Flow-Unit Architecture

GR Phi GR Phi GR Phi GR Phi GR Phi

High porosity Low porosity

1960 - Onset of waterflooding

1942 - Discovery

1980 - 20 years of waterflooding

2002 - 40 years of waterflooding

Simulation Results

<20

Oil Saturation(percent)

>85

Wichita

L2.1L2.2

<20

Oil Saturation(percent)

>85

Wichita

L2.1L2.2

<20

Oil Saturation(percent)

>85

Wichita

L2.1L2.2

<20

Oil Saturation(percent)

>85

Wichita

L2.1L2.2

Initial oil saturation was highest inhigh-porosity (25 - 30%) tidal-flatdolostone facies at top of Wichita andin subtidal dolostone and limestonefacies of the Lower Clear Fork.

Line-drive waterflooding patterninitiated in 1960. Prior primaryproduction came dominantly fromhigh-porosity Wichita tidal-flat faciesand L2.1 Lower Clear Fork.

Effects of line-drive injection areobvious afer 20 years.

Effects of conversion to patternflooding and infill drilling are apparent.Considerable bypassed oil exists,especially in Lower Clear Fork L2.2.

Well 2Well 1

A

B

Tubb

Wichita

HFS2.2HFS2.1

Top Tubb

HFS2.3

Wichita 8

Tubb

Wichita

HFS2.2HFS2.1

Top Tubb

HFS2.3

Wichita 8

A

B

Inversion modelWell 1 Well 2 (removed)

Tubb

Wichita

HFS2.2HFS2.1

Top Tubb

HFS2.3

Wichita 8

A

B

Tubb

Wichita

HFS2.2HFS2.1

Top Tubb

HFS2.3

Wichita 8

A

B

Inversion model with removed well

QAd4360(b)x

>15<3

Porosity

<-0.5 (0.3) >1 (10)

Log permeability

>70<20

So (%)

Low

High

Negativeamplitude

Low

High

Negativeamplitude

L2.2

L2.1

Wichita

3-D Seismic Inversion Modeling

Progressive Inversion

3-D Seismic for Porosity Imaging

3-D Seismic for Defining Architecture

ROLE OF 3-D SEISMIC IN RESERVOIR CHARACTERIZATION

Abo

Wichita

LCF

Tubb

L1

L2

L3

West East

UCF

RESERVOIR MODELING AND SIMULATION

ACKNOWLEDGMENTS

3-D Seismic for Siting Infill Wells

3-D seismic data are important for distinguishing the subhorizontal nature of theLower Clear Fork from the clinoformal architecture of the Abo but are misleadingwith respect to the position of the L1 - L2 sequence boundary.

Simple amplitude extractions provide superior imaging of porositydistribution in both interwell areas and areas where modern wireline dataare lacking. Note that contours based on wireline data are inaccurate.

Blind Testing of Inversion Model

Low AI High

High Porosity Low

1 km

0 1 mi

0

Blind testing demonstrates predictive ability of 3-D seismic in areas where wells are absent.When well 2 (above) is removed, wireline log model predicts too little porosity at A and toomuch at B. Inversion model does a much better job of defining true porosity in both areas.

Progressive inversion entails four steps:1. construction of detailed, cycle-based geological model.2. development of high-resolution, wireline log model.3. definition of geological framework in tuned, 3-D seismic data volume.4. Merging of wireline model and 3-D seismic volume.

r = 0.9

0 5 10 15 20 25 30

AI (

[m/s

] [g/

cm3]

)

Wireline Log Porosity (%)

12,000

15,000

18,000

21,000

Full-field Reservoir Modeling Flow Simulation

Study area

Sensitivity flow modeling was performed on a 2,000-acrearea selected for high-data quality and representative setting.The area includes 140 wells on 10- and 20-acre spacing.Modeling and simulation were based on 86 wells, all withmodern log suites. Model framework contains 38 cycle-based flow-unit layers.

The flow-simulation model area exhibits nearly the entire range of petrophysical andstratigraphic relationships observed in the full-field model. Results are thus applicable tomost of the reservoir.

Porosity is a function of diagenesis. Porosity is high in much of the Wichita tidal-flat succession and in theLower Clear Fork L2.1 subtidal. Lateral changes in porosity are more the result of structurally controlledearly diagenesis than changes in depositional facies.

Saturation is a function of rock fabric. Like permeability, oil saturation is highest in limestone-rich,ramp-crest facies of the subtidal Lower Clear Fork L2.1. Compare with porosity maps above.

Field-wide Porosity Trends

Field-wide Saturation Trends

Model Area Petrophysics

Field-wide Permeability Trends

Porosity (%)

<5 >20

L 2.1

Porosity

5 20(%)

L 2.2

Water Saturation

<0.05 >0.2

Water Saturation

<0.05 >0.2

Water Saturation

<0.05 >0.2

L 2.1

L 2.2

Water saturation

8020 (%)

Wichita

LowerClearFork

This research was co-funded by the U.S. Department of Energy, TheUniversity of Texas System, ExxonMobil Corporation, and the ReservoirCharacterization Research Laboratory of the Bureau of EconomicGeology. We thank John Stout, Jeff Simmons, and Craig Kemp ofOccidental Permian for providing important data for the study and forproviding thoughtful insights into project planning. Steve Krohn andAmy Powell helped develop and promote the project within ExxonMobil.Terry Anthony and David Smith, also with ExxonMobil provided ongoingoperator liaison, data exchange, and input on project activities anddeliverables. Tim Hunt, University of Texas System West TexasOperations Office, helped with technology transfer activities andgathered new GPS location data for wells in the field. Dan Fergusonserved as contract officer for DOE. The Bureau acknowledges supportof this research by Landmark Graphics Corporation via the LandmarkUniversity Grant Program.

PUBLISHED RESULTSA complete report on the findings of this study, entitled“Multidisciplinary imaging of rock properties in carbonatereservoirs for flow unit targeting”, has been released by theBureau of Economic Geology. Contents are listed below.

Ruppel, S. C. and Jones, R. H., Facies, Sequence Stratigraphyand Porosity Development in the Fullerton Clear ForkReservoir, p. 1-124, 52 figures.

Jones, R. H. and Lucia, F. J., Integration of Rock Fabric,Petrophysical Class, and Stratigraphy for PetrophysicalQuantification of Sequence-Stratigraphic Framework, FullertonClear Fork Field, Texas, p. 125-162, 23 figures.

Kane, J. A. and Jennings, J. W. Jr., Development of the Wireline-Log Database and Determination of Porosity Using WirelineLogs, p. 163-187, 9 figures.

Lucia, F. J., and Kane, J. A., Calculations of Permeability andInitial Water Saturations from Wireline Logs, p. 189-218, 16figures.

Wang, Fred and Lucia, F. J., Reservoir Modeling and Simulationof Fullerton Clear Fork Field, Andrews County, Texas, p. 219-304, 56 figures.

Zeng, Hongliu, Construction and Analysis of 3-D Seismic PorosityInversion Models, p. 305-342, 17 figures.

A pdf copy of this report is available from:

The Bureau of Economic Geology website <http://www.beg.utexas.edu/resprog/fullerton/index.htm>

The U.S. Department of Energy website <http://www.osti.gov/servlets/purl/837750-7tCynd/webviewable/>

A more complete CD publication of all results is being preparedand will be published and distributed by the Bureau of EconomicGeology in 2006.

Flow Modeling Area Flow-Model Attributes

100

GR Phi

Wic

hita

Low

er C

lear

For

k

Sequences Cycles andflow layers

Stratigraphic Architecture• 3-D seismic data can provide inaccurate image of reservoir

architecture• Sequence boundary karst is widespread• Porosity is a function of diagenesis, but permeability is a

function of facies• Porosity development is partially controlled by deep structurePetrophysics• Peritidal successions contain high-porosity, low-permeability

rock fabrics• Subtidal successions contain higher permeability rock fabrics• Peritidal limestones are flow baffles; subtidal limestones are

high-flow zones• A single porosity cutoff or permeability transform is inadequate• Permeability and water saturation models must consider rock-

fabric distribution• Petrophysical classes and rock fabrics can be mapped

throughout the field using stratigraphic framework3-D Geophysics• Strong relationship between amplitude and porosity• Amplitude extractions provide robust qualitative guide to

interwell and extrawell porosity distribution• Progressive 3-D inversion provides quantitative measure of

interwell and extrawell porosity distribution• Seismic response controlled by porosity (facies and diagenesis)• Seismic architecture must be vetted by geological models• Continuing and differential fault motion through PermianFlow Modeling and Simulation• Key insights to sweep and remaining oil distribution• Recovery efficiency controlled by rock fabric, continuity, and

completion coverage Reservoir Resources• Peritidal facies contains 55 percent of the total pore volume

but only 43 percent of the original hydrocarbon pore volume

P e rm e ab il i t y

0.1 10(md)

P e rm e ab il i t y

0.1 10(md)

P e rm e ab il i t y

0.1 10(md)

Log permeability

-1 (0.1) 1 (10)(md)

Permeability is a function of facies. Note that in contrast to porosity, permeability is much higher in thesubtidal Lower Clear Fork than in Wichita tidal-flat facies. The Wichita exhibits much lower permeabilitiesbecause of the dominance of Type 3 (mud-rich) rock fabrics.

L2.2 L2.1

Lower Clear Fork Wichita

upper L2.0 lower L2.0 upper L1.0 lower L1.0

L2.2 L2.1Lower Clear Fork Wichita

upper L2.0 lower L2.0 upper L1.0 lower L1.0

L2.2 L2.1Lower Clear Fork Wichita

upper L2.0 lower L2.0 upper L1.0 lower L1.0

28112813

2913

30133015

31133115

2910

Producing well

Proposed infill well

3-D seismic

Unitboundary

Amplitude extractions provide powerful constraints fordefining optimal locations for infill wells where otherporosity control is lacking. Note that half of the initiallyproposed wells (yellow stars above) would have beendrilled into low-porosity areas without 3-D data.

Lower Clear Fork HFS L2.1AMPLITUDE EXTRACTIONS

12

8

4

68

4

12

12

3-D seismic

LowerClear Fork

Outer Ramp

Core

Image log

Wichitainner ramp

PorosityDistribution

fromWireline

Logs

3-D seismic

Wireline Log Data 3-D Amplitude Extraction

Unitboundary

(PhiH)

ROLE OF 3-D SEISMIC IN RESERVOIR CHARACTERIZATION RESERVOIR MODELING AND SIMULATIONIMPORTANT FINDINGS

Lower Clear Fork (L2.1, L2.2) and upper Wichita (L2.0)sequences comprise cycle-based, subparallel layers (7, 7and 5, respectively) that extend across the model area. The

AboOWC-4100

LowerClearFork

Wichita L1

L2.1

L2.0

L2.2

Model Framework Cycle-based Layers

lower Wichita contains five layers that terminate at theWichita-Abo facies transition. Abo layers are conceptuallymodeled clinoforms.

L2.2

L2.1

Wichita

L2.2

L2.1

Wichita

Initial wireline modelWell 2Well 1

A

B

A

B

Well 1 Well 2 (removed)

A

B

A

B

Initial wireline model with removed well

L 1

Geologic Model

High-Resolution Wireline Model

L

L

L

W E

Combined Wireline and 3-D ModelW E

L

L

L

Seismic DataW E

L

L

L

Tubb

Wichita

HFS2.2HFS2.1

HFS2.3

Abo

Tubb

Wichita

HFS2.2HFS2.1

HFS2.3

Abo

Model Dimensions

X

Y

Z

Total

103 140 200 73

48

39

136,656

105

2.2 (0-16)

90 200

254

3.16 million

Fine-grid Model Simulation ModelCellsDimension (ft) Cells Dimension (ft)

2,000-acre model

Model Dimensions

X

Y

Z

Total

150 140 150 239

462

36

150 462 150

380

42 million

Fine-grid Model Simulation Model

CellsDimension (ft) Cells Dimension (ft)

35,000-acre Model

3.7 million

Model Attributes

>3

<1

logKH(mD-ft)

>0.12

<0.05Porosity

>12

<2SoPhiH

3-D seismic inversion has value only if astrong relationship exists between 3-D dataand porosity. At Fullerton field, there is a

good relationship between acousticimpedance (AI) and wireline log porosity.

L 2.2L 2.1

Wichita

LowerClearFork

Wichita

LowerClearFork

BureauofEconomic

Geology

N

0 8000 ft

2000 m0

0 8000 ft

2000 m0

0

0 2000 ft

500 m

NN