occurredapproximately on January 2, 1997. Mr. Steve Jones, the TAG representative, has investigated thisoccurrence. The trial burns did not occur until January 20, 1997. On that date, two U.S. ArmyCorp of Engineer employees inspected the feed building to assure that no previously treated ashwas present. It is also ATSDR's understanding that Mr. Valentine, who originally brought up thisissue was no longer employed at the Drake site past November 30, 1996, so it is unclear how hewould have knowledge of what was burned during the January 1997 trial burns. ATSDR is unableto prove or disprove soil mixing, but it appears that reasonable precautions were in place toassure that only contaminated on-site soils were fed into the incinerator during the trial bums.

Issue #3: The bottom ash samples for both trial burn runs on 25 January 1997, should beconsidered invalid because of alleged chain-of-custody violations.

ATSDR Response: Since the public is not exposed to the bottom ash, chain-of-custody of thebottom ash samples is not a public health concern. This issue will not be addressed by ATSDR.

Issue #4: Beer drinking on-site is alleged to have occurred on 27 January 1997.

ATSDR Response: ATSDR agrees that beer drinking should not be allowed on the Drake site orany other job site.

Issues #5 and #6 are comments regarding EPA procedural issues; not public health issues. Theywill not be addressed here.

ATSDR has received and reviewed EPA Ombudsman Bob Martin's draft report and consideredhis findings. Some of these issues have already been addressed in ATSDR's Health Consultations#1 and #2, including review of the trial burn risk assessment, detection problems with BNA,weather data availability, and site operations. Other issues have been discussed with EPA and willbe addressed in the third health consultation, after the data have been reviewed.

17

A R 3 I 9 I 5 I


Twenty-two newspaper articles were reviewed to determine if they raised any additional issuesthat are relevant to Health Consultation #2. None were specifically related to this consultation,however quite a few articles raised issues related to the potential for public health effects due toincinerator emissions. This issue is addressed in Health Consultation #3. Several articles alsoraised the issue of soil blowing off-site. This issue is also addressed in Health Consultation #3. Acouple of articles discussed an odor problem that existed in 1995. Apparently that problem hasabated since there are no recent articles regarding this issue and we have not noticed a strongodor problem during any of our site visits over the last couple of years. One article referenced anATSDR health study that found increased neurological and respiratory symptoms among residentswho lived near or downwind of an unnamed hazardous waste incinerator. The author suggestedthat this is what the residents of Lock Haven could expect to happen to them. That health studywas a symptom and disease prevalence study conducted in the community around the CaldwellSystems incinerator in North Carolina. That incinerator had been hi operation for a number ofyears without any air pollution control equipment. It was a very poorly run facility, that burnedsome very toxic Navy wastes under conditions that were indicative of incomplete combustion.The results of that study are not relevant to a state-of-the-art incinerator like the Drake Chemicalincinerator treating soil contaminated with parts per million concentrations of chemicals ratherthan pure products. Other issues raised in these articles which are not within ATSDR'sjurisdiction to address are EPA regulatory and legal issues related to the AIR lawsuit, whether theEPA record of decision should be reopened, and contracting costs.

EPA's RESPONSE TOATSDR COMMENTS, CONCLUSIONS, AND RECOMMENDATIONS

COMMENTS:

Comment I: Modeling for potential acute effects should reflect the maximum hourly andmaximum daily ambient concentrations.

EPA Response: The air quality modeling for acute effects will utilize the maximum hourlyestimated emissions for the normal operations of the incinerator. The modeling will utilize themaximum estimated emissions to determine the annual average ambient concentrations for eachpollutant expected to be emitted from the incinerator. As the incinerator is not expected tooperate in a fashion that would generate the maximum emissions, the maximum values provide themost conservative approach for the modeling and would overestimate the predicted ambientconcentrations and resultant deposition rates.

Comment 2: A comparison of the Williamsport -wind data and the actual site-specific -wind datafor the months that the data are available -willprovide additional scientific support to the use ofthe Williamsport meteorological data. A time histogram or auto-correlation may be useful forcomparing the two data sets.

18 .

AR319152


EPA Response: A comparison of the Williamsport, PA and onsite meteorological data ispresented in Section 4.1 of Volume ITJ of the Drake Chemical Site Incinerator Trial Burn RiskAssessment. A-series of monthly wind roses for each data set has been developed to allow for aside-by-side comparison of the wind patterns between Lock Haven and Williamsport, PA.

ATSDR Response: The wind roses do not adequately address our comments; therefore, nochanges were made.

' Comment 3: The use of 5 years of meteorological data -will help assure that likely worst casepublic exposure concentrations are part of the evaluation.

EPA Response: One (1) year of a combined onsite and Williamsport, PA meteorological data isused in the ah" quality modeling analysis for the full operation risk assessment. This approachmaximizes the influence of local conditions on the meteorological modeling. The use of thecombined database maximizes the available and quality assured onsite data for the air qualitymodeling. However, four and one-half (4.5) years of Williamsport data were also evaluated toensure that temporal changes in the meteorological data were considered in the risk assessment.This combined data set, including the six months of local Lock Haven data with the four and on-half years of Williamsport data, will be evaluated hi a sensitivity analysis in Section 8 of thecurrent risk assessment for full operations.

Comment 4: Are there other sources of local meteorological data, such as the Piper MemorialAirport, the state college, other industrial air sources, river and stream flow stations, etc. ? Adefinitive statement as to the unavailability or unsuitability of any such data (and-what attemptswere made to locate other local data sources) -would strengthen the full operation riskassessment.

EPA Response: WESTON has investigated other possible sources of meteorological data in theLock Haven, PA data and there are no known sources of data meeting the siting or datarequirements for air quality modeling. A statement on the unavailability of local meteorologicaldata will be added to the full operation risk assessment.

Comment 5: The actual number of receptors used in the modeling for the trial burn is unclear. Ifmore than one model is used in the operational risk assessment, ensure that the same receptorsare used in each of the models.

EPA Response: A total of 8,571 receptors were utilized hi the air quality modeling analysis. Afigure displaying the location of these receptors is Figure 2.5-1 of Volume ITJ of the June 1996risk assessment. The same receptor grid was used with the ISC3 and CALPUFF models.

ATSDR Response: In the Emissions Modeling, this comment (the fifth bullet) was deleted. EPAhas clarified that the same receptor grid was used in all the models.

19

AR319I53


Comment 6: The rationale for the use of CALPUFF, ISC3, INPUFF, and VALLEY is not clearlyexplained. If all four models are used in the modeling protocol for full operation, it would behelpful to include a clearer discussion of the rationale for the use of each model

EPA Response: The rationale for use of each of the air quality models is described in Section 2.1,General Methodology, of Volume HI hi the June 1996 Trial Burn Risk Assessment. The airquality modeling analysis consisted of a primary and supplemental analysis for both the normaltrial burn emissions and upset conditions. The primary analysis for normal trial bum emissionsused the ISC3 air quality dispersion model, and the primary analysis for the upset condition usedthe INPUFF model. The supplemental analysis for both normal emissions and upset conditionsused the state of the science model CALMET/CALPUFF. The VALLEY model was not used inthe analyses.

The ISC3 model was used to provide the primary estimate of all air concentration and depositionfluxes required for the exposure assessment. The CALMET/CALPUFF model was used toprovide the concentration and deposition estimates using a more-comprehensive representation ofthe local wind field and to provide a complete set of concentration and deposition estimates underinversion/stagnation conditions. CALPUFF was used to provide a complete set of concentrationand deposition values. As with the ISC3 estimates, these values were used to produce a secondcomplete and independent set of risk estimates. The INPUFF model was used to estimate theambient concentrations due to a relief valve release, since it could appropriately represent thevarying emission rates and exit velocity of the relief valve release.

ATSDR Response: In the Emissions Modeling section, this comment (the sixth bullet) wasdeleted. When ATSDR staff read Section 2.1 of Volume IE of the June 1996 Trial Burn RiskAssessment, they agreed that it provided a clear discussion of the rationale for the use of eachmodel.

Comment 7: The sensitivity discussion in Volume I of the trial burn risk assessment, Section 8,compares ISC3 results with CALPUFF results, presumably at the same maximum concentrationlocations derived from ISC3 results. CALPUFF might yield significantly higher concentrationselsewhere. It is not clear whether each model was run independently to allow each one to projectmaximum~impact locations. If one forces the non-steady-state models to predict concentrationsat the same location as the ISC3 maximum location(s), the true maximum concentrations may beunderestimated for both normal and upset conditions.

EPA Response: The ISC3 and CALPUFF models were run separately using the same receptorgrid network. Maximum concentrations locations were determined independently for both models-and compared independent of location.

ATSDR Response: In the Emissions Monitoring section, this comment (the seventh bullet) wasdeleted since EPA has clarified that both models were run independently.

20

AR3I9I51*


COMMENT 8: The uncertainty and sensitivity analyses do not present quantitative assessmentof the air quality modeling results. One approach to quantifying the uncertainty and sensitivityissues would be to exercise one of the models at a few locations, systematically varying thevarious input assumptions. This would yield an ensemble of model predictions that could helphealth scientists assess the values used in the risk assessment.

EPA Response: The Uncertainty Assessment does provide for a comparison of the two modelsused for the air quality analysis. The differences and variability in the two models was consideredinsignificant. The levels of conservatism embedded in the modeled assumptions was considered tooverestimate risks (e.g. use of maximum emission rates and consideration of the duration ofoperation). Since the ISC3 and CALPUFF models were both run independently, and thedifferences in the respective results were determined to be insignificant, other changes in Variablesfor subsequent model comparisons were deemed unnecessary.

ATSDR Response: Although we agree with EPA that comparing the two models was perhaps away to evaluate the relative uncertainty of the models, we still think a more quantitative sensitivityanalysis would be more convincing. When reviewing the comparison of ISC3 andCALMET/CALPUFF modeling results in Volume HI, Section 5.2 of the 21 June 1996, riskassessment it is easy to agree that a 8% to 20% difference is not likely to affect the final risknumbers. However, when you see 40% to 132% differences, it is not easy to say these differencesplus a factor of four difference for downwash and calm winds (Section 5.3) are insignificant whenconsidered hi an aggregate form.

Comment 9: Volume III of the risk assessment assumes that the particulate released during aTRV event would be primarily metals (not organics) and that they would be relatively large andthus would deposit on site. This assumption is contrary to the information in Table 2.4-2 ofVolume I of the risk assessment. That table says all the metals except barium, beryllium, andchromium will be 100% volatilized during normal and upset conditions. Barium will be 50%volatilized, and beryllium and chromium will be 5% volatilized. It seems reasonable to assumethat most of the metals will be in the vapor state when exiting the TRV and will be dispersed inthe plume rather than being large particles that fall on site.

EPA Response: The air quality modeling for the TRV release utilized the emissions developed forthe process upset conditions. These emission calculations are in Volume TJ, Section 3B.3.3 whichpresents the metal emission considering volatilization partition factors without emission controls.The emission used hi the INPUFF modeling were simulated as vapor emissions since noparticulate deposition was accounted for.

ATSDR Response: We have added references to our original comment to clarify which sectionswe think are in conflict. Since Volume HI of the risk assessment still appears to be hi conflict withother sections of the risk assessment, the comment has not been deleted. However, we doacknowledge that it appears from Volumes I and JJ that EPA did use more reasonable metal

21

AR319I55


volatilization factors; contrary to the section in Volume in which discusses the uncertaintiesassociated with the INPUFF model that they used for TRV release modeling. If EPA did modifythe INPUFF model to assume the metals were vapor, then why doesn't Volume HI, Section 5.3,state that instead of saying the INPUFF model assumed all the particulate dropped on site?

Comment 10: EPA assumed that during TRV events, the particulate emissions would be JO,000times those occurring during normal operations. If this is equivalent to zero removal efficiency,the document should indicate that you are assuming no removal of particulate.

EPA Response: Section 3B.3 of Volume H describes the Emissions Survey process upsetconditions. Metals (non-volatilized fraction) and particulate emissions were assumed to beuncontrolled (e.g., no removal of particulate).

ATSDR Response: This comment was deleted (the tenth bullet). EPA is correct, their assumptionis clearly explained in Volume H, Section 3B.3.

Comment 11: ATSDR should review the draft proposal on how the full operation risk assessmentwill project TRV emissions. The proposal should provide references to support any assumptions.

EPA Response: The plan for the consideration TRV emissions in the full operation riskassessment is presented hi the Work Plan for the full operation risk assessment, and is based uponthe same approach used in the June 1996 analysis.

ATSDR Response: ATSDR was not provided with the draft proposal on how the TRV emissionswould be projected or the "Work Plan for the full operation risk assessment". This comment wasreworded to clarify that we are requesting references for (1) the assumption that after an initialpuff of one minute duration, the emissions will decrease exponentially to zero within 25 minutesand (2) the assumption that during the first minute the DRE will decrease from 99.99% to 99%.

CONCLUSIONS:

Conclusion I: Automatic waste feed shut-offs (AWFSOs) were set on the key operatingconditions, and the documents specified that the flame was to be maintained whenever it wassafe to do so. The operating conditions and AWFSOs should ensure safe operation of theincinerator and prevent or minimize to the maximum extent possible exposure of the community.

EPA Response: No response required.

Conclusion 2: The continuous monitoring and recording ofthg. key operating conditions andstack emissions will provide continuous documentation of the operation of the incinerator andthe causes of any releases that may occur.

22

AR3I9I56



Conclusion 3: The circumstances specified when the thermal relief valve (TRV) would be openedwere only those that truly constituted emergency situations, when it would be more protective ofpublic health to vent the hot gases through the TRV than to allow catastrophic equipment failurethat would release the hot gases at ground level.


Conclusion 4: Estimates of TRV emissions for full operation and modeling of those valueswould help investigators predict the public health impacts of upset conditions.

EPA Response: The full operation risk assessment currently being drafted will include anevaluation of upset conditions from TRV emissions. This evaluation will include the experiencegained in the trial burn as well as existing data from other similar sources and will involve both anestimate of emissions and short term ambient dispersion modeling to appropriately estimate anypotential public health impacts.

Conclusion 5: New modeling would be useful for evaluating the stack emissions measuredduring the trial burn. Modeling to date is not adequate to project the ground levelconcentrations in the communities around the Drake incinerator during full operation.

EPA Response: New modeling of full operation emissions is currently being conducted as part ofthe full burn risk assessment.

RECOMMENDATIONS:

Recommendation 1: Conduct new modeling using 5 years of meteorological data and stackemissions rates measured during the trial burn and risk burn. Agency for Toxic Substances andDisease Registry (ATSDR) staff should review the draft modeling protocol before the modeling isconducted so that our input can be more pertinent and timely.

EPA Response: Modeling for the full operation risk assessment is underway pursuant to the workplan for this project. As discussed above, one (1) year of meteorological data (six-months of LockHaven and six months of Williamsport) are used as the basis of the modeling for the body of therisk assessment. This approach maximizes the influence of local conditions in the meteorologicalmodeling. A separate sensitivity analysis will be conducted to evaluate six months of on-site datacombined with four and one-half years of Williamsport data. The resulting modeled ambientconcentrations and depositions will be compared.

Measured stack emissions are being used as the basis for the calculation of emission rates duringthe full operation.

23

A R 3 I 9 I 5 7


ATSDR Response: No changes have been made in this recommendation. The draft modelingprotocol was not provided to ATSDR staff.

Recommendation 2: Reevaluate the assumptions used in projecting the emissions that will bereleased when operating conditions cause the thermal relief valve (TRV) to open. ATSDR shouldreview and provide comments on the draft proposal on how TRV emissions will be projected inthe full operation risk assessment.

EPA Response: The emissions from the potential TRV openings are being reevaluated as part ofthe full burn risk assessment. This revaluation includes the experience gained in the trial burnalong with information from other similar facilities. The approach and results will be included inthe risk assessment report.

ATSDR Response: ATSDR was not provided with the draft proposal on how the TRV emissionswould be projected. This comment was reworded to clarify that we are recommending that thefollowing assumptions be reevaluated and references provided: (1) that after an initial puff of oneminute duration, the emissions will decrease exponentially to zero within 25 minutes and, (2) thatduring the first minute the DRE will decrease from 99.99% to 99%.

24

AR3I9I58


APPENDIX B

25

AR3I9I59


ATSDR Supported Health Studies Related To Incinerators

CALDWELL SYSTEMS, INC., N.C.

Background _ _ •Caldwell Systems, Inc. (CSI) operated a hazardous waste incinerator from 1977 to 1988, inCaldwell County, NC. The hazardous substances burned included: varnish, paint, glue, lacquer,toluene, xylene, and other solvents used by the regional furniture and pipeline industries, as wellas waste torpedo fuel from the U.S. Navy. This incinerator operated without air pollution control

, equipment until 1987. Emissions from the site included smoke from the incinerator and fugitiveemissions from the handling of hazardous wastes.

Original StudyStudy of Symptom and Disease PrevalenceCaldwell Systems, Inc. Hazardous Waste IncineratorCaldwell County, North Carolina, September 1993 •

During July 1991, ATSDR conducted a cross-sectional symptom and disease prevalence study,collecting questionnaire data from 713 residents living within 1.5 miles of the incinerator (targetarea) and 588 residents of a comparison area. Subsequent analysis of this data revealed thatresidents of the target area were more likely than residents of the comparison area (after adjustingfor potential confounders) to report recurrent wheezing or coughing, neurologic symptoms(dizziness and poor coordination), neurologic diseases, and irritative symptoms.

Follow-up StudyHealth Outcome Follow-up Study of ResidentsLiving Near the Caldwell Systems, Inc. SiteCaldwell County, North Carolina(a draft final report)

In a follow-up study conducted during August 1993, ATSDR administered questionnaire's,pulmonary function tests (PFT), neurobehavioral tests, and tests of the immune system toparticipants selected from the original study (above). The follow-up study included 164 targetarea participants and 96 comparison area residents. One finding was that pulmonary function testresults were worse among target area participants who reported respiratory symptoms at follow-up than among those who did not. A draft final report is available for review and is expected tobe published early in FY98.

26

A R 3 I 9 1


VERTAC/HERCULES SITE, AK

Background

In 1988, the Arkansas Department of Pollution Control and Ecology (ADPC&E) contracted forthe incineration of the drummed waste, using a $10.7 million combined trust fund and letter ofcredit obtained from Vertac during bankruptcy litigation. A contract for incineration of thedrummed waste was signed in 1989 between ADPC&E and Vertac Site Contractors (VSC). VSCis a joint venture of MRK Incineration and Morrison-Knudsen Environmental Services.

In January 1992, the ADPC&E approved the VSC trial burn, and production scale incinerationbegan. Because the drums were difficult to handle, incineration took longer than expected. Therewere extensive legal delays due to lawsuits. In May 1993, the trust fund money had been spentwith about one half of the waste destroyed.

In 1990, the Arkansas Department of Health and the Agency for Toxic Substances and DiseaseRegistry began planning the first of a series of studies addressing problems due to the presence ofthe site. The Historical Exposure Assessment Study and the Inhalation Exposure AssessmentStudy were initiated hi 1991. The Health Outcome and the Reproductive Health Monitoringstudies were initiated hi 1993.

Reproductive Health Monitoring StudyPurposes: 1) To determine if there is a temporal association of birth prevalence rates to theJacksonville Superfund sites, and 2) To determine if the birth prevalence rates differ betweenJacksonville and the remainder of Pulaski County.

Results: Examination of adverse reproductive outcomes in Pulaski County over the period of1980-1990, was performed to assess possible relationships to the past chemical manufacturingactivities at the Vertac Superfund site in Jacksonville, Arkansas. Analyses included statisticalcomparisons of occurrences of fetal loss, birth defects, developmental disabilities and low birthweight between Jacksonville and the remainder of the county. Other studies included time trendsand spatial analyses. The findings were:

- The increased fetal loss rates observed hi the Jacksonville area in the early 1980s was notassociated with spatial proximity to the Vertac Site.

- No indication of an excess in birth defects related to the Vertac site was observed.

Weak associations were found for a few developmental problems, (such as seizures andneonatal seizures) and the Vertac site.

27

A R 3 I 9 I 6 I


- Several clusters of low birth weight were noted in Pulaski County, including one to thesoutheast of the Vertac site.

No indication of an association of fetal loss, birth defects or developmental disabilities withthe passage of time was demonstrated.

The localized excess in low birth weight may account for the weak clusters of developmentaldisabilities. There is no direct evidence relating the low birth weight cluster to the southeast ofthe Vertac site to site activities. Further spatial studies of low birth weight and developmentaldisabilities are recommended, including evaluating possible relationships to body burdens andenvironmental data.

On going studies -

Historical Exposure AssessmentPurposes: 1) To determine if exposure to persistent chemicals at the site, including Dioxin andDioxin-like compounds for more than 15 years, results in higher body burdens by conductingbiomonitoring activities which include collection of medical history and blood samples; 2) Todetermine if exposure to non-persistent chemicals at the site including 2,4-D, 2,4-dichlorophenol,2,4,5-T and 2,4,5-trichlorophenol for more than one year (as of 1991) results in higher bodyburdens by conducting biomonitoring activities which include collection of medical history andblood samples.

Incinerator Exposure AssessmentPurpose: To determine if remediation activities associated with the site clean-up, particularlyhandling and incineration of drummed wastes, results in increases in body burdens of site relatedchemicals by conducting pre- and post-incineration biomonitoring activities for estimating bloodlipid concentrations and urine concentrations of site contaminants in nearby residents.

Health Outcome StudyPurpose: To test the hypothesis that "Group living for extended periods hi a defined area near oradjacent to the site have a higher prevalence of acute or chronic illness, plausibly associated withexposure to persistent pollutants related to the site such as TCDDX than similar groups who live hian area far away from the site."

Effects of Dioxin-like Compounds on Glucose Regulation and Insulin Sensitivity (DiabetesReport)Purposes: 1) To determine if there is a positive association between excess blood lipid levels ofTCDD (and TEQ-TCDD) and an increased likelihood that elevated levels of insulin are necessaryto control blood glucose concentrations within desirable physiological limits; and 2) To test if theassociation between excessive exposure to TCDD (TEQ-TCDD) and increased insulinconcentrations follows a dose response.

28

AR3I9162


Do Waste Incinerators Induce Adverse Respiratory Effects? An Air Quality andEpidemiological Study of Six Communities in North Carolina.

The University of North Carolina conducted a three-year epidemiological study of the prevalenceand incidence of respiratory effects among residents of communities surrounding three types ofwaste incinerators (a biomedical incinerator, a municipal waste incinerator, and a liquid hazardouswaste-burning industrial furnace), and three matched comparison communities.

Purposes: 1) To compare the prevalence of chronic respiratory symptoms, respiratoryhypersensitivity, diminished lung function, upper respiratory tract inflammatory reactions, andupper and lower respiratory tract diseases hi exposed and non-exposed communities, adjusting forthe distribution of known risk factors for these conditions; 2) To select subcohorts of normal andof hypersensitive adults in these exposed and control communities and to obtain dailymeasurements of lung function and respiratory symptoms in these persons over a one monthperiod, annually, for three years, with simultaneous daily measurements of air quality in eachcommunity; 3) To identify whether subgroups of the population are at higher risk of lung andrespiratory disease from exposure to fugitive or stack emissions from incinerators.

Results from first year: differences hi concentrations of particulate matter were detected amongany of the three pairs of study communities. Average fine particulate (PM2 5) concentrationsmeasured for 35 days varied across study communities from 16 to 32 ug/m3. Within the samecommunity, daily concentrations of the fine particulate varied by as much as eightfold, from 10 to80 ug/m3, and were nearly identical within each pair of communities. Direct measurements of airquality and estimates based on a chemical mass balance receptor model showed that incineratoremissions did not have a major or even a modest impact on routinely monitored air pollutants. Aonetime baseline descriptive survey (n=6963) did not reveal consistent community differences inthe prevalence of chronic or acute respiratory symptoms between incinerator and comparisoncommunities, nor were differences seen in baseline lung function tests or in the average peakexpiratory flow rate measured over a period of 35 days. Based on this analysis of the first year ofour study, the investigators concluded that we have no evidence to reject the null hypothesis of noacute or chronic respiratory effects associated with residence in any of the three incineratorcommunities.

Results from subsequent years and the final report of this project are yet to be finalized.

29

AR3I9I63

Petitioned Health Consultation No. 3

Public Comment Release

Air Emissions EvaluationThermal Destruction Facility

DRAKE CHEMICAL



MARCH 25, 1998

Comment Period End Date: May 3, 1998

U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICESAgency for Toxic Substances and Disease RegistryDivision of Health Assessment and Consultation

Atlanta, Georgia 30333

AR3I9I61*

Health Consultation: A Note of Explanation

An ATSDR health consultation is a verbal or written response from ATSDR to a specificrequest for information about health risks related to a specific site, a chemical release, or thepresence of hazardous material. In order to prevent or mitigate exposures, a consultation maylead to specific actions, such as restricting use of or replacing water supplies; intensifyingenvironmental sampling; restricting site access; or removing the contaminated material. Inaddition, consultations may recommend additional public health actions, such as conductinghealth surveillance activities to evaluate exposure or trends in adverse health outcomes;conducting biological indicators of exposure studies to assess exposure; and providing healtheducation for health care providers and community members.

The Public Comment Period is an opportunity for the general public to comment on Agencyfindings or proposed activities for this written consultation. The purposes of the commentperiod are to 1) provide the public, particularly the community associated with a site, theopportunity to comment on the public health findings, 2) evaluate whether the communityhealth concerns have been adequately addressed, and 3) provide ATSDR with additionalinformation.

The conclusions and recommendations presented in this health consultation are the result ofsite specific analyses and are not to be cited or quoted for other evaluations or healthconsultations.

You May Contact ATSDR TOLL FREE at1-800-447-1544

orVisit our Home Page at: http://atsdrl.atsdr.cdc.gov:8080/

A R 3 I 9 I 6 5

PUBLIC COMMENT RELEASE

PETITIONED HEALTH CONSULTATION NO. 3

Air Emissions EvaluationThermal Destruction Facility

DRAKE CHEMICAL



Prepared by:

Petition Response SectionExposure Investigation and Consultation BranchDivision of Health Assessment and ConsultationAgency for Toxic Substances and Disease Registry

AR3I9I66

Drake Chemical - Public Comment

BACKGROUND AND STATEMENT OF ISSUES

This health consultation is the third in a series the Agency for Toxic Substances and DiseaseRegistry (ATSDR) is conducting to address public health issues relating to the Drake ChemicalSuperfund site in Lock Haven, Pennsylvania. The consultation evaluates three sets of air data;1) on-site perimeter and off-site ambient air data collected before the trial burn (September1995 - February 1996, Appendix D), 2) stack emissions, on-site perimeter, and off-siteambient air data sampled during the trial burn period (September 1996 - February 1997), and3) on-site perimeter data (August - December 1997)(1'5) and off-site ambient air data (June -September 1997)(6-9) collected after the trial burn period. In Health Consultation #1(10) andHealth Consultation #2(11), ATSDR has previously evaluated information contained in the trialburn risk assessment including, ambient air sampling/monitoring plans and activities, andincinerator operational plans and trial burn protocols. ATSDR refers readers to thosedocuments for general background information and detailed discussions of those issues.

The purpose of this health consultation is to determine whether the levels of stack emissionsand fugitive emissions produced at the Drake site during the trial burn period were sufficient tocause any adverse health effects in the surrounding community. ATSDR's evaluation will alsoindicate whether future adverse health effects would be likely if the incinerator operationproduced similar emissions during a projected 13-16 months of full burn operation.

DISCUSSION

Evaluation of Ambient Air Data Quality

ATSDR's contractors (Eastern Research Group, Inc., Morrisville, NC)(12) evaluated qualitycontrol data for the ambient air analytical data that were collected by the EnvironmentalProtection Agency (EPA) and Army Corps of Engineers' contractors during the trial burnperiod. That evaluation covered both the site perimeter and community ambient air samplingdata. The following discussions are based upon ATSDR's review and evaluation of thecontractor's findings.

All necessary analytical and quality control information was properly documented. Thelaboratory analytical quality control data reviewed reveals appropriate and acceptable qualitycontrol measures. Laboratory equipment performed appropriately with regard to calibrationand tuning criteria.(12)

AR3I9I67


P-naphthylamine and Fenac were not detected in air samples at the current detection limit of0.001 M-g/m3, due to poor recoveries using the EPA TO-13 method of detection. In addition,low recoveries, below acceptance criteria, were noted for several surrogate compounds(particularly surrogates for P-naphthylamine and Fenac) despite evidence that the analyticalmethod was appropriately executed in the laboratory/13 ATSDR noted this issue in HealthConsultation #1(10) and recommended that efforts be made to improve the recoveries or select amore appropriate method. The National Institute for Occupational Safety and Health (NIOSH)method 5518 was used to monitor for p-naphthylamine from November 3 to November 25,1997.(13) The detection limit set for p-naphthylamine using method 5518 was 2.0 |ig/m3compared to a detection limit of 0.001 Jig/m3 for the TO-13 method. ATSDR does notconsider the NIOSH 5518 method to be acceptable, since it is 2,000 times less sensitive indetecting p-naphthylamine than the TO-13 method which is only 20 times higher than the risk-based concentration for cancer (RBCc) of 0.000048 U.g/m3. Therefore, ATSDR suggested thata different sampling and/or analytical method for P-naphthylamine should be investigated byEPA. The EPA tested a modified Occupational Safety and Health Administration (OSHA) 93method during laboratory and field trials, collecting air samples from December 1997 toFebruary 1998.(14-18)

EPA sampling results demonstrated that while P-naphthylamine and its surrogate 4-aminobiphenyl were not detected in air samples on or off-site, control samples spiked withknown quantities of either p-naphthylamine or its surrogate, were detected. ATSDR considersthe OSHA 93 method to provide improved sampling and analytical methodology required todetect potential contaminants present with an adequate method detection limit of 0.001 }ig/m3and a practical detection of 0.02 |ig/m3. ATSDR considers that exposure to 0.02 M-g/m3 p-naphthylamine for 16 months would not represent a carcinogenic hazard to residents livingwithin a mile of the Drake site (Appendix C). However, a detection limit of 0.001 u.g/m3 forP-naphthylamine would offer a larger margin of safety. ATSDR considers a detection limit of0.02 jig/m3 set for Fenac, a noncarcinogen, to be protective of public health.

Ambient Air TVfnnitnring nf "Emissions from the Drakp ("!hpmiral Site

Ambient air monitoring continues to be conducted and evaluated for stack emissions from theincinerator at four stations (Appendix A) within the communities surrounding the Drake site.These stations are located approximately one mile from the site and designated Hospital (west),School (northwest), Airport (northeast), and Township building (southeast). In addition, fourair monitoring stations are located on the Drake site perimeter (Appendix B) at approximately90"intervals and are designated stations 1, 2, 3, and 4. These sampling locations primarilyevaluate possible fugitive emissions from the Drake site. Samples from the community werecollected for a 24-hour period every sixth day and analyzed for volatile organic compounds(VOCs), semi-volatile organic compounds (SVOCs), metals, dioxins and furans, and total

AR3I9I68


suspended particulates (TSP). Site perimeter samples were collected for a 24-hour period onceevery month and analyzed for semi-volatile organic compounds, TSP and metals. In addition,real-time monitoring for VOCs was conducted at the perimeter stations for, 5 minutes out ofevery hour on a continuous basis during site operations .

Ambient air, sampled from all eight locations, was collected in the breathing zone to identifychemical concentrations to which people in those or adjacent areas may have been exposed.Sampling and monitoring at the perimeter stations focused on fugitive emissions, whichresulted primarily from direct volatilization and wind blown dust from the site. It is intendedto detect concentrations of contaminants that may exist on-site and may be migrating off-site.

Community sampling locations were chosen based on models developed to determine areas ofmaximum deposition of incineration products and public accessibility. (19) Monitors areintended to detect the maximum concentration of contaminants that reach ground level afterexiting the incinerator stack. While all of the sampling stations were targeting the incineratorand site-related activities, the impact from other air pollution sources was also detected. Todetermine the level of air pollution from existing sources, sampling was conducted prior to theinitial operation of the incinerator and after completion of the trial bum. In this consultation,ATSDR has not only evaluated the impact of ambient air contamination during the trial burnprocess, but has also compared that data with air sampling results for time periods prior to theincinerator's initial operation and after the trial burn period to determine the incinerator'soverall impact on local air quality.

ATSDR evaluated site perimeter ambient air data, off-site ambient air data, and stackemissions data. Chemicals of concern were selected for further toxicological evaluation. A"chemical of concern" (COC) is defined as any chemical that is detected in air atconcentrations that exceed one or more of ATSDR' s comparison values (CVs). Thesecomparison values are health guidelines, which are based on toxicological information aboutthe chemical, assumptions about the rate of exposure, and include ample safety factors to /ensure protection of public health, including sensitive populations. However, the presence ofa contaminant at concentrations that exceed one of these conservative health guidelines doesnot indicate the existence of a public health hazard. ATSDR' s comparison values are used toscreen contaminant concentrations at a site and to select "chemicals of concern" for furtherevaluation by agency health scientists.

Off-Site Ambient Air Data F.

Analysis of the off-site ambient air data demonstrated that none of the 22 COCs represents apublic health hazard. Only the Cancer Risk Evaluation Guides (CREGs) or Risk-BasedConcentrations (RBCc) for cancer were exceeded by the average and/or the maximumconcentrations of these COCs. However, CREGs and RBCc are based on lifelong exposure,

AR3I9I69


while the burn is projected to last for 16 months. In addition, cancer-based CVs assume theabsence of a threshold, which makes them extremely conservative. More importantly, thesesame chemicals (and no others) exceeded the same comparison values before the test burn ofthe incinerator was conducted. Note that a similar situation existed with the dioxin/furantoxicity equivalency quotients (TEQs). Of the 22 contaminants selected for further evaluation,17 were not actually detected but were selected because the applicable detection limitsthemselves exceeded cancer guidelines. There was no detectable difference between pre-burn,trial burn, or post trial burn air data. Average concentrations of benzene and arsenic and themaximum concentration of benzo(a)pyrene were actually higher before the test burn. Only theaverage concentration of dibenzo(a,h)pyrene at station #1 (the airport) was higher during thetest burn than before it, and this was due to a single high value.

Ambient air data collected from the four off-site locations identified several dioxin-likecompounds. The dioxin-like compounds detected included various chlorodibenzo-p-dioxins(CDDs) and chlorodibenzofuran (CDFs). The concentrations of all of the dioxin-likecompounds detected at each of the 4 monitoring stations were expressed in terms of the mosttoxic congener, 2,3,7,8-tetrachlorodibenzo-p-dioxin or TCDD, using the Toxicity EquivalencyFactor (TEF)00. TEFs are conservative estimates of relative toxicity. The resultingcomposite concentrations before, during, and after the test burn, were expressed as ToxicityEquivalency Quotients (TEQs) and were then compared to EPA Region ffl's Risk-BasedConcentration (RBCc) for potential cancer effects (5.4 X 10'8 |ig TCDD/m3, or 0.000054 ngTCDD/m3, or 0.054 pg TCDD/m3). This RBCc is specific for the most toxic congener2,3,7,8-TCDD and applying it to all dioxin-like compounds as total TEQs, introducesadditional conservatism into the assessment.

There was no significant difference between the background values and the trial burn data,suggesting that the Drake incinerator had no discernable impact on background concentrationsof dioxin-like compounds in the air at Lock Haven. In fact, total TEQs actually declined,relative to background levels, during the risk burn (Appendix D) at 3 of the 4 off-site ambientair monitoring stations. This suggests that the impact of the Drake site incinerator may besmaller than the seasonal variation of background CDD/CDF concentrations at Lock Haven(the background data for CDD/CDF was collected between January and August 1996, whilethe risk burn data was collected between September 1996 and February 1997). However, nomatter what the actual seasonal variation in background dioxin levels may be, the availabledata indicate that the contribution of the Drake incinerator is of no toxicological significance.

Site Perimeter Ambient Air Data F.valnatinn

Four contaminants (arsenic, P-naphthylamine, cadmium, and nickel) analyzed for in the siteperimeter ambient air monitoring program were selected for further toxicological evaluation.Benzo(a)pyrene also exceeded its cancer based RBCc, but was not selected as a site-related

AR3I9I70


chemical of concern because this value was exceeded only at stations 1 and 3, and only beforethe risk burn. At the levels detected, none of these four contaminants were determined torepresent a public health hazard. Only CREGs or RBCs for cancer (RBCc) were exceeded bythe average and/or the maximum concentrations. However, CREGs and RBCc's are based onlifelong exposure, while the burn is projected to last for only 16 months. In addition, cancer-based comparison values assume zero-threshold, which makes them extremely conservative.Moreover, these four contaminants exceeded the same CVs before the risk burn, suggestingthat the Drake incinerator was not a significant contributor to these levels, but may reflect soil-disturbing activities at the Drake site and would be expected to be limited in duration.ATSDR has concluded that, due to the low magnitude and duration of projected exposures atLock Haven, none of these four contaminants selected as chemicals of concern represent apotential hazard to public health.

p-naphthylamine was not detected on-site using the EPA TO-13 method. It is a COC onlybecause 1/2 the detection limit (the default concentration used by EPA for non-detectedchemicals) was a higher concentration than the RBCc of 0.000048 M-g/ni3. P-naphthylaminewas not detected on-site using the same detection limit for measuring off-site emissions. Theidentical assertions concerning cancer risks apply equally well on and off-site.

The average concentrations of arsenic, cadmium, and nickel were actually lower at stations 1,2, and 4 during the burn than before it. Concentrations of all three metals and the TSP levelincreased at station 3 during the trial burn, but are not at levels of health concern. Theincrease in metals and TSP is probably due to emissions from the bottom ash and emissionsmay be limited by wetting or covering the bottom ash storage area.

Stack Hmissinns Air Data Evaluation

Contaminants exiting the stack are quickly dispersed into the environment, reducing theconcentrations by various amounts, depending on where a person is located and weatherconditions. Residents are not breathing air contaminants at concentrations exiting the stackand their levels are not directly relevant to the potential for adverse health effects. Therefore,ATSDR prefers to use ambient air data rather than stack data in the assessment of potentialpublic health hazards, since no one will be directly exposed to stack gases. ATSDR evaluatedthe health implications of the stack data as a "worst-case" exposure potential, and used it toidentify contaminants exiting the stack that are not monitored for in ambient air at the siteperimeter or off-site within the community.

To assess stack emissions, two risk burns were conducted during the trial burn operation of theincinerator at the Drake site. Data from Risk Bum No. 1 was collected from January 20-22,1997 and data from Risk Burn No. 2 was collected from February 7-9, 1997. ATSDR hasreviewed and evaluated this stack testing data. Air samples were collected from the incinerator

A R 3 I 9 I 7 I


stack to determine whether concentrations of chemicals present during the trial burn(incineration of site soil spiked with known quantities of specific chemicals) and the risk burn(incineration of site soil) would be at levels of public health concern.

Data indicated that the concentrations of all inorganics, VOCs, pesticides, polychlorinatedbiphenyls (PCBs), and other SVOCs detected in the stack are substantially lower than theirrespective Threshold Limit Values (TLVs), even before these substances left the stack andwere diluted in the ambient air. Several substances detected in the stack gases did exceed -theirrespective CREGs, and chlordane exceeded its chronic Environmental Media Evaluation Guide(EMEG). However, the projected duration of exposure (approximately 16 months) does notjustify using comparison values based on lifetime exposures for evaluating either ground levelambient air or stack emissions exposures.

Benzene (average concentration 335 |ig/m3) was the only detected substance that exceeded anon-chronic comparison value at the stack. Benzene exceeded its acute EMEG (50 ppb or 160M-g/m3), as well as its CREG (0.1 |J.g/m3), but only 10% of its TLV (1 ppm or 3.2 |ig/m3 ,OSHA), even before leaving the stack. In off-site ambient air during the trial burn, benzeneconcentrations (0.2-0.95 ppb) were less than 2% of ATSDR's acute EMEG or 50 ppb and lessthan 0. 1 % of the OSHA TLV (1 ppm or 3.2 (ig/m3). No inhalation values are available forcalcium, magnesium, potassium, or sodium, however, these essential nutrients would not bechemicals of toxicological concern. Therefore, ATSDR concludes that none of the detectedsubstances would be likely to produce adverse health effects at the concentrations detected hiambient air (i.e., hi the breathing zone) or under the unlikely "worst-case" scenario ofbreathing stack gases directly at the stack.

Other Ambient Air Issues

Action

Real time fence line air monitoring is being conducted at 4 stations around the site for totalVOCs. If total VOC concentrations exceed 1 ppm, air samples from each station will beanalyzed using an on-site gas chromatograph for the following three "indicator compounds";toluene, tetrachloroethene, and chlorobenzene. If any of these compounds exceed theirindividual action level (See table), site activities will stop until the source is found and abated.

Previous soil, sediment, and on-site air data were reviewed by ATSDR to determine ifappropriate compounds were selected for site perimeter "indicator compounds" to monitor forfugitive VOC emissions from on-site activities. (23"26) The data reviewed was not sufficient toadequately determine if other indicator compounds would be more appropriate. To ensure thatVOCs other than the indicator compounds are not exceeding levels of health concern, ATSDRrecommends the following:

A R 3 I 9 I 7 2


1) If total VOC concentrations exceed 1 ppm for 10 -15 minutes, a grab air sample should becollected and sent for laboratory analysis of VOCs (i.e., EPA method TO-14 or equivalent).Review of this data should ensure that the current indicator compounds and their respectiveaction levels are appropriate and protective.

i

2) If total VOC concentrations exceed 5 ppm for 10 - 15 minutes, in addition to collecting agrab sample, the source of VOCs should be determined and stopped. This may includecovering exposed soils, stopping work, etc.

3) If any of the indicator compounds reach 50% of their action levels, the source of VOCsshould be determined and stopped. This may include covering exposed soils, stopping work,etc.

4) If any of the indicator compounds reach or exceed their action levels, on-site activitiesshould cease until the source is found and abated. Ambient air samples should be analyzed forVOCs, SVOCs, metals, and particulates, to identify and determine the maximum concentrationof the contaminants present at the perimeter of the site.

Perimeter Air Action Levels and Corresponding Actions

Chemical

Total VOCs

Toluene

Tetrachloroethene

Chlorobenzene

Reduce Site ActivityAction Level

1 ppm* (10-15 minutes)

4.6 ppm

1.3 ppm

3.9 ppm

Cease Site Activity ActionLevel**

5 ppm (10 - 15 minutes)

9.2 ppm

2.5 ppm

7.8 ppm

* Collect grab air sample for VOC laboratory analysis** Collect air samples for VOCs, SVOCs, metals, and particulates for laboratory analysis

Fugitive Emissions (Air Rlnwn Dust)

Residents of Lock Haven have expressed concern over the potential exposure fromcontaminated soil blown off-site during periods of soil excavation on the Drake site andoccasional high winds that may blow soil off-site from uncovered soil stock piles. ATSDRreviewed the on-site and off-site ambient air data for total suspended particulates (TSP).

AR3I9I73


The maximum on-site TSP for monitoring stations 1,2, and 4 was approximately two-threetimes higher before the trial burn than during the trial burn period, but did not exceed EPA'sNational Ambient Air Standard for PM10 of 150 p.g/m3, based on a 24 hour average not toexceed more than 3 times in a year. A maximum TSP of 436.4 |ig/m3 was detected on-siteonly at station 4 and before the trial burn period. Only at station 3 and for one sampling dateduring the trial burn period was the maximum TSP (199.2 p.g/m3) above the comparisonvalue. The TSP exceeded 150 |ig/m3 only during one air sampling period (May 5, 1997) atstation 4, after the trial bum period ended. No difference in the maximum TSP was observedeither before or during the trial burn period at off-site air monitors located at the hospital,school, or hi Castanea. Only the maximum TSP monitored at the airport, increased during thetrial burn period. It is unlikely that any health effects would be expected, from the totalsuspended particulate levels detected at the perimeter or off-site within the community.

CHBLD HEALTH

Children may be at greater risk than adults from certain kinds of exposures to hazardoussubstances emitted from waste sites. ATSDR has evaluated the likelihood that children livingnear the Drake Chemical site would be exposed to contaminants at levels that may affect theirhealth. Ambient air data reviewed to date did not demonstrate an exposure of health concernduring the trial burn period at the Drake site and health effects are not likely to occur duringthe full burn operation.

CONCLUSIONS

1. With the exception of the P-naphthylamine and Fenac, the analytical data for ambient airsamples is of generally high quality and acceptable for the evaluation of potential publichealth impacts. Due to poor surrogate recoveries, P-naphthylamine and Fenac were notdetected on or off-site. Results suggest that concentrations of 0.02 }ig/m3 for bothcompounds would be detected during laboratory analysis of air samples. A detection limitset at 0.02 jig/m3 for Fenac and P-naphthylamine would be protective of public healthduring the full burn period of operation.

2. Off-site ambient air data collected at all four air monitoring stations within the communitydemonstrated that contaminants were not present at levels of health concern.

3. There was no detectable difference in the concentration of 17 out of the 22 chemicalsdetected off-site from air collected either before, during, or after the trial burn period.

AR3I9I71*


4. Average off-site concentrations of benzene and arsenic, and the maximum concentrations ofbenzo(a)pyrene were higher before the test burn than during the test burn. None of thedetected concentrations are at levels of health concern.

5. No significant difference between total dioxin/furan toxicity equivalency quotients(TEQs) was found at 3 of the 4 off-site air monitoring stations.

6. Total chlorodibenzofurans (CDF) toxicity equivalency quotients and total toxicityequivalency quotients were significantly elevated at station 3 (the hospital monitoringstation) before, during, and after the trial burn, compared to the other 3 stations. Theseconcentrations were below levels of health concern.

7. All maximum concentrations of individual dioxin congeners and the total chlorodibenzo-p-. dioxins (CDDs) at every station off-site were below comparison values and were of nopublic health concern.

8. Levels of arsenic, cadmium, and nickel detected by site perimeter air monitors exceededcancer-based comparison values only, and were determined not to be of any public healthconcern under site-specific conditions of exposure.

9. P-naphthylamine was not detected in ambient air, due to poor recoveries using previousanalytical methods of detection. Therefore, ATSDR suggested a differentsampling/analytical method be investigated by EPA. ATSDR considers that the modifiedOSHA 93 method proposed by EPA would be adequate in detecting p-naphthylamine inambient air at Lock Haven and the detection limits established for this method would be atlevels protective of public health.

10. The site sampling data reviewed was not sufficient to adequately determine if the current"indicator compounds" are the most appropriate to ensure that other VOCs are notexceeding levels of health concern.

11. ATSDR reviewed the on-site and off-site ambient air data for total suspended particulates.It is unlikely that any health effects would be expected, from the total suspended particulatelevels detected at the perimeter or off-site within the community.

AR3I9I75


RECOMMENDATIONS

1. Continue to use the modified OSHA 93 analytical method with appropriate field andlaboratory controls for detecting P-naphthylamine, so that low ambient air levels of thiscontaminant can be reliably identified.

2. To ensure that VOCs other than the indicator compounds are not exceeding levels ofhealth concern, ATSDR recommends that the following actions be added to EPA's siteperimeter sampling plan for protection of public health.

a) If total VOC concentrations exceed 1 ppm for 10 -15 minutes, a grab air sampleshould be collected and sent for laboratory analysis of VOCs (i.e., EPA method TO-14or equivalent). Review of these data should ensure that the current indicatorcompounds and their respective action levels are appropriate and protective.

b) If total VOC concentrations exceed 5 ppm for 10 - 15 minutes, in addition tocollecting a grab sample, the source of VOCs should be determined and stopped. Thismay include covering exposed soils, stopping work, etc.

c) If any of the indicator compounds reach 50% of their action levels, the source ofVOCs should be determined and stopped. This may include covering exposed soils,stopping work, etc.

d) If any of the indicator compounds reach or exceed their action levels, on-siteactivities should cease until the source is found and abated. Ambient air samplesshould be taken for VOCs, SVOCs, metals, and particulates should be taken to identifyand determine the maximum concentration of the contaminants present at the perimeterof the site.

3. Continue sampling for volatile, semi-volatile organic compounds, metals, dioxin/furans,and total.suspended particulates off-site during the full burn operation.

4. Cover all stock piles of soil and ash to reduce dispersion of particulates off-site.

10

AR3I9I76


Preparers nf the. Health Consultation

Adele M. Childress, PhD, MSPHEnvironmental Health ScientistPetitions Response BranchDivision of Health Assessment and Consultation

Jeffrey A. Church, MPH, REHSEnvironmental Health SpecialistPetitions Response BranchDivision of Health Assessment and Consultation

Frank Schnell, PhD, DABTToxicologistPetitions Response BranchDivision of Health Assessment and Consultation

Lynn Wilder, MSHyg, CfflEnvironmental Health ScientistExposure Investigation and Consultation BranchDivision of Health Assessment and Consultation

Betty C. Willis, MSEnvironmental Health ScientistProgram Evaluation, Records, and Information Services BranchDivision of Health Assessment and Consultation

11

A R 3 I 9 I 7 7


REFERENCES

1. Specification Section 13600/1005 Perimeter Air Monitoring Report. August 1997.Prepared by OHM Remediation Services Corporation, 180 Myrtle Street, Lock Haven, PA17745

2. Specification Section 13600/1005 Perimeter Air Monitoring Report. September 1997.Prepared by OHM Remediation Services Corporation, 180 Myrtle Street, Lock Haven, PA17745

3. Specification Section 13600/1005 Perimeter Air Monitoring Report. October 1997.Prepared by OHM Remediation Services Corporation, 180 Myrtle Street, Lock Haven, PA17745

4. Specification Section 13600/1005 Perimeter Air Monitoring Report. November 1997.Prepared by OHM Remediation Services Corporation, 180 Myrtle Street, Lock Haven, PA17745 . ,

5. Specification Section 13600/1005 Perimeter Air Monitoring Report. December 1997.Prepared by OHM Remediation Services Corporation, 180 Myrtle Street, Lock Haven, PA17745

6. Off-Site Air Monitoring Program. June 1997. Prepared by OHM Remediation ServicesCorporation, 180 Myrtle Street, Lock Haven, PA 17745

7. OffrSite Air Monitoring Program. July 1997. Prepared by OHM Remediation ServicesCorporation, 180 Myrtle Street, Lock Haven, PA 17745

8. Off-Site Air Monitoring Program. August 1997. Prepared by OHM RemediationServices Corporation, 180 Myrtle Street, Lock Haven, PA 1774527

9. Specification Section 13600/1005 Off-Site Air Monitoring Report. September 1997.Prepared by OHM Remediation Services Corporation, 180 Myrtle Street, Lock Haven, PA17745

10. Agency for Toxic Substances and Disease Registry. Health Consultation #1 for DrakeChemical Superfund site thermal destruction facility, Lock Haven, Pennsylvania. Atlanta:US Department of Health and Human Services, Public Health Service. 1997 July 24.

12

AR3I9I78


11. Agency for Toxic Substances and Disease Registry. Health Consultation #2 for DrakeChemical Superfund site thermal destruction facility, Lock Haven, Pennsylvania. Atlanta:U.S. Department of Health and Human Services, Public Health Service. 1997 July 24.

12. Eastern Research Group. Drake Chemical Superfund site air sampling program qualitycontrol data review. Morrisville, NC. 1997 June 20

13. NOSH Method 5518 Air Monitoring Report for 2-Naphthylamine: November 1997.Prepared by OHM Remediation Services Corporation, 180 Myrtle Street, Lock Haven, PA17745

14. OHM Remediation Services Corporation, 180 Myrtle Street, Lock Haven, PA. 17745.Drake Chemical Superfund Site. OSHA 93 Program Approach and SOPs. Enclosure A.December 1997

15. OHM Remediation Services Corporation, 180 Myrtle Street, Lock Haven, PA. 17745.Drake Chemical Superfund Site. Analytical Results for OSHA 93. Method Detection Limit(MDL) Study. December 1997.

16. OHM Remediation Services Corporation, 180 Myrtle Street, Lock Haven, PA. 17745.Drake Chemical Superfund Site. Analytical Results for OSHA 93. Recovery StudyExperiment. December 1997.

17. OHM Remediation Services Corporation, 180 Myrtle Street, Lock Haven, PA. 17745.Drake Chemical Superfund Site. Analytical Results for OSHA 93. Air Exposure StudySamples. December 1997.

18. Specification Section 13600/1008. OSHA 93 Preliminary Summary Report. January 1998.Prepared by OHM Remediation Services Corporation, 180 Myrtle Street, Lock Haven, PA17745

19. Midwest Research Institute and RUST International, Inc. Drake Chemical SuperfundSite Off-Site Ambient Air Sampling Plan. Gary, NC. April 20, 1995.

20. EPA (1996). Risk-Based Concentration Table. EPA Region JH, Philadelphia, PA.September 25, 1996.

21. USEPA, 1989. Exposure Factors Handbook. Exposure Assessment Group, Office ofHealth and Environmental Assessment, Office of Research and Development, U.S.Environmental Protection Agency. EPA/600/8-89/043. July 1989.

13

A R 3 I 9 I 7 9


22. EPA (1994). Estimating Exposure to Dioxin-Like Compounds, Vol U: Properties,Sources, Occurrence and Background Exposures (External Review Draft). EPA/600/6-88/005Cb. Exposure Assessment Group, Office of Health and Environmental Assessment,United States Environmental Protection Agency, Washington, D.C., June 1994. Page 4-47.

23. Summa Canister Monitoring Report for Drake Chemical Superfund Site. July 1997.Prepared by OHM Remediation Services Corporation, 180 Myrtle Street, Lock Haven,PA 17745.

24. Inspectors Quality Assurance Report (QAR) Daily Log of Construction. July 2 & 3,1997, OHM Remediation Services Corporation.

25. Background Air Quality Report for On-Site Soil Incineration, Drake ChemicalSuperfund Site, Lock Haven, PA. DACW45-93-C-0200. March 1995. Prepared byMRI and RUST International, Inc., for U.S. Army Engineer District, Corps ofEngineers.

26. Test Report, Ambient and Borehole Air Monitoring at Drake Chemical Superfund Site,Lock Haven, PA, January 21, 1991. Prepared by Engineering Science, Inc., for James M.Montgomery Consulting Engineers, Inc.

27. Hazardous Substances Database (HSDB). 1996. Fenac (2,3,6-trichhlorophenylaceticacid). National Library of Medicine, National Toxicology Program, Bethesda, MD.

28. St. John Jr., L.E. and Lisk, D.J. 1970. Metabolism of Fenac herbicide in a lactating cow.Journal of Dairy Sci. 53(2): 161-4.

29. Arena, J.M. 1986. PDTSONTNflr Tmcicnlngy, Symptoms, Treatments, 5th Edition.Charles C.- Thomas, Publisher, Springfield, ILL.

30. Worthing, C.R. and Walker, S.B. 1987. The Pesticide Manual•_ A WnrlH Compendium,8th Edition. Thorton Heath, Ed. The British Crop Protection Council, UK.

31. ACGJJEL 1995. 1995-1996 Threshold Limit Values (TLVs) for Chemical Substances and.Physical Agents and Biological Exposure Indices (BEIs). The American Conference ofGovernmental Industrial Hygienists, Cincinnati, OH.

32. Federal Register. 51:33997-8, September 24, 1986.

14

AR3I9I80


33. Re-evaluation of theEDOl Study. 1981. Fundal. Appl. Toxicol. 1:27-128. Society ofToxicology.

15

A R 3 I 9 I 8 I


APPENDIX A

Location of Off-Site Ambient Air Monitors

Al

AR3I9I82

^ _————~—;(- ff Interstate Highway

StateflJical Highway

Local Street

Railroad

/ N River/Creek

// Sita BoundaryD One Mile Buffer

Community Air Sampling Stations

Perimeter Air Monitoring StationsSOWM: 1992 T1G5R UM Files

Drake Chemical Co.Lock Haven, PAPAD003058047

Sit* Location

AR3I9I83


APPENDIX B

Location of On-Site Perimeter Ambient Air Monitors

Bl

AR3I918U

co

Ca

qr

uu£•cV

AR319185


APPENDIX CToxicological Information

Cl

AR3I9I&6


Tnyicnlngiral Information

FenacFenac (2,3,6-trichhlorophenylacetic acid) is a persistent herbicide that is not known tobioaccumulate due to effective excretion in the urine028'. Due in part to this effectiveelimination mechanism, Fenac is only slightly to moderately toxic to mammalian speciesC9).Inhalation exposure standards for Fenac are not available. Although oral and inhalation dosesare not strictly comparable in terms of effect, such a comparison can at least provide someperspective as to the magnitude of estimated doses at Lock Haven. For example, the lowestrecorded oral LD50 in mammals is 576-1780 mg/kg in rats005. This is more than four billiontimes higher than the daily inhalation dose (0.000000143 mg/kg/day) that would result frombreathing 20 m3 air/day containing 0.0005 fig/m3 in the community surrounding the Drakesite.

A useful comparison can be made between the levels of Fenac in air at Lock Haven andOSHA's previous occupational exposure limit (TLV-TWA) for a more toxic herbicide, 2,4,5,-T, which is also no longer in use.. The previous TLV-TWA for 2,4,5-T was 10 mg/m3.031'This concentration was twenty million times higher than the average concentration of Fenacmeasured at the Lock Haven ambient air monitoring stations (0.0005 (ig/m3) during the trialburn. The margin of safety for the less toxic pesticide, Fenac, could be even greater.

Thus, in spite of an absence of detailed toxicological data, ATSDR concludes that the levels ofFenac in air resulting from incineration of contaminated soil at the Drake site will pose nothreat to human health at Lock Haven.

[\-naphrhy1amineBeta naphthylamine or 2-naphthylamine was formerly used as an intermediate in themanufacture of dyes and rubber, but is no longer manufactured in the U.S. for commercial usebecause it was found to cause bladder cancer in laboratory animals and in occupationallyexposed humans. Beta naphthylamine was not actually detected on- or off-site. However, thedetection limit (0.001 }ig/m3) of the method used was too high to enable ATSDR to determinewhether or not the "true" concentration of p-naphthylamine in air exceeded the comparisonvalue of 0.000048 M-g/m3 (EPA Region Ill's RBCc). If one were to make several highlyconservative assumptions, including: 1) the "true" average concentration of P-naphthylaminein air was the maximum possible, i.e., just barely below the detection limit or, essentially,0.001 p.g/m3, 2) the human bladder carcinogen, 2-naphthylamine, has no threshold of effectother than zero, and 3) the estimated lifetime "risk" of one "excess" cancer in a million at anexposure level of 0.00005 fig/m3 is a reasonable prediction of the resulting adverse healtheffects, then the corresponding theoretical lifetime cancer risk would be 20 per million or 0.2person in an exposed population of 10,000. The potentially exposed population living in the

C2

AR3I9I87


vicinity of the Drake Incinerator is about 10,000, rather than 1 million. Thus, not a single"excess" case of bladder cancer would be expected to occur in an exposed population of10,000, even if the exposure lasted for a lifetime of 70 years, instead of only 2 years or less atthe Drake site.

In reality, the "true" risk would be considerably lower. The assumption that chemicalcarcinogens exhibit no threshold of effect other than zero is a useful one from the regulatorypoint of view. EPA's 1986 cancer guidelines clearly warn that a cancer risk estimate based onsuch assumptions (which are built into the Linear Multistage Model that the Agency uses forcancer risk assessment) "does not necessarily give a realistic prediction of risk". Indeed, EPAhas stated that "the true risk is unknown and may be as low as zero".032* Furthermore, judging,by analogy, from the carcinogenic response of mice to 2-acetylaminofluorene,(33) anotherarlyamine bladder carcinogen, it seems reasonable to expect that P-naphthylamine wouldexhibit a practical threshold for cancer well in excess of the inhalation dose (0.006 and 0.0003mg/kg/day, respectively) associated with concentrations of p-naphthylamine equivalent toeither the practical detection limit of 0.02 ng/m3 or method detection limit of 0.001 |ig/m3.All of the above considerations apply equally well both on and off-site, since p-naphthylaminewas undetected at the same detection limit in both places.

C3

AR3I9I87/A

APPENDIX DAbbreviations and Terms

Dl

AR3I9I88

Listed and described below are the various comparison values that ATSDR uses to selectchemicals for further evaluation, along with the abbreviations for the most common units ofmeasure.

ATSDR = Agency for Toxic Substances and Disease RegistryCDD = chlorodibenzo-p-dioxinCDF = chlorodibenzofuran

i,

COC = Contaminants of ConcernCREG = Cancer Risk Evaluation GuidesDWEL = Drinking Water Equivalent LevelEMEG = Environmental Media Evaluation GuideEPA in = EPA Region IENIOSH = National Institute for Occupational Safety and HealthOSHA = Occupational Safety and Health AdministrationPEL = Permissible Exposure Limit (OSHA)REL = Recommended Exposure Limits (NOSH)SVOC = Semi-Volatile Organic CompoundsTCDD = 2,3,7,8-tetrachlorodibenzo-p-dioxinTEF = Toxicity Equivalency FactorTEQ = Toxicity Equivalency QuotientTLV = Threshold Limit Value (ACGffl)TSP = Total Suspended ParticulateVOC = Volatile Organic Compoundsppm = parts per million, e.g., mg/L or mg/kgppb = parts per billion, e.g., //g/L or y g/kgkg = kilogram (1,000 grams)mg = milligram (0.001 grams)jug = microgram (0.000001 grams)L = literm3 = cubic meter (used in reference to a volume of air equal to 1,000 liters)

Cancer Risk Evaluation Guides (CREGS) are estimated contaminant concentrations inwater, soil, or air that would be expected to cause no more than one excess cancer in a millionpersons exposed over a lifetime. CREGS are calculated from EPA's cancer slope factors.

ATSDR1 s Comparison Values (CVs) are media-specific concentrations that are considered tobe "safe" under default conditions of exposure. They are used as screening values in thepreliminary identification of "contaminants of concern" at a site. Generally, a chemical isselected as a contaminant of concern because its maximum concentration in air, water, or soilat the site exceeds one of ATSDR's comparison values. However, it cannot be emphasizedstrongly enough that comparison values are not thresholds of toxicity. While concentrations ator below the relevant comparison value may reasonably be considered safe, it does not

D2

AR3I9I89

automatically follow that any environmental concentration that exceeds a comparison valuewould be expected to produce adverse health effects.

Environmental Protection Agency Region ELI (EPA HI) values are similar to ATSDR'sEMEGs in that they are risk-based concentrations derived for carcinogens and non-carcinogensfrom RfDs and Cancer Slope Factors, respectively, assuming default values for body weight,exposure duration and frequency, etc. Unlike EMEGs, however, they are available for fish,as well as for water, soil, and air.

Risk-Based Concentrations (RBC) are derived by the Region m Office of the EnvironmentalProtection Agency. They represent levels of contaminants (non-carcinogens) in air, water,soil, and fish that are considered safe, assuming default values for body weight, exposureduration, and ingestion/inhalation rates.

Risk-Based Concentrations for Cancer (RBCc) are derived by the Region III Office of theEnvironmental Protection Agency. They represent levels of contaminants (carcinogens) in air,water, soil, and fish that are considered safe, assuming default values for body weight,exposure duration, and ingestion/inhalation rates.

Risk Burn -incineration of contaminated soil on the Drake Chemical site. ATSDR used thisdata to predict potential exposures to the community that might occur during remediation ofthe site.

Threshold Limit Values (TLV) are established by the American Conference of GovernmentIndustrial Hygienists (ACGIH). The TLV is the time-weighted average concentrations for anormal 8-hour workday and a 40-hour workweek, to which nearly all workers may berepeatedly exposed, day after day, without adverse effect. Many of ACGIH's TLVs wereadopted by OSHA for use as PELs. TLVs and PELs, which are designed to protect healthyworkers, are usually much higher than the health-based values of ATSDR and EPA, whichwere designed to protect the health of the general population, including the very young and theelderly. (Note: OSHA is required by statute to make rules that are economically feasible.)Although the ATSDR does not base any of its community health decisions on TLVs or PELs,it sometimes cites such values in Public Health Assessments merely as a means of puttingconcentrations of site-specific contaminants into a meaningful perspective for the reader.

Trial Burn Test -incineration of site soil spiked with known quantities of specific chemicals.ATSDR used this data to evaluate the performance of the incinerator.

D3

AR3I9I89A

FACT SHEET ATSDR's Health Consultations #2 and #3' f on Drake Chemical Thermal Destruction Facility

The Agency for Toxic Substances and Disease Registry (ATSDR) developed this fact sheet to providecommunities near the Drake Chemical Superfund Site information about health consultations #2 and #3 onthe site because the consultations are very technical. Comments were received from one resident and EPAon the public comment version of Health Consultation #2. Those comments and our responses are includedin an appendix of the final version. If you need more information about the site or about the agency andits activities you can call the agency at 1-800-447-1544.

Why is ATSDR involved at Drake?ATSDR conducted this technical review of the site-related documents and ambient air monitoring and stacksampling data because community members around the site raised concerns about the risk assessment, theambient air monitoring system, operation of the incinerator, and the modeling and weather data theEnvironmental Protection Agency (EPA) used in their risk assessments. ATSDR has completed theevaluation of actual stack and ambient air sampling data related to the Drake Chemical Superfund Site.

What does Health Consultation #2 cover?• It evaluated the incinerator design and operating conditions and the modeling that was done to

predict ground-level concentrations of stack emissions during the trial burn period.

What did ATSDR conclude in Health Consultation #2?• There is a need for new modeling before restarting the incinerator for full operation; modelers

should consider all seasons of the year and a longer history of meteorological data.

• The incinerator design includes a number of safety features that will help assure the properoperation of the incinerator. Good operation does not depend solely on the operator on duty "doingthe right thing"; the incinerator will automatically shut off the waste feed when the key operatingconditions exceed specific numbers or when certain conditions occur. All these conditions arecontinuously monitored and recorded by a computer or other instruments. Our staff have concludedthat all the key safety operating conditions that would protect the public from exposures due toemissions are included in the automatic waste feed shut-off system and are continuously monitoredand recorded.

• If the safety valve (thermal relief valve or TRV) is not opened, four events could result in anemergency situation that would release extremely hot incinerator gases at ground level (andpotentially cause a fire or explosion). It is much more protective of the public and workers torelease those hot gases (2,600°F) through the TRV stack than to allow a catastrophic equipmentfailure and the subsequent release of the gases at ground level. ATSDR concurs with EPA andPennsylvania staff in approving the use of the TRV if one of the designated events occurs.

• To evaluate the potential public health impacts, ATSDR needs actual stack emission data to predictemissions that may occur if the incinerator does not operate properly and the TRV opens.

What did ATSDR recommend to EPA in Health Consultation #2?• Conduct new modeling of the stack emissions measured during the trial bum and risk bum. We also

recommended that 5 years of meteorological data be used in the modeling.

RR3I9I90

• Reevaluate the assumptions used in projecting the emissions that will be released when operatingconditions cause the TRV to open.

What does Health Consultation #3 cover?• It evaluates three sets of air data; (1) on-site perimeter and off-site ambient air data collected before

the trial burn period, (2) stack emissions, on-site perimeter, and off-site ambient air data collectedduring the trial burn period, and (3) on-site perimeter and off-site ambient air data collected afterthe trial burn period.

• It evaluates whether the levels of stack emissions and fugitive emissions produced at the Drake siteduring the trial burn period were sufficient to cause any adverse health effects in the surroundingcommunity. It also evaluates whether future adverse health effects would be likely if the incineratoroperation produced similar emissions during 16 months of full operation.

What did ATSDR conclude in Health Consultation #3?• Off-site ambient air data collected at all four air monitoring stations (including dioxin and furan

measurements) within the community demonstrated that contaminants were not present at levels ofhealth concern.

• With the exception of the P-naphthylamine and Fenac, the analytical data for ambient air samplesare of generally high quality and acceptable for the evaluation of potential public health impacts.Due to poor surrogate recoveries using previous methods of detection, P-naphthylamine and Fenacdata were not detected in ambient air. The OSHA 93 analytical method appears to have good fieldand laboratory control data and to have a detection limit for Fenac and P-naphthylamine whichwould be protective of public health during the full burn period of operation.

What did ATSDR recommend to EPA in Health Consultation #3?• Continue sampling for volatile, semi-volatile organic compounds, metals, dioxin/furan, and total

suspended particulate off-site during the full burn operation.

• Use a method for detecting P-naphthylamine and Fenac with appropriate detection limits, field andlaboratory controls, so low ambient air levels can be reliably identified.

• Cover all stock piles of soil and ash to reduce dispersion of particulates off-site.

• Revise the site perimeter sampling plan to add several additional sampling protocols and actionlevels to ensure that other VOCs are not exceeding levels of public health concern.

What can be expected to follow?• After the public comment period on Health Consultation #3 closes on April 20, 1998, ATSDR will

address all comments received and issue the final Health Consultation #3.

• ATSDR will continue to evaluate the perimeter and off-site ambient air data for the first threemonths of full incineration operation to assure that ambient air concentrations to which thecommunity is being exposed are staying below levels of health concern.

A R 3 I 9 I 9 I

rr3i9i3i - semspub.epa.gov

Documents