Meeting the World's Energy NeedsJune 25, 2004
Figure 1. M. King Hubbert's original 1956 graph, with actual U.S. oil production for the years from 1956 to 2000 superimposed as small circles. The lower dashed curve gives Hubbert's estimate of U.S. oil production rates if the ultimate amount of discoverable oil is 150 billion barrels. The upper dashed line, for 200 billion barrels, is his famous prediction that U.S. oil production would peak in the early 1970s. Since 1985, the U.S. has produced slightly more oil than Hubbert predicted, largely because of successes in Alaska and in the far-offshore Gulf Coast. From Hubbert's Peak.
Out of Gas: The End of the Age of Oil. By David Goodstein, W.W. Norton, New York, 2004, 140 pages, $21.95.
Hubbert's Peak: The Impending World Oil Shortage. By Kenneth S. Deffeyes, Princeton University Press, Princeton, New Jersey, 2001, xii + 208 pages, $35.00 (cloth); 2003, 224 pages, $16.95 (paper).
Thermodynamicist David Goodstein is the vice provost and the Frank J. Gilloon Distinguished Teaching and Service Professor at Caltech. Petroleum geologist Kenneth Deffeyes spent many years in the oil industry before joining the Princeton faculty in 1967. Both are anxious to warn the public about the dwindling supply of fossil fuels, and to help concerned citizens decide how the gradual depletion of such fuels is likely to affect the world around them. While Deffeyes dwells mainly on oil and natural gas, Goodstein concentrates on alternatives to which we shall soon have to turn: nuclear, tidal, hydroelectric, geothermal, wind, solar, gasohol, and hydrogen fuel cell technologies. He sees little likelihood that any of the above, save nuclear energy, will be able to meet the world's energy needs.
Being well acquainted with Deffeyes's book, and others like it, Goodstein furnishes an annotated bibliography of relevant books, papers, and Web sites. He fails to mention his March 2002 op-ed column in The Los Angeles Times, in which he argued that retired petroleum geologists like Deffeyes---whom he mentioned by name in the column---have become the sole source of reliable estimates of oil and gas reserves; experts still active in the industry, he wrote, are discouraged by their employers from suggesting that recoverable hydrocarbons constitute anything less than an unlimited resource. It is by quoting authorities like Deffeyes that Goodstein is able to summarize a great deal of material in a mere 140 pages.
The first part of Deffeyes's book constitutes a primer on oil production. Beginning with photos of black and white ping-pong ball models of common hydrocarbon molecules, he explains how oil began to form almost as soon as life appeared on Earth and why source rocks, to contain useful quantities of "black gold,"need to have been buried in the "oil window" between 7500 and 15,000 feet below the Earth's surface at some time during their geologic histories.
Temperatures increase with depth underground, he writes, reaching the boiling point of water at about 15,000 feet. Below that depth, organic molecules embedded in sedimentary rock are broken down into the constituents of natural gas---of which methane is the most plentiful---and begin to diffuse away to parts unknown. Above 7500 feet, the organic matter embedded in such rocks does not break down into flowing oil. The oil shales of western Colorado are examples of carboniferous rocks that have never been buried deeply enough to contain flowing oil. Fuel can be extracted from them, by crushing and heating, which add greatly to pollution and the cost of production. Much of the flowing oil to be found in the oil window becomes essentially solid at the Earth's surface. Only at the temperatures and pressures found far below ground does it flow through porous rock and/or pool in wells.
Col. Edwin Drake's original oil well in Titusville, Pennsylvania, was only 75 feet deep, an indication that rock once buried 7500 feet below ground need not stay at that depth. It can be thrust up toward ground level, and overlying rock can be eroded until the oil within actually seeps to the surface, as at the La Brea tar pits near downtown Los Angeles, at Titusville, and at numerous places in the Middle East. Deffeyes observes that Kansas, together with a little bit of northern Oklahoma, constitutes the most thoroughly explored "oil province" in the world. There is absolutely no chance, he says, that a "supergiant" oil field remains to be discovered there, because there is no way to "gerrymander" one in between all the dry holes already drilled.
The fact that there are some thirty one-well oil fields in Kansas confirms that little is to be expected from further drilling there. Whenever a well becomes a producer, in Kansas or anyplace else, additional wells are quickly drilled to the north, south, east, and west of it. If all four come up dry, the field is uneconomic: One producing well won't pay for the drilling of four dry holes. Although the rest of the world is less well explored than Kansas, an undiscovered supergiant field could lie hidden in only a few remaining places. Much of the rock in the oil window is devoid of organic matter, and most of the rest has been drilled. Western Siberia and offshore sites in the South China Sea are among the few places left where large finds could realistically be expected. Jurisdictional disputes have long delayed the evaluation of prospects in the South China Sea.
It is hard to exaggerate the importance of the large finds that contain most of the world's oil. The North Sea, discovered at the end of World War II, is the world's youngest major oil and gas province; Norway is now the sixth largest oil producer in the world, and England the tenth. The North Sea reserves will apparently be exhausted in about half the time it takes to deplete an ordinary petroleum province, because the drilling targets in the area are relatively homogeneous-so that one exploratory method can be used to find all deposits-and because seismic techniques are particularly effective offshore.
The homogeneity of the deposits in the North Sea field contrasts with the diversity found in other fields, such as the one in West Texas, which actually consists of several different fields discovered at different times and by different means. None of them rival the East Texas field, discovered in 1930 by the legendary "Dad" Joiner, acting on a tip from a local veterinarian. By drilling in a spot that geologists of the day would never have recommended, he found the largest oil field (seven billion barrels) in the Lower 48.
Neither author has much use for economists. Goodstein criticizes them for insisting that there's no need to worry about running out of oil; the inevitable price increases, economists say, will render alternative fuels economically competitive, paving the way for oil's replacement by "something else." Deffeyes's criticism stems from economists' claim that a growing scarcity of oil will stimulate investment in oil exploration, which will bring about the discovery of more and more oil. He cautions that the flurry of exploration brought on by the OPEC embargos of 1973 and 1978 resulted in few new discoveries, and that more oil was discovered during the 1930s---a time when money for exploration was all but unavailable---than at any other time in history. Both regard reliance on "market magic" as irresponsible in a situation that calls for planning and foresight.
Goodstein complements Deffeyes's primer on "the oil patch" with a brief history of the energy concept. It is a task for which he is well suited, having taught freshman physics at Caltech for a number of years and co-authored several introductory books on the subject. He does so in the belief that energy issues are impossible to understand without a nodding acquaintance with the physical principles involved.
Goodstein is particularly interested in the greenhouse effect. Though personally convinced that global warming is under way, he does not disparage the minority opinion among climatologists that a new ice age is a more realistic threat. After referring several times to a Russian probe that found the surface temperature on Venus to be hot enough to melt lead, Goodstein concludes that "with the same solar flux, the Earth could be a ball of ice or a Venusian inferno." We have, at present, no reliable estimate of the effort required to tip the planet---perhaps quite suddenly---into one or the other of those two extremes.
Both authors emphatically reject official (U.S. Geological Survey) assurances that known fossil fuel reserves will allow present rates of consumption to continue for another 40 years in the case of oil, 60 years in the case of natural gas, and 100-200 years in the case of coal. For one thing, the four to five billion people who currently consume very little fossil fuel are sure to want more in coming years. For another, it won't be possible to produce oil and gas at anything like current rates for another forty to sixty years, because of a phenomenon known as "Hubbert's peak."
M. King Hubbert was a somewhat curmudgeonly geophysicist who, during the mid-1950s, annoyed his superiors at the Shell Oil Research Labs in Houston by predicting publicly that U.S. oil production would peak in or about 1970, and then begin an inexorable decline. After retiring from Shell, he went to work for USGS, where he continued to refine his estimates of global oil and gas reserves. In fact, U.S. production did peak in 1970, at about 9 million barrels a day, and has since declined more or less steadily, to the current level of about 6 million barrels a day.
Hubbert originally gave two estimates of the domestic production peak, based on two different estimates of the ultimate amount of recoverable oil. The graph on which he based his conclusions is shown in Figure 1. The two smooth broken curves are the results of fitting standard logistic functions to the production figures available in 1956, augmented by two separate estimates of the amounts ultimately recoverable by conventional means. For comparison, Deffeyes appended the small black dots representing the actual production history for 1956 through 2000 near the top of the figure. The total amount of oil in the ground---most of which will never be recovered---is at least three times greater than the amount produced, given that conventional means of production rarely recover more than 30% of the oil in any given field. Secondary and tertiary means of recovery have had only marginal impact on what qualifies as "ultimately recoverable."
Since 1985, the U.S. has produced slightly more oil than Hubbert predicted, largely because of successes in Alaska and in the far-offshore Gulf Coast. But those welcome developments have left the big picture remarkably unchanged: Hubbert is only a little less right than he would have been had neither development occurred. The Arctic National Wildlife Refuge---which eventually will be developed---will make even less difference. The oft-made suggestion that it could confer energy independence on the U.S. is simply irresponsible.
Deffeyes has made a heroic effort to improve on Hubbert's techniques while seeking to identify the global Hubbert's peak---the date after which worldwide oil production will begin to decline. In particular, he fits normal and Lorentzian functional forms to the data, as well as Hubbert's logistic ones, and uses least-absolute-value instead of least-squares estimation techniques. Goodstein cites four additional studies that apply similar methods to similar data in order to predict the same thing. All five agree that the peak will soon be attained, most probably within the present decade.
The details of the argument are open to criticism, on the ground that the best fits obtained are not particularly good fits. The reason becomes apparent on close examination of the actual production history illustrated in Figure 1 and even more so of the historic production of anthracite coal in Pennsylvania, shown in Figure 2.
In neither case does the curve appear to level off appreciably as it approaches its peak from the left, as any bell-shaped curve must. On the contrary, the data seem better explained by the pair of curves shown in Figure 3, in which the bell-shaped curve represents production capacity, while the steadily rising curve represents---at least until surplus production capacity ceases to exist---both sales and actual production. In the language of optimization, production capacity is just a constraint in the oil industry's profit-maximization problem. It does not become binding (or even accurately observable) until the resource becomes scarce.
The vertical separation between the two curves in Figure 3 represents the surplus production capacity at any given time. Initially small, it grows almost until peak production is attained, then plunges rapidly and permanently into negative territory. The peak is not predictable without knowledge of surplus capacity. The Texas Railroad Commission did have such knowledge in 1970, and did warn (in code) all who needed to know several months in advance of the U.S. peak. But unless OPEC proves equally candid and astute, the world may receive even less advance warning of the global peak.
Technical quibbles aside, Goodstein says that it is all but impossible to deny that oil, gas, and (probably) coal production rates in any one region---or in all regions combined---are likely to rise to a single Hubbert-like peak, and then dwindle (more or less steadily) to nothing. That unwelcome fact, Goodstein and Deffeyes agree, reduces to irrelevance the seemingly endless stream of com-placent official assurances that there is enough oil and gas in the ground to continue producing at current rates for another forty to sixty years. Long before then, as they have shown, it will be physically impossible to produce at anything like current rates from the played-out fields that remain.
The global Hubbert's peak is coming, whether in the present decade, the next, or possibly the one after that, and with it will come severe disruption in the lands of the unprepared. It will arrive not in forty to sixty years, as so many would have us believe, but as soon as the production capacity constraint becomes binding. Though the fossil fuel future is impossible to predict in detail, hard science leaves little doubt concerning its broad outlines. Whereas lasting and severe shortages are all but inevitable, the research and development programs that could at least cushion their impact are underfunded and behind schedule.
Figure 2. Annual production of anthracite coal in Pennsylania. Major deviations from the bell-shaped curve include a dip during the Great Depression (1928-1939) and a peak during World War II. From the AAPG and Hubbert's Peak.
Figure 3. Alternative explanation of production data for fossil fuels, with the bell-shaped curve representing production capacity and the steadily rising curve representing sales and actual production.
James Case writes from Baltimore, Maryland.