How To Evaluate Hard-To-Evaluate Reserves (includes associated papers 23545 and 23553 )

Abstract

Summary. Traditional U.S. reserveevaluations are based on tried andtested engineering principles, awealth of local and general experience, and a set of reserve definitionsthat have evolved to become theindustry standard. For the most part, these definitions work well. However, for some of the emerging technologyplays (sometimes called statisticalplays) where individual wellperformances are characterized bysignificant variability of recoveries, applying these definitions alone isinsufficient. The problem is how bestto evaluate tight sands, coalbed methane, Devonian shales, horizontaldrilling in fractured reservoirs, andredevelopment of depleted fields. This paper presents a method for evaluating plays that involve asignificant variability (uncertainty) component. This probability analysisis not new; indeed, it has been formalized to a stage of definitionsof proven, probable, and possible reserve categories. Although thismethod has been applied in manyparts of the world, use in U.S. based reserve evaluations has beenminimal. Case histories compare reserveevaluation methods in two hard-to-evaluate U.S. plays, the Austin Chalkhorizontal drilling play and the SanJuan basin coalbed-methane play. These case histories illustrate thebenefit of coupling classic deterministictechniques with probability analysis to express uncertainty consistentlyand meaningfully. Background The reserve definitions most commonly usedin the U.S. are those published by the Securities and Exchange Commission (SEC)and SPE in conjunction with the Soc. of Petroleum Evaluation Engineers (SPEE). The 1988 SPEE monograph comprehensivelydescribes the most recent version ofthe definitions published by SPE in May1987. Definitions are subdivided into threecategories(proved, probable, and possible), and four status modifiers (producing, shutin, behind pipe, and undeveloped) areavailable for each category, resulting in 12composite descriptions. For any estimate, the category assignedreflects the degree of certainty of theestimate because the category definitionsare based on the results of a test of "reasonable certainty." The status assignmentprovides an indirect measure of confidence. Theclassifications at the top of the list benefitfrom hard production data, while thoselower on the list rely on more inferentialdata and assumptions to derive an estimate. These definitions are strictly deterministic. That is, a single figure is estimated as to thefuture recovery of oil and gas from a welllease field or for a company as a whole. That such estimates are imprecise isacknowledged by all professional reserveevaluators: "In the final analysis, thereliability of reserve estimates is the direct function of the available data and theconfidence and integrity of the estimator." The SEC and SPE reserve definitions donot mention the use of probability analysisin reserve evaluations. In the SPEEmonograph, use of probability analysis isexplicitly rejected unless specifically requested by the client. In contrast, a 1983 World Petroleum Congress (WPC) report addressedthe probabilistic approach. This reportdiscussed how probability might be used inreserve definitions, particularly where thedegree of uncertainty associated with theestimate is large. The resulting definitions may be summarized as follows. Proven reserves--those that have aprobability of existence greater than 85 to95% (a 90% value is used in the subsequent discussion). Probable reserves--the quantities addedto proven reserves that extend the overallprobability of existence to more than 50%. Possible reserves--the quantities added toproven and probable reserves that extend theprobability of overall existence to more than5 to 15% (a 10% value is used in thesubsequent discussion). Note the terminology differences. Provenis used for the highest coincidence categoryinstead of proved. As discussed later, the highest-confidence categories are notequivalent. The evolution of these reserve categorieswas influenced by oil companies'need fora better idea of ultimate potential recoverythan available through deterministic definitionsalone. Such information was especiallynecessary when substantial capitalinvestment decisions were required with only limited reservoir information. A typicalexample is the North Sea. In certain areas ofthe world, the use of probabilistic analysisin reserve evaluation is commonly acceptedboth by the oil companies and regulatoryauthorities(e.g., the London Stock Exchange). We believe that reserve evaluation using SEC/SPE/SPEE deterministic reserve definitionsis the most appropriate approach forthe majority of U.S. oil and gas plays. Where reserve evaluation is required fortechnology or statistical plays (which, bytheir nature, yield wide variations inindividual well performance and consequently are more suited to significant multiwelldevelopment program commitment), theprobabilistic approach enhances traditionalreserve determination methods. Technology/Statistical Plays Technology/statistical plays are not new. Volatility in oil price, the U.S. Tax Reform Act of 1986, Sec. 29 tax credits, and theunrelated high potentials of certain horizontalwell developments have led to a recent emphasis on such plays. In many cases, theseplays are characterized by a few excellentwells with a substantial number of averageand marginal wells. Copyright 1991 Society of Petroleum Engineers JPT P. 998⁁