assessing spatial uncertainty in reservoir … › ... › uploaded › eg09008.pdfassessing spatial...

AUTHORS

Erik R Venteris Ohio Department of Nat-ural Resources Division of Geological Survey2045 Morse Rd Bldg C-1 Columbus Ohio 43229erikventerisdnrstateohus

Erik Venteris a senior geologist for the OhioDepartment of Natural Resources Division ofGeological Survey has worked on various car-bon sequestration projects involving soil-basedand geology-based approaches and has beencollaborating with the MRCSP since 2004 His in-terests include applying statistics and geostatis-

Assessing spatial uncertaintyin reservoir characterizationfor carbon sequestrationplanning using public well-logdata A case studyErik R Venteris and Kristin M Carter

tics to earth science problems Erik holds a PhDin soil science and an MS degree in geologyfrom the Ohio State University and a BS degreein geology from Western Illinois University

Kristin M Carter Pennsylvania GeologicalSurvey 400 Waterfront Dr Pittsburgh Pennsyl-vania 15222 krcarterstatepaus

Kristin Carter joined the Pennsylvania GeologicalSurvey in 2001 and currently serves as the chiefof the Carbon Sequestration Section Kristin re-searches oil gas and subsurface geology inPennsylvania and surrounding states particularlyas they relate to geologic carbon sequestrationopportunities Kristin received an MS degreein geological sciences from Lehigh Universityand a BS degree in geology and environmentalscience from Allegheny College

ACKNOWLEDGEMENTS

The authors thank both Christopher Laughrey(Pennsylvania Geological Survey) and James Castle(Clemson University) for their insightful and con-structive comments regarding the petrophysicsof the Medina play in northwestern PennsylvaniaIn addition we thank Nathan Yancheff Kelly Sagerand Brooke Molde three of the Pennsylvania

ABSTRACT

Mapping and characterization of potential geologic reservoirs are keycomponents in planning carbon dioxide (CO2) injection projectsThe geometry of target and confining layers is vital to ensure thatthe injected CO2 remains in a supercritical state and is confined tothe target layer Also maps of injection volume (porosity) are neces-sary to estimate sequestration capacity at undrilled locations Ourstudy uses publicly filed geophysical logs and geostatistical modelingmethods to investigate the reliability of spatial prediction for oil andgas plays in the Medina Group (sandstone and shale facies) in north-western Pennsylvania Specifically the modeling focused on twotargets the Grimsby Formation and Whirlpool Sandstone For eachlayer thousands of data points were available to model structureand thickness but only hundreds were available to support volu-metric modeling because of the rarity of density-porosity logs inthe public records Geostatistical analysis based on this data resultedin accurate structure models less accurate isopach models and in-consistent models of pore volume Of the two layers studied only theWhirlpool Sandstone data provided for a useful spatial model of porevolume Where reliable models for spatial prediction are absent thebest predictor available for unsampled locations is the mean value ofthe data and potential sequestration sites should be planned as closeas possible to existing wells with volumetric data

Geological Surveyrsquos summer student workers whohelped to gather and interpret geologic data inthe study area Gary Weismann is thanked for pro-viding a thoughtful review of the original manu-script Acknowledgement is also made to the MidwestRegional Carbon Sequestration Partnership (MRCSP)managed by Battelle Memorial Institute andfunded in large part by the US Department of En-

INTRODUCTION

Geologic sequestration of carbon dioxide (CO2) is a promising ap-proach for reducing atmospheric greenhouse gas levels Sequestra-tion of CO2 in geologic units may be accomplished using a variety of

ergy As part of the MRCSP the Ohio and Penn-sylvania Geological Surveys have been able to gatherinterpret and evaluate reservoir and geostatisticaldata relative to the geologic sequestration of sev-eral subsurface units in the Appalachian Basinincluding the Medina GroupldquoClintonrdquo Sandstone

Copyright copy2009 The American Association of Petroleum GeologistsDivision of EnvironmentalGeosciences All rights reservedDOI101306eg04080909008

Environmental Geosciences v 16 no 4 (December 2009) pp 211ndash234 211

different storage mechanisms The commonly discussedstorage types include volumetric solution adsorptionand mineral Volumetric storage refers to the amountof CO2 that is retained in the pore space of a geologicunit generally as a supercritical phase retained by struc-tural or stratigraphic traps or by overlying cap rock(Wickstrom et al 2005) Mapping the location and es-timating the total capacity of this storage type for sev-eral geologic units in the north central AppalachianBasin have been the focus of our research

As in oil-and-gas exploration reservoir characteri-zation is a key task in planning CO2 injection projectsSpatial models and maps are needed to ensure (1) suf-ficient overburden to maintain CO2 in a supercriticalstate (approximate depth of 2500 ft [762m]) (2) largeenough porosity and permeability to support a practicalinjection rate and (3) sufficient total CO2 capacity toensure the economic viability of the project In addi-tion selection and evaluation of potential injection sitesrequire a detailed understanding of the overlying andunderlying formation(s) that would serve as seals (a topicbeyond the scope of this contribution) Accordinglyrock characteristics critical to characterizing a potentialsequestration reservoir include depth thickness poros-ity water saturation permeability and the lateral andvertical heterogeneity of these parameters Petrophysi-cal analyses of geophysical well logs and core samplesare used to determine these properties at individualboreholes and the spatial models are used to estimatevalues for these throughout the area of interest

The selection of sites for carbon sequestration proj-ects includes a wide range of geologic and nongeologicfactors (Wickstrom et al 2005 Venteris et al 2008)Because of the complexity of site selection and planningpredicting the sequestration capacity at unsampled lo-cations is advantageous Maps and other spatial modelsof reservoir characteristics are the most convenientmethods available to provide such information Geosta-tistical reservoir characterization methods (Deutsch2002) provide an objective optimized and scientificallydefensible method for estimating reservoir characteris-tics at unsampled locations providing measures of theaccuracy of such estimates and facilitating the genera-tion of continuous models (rasters) for use in regionalcapacity calculations (Venteris et al 2008)

As part of our early research efforts we used krigingto map structure elevation data and thickness for theMedina GroupldquoClintonrdquo Sandstone in PennsylvaniaWest Virginia and Ohio The Medina was chosen formore detailed study because a large number of borings

212 Spatial Uncertainty in Reservoir Characterization for Carbon

are available in northwestern Pennsylvania and our pre-viousmapping effort showed relatively high accuracy instructure and isopach mapping relative to other poten-tial reservoirs (Carter and Venteris 2005 Wickstromet al 2005) The lack of prominent faulting in the for-mation also simplified the application of geostatisticalmethods In this study conducted at a higher resolutionwith more data than the previous works we seek morefully to characterize the reservoir through refinementof the structure and isopachmodels and through the ad-dition of spatial models of the available pore space

Our case study uses geostatistical methods (krigingand stochastic simulation) to quantify the strength ofspatial prediction for a range of reservoir parametersbased on a typical public oil-and-gas data set for the Ap-palachian Basin Amajor goal of this study was to inves-tigate the usefulness of the available public data for eachof the various reservoir parameters to identify knowl-edge gaps and design future data collection efforts Thiscontribution illustrates the dangers of mapping withblack box approaches typical of some reservoir charac-terization software that do not encourage or allow themodeler to explore the data for outliers investigate andquantify the spatial structure of the data for use in pre-diction and evaluate the error of the predictions Amap whether drawn by computer algorithm or handcontouring implies that the person creating the mapscan predict values at unsampled locations The qualityof that prediction must be evaluated For accurate spa-tial prediction (1) the error in the measurements mustbe small compared to the magnitude of spatial variabil-ity (2) that spatial variability must be at least partiallyordered in space (not random) and (3) the distance be-tween the samples (wells) must be less than the range ofthe spatial structure If these conditions are not metpredicted values beyond the data points are highly sus-pect and the map is of little value This work seeks toillustrate these spatial modeling issues using various res-ervoir parameters of the Medina Group

STUDY AREA

The current study area encompasses twomajor oil-and-gas fields in northwestern Pennsylvania and incorpo-rates publicly available geologic data from more than5600 deep wells that penetrate the Early SilurianndashageMedina Group throughout Erie Crawford Mercerand Venango counties (Figure 1) In general oil and gasfields are of interest in sequestration projects because

Sequestration Planning

Figure 1 Map of northwestern Pennsylvania showing the well locations used in this study

Venteris and Carter 213

of demonstrated fluid and gas storage capacity and thepotential at some locations for CO2-enhanced oil recov-ery The Medina Group has been extensively exploredthroughout these and surrounding areas but perhaps themost notable Medina producing areas are the Athensand Conneaut fields which are discussed in further de-tail below as examples of the productivity of this play innorthwestern Pennsylvania

Oil and Gas Fields

TheConneaut field inwesternCrawford and Erie coun-ties encompasses 128050 ac (52E8 m2) (McCormacet al 1996) (Figure 2) The field was discovered in1957 and produces oil and gas primarily from theGrims-by Formation and Whirlpool Sandstone (Lytle et al1961) A part of the field features a thin sandstone lenswithin the upper part of the Cabot Head Shale thatthickens considerably and provides additional produc-


tion Operators call this the Tracy sandWhere the Tracysand is well developed the Pennsylvania GeologicalSurvey typically includes it within the Grimsby Forma-tion The average producing depth of Medina units isapproximately 3600 ft (1097m) and pay thicknesses av-erage 12 ft (4 m) Reservoir porosity ranges from 3 to18 averaging 10 Reservoir temperature and initialpressure were reported at 104degF (40degC) and 1100 psi(7584 kPa) respectively (McCormac et al 1996) Thetotal cumulative gas production is 981 bcf (27E12 L)through 2006 and the total estimated cumulative oilproduction for the past 22 yr is roughly 16 millionbbl (25E8 L) (WIS 2009)

The Athens field is located in eastern CrawfordCounty and includes 23456 ac (95E7 m2) (McCormacet al 1996) (Figure 2) The field was discovered in1974 and produces gas from both the Grimsby Forma-tion and Whirlpool Sandstone (Laughrey 1984) Theaverage producing depth of these reservoir rocks is

Figure 2 Map of northwestern Pennsylvania showing the locations of the Athens and Conneaut fields and the locations of the crosssections in Figures 4 and 5


approximately 4800 ft (1463 m) and pay thicknesses av-erage 42 ft (13 m) Porosity ranges from 2 to 9 aver-aging 56 Reservoir temperature and initial pressurewere reported at 106degF (41degC) and 1220 psi (8412 kPa)respectively (McCormac et al 1996) The total cumu-lative gas production through 2006 for the Athens fieldis 305 bcf (86E11 L) and the total estimated cumula-tive oil production through 2002 the last year oil pro-duction was reported is less than 1000 bbl (WIS 2009)

Lithostratigraphy of the Medina Group

The Medina Group of northwestern Pennsylvania con-sists of threemajor lithostratigraphic units in ascending

order the Whirlpool Sandstone the Cabot Head Shale(sometimes called the Power Glen Shale) and theGrimsby Formation (Figure 3) The Whirlpool Sand-stone forms the basal unit of this lithostratigraphic inter-val and throughout much of the basin is composed ofa white to light-gray to red fine- to very fine-grainedmoderately well-sorted quartzose sandstone with suban-gular to subrounded grains (Piotrowski 1981 Brett et al1995 McCormac et al 1996) The Cabot Head Shaleis a dark-green to black marine shale with thin quartz-ose siltstone and sandstone laminations that increasein number and in places thicken upward in the unit(Piotrowski 1981 Laughrey 1984) The sandstones oftheGrimsby Formation are very fine- tomedium-grained

Figure 3 Stratigraphic columnshowing the rocks of the MedinaGroup in northwestern Penn-sylvania and equivalent rocks inOhio

V

enteris and Carter 215

monocrystalline quartzose rocks with subangular tosubrounded grains variable sorting and thin discontin-uous silty shale interbeds Cementingmaterials includesecondary silica evaporites hematite and carbonateminerals (Piotrowski 1981 McCormac et al 1996)

We prepared lithostratigraphic cross sections(Figures 4 5) parallel and perpendicular to strike with-in the Conneaut field area to further illustrate Medinageology thickness and geophysical log responses Thesestructural cross sections are hung on the top of theRochester Shale Gross thicknesses of the Grimsby For-mation andWhirlpool Sandstone range from85 to 170 ft(26 to 52 m) and 10 to 30 ft (3 to 9 m) respectivelyGamma-ray logs are included for all wells (Track 1) andwhere available density and neutron porosity logs(Track 2) are provided

The depth to the top of the Medina Group rangesfrom approximately 2000 to 6400 ft (610 to 1951 m)throughout the study area The structure on top of theMedina Group strikes northeastndashsouthwest and dipssoutheastward at a rate of 30 to 50 ftmi (6 to 9 mkmFigure 6) Figures 7 and 8 illustrate the thickness of theGrimsby Formation and Whirlpool Sandstone acrossthe study area respectively the gross thicknesses of theentire Medina Group ranges from just less than 80 ft(24 m) to about 400 ft (122 m) Early studies of theMedina and equivalent units were performed in the1960s through early 1980s (Yeakel 1962 Knight1969 Martini 1971 Piotrowski 1981 Cotter 19821983 Laughrey 1984) A summary of these and re-lated works is provided by McCormac et al (1996)

The depositional history of the Medina GroupldquoClintonrdquo Sandstone began near the end of the Taconicorogeny in the Early Silurian During this period clasticmaterial was being eroded from both foreland fold-belthighlands adjacent to the eastern edge of the Appala-chian Basin and the plutonic igneous rocks of the islandarc orogen (Laughrey 1984 Laughrey and Harper1986 McCormac et al 1996) The directions of sedi-ment transport from these highlands were both parallel(ie northeastndashsouthwest trending) and perpendicular(northwestward) to the shoreline (Figure 9) (Laughreyand Harper 1986) which ran from what is now north-ern Beaver County to central Warren County in Penn-sylvania (Piotrowski 1981) Martini (1971) interpretedthe Medina depositional system as a shelf longshore-bar tidal-flat and delta complex based on outcrop stud-ies conducted in western New York and southern On-tario TheWhirlpool Sandstone is the basal transgressiveunit of this system and is overlain by shelf muds tran-


sitional silty sands and lower shoreface sands of theCabot Head Shale These sediments were in turn over-lain by shoreface and nearshore sands of the lower Grim-sby Formation and later argillaceous sands at the top ofthis unit (Laughrey 1984 Laughrey and Harper 1986McCormac et al 1996) Laughrey (1984) dividedthe Medina Grouprsquos depositional system into five fa-cies (1) tidal-flat tidal-creek and lagoonal sediments(2) braided fluvial-channel sediments (3) littoral de-posits (4) offshore bars and (5) sublittoral sheet sandsFacies 1 2 and 3 sediments comprise the Grimsby For-mation which was deposited in a complex deltaic toshallow-marine environment The deeper offshore-mudand sand-bar deposits of Facies 4 were reworked by bothstorm and tidal currents to become transitional sand-stones of the Cabot Head Shale The Whirlpool Sand-stone is included in facies 5 which formed in nearshoremarine and fluvial braided-river environments that existedat the beginning of a marine transgression (Piotrowski1981 Laughrey 1984 McCormac et al 1996) Withthe advent of sequence stratigraphy as an important res-ervoir rock interpretation tool in the 1990s several stud-ies reevaluated Early Silurianndashage rocks in the northernAppalachian Basin including those of Castle (19982001) Hettinger (2001) and Ryder (2004) Researchperformed byCastle (1998 2001) onMedina and equiv-alent cores and outcrops throughout the basin identifiedsix different depositional facies for this rock sequenceincluding fluvial estuarine upper shoreface lower shore-face tidal channel and tidal flat Furthermore Castleidentified three types of sequences in these rocks (1) afining-upward sequence which includes upper and lowershoreface facies associated with incised valley-fill de-posits (2) a coarsening-upward sequence (type A) thatcomprises several tidal facies and is representative of aprogradational shoreline and (3) a coarsening-upwardsequence (type B) which comprises fluvial and estua-rine facies that are interpreted as aggradational Of thesethree the coarsening-upward sequence (type B) prevailsin oil and gas wells of the study area (Castle 1998 2001)

METHODOLOGY

Geophysical Log Interpretation

We evaluated two types of geophysical logs for the cur-rent study gamma ray and density-porosity logs Eachof these provides information critical to performingvolumetric calculations of potential CO2 sequestration


Figure 4 Cross section A along strike showing geophysical logs and picks for the major rock layers from the Rochester Shale to the Queenston Shale

Venterisand

Carter217

Figure 5 Cross section B along dip showing geophysical logs and picks for the major rock layers from the Rochester Shale to the Queenston Shale

218SpatialUncertainty

inReservoirCharacterization

forCarbonSequestration

Planning

Figure 6 Structure map (interpolated using ordinary kriging) drawn on the top of the Medina Group in northwestern Pennsylvania


Figure 7 Results from 100 SGSIM runs showing the average thickness and standard deviation (inset) of the Grimsby Formation

220 Spatial Uncertainty in Reservoir Characterization for Carbon Sequestration Planning

Figure 8 Results from 100 SGSIM runs showing the average thickness and standard deviation (inset) of the Whirlpool Sandstone


capacity for Medina Group sands The number of logsavailable for this study was ultimately limited by theavailability of digital log curves for wells completed inthe study area The Commonwealth of Pennsylvaniadoes not require the submittal of digital geophysicallogs with completion reports therefore only paper logsare received by the Pennsylvania Geological SurveyConsequently all of the digital-log curves used in thisstudy were generated in house by the Pennsylvania Geo-logical Survey

Gamma-ray logs were evaluated for 360 wellsthroughout the study area For those completely pene-trating the Medina Group interval a single clean sandcutoff of 60 (ie 80 API units) was selected as anaverage of what has been used previously for so-calledcleanMedinawells in northwesternCrawford andwest-ern Erie counties (Kelley 1966 Kelley and McGlade


1969 Piotrowski 1981) and dirty Medina wells fartherto the south and east inCrawfordMercer andVenangocounties (Laughrey 1984)

Density-porosity logs were available for 176 wellsin the study area all of which penetrated the entireMedinaGroup intervalMost of these were digitized di-rectly from density-porosity curves provided by the op-erator but in those instances where only bulk densitycurves were available density porosity (f) was calcu-lated at 05-ft (015-m) intervals using the followingformula per Asquith and Krygowski (2004)

f frac14 rma rbrma rf

(1)

where rma and rf are given as the grain density and fluiddensity respectively and rb is the bulk density as recorded

Figure 9 Regional paleoenvironmental interpretations for the Grimsby Formation (and equivalent units) of the Medina GroupldquoClintonrdquoSandstone play Four different shoreline configurations (labeled 1 through 4) have been proposed by Knight (1969) Pees (1987) Keltchet al (1990) and Coogan (1991) respectively (modified from Castle 1998)


on the geophysical log For this work we used grain andfluid densities of 267 and 09 gcm3 respectively asrecommended by Castle and Byrnes (1998) Minimumporosity cutoffs as are necessary for any volume-basedsequestration calculations were established at 6 whichis on the conservative side of previously reported values(ie 3 to 6 per Laughrey 1984 andCastle and Byrnes1998)

Pore space was estimated from the logs using a vari-ety of approaches common in oil and gas explorationpractice which provided a range of parameters to in-vestigate for spatial structure Parameters calculatedfrom the logs for further geostatistical evaluation in-clude thickness of sand exceeding 60 (ISOCSWHIRLISOCSGRIM for Whirlpool Sandstone and GrimsbySandstone respectively) average porosity in the cleansandpart (MPWHIRLMPGRIM) feet of porosity in theclean sandexceeding6cutoff (PFTWHIRLPFTGRIM)and the pore volume (PVWHIRL PVGRIM) calcu-lated by multiplying the thickness of clean sand by theaverage porositywithin the clean sand The global use ofa 60 sand cutoff combined with a 6 minimum po-rosity cutoff in our porosity-feet calculations necessarilyexcludes certain porous zones in dirty Medina wells ofthe study area Consequently the porosity-feet statisticsin Table 1 particularly for the Grimsby Formation may

appear low to reservoir geologists who know the produc-tive nature of theGrimsby in eastern CrawfordMercerand Venango counties

Geostatistical Modeling

The prediction at unsampled locations and the genera-tion of continuous maps (rasters) were conducted usingkriging (ordinary or simple kriging with or without de-trending) In addition geostatistical simulation (sequen-tialGaussian simulation [SGSIM] Isaaks 1990Deutsch2002) was used to generate maps of uncertainty for caseswhere the spatial structure was weak In addition to theavailability of stochastic simulation kriging was favoredover other interpolation methods because of the addi-tional information available through variogrammodelingThe ratio of the partial sill to the nugget effect providedinformation on the strength and consistency of spatialautocorrelation within the data set Experimental vario-grams without structure (pure nugget effect) demon-strate that the scale of spatial variation is less than thesample spacing andor that the measurement error islarge compared to spatial variability In such situations(also revealed through cross validation) spatial predic-tion through interpolation provides little to no addi-tional information and the sample mean (or median

Table 1 Univariate Statistics for Spatial Models Used in this Study

Data

N

Min

Median

Max

Average

Venteris and C

Standard Deviation

Depth top Medina

5598

2082

4593

6388

4542

7837

Elevation top Medina

5598

minus4933

minus3315

minus1512

minus3237

7350

Depth top Whirlpool (ft)

5295

2226

4784

6567

4731

77595

Elevation top Whirlpool (ft)

5297

minus5137

minus3508

minus1656

minus3420

7389

Elevation bottom Whirlpool (ft)

5375

minus5153

minus3508

minus1668

3424

7422

Isopach Whirlpool (ft)

5278

1

11

47

1165

403

ISOCSWHIRL (ft)

329

1

85

23

848

398

MPWHIRL ()

160

152

554

2013

574

218

PFTWHIRL (ft)

122

003

033

169

041

032

PVWHIRL (ft3)

160

004

047

179

047

032

Depth top Grimsby (ft)

5375

2082

4597

6388

4549

7762

Elevation top Grimsby (ft)

5375

minus4922

minus3320

minus1512

minus3242

7330

Elevation bottom Grimsby (ft)

3810

minus5097

minus3724

minus1623

minus3530

7431

Isopach Grimsby (ft)

3798

5

120

195

1181

220

ISOCSGRIM (ft)

243

3

418

1285

452

216

MPGRIM ()

155

149

625

2278

642

216

PFTGRIM (ft)

152

007

169

697

194

130

PVGRIM (ft3)

155

013

250

732

278

126

ISOCSWHIRL ISOCSGRIM = thickness clean sand (exceeding 60 sand) for Whirlpool Sandstone and Grimsby Formation respectively MPWHIRL MPGRIM = averageporosity in the clean sand portion PFTWHIRL PFTGRIM = feet of porosity in the clean sand exceeding 6 cutoff PVWHIRL PVGRIM = pore volume

arter 223

for highly skewed data) is the best predictor at un-sampled locations

Geostatistical analysis proceeded using the follow-ing steps for each parameter modeled Summary statis-tics (Table 1) histograms and box plots were calcu-lated using JMPreg (SAS Institute Inc 2008) softwareto identify outliers check for skewed distributionsand obtain measures of center and spread Spatial andgeostatistical analyses were conducted using ArcMapand Geostatistical Analyst (GA) (ESRI 2009) respec-tively Each property was mapped as point data and in-vestigated for broad trends using the trend analysis soft-ware available in GA Trends if present were fitted andremoved from the data using global first-order polyno-mials (linear) inGAWhere simple kriging was used thedata were declustered using the cell method (Deutschand Journel 1998) and a normal score transformationwas applied Variogrammodeling and kriging were con-ducted using GA Experimental variograms were calcu-lated and model variograms were fitted For each dataset a heuristic procedure was followed where a rangeof variogrammodels were calculated and tested by one-out cross validation Variograms were calculated over awide range of scales by altering lag spacings (lag spacingsfrom 500 to 50000 ft [152 to 15244 m]) and numberof lags Anisotropic effects using or not using the nuggeteffect and detrending or not detrending were testedWeak anisotropy was detected in a few of the variogrammodels but inclusion in the kriging models did not im-prove cross validation results Without prior evidenceto justify the more complex anisotropic model isotro-pic variograms were used We tested a range of modelvariogram functions as well a spherical model was usedfor all because exponential and k-Bessel variogrammodels did not improve the results The neighborhoodsearch algorithm was held constant over all parametersThe points for each kriging estimate were chosen usinga quadrant scheme with five data points chosen perquadrant (for a total of 20 for each estimate) Themaxi-mum search rangewas set to the range of themodel vario-gram The choice of final variogram was not formallyoptimized but the variogram providing the lowest rootmean squared error (RMSE) from cross validation waschosen for the final kriging model Side investigationsshowed similar error estimates between cross and exter-nal validation for the structure and isopach data but thepetrophysical parameters had too few data to permit ex-ternal validation

In addition to kriging SGSIM was used to modelthe isopachs for both theGrimsby Formation andWhirl-


pool Sandstone and the pore volume for the WhirlpoolSandstone This was done to further evaluate the uncer-tainty of these geostatistical models which exhibited alarge nugget effect and large cross validation RMSE(uncertainty mapping based only on the kriging vari-ance provided information of limited use because thekriging variance is independent of the data valuesDeutsch and Journel 1998) The SGSIM uses krigingin conjunction withMonte Carlo simulation to producea range of statistically compatible realizations each re-producing the spatial structure (instead of the smoothkriging estimates) and the cumulative distribution ofthe data (Deutsch 2002) In SGSIM grid cells to be in-formed are selected along a random path For each cellthe value is estimated from the simple kriging estimateplus a random residual drawn from normal distribution(with a mean of zero and a variance equal to the simplekriging variance) Each estimated cell is treated as a datapoint for subsequent estimates ensuring that themodelfaithfully reproduces the variogram and distributionSummary statistics calculated from the multiple spatialmodels can then be used to evaluate the quality of thespatial prediction The SGSIM was performed in GAbased on the simple kriging model (Table 2) The meanand standard deviation for each simulated cell was cal-culated from 100 simulation runs and mapped

RESULTS

Results from exploratory and geostatistical analyses arepresented in Tables 1 and 2 respectively The spatialmodeling results for each reservoir parameter are dis-cussed below

Formation Top and Overburden Thickness

The thickness of the rock above the target formation iscrucial for sequestration planning An overburden thick-ness of at least 2500 ft (762m) (Wickstrom et al 2005)is needed tomaintain the injectedCO2 in a dense super-critical state Overburden thickness can be modeledeither by kriging the depth data directly or by krigingthe subsea elevation and then computing the differencebetween a digital elevationmodel (DEM) of the groundsurface and the subsea elevation surface For this exer-cise the latter approach was preferred because it wasmore accurate The variogram is stronger for the eleva-tion of the top of theMedinaGroup (Figure 10 Table 2)


than for the depth data (Table 2) The cross validationerror for the depth data is 523 ft (159 m RMSE)whereas the error for the subsea top is 282 ft (86 mRMSE) primarily because the subsea data contain onlythe variation in the stratigraphic boundary but the depthdata contain variability from both the stratigraphic topand the ground topography The accuracy of the DEMin northwestern Pennsylvania was not measured di-rectly but for a DEM in eastern Ohio in similar topogra-phy Venteris and Slater (2005) determined an accuracyof approximately 10 ft (3 m) at 30-m (98-ft) resolutionwhen compared to elevation data obtained by precisionglobal positioning systems Assuming no covariance be-tween the error sources the total error in the second ap-proach to determine the overburden thickness is 299 ft(91 m RMSE)

The overburden thickness map (Figure 11) showsthat for most of the study area the Medina Group hassufficient overburden for injection projects Thicknesssteadily increases from the Lake Erie shoreline to thesoutheast toward the basin center Only along the LakeErie shoreline has the overburden thickness become in-sufficient to support sequestration projects

Isopach

As a first step in estimating the volume available for se-questration we mapped the gross interval thickness forthe Grimsby Formation and Whirlpool Sandstone thetwo main targets within the Medina Group Althoughnot typically used directly in volumetric estimates theisopachs were mapped because their calculation re-quires only two picks off geophysical logs Thus the iso-pach potentially contains less noise than the volumetriccalculations forwhich themeasurements aremore com-plex Spatial variation in volume could potentially bemapped using a spatial isopach model and the averageporosity if porosity proved difficult to map

With isopachmodeling the question arises as to thebest method of calculation Two possible approachesexist one involves calculating the thickness at each welland kriging the thickness the other is to krig the top andbottom surfaces and then calculate the difference be-tween them The choice of method is based on the dif-ference in error between the two approaches

The kriging of the top and bottom of the sandstonecontacts was not problematic with strong variograms

Table 2 Results from Variogram Modeling

Data

Detrended

Kriging

Lag

Partial Sill

Nugget

Range

V

StandardDeviation Mean

enteris and Carter

RMSE

Depth top Medina

Yes

OK

2500

13146

1808

22807

7837

5232


Yes

OK

5000

18528

4705

59266

7350

282


Yes

OK

2500

13943

1416

23428

77595

4986


Yes

OK

5000

16579

2695

59266

7395

2517


Yes

OK

5000

16643

3824

59266

7422

2473


No

SK

20000

045

052

174285

403

325

ISOCSWHIRL (ft)

No

SK

5000

045

030

59266

398

323

MPWHIRL ()

No

SK

5000

038

050

59266

218

217

PFTWHIRL (ft)

No

SK

5000

064

030

41977

032

029

PVWHIRL (ft3)

No

SK

10000

068

024

118533

032

026


Yes

OK

2500

13716

1811

23521

7762

5178


Yes

OK

5000

18136

4527

59266

7330

2686


Yes

OK

5000

16915

5851

59266

7431

295


No

SK

20000

039

062

59266

220

1906191

ISOCSGRIM (ft)

No

SK

20000

018

075

160975

216

21772212

MPGRIM ()

No

SK

20000

032

070

237065

216

213231

PFTGRIM (ft)

No

SK

20000

017

081

99010

130

135154

PVGRIM (ft3)

No

SK

10000

003

094

No

126

125136

ISOCSWHIRL ISOCSGRIM = thickness clean sand (exceeding 60 sand) for Whirlpool Sandstone and Grimsby Formation respectively MPWHIRL MPGRIM = averageporosity in the clean sand portion PFTWHIRL PFTGRIM = feet of porosity in the clean sand exceeding 6 cutoff PVWHIRL PVGRIM = pore volume RMSE = root meansquared error OK = ordinary kriging SK = simple kriging

Denotes cross validation results for IDW instead of kriging because a reliable variogram was not obtained for the variable in question

225

and small RMSE cross validation errors of around 25 ft(8 m) (only data points with both top and bottom pickswere used in this part of the analysis) By contrast thekriging models of thickness data showed relatively weakspatial structure (Table 2) Oneway to assess the strengthof the spatial structure is the ratio of the partial sill (part


of the semivariance with spatial structure) to the nuggeteffect (part without spatial structure because of spatialvariation below the sampling distance andor measure-ment error) For a good predictive model (elevation topMedina) this ratio is 36 but for the thickness data it isless than 10 Another measure of the strength of the

Figure 10 Variogram model(top) and cross validation results(bottom) from ArcGIS Geosta-tistical Analyst (ESRI 2009) forthe elevation of the top of theMedina Group in northwesternPennsylvania This figure is illus-trative of data with strong spatialstructure showing a distinctvariogram and small RMSE fromcross validation


Figure 11 Thickness of overburden on top of the Medina Group calculated from the difference in the US Geological Survey 984-ft(30-m) resolution DEM and the structure map drawn on the top of the Medina Group (Figure 6)


model is the comparison of the standard deviation ofthe mean compared to the RMSE obtained from crossvalidation For the isopach models these statistics arenearly equal indicating that the spatial model is addingonly a small amount of additional information com-pared to using themean thickness to predict unsampledlocations Also note that the issue is not a problemwithgeostatistical interpolation or with the attendant as-sumptions (stationarity) cross validation results frominverse distance weighting (IDW for the Grimsby For-mation Table 2) gave nearly identical cross validationresults

The choice of approach to calculate the isopach ismade on the basis of the relative error between the op-tions For the kriged thickness data the error is estimateddirectly from the cross validation results When calcu-lating the isopach from the difference by the two surfacemethods the errors propagate by

s2iso frac14 s2top thorn s2bot 2covethtop botTHORN (2)

where sx2 is the variance of the errors (RMSE for top

and bottom) and cov(top bot) is the covariance be-tween the errors for the top and bottom surfaces Forthe Whirlpool Sandstone the cross validation errors(calculated for data points where both bottom and topsurfaces are present) are as follows stop

2 is 6336 ft2

(589 m2) and sbot2 is 6013 ft2 (559 m2) The cross

validation errors between the surfaces are highly posi-tively correlated (Pearson coefficient r = 094) with acovariance of 58104 ft2 (54 m2) By equation 2 theerror (square root of siso

2 ) in determining the isopachfrom the difference of the top and bottom surfaces is85 ft (26 m) In this case the error from the two-layermethod is greater than the cross validation error fromsimple kriging (331 ft [101 m]) In contrast the errorfor the Grimsby Formation isopach from equation 2 is2003 ft (611 m) which is not meaningfully differentthan the cross validation error (1906 ft [581 m]) Ac-cordingly isopachs presented in this work (Figures 7 8)are based on simple kriging models of thickness (pre-sented as averages from SGSIM runs) Kriging of thick-ness data is favored over the two-layer method whenthe covariance (equation 2) is low (or negative) and thelayer is thin (assuming as in this case a positive corre-lation between the magnitude of thickness and crossvalidation error) Such calculations could also be basedon error estimates from SGSIM which is a subject forfuture investigation


Spatial Modeling of Petrophysical Parameters

We attempted to calculate spatial models for pore vol-ume and its individual components (thickness of cleansand average porosity within clean sand) and porosityfeet (a metric popular in industry) for the Grimsby For-mation and the Whirlpool Sandstone (Figure 12) Aswith the structure and isopach modeling we adjusteda wide range of variogram calculation parameters in anattempt to find models of spatial structure To be thor-ough experimental variography was conducted in bothGA and Stanford Geostatistical Modeling Software(SGeMS Remy et al 2009) because these softwaresoffer subtle differences in the variography approachMost of these parameters exhibited a very small partialsill and large nugget effect (Table 2) The kriging modeltherefore provided little to no spatially predictive infor-mation for these parameters As further verification theparameters for the Grimsby Formation were also mod-eled using IDW in GA which depends only on the rela-tion of local data values to one another instead of aglobally applicable model of spatial structure as krigingdoes In each case cross validation results from IDWdidnot improve the model over using the mean of the datato predict at unsampled locations (Table 2) As a finalcheck potential correlations between the isopachs andthe measures of pore space were sought for use in map-ping but a predictive relationship was not found Withthe exception of pore volume for the Whirlpool Sand-stone (explored further below) the current data set isinsufficient to make reliable spatial predictions of volu-metric variables

Themodel of pore volume for theWhirlpool Sand-stone warranted further investigation because a clearvariogram was observed and the comparison betweenthe standard deviation of the mean and the RMSE fromcross validation (Table 2) showed at least some potentialpredictive value At question was the value of a ratherweak predictive model for planning purposes To ad-dress this issue 100 conditional SGSIM runs were cal-culated (1000-ft [305-m] grid resolution) as weresummary statistics for each grid cell The average of100 runs (the E type in geostatistical literature Deutschand Journel 1998) is displayed in Figure 13 The modelshows distinct areas of low (less than 05 ft3 [0014m3])porosity volume in northwest Crawford and westernErie counties as well as the central part of VenangoCounty Areas of relatively large porosity volume(gt10 ft3 [0028 m3]) exist in central Crawford and eastcentral Mercer counties At question is the reliability


Figure 12 Maps showing the point values of parameters dealing with the estimation of pore volume for the Grimsby Formation andWhirlpool Sandstone (A) MPWHIRL (B) PVWHIRL (C) MPGRIM and (D) PVGRIM


Figure 13 Results from 100 SGSIM runs showing the average porosity volume (ft3) and the standard deviation (inset) for the WhirlpoolSandstone


Figure 14 Map of porosity volume (ft3) for the Grimsby Formation drawn using IDW The interpolated values around the markedlocation (number 1) appear to predict an area of elevated pore volume


of prediction for unsampled locations in these areasThe map of the standard deviation of the model runsshows that uncertainty increases with pore volume andto a lesser extent as the data become sparser Low-porosity areas do not significantly exceed zero (2 sigmaor 95 confidence level) whereas the areas of largerporosity volume exceed zero at a minimum of 04 to08 ft3 (0011 to 0023 m3) Despite the noise presentin the model it provides useful information on the spa-tial distribution of pore volume within the WhirlpoolSandstone

DISCUSSION

This case study shows the difficulty in mapping reser-voir parameters with a typical amount of data availablefrom existing public sources available to state geologicalsurveys in theAppalachianBasin For theMedinaGroupthis study and others (Wickstrom et al 2005 Venteriset al 2008) demonstrate that the data are sufficient toreliably predict reservoir architecture (especially in areaslacking intensive faulting) The spatial prediction of vol-umetric parameters however is not always possibleThe key issue is the interaction between the scale(s) ofspatial variability and the distances between well dataThe prediction of the elevation of stratigraphic tops istypically more accurate because there are many moreformation top and bottom picks than porosity measure-ments This is because formation contacts can be derivedfromdetaileddrillersrsquo logs andor gamma-ray logswhichare collected for most wells bymost operators howeverporosity measurements require that the rarer porositylog be obtained Accordingly the median distance tothe closest well is 1643 ft (501 m) for the top of the Me-dina but for the porosity volumemeasurements theme-dian distance is 7934 ft (2419 m) The magnitude andscale of spatial variation are also important For examplethe quality of prediction for the two isopach models isless than the quality of the structure models despitebeing based on similar amounts of data Possible reasonsinclude the decreased range in thickness values com-pared to the elevation of stratigraphic tops (signal tonoise) and possiblymore spatial variation below the sam-pling distance because of sedimentary processes Simi-larly our geostatistical analysis for the Grimsby Forma-tion shows that the spatial scale of variability for porespace occurs at distances below the sample spacing orthatmeasurement errors are too large to detect an exist-


ing structure Castle and Byrnes (1998) investigated thecalibration between porosity measured in lab samplesand from logs and showed strong correlation thus largemeasurement errors are not expected In contrast spa-tial variation occurred at a broad enough scale in theWhirlpool Sandstone to support meaningful mappingwith the samedata spacing The variety inmodel qualityillustrates the need for best statistical and geostatisticalpractices to ensure that spatial predictions are meaning-ful and potential spatial structures are not missed Carein choosing geologic modeling units is also importantbecause initial pore volume modeling in this contribu-tionwas based on theMedinaGroup as awholewithoutsuccess Dividing the model intoWhirlpool and Grims-by components resulted in the exposure of spatial pat-terns in pore volume for the Whirlpool Sandstone

Furthermore these results illustrate thepotential dan-gers in using black-box mapping algorithms or hand con-touringGenerating amap of pore volume for theGrims-by Formation is possible (Figure 14) but the map givesunsubstantiated and potentially misleading informationas cross validation results (Table 2) cast doubt on the va-lidity of the spatial trends implied by the map Figure 14features a map drawn using IDW At location 1 a wellwith a large PV value (73 ft3) is observed The IDWmodel shows a circular area of elevated PV surroundingthis point suggesting that this area is one of the mostfavorable locations for a project The spatial model itis based on is unreliable however and no other infor-mation on the spatial distribution of geologic processescontrolling thickness and porosity exists Therefore thelateral extent of the increased porosity volume surround-ing location 1 is unknown and the map is misleading

For cases such as the Grimsby Formation the bestspatial predictor of pore volume at unsampled locationsis the sample mean When a reliable spatial model isabsent site locations should be selected in the immedi-ate vicinity of existing wells with porosity measure-ments When choosing sites for targets such as theGrimsby Formation a question remains (how close isclose enough) for which the currently available dataoffer little guidance

CONCLUSIONS

Site selection for carbon sequestration projects is a com-plex process involving a wide range of geologic logisti-cal and political issues The current case study and


earliermapping efforts (Wickstromet al 2005Venteriset al 2008) demonstrate the complex range of issuesinvolved in the making of geologic spatial models andmaps to support carbon sequestration planning Thesestudies show that careful consideration of geologicalstatistical and geostatistical issues is needed to addressthe challenges presented by the limitations of the datarelative to the complexity of the sedimentary geologyof the north central Appalachian Basin

REFERENCES CITED

Asquith G B and D Krygowski 2004 Porosity logs in G BAsquith and D Krygowski eds Basic well log analysis 2d edAAPG Methods in Exploration Series 16 p 37ndash76

Brett C E D H Tepper W M Goodman S T LoDuca and BEckert 1995 Revised stratigraphy and correlations of the Nia-garan provincial series (Medina Clinton and Lockport groups)in the type area of western New York US Geological SurveyBulletin 2086 66 p

Carter K M and E R Venteris 2005 A multi-disciplinary ap-proach for mapping rock units with geologic carbon sequestra-tion potential AAPG Eastern Section Meeting MorgantownWest Virginia Abstracts with Programs p 18 httpkarlnrccewvueduesaapgabstracts2005abspdf (accessed July 312009)

Castle J W 1998 Regional sedimentology and stratal surfaces ofa Lower Silurian clastic wedge in the Appalachian forelandbasin Journal of Sedimentary Research v 68 no 6 p 1201ndash1211

Castle J W 2001 Foreland-basin sequence response to collisionaltectonism Geological Society of America Bulletin v 113no 7 p 801ndash812 doi1011300016-7606(2001)113lt0801FBSRTCgt20CO2

Castle JW andA P Byrnes 1998 Petrophysics of low-permeabilityMedina Sandstone northwestern Pennsylvania AppalachianBasin The Log Analyst v 39 no 4 p 36ndash46

Coogan AH 1991 A fault-relatedmodel for the facies of the LowerSilurian Clinton Sandstone interval in the subsurface of easternOhio Northeastern Geology v 13 p 110ndash129

Cotter E 1982 Tuscarora Formation of Pennsylvania SEPM East-ern Section Field Trip Guidebook 105 p

Cotter E 1983 Shelf paralic and fluvial environments and eustaticsea-level fluctuations in the origin of the Tuscarora Formation(Lower Silurian) of central Pennsylvania Journal of Sedimen-tary Research v 53 no 1 p 25ndash49

Deutsch C V 2002 Geostatistical reservoir modeling OxfordOxford University Press 384 p

Deutsch C V and A G Journel 1998 GSLIB Geostatistical soft-ware library and userrsquos guide Oxford Oxford University Press369 p

Environmental SystemsResearch Institute Inc (ESRI) 2009ArcGISand Geostatistical Analyst software httpwwwesricom (ac-cessed July 31 2009)

Hettinger R D 2001 Subsurface correlations and sequence strati-graphic interpretations of Lower Silurian strata in the Appala-chian Basin of northeast Ohio southwest New York andnorthwest Pennsylvania US Geological Survey Geologic In-vestigations Series I-274 22 p

Isaaks E 1990 The application of Monte Carlo methods to the

analysis of spatially correlated data PhD thesis Stanford Uni-versity Stanford California 213 p

Kelley D R 1966 The Kastle Medina gas field Crawford CountyinW S Lytle J H Goth Jr D R Kelley W GMcGlade andW R Wagner eds Oil and gas developments in Pennsylvaniain 1965 Pennsylvania Geological Survey 4th ser Progress Re-port 172 p 30ndash44

Kelley D R and W G McGlade 1969 Medina and Oriskany pro-duction along the shore of Lake Erie Pierce field Erie CountyPennsylvania Pennsylvania Geological Survey 4th ser Min-eral Resource Report 60 38 p

Keltch B W D A Wilson and P E Potter 1990 Deltaic deposi-tional controls on Clinton Sandstone reservoirs Senecaville gasfield Guernsey County Ohio in J H Barwis J R J Studlickand J G McPherson eds Sandstone petroleum reservoirsNew York Springer-Verlag p 263ndash280

Knight W V 1969 Historical and economic geology of LowerSilurian Clinton Sandstone of northeastern Ohio AAPG Bul-letin v 53 no 7 p 1421ndash1452

Laughrey C D 1984 Petrology and reservoir characteristics ofthe Lower Silurian Medina Group sandstones Athens andGeneva fields Crawford County Pennsylvania PennsylvaniaGeological Survey 4th ser Mineral Resource Report 85126 p

Laughrey C D and J A Harper 1986 Comparisons of Upper De-vonian and Lower Silurian tight formation in PennsylvaniamdashGeological and engineering characteristics in C W Spencerand R F Mast eds Geology of tight gas reservoirs AAPGStudies in Geology 24 p 9ndash43

Lytle W S A S Cate V M Fairall L A Heeren and W RWagner 1961 A summary of oil and gas developments in Penn-sylvania 1955 to 1959 Pennsylvania Geological Survey 4thser Mineral Resource Report 45 133 p

Martini I P 1971 Regional analysis of sedimentology of MedinaFormation (Silurian) Ontario and New York AAPG Bulletinv 55 no 8 p 1249ndash1261

McCormac M P G O Mychkovsky S T Opritza R A RileyM E Wolfe G E Larsen and M T Baranoski 1996 PlayScmmdashLower Silurian cataractMedina Group (ldquoClintonrdquo) sand-stone play in J B Roen and B JWalker eds The atlas of majorAppalachian gas plays West Virginia Geological and EconomicSurvey Publication 25 p 156ndash163

Pees S T 1987 Mapping Medina (Clinton) play by lumping cleansandstones Oil amp Gas Journal v 85 no 10 p 42ndash45

Piotrowski R G 1981 Geology and natural gas production of theLower Silurian Medina Group and equivalent rock units inPennsylvania Pennsylvania Geological Survey 4th ser Min-eral Resource Report 82 21 p

Remy N A Boucher and J Wu 2009 Applied geostatistics withSGeMSmdashA userrsquos guide New York Cambridge UniversityPress 288 p

Ryder R T 2004 Stratigraphic framework and depositional se-quences in the Lower Silurian regional oil and gas accumula-tion Appalachian BasinmdashFrom Ashland County Ohiothrough southwestern Pennsylvania to Preston County WestVirginia US Geological Survey Geologic Investigations SeriesMap I-2810 2 sheets

SAS Institute Inc 2008 JMP 70 httpwwwjmpcom (accessedJuly 31 2009)

Venteris E R and B K Slater 2005 A comparison between con-tour elevation data sources for DEM creation and soil carbonprediction Coshocton Ohio Transactions in GIS v 9 no 2p 179ndash195 doi101111j1467-9671200500212x

Venteris E R R R Riley J McDonald L H Wickstrom J ADrahovzal and J A Harper 2008 Establishing a regional geo-logic framework for carbon dioxide sequestrationmdashA case


study inM Grobe J C Pashin and R L Dodge eds Carbondioxide sequestration in geological mediamdashState of the scienceAAPG Studies 59 p 1ndash35

Wells Information System (WIS) 2009 Petroleum productiondata retrieved from database of the Pennsylvania GeologicalSurvey httpwwwpairisstatepaus (accessed January 292009)

Wickstrom L H et al 2005 Characterization of geologic se-


questration opportunities in the MRCSP region phase I taskreport period of performance October 2003ndashSeptember 2005DOE Cooperative Agreement No DE-PS26-05NT42255152 p

Yeakel L S 1962 Tuscarora Juniata and Bald Eagle paleocurrentsand paleogeography in the central Appalachians Geological So-ciety of America Bulletin v 73 no 12 p 1515ndash1540doi1011300016-7606(1962)73[1515TJABEP]20CO2


different storage mechanisms The commonly discussedstorage types include volumetric solution adsorptionand mineral Volumetric storage refers to the amountof CO2 that is retained in the pore space of a geologicunit generally as a supercritical phase retained by struc-tural or stratigraphic traps or by overlying cap rock(Wickstrom et al 2005) Mapping the location and es-timating the total capacity of this storage type for sev-eral geologic units in the north central AppalachianBasin have been the focus of our research

As in oil-and-gas exploration reservoir characteri-zation is a key task in planning CO2 injection projectsSpatial models and maps are needed to ensure (1) suf-ficient overburden to maintain CO2 in a supercriticalstate (approximate depth of 2500 ft [762m]) (2) largeenough porosity and permeability to support a practicalinjection rate and (3) sufficient total CO2 capacity toensure the economic viability of the project In addi-tion selection and evaluation of potential injection sitesrequire a detailed understanding of the overlying andunderlying formation(s) that would serve as seals (a topicbeyond the scope of this contribution) Accordinglyrock characteristics critical to characterizing a potentialsequestration reservoir include depth thickness poros-ity water saturation permeability and the lateral andvertical heterogeneity of these parameters Petrophysi-cal analyses of geophysical well logs and core samplesare used to determine these properties at individualboreholes and the spatial models are used to estimatevalues for these throughout the area of interest

The selection of sites for carbon sequestration proj-ects includes a wide range of geologic and nongeologicfactors (Wickstrom et al 2005 Venteris et al 2008)Because of the complexity of site selection and planningpredicting the sequestration capacity at unsampled lo-cations is advantageous Maps and other spatial modelsof reservoir characteristics are the most convenientmethods available to provide such information Geosta-tistical reservoir characterization methods (Deutsch2002) provide an objective optimized and scientificallydefensible method for estimating reservoir characteris-tics at unsampled locations providing measures of theaccuracy of such estimates and facilitating the genera-tion of continuous models (rasters) for use in regionalcapacity calculations (Venteris et al 2008)

As part of our early research efforts we used krigingto map structure elevation data and thickness for theMedina GroupldquoClintonrdquo Sandstone in PennsylvaniaWest Virginia and Ohio The Medina was chosen formore detailed study because a large number of borings


are available in northwestern Pennsylvania and our pre-viousmapping effort showed relatively high accuracy instructure and isopach mapping relative to other poten-tial reservoirs (Carter and Venteris 2005 Wickstromet al 2005) The lack of prominent faulting in the for-mation also simplified the application of geostatisticalmethods In this study conducted at a higher resolutionwith more data than the previous works we seek morefully to characterize the reservoir through refinementof the structure and isopachmodels and through the ad-dition of spatial models of the available pore space

Our case study uses geostatistical methods (krigingand stochastic simulation) to quantify the strength ofspatial prediction for a range of reservoir parametersbased on a typical public oil-and-gas data set for the Ap-palachian Basin Amajor goal of this study was to inves-tigate the usefulness of the available public data for eachof the various reservoir parameters to identify knowl-edge gaps and design future data collection efforts Thiscontribution illustrates the dangers of mapping withblack box approaches typical of some reservoir charac-terization software that do not encourage or allow themodeler to explore the data for outliers investigate andquantify the spatial structure of the data for use in pre-diction and evaluate the error of the predictions Amap whether drawn by computer algorithm or handcontouring implies that the person creating the mapscan predict values at unsampled locations The qualityof that prediction must be evaluated For accurate spa-tial prediction (1) the error in the measurements mustbe small compared to the magnitude of spatial variabil-ity (2) that spatial variability must be at least partiallyordered in space (not random) and (3) the distance be-tween the samples (wells) must be less than the range ofthe spatial structure If these conditions are not metpredicted values beyond the data points are highly sus-pect and the map is of little value This work seeks toillustrate these spatial modeling issues using various res-ervoir parameters of the Medina Group

STUDY AREA

The current study area encompasses twomajor oil-and-gas fields in northwestern Pennsylvania and incorpo-rates publicly available geologic data from more than5600 deep wells that penetrate the Early SilurianndashageMedina Group throughout Erie Crawford Mercerand Venango counties (Figure 1) In general oil and gasfields are of interest in sequestration projects because





Oil and Gas Fields












V








METHODOLOGY





Venterisand

Carter217





Planning













(1)










Data

N

Min

Median

Max

Average

Venteris and C

Standard Deviation

Depth top Medina

5598

2082

4593

6388

4542

7837


5598

minus4933

minus3315

minus1512

minus3237

7350


5295

2226

4784

6567

4731

77595


5297

minus5137

minus3508

minus1656

minus3420

7389


5375

minus5153

minus3508

minus1668

3424

7422


5278

1

11

47

1165

403

ISOCSWHIRL (ft)

329

1

85

23

848

398

MPWHIRL ()

160

152

554

2013

574

218

PFTWHIRL (ft)

122

003

033

169

041

032

PVWHIRL (ft3)

160

004

047

179

047

032


5375

2082

4597

6388

4549

7762


5375

minus4922

minus3320

minus1512

minus3242

7330


3810

minus5097

minus3724

minus1623

minus3530

7431


3798

5

120

195

1181

220

ISOCSGRIM (ft)

243

3

418

1285

452

216

MPGRIM ()

155

149

625

2278

642

216

PFTGRIM (ft)

152

007

169

697

194

130

PVGRIM (ft3)

155

013

250

732

278

126


arter 223






RESULTS







Isopach





Data

Detrended

Kriging

Lag

Partial Sill

Nugget

Range

V


enteris and Carter

RMSE

Depth top Medina

Yes

OK

2500

13146

1808

22807

7837

5232


Yes

OK

5000

18528

4705

59266

7350

282


Yes

OK

2500

13943

1416

23428

77595

4986


Yes

OK

5000

16579

2695

59266

7395

2517


Yes

OK

5000

16643

3824

59266

7422

2473


No

SK

20000

045

052

174285

403

325

ISOCSWHIRL (ft)

No

SK

5000

045

030

59266

398

323

MPWHIRL ()

No

SK

5000

038

050

59266

218

217

PFTWHIRL (ft)

No

SK

5000

064

030

41977

032

029

PVWHIRL (ft3)

No

SK

10000

068

024

118533

032

026


Yes

OK

2500

13716

1811

23521

7762

5178


Yes

OK

5000

18136

4527

59266

7330

2686


Yes

OK

5000

16915

5851

59266

7431

295


No

SK

20000

039

062

59266

220

1906191

ISOCSGRIM (ft)

No

SK

20000

018

075

160975

216

21772212

MPGRIM ()

No

SK

20000

032

070

237065

216

213231

PFTGRIM (ft)

No

SK

20000

017

081

99010

130

135154

PVGRIM (ft3)

No

SK

10000

003

094

No

126

125136



225













2 is 6336 ft2
















DISCUSSION






CONCLUSIONS




REFERENCES CITED











































Oil and Gas Fields












V








METHODOLOGY





Venterisand

Carter217





Planning













(1)










Data

N

Min

Median

Max

Average

Venteris and C

Standard Deviation

Depth top Medina

5598

2082

4593

6388

4542

7837


5598

minus4933

minus3315

minus1512

minus3237

7350


5295

2226

4784

6567

4731

77595


5297

minus5137

minus3508

minus1656

minus3420

7389


5375

minus5153

minus3508

minus1668

3424

7422


5278

1

11

47

1165

403

ISOCSWHIRL (ft)

329

1

85

23

848

398

MPWHIRL ()

160

152

554

2013

574

218

PFTWHIRL (ft)

122

003

033

169

041

032

PVWHIRL (ft3)

160

004

047

179

047

032


5375

2082

4597

6388

4549

7762


5375

minus4922

minus3320

minus1512

minus3242

7330


3810

minus5097

minus3724

minus1623

minus3530

7431


3798

5

120

195

1181

220

ISOCSGRIM (ft)

243

3

418

1285

452

216

MPGRIM ()

155

149

625

2278

642

216

PFTGRIM (ft)

152

007

169

697

194

130

PVGRIM (ft3)

155

013

250

732

278

126


arter 223






RESULTS







Isopach





Data

Detrended

Kriging

Lag

Partial Sill

Nugget

Range

V


enteris and Carter

RMSE

Depth top Medina

Yes

OK

2500

13146

1808

22807

7837

5232


Yes

OK

5000

18528

4705

59266

7350

282


Yes

OK

2500

13943

1416

23428

77595

4986


Yes

OK

5000

16579

2695

59266

7395

2517


Yes

OK

5000

16643

3824

59266

7422

2473


No

SK

20000

045

052

174285

403

325

ISOCSWHIRL (ft)

No

SK

5000

045

030

59266

398

323

MPWHIRL ()

No

SK

5000

038

050

59266

218

217

PFTWHIRL (ft)

No

SK

5000

064

030

41977

032

029

PVWHIRL (ft3)

No

SK

10000

068

024

118533

032

026


Yes

OK

2500

13716

1811

23521

7762

5178


Yes

OK

5000

18136

4527

59266

7330

2686


Yes

OK

5000

16915

5851

59266

7431

295


No

SK

20000

039

062

59266

220

1906191

ISOCSGRIM (ft)

No

SK

20000

018

075

160975

216

21772212

MPGRIM ()

No

SK

20000

032

070

237065

216

213231

PFTGRIM (ft)

No

SK

20000

017

081

99010

130

135154

PVGRIM (ft3)

No

SK

10000

003

094

No

126

125136



225













2 is 6336 ft2
















DISCUSSION






CONCLUSIONS




REFERENCES CITED













































V








METHODOLOGY





Venterisand

Carter217





Planning













(1)










Data

N

Min

Median

Max

Average

Venteris and C

Standard Deviation

Depth top Medina

5598

2082

4593

6388

4542

7837


5598

minus4933

minus3315

minus1512

minus3237

7350


5295

2226

4784

6567

4731

77595


5297

minus5137

minus3508

minus1656

minus3420

7389


5375

minus5153

minus3508

minus1668

3424

7422


5278

1

11

47

1165

403

ISOCSWHIRL (ft)

329

1

85

23

848

398

MPWHIRL ()

160

152

554

2013

574

218

PFTWHIRL (ft)

122

003

033

169

041

032

PVWHIRL (ft3)

160

004

047

179

047

032


5375

2082

4597

6388

4549

7762


5375

minus4922

minus3320

minus1512

minus3242

7330


3810

minus5097

minus3724

minus1623

minus3530

7431


3798

5

120

195

1181

220

ISOCSGRIM (ft)

243

3

418

1285

452

216

MPGRIM ()

155

149

625

2278

642

216

PFTGRIM (ft)

152

007

169

697

194

130

PVGRIM (ft3)

155

013

250

732

278

126


arter 223






RESULTS







Isopach





Data

Detrended

Kriging

Lag

Partial Sill

Nugget

Range

V


enteris and Carter

RMSE

Depth top Medina

Yes

OK

2500

13146

1808

22807

7837

5232


Yes

OK

5000

18528

4705

59266

7350

282


Yes

OK

2500

13943

1416

23428

77595

4986


Yes

OK

5000

16579

2695

59266

7395

2517


Yes

OK

5000

16643

3824

59266

7422

2473


No

SK

20000

045

052

174285

403

325

ISOCSWHIRL (ft)

No

SK

5000

045

030

59266

398

323

MPWHIRL ()

No

SK

5000

038

050

59266

218

217

PFTWHIRL (ft)

No

SK

5000

064

030

41977

032

029

PVWHIRL (ft3)

No

SK

10000

068

024

118533

032

026


Yes

OK

2500

13716

1811

23521

7762

5178


Yes

OK

5000

18136

4527

59266

7330

2686


Yes

OK

5000

16915

5851

59266

7431

295


No

SK

20000

039

062

59266

220

1906191

ISOCSGRIM (ft)

No

SK

20000

018

075

160975

216

21772212

MPGRIM ()

No

SK

20000

032

070

237065

216

213231

PFTGRIM (ft)

No

SK

20000

017

081

99010

130

135154

PVGRIM (ft3)

No

SK

10000

003

094

No

126

125136



225













2 is 6336 ft2
















DISCUSSION






CONCLUSIONS




REFERENCES CITED














































METHODOLOGY





Venterisand

Carter217





Planning













(1)










Data

N

Min

Median

Max

Average

Venteris and C

Standard Deviation

Depth top Medina

5598

2082

4593

6388

4542

7837


5598

minus4933

minus3315

minus1512

minus3237

7350


5295

2226

4784

6567

4731

77595


5297

minus5137

minus3508

minus1656

minus3420

7389


5375

minus5153

minus3508

minus1668

3424

7422


5278

1

11

47

1165

403

ISOCSWHIRL (ft)

329

1

85

23

848

398

MPWHIRL ()

160

152

554

2013

574

218

PFTWHIRL (ft)

122

003

033

169

041

032

PVWHIRL (ft3)

160

004

047

179

047

032


5375

2082

4597

6388

4549

7762


5375

minus4922

minus3320

minus1512

minus3242

7330


3810

minus5097

minus3724

minus1623

minus3530

7431


3798

5

120

195

1181

220

ISOCSGRIM (ft)

243

3

418

1285

452

216

MPGRIM ()

155

149

625

2278

642

216

PFTGRIM (ft)

152

007

169

697

194

130

PVGRIM (ft3)

155

013

250

732

278

126


arter 223






RESULTS







Isopach





Data

Detrended

Kriging

Lag

Partial Sill

Nugget

Range

V


enteris and Carter

RMSE

Depth top Medina

Yes

OK

2500

13146

1808

22807

7837

5232


Yes

OK

5000

18528

4705

59266

7350

282


Yes

OK

2500

13943

1416

23428

77595

4986


Yes

OK

5000

16579

2695

59266

7395

2517


Yes

OK

5000

16643

3824

59266

7422

2473


No

SK

20000

045

052

174285

403

325

ISOCSWHIRL (ft)

No

SK

5000

045

030

59266

398

323

MPWHIRL ()

No

SK

5000

038

050

59266

218

217

PFTWHIRL (ft)

No

SK

5000

064

030

41977

032

029

PVWHIRL (ft3)

No

SK

10000

068

024

118533

032

026


Yes

OK

2500

13716

1811

23521

7762

5178


Yes

OK

5000

18136

4527

59266

7330

2686


Yes

OK

5000

16915

5851

59266

7431

295


No

SK

20000

039

062

59266

220

1906191

ISOCSGRIM (ft)

No

SK

20000

018

075

160975

216

21772212

MPGRIM ()

No

SK

20000

032

070

237065

216

213231

PFTGRIM (ft)

No

SK

20000

017

081

99010

130

135154

PVGRIM (ft3)

No

SK

10000

003

094

No

126

125136



225













2 is 6336 ft2
















DISCUSSION






CONCLUSIONS




REFERENCES CITED









































Venterisand

Carter217





Planning













(1)










Data

N

Min

Median

Max

Average

Venteris and C

Standard Deviation

Depth top Medina

5598

2082

4593

6388

4542

7837


5598

minus4933

minus3315

minus1512

minus3237

7350


5295

2226

4784

6567

4731

77595


5297

minus5137

minus3508

minus1656

minus3420

7389


5375

minus5153

minus3508

minus1668

3424

7422


5278

1

11

47

1165

403

ISOCSWHIRL (ft)

329

1

85

23

848

398

MPWHIRL ()

160

152

554

2013

574

218

PFTWHIRL (ft)

122

003

033

169

041

032

PVWHIRL (ft3)

160

004

047

179

047

032


5375

2082

4597

6388

4549

7762


5375

minus4922

minus3320

minus1512

minus3242

7330


3810

minus5097

minus3724

minus1623

minus3530

7431


3798

5

120

195

1181

220

ISOCSGRIM (ft)

243

3

418

1285

452

216

MPGRIM ()

155

149

625

2278

642

216

PFTGRIM (ft)

152

007

169

697

194

130

PVGRIM (ft3)

155

013

250

732

278

126


arter 223






RESULTS







Isopach





Data

Detrended

Kriging

Lag

Partial Sill

Nugget

Range

V


enteris and Carter

RMSE

Depth top Medina

Yes

OK

2500

13146

1808

22807

7837

5232


Yes

OK

5000

18528

4705

59266

7350

282


Yes

OK

2500

13943

1416

23428

77595

4986


Yes

OK

5000

16579

2695

59266

7395

2517


Yes

OK

5000

16643

3824

59266

7422

2473


No

SK

20000

045

052

174285

403

325

ISOCSWHIRL (ft)

No

SK

5000

045

030

59266

398

323

MPWHIRL ()

No

SK

5000

038

050

59266

218

217

PFTWHIRL (ft)

No

SK

5000

064

030

41977

032

029

PVWHIRL (ft3)

No

SK

10000

068

024

118533

032

026


Yes

OK

2500

13716

1811

23521

7762

5178


Yes

OK

5000

18136

4527

59266

7330

2686


Yes

OK

5000

16915

5851

59266

7431

295


No

SK

20000

039

062

59266

220

1906191

ISOCSGRIM (ft)

No

SK

20000

018

075

160975

216

21772212

MPGRIM ()

No

SK

20000

032

070

237065

216

213231

PFTGRIM (ft)

No

SK

20000

017

081

99010

130

135154

PVGRIM (ft3)

No

SK

10000

003

094

No

126

125136



225













2 is 6336 ft2
















DISCUSSION






CONCLUSIONS




REFERENCES CITED












































Planning













(1)










Data

N

Min

Median

Max

Average

Venteris and C

Standard Deviation

Depth top Medina

5598

2082

4593

6388

4542

7837


5598

minus4933

minus3315

minus1512

minus3237

7350


5295

2226

4784

6567

4731

77595


5297

minus5137

minus3508

minus1656

minus3420

7389


5375

minus5153

minus3508

minus1668

3424

7422


5278

1

11

47

1165

403

ISOCSWHIRL (ft)

329

1

85

23

848

398

MPWHIRL ()

160

152

554

2013

574

218

PFTWHIRL (ft)

122

003

033

169

041

032

PVWHIRL (ft3)

160

004

047

179

047

032


5375

2082

4597

6388

4549

7762


5375

minus4922

minus3320

minus1512

minus3242

7330


3810

minus5097

minus3724

minus1623

minus3530

7431


3798

5

120

195

1181

220

ISOCSGRIM (ft)

243

3

418

1285

452

216

MPGRIM ()

155

149

625

2278

642

216

PFTGRIM (ft)

152

007

169

697

194

130

PVGRIM (ft3)

155

013

250

732

278

126


arter 223






RESULTS







Isopach





Data

Detrended

Kriging

Lag

Partial Sill

Nugget

Range

V


enteris and Carter

RMSE

Depth top Medina

Yes

OK

2500

13146

1808

22807

7837

5232


Yes

OK

5000

18528

4705

59266

7350

282


Yes

OK

2500

13943

1416

23428

77595

4986


Yes

OK

5000

16579

2695

59266

7395

2517


Yes

OK

5000

16643

3824

59266

7422

2473


No

SK

20000

045

052

174285

403

325

ISOCSWHIRL (ft)

No

SK

5000

045

030

59266

398

323

MPWHIRL ()

No

SK

5000

038

050

59266

218

217

PFTWHIRL (ft)

No

SK

5000

064

030

41977

032

029

PVWHIRL (ft3)

No

SK

10000

068

024

118533

032

026


Yes

OK

2500

13716

1811

23521

7762

5178


Yes

OK

5000

18136

4527

59266

7330

2686


Yes

OK

5000

16915

5851

59266

7431

295


No

SK

20000

039

062

59266

220

1906191

ISOCSGRIM (ft)

No

SK

20000

018

075

160975

216

21772212

MPGRIM ()

No

SK

20000

032

070

237065

216

213231

PFTGRIM (ft)

No

SK

20000

017

081

99010

130

135154

PVGRIM (ft3)

No

SK

10000

003

094

No

126

125136



225













2 is 6336 ft2
















DISCUSSION






CONCLUSIONS




REFERENCES CITED




















































(1)










Data

N

Min

Median

Max

Average

Venteris and C

Standard Deviation

Depth top Medina

5598

2082

4593

6388

4542

7837


5598

minus4933

minus3315

minus1512

minus3237

7350


5295

2226

4784

6567

4731

77595


5297

minus5137

minus3508

minus1656

minus3420

7389


5375

minus5153

minus3508

minus1668

3424

7422


5278

1

11

47

1165

403

ISOCSWHIRL (ft)

329

1

85

23

848

398

MPWHIRL ()

160

152

554

2013

574

218

PFTWHIRL (ft)

122

003

033

169

041

032

PVWHIRL (ft3)

160

004

047

179

047

032


5375

2082

4597

6388

4549

7762


5375

minus4922

minus3320

minus1512

minus3242

7330


3810

minus5097

minus3724

minus1623

minus3530

7431


3798

5

120

195

1181

220

ISOCSGRIM (ft)

243

3

418

1285

452

216

MPGRIM ()

155

149

625

2278

642

216

PFTGRIM (ft)

152

007

169

697

194

130

PVGRIM (ft3)

155

013

250

732

278

126


arter 223






RESULTS







Isopach





Data

Detrended

Kriging

Lag

Partial Sill

Nugget

Range

V


enteris and Carter

RMSE

Depth top Medina

Yes

OK

2500

13146

1808

22807

7837

5232


Yes

OK

5000

18528

4705

59266

7350

282


Yes

OK

2500

13943

1416

23428

77595

4986


Yes

OK

5000

16579

2695

59266

7395

2517


Yes

OK

5000

16643

3824

59266

7422

2473


No

SK

20000

045

052

174285

403

325

ISOCSWHIRL (ft)

No

SK

5000

045

030

59266

398

323

MPWHIRL ()

No

SK

5000

038

050

59266

218

217

PFTWHIRL (ft)

No

SK

5000

064

030

41977

032

029

PVWHIRL (ft3)

No

SK

10000

068

024

118533

032

026


Yes

OK

2500

13716

1811

23521

7762

5178


Yes

OK

5000

18136

4527

59266

7330

2686


Yes

OK

5000

16915

5851

59266

7431

295


No

SK

20000

039

062

59266

220

1906191

ISOCSGRIM (ft)

No

SK

20000

018

075

160975

216

21772212

MPGRIM ()

No

SK

20000

032

070

237065

216

213231

PFTGRIM (ft)

No

SK

20000

017

081

99010

130

135154

PVGRIM (ft3)

No

SK

10000

003

094

No

126

125136



225













2 is 6336 ft2
















DISCUSSION






CONCLUSIONS




REFERENCES CITED














































Data

N

Min

Median

Max

Average

Venteris and C

Standard Deviation

Depth top Medina

5598

2082

4593

6388

4542

7837


5598

minus4933

minus3315

minus1512

minus3237

7350


5295

2226

4784

6567

4731

77595


5297

minus5137

minus3508

minus1656

minus3420

7389


5375

minus5153

minus3508

minus1668

3424

7422


5278

1

11

47

1165

403

ISOCSWHIRL (ft)

329

1

85

23

848

398

MPWHIRL ()

160

152

554

2013

574

218

PFTWHIRL (ft)

122

003

033

169

041

032

PVWHIRL (ft3)

160

004

047

179

047

032


5375

2082

4597

6388

4549

7762


5375

minus4922

minus3320

minus1512

minus3242

7330


3810

minus5097

minus3724

minus1623

minus3530

7431


3798

5

120

195

1181

220

ISOCSGRIM (ft)

243

3

418

1285

452

216

MPGRIM ()

155

149

625

2278

642

216

PFTGRIM (ft)

152

007

169

697

194

130

PVGRIM (ft3)

155

013

250

732

278

126


arter 223






RESULTS







Isopach





Data

Detrended

Kriging

Lag

Partial Sill

Nugget

Range

V


enteris and Carter

RMSE

Depth top Medina

Yes

OK

2500

13146

1808

22807

7837

5232


Yes

OK

5000

18528

4705

59266

7350

282


Yes

OK

2500

13943

1416

23428

77595

4986


Yes

OK

5000

16579

2695

59266

7395

2517


Yes

OK

5000

16643

3824

59266

7422

2473


No

SK

20000

045

052

174285

403

325

ISOCSWHIRL (ft)

No

SK

5000

045

030

59266

398

323

MPWHIRL ()

No

SK

5000

038

050

59266

218

217

PFTWHIRL (ft)

No

SK

5000

064

030

41977

032

029

PVWHIRL (ft3)

No

SK

10000

068

024

118533

032

026


Yes

OK

2500

13716

1811

23521

7762

5178


Yes

OK

5000

18136

4527

59266

7330

2686


Yes

OK

5000

16915

5851

59266

7431

295


No

SK

20000

039

062

59266

220

1906191

ISOCSGRIM (ft)

No

SK

20000

018

075

160975

216

21772212

MPGRIM ()

No

SK

20000

032

070

237065

216

213231

PFTGRIM (ft)

No

SK

20000

017

081

99010

130

135154

PVGRIM (ft3)

No

SK

10000

003

094

No

126

125136



225













2 is 6336 ft2
















DISCUSSION






CONCLUSIONS




REFERENCES CITED













































RESULTS







Isopach





Data

Detrended

Kriging

Lag

Partial Sill

Nugget

Range

V


enteris and Carter

RMSE

Depth top Medina

Yes

OK

2500

13146

1808

22807

7837

5232


Yes

OK

5000

18528

4705

59266

7350

282


Yes

OK

2500

13943

1416

23428

77595

4986


Yes

OK

5000

16579

2695

59266

7395

2517


Yes

OK

5000

16643

3824

59266

7422

2473


No

SK

20000

045

052

174285

403

325

ISOCSWHIRL (ft)

No

SK

5000

045

030

59266

398

323

MPWHIRL ()

No

SK

5000

038

050

59266

218

217

PFTWHIRL (ft)

No

SK

5000

064

030

41977

032

029

PVWHIRL (ft3)

No

SK

10000

068

024

118533

032

026


Yes

OK

2500

13716

1811

23521

7762

5178


Yes

OK

5000

18136

4527

59266

7330

2686


Yes

OK

5000

16915

5851

59266

7431

295


No

SK

20000

039

062

59266

220

1906191

ISOCSGRIM (ft)

No

SK

20000

018

075

160975

216

21772212

MPGRIM ()

No

SK

20000

032

070

237065

216

213231

PFTGRIM (ft)

No

SK

20000

017

081

99010

130

135154

PVGRIM (ft3)

No

SK

10000

003

094

No

126

125136



225













2 is 6336 ft2
















DISCUSSION






CONCLUSIONS




REFERENCES CITED










































Isopach





Data

Detrended

Kriging

Lag

Partial Sill

Nugget

Range

V


enteris and Carter

RMSE

Depth top Medina

Yes

OK

2500

13146

1808

22807

7837

5232


Yes

OK

5000

18528

4705

59266

7350

282


Yes

OK

2500

13943

1416

23428

77595

4986


Yes

OK

5000

16579

2695

59266

7395

2517


Yes

OK

5000

16643

3824

59266

7422

2473


No

SK

20000

045

052

174285

403

325

ISOCSWHIRL (ft)

No

SK

5000

045

030

59266

398

323

MPWHIRL ()

No

SK

5000

038

050

59266

218

217

PFTWHIRL (ft)

No

SK

5000

064

030

41977

032

029

PVWHIRL (ft3)

No

SK

10000

068

024

118533

032

026


Yes

OK

2500

13716

1811

23521

7762

5178


Yes

OK

5000

18136

4527

59266

7330

2686


Yes

OK

5000

16915

5851

59266

7431

295


No

SK

20000

039

062

59266

220

1906191

ISOCSGRIM (ft)

No

SK

20000

018

075

160975

216

21772212

MPGRIM ()

No

SK

20000

032

070

237065

216

213231

PFTGRIM (ft)

No

SK

20000

017

081

99010

130

135154

PVGRIM (ft3)

No

SK

10000

003

094

No

126

125136



225













2 is 6336 ft2
















DISCUSSION






CONCLUSIONS




REFERENCES CITED




















































2 is 6336 ft2
















DISCUSSION






CONCLUSIONS




REFERENCES CITED















































DISCUSSION






CONCLUSIONS




REFERENCES CITED









































REFERENCES CITED








































assessing spatial uncertainty in reservoir … › ... › uploaded › eg09008.pdfassessing spatial...

Documents