“For a successful technology, reality must take precedence over public relations, for Nature cannot be fooled.”

Richard Feynman

ENERGY AND RESOURCE QUALITY

Oil And Gas Availability: A History Of Federal Government Overestimation

[[ pp. 343-349 ] , ENERGY AND RESOURCE QUALITY,  by Charles A.S. Hall, Cutler J. Cleveland, Robert Kaufmann, Univ Pr Colorado, 1992

OIL AND GAS AVAILABILITY: A HISTORY OF FEDERAL GOVERNMENT OVERESTIMATION

The citizens of a democratic nation require relevant information to guide business investments, make political choices, and understand other economic and social processes. One function of government is to supply the citizenry and its political representatives with information necessary to make informed decisions, and a certain proportion of tax revenues is diverted toward that end. Generally, relatively little concern is expressed about the accuracy of the data collected and published by the U.S. government and the vast majority of economic and natural resource-related data are supplied by government agencies. Presumably most of these data are reasonably accurate. Various branches of the government also routinely provide us with predictions of economic conditions and resource supplies — for example, predictions of GNP, inflation rates, and energy availability. We would like to examine in detail what we believe is a very important subset of government predictions — those of the future availability of liquid and gaseous petroleum.

Predictions of future petroleum availability have been, and still are, very important with respect to influencing government policies toward energy pricing, energy imports and exports, conservation, synthetic fuel development, and other energy-related programs. Yet we believe that until very recently official government estimates of the future availability of petroleum have been grossly in error and, more important, have been received extremely uncritically despite the ongoing availability of much more explicit and more accurate alternative methods of assessment. We do not know why this state of affairs has existed beyond the observation that it is the desire of governmental entities to appear effective, and hence optimistic, although one might argue that for at least the first two-thirds of this century resource optimists have been more or less on target. The following is a more specific history of the official and unofficial estimates of U.S. petroleum reserves and the procedures used to make them.

METHODS OF ESTIMATING UNDISCOVERED PETROLEUM RESOURCES

There are three general methods for estimating yet to be discovered petroleum resources, although each method has many variations: First are extrapolations of historic discovery rates or performance patterns. Davis’ (1958) discovery rate curve and Hubbert’s (1962) growth curve projections and discovery rate curves (1967) are examples of this type of analysis. Second are volumetric yield methods which have been the most popular and widely used methods of assessment, particularly by the U.S. Geological Survey (USGS). Examples of this type of approach are the pioneering work by Weeks (1948) and various official USGS estimates (Zapp, 1962; Hendricks, I965; Miller et al., 1975). Third are combined methods that incorporate both subjective geological evaluations and statistical models. Hubbert (1974) and the most recent USGS estimate (Dolton et al., 1981) are examples of this method.

The two primary methods, extrapolation of past discovery rates and volumetric yield, often produce different estimates even when applied to the same region because different assumptions are employed by each. The discovery rate approach is based on the existing data and behavior of the industry and on the statistical extrapolation of past performances into the future. As such, this approach is most appropriate in mature areas of development and not in frontier areas where geology and economic conditions may be distinctly different from past experience. The volumetric yield approach is more subjectively based in that it involves extrapolating average yields (barrels of oil per cubic volume of sedimentary rock) from known areas to frontier areas that have been determined to be similar to the original area. Obviously, when using this method, there is often a range of opinions of not only what yield ratio is appropriate but also what constitutes a geologically similar region.

In 1956 Hubbert published an estimate that the ultimate recovery of crude oil from the lower 48 states would be about 150 billion bbl (Figure III.5). In 1962 Hubbert estimated about 170 billion bbl of oil, based on the extrapolation of past relations between production, discoveries, and proven reserves, a technique discussed in detail in Chapter 7. Hubbert’s analyses were based on the available data for the oil industry and were selected partly on the belief that the domestic oil and gas industry was in a more or less mature stage of development so that the extrapolation of past trends was a reasonable way to approach the problem of resource assessment.

During the same time period that Hubbert was publishing his 1962 and 1967 analyses, a series of official government estimates of future petroleum availability were released, primarily by the USGS (Zapp, 1961, 1962; Hendrick, 1965), which were many times higher than Hubbert’s projections (Figure III.5). The USGS method of assessment throughout the 1960s and early 1970s was primarily a form of a volumetric yield model developed by Zapp (1962). The so-called Zapp hypothesis is based on the assumption that since oil is discovered only through drilling, exploration for oil would not be complete until all potential oil-bearing regions had been drilled intensively enough to reach a well density that would leave virtually no fields undiscovered. Zapp and his colleagues estimated that this would require an overall density of one well per 2 mi2, drilled to either the basement of the sedimentary rock or 20,000 ft. Implicit in this approach is the important assumption that oil would continue to be found at a constant rate of about 118 bbl per foot of exploratory drilling, the mean rate up to that time. Hence, the validity of Zapp’s and the USGS estimates is dependent on the validity of the hypothesis that, on the average, the finding rate for oil would continue to remain relatively constant over time. Based on this model, Zapp estimated that about 590 billion bbl of crude oil would be produced in the United States. It is possible, however, that this high value should not be attributed to Zapp for he died at about the time these estimates were released and hence had no chance to revise or update his original analysis. Thus Zapp, a serious and scholarly scientist, may have been treated poorly by history because of the actions of others-something we may never know because of his untimely death. The Zapp hypothesis, with slight modifications, was the primary theoretical basis for all USGS estimates until the mid-1970s. The then assistant chief geologist (and later chief) for the USGS, Vincent E. McKelvey, stated in 1962 that “those who have studied Zapp’s method are much impressed with it, and we in the Geological Survey have much confidence in his estimates.”

A test of the validity of the Zapp hypothesis is to see if in fact oil is found at a constant rate of success per unit of drilling. This test is readily performed with the existing data base of the oil and gas industry. As we have seen already, Davis (1958), Hubbert (1967), and Hall and Cleveland (1981) clearly documented that oil is not discovered at a constant rate per unit drilling through time but rather exhibits a trend over time of decreasing returns per drilling effort (Figure 7.19). This undeniable fact illustrates clearly how the Zapp hypothesis was not even a reasonably good approximation of reality and led to unwarranted optimism for any resource estimate based on it (Figure III.6). The Zapp hypothesis does not take into account the important and inescapable trends in oil and gas exploration described in detail in Chapter 7, namely that the few very large fields that contain most of the recoverable petroleum are discovered first with relatively little drilling effort. As drilling effort accumulates, increasing levels of drilling are required to locate similar volumes of oil because the fields that are found are smaller.

Despite the undeniable observation that oil was being found at much less than the 118 bbl per foot implicit in Zapp’s analysis, and even the quite startling fact that Hubbert’s original 1956 prediction that U.S. production would peak around 1970 occurred as predicted, the Zapp-type method or other variations of the volumetric yield approach prevailed as the official U.S. estimate of petroleum reserves throughout the 1960s and early 1970s. In 1961 a USGS estimate based on Zapp’s methodology put the ultimate recovery of crude oil in the United States at almost 600 billion bbl (Figure III.5). Thirteen years later the USGS still estimated a minimum of about 310 billion bbl of ultimately producible crude oil. As recently as 1975, the USGS estimated between 250 and 300 billion bbl of oil (Miller et al., 1975). The irreconcilability of the estimate made in this Circular 725 with the existing data of the industry was noted by, among others, Hubbert (1978) and Nehring (1981)· At the conclusion of the most exhaustive empirical analysis of U.S. petroleum resources to date, Nehring (1981) stated:

“… we do not believe that anyone could develop a plausible list of geologic prospects.., in the lower 48 onshore containing an amount of petroleum even approaching the estimates of Circular 725 ….”

As Hubbert (1978) notes, the publication in 1975 of USGS Circular 725 was a “significant historical event,” despite the fact that it was still considerably higher than Hubbert’s estimate· Circular 725 marked the end of a 14-year period during which the succession of USGS estimates of the ultimate amounts of oil were two to three times higher than could be justified based on existing petroleum industry data on drilling, production, and discoveries. One reason for the dramatic downward revision was the increasing pressure being brought to bear on the unwarranted high USGS estimates of the 1950s, 1960s, and early 1970s brought about in part by the realities of poor drilling success of the 1970s. The new revised USGS estimates were part of the basis for President Carter’s new emphasis on the energy problem in general and the need for conservation in particular· Some cynics have suggested that the new low estimates followed Carter’s political needs for a national rallying point — an energy crisis — but we believe that reality had finally achieved political respectability and that President Carter had correctly assessed the situation·

In its most recent estimate the USGS’s 95 percent probability estimate (meaning there was 95% chance of the amount actually being greater) was 218 billion bbl of ultimately recoverable crude oil (Dolton et al., 1981; Table III.l). The 5% probability, or highest, estimate of the USGS group would lead to about 280 billion bbl of oil, whereas the mean estimate was about 260 billion bbl. The lower figure corresponds rather closely with Hubbert’s latest estimate of 213 billion bbl of oil from the lower 48 states, Alaska, and bordering continental shelves, and is consistent with what he has been saying since 1956 (Figure III.5). It is interesting to note that the approach used by Dolton et al. was a combination of exploration history, finding rate studies, petroleum geology, and volumetric yield procedures, and it was the first time that statistical finding rate methods were used explicitly in the USGS methodology, some 25 years after Hubbert had made his initial resource estimates. Meanwhile history continues to substantiate Hubbert’s analysis as one of the single most accurate and consistent major economic predictions ever made (Figure III.7).

The use of volumetric yield methods by USGS for its petroleum resource assessments isreally not surprising for they are consistent with the attitude that many resource analysts and politicians have held for many years — namely that human ingenuity, both technical and political, has no imaginable limits and that our ever-increasing technical cleverness at performing economic tasks will overcome any implications of increasing resource scarcity, even by providing more and more oil. This attitude is typified by the following statement made by V. E. McKelvey, former director of the USGS, in 1972 (italics added):

“Personally, I am confident that for millennia to come we can continue to develop the mineral supplies needed to maintain a high level of living for those who enjoy it now and to raise it for the impoverished people of our country and the world. My reasons for thinking so are that there is visible undeveloped potential of substantial proportions in each of the process by which we create resources and that our experience justifies the belief that these processes have dimensions beyond our knowledge and even beyond our imagination at any given time.”

Dr. McKelvey reiterated his cornucopian outlook toward future resource availability in a 1977 speech in Boston where he stated that natural gas reserves were so large that they amounted to “about 10 times the energy value of all previous oil, gas and coal reserves in the United States combined.” No one knows of course how much gas we will eventually find, but this attitude appears to us to be inconsistent with the peaking of gas production in 1973 and the fact that in 1985 proven natural gas reserves are but 12 years of current rates of use. As was described in Chapter 8, the outlook for future domestic gas availability is somewhat more promising than it is for crude oil. It is known that large amounts of gas exist in unconventional formations such as deep gas and geopressured gas. But the energy costs of developing these deeply buried and/or dilute natural gas sources must be evaluated before we include them in our estimates of future energy supplies. In the case of geopressured gas in the U.S. Gulf Coast region, preliminary evidence indicates that the average geopressured reservoir may have only a modest EROI, much less than conventional natural gas (Cleveland and Costanza, 1984).

Government attitudes toward future supplies of our natural resources affects all of us directly. Natural gas regulations from the mid-1950s to the late 1970s kept the price of gas artificially low and encouraged its irresponsible use and depletion. The points made in Chapters 2 and 3 suggest that these optimistic attitudes were adopted during a time of abundance of fossil fuels and other resources and perhaps were appropriate for the conditions that existed at that time. During times of decreasing availability of some important natural resources, however, existing attitudes and policies must be scrutinized closely to determine whether they still yield rational and logical policy relative to changing resource conditions. Government subsidies and price regulations cannot always compe0sate for changing resources realities, and in fact they can exacerbate the problem by giving false promises of future abundance of conventional petroleum.

Despite the increasing awareness of the trends in the quality of our energy resources, many individuals remain very optimistic about our ability to increase our domestic supply of conventional oil and gas. The attitude of the Reagan administration is that increased financial incentives to the oil industry will lead to more intensive searching for petroleum, and ultimately to substantially more oil and gas being found and produced. While this may prove to be true, it is, we believe, more likely that it will not be true at all. In fact virtually all evidence from the oil and gas industry explicitly contradicts this notion. The facts are that the energy crisis that followed the oil embargo in 1973 occurred when the official USGS estimate of that period was that we still had over 500 billion bbl of oil to be found in the United States. Why have we discovered only 11.8 billion bbl of new oil between 1974 and 1980, the equivalent of a little over two years’ use, despite unprecedented drilling? We believe that the reasons for this decline are simple and can be found in the physical nature of the resource itself, and not only in the economics of the industry that extracts it. Although it is true that allowing market prices to rise assists in discouraging consumption (an excellent goal, but not without its price in economic growth and material well-being), its effects on encouraging new discovery and production is a two-edged sword. Any present-day production simply hasten the inevitable point in time when the energy cost of domestic production equals the energy gains.

It is possible that large new oil reserves may be found in relatively untested provinces, such as off the coast of Alaska or even Iowa, or that large new gas deposits may be found in deep or unconventional beds, but there is little or no empirical evidence that indicates that the last 10 years of substantially greater price incentives and exploratory efforts is substantially increasing annual finds, especially of new oil. In Texas, for example, 90% of the rather modest 6 billion barrels of oil added to reserves from 1973 to 1984 came from existing fields, rather than from new fields (Oil and Gas Journal, July 22, 1985). In the meantime the much larger drilling and other exploration rates are greatly increasing energy costs. Most new provinces that remain will be extremely, perhaps critically, energyexpensive to develop. We should not decrease our search for new petroleum, but neither should we be deluded as to the costs of developing new oil deposits. Alternatively, if future exploratory efforts in frontier regions become more effective, we can reduce our pessimism. At this time we believe such optimism is not warranted, for the fact remains that most of the domestic oil we use in the United States in the 1970s and 1980s is from oil fields discovered before 1940. This reality is catching up with official government estimates. For example, based on the failure of the past 200 offshore wells to find significant quantities of oil the U.S. Dept. of Interior decreased its estimates of oil to be found there from 28 to 12 billion barrels, less than 5 years’ use (Norman, 1985).

Surely those who had been listening to Hubbert and Davis all along were neither shocked by the energy crisis nor by our large dependence on foreign sources of oil. On the other hand, the average citizen was quite unaware of Hubbert’s analysis and was dependent on official government estimates that we believe were outrageously optimistic. Reputable, although conservative, estimates from some oil industry analysts were discounted by many because they appeared (and may have been) self-serving to an industry trying to raise its domestic oil prices. Even now there is very little recognition or appreciation by the general populace of the precariousness of our petroleum situation and even less government assistance in understanding this problem or dealing with the inevitable future economic and social repercussions.

pp. 138-139, LIVING WITHIN LIMITS, Garrett Hardin; Oxford, 1993

Petroleum prophecy took a new turn in 1948, the 150th anniversary of Malthus’s essay. M. King Hubbert (1903-1989), a petroleum geologist employed at that time by Shell Oil, introduced a sophisticated new method of analysis. His method was based on effort per barrel the drilling effort expended per barrel of oil discovered and brought to the surface. The money price of oil will, “other things being equal,” increase with scarcity; but as long as people are willing to pay the price, oil companies have no reason to stop drilling. Ultimately, however, the energy price of obtaining oil will exceed the energy derivable from the product; beyond that point there is no rational defense for “producing” more fuel oil. Hubbert noted that the barrels of oil produced per unit effort required for the discovery of the reserves had been decreasing regularly for a long time. Projecting the curve into the future he predicted the “end of oil” for the United States and for the world. These terminal dates were much closer to hand than the ones assumed in the front office of the major oil companies. Understandably, Hubbert was Promptly labeled as the latest reincarnation of Chicken Little.

Hubbert persisted, extending and refining his methods during the next two decades. He predicted that in the early 1970s the price of petroleum would take a sharp turn upward. When the oil shock of 1973 came — the first oil shock we now call it — Hubbert was vindicated. We cannot ignore the fact that international politics played a role in producing the oil shock of 1973: price fixing by OPEC, the cartel of the major oil-exporting nations, touched it off. But the cartel could not have made its high prices stick in the absence of the relative shortage predicted by Hubbert.

The decade preceding the 1973 oil crisis was marked by sharp debates between the supporters of Hubbert (the “pessimists”) and his opponents, the “cornucopists” — who occupied positions of power in industry and government. While Hubbert estimated that the lifetime production of petroleum in the United States would be from 150 to 200 billion barrels, A. D. Zapp, of the U.S. Geological Survey, estimated 590 billion. The opponents were working with the same data.

A significant difference in their methods involved the estimate of oil found per foot of drilling in the future. Zapp assumed that the future would be like the past. This approach no doubt seemed conservative to many people, but it was not: Zapp was assuming that the extended future would be like the immediate past, which is a mere moment in time. A true conservative would use not the moment but the trend in constructing a telescope for looking into the future.

Hubbert was a conservative of the second sort. Extrapolating the trend of increasing effort that was apparent in the history of drilling, Hubbert concluded that the future will be worse than the past. Cornucopists of course called this attitude “pessimism.” Perhaps it is: but history has vindicated Hubbert. The yield per effort has gotten steadily worse. To expect otherwise would be to assert that petroleum geologists are incompetent. Faced with many possible drilling sites, a company geologist will advise his firm to drill the most hopeful ones first. If he is competent, the potential of the yet undrilled sites will diminish steadily with the passage of time. Productivity per foot drilled goes down, cost goes up. (If Zapp were right — if petroleum geologists were incompetent — then oil companies might as well save money by firing their geologists and choosing their drill sites by flipping coins.)

For a decade the influential director of the USGS “bought” Zapp’s estimate and opposed Hubbert. When the first oil shock came, two national committees (one within the USGS) were appointed to evaluate the situation. Both committees endorsed Hubbert. Finding his professional authority undermined, the director of the USGS resigned. Hubbert, for so long a “prophet without honor in his own country,” was fully vindicated.

Yet the biblical description of a “prophet without honor” is not entirely appropriate in this case. Pessimistic prophets and whistle-blowers are often given a hard time by their bosses. It is, therefore, a pleasure to report that the executives of Hubbert’s own company, the Shell Oil company, though not pleased with what he was saying, supported him during his “years in the wilderness.” In 1963 M. King Hubbert joined the faculty of Stanford University, from which he retired in 1968. After his victory in 1973 the “retired” prophet was in great demand as a speaker on the significance of physical resources for the survival of civilization.