February 24, 2005
Mr. Philip Budzik
Office of Integrated Analysis and Forecasting
Energy Information Administration
Washington, DC 20585
Re: Follow-up to My Jan. 26, 2005 Query and Our Subsequent Discussions Regarding Production Scenarios in EIA Analysis of Oil and Gas Production in the Arctic National Wildlife Refuge (EIA Report No. SR/OIAF/2004-04)
Dear Philip Budzik:
While I appreciate the time you spent briefing me February 9 on EIA’s model for estimating potential production from the Arctic National Wildlife Refuge and the information that you subsequently provided, the central questions from my January 26 letter remain largely unresolved. The information you have shared suggests that EIA’s March 2004 report, Analysis of Oil and Gas Production in the Arctic National Wildlife Refuge (Report No. SR/OIAF/2004-04) overstates the estimated annual production that would result from the agency assumptions regarding total undiscovered economically recoverable reserves that may lie beneath the Arctic Refuge Coastal Plain. In the interest of providing policy makers with most accurate information possible regarding Arctic Refuge production potential, this letter revisits and recasts the questions in my letter of January 26 in light of the information you have provided.
The annual production estimates in the EIA March 2004 report are derived from production profiles for the life of the individual fields that, combined, would represent total production from assumed Arctic Refuge discoveries. As I noted in my January 26 letter, profiles that reflect North Slope development experience resulted in significantly lower annual production rates than those indicated in the EIA report. From a public policy standpoint, this discrepancy is significant because annual production rates directly affect estimates of potential effects of Arctic Refuge development on petroleum imports and other aspects of the economy.
Based on the information you have provided, it appears that the root problem is that EIA based its March 2004 estimates of production from the Arctic Refuge Coastal Plain on an outdated model for establishing field production profiles on Alaska’s North Slope. In addition to this problem, I believe that the throughput in the report’s high production scenario may be further inflated by arbitrary assumptions regarding the discovery of larger oil fields. These two problems are discussed in more detail below.
The 1991 Field Production Profile Algorithm
Production profiles estimate annual production rates over the life of a field. You have advised that the model used to determine field production profiles for the 2004 report was published in a 1991 Department of Energy report. In a production profile, annual production rates are typically determined by a combination of factors that include timing of development, peak production rate, years to peak production, years at peak production and the rate of decline from peak. Of particular interest here are peak production rate (achieved early in field life and measured as a fraction of total field production) and rate of decline (measured as an annual percentage decrease in production). When the model EIA used in its 2004 report was published in 1991, Alaska North Slope production was in its 14th year of production and there was little experience with these critical estimating factors. In the intervening 14 years, it has become clear that the artificially high peak production and the assumption of steep decline rates in the 1991 algorithm compress total field production into a shorter life span than actual North Slope fields have exhibited.
The artificial elevation of annual production rates in the 2004 report was not apparent to the casual reader because EIA did not attempt to reconcile field production profiles with the total production anticipated from each field that EIA hypothesized might be put into production. Instead, EIA simply ran the model for the early years of assumed Arctic Refuge production, ending its analysis in 2025 – the final year of published data from the agency’s National Energy Modeling System. I discovered the apparent gap between estimated annual production and the total production only when I tried – without success – to replicate EIA’s annual production rates.
The usefulness of a model depends on how well it represents reality. The following facts suggest that EIA’s outdated model for Alaska’s North Slope may not fare well in this regard: As noted above, the 1991 report from which the model was taken significantly underestimated production from the seven fields then operating or under development. Less than a year after that report was released, an updated version was already in process with significant increases to the production predicted from these seven fields. Moreover, a North Slope model, published by a U.S. Geological Survey economist in 2003, adopts different parameters that have the effect of reducing annual production levels, relative to total production. For example, where the 1991 model used by EIA assumes a field decline rate of 15% per year from peak for most fields, a 2003 report on North Slope development issues by a USGS economist uses a 12% decline rate and notes that even this reduced rate is steeper than observed North Slope decline rates, which have been reduced by the application of recovery techniques to prolong field life.
In sum: identification of the algorithm EIA used in its March 2004 report may resolve the apparent discrepancy between projected total production for the Arctic National Wildlife Refuge shown in EIA report Table 1 (page 6), on the one hand, and the annual production forecasts for that province (if oil is discovered) from 2013 through 2025, shown in report Figure 2 (page 7), on the other. However, EIA’s use of outdated algorithm factors from a 1991 model resulted in annual production rates through 2025 (the year the model abruptly stops) that are artificially elevated, relative to the total production volumes derived from field profiles that reflect 14 years of subsequent experience on Alaska’s North Slope (see Attachment A to my Jan. 26 letter). If EIA intends to provide future analysis of Arctic Refuge production potential and the economic impacts of that production, I respectfully suggest that the report should identify the source of the model employed and discuss EIA’s reasons for selecting that model.
EIA’s “High Oil Resource Case”
The analysis and spreadsheets I provided with my Jan. 26 letter to support these concerns focused primarily on EIA’s “Mean Oil Resource Case,” whose production profile was stated to be based on total production of 4.21 billion barrels of oil. That analysis also raises questions about the reported results of EIA’s “High Oil Resource Case,” which was based on total production of 7.48 billion barrels of oil. The EIA report high case consisted of production from seven fields: one 2.0 billion barrel field, two 1.34 billion barrel fields and four 0.7 billion barrel fields.
Paralleling the discussion of the mean resource case above, my first question relates to high case production profiles. Based on an additional 14 years of experience with North Slope fields and the 2003 model mentioned above, it appears the probable field decline rates from peak would be significantly lower than the 15% rate the 1991 algorithm assigns go six of the seven fields. For the seventh field, algorithm assigns a 12 percent decline rate to the 2.0 billion barrel field; the only field on the North Slope in that size range – Kuparuk – has exhibited a decline rate of less than six percent. If the model were run through the anticipated life of the discovered fields, would an algorithm that more accurately reflects North Slope experience result in longer field life, with correspondingly lower annual production rates than those stated in the EIA report?
A second set of questions relates to the field size assumptions in the EIA high case. In my January 26 letter, I posed the following question regarding EIA assumptions about field size: “If the field size assumptions on which the Arctic Refuge production profiles are based are too optimistic, this, too, would result in overstatement of the potential effects of development of the Arctic Refuge Coastal Plain. For this reason, I would like to get a more precise understanding of the EIA’s assumption that discovered fields on the Arctic Refuge Coastal Plain are likely to be larger, on average, than their geologic counterparts already discovered on the North Slope.” In this regard, two questions are critical to understanding the high scenario results reported in EIA’s March 2004 analysis: (1) Except for the super-giant Prudhoe Bay, to date Alaska’s North Slope has only one producing field larger than 0.7 billion barrels (Kuparuk); in light of this experience, is it reasonable to assume that three out of seven fields discovered beneath the Arctic Refuge Coastal Plain will be larger than 0.7 billion barrels? (2) On what basis did EIA conclude that two fields would come in at 1.34 billion barrels – just 0.01 billion barrels below the level at which the 1991 algorithm would have required reduced peak and decline rates?
As we discussed, due to geologic and economic unknowns, a great deal of uncertainty inevitably attaches to any estimates of Arctic Refuge production potential. Nevertheless, policy makers require information that is policy-neutral as a starting point for consideration of issues related to the status of the Arctic Refuge Coastal Plain. To this end, resolution of the questions posed here is both important and necessary. For this reason, I look forward with keen interest to your response to the preceding questions regarding Arctic Refuge field size and production profile assumptions.
Richard A. Fineberg
 U.S. Department of Energy, Alaska Oil and Gas: Energy Wealth or Vanishing Opportunity? January 1991 (Report No. DOE/ID/01570-H1; in cooperation with the State of Alaska), p. 3-42.
 That the 1991 report underestimated actual North Slope production can be seen by comparing Table A-1 of that report (“Yearly Average Field Production Rates for Most Likely Cases [Values in parentheses are beyond the economic limit],” p. 3-96) to production history and forecast data, as reported by the State of Alaska at: Alaska Department of Natural Resources, Division of Oil and Gas 2004 Report, pp. 4-4 - 4-7 (Table IV.3, “Oil Production – Historic”) and the Alaska Department of Revenue, Revenue Sources Book, Fall 2004, p. 84 (“Historical and Projected Crude Oil Production”), supplemented by forecast data provided by Alaska Department of Revenue.
 My assumptions regarding factors such as peak production rates, rates of decline and stated field volumes were provided with my production profile worksheet January 26.
 Charles P. Thomas, et al., Alaska North Slope National Energy Strategy Initiative (Analysis of Five Undeveloped Fields), March 1992 draft, passim (final report published in May 1993 [U.S. Department of Energy, Office of Fossil Energy, Report No. DOE/ID/01570-T-164]).
 Emil D. Attanasi, “Economics of Undiscovered Oil in Federal Lands on the National Petroleum Reserve, Alaska,” January 2003, Table A-5 and p. 46 (U.S. Geological Survey Open-File Report 03-44).