Reliability and Uncertainty in Reserves: How the Industry Fails, and a Vision for Improvement

Abstract

Review of field reserves from different parts of the world reveals that reliability of reserves estimates is rather poor. Many fields show wide fluctuations in reserves estimates through time, both at pre- and post-production stage, with a tendency toward underestimation. Accompanying this trend, uncertainty attached to the "best" reserve estimate does not shrink with additional data, defying intuition. Reasons for such phenomena are varied, but one fundamental reason is the approach the industry generally takes toward reserves estimation. Reserves variations, whether in upward or downward direction, adversely affect project economics. Factors that affect reserves evaluation include reservoir aspects, development scheme, operations and technology, economic and regulatory aspects, and "intangibles" that relate directly to human input and conduct. In the last-named category, proclivity to ignore statistical rules leads to distortion in reserves and associated uncertainty estimates. The vision toward improving reserves reliability involves a multi-prong approach that tackles all these root causes. An optimal appraisal strategy that addresses the key subsurface uncertainty, due regard to statistical rules, and use of volumetric methods that consider different scenarios and put the evaluator in contact with "real" data, are viewed, among others, as providing tangible improvements. Introduction Reserves make up major assets of oil and gas companies. The companies - on the upstream side - make development and production plans based on reserves, and investors and the public value oil and gas companies largely based on reserves held. Reliability and consistency of reserves, therefore, is a major issue in the oil industry. Revisions made by a number of oil companies last year to proved reserves already booked with the U.S. Securities and Exchange Commission (SEC) have heightened public and regulatory concern on reliability of reserves. This paper deals with reliability and uncertainty of reserves estimates. The topic will be addressed from a broad perspective, without a particular focus on SEC bookings. It will be shown, through a review of field reserves histories from the North Sea, the Gulf of Mexico (GOM) and other areas, that the industry's record of accurately estimating reserves is, generally speaking, unsatisfactory. While there are exceptions, the record starts to be poor at the pre-production phase and continues to be so for years after production. The economic impact and causes of the problem are investigated and some remedies proposed. "Reserves Accuracy": A Misnomer While it may seem trite, it is well to clarify at the outset that the commonly used term "reserves accuracy" - also used in this paper for reasons of convenience - is a misnomer. Accuracy refers to the degree to which an estimated quantity represents the true value. Accuracy is assessed by comparison with the true value. In the case of reserves, however, the true value will be known when the last barrel of oil or the last cubic foot of gas is produced, and that means many years later. Because there is no standard with which to compare an estimate before and during production, accuracy becomes an immeasurable and illusory target until field abandonment. A less ambitious term, ‘reliability,’ seems better suited for reserves. Reserves reliability is inherently poor if reserves are inconsistent and change significantly over time. It is further clarified that the term "reserves," as herein used, refers to total recoverable volumes, including production if applicable. It is thus distinguished from "remaining reserves."

Assessment of the Reliability of Reserves Estimates of Public Companies in the US and Canada

“…Previous authors (Franzen and Sawyer, 1980;Demirmen, 2005;Robinson and Elliot, 2005) analyzed different sets of data in different periods of time to determine the reliability of reserves estimates. Franzen and Sawyer (1980) analyzed biases in the initial estimates made for 40 off-shore field development projects over a ten-year period, and sought to determine the reliability of variables such as project time, drilling time, oil production, and oil reserves.…”

“…However, regarding reserves estimates, the authors could not draw any clear conclusions due to the large variance presented on the recoverable volumes and hence, there was a large uncertainty of what the true or actual reserves would be (Franzen and Sawyer 1980). Demirmen (2005) reviewed field Estimated-Ultimate-Recovery (EUR) variations for the North Sea and for the Gulf of Mexico (GOM) (the author used the term "reserves" rather than "EUR," defining "reserves" as the best estimate of total recoverable volume including production). For the North Sea, the author reviewed 15 major oil fields for a period from 1974 to 2003, observing a clear tendency that on average, EUR from these fields grew over this time period by a factor of 2.7.…”

“…For the GOM, the author normalized pre-and post-production EUR variations for 14 large deep-water fields and concluded that, in general, from year one of production, EUR variations in the GOM were more pronounced than those in the North Sea, as the majority of large GOM deep-water fields had reached the same amount of EUR growth within a much shorter period after production start. The author concluded that the reliability of reserves estimates is poor, as many fields show wide fluctuations in reserves estimates over time, with a tendency toward underestimation of reserves (Demirmen 2005). Robinson and Elliot (2005) reviewed "technical revisions" in annual information forms.…”

Application of Hubbert Peak Theory to Stimulate Biogas Production

“…The level of uncertainty lessens as more data becomes available. Therefore, as resources decrease, reserves increase at the same time when continuing from dis covery to production [13]. Few major uncertainties attached to these kinds of studies are as follows:…”

Application of the Monte Carlo Simulation in Calculating HC-Reserves

“…However, when Demirmen (2005) analyzed the relevant statistics he came to the conclusion that the said decreasing uncertainty was a myth. The authors also share this view pursuant to their practical experiences.…”