Process evaluateprospect97 1087

ABSTRACT

In 1989, Chevron Overseas Petroleum, Inc.,developed a process to allow management to com-pare a wide variety of global exploration opportu-nities on a uniform and consistent basis. Over thenext five years, the process evolved into an effec-tive method to plan exploration programs on abasis of value incorporating prospect ranking, bud-get allocation, and technology management. Thefinal product is a continuous process and includes,within a single organizational unit, the integrationof geologic risk assessment, probabilistic distribu-tion of prospect hydrocarbon volumes, engineeringdevelopment planning, and prospect economics.

The process is based on the concepts of the playand hydrocarbon system. Other steps of the pro-cess (geologic risk assessment, volumetric estima-tion, engineering support, economic evaluation,and postdrill feedback) are considered extensionsof fundamental knowledge and understanding ofthe underlying geological, engineering, and fiscalconstraints imposed by these concepts. A founda-tion is set, describing the geologic framework andthe prospect in terms of the play concept—source,reservoir, trap (including seal), and dynamics(timing/migration). The information and data fromthis description become the basis for subsquentsteps in the process. Risk assessment assigns aprobability of success to each of these four ele-ments of the play concept, and multiplication of

these probabilities yields the probability of geolog-ic success. A well is considered a geologic successif a stabilized flow of hydrocarbons is obtained ontest. Volumetric estimation expresses uncertaintyin a distribution of possible hydrocarbon volumesfor the prospect constructed from ranges of param-eters obtained from information specific to theprospect, and data described by the parent playconcept. With this distribution, engineering sup-port provides development scenarios for threecases—a pessimistic case (10%), the mean, and anoptimistic case (90%). Economic evaluation is runfor each of the three cases, thus providing a rangeof economic consequences of the geological, engi-neering, and fiscal framework. Commercial risk isbased on the results of this evaluation, and overallprobability of success is the multiplication of theprobability of geologic success and probability ofcommercial success. Postdrill feedback determineswhether the individual processes are providing pre-dicted results consistent with actual outcomes.

INTRODUCTION

The topic of prospect evaluation has been dis-cussed in the literature for many years and has beenrecently described in a sequence of reviews byRobert Megill in the AAPG Explorer. In recent years,AAPG has encouraged discussions on this subject bysponsoring Hedberg research conferences and con-vention sessions at which we presented parts of theChevron system (Otis and Schneidermann, 1994;Otis, 1995). Many of the conference participantsrequested that we summarize our process in print.This paper is a summary of the exploration evalua-tion process that has been used to provide estimatesof exploration prospect value for the last 7 yr atChevron Overseas Petroleum, Inc. For obvious rea-sons, this summary does not include all of thedetails; however, we hope this paper will stimulatefurther discussions and encourage the release of sim-ilar summaries by other companies.

The foundation of the process is knowledge ofgeology; in particular, the concepts of hydrocarbonsystems and the play concept as developed overthe years by Dow (1972, 1974), Nederlof (1979),Perrodon (1980, 1983, 1992), Demaison (1984),

1087AAPG Bulletin, V. 81, No. 7 (July 1997), P. 1087–1109.

©Copyright 1997. The American Association of Petroleum Geologists. Allrights reserved.

1Manuscript received February 16, 1996; revised manuscript receivedSeptember 26, 1996; final acceptance February 4, 1997.

2Chevron Overseas Petroleum, Inc., P.O. Box 5046, San Ramon,California 94583-0946.

We acknowledge the champion of this process, M. W. Boyce, withoutwhose continuing, senior-management support this process would not havebeen possible. We acknowledge the pioneering efforts of C. L. Aguilera, G. A.Demaison, E. J. Durrer, F. R. Johnson, W. E. Perkins, J. L. Reich, and R. A.Seltzer, who established the framework for the process in its early stages. Wealso acknowledge the efforts to refine, document, and teach the processduring the later stages by S. D. Adams, A. O. Akinpelu, G. A. Ankenbauer, G. L. Bliss, T. J. Humphrey, E. McLean, and D. B. Wallem. Finally, weacknowledge all the people who, over the past several decades, havechampioned such a process, but fell victim to deaf ears because of high oilprices or dumb luck. These people provided the well-founded basis for thetheoretical and practical application of evaluation principles. We also wish toextend special thanks to Gerard Demaison and Erwin Durrer for theircontinuous support, guidance, and friendship.

A Process for Evaluating Exploration Prospects1

Robert M. Otis and Nahum Schneidermann2

Ulmishek (1986), White (1988, 1993), Demaisonand Huizinga (1991), Magoon (1987, 1988, and1989, Magoon and Dow (1994). Ultimately, all esti-mates of value are based on hydrocarbon volumes,geological risk, and reservoir productivity and per-formance, which, in turn, are based on the geologi-cal characteristics of the hydrocarbons present andthe geological nature of the reservoir and trap char-acteristics. The process, therefore, focuses on esti-mating the range of resources that may be possible(what nature has provided), the chances of findinga hydrocarbon accumulation, and the requirementsfor producing the hydrocarbons to add significantvalue at an acceptable rate of return.

The full process, illustrated in Figure 1, beginsby establishing the play concept, described byfour elements: source rock, reservoir, trap (includ-ing seal), and dynamics (timing and migration).Based on this descr iption, geological r isk isassessed, and the probability of finding produciblehydrocarbons is assigned a value between 0.01and 0.99. At the same time, the volume of hydro-carbons present is estimated as a probability distri-bution of recoverable volumes. The engineeringdepartment provides estimates of production pro-files and facilities and transportation costs, whichare then incorporated with a country economicmodel and risk to generate economics that corre-spond to pessimistic, mean, and optimistic esti-mates from the distribution. If a decision is madeto go ahead with the project, results are docu-mented so that predicted and actual outcomes canbe compared, added to the knowledge base, andused for process improvement.

Methods used in the process are not new. Theyare based on pioneering publications by Haun(1975), Newendorp (1975), White (1980, 1988,

1993), Megill (1984), and Rose (1987, 1992), aswell as in-house work by both Chevron (Jones,1975) and Gulf. The ideas of hydrocarbon systemand play concept, as well as descriptive tools, aredescribed fully by Magoon (1987, 1988, 1989),Magoon and Dow (1994), and Demaison andHuizinga (1991). The breakdown of geologic riskinto basic risk factors, preparing production pro-files, estimating facilities and transportationcosts, and developing economic models are prac-ticed throughout the industry. Probabilistic tech-niques are well known from elementary probabil-ity and statistics. The three-point method wasdeveloped by J. E. Warren of Gulf Oil Corporationin the late 1970s (Warren, 1980–1984, personalcommunication) and used in the years before theChevron-Gulf merger. The three-point method isbased on an operator for estimating moments ofdistributions described by Pearson and Tukey(1965) and Keefer and Bodily (1983). Anapproach similar to Warren’s was also discussedby Bourdaire et. al. (1985).

This process was introduced to ChevronOverseas Petroleum, Inc., in mid-1989 and hassince been adopted by the other operating com-panies upstream in Chevron. Because of its easeof use, transparency, and the built-in mechanismof postdrill feedback, the process has been wide-ly accepted by explorationists and senior manage-ment to provide consistent, credible estimates ofvalue that can be used to compare and rankexploration projects across business unit andoperating company boundaries. The use of thisprocess to provide risk, volumetric, and econom-ic input to exploration decision making has allbut eliminated the previous gap between predict-ed and actual results.

1088 Evaluating Prospects

POSTDRILLREVIEWIf Success,

Compare ActualParameters to

Predicted;If Failure,

Reason Why

ECONOMICANALYSISCash Flow

Modeland ValueMeasures

PLAY CONCEPTSource Rock,

Reservoir, Trap,Timing, andMigration

RISKTesting a Stabilized

Flow ofHydrocarbons

ENGINEERINGConceptual

Development PlanFacilities Costs

Production ProfileRecovery Factor

VOLUMETRICSVolumetric Distribution

of Hydrocarbons(In-Place and Estimated

Recoverable)

DECISION

OPTIMIZATION

953009 fre

DECISION

Figure 1—The exploration evaluationprocess incorporatesspecification of geologicplay concept, assessmentof geologic risk, estimation of hydrocarbon volumes,conceptual engineering,and a development plan for economic analysis. The process includes afeedback loop for process improvementbased on results of comparisons betweenpredrill and postdrillresults.

PLAY CONCEPT

The distribution of hydrocarbons in the Earth’scrust follows a lognormal distribution typical ofmany other natural resources. Such a distributionimplies that hydrocarbons are concentrated in rela-tively few basins, and that exploration is not anequal-chance game. In our assessment process, weevaluate four different concepts of exploration as afunction of the degree of knowledge about the spe-cific project: basin framework, petroleum systemframework, play, and prospect.

Basin Framework

Is there a volume of sedimentary rocks capableof containing potential ingredients of a working“hydrocarbon machine”: source, reservoir, trap andseal, and proper timing and migration? This assess-ment is a screening device only, and does notinclude economic considerations.

Petroleum System Framework

The petroleum system framework is defined asa volume of sedimentary rocks containing hydro-carbons and charged by a single source rock. Thedefinition requires manifestations of hydrocar-bons (seeps, shows, or a producing well) and isapplicable in many frontier basins only by analogy.

Recognition of an active petroleum system alsoserves only as a screening device because it car-ries no volumetric (and therefore, no economic)value.

Play

In our definition, the play is the elemental partof a petroleum system, and is recognized as hav-ing one or more accumulations of hydrocarbonsidentified by a common geological character ofreservoir, trap, and seal; timing and migration;preservation; a common engineering character oflocation, environment, and fluid and flow proper-ties; or a combination of these. Individual plays,therefore, have unique geological and engineeringfeatures, and can be used as a basis for economiccharacterization.

Prospect

Prospect represents an individual, potentialaccumulation. Each prospect is perceived asbelonging to an individual play, characterized byrisk components and a probabilistic range distri-bution of potential hydrocarbon volumes withinits trap confines.

In frontier areas, geological analogs provide thebest models for assessing the capability of the eval-uated basin to yield commercial accumulations of

Otis and Schneidermann 1089

Figure 2—The timing riskchart (Magoon, 1987)helps to integrate geological knowledge and factual information for risk assessment, volumetric parameterranges, and engineering considerations.

hydrocarbons. In more mature areas, the presenceof a petroleum system has been proven, and theassessment focuses on play types. Regardless ofthe maturity of exploration or the amount of exist-ing production, however, each prospect requires adetailed review of the individual risk components.A timing risk chart (Figure 2), modified from theoriginal ideas of Magoon (1987), provides a veryuseful and user-friendly summary and display ofthe play concept.

RISK ASSESSMENT

Within the evaluation process, the risk consid-ered is geologic risk; i.e., the risk that a produciblehydrocarbon accumulation exists. We consider aproducible accumulation to be one capable of test-ing a stabilized flow of hydrocarbons. Geologic riskis assessed by considering the probability that thefollowing four independent factors of the play con-cept exist.

(1) Presence of mature source rock (Psource)(2) Presence of reservoir rock (Preservoir )(3) Presence of a trap (Ptrap)(4) Play dynamics (Pdynamics ) or the appropriate

timing of trap formation relative to timing of migra-tion, pathways for migration of hydrocarbons fromthe source to the reservoir, and preservation ofhydrocarbons to the present day.

The probability of geologic success (Pg) isobtained by multiplying the probabilities ofoccurrence of each of the four factors of the playconcept.

If any one of these probability factors is zero, theprobability of geologic success is zero.

Geological success is defined as having a sus-tained, stabilized flow of hydrocarbons on test. Wedo not consider the oil machine to work with onlyoil and gas shows or flows of hydrocarbons with-out pressure stabilization. This definition elimi-nates very low-permeability reservoirs, reservoirsof limited areal extent, biodegraded oils, and othermarginal cases that cannot deliver a stabilized flowof hydrocarbons from the success case. In practice,this definition has been easily applied to the rangeof prospects drilled during the time the process hasbeen used.

The probabilities that any of the play (or risk)factors occur are estimated by first analyzing theinformation available. The risk assessment checklist(Figure 3) was designed to assist the earth scientistin examining as much information as possible. The

checklist has been compiled over several years,with input from personnel inside and outside ofChevron to ensure all aspects of each play factorare considered. The checklist categorizes the fourrisk factors with following elements.

The r isk assessment worksheet (Figure 4)records our assessments of the elements of therisk factors, which are expressed as unfavorable,questionable, neutral, encouraging, and favorable.With little or no data, assessment is based on eval-uating the analogs and the likelihood that themodel will ref lect the analog. As data areacquired, we begin to develop opinions support-ed by the data. These opinions may be positive(encouraging or favorable) or negative (question-able or unfavorable). Factors with equal probabili-ty of positive or negative outcomes are given aprobability of occurrence of 0.5.

Assessments of encouraging or questionable arebased on indirect data that support or do not sup-port the model. Examples of indirect data for anassessment of encouraging include shows, seeps,and presence of direct analogies. Examples of indi-rect data for an assessment of questionable includelack of shows in nearby wells, thin or poor reser-voirs, and evidence of recent faulting. With indirectdata, we are more dependent on the model than onthe data, and our opinions are supported, but notconfirmed, with data. With indirect data support-ing the model, probability of occurrence is encour-aging, with values between 0.5 and 0.7. When indi-rect data do not support the model, probability ofoccurrence is questionable, with values between0.3 and 0.5.

Assessments of favorable or unfavorable arebased on direct data that tend to confirm or dis-prove the model. Examples of direct data for anassessment of favorable include nearby producingfields or wells with stabilized f lows on test,proven hydrocarbon systems with moderate tohigh source potential index (>5, based on high-quality Rock-Eval data) (Demaison and Huizinga,1991), and maturation models with parameterssupported by data from nearby wells. Examplesof direct data for an assessment of unfavorableinclude dry wells testing similar structuresdefined by good-quality seismic, lack of reservoirin wells, and a hydrocarbon system with very lowsource potential index (<2, based on high-qualityRock-Eval data). With direct data supporting themodel, probability of occurrence is favorable,with values between 0.7 and 0.99. When directdata do not support the model, probability ofoccurrence is unfavorable, with values between0.01 and 0.3.

We record our assessments on the worksheet,and as we complete each factor, we assign a valuecorresponding to the key at the bottom of the

P P P P Pg source reservoir trap dynamics= × × ×



Figure 3—The risk assessment checklist lists the critical aspects of geologic risk assessment to help ensure allaspects have been considered.


Figure 4—The risk assessment worksheet provides a method for transferring qualitative judgments on geologic riskto quantitative probability of geologic success.

worksheet (Figure 4). Note that the probability ofoccurrence for each element depends on the least-favorable assessment.

During the past 5 yr, an understanding of riskhas evolved into five broad categories and general“rules of thumb” that allow characterization of riskand reduce impractical arguments over specificnumbers.

(1) Very low risk (Pg between 0.5 and 0.99, bet-ter than 1:2). All risk factors are favorable. This cat-egory is associated with wells that test proven playsadjacent to (<5 km) existing production.

(2) Low risk (Pg between 0.25 and 0.5, between1:4 and 1:2). All risk factors are encouraging to favor-able. This category is associated with wells that testproven plays near (5–10 km) existing production.

(3) Moderate risk (Pg between 0.125 and 0.25,between 1:8 and 1:4). Two or three risk factors areencouraging to favorable—one or two factors areencouraging or neutral. This category is associatedwith wells testing new plays in producing basinsor proven plays far from (>10 km) existing produc-tion.

(4) High risk (Pg between 0.063 and 0.125,between 1:16 and 1:8). One or two risk factors areencouraging—two or three factors are neutral orencouraging to neutral. This category is often asso-ciated with wells testing new plays in producingbasins far from (>20 km) existing production orproven plays in an unproved area.

(5) Very high risk (Pg between 0.01 and 0.063,worse than 1:16). Two to three risk factors are nobetter than neutral, with one or two factors ques-tionable or unfavorable. This category is usually asso-ciated with wells testing new plays in an unprovedarea far from (>50 km) existing production.

This categorization is summarized in Figure 5.

VOLUMETRICS

Oil and gas volumes are expressed as a productof a number of individual parameters. Because ofuncertainty in the value of each of the individualparameters, oil and gas volumes can be represent-ed as a distribution. The distribution is generallyassumed to be lognormal (Capen, 1993). In ourprocess, the distribution represents the range ofrecoverable hydrocarbons (or reserves, in theirmost general sense) expected to be found whenthe well is drilled, assuming geologic success (sta-bilized flow of hydrocarbons on test). It is not thedistribution representing the range of commercialreserves, proven reserves, or any other type ofreserves tied to economic considerations. Notethat we use the term reserves as being inter-changeable with recoverable volumes throughoutthis text based on the general definition ofreserves being “those quantities of hydrocarbonsthat are anticipated to be recovered from a givendate forward.” (Journal of Petroleum Technology,1996, p. 694). We address commerciality duringthe economics phase of the process.

One method that can be used to obtain this dis-tribution of reserves is Monte Carlo simulation. Thedistribution is obtained by specifying distributionsfor each of the individual parameters and then mul-tiplying randomly selected values together manytimes, thereby creating a highly sampled histogramthat approximates the actual distribution. Thenumber of estimates (iterations) necessary toobtain a satisfactory representation of the distribu-tion ranges from a few hundred to several thou-sand. Monte Carlo simulation programs are widelyavailable and the calculation can be done in afew minutes, depending on the number of itera-tions used.


Same PlayAdjacent Structure

Same PlayNearby Structure

New Play - Same TrendOld Play - New Trend

New Play - New Basinor Play with Negative Data

Avg. Pg= 0.75 Avg. Pg= 0.375 Avg. Pg= 0.183 Avg. Pg= 0.092 Avg. Pg= 0.05Pg= Probability of Geological Success

VERYLOWRISK

LOWRISK

MODERATERISK

HIGHRISK

VERYHIGHRISK

1:2 1:4 1:8 1:16

Producing Area Emerging Area Frontier Area

Delineation Prospect Play Hydrocarbon System

Evaluation FrontierConventionalFigure 5—Risk categorizationof “rules of thumb” for geologicrisk assessment based on feedback from five years ofdrilling history.

An alternative method to Monte Carlo simula-tion was developed by J. E. Warren of Gulf OilCorporation (Warren, 1980–1984, personal commu-nication). This method produces distributions thatare essentially identical to Monte Carlo simulations,but requires no iterations and no assumptions aboutthe distributions of the reserve parameters. We callthe method the three-point method; it is explainedin detail in Appendix 1. Briefly, the method uses asinput a range for each parameter by specification ofvalues corresponding to the 5, 50, and 95% proba-bility of occurrence. From these ranges, a meanand variance are estimated for each parameterusing the Pearson-Tukey operator (Pearson andTukey, 1965). The means and variances are com-bined to provide the mean and variance of theresultant reserve distribution. A lognormal distribu-tion is assumed for the reserves distribution andcan be calculated from the estimated mean andvariance.

Advantages of this method are the speed of thecalculation, which is essentially instantaneous onany spreadsheet computer program, and that it hasno requirement for specifying the parameter distri-bution. The key to success with this method, there-fore, is correctly specifying the ranges. Guidelinesinclude the following:

(1) Selecting the 5% value, which is generallynear the minimum value expected. For example,for porosity the 5% value would be near the mini-mum porosity observed in nearby wells; for area,the 5% value would be the area corresponding tothe minimum hydrocarbon column expected.The explorationist should keep in mind that theodds of finding a value less than the selection are1 in 20.

(2) Selecting the 95% value, which is generallynear the maximum value expected. For example,for porosity the 95% value would be near the maxi-mum porosity observed in nearby wells; for area,the 95% value would be the area corresponding to a maximum hydrocarbon column expected.Likewise, the explorationist should keep in mindthat the odds of finding a value greater than theselection are 1 in 20.

(3) Selecting the 50% value, which is generallynear the middle of the expected range of values.The median is often the most difficult to chooseand requires the support of data associated withthe play or with an appropriate analog. Analogsshould be used with caution. For example, in apurely continental basin, a partial analog withlacustrine source and marine reservoir does notapply. The explorationist should keep in mind thatthe odds of finding a value less than the selectionis equal to the odds of finding a value greater thatthe selection.

After the ranges for the reserve parameters havebeen specified, the mean and variance for thereserve distribution are calculated. Figure 6 showsa spreadsheet with an example for a typical smallprospect in a deltaic environment, such as theNiger Delta or the Mississippi Delta. The inputranges are as shown, and the output informationincludes the mean reserves and cases for a pes-simistic result (10% or P10) and an optimistic case(90% or P90). In addition to reserves, the spread-sheet calculates values for individual reservoirparameters, including porosity, area, and net pay,that, when multiplied together, will total the pes-simistic or optimistic reserve value for use duringthe engineering and economics phases of the pro-cess. These pessimistic and optimistic parametervalues are consistent with the variances specifiedby their corresponding input ranges. Note that theparameter values are not the 10 and 90% values ofthe input ranges. Figure 6 also shows the cumula-tive reserve distribution and values for specific per-centiles, as well as the mean, median, and mode.

In practice, the mean value for the distribution iscommonly less than the explorationist’s expecta-tion. At this point it is critical to keep in mind thatthis result is the consequence of the input parame-ter ranges. If the input ranges are based on goodavailable data, it may be difficult to alter them sig-nificantly, and the explorationist may have to adjustexpectations. This dilemma can be resolved bycomparing the prospect reserve distribution tofield-size distributions of the play or analogs.Questions that arise and responses to them ofteninclude the following:

(1) Are the predicted values reasonably consis-tent with reserves found in analogs to date? If so,use the numbers obtained from the input parame-ter ranges.

(2) Are the predicted reserves significantly small-er or larger than those found in analogs to date? Ifyes, then

(3) Are there technical reasons to justify the dif-ference? If so, use the ranges as stated.

(4) Are technical reasons for the difference lack-ing? If so, reconsider values assigned in previoussteps and recalculate reserves.

When the final reserve distribution is obtained,the information from the process moves to theengineering support and economics stages.

ENGINEERING SUPPORT AND ECONOMICS

The amount of time spent making a conceptualdevelopment plan for an exploration prospect isminimal. With the small amount of informationavailable concerning the nature and extent of the



Figure 6—Three-point-method spreadsheet illustrates volumetric parameter ranges and shows calculations basedon Pearson-Tukey estimator and the three-point method. M = million.

Figure 7—An economic summary sheet provides critical economic and geologic information and provides a mechanism for estimation of commercial or economic risk. M = million.

reservoir (or even if there is a reservoir), fluid prop-erties, or amount of resource present, our experi-ence indicates the time and costs of preparing adetailed development plan for a specific case aregenerally not justified. However, significant atten-tion is given to the credibility of general plans cov-ering a range of cases that rely heavily on analogsor nearby producing examples. This approach isdiscussed in the following paragraphs.

The first step is to take the mean reserve casefrom the volumetric distribution and construct a“mean” development plan. This plan uses the meanparameters from the volumetrics and mean param-eters for reservoir fluid and flow properties to con-struct a mean production profile. This becomes themean case (base case) for which facilities, drilling,and transportation costs are estimated. From thisinformation, the revenue profile, based on the pro-duction profile and a product price assumption; aninvestment profile, based on the phasing of drilling,facilities, and transportation costs; an operatingcost profile, based on an expected opex/bbl as afunction of time; and a miscellaneous expense pro-file characterize the “mean” development plan andare used as input for the economic model preparedfor the prospect.

The economic model is then prepared based onthe host country contract, if available. If no con-tract is available, the economic model is based onother known contracts or other published infor-mation pertinent to the country. The economicmodel takes as input the production, investment,operating cost, and miscellaneous profiles andapplies the contract terms, resulting in outputprofiles of net income to the company and othertax-related profiles, such as depreciation, royalty,and income tax. The model remains f lexible; ifnegotiations are not complete, the contract usual-ly becomes a subject of the negotiations and com-monly changes.

The engineering and economic phases general-ly require refinement and involve a feedback loopto mature the mean case. In other words, theengineer constructs the conceptual developmentplan and economics are run. Economic output isexamined, and an optimization loop among earthscientist, engineer, and economist generally takesplace, resulting in modifications or refinementsto the plan and subsequent economic output.Modifications are generally applied to facilities anddrilling plans because of preliminary poor econom-ic indicators. If modifications do not result in eco-nomics acceptable for a commercial project, theprospect is generally abandoned at this stage. Theconstruction of this “mean” development plan gen-erally takes from 1 day to 2 weeks, depending onthe time available before a decision point and theinformation available.

Once the mean case is completed, pessimistic(P10) and optimistic (P90) cases are run by modify-ing the mean case input profiles to the economicmodel. Modifications are based on the pessimisticand optimistic reserve cases from the reserve distri-bution. Economics are run for these two additionalcases, and a range of economic outcomes is estab-lished. Volumetrics, development and contractassumptions, and economic results are summarizedon a 1-page summary data sheet, as shown in Figure7. The basic layout of the summary is a synopsis ofterms, development assumptions, and a range ofvolumetric parameters and their impact on eco-nomic results. Two graphs are displayed that show(1) the volumetric distribution, both cumulative anddensity, and (2) the resultant ROR (rate of return) forthe unrisked case and several risked cases. Fromthese graphs, one can easily see the economic con-sequences of the expected distribution of reserves,development plans associated with that distribution,and the contract. Additional information, such asNPV (net present value) and NCF (net cash flow), is


RISK

NU

MB

ER

OF

WE

LLS

2

4

6

8

10

1:2 1:4 1:6 1:8 1:10 1:12 1:16 >1:160

1:14

Figure 8—A risk histogram ofevalution wells, 1989–1990, illustrates predicted and actualresults for feedback into the riskassessment process.


246810

Number of Wells

12

Number of Wells

Number of Wells Number of Wells

24681012

1:2

1:4

1:6

1:8

1:10

1:12

1:14

1:16

>1:1

6

24681019

94

Ris

k

12

1:2

1:4

1:6

1:8

1:10

1:12

1:14

1:16

>1:1

6

1993

Ris

k

24681012

1991

1:2

1:4

1:6

1:8

1:10

1:12

1:14

1:16

>1:1

6R

isk

1:2

1:4

1:6

1:8

1:10

1:12

1:14

1:16

>1:1

6

1992

Ris

k

Fig

ure

9—

Ris

k h

isto

gram

s fo

r 1991–1

994 s

ho

w p

rogr

ess

of

imp

rove

men

t in

ass

essm

ent

of

risk

ove

r a

per

iod

of

4 y

r.

also plotted at the P10, mean, and P90 cases to illus-trate results for those parameters as well.

Given the range of possible outcomes for the vol-umetrics and their economic consequences, an esti-mate of commercial risk is easily determined. Giventhe conditions of commerciality, usually a minimumROR, the probability of a commercial prospect canbe read directly from the two graphs. In Figure 7, if a20% ROR is considered a minimum for a commercialproject, from the bottom graph a 20% ROR corre-sponds to a reserve of 11 MBO (million barrels ofoil). From the top graph, 11 MBO corresponds to a50% probability of finding that reserve or more.Thus, the probability of commercial success isapproximately 50%. This will vary from prospect toprospect, but this link is the fundamental driver forthis process. In other words, we need to understandwhat nature has provided, which is the volumetricdistribution that describes what we might find whenwe drill the well. We must also understand the eco-nomic consequences; that is, what nature has pro-vided may or may not yield satisfactory economics.Analysis of both geologic and commercial risk in thismanner allows appropriate decisions regarding risktolerance and potential reward.

POSTDRILL REVIEW

Postdrill information is primarily used as feed-back to the risk assessment and volumetric estima-tion phases of the process. Feedback to the engi-neering and economics sections generally does notoccur within a time frame that can impact the pro-cess. In other words, by the time a discovered fieldis developed and feedback is obtained, the processhas already changed because of other, more timely,reasons.

Postdrill information is obtained from a postdrillwell review conducted within a few months aftercompleting the well. Data analyses are collectedand reviewed to (1) determine reasons for failure if

the well is unsuccessful, (2) compare predictedand actual reserves parameters if the well is suc-cessful, and (3) review lessons learned regardless ofthe result. Individual postdrill well reviews arecompiled on an annual basis to provide statisticalfeedback, using simple histograms for both riskassessment and volumetric estimation.

The first tool is the risk histogram, a simple plotof well results vs. risk expressed as a fraction ofprobability of success. Figure 8 shows a risk his-togram from an actual 1989–1990 drilling programof wells drilled in producing areas on producingplays (evaluation wells). As is evident from the plot,the bulk of the wells had predrill probability of geo-logical success between 1:3 and 1:6 (30–15%).From the histogram, it was immediately obviousthat the number of successful wells is inconsistentwith the assessed r isk. For those wells withassessed risk of 1:2, or 50%, 100% of the wells weresuccessful. For those wells with assessed risk of1:3, or 33%, 87% of the wells were successful, andso on. In fact, the average success rate for all wellsdrilled was 50% rather than the 20–25% predictedby the mode of the histogram.

For this type of well (proven play in a producingarea), our first modification to the process was tomodify our process of assessing r isk to betterreflect our actual success rate. Figure 9 shows therisk histogram for each of the subsequent years(1991–1994). Although our efforts to more correct-ly assess risk were not immediately successful, overthe 4-yr period improvement is evident, and by1994 our predicted success rate is more consistentwith that observed.

As a side note, examining drilling results prior to1989 indicated a similar trend. The success rate forwells drilled on proven plays in producing areas isabout 50%, or 1:2, whereas the predicted rate wasabout 0.3–0.2, or 1:3 to 1:5. However, no attemptwas made to adjust risk assessment methods until theprocess was implemented in 1989. Apparently, every-one knew the answer, but without a methodical,


Reserves (MBO)

Pro

babi

lity

of F

indi

ngR

eser

ves

Less

Tha

n (%

)80

100

20

60

40

0100 200 3000

Actual Reserves,190 MBO,

corresponds to64th percentile

Predrill Reserve Distribution Figure 10—Predicted distributionof reserves with actual results atthe indicated percentile. In thiscase, the actual reserves of 190MBO fell on the 64th percentile ofthe distribution.

periodic performance review, little was done tomodify the process. Thus, the feedback step isconsidered critical to the success of any process;without it , no process will be modified andimproved.

Volumetric estimation feedback is somewhatmore complicated because it requires a method todetermine whether distributions are being accu-rately estimated. Our volumetric feedback processconsists of two steps. The first step is to determinewhether reserve distributions are accurate. Thesecond step is to determine whether the individualreserve parameters are accurate. The method is thesame for both steps and uses a second tool, the per-centile histogram. The percentile histogram is con-structed in the following way.

Given a set of successful wells, each with a pre-dicted distribution of reserves, calculate the proba-bility of occurrence for the actual reserves on thepredicted parameter distribution. For example, inFigure 10 a predicted distribution of reserves isshown where the actual reserves of 190 MBO cor-respond to the 64% probability of occurrence.Extending this to the set of four wells, as shown inFigure 11, the percentiles of the actual reserves onthe predicted reserve distributions 1–4 are 25, 75,21, and 91%, respectively. If these probabilities ofoccurrence for the four distributions are plotted asa histogram of occurrences in the ten dectiles (ten10% intervals), the result is a percentile histogram,also shown in Figure 11.

The percentile histogram can be used to diag-nose a variety of problems, as shown in Figure 12.The desired response is “flat.” In other words, ifwe are estimating distributions correctly there isan equal probability that the actual reserves willfall within any one of the ten dectiles (ten 10%intervals). It is analogous to rolling a ten-sideddie, because each side (a 10% interval) has anequal probability of occurrence. Diagnostics arerelatively simple. If the histogram is heavy to thelow, or downside, we are tending to overestimatepotential. In other words, most of the actualresults are on the downside of the distribution. Ifthe histogram is heavy to the high, or upside, theopposite is true; most of the actual results are onthe upside of the distribution, indicating a ten-dency to underestimate reserves. If the histogramis heavy on the ends and light in the middle,prospect reserve ranges are too narrow and needto be broadened. If the histogram is heavy in themiddle, ranges need to be reduced.

Figure 13 shows the percentile histogram forreserves for Chevron Overseas Petroleum, Inc.,in 1989–1990. The histogram is heavy to thedownside; thus, we had overestimated potentialin the majority of cases and needed to accountfor the large number of small discoveries we hadmade. We knew we had to correct this problem,but the primary cause required additional analy-sis. To determine what was causing the overesti-mation of reserves, we applied the same method


80

6

20 40 60 100

4

2

00

25 50 750 100

100

50

0

100

50

20 40 600

0 80 100 2000 300 400

100

50

02010 300 40

100

50

0

100

50

0

Percentile HistogramNumber of Occurrences

31 MBO25%

250 MBO75%

9 MBO22%

75 MBO91%

Figure 11—Example of percentile histogram with four predicted distributions and actual results. This histogram isused to calibrate estimation of predrill volumetric parameters with actual results.

to individual parameters. The percentile histogramsfor the individual parameters are shown in Figure14. The following observations were made:

(1) Estimates for gross pay and area were consis-tently overestimated.

(2) Estimates of net-to-gross ratio (N:G), porosity,hydrocarbon saturation, and formation volume fac-tor (FVF) were too narrow.

(3) The geometry factor was not being estimatedcorrectly.

Modifications were made to tie ranges of grosspay and area to the expected hydrocarbon column.Research indicated columns associated with previ-ous ranges of gross pay and areal extent weregrossly overestimated, so considerable attentionwas given to hydrocarbon columns expected fordifferent seals, especially fault seals. Other modifi-cations included widening ranges for N:G, porosity,hydrocarbon saturation, and formation volume fac-tor, as well as introducing a different approach toestimating geometry factor.

Figure 15 shows the reserve histogram andFigure 16 shows the parameter histograms for1993–1994. The reserves and all parameters havepercentile histograms that are within the statisticaltolerance of being acceptable for the number ofsamples, and it is obvious they are being estimatedwith improved accuracy. The histograms are muchcloser to the desired “flat” response.

Based upon this feedback for both risk assess-ment and volumetric estimation, we observed a dis-crepancy between predicted and actual results,analyzed the data to determine where improve-ments could be made, implemented those changes,and observed a favorable response when predictedand actual results were in better agreement. Thefeedback was absolutely necessary to establishcredibility and build support for the continued useof the process.

CONCLUSION

Since its inception in 1989, application of thisprocess has resulted in a consistent method ofassessing risk, estimating volumes of hydrocarbons,and, thus, calculating economic indicators that canbe used to judge the potential of explorationprospects. Through yearly feedback and modifica-tions, credibility has improved, and the process hasbeen accepted by Chevron upstream operatingcompanies as a basis to assess the potential ofopportunities in Chevron’s worldwide explorationprospect inventory. The process is used routinelyin international exploration activities and has beenthe subject of numerous training sessions withpartners and host countries.


• S

kew

to

lo

wsi

de

• D

istr

ibu

tio

n t

oo

op

tim

isti

c o

n d

ow

nsi

de

• S

atis

fact

ory

on

up

sid

e

• S

kew

to

hig

hsi

de

• D

istr

ibu

tio

n t

oo

pes

sim

isti

c o

n u

psi

de

• S

atis

fact

ory

on

do

wn

sid

e

• B

imo

dal

on

lo

w-

and

hig

hsi

des

• D

istr

ibu

tio

n is

to

o n

arro

w

• C

ente

r w

eig

hte

d

• D

istr

ibu

tio

n is

to

o w

ide

• D

esir

ed u

nif

orm

dis

trib

uti

on

• D

istr

ibu

tio

ns

are

sat

isfa

cto

ry

Fig

ure

12

—E

xam

ple

s o

f p

erce

nti

le h

isto

gram

s w

ith

dia

gno

stic

in

terp

reta

tio

ns.


Figure 13—Actual percentile histogram for years 1989–1990. Diagnostics indicate distribution estimates were too optimistic ondownside uncertainty(downside and medianestimates were too large).

Figure 14—Actual percentile histograms for parameters of reserve distribution for years 1989–1990. Note problemswith area, gross pay, geometry factor, porosity, and hydrocarbon saturation.


Figure 15—Actual percentile histogram foryears 1993–1994 after modifications to process.Note distributions are more consistent with desirableuniform distribution.

Figure 16—Actual percentile histograms for parameters for years 1993–1994 after modification to process. Noteproblems have essentially been eliminated and distributions are consistent with desirable uniform distribution.

APPENDIX 1: THREE-POINT METHOD

The three-point method, as developed by J. E. Warren(1980–1984, personal communications) for reserve estimation,uses the general equation shown below, which combines individ-ual parameters in calculating recoverable reserves, R.

where A = areal extent of prospect in acres, h = average net pay infeet, f = average porosity, Sh = hydrocarbon saturation (1 – Sw,where Sw = water saturation), Boi = initial oil formation volume fac-

tor in reservoir barrels/stock tank barrels (STB), Bgi = initial gas for-mation volume factor in reservoir cubic feet/surface cubic feet, Rfo= recovery factor for oil, Rfg = recovery factor for gas, CR = con-densate recovery factor in STB/ft3, 7758 = conversion factor fromacre-feet to barrels, and 43560 = conversion factor from acre-feetto cubic feet.

The parameters are combined by multiplication; therefore, ifthe parameters are assumed to be probabilistically independent,the reserve distribution, R, will be lognormal in the limit as provid-ed by the central limit theorem. Likewise, the first and secondmoments of R [m(R) and m2(R)], respectively, will be the productof the first and second moments of the parameter distributions,respectively, as shown. Note that the first moment of the distribu-tion is the mean.

(1)m R oil 7758 m A m h m

m Sh m 1 B m Roi fo

( )[ ] = × ( ) × ( ) × ( ) ×

( ) × ( ) × ( )φ

R(condensate) 4 A h S 1 B R CRh gi fg= × × × × × ( ) × ×3560 φ

R(gas) 4 A h S 1 B Rh gi fg= × × × × × ( ) ×3560 φ

R(oil) 7758 A h S 1 B Rh oi fo= × × × × × ( ) ×φ


Figure 17—Step 1 of three-point method for calculating reserve distributions: specify parameter ranges. M = million.

(2)

With the first and second moments of R, the lognormalreserve distribution is completely specified. Even if probabilis-tic independence is not strictly valid, the results are a usefulapproximation, given the level of information generally avail-able to an exploration project. In practice, the uncertainty inspecifying the ranges of input parameters is far greater than theamount of uncertainty introduced by assuming parameter inde-pendence.

The first and second moments of R are calculated using equa-tions 1 and 2 and estimates of the first and second moments of the

input parameter distributions. These estimates are obtained usingthe Pearson-Tukey estimator (Pearson and Tukey, 1965; Keeferand Bodily, 1983). An example for the area, A, is

where P5 = the 5% probability of occurrence of the area distribu-tion, P50 = the median of the area distribution, and P95 = the 95%probability of occurrence of the area distribution.

m A 0.185 P5 A 0.63 P50 A 0.185 P95 A2( ) = × ( ) + × ( ) + × ( )2 2 2

m A 0.185 P5 A 0.63 P50 A 0.185 P95 A( ) = × ( ) + × ( ) + × ( )

m R oil 7758 m A m h m

m Sh m 1 B m R

2 2 2 2

2 2 oi 2 fo

( )[ ] = × ( ) × ( ) × ( ) ×

( ) × ( ) × ( )φ


Figure 18—Step 2 of three-point method for calculating reserve distributions: calculate parameter means and vari-ances. M = million.

The Pearson-Tukey estimator is used because of its robustnessin estimating mean values from a wide variety of nonsymmetricdistributions, including the popularly used triangular distribution.Thus, the estimated mean values estimated are not restricted toany assumptions of distribution, such as those necessary for aMonte Carlo simulation, and allow the Earth scientist a reasonableamount of freedom in choosing the input values for the P5, P50,and P95 estimates.

At this point it is useful to introduce a more convenient param-eterization, ∂2, the variance of the natural logarithm of R. ∂2 is cal-culated using the following formula.

It is easy to show that the variance of the natural logarithm of Ris the sum of the ∂2 of the individual parameters. Thus,

∂ ∂ ∂ ∂ φ

∂ ∂ ∂

2 2 2 2

2h

2oi

2fo

R oil A h

S 1 B R

( )[ ] = ( ) + ( ) + ( ) +

( ) + ( ) + ( )

∂2 = ( ) ( )[ ]ln m R m R2

2


Figure 19—Step 3 of three-point method for calculating reserve distributions: calculate mean and variance ofreserve distribution. M = million.

and any percentile value of the lognormal distribution can be cal-culated using the formula

where P50(R) = m(R) * e-0.5∂2 (the median of the distribution), x= the probability of occurrence desired, z(x) = the value or z-factor corresponding to the x-percentile of the standard normal

distribution (obtained from tables given in most probability text-books).

Figures 17–20 show a spreadsheet with the example from thetext and illustrate the calculation process.

Step 1: Specify the parameter ranges.Step 2: Calculate a mean and ∂ (variance) for each parameter.Step 3: Multiply the parameter means and sum the ∂ to obtain

the mean and ∂ of the reserve distribution.Step 4: Calculate values for different probabilities of occurrence

as listed in the table and plotted on the cumulative distribution.

R P50 R exz x= ( ) × ( )∂


Figure 20—Step 4 of three-point method for calculating reserve distributions: calculate values for different probabil-ities of occurrence. M = million.

REFERENCES CITEDBourdaire, J. M., R. J. Byramjee, and R. Pattinson, 1985, Reserve

assessment under uncertainty—a new approach: Oil & GasJournal, June 10, v. 83, no. 23, p. 135–140.

Capen, E. C., 1993, A consistent probabilistic approach to reservesestimates: Society of Petroleum Engineers HydrocarbonEconomics and Evaluation Symposium, SPE Paper 25830, p. 117–122.

Demaison, G., 1984, The generative basin concept, in G. Demaisonand R. J. Murris, eds., Petroleum geochemistry and basin evalua-tion: AAPG Memoir 35, p. 1–14.

Demaison, G., and B. J. Huizinga, 1991, Genetic classification ofpetroleum systems: AAPG Bulletin, v. 75, p. 1626–1643.

Dow, W. G., 1972, Application of oil correlation and source rockdata to exploration in Williston basin (abs.): AAPG Bulletin, v. 56, p. 615.

Dow, W. G., 1974, Application of oil correlation and source rockdata to exploration in Williston basin: AAPG Bulletin, v. 58, no. 7, p. 1253–1262.

Haun, J. D., ed., 1975, Methods of estimating the volume of undis-covered oil and gas resources: AAPG Studies in Geology 1, 206 p.

Jones, R. W., 1975, A quantitative geologic approach to predictionof petroleum resources, in J. D. Haun, ed., Methods of estimat-ing the volume of undiscovered oil and gas resources: AAPGStudies in Geology 1, p. 186–195.

Journal of Petroleum Technology, 1996, SPE/WPC draft reservesdefinitions: Journal of Petroleum Technology, v. 48, no. 8, p. 694–695.

Keefer, D. L., and S. E. Bodily, 1983, Three-point approximationsfor continuous random variables: Management Science, v. 29,no. 5, p. 595–609.

Magoon, L. B., 1987, The petroleum system—a classificationscheme for research, resource assessment, and exploration(abs.): AAPG Bulletin, v. 71, p. 587.

Magoon, L. B., 1988, The petroleum system—a classificationscheme for research, exploration, and resource assessment, inL. B. Magoon, ed., Petroleum systems of the United States: U.S.Geological Survey Bulletin 1870, p. 2–15.

Magoon, L. B., 1989, The petroleum system—status of research

and methods, in L. B. Magoon, ed., The petroleum system—status of research and methods, 1990: U. S. Geological SurveyBulletin 1912, p. 1–9.

Magoon, L. B., and W. G. Dow, eds., 1994, The petroleum sys-tem—from source to trap: AAPG Memoir 60, 655 p.

Megill, R. E., 1984, An introduction to risk analysis: Tulsa,Oklahoma, PennWell Books, 274 p.

Nederlof, M. H., 1979, The use of habitat of oil models in explo-ration prospect appraisal: Proceedings of the 10th WorldPetroleum Congress, p. 13–21.

Newendorp, P. D., 1975, Decision analysis for petroleum explo-ration: Tulsa, Oklahoma, PennWell, 668 p.

Otis, R. M., 1995, Five year look back at risk assessment and esti-mation of hydrocarbon volumes (abs.): AAPG 1995 AnnualConvention Program, p. 73A.

Otis, R. M. and N. Schneidermann, 1994, A process for valuation ofexploration prospects (abs.): AAPG 1994 Annual ConventionProgram, p. 228.

Pearson, E. S., and J. W. Tukey, 1965, Approximate means andstandard deviations based on distances between percentagepoints of frequency curves: Biometrika, v. 52, no. 3–4, p. 533–546.

Perrodon, A., 1980, Géodynamique pétrolière. Genèse et répartitiondes gisements d’hydrocarbures: Paris, Masson-Elf Aquitaine, 381 p.

Perrodon, A., 1983, Dynamics of oil and gas accumulations: Pau,Elf Aquitaine, p. 187–210.

Perrodon, A., 1992, Petroleum systems: models and applications:Journal of Petroleum Geology, v. 15, no. 3, p. 319–326.

Rose, P. R., 1987, Dealing with risk and uncertainty in exploration:how can we improve?: AAPG Bulletin, v. 77, no. 3, p. 485–490.

Rose, P. R., 1992, Chance of success and its use in petroleumexploration, in R. Steinmetz, ed., The business of petroleumexploration: AAPG Treatise of Petroleum Geology, Handbookof Petroleum Geology, p. 71–86.

White, D. A., 1980, Assessing oil and gas plays in facies-cyclewedges: AAPG Bulletin, v. 64, no. 8, p. 1158–1178.

White, D. A., 1988, Oil and gas play maps in exploration andassessment: AAPG Bulletin, v. 72, no. 8, p. 944–949.

White, D. A., 1993, Geologic risking guide for prospects and plays:AAPG Bulletin, v. 77, p. 2048–2061.



Robert M. Otis

Bob Otis is supervisor for Cabin-da B/C Exploration, Chevron Over-seas Petroleum, Inc. PreviousChevron experience includes man-ager, exploration evaluation divi-sion, coordinator Argentina explo-ration, and coordinator MiddleEast exploration. Before joiningChevron, Bob worked one year forthe Western Division of Sohio(California and Alaska) and eight years for Mobil in GulfCoast and Alaska exploration. He received a B.S. degreein 1969 and a Ph.D. in 1975, both from the University ofUtah.

Nahum Schneidermann

Nahum Schneidermann is direc-tor of international technical rela-tions, executive staff, ChevronOverseas Petroleum, Inc., SanRamon, California. A native ofZayadin, former Soviet Union(now Uzbekistan), Schneidermannreceived his bachelor’s and mas-ter’s degrees from the HebrewUniversity of Jerusalem, Israel, in1967 and 1969, respectively, andhis Ph.D. from the University of Illinois, Urbana, Illinois,in 1972. His career in the industry started in 1974 withGulf Oil, where he held various positions at theHouston Technical Services Center. In 1985 he startedhis tenure with Chevron Overseas Petroleum in SanRamon, serving as manager, basin studies and geochem-istry, for the exploration department prior to beingnamed to his present position.

ABOUT THE AUTHORS

Process evaluateprospect97 1087

Documents

continuous process

prospect ranking

prospect economics

economic evaluation

prospect hydrocarbon

geologic successif

exploration programs

exploration prospects1robert