The US military seemed to be acting as a global oil cop. Almost everywhere there were significant oil and gas pipelines or fields, one could expect to find a US base close by. American troops were stationed in 120 countries, now including most of the major oil producers (with the remaining notable exceptions of Iran and Russia). The US appeared to be the center of an empire of oil, facing an elusive enemy whose unpredictable attacks were capable of justifying intervention anywhere, at any time.

Was the price runup of 2004 in any way predictable? Was there some deeper reason for it? And could American officials' ability to foresee major oil price increases, with disastrous eventual impacts, have led them to undertake desperate measures?

The Ground Giving Way.

In nearly every year since 1859, the total amount of oil extracted from the world's ancient and finite underground reserves had grown - from a few thousand barrels a year to 65 million barrels per day by the end of the 20th century, an increase averaging about two percent per annum. Demand had grown just as dramatically, sometimes lagging behind the erratically expanding supply. The great oil crises of the 1970s - the most significant occasions when demand exceeded supply - had been politically-based interruptions in the delivery of crude that was otherwise readily available; there had been no actual physical shortage of the substance then, or at any other time.

In the latter part of the year 2000, as Al Gore and George W. Bush were crisscrossing the nation vying for votes and campaign contributions, the world price of oil rose dramatically from its low point of $10 per barrel in February 1999 to $35 per barrel by mid-September of 2000. Essentially, Venezuela and Mexico had convinced the other members of OPEC to cease cheating on production quotas, and this resulted in a partial closing of the global petroleum spigot. Yet while Saudi Arabia, Iraq, and Russia still had excess production capacity that could have been brought on line to keep prices down, most other oil-producing nations were pumping at, or nearly at, full capacity throughout this period.

Meanwhile, a wave of mergers had swept the industry. Exxon and Mobil had combined into Exxon-Mobil, the world's largest oil company; Chevron had merged with Texaco; Conoco had merged with Phillips; and BP had purchased Amoco-Arco. Small and medium-sized companies - such as Tosco, Valero, and Ultramar Diamond Shamrock Corp - also joined in the mania for mergers, buyouts, and downsizing. Nationally, oil-company mergers, acquisitions, and divestments totaled $82 billion in 1998 and over $50 billion in 1999.

Altogether, the oil industry appeared to be in a mode of consolidation, not one of expansion. As Goldman Sachs put it in an August 1999 report, "The oil companies are not going to keep rigs employed to drill dry holes. They know it but are unable ... to admit it. The great merger mania is nothing more than a scaling down of a dying industry in recognition that 90 percent of global conventional oil has already been found."4 Meanwhile the Energy Information Agency (EIA) predicted that global demand for oil would continue to grow, increasing 60 percent by the year 2020 to roughly 40 billion barrels per year, or nearly 120 million barrels per day.5 The dramatic price hikes of 2000 soon triggered a global economic recession. The link between energy prices and the economy was intuitively obvious and had been amply demonstrated by the oil crises and accompanying recessions of the 1970s. Yet, as late as mid-2000, many pundits were insisting that the new "information economy" of the 1990s was impervious to energy-price shocks. This trend of thought was typified in a comment by British Prime Minister Tony Blair, who in January 2000 stated that "[t]wenty years on from the oil shock of the '70s, most economists would agree that oil is no longer the most important commodity in the world economy. Now, that commodity is information."6 Yet when fuel prices soared in Britain during the last quarter of the year, truckers went on strike, bringing commerce within that nation to a virtual standstill. Though energy resources now directly accounted for only a small portion of economic activity in industrialized countries - 1.2 percent to 2 percent in the US - all manufacturing and transportation still required fuel. In fact, the entire economy in every industrial nation was completely dependent on the continuing availability of energy resources at low and stable prices.

As the world economy slowed, demand for new goods also slowed, and manufacturing and transportation were scaled back. As a result, demand for oil also decreased, falling roughly five percent in the ensuing year. Prices for crude began to soften. Indeed, by late 2001, oil prices had plummeted partly as the result of market-share compet.i.tion between Russia and Saudi Arabia. Gasoline prices at the pump in California had topped $2 in late 2000, but by early 2002 they had drifted to a mere $1.12 per gallon.

Such low prices tended to breed complacency. The Bush administration warned of future energy shortages, but proposed to solve the problem by promoting exploration and production within the US and by building more nuclear power plants - ideas that few with much knowledge of the energy industry took seriously. Now that gasoline prices were low again, not many citizens contemplated the possible future implications of the price run-ups of 2000. In contrast, industry insiders expressed growing concern that fundamental limits to oil production were within sight.

This concern gained public recognition in 2004, as oil prices again shot upward, this time attaining all-time highs of over $55 per barrel. National Geographic proclaimed in its cover story that this was "The End of Cheap Oil"; Le Monde announced "The Petro-Apocalypse;" while Paul Erdman, writing for the CBS television magazine Market.w.a.tch, proclaimed that "the looming oil crisis will dwarf 1973." In article after article, a.n.a.lysts pointed to dwindling discoveries of new oil, evaporating spare production capacity, and burgeoning global demand for crude. The upshot: world oil production might be near its all-time peak.

If this were indeed the case - that world petroleum production would soon no longer be able to keep up with demand - it would be the most important news item of the dawning century, dwarfing even the atrocities of September 11. Oil was what had made 20th-century industrialism possible; it was the crucial material that had given the US its economic and technological edge during the first two-thirds of the century, enabling it to become the world's superpower. If world production of oil could no longer expand, the global economy would be structurally imperiled. The implications were staggering.

There was every reason to a.s.sume that the Bush administration understood at least the essential outlines of the situation. Not only were many policy makers themselves - including the President, Vice President, and National Security Advisor - former oil industry executives; in addition, Vice President d.i.c.k Cheney's chief petroleum-futures guru, Matthew Simmons, had warned his clients of coming energy-supply crises repeatedly. Moreover, for many years the CIA had been monitoring global petroleum supplies; it had, for example, subscribed to the yearly report of Switzerland-based Petroconsultants, published at $35,000 per copy, and was thus surely also aware of another report, also supplied by Petroconsultants, t.i.tled "The World's Oil Supply 1995," which predicted that the peak of global oil production would occur during the first decade of the new century.

It would be an understatement to say that the general public was poorly prepared to understand this information or to appreciate its gravity. The New York Times had carried the stories of the oil company mergers on its front pages, but offered its readers little a.n.a.lysis of the state of the industry or that of the geological resources on which it depended. Ma.s.s-audience magazines Discover and Popular Science blandly noted, in buried paragraphs or sidebars, that "early in [the new century] ... half the world's known oil supply will have been used, and oil production will slide into permanent decline"7 and that "experts predict that production will peak in 2010, and then drop over subsequent years"8 - but these publications made no attempt to inform readers of the monumental implications of these statements. It would be safe to say that the average person had no clue whatever that the entire world was poised on the brink of an economic cataclysm that was as vast and unprecedented as it was inevitable.

Yet here and there were individuals who did perfectly comprehend the situation. Many were petroleum geologists who had spent their careers searching the globe for oil deposits, honing the theoretical and technical skills that enabled them to a.s.sess fairly accurately just how much oil was left in the ground, where it was located, and how easily it could be accessed.

What these people knew about the coming production peak - and how and when they arrived at this knowledge - const.i.tutes a story that centers on the work of one extraordinary scientist.

M. King Hubbert: Energy Visionary.

During the 1950s, '60s, and '70s, Marion King Hubbert became one of the best-known geophysicists in the world because of his disturbing prediction, first announced in 1949, that the fossil-fuel era would prove to be very brief.

Of course, the idea that oil would run out eventually was not, in itself, original. Indeed, in the 1920s many geologists had warned that world petroleum supplies would be exhausted in a matter of years. After all, the early wells in Pennsylvania had played out quickly; and extrapolating that initial experience to the limited reserves known in the first two decades of the century yielded an extremely pessimistic forecast for oil's future. However, the huge discoveries of the 1930s in east Texas and the Persian Gulf made such predictions laughable. Each year far more oil was being found than was being extracted. The doomsayers having been proven wrong, most people a.s.sociated with the industry came to a.s.sume that supply and demand could continue to increase far into the future, with no end in sight. Hubbert, armed with better data and methods, doggedly challenged that a.s.sumption.

M. King Hubbert had been born in 1903 in central Texas, the hub of world oil exploration during the early 20th century. After showing a childhood fascination with steam engines and telephones, he settled on a career in science. He earned BS, MS, and Ph.D. degrees at the University of Chicago and, during the 1930s, taught geophysics at Columbia University. In the summer months, he worked for the Amerada Petroleum Corporation in Oklahoma, the Illinois State Geological Survey, and the United States Geological Survey (USGS). In 1943, after serving as a senior a.n.a.lyst at the Board of Economic Warfare in Washington, DC, Hubbert joined Sh.e.l.l Oil Company in Houston, where he directed the Sh.e.l.l research laboratory. He retired from Sh.e.l.l in 1964, then joined the USGS as a senior research geophysicist, a position he held until 1976. In his later years, he also taught occasionally at Stanford University, the University of California at Los Angeles, the University of California at Berkeley, the Ma.s.sachusetts Inst.i.tute of Technology, and Johns Hopkins University.

During his career, Hubbert made many important contributions to geophysics. In 1937 he resolved a standing paradox regarding the apparent strength of rocks that form the Earth's crust. Despite their evident properties of hardness and brittleness, such rocks often show signs of plastic flow. Hubbert demonstrated mathematically that, because even the hardest of rocks are subject to immense pressures at depth, they can respond in a manner similar to soft muds or clays. In the early 1950s, he showed that underground fluids can become entrapped under circ.u.mstances previously not thought possible, a finding that resulted in the redesign of techniques employed to locate oil and natural gas deposits. And by 1959, in collaboration with USGS geologist William W. Rubey, Hubbert also explained some puzzling characteristics of overthrust faults - low-angle fractures in rock formations in which one surface is displaced relative to another by a distance on the order of kilometers.

These scientific achievements would have been sufficient to a.s.sure Hubbert a prominent place in the history of geology. However, his greatest recognition came from his studies of petroleum and natural gas reserves - studies he had begun in 1926 while a student at the University of Chicago. In 1949, he used statistical and physical methods to calculate total world oil and natural gas supplies and doc.u.mented their sharply increasing consumption. Then, in 1956, on the basis of his reserve estimates and his study of the lifetime production profile of typical oil reservoirs, he predicted that the peak of crude-oil production in the United States would occur between 1966 and 1972. At the time, most economists, oil companies, and government agencies (including the USGS) dismissed the prediction. The actual peak of US oil production occurred in 1970, though this was not apparent until 1971.9 Let us trace just how Hubbert arrived at his prediction. First, he noted that production from a typical reservoir or province does not begin, increase to some stable level, continue at that level for a long period, and then suddenly drop off to nothing after all of the oil is gone. Rather, production tends to follow a bell-shaped curve. The first exploratory well that punctures a reservoir is capable of extracting only a limited amount; but once the reservoir has been mapped, more wells can be drilled.

Figure 5. US oil production, history and projection, including lower 48, Alaska and Gulf of Mexico (deep water). Source: ASPO.

During this early phase, production increases rapidly as the easiest-accessed oil is drained first. However, beyond a certain point, whatever remains is harder to get at. Production begins to decline, even if more wells are still being drilled. Typically, the production peak will occur when about half of the total oil in the reservoir has been extracted. Even after production has tapered off, some oil will still be left in the ground: it is economically impractical - and physically impossible - to remove every last drop. Indeed, for some reservoirs only a few percent of the existing oil may be recoverable (the average is between 30 and 50 percent).

Hubbert also examined the history of discovery in the lower-48 United States. More oil had been found in the 1930s than in any decade before or since - and this despite the fact that investment in exploration had increased dramatically in succeeding decades. Thus discovery also appeared to follow a bell-shaped curve. Once the history of discovery had been charted, Hubbert was able to estimate the total ultimately recoverable reserves (URR) for the entire lower-48 region. He arrived at two figures: the most pessimistic reasonable amount (150 billion barrels) and the most optimistic reasonable amount (200 billion barrels). Using these two estimates, he calculated future production rates. If the total URR in the lower-48 US amounted to 150 billion barrels, half would be gone - and production would peak - in 1966; if the figure were closer to 200 billion barrels, the peak would come in 1972.

These early calculations involved a certain amount of guesswork. For example, Hubbert chose to chart production rates on a logistic curve, whereas he might have employed a better-fitting Gaussian curve.10 Even today, according to Princeton University geophysicist Kenneth S. Deffeyes, author of Hubbert's Peak: The Impending World Oil Shortage, the "numerical methods that Hubbert used to make his prediction are not crystal clear."11 Despite many conversations with Hubbert and ensuing years spent attempting to reconstruct those original calculations, Deffeyes finds aspects of Hubbert's process obscure and "messy." Nevertheless, Hubbert did succeed in obtaining important, useful findings.

Following his prediction of the US production peak, Hubbert devoted his efforts to forecasting the global production peak. With the figures then available for the likely total recoverable world petroleum reserves, he estimated that the peak would come between the years 1990 and 2000. This forecast would prove too pessimistic, partly because of inadequate data and partly because of minor flaws in Hubbert's method. Nevertheless, as we will see shortly, other researchers would later refine both input data and method in order to arrive at more reliable predictions - ones that would vary only about a decade from Hubbert's.

Hubbert immediately grasped the vast economic and social implications of this information. He understood the role of fossil fuels in the creation of the modern industrial world, and thus foresaw the wrenching transition that would likely occur following the peak in global extraction rates. In lectures and articles, starting in the 1950s, Hubbert outlined how society needed to change in order to prepare for a post-petroleum regime. The following pa.s.sage, part of a summary by Hubbert of one of his own lectures, conveys some of the breadth and flavor of his macrosocial thinking: The world's present industrial civilization is handicapped by the coexistence of two universal, overlapping, and incompatible intellectual systems: the acc.u.mulated knowledge of the last four centuries of the properties and interrelationships of matter and energy; and the a.s.sociated monetary culture which has evolved from folkways of prehistoric origin.

The first of these two systems has been responsible for the spectacular rise, princ.i.p.ally during the last two centuries, of the present industrial system and is essential for its continuance. The second, an inheritance from the prescientific past, operates by rules of its own having little in common with those of the matter-energy system. Nevertheless, the monetary system, by means of a loose coupling, exercises a general control over the matter-energy system upon which it is superimposed.

Figure 6. M. King Hubbert's projected cycles for world crude production for the extreme values of the estimated total resource. (Source: M. K. Hubbert, Resources and Man) Despite their inherent incompatibilities, these two systems during the last two centuries have had one fundamental characteristic in common, namely exponential growth, which has made a reasonably stable coexistence possible. But, for various reasons, it is impossible for the matter-energy system to sustain exponential growth for more than a few tens of doublings, and this phase is by now almost over. The monetary system has no such constraints, and, according to one of its most fundamental rules, it must continue to grow by compound interest.12 Hubbert thus believed that society, if it is to avoid chaos during the energy decline, must give up its antiquated, debt-and-interest-based monetary system and adopt a system of accounts based on matter-energy - an inherently ecological system that would acknowledge the finite nature of essential resources.

Hubbert was quoted as saying that we are in a "crisis in the evolution of human society. It's unique to both human and geologic history. It has never happened before and it can't possibly happen again. You can only use oil once. You can only use metals once. Soon all the oil is going to be burned and all the metals mined and scattered."13 Statements like this one gave Hubbert the popular image of a doomsayer. Yet he was not a pessimist; indeed, on occasion he could a.s.sume the role of utopian seer. We have, he believed, the necessary know-how; all we need do is overhaul our culture and find an alternative to money. If society were to develop solar-energy technologies, reduce its population and its demands on resources, and develop a steady-state economy to replace the present one based on unending growth, our species' future could be rosy indeed. "We are not starting from zero," he emphasized. "We have an enormous amount of existing technical knowledge. It's just a matter of putting it all together. We still have great flexibility but our maneuverability will diminish with time."14 Reading Hubbert's few published works - for example, his statement before the House of Representatives Subcommittee on the Environment on June 6, 1974 - one is struck by his ability to follow the implications of his findings on oil depletion through the domains of economics and ecology.15 He was a holistic and interdisciplinary thinker who deserves, if anyone does, to be called a prophet of the coming era.

Hubbert died in 1989, a few years before his predicted date for the global production peak. That all-important forecast date was incorrect, as the rate of world oil production continued to increase through the first months of 2005. But by how far did he miss the mark? It would be up to his followers to find out.

Hubbert's Legacy.

Since Hubbert's death, several other prominent petroleum geologists have used their own versions of his method to make updated predictions of the world's oil production peak. Their results diverge only narrowly from one another's. Since these scientists have been able to maintain updated data on reserves and production rates and since their work figures prominently in the current discussion about petroleum depletion, it will be helpful to introduce some of these individuals.

Colin J. Campbell is by most accounts the dean among Hubbert's followers. After earning his Ph.D. at Oxford in 1957, Campbell worked first for Texaco and then Amoco as an exploration geologist, his career taking him to Borneo, Trinidad, Colombia, Australia, Papua New Guinea, the US, Ecuador, the United Kingdom, Ireland, and Norway. He later was a.s.sociated with Petroconsultants in Geneva, Switzerland, and in 2001 brought about the creation of the a.s.sociation for the Study of Peak Oil (ASPO), which has members affiliated with universities in Europe. He has published extensively on the subject of petroleum depletion, and is author of the book The Coming Oil Crisis.16 Campbell's most prominent and influential publication was the article "The End of Cheap Oil?", which appeared in the March 1998 issue of Scientific American. The co-author of that article, Jean Laherrere, had worked for the oil company Total (now Total Fina Elf) for thirty-seven years in a variety of roles encompa.s.sing exploration activities in the Sahara, Australia, Canada, and Paris. Like Campbell, Laherrere had also been a.s.sociated with Petroconsultants in Geneva.

The Scientific American article's most arresting features were its sobering t.i.tle and its conclusion: From an economic perspective, when the world runs completely out of oil is ... not directly relevant: what matters is when production begins to taper off. Beyond that point, prices will rise unless demand declines commensurately. Using several different techniques to estimate the current reserves of conventional oil and the amount still left to be discovered, we conclude that the decline will begin before 2010.17 Figure 7a. World oil production, history and projection. (Source: ASPO).

Figure 7b. Estimated world oil production to 2100. (Source: ASPO).

From the standpoint of the article's contribution to advancing the discussion beyond Hubbert's initial projections, its explanation of the methods and problems of estimating the world URR deserves treatment here. Many oil a.n.a.lysts have discounted warnings from Hubbert and his followers because official figures suggest that world oil reserves have grown substantially over the past 20 years. Campbell and Laherrere point out that such figures contain systematic errors arising from the fact that OPEC countries are often motivated to inflate reserve figures because the higher their reserves, the more oil they are allowed to export.

"There is thus good reason to suspect that when, during the late 1980s, six of the 11 OPEC nations increased their reserve figures by colossal amounts, ranging from 42 to 197 percent, they did so only to boost their export quotas," according to Campbell and Laherrere, who call such reserve growth "an illusion." They note that: about 80 percent of the oil produced today flows from fields that were found before 1973, and the great majority of them are declining.

In the 1990s oil companies have discovered an average of seven Gbo [billion barrels of oil]; last year they drained three times that much. Yet official figures indicated that proved reserves did not fall by 16 Gbo, as one would expect; rather, they expanded by 11 Gbo. One reason is that several dozen governments opted not to report declines in their reserves, perhaps to enhance their political cachet and their ability to obtain loans. A more important cause of the expansion lies in revisions: oil companies replaced earlier estimates of the reserves left in many fields with higher numbers. For most purposes, such amendments are harmless, but they seriously distort forecasts extrapolated from published reports.

Figure 8. Published estimates of global ultimately recoverable oil, in trillions of barrels (Source: C. J. Campbell) Campbell and Laherrere suggest that one way to avoid such distortions is to backdate every revision to the year in which the field in question was first discovered. When this is done, it becomes apparent that global oil discovery peaked in the early 1960s and has been falling ever since. If that trend in discovery is extrapolated, it is possible to make a good guess at how much oil will ultimately be found. Even if this guess is off by two or three hundred billion barrels, the error will not affect the timing of the production peak by more than a few years.

The authors also discussed "nonconventional" oil - including heavy oil in Venezuela and oil sands in Canada - of which vast quant.i.ties are known to exist. "Theoretically," they write, "these unconventional oil reserves could quench the world's thirst for liquid fuels as conventional oil pa.s.ses its prime. But the industry will be hard-pressed for the time and money needed to ramp up production of unconventional oil quickly enough." (Later in this chapter we will examine some other problems with nonconventional petroleum resources.) Figure 9. Dubious reserve revisions by OPEC countries in 1986 and 1987, in billions of barrels (Source: C. J. Campbell) Campbell is currently predicting the peak of global production of conventional oil to occur by about 2008.18 Kenneth S. Deffeyes (whom we quoted earlier in this chapter), in his book Hubbert's Peak: The Impending World Oil Shortage (2001), discusses the work of a petroleum geologist in layman's terms. The reader learns how oil was formed, where it is likely to be found, and what techniques and machinery geologists use to find it. Deffeyes also devotes two chapters to a detailed a.n.a.lysis of Hubbert's predictive method, offering mathematical refinements that yield more accurate forecasts. At the conclusion of the first of those chapters he writes: The resulting estimate gives a peak production year of 2003 and a total eventual oil recovery of 2.12 trillion barrels. The peak year, 2003, is the same year that we got by fitting [Colin] Campbell's 1.8-trillion-barrel estimate to the production history. Other published estimates, using variations on Hubbert's methods, give peak years from 2004 to 2009. I honestly do not have an opinion as to the exact date for two reasons: (1) the revisions of OPEC reserves may or may not reflect reality; (2) OPEC production capacities are closely guarded secrets .... This much is certain: no initiative put in place starting today can have a substantial effect on the peak production year. No Caspian Sea exploration, no drilling in the South China Sea, no SUV replacements, no renewable energy projects can be brought on at a sufficient rate to avoid a bidding war for the remaining oil. At least, let's hope that the war is waged with cash instead of with nuclear warheads.19 The late L. F. Ivanhoe was the founder of the M. King Hubbert Center for Petroleum Supply Studies at the Colorado School of Mines, whose mission is to a.s.semble, study, and disseminate global petroleum supply data. He was a registered geologist, geophysicist, engineer, and oceanographer with 50 years of domestic and international experience in petroleum exploration with various private and government oil companies. He was a.s.sociated first with Chevron and then with Occidental Petroleum, where he was senior advisor of worldwide evaluations of petroleum basins from 197480. Ivanhoe was the author of many papers on technical subjects, including roughly 50 on the evaluation of foreign prospective basins and the projection of future global oil supplies.

Ivanhoe called Hubbert's followers "Ca.s.sandras," after the mythological Trojan princess who could foretell the future but was doomed never to be believed.

In 1997, in a paper ent.i.tled "King Hubbert - Updated," Ivanhoe presented the following scenario: Hubbert wrote virtually nothing about details of the "decline side" of his Hubbert Curve, except to mention that the ultimate shape of the decline side would depend upon the facts and not on any a.s.sumptions or formulae. The decline side does not have to be symmetrical to the ascending side of the curve - it is just easier to draw it as such, but no rules apply. The ascending curve depends on the skill/luck of the explorationists while the descending side may fall off more rapidly due to the public's acquired taste for petroleum products - or more slowly due to government controls to reduce consumption .... 20 In his summary at the end of that paper, Ivanhoe concluded that the critical date ... when global oil demand will exceed the world's production will fall somewhere between 20002010, and may occur very suddenly due to unpredictable political events .... This foreseeable energy crisis will affect everyone on earth.

Walter Youngquist, retired Professor of Geology at the University of Oregon, is the author of Geodestinies: The Inevitable Control of Earth Resources over Nations and Individuals (1997). During his career, he led or partic.i.p.ated in on-the-ground geological studies in the US and abroad, and studied populations and resources in 70 countries. In his book, Youngquist discusses the important concept of net energy. He writes: All this energy expended in thousands of ways used to finally discover oil and produce it has to be added up and compared with the amount of energy in the oil which these efforts produce. This ratio - of energy produced compared to the energy used - is the all-important energy/ profit ratio. As we have to drill deeper to find oil, and as we have to move into more difficult and expensive areas in which to operate, the ratio of [energy] profit to energy expended declines. Already, in some situations energy in the oil found is not equal to the total energy expended. Also, although some wells flow initially, all wells eventually have to be pumped. Pumping oil is expensive, particularly if it is being pumped from a considerable depth. It takes energy to move steel pumping rods up and down, in some cases as much as three miles of them ....

The most significant trend in the US oil industry has been the decline in the amount of energy recovered compared to energy expended. In 1916 the ratio was about 28 to 1, a very handsome energy return. By 1985, the ratio had dropped to 2 to 1, and is still dropping. The Complex Research Center at the University of New Hampshire made a study of this trend and concluded that, by 2005 at the latest, it will take more energy, on the average, in the United States to explore for, and drill for, and produce oil from the wells than the wells will produce in energy.21 Figure 10. Yield per effort (YPE) for the US lower 48 states (onsh.o.r.e plus offsh.o.r.e). Yield per effort is the ratio of total annual additions to proved oil reserves to total oil footage drilled. The dark circles are actual observations of YPE. predicted by a regression that predicts YPE as a function of c.u.mulative drilling effort alone. The solid line is YPE predicted by the regression that predicts YPE as a function of c.u.mulative driling, the rate of drilling, revisions as a fraction of total additions, and new field discoveries as a fraction of the sum of new field discoveries, new discoveries in old fields, and extensions. (Source: Oil a.n.a.lytics) Figure 11. Costs per well of oil and gas wells drilled in the US, 19602000 (Source: US Energy Information Administration.) About the end of oil production, Youngquist has this to say:.

Most likely the end of the Petroleum Interval will be gradual wherein no crisis point is reached, just slow change. But, especially with continually rising populations, and no sufficient subst.i.tutes for oil at hand, there is the possibility of a chaotic breakdown of society.22 Some of the essential elements of Hubbert's message have been taken up by others who are not petroleum geologists. A prominent example is Matthew Simmons, the founder of Simmons & Company International, an independent investment bank specializing in the energy industry. Simmons, has highlighted the reality and significance of petroleum depletion in many of his writings and public presentations.

Figure 12. In recent years, the cost of exploration for oil has been exceeding the net present value of the discoveries in absolute terms. In simple terms, these days it usually costs more to explore for oil than consequent oil discoveries warrant. This trend appears to be accelerating. (Source: ASPO) Simmons describes himself as a lifelong Republican with 30 years of experience in investment banking. In a lecture called "Digging Out of Our Energy Mess," delivered to the American a.s.sociation of Petroleum Geologists on June 5, 2001, Simmons noted: Figure 13a. The growing gap. (Source: ASPO).

Figure 13b. Additions to world proven reserves from the discovery of new fields, and production. (Source: World Energy Outlook 2004 [Figure 3.14]) A simple check of the facts quickly reveals that almost every sc.r.a.p of spare energy [production] capacity around the globe is now either gone or just about to disappear ....

Even the Middle East is now beginning to experience, for the first time ever, how hard it is to grow production once giant fields roll over and begin to decline. There is so little data on field-by-field production statistics in the Middle East that any guesses on average decline rates are simply speculation. But there is growing evidence that almost every giant field in the Middle East has already pa.s.sed its peak production.

It is also interesting to see how few truly giant oil and gas fields have been discovered over the past 40 years .... Even the newest giant field in the Middle East, Saudi's Shaybah field, was discovered in the 1970s, though it only began production two years ago.23 Simmons went on to point out that two of the world's other giant fields - Alaska's Prudhoe Bay and Western Siberia's Samatlor field, both discovered in 1967 - peaked in the 1970s, Prudhoe at 1.5 million barrels per day and Samatlor at over 3.5 million barrels per day: Today, Prudhoe Bay struggles to stay around 500,000 barrels per day while Samatlor's production averaged just under 300,000 barrels per day in 2000. To think that two giant fields which collectively topped 5 million barrels per day 12 years ago could now be down to 800,000 barrels per day is a staggering example of the power of the decline curve ....

Could the days of 1 million barrel a day or greater oil and gas fields be over? There are over 140 oil and gas fields under development through the end of 2005. Only a handful of these projects have plans to produce over 200,000 barrels per day when they peak and none are expected to exceed 250,000 barrels per day.

Matt Simmons has recently auth.o.r.ed a book, Twilight in the Desert, in which he a.n.a.lyzes over 200 technical papers by geologists working the Saudi fields, and concludes that Saudi Arabia's oil production may already be near or at its all-time peak. Since virtually all of the world's spare production capacity is in Saudi Arabia, that country's peaking date will also signal the global peak.

Zeroing in on the Date of the Peak.

As a growing community of scientists applies itself to the task of determining exactly when the world's oil production will begin to decline, four princ.i.p.al methods for doing so are emerging.

1. Estimate the total ultimately recoverable resource (URR) and calculate when half will have been extracted. This is the original method developed and used by Hubbert himself, beginning in the 1950s. As we have seen, Hubbert noted that for a typical oil-producing region, a graph of extraction over time tends to take on a bell-shaped curve: the more cheaply and easily accessed portion of the resource is depleted first, so that when about half is gone the rate at which extraction can proceed tends to diminish.

The method is relatively simple, and it worked well for predicting the US production peak. However, it relies on the availability of accurate discovery and reserve data so that URR can be accurately estimated, and on accurate extraction data as well. For the US, these figures were not problematic, but for some other important oil-producing regions this is not the case: reserve data for Saudi Arabia, for example, are controversial.

Moreover, there is no natural law stating that the extraction curve must be precisely bell-shaped or symmetrical. Indeed, political, economic, and technological factors can deform the curve in an infinite number of ways. In reality, the actual production curves from producing nations never conform to a simple mathematical curve, and are characterized by b.u.mps, plateaus, valleys, and peaks of differing sizes and durations. A war, a recession, the application of a new recovery technology, or the decision by a government to restrain extraction can reshape the curve arbitrarily.

Nevertheless, what comes up must come down: for every oil-producing region, extraction starts at zero, increases to a maximum, and declines, regardless of the tortuous b.u.mps in between. And, in general, the latter half of the resource will require more effort and time to extract than the first half. Moreover, experience shows that if actual production strays far from the predicted curve because of political or economic factors, it will tend to return to it once the influence of those factors subsides.

The Hubbert model is therefore a good way of providing a first-order approximation: it gives us a general overview of the depletion process, and (depending on the accuracy of the available reserve and production data) yields a likely peak year, with a window of uncertainty. However, it cannot be used to forecast actual production for the next month or even the next decade.

Using this method (with various refinements), Kenneth Deffeyes, in his book Hubbert's Curve, arrives at a peak date of 2005. Other researchers, such as Jean Laherrere, using more optimistic reserve estimates, place the peak further out, up to 2020.

2. Count the number of years from peak of discovery. Hubbert realized that, when graphed over time, the discovery of oil within any given region tends to peak and decline, just as production does. Understandably, discovery always peaks first - since it is necessary to find oil before it can be extracted.

In the US, discoveries of oil peaked in the early 1930s with the stupendous finds in east Texas; production peaked almost exactly 40 years later. Are we likely to find a similar time lag for other oil producing regions? If so, this could provide a basis for predicting the timing of the global production peak.

The duration of the lag between discovery and production peaks depends on a number of factors: the geological conditions (some fields can be depleted more quickly than others); the extraction technology being used (new recovery methods can deplete a reservoir more quickly and efficiently, but can also increase the total amount recoverable and thus extend the life of the field); and whether the resource is being extracted at maximum possible rate (as already noted, economic or political events can intervene to reduce production rates).

The North Sea provides an example of a relatively brief lag between discovery and extraction peaks: there, discoveries peaked in the early 1970s, while production peaked only 30 years later, at the turn of the new century. The latest exploration and extraction technologies were applied, and the resource base was drawn down at virtually the maximum possible rate because North Sea oil was in high demand throughout this period.

Iraq provides a counterexample: there, two princ.i.p.al periods of major discovery occurred - in the early 1950s and the mid-1970s. For that country, political and economic events have constrained production to a very significant degree: first, the Iran-Iraq war of the 1980s, then the US-led embargo of the 1990s, and finally the turmoil surrounding the US invasion and occupation have reduced extraction rates well below levels that would otherwise have been achieved. Consequently, Iraqi oil production may not peak until 2015 at the earliest, though more likely a decade or so later, yielding a discovery-to-production-peak lag of 45 to 60 years.

Will other discovery-to-production-peak lag times tend to more closely match those of the North Sea countries, or that of Iraq? Chances are that, as individual anomalies cancel each other out, lag times are on average likely to cl.u.s.ter around that of the US - roughly 40 years.

The US Department of Energy Discusses Peak Oil.

When The Party's Over was originally published, no official US or international agency had formally acknowledged the likelihood that global oil production will begin its historic decline within the next few years. That situation has recently changed. In March 2004, the Department of Energy published a little-heralded doc.u.ment on the strategic importance of oil shale; roughly a quarter of the 45-page report is devoted to the subject of oil depletion and its likely consequences. Here are just a few excerpts: [World] Discoveries did peak before the 1970s, as shown in Figure 6. This figure also shows that no major new field discoveries have been made in decades. Presently, world oil reserves are being depleted three times as fast as they are being discovered . . . .

The disparity between increasing production and declining reserves can have only one outcome: a practical supply limit will be reached and future supply to meet conventional oil demand will not be available. The question is when peak production will occur and what will be its ramifications. Whether the peak occurs sooner or later is a matter of relative urgency . . . .

In spite of projections for growth of non-OPEC supply, it appears that non-OPEC and non-Former Soviet Union countries have peaked and are currently declining. The production cycle of countries . . . and the c.u.mulative quant.i.ties produced reasonably follow Hubbert's model. . . . Although there is no agreement about the date that world oil production will peak, forecasts presented by USGS geologist Thomas Magoon, the OGJ [Oil & Gas Journal], and others expect the peak will occur between 2003 and 2020 . . . . What is notable about these predictions is that none extend beyond the year 2020 . . . . [pp. 7-8]

The Nation must start now to respond to peaking global oil production to offset adverse economic and national security impacts. [p. 26]

(Source: "Strategic Significance of America's Shale Oil Resource," Vol. 1, "a.s.sessment of Strategic Issues," Office of Deputy a.s.sistant Secretary for Petroleum Reserves, Office of Naval Petroleum and Oil Shale Reserves, U.S. Department of Energy, March 2004 ) Global oil discoveries peaked in 1963. This is not a controversial fact: both the oil industry and the US Department of Energy acknowledge that this is the case.Given this, we might expect that the global peak in the rate of oil extraction would occur roughly 40 years later - i.e., in 2003.However, we must take into account intervening economic and political events that might have tended to reduce extraction rates below their potential, and thus increase the time lag. The princ.i.p.al such events were the Arab OPEC embargo of the early 1970s, the fall of the Shah of Iran, and the subsequent Iran-Iraq war a few years later. The consequent oil price spikes reduced demand for oil, and led to a decline in extraction rates. The effect may have been to add up to ten years to the global discovery-to-production-peak time lag, yielding a likely peak date window of 2005 to 2013.3. Track the reserve and production data of individual countries. For the past few years, both Colin Campbell (a.s.sociation for the Study of Peak Oil) and Richard C. Duncan (Inst.i.tute on Energy and Man) have been keeping close track of production data for individual producing nations.Campbell's detailed discussions of oil statistics nation by nation are available in the archived newsletters of the a.s.sociation for the Study of Peak Oil (www.asponews.org), and in his book, The Essence of Oil & Gas Depletion (MultiScience, 2003).Duncan uses a "graphical-heuristic-iterative (GHI)" method to forecast world oil production, repeating the entire modeling and forecasting process annually to give a series of consistent but unique world oil forecasts. According to Duncan, heuristic means "a method of computer programming in which the modeler and machine proceed along empirical lines, using data, other information, and rules of thumb to find solutions or answers." Iterative means "repet.i.tious; repeating or repeated." The Graphical Input Device (used in system dynamics programs such as Stella) enables the modeler "to quickly create and/or edit an oil production forecast of a nation just before each trial run (iteration) of the model." The Scatter Graph (system dynamics) is used to depict "the forecasted peak year of oil production (x-axis) versus the forecasted peak production rate (y-axis) of our ongoing series of world oil forecasts." Duncan describes this as a work in progress "that will eventually converge on Peak Oil - whether the Peak is near at hand or far in the future."24 Many countries are now clearly past their individual all-time extraction peaks. The list includes not only the US, but also Indonesia, Gabon, Great Britain, and Norway. Altogether, according to Duncan, of 45 significant oil-producing countries, 25 are past-peak (BP Statistical Review of World Energy currently estimates the latter number at 18, indicating that there is some uncertainty on this point, but also that oil companies are keenly aware of the peaking phenomenon and are keeping score).25 Some of the pre-peak nations are major producers with huge reserves (e.g., Iraq and Saudi Arabia). Thus it would be unwise to a.s.sume that the global peak will occur when exactly half of all producing nations have undergone their individual peaks. Clearly, more complex calculations are necessary, and this is the work that Duncan and Campbell are undertaking.The countries in decline account for about 30 percent of the world's total oil production. Further, according to Oil & Gas Journal, as demand for oil expanded and prices rose during 2004, all of the added supply came from Russia and a few OPEC nations. Evidently, all of the nations outside of Russia and OPEC, when taken together, have already peaked in production (though there are individual exceptions, such as Brazil).By examining the geology, history, and economic-political circ.u.mstances of each oil-producing country, it is possible to encircle the remaining uncertainties and pick away at them. How much oil has been discovered in each given nation? How long ago did discoveries peak? Are significant future discoveries likely? What kinds of recovery methods are being used?Duncan summarizes his method as follows:.I make a separate computer-based model to forecast the oil production for each of the major oil-producing nations in the world; 2) The latest oil data and related information on each nation are gathered from journals, the internet, and colleagues just before each national model is run; 3) Then both a Low oil forecast and a Medium oil forecast are made for each nation; 4) Next all of the Medium oil forecasts are combined (added up) to give the world oil forecast; 5) This process is repeated annually as soon as new oil data and related information become available; 6) A Scatter Graph indicates that the eight world oil forecasts that we've completed so far seem to be converging on Peak Oil in 2006 or 2007; 7) One new forecast (point) will be added to the Scatter Graph each year until Peak Oil is confirmed.Campbell's a.n.a.lysis of likely future oil production by individual producing nations yields a global peak date of 2008.4. Compare the amount of new production capacity likely to be available over the coming years with the amount of production capacity needed to offset decline rates from existing fields. The global oil industry needs to develop new production capacity yearly, in order to meet new demand and offset declines in production rates from individual wells and producing regions already past their all-time peaks. Currently, the world produces about 83 million barrels per day of all petroleum liquids combined (conventional oil plus oil from tar sands, natural gas liquids, and so on). The IEA estimates that in 2005 the world will need another 1.5 million barrels per day of new production capacity in order to meet new demand, plus another 4 mb/d to offset declines from existing fields - a total of about 5.5 mb/d. In 2006, a slightly greater new quant.i.ty will be needed, and in 2007, more still. In the five years from 2005 to 2010 a total of over 35 mb/d of new production capacity will need to come online. (These figures are agreed upon by both industry and various governmental agencies.) A substantial effort is necessary, to say the least.Many Little Peaks, One Big One.Richard Duncan, of the Inst.i.tute on Energy and Man, has compiled the following forecasts on oil production peaks for 45 nations comprising seven regions (combined they accounted for more than 98 percent of the world's oil production, as of yearend 2003). The data are extracted from Duncan's World Oil Forecast #8, and (along with his unique method of world oil forecasting discussed previously) they are published for the first time here.Note that forecast #8 includes oil production as defined by the BP Statistical Review of World Energy, June 2004, p. 6: "Includes crude oil, oil sands, NGLs (natural gas liquids - the liquid content of natural gas where this is recovered separately). Excludes liquid fuels from other sources such as coal derivatives."But where will all this new production capacity come from?In general, new production capacity arises from three sources: the discovery of new resources; the development of previously discovered resources (including reserve growth and infill drilling); or the development of unconventional resources (which sometimes depends on the invention and implementation of new technologies).It takes time and investment to develop new production capacity. Thus it is possible - though no simple matter! - to gather the necessary data, a.n.a.lyze it, and project how much new production capacity is likely to emerge over the next five years, given current rates of investment, the available technology, and the discoveries in place. (Even if a huge new discovery were to be made next year, it would probably be impossible to bring the oil from it into production before 2010.) Chris Skrebowski, editor of Petroleum Review, has done just that in his 2004 report, "Oil Field Megaprojects," sponsored by the Oil Depletion a.n.a.lysis Centre (ODAC).Skrebowski compiles and regularly updates the details of planned major production projects, as reported by the oil companies. The list contains data on all announced fields with at least 500 million barrels of estimated reserves, and on projects with the claimed potential to produce 100,000 barrels a day or more.Skrebowski and ODAC a.n.a.lyzed 68 production projects with announced start-up dates ranging from 2004 through 2010, and found that they are likely to add about 12.5 million barrels per day of new production capacity. In a press release, he stated: "This new production would almost certainly not be sufficient to offset diminishing supplies from existing sources and still meet growing global demand," and that "even with relatively low demand growth, our study indicates a seemingly unbridgeable supply-demand gap opening up after 2007."27 "It is disturbing to see such a dramatic fall-off of new project commitments after 2007, and not more than a handful of tentative projects into the next decade," Skrebowski said. "This could very well be a signal that world oil production is rapidly approaching its peak, as a growing number of a.n.a.lysts now forecast, especially in view of the diminishing prospects for major new oil discoveries."At the end of the day, there are still uncertainties. Major new oil discoveries are always possible, though increasingly unlikely. Probably the greatest uncertainty with respect to the timing of the global oil production peak is future demand. If the global economy fares well, then demand will increase and the peak will come sooner; if the economy falters, then the peak will come later. If the world stumbles into a full-fledged depression, the peak could be delayed significantly, and the effects of the phenomenon could be masked by other events.Nevertheless, as we have seen, the results of the possible forecasting methods tend to converge. We are within only a few years of the all-time global oil production peak. We are virtually at the summit now, with almost no time left for maneuvering before the event itself is upon us.Hubbert's Critics: The Cornucopian Argument.If, as Hubbert and his followers have said, the future of oil production could spell disaster for industrial societies, then it is vital that we examine the geologists' claims from every possible angle to determine whether or not they are correct. Are there critics who dispute Hubbert, Campbell, et al., and are their critiques valid?There is a school of thought, whose ideas are voiced mostly by economists, that says there is plenty of oil. In this section we will examine the arguments of three such "cornucopians": Peter Huber, Bjrn Lomborg, and Michael C. Lynch.It is important to point out, however, that the cornucopian perspective is not limited to a few economists or industry lobbyists. As we will see, the USGS and Department of Energy (DoE) have posted petroleum production forecasts that are far more optimistic than those of Hubbert and his followers. These organizations present "official" projections, which are presumably supported by hard evidence.Who is right? Sorting out the arguments is no small task, but the stakes are high enough to warrant whatever intellectual effort is required.Let us begin with the most extravagant and general cornucopian claims, and work our way toward more specific and technical arguments.Peter Huber, author of Hard Green: Saving the Environment from the Environmentalists, is a lawyer and writer. He earned his doctorate in mechanical engineering from MIT and served as an a.s.sistant and later a.s.sociate Professor at MIT for six years. He is currently a senior fellow at the Manhattan Inst.i.tute.