The End of Cheap Oil Global Production of Conventional Oil Will Begin to Decline Sooner Than Most People Think, Probably Within 10 Years by Colin J

Home , Unconventional oil

The End of Cheap Oil Global production of conventional oil will begin to decline sooner than most people think, probably within 10 years by Colin J. Campbell and Jean H. Laherrère

n 1973 and 1979 a pair of sudden Middle Eastern members of the Orga- could remain plentiful and cheap for 43 price increases rudely awakened the nization of Petroleum Exporting Coun- more years—probably longer, because Iindustrial world to its dependence on tries (OPEC) were able to hike prices official charts show reserves growing. cheap crude oil. Prices first tripled in re- not because oil was growing scarce but Unfortunately, this appraisal makes sponse to an Arab embargo and then because they had managed to corner 36 three critical errors. First, it relies on dis- nearly doubled again when Iran de- percent of the market. Later, when de- torted estimates of reserves. A second throned its Shah, sending the major mand sagged, and the flow of fresh mistake is to pretend that production economies sputtering into recession. Alaskan and North Sea oil weakened will remain constant. Third and most Many analysts warned that these crises OPEC’s economic stranglehold, prices important, conventional wisdom erro- proved that the world would soon run collapsed. neously assumes that the last bucket of out of oil. Yet they were wrong. The next oil crunch will not be so oil can be pumped from the ground just Their dire predictions were emotion- temporary. Our analysis of the discov- as quickly as the barrels of oil gushing al and political reactions; even at the ery and production of oil fields around from wells today. In fact, the rate at time, oil experts knew that they had no the world suggests that within the next which any well—or any country—can scientific basis. Just a few years earlier decade, the supply of conventional oil produce oil always rises to a maximum oil explorers had discovered enormous will be unable to keep up with demand. and then, when about half the oil is gone, new oil provinces on the north slope of This conclusion contradicts the picture begins falling gradually back to zero. Alaska and below the North Sea off the one gets from oil industry reports, which From an economic perspective, when coast of Europe. By 1973 the world had boasted of 1,020 billion barrels of oil the world runs completely out of oil is consumed, according to many experts’ (Gbo) in “proved” reserves at the start thus not directly relevant: what matters best estimates, only about one eighth of of 1998. Dividing that figure by the is when production begins to taper off. its endowment of readily accessible crude current production rate of about 23.6 Beyond that point, prices will rise un- oil (so-called conventional oil). The five Gbo a year might suggest that crude oil less demand declines commensurately.

HISTORY OF OIL PRODUCTION, from the first commercial American well in 1859 Titusville, Pa. (left), to derricks bristling above the Los Angeles basin (below), began with steady growth in the U.S. (red line). But domestic production began to decline after 1970, and restrictions in the flow of Middle Eastern oil in 1973 and 1979 led to inflation and shortages (near and center right). More recently, the Persian Gulf War, with its burning oil fields (far right), reminded the industrial world of its dependence on Middle Eastern oil production (gray line).

1920s CORBIS-BETTMANN CORBIS-BETTMANN

78 Scientific American March 1998 Copyright 1998 Scientific American, Inc. The End of Cheap Oil Using several different techniques to es- tistics because companies meter the oil In practice, companies and countries timate the current reserves of conven- as it flows from their wells. The record are often deliberately vague about the tional oil and the amount still left to be of production is not perfect (for exam- likelihood of the reserves they report, discovered, we conclude that the de- ple, the two billion barrels of Kuwaiti preferring instead to publicize whichev- cline will begin before 2010. oil wastefully burned by Iraq in 1991 is er figure, within a P10 to P90 range, usually not included in official statistics), best suits them. Exaggerated estimates Digging for the True Numbers but errors are relatively easy to spot can, for instance, raise the price of an and rectify. Most experts agree that the oil company’s stock. e have spent most of our careers industry had removed just over 800 The members of OPEC have faced an W exploring for oil, studying reserve Gbo from the earth at the end of 1997. even greater temptation to inflate their figures and estimating the amount of oil Getting good estimates of reserves is reports because the higher their reserves, left to discover, first while employed at much harder, however. Almost all the the more oil they are allowed to export. major oil companies and later as inde- publicly available statistics are taken National companies, which have exclu- pendent consultants. Over the years, from surveys conducted by the Oil and sive oil rights in the main OPEC coun- we have come to appreciate that the rel- Gas Journal and World Oil. Each year tries, need not (and do not) release de- evant statistics are far more complicat- these two trade journals query oil firms tailed statistics on each field that could ed than they first appear. and governments around the world. be used to verify the country’s total re- Consider, for example, three vital They then publish whatever production serves. There is thus good reason to sus- numbers needed to project future oil and reserve numbers they receive but pect that when, during the late 1980s, production. The first is the tally of how are not able to verify them. six of the 11 OPEC nations increased much oil has been extracted to date, a The results, which are often accepted their reserve figures by colossal amounts, figure known as cumulative production. uncritically, contain systematic errors. ranging from 42 to 197 percent, they The second is an estimate of reserves, For one, many of the reported figures did so only to boost their export quotas. AHERRÈRE the amount that companies can pump are unrealistic. Estimating reserves is an Previous OPEC estimates, inherited

out of known oil fields before having to inexact science to begin with, so petro- from private companies before govern- CE: JEAN H. L abandon them. Finally, one must have leum engineers assign a probability to ments took them over, had probably an educated guess at the quantity of their assessments. For example, if, as been conservative, P90 numbers. So conventional oil that remains to be dis- geologists estimate, there is a 90 percent some upward revision was warranted. covered and exploited. Together they chance that the Oseberg field in Norway But no major new discoveries or tech- add up to ultimate recovery, the total contains 700 million barrels of recover- nological breakthroughs justi- number of barrels that will have been able oil but only a 10 percent fied the addition of a stag- JENNIFER C. CHRISTIANSEN; SOUR extracted when production ceases many chance that it will yield 2,500 mil- gering 287 Gbo. That in- decades from now. lion more barrels, then the lower crease is more than all The obvious way to gather these num- figure should be cited as the so- the oil ever discovered bers is to look them up in any of several called P90 estimate (P90 for in the U.S.—plus 40 publications. That approach works well “probability 90 percent”) and percent. Non-OPEC enough for cumulative production sta- the higher as the P10 reserves. countries, of course, 1973 1979

1973 1991 MANN T ORBIS-BET I/C UP

OIL PRODUCTION

IN THE U.S. os

AND CANADA hot agnum P MANN M T Y OIL PRODUCTION 1979

1979 ARBE

IN THE PERSIAN GULF ORBIS-BET I/C UNO B 1991 UP BR

from any large region when A

SICIST TOTAL FOR

HY about half the crude is gone. ENTIRE REGION URIE GR Adding the output of ﬁelds of A various sizes and ages (green TION GEOP A curves at right) usually yields OR a bell-shaped production curve for the region as a whole. M. Y OF EXPL King Hubbert (left), a geolo- INDIVIDUAL

gist with Shell Oil, exploited WELLS JENNIFER C. CHRISTIANSEN; L this fact in 1956 to predict cor- (INCREASING YIELD) rectly that oil from the lower ANNUAL OIL PRODUCTION TESY OF THE SOCIET 48 American states would peak 0 OUR C around 1969. 0 YEARS 40

are not above fudging their numbers ei- answer, because adding what is 90 per- our value represents the amount most ther: 59 nations stated in 1997 that their cent likely for each field, as is done in likely to come out of known oil fields, reserves were unchanged from 1996. the U.S., does not in fact yield what is whereas the larger number is supposedly Because reserves naturally drop as old 90 percent likely for a country or the en- a cautious estimate of proved reserves. fields are drained and jump when new tire planet. On the contrary, summing For the purposes of calculating when fields are discovered, perfectly stable many P90 reserve estimates always un- oil production will crest, even more crit- numbers year after year are implausible. derstates the amount of proved oil in a ical than the size of the world’s reserves region. The only correct way to total is the size of ultimate recovery—all the Unproved Reserves up reserve numbers is to add the mean, cheap oil there is to be had. In order to or average, estimates of oil in each field. estimate that, we need to know wheth- nother source of systematic error in In practice, the median estimate, often er, and how fast, reserves are moving up A the commonly accepted statistics is called “proved and probable,” or P50 or down. It is here that the official sta- that the definition of reserves varies reserves, is more widely used and is good tistics become dangerously misleading. widely from region to region. In the enough. The P50 value is the number of U.S., the Securities and Exchange Com- barrels of oil that are as likely as not to Diminishing Returns mission allows companies to call re- come out of a well during its lifetime, serves “proved” only if the oil lies near assuming prices remain within a limited ccording to most accounts, world a producing well and there is “reason- range. Errors in P50 estimates tend to A oil reserves have marched steadily able certainty” that it can be recovered cancel one another out. upward over the past 20 years. Extend- profitably at current oil prices, using We were able to work around many ing that apparent trend into the future, existing technology. So a proved reserve of the problems plaguing estimates of one could easily conclude, as the U.S. estimate in the U.S. is roughly equal to conventional reserves by using a large Energy Information Administration has, a P90 estimate. body of statistics maintained by Petro- that oil production will continue to rise Regulators in most other countries do consultants in Geneva. This informa- unhindered for decades to come, in- not enforce particular oil-reserve definition, assembled over 40 years from myr- creasing almost two thirds by 2020. tions. For many years, the former Soviet iad sources, covers some 18,000 oil fields Such growth is an illusion. About 80 countries have routinely released wildly worldwide. It, too, contains some dubi- percent of the oil produced today flows optimistic figures—essentially P10 re- ous reports, but we did our best to cor- from fields that were found before 1973, serves. Yet analysts have often misinter- rect these sporadic errors. and the great majority of them are de- preted these as estimates of “proved” According to our calculations, the clining. In the 1990s oil companies have reserves. World Oil reckoned reserves world had at the end of 1996 approxi- discovered an average of seven Gbo a in the former Soviet Union amounted mately 850 Gbo of conventional oil in year; last year they drained more than to 190 Gbo in 1996, whereas the Oil P50 reserves—substantially less than three times as much. Yet official figures and Gas Journal put the number at 57 the 1,019 Gbo reported in the Oil and indicated that proved reserves did not Gbo. This large discrepancy shows just Gas Journal and the 1,160 Gbo esti- fall by 16 Gbo, as one would expect— how elastic these numbers can be. mated by World Oil. The difference is rather they expanded by 11 Gbo. One Using only P90 estimates is not the actually greater than it appears because reason is that several dozen governments

EARTH’S CONVENTIONAL CRUDE OIL is almost half gone. out of known ﬁelds) and future discoveries together will provide Reserves (deﬁned here as the amount as likely as not to come little more than what has already been burned.

UNDISCOVERED: RESERVES: 150 BILLION BARRELS 850 BILLION BARRELS CE A URIE GR A L

80 Scientific American March 1998 Copyright 1998 Scientific American, Inc. The End of Cheap Oil GLOBAL PRODUCTION OF OIL, both 30 conventional and unconventional (red), WORLD recovered after falling in 1973 and 1979. WORLD OUTSIDE PERSIAN GULF AHERRÈRE But a more permanent decline is less than 25 PERSIAN GULF 10 years away, according to the authors’ U.S. AND CANADA

FORMER SOVIET UNION CE: JEAN H. L model, based in part on multiple Hubbert U.K. AND NORWAY curves (lighter lines). U.S. and Canadian 20 CE; SOUR oil (brown) topped out in 1972; produc- A tion in the former Soviet Union (yellow) URIE GR

15 A has fallen 45 percent since 1987. A crest in L the oil produced outside the Persian Gulf region (purple) now appears imminent. 10 opted not to report declines in their re- 5 serves, perhaps to enhance their political cachet and their ability to obtain loans. ANNUAL OIL PRODUCTION (BILLIONS OF BARRELS 0 A more important cause of the expan- 1930 1940 1950 1960 1970 1980 1990 2000 2010 2020 2030 2040 2050 sion lies in revisions: oil companies re- placed earlier estimates of the reserves left in many fields with higher numbers. rate continued its decline uninterrupt- tions continued producing at full capac- For most purposes, such amendments ed. There is only so much crude oil in ity. Our analysis reveals that a number are harmless, but they seriously distort the world, and the industry has found of the largest producers, including Nor- forecasts extrapolated from published about 90 percent of it. way and the U.K., will reach their peaks reports. around the turn of the millennium un- To judge accurately how much oil ex- Predicting the Inevitable less they sharply curtail production. By plorers will uncover in the future, one 2002 or so the world will rely on Mid- has to backdate every revision to the redicting when oil production will dle East nations, particularly five near year in which the field was first discov- P stop rising is relatively straightfor- the Persian Gulf (Iran, Iraq, Kuwait, ered—not to the year in which a compa- ward once one has a good estimate of Saudi Arabia and the United Arab Emi- ny or country corrected an earlier esti- how much oil there is left to produce. rates), to fill in the gap between dwin- mate. Doing so reveals that global dis- We simply apply a refinement of a tech- dling supply and growing demand. But covery peaked in the early 1960s and nique first published in 1956 by M. once approximately 900 Gbo have been has been falling steadily ever since. By King Hubbert. Hubbert observed that consumed, production must soon begin extending the trend to zero, we can in any large region, unrestrained extrac- to fall. Barring a global recession, it make a good guess at how much oil the tion of a finite resource rises along a bell- seems most likely that world produc- industry will ultimately find. shaped curve that peaks when about tion of conventional oil will peak dur- We have used other methods to esti- half the resource is gone. To demon- ing the first decade of the 21st century. mate the ultimate recovery of conven- strate his theory, Hubbert fitted a bell Perhaps surprisingly, that prediction tional oil for each country [see box on curve to production statistics and pro- does not shift much even if our esti- next two pages], and we calculate that jected that crude oil production in the mates are a few hundred billion barrels the oil industry will be able to recover lower 48 U.S. states would rise for 13 high or low. Craig Bond Hatfield of the only about another 1,000 billion bar- more years, then crest in 1969, give or University of Toledo, for example, has rels of conventional oil. This number, take a year. He was right: production conducted his own analysis based on a though great, is little more than the 800 peaked in 1970 and has continued to 1991 estimate by the U.S. Geological billion barrels that have already been follow Hubbert curves with only minor Survey of 1,550 Gbo remaining—55 extracted. deviations. The flow of oil from several percent higher than our figure. Yet he It is important to realize that spend- other regions, such as the former Soviet similarly concludes that the world will ing more money on oil exploration will Union and the collection of all oil pro- hit maximum oil production within the not change this situation. After the price ducers outside the Middle East, also fol- next 15 years. John D. Edwards of the of crude hit all-time highs in the early lows Hubbert curves quite faithfully. University of Colorado published last 1980s, explorers developed new tech- The global picture is more complicat- August one of the most optimistic re- nology for finding and recovering oil, ed, because the Middle East members of cent estimates of oil remaining: 2,036 and they scoured the world for new OPEC deliberately reined back their oil Gbo. (Edwards concedes that the indus- fields. They found few: the discovery exports in the 1970s, while other na- try has only a 5 percent chance of at-

PRODUCED: 800 BILLION BARRELS

The End of Cheap Oil Copyright 1998 Scientific American, Inc. Scientific American March 1998 81 How Much Oil Is Left to Find? off the coast of Britain, for example, will yield about 420 million barrels (a). Second, we plotted the amount of oil discov- e combined several techniques to conclude that about ered so far in some regions against the cumulative number of W1,000 billion barrels of conventional oil remain to be exploratory wells drilled there. Because larger fields tend to produced. First, we extrapolated published production figures be found first—they are simply too large to miss—the curve for older oil fields that have begun to decline. The Thistle field rises rapidly and then flattens, eventually reaching a theo-

We can predict the amount of remaining oil ...from the diminishing returns on exploration in larger regions... a from the decline of aging fields... D b 50 300 FORMER SOVIET UNION 40 THEORETICAL 200 MAXIMUM 30 20 100

TIVE OIL DISCOVERE AFRICA 10 PRACTICAL LIMIT

0 (BILLIONS OF BARRELS) 0

(MILLIONS OF BARRELS) 0 5,000 10,000 15,000 20,000 ANNUAL OIL PRODUCTION 0 50 100 150 200 250 300 350 400 450 500 CUMULA OIL PRODUCED TO DATE (MILLIONS OF BARRELS) CUMULATIVE NUMBER OF EXPLORATORY WELLS taining that very high goal.) Even so, new technologies have steadily increased Last, economists like to point out that his calculations suggest that convention- the fraction of oil that can be recovered the world contains enormous caches of al oil will top out in 2020. from fields in a basin—the so-called re- unconventional oil that can substitute covery factor. In the 1960s oil compa- for crude oil as soon as the price rises Smoothing the Peak nies assumed as a rule of thumb that high enough to make them profitable. only 30 percent of the oil in a field was There is no question that the resources actors other than major economic typically recoverable; now they bank are ample: the Orinoco oil belt in Vene- F changes could speed or delay the on an average of 40 or 50 percent. That zuela has been assessed to contain a point at which oil production begins to progress will continue and will extend staggering 1.2 trillion barrels of the decline. Three in particular have often global reserves for many years to come, sludge known as heavy oil. Tar sands led economists and academic geologists the argument runs. and shale deposits in Canada and the to dismiss concerns about future oil Of course, advanced technologies will former Soviet Union may contain the production with naive optimism. buy a bit more time before production equivalent of more than 300 billion First, some argue, huge deposits of oil starts to fall [see “Oil Production in the barrels of oil [see “Mining for Oil,” by may lie undetected in far-off corners of 21st Century,” by Roger N. Anderson, Richard L. George, on page 84]. Theo- the globe. In fact, that is very unlikely. on page 86]. But most of the apparent retically, these unconventional oil re- Exploration has pushed the frontiers improvement in recovery factors is an serves could quench the world’s thirst back so far that only extremely deep wa- artifact of reporting. As oil fields grow for liquid fuels as conventional oil pass- ter and polar regions remain to be fully old, their owners often deploy newer es its prime. But the industry will be tested, and even their prospects are now technology to slow their decline. The hard-pressed for the time and money reasonably well understood. Theoreti- falloff also allows engineers to gauge needed to ramp up production of un- cal advances in geochemistry and geo- the size of the field more accurately and conventional oil quickly enough. physics have made it possible to map to correct previous underestimation—in Such substitutes for crude oil might productive and prospective fields with particular P90 estimates that by defini- also exact a high environmental price. impressive accuracy. As a result, large tion were 90 percent like-

tracts can be condemned as barren. ly to be exceeded. Y

1,200 OIL VE Much of the deepwater realm, for ex- Another reason not to BACKDATED RESERVES S, ANT T ample, has been shown to be absolutely pin too much hope on 1,000 AL SUR OGIC nonprospective for geologic reasons. better recovery is that oil 800 ONSUL What about the much touted Caspi- companies routinely REPORTED OC ETR

RESERVES P an Sea deposits? Our models project count on technological 600 CE: AND U.S. GEOL that oil production from that region progress when they com- 400

will grow until around 2010. We agree pute their reserve esti- CE; SOUR

DISCOVERIES A 200 with analysts at the USGS World Oil As- mates. In truth, advanced BILLIONS OF BARRELS URIE GR sessment program and elsewhere who technologies can offer lit- A AND GAS JOURNAL L rank the total resources there as roughly tle help in draining the 1940 1950 1960 1970 1980 1990 2000 equivalent to those of the North Sea— largest basins of oil, those GROWTH IN OIL RESERVES since 1980 is an illusion that is, perhaps 50 Gbo but certainly not onshore in the Middle caused by belated corrections to oil-ﬁeld estimates. Back- several hundreds of billions as some- East where the oil needs dating the revisions to the year in which the ﬁelds were dis- times reported in the media. no assistance to gush from covered reveals that reserves have been falling because of a A second common rejoinder is that the ground. steady decline in newfound oil (blue).

82 Scientific American March 1998 Copyright 1998 Scientific American, Inc. The End of Cheap Oil retical maximum: for Africa, 192 Gbo. But the time and cost of terestingly, galaxies, urban populations and other natural ag- exploration impose a more practical limit of perhaps 165 Gbo glomerations also seem to fall along such parabolas.) Finally, (b). Third, we analyzed the distribution of oil-field sizes in the we checked our estimates by matching our projections for oil Gulf of Mexico and other provinces. Ranked according to size production in large areas, such as the world outside the Per- and then graphed on a logarithmic scale, the fields tend to sian Gulf region, to the rise and fall of oil discovery in those fall along a parabola that grows predictably over time (c). (In- places decades earlier (d ). —C.J.C. and J.H.L.

...by extrapolating the size of new fields into the future...... and by matching production to earlier discovery trends. c d 1000 DISCOVERY YEAR

1935 1955 1975 1995 AHERRÈRE 100 40 ARRELS) PROJECTED 30 DISCOVERY Y IN FIELD 10 CE: JEAN H. L 1959 20 HUBBERT 1969 1 PRODUCTION CE; SOUR MODEL A 1979 10 1993 OIL INITIALL URIE GR

(MILLIONS OF B 0.1 A

0 L OUTSIDE MIDDLE EAST LARGEST 10TH 100TH 1,000TH SMALLEST (BILLIONS OF BARRELS) 1930 1950 1970 1990 2010 2030 FIELD LARGEST LARGEST LARGEST FIELD OIL FOUND OR PRODUCED PRODUCTION YEAR

Tar sands typically emerge from strip mand for oil will increase 60 percent (to tration to increase funding for energy mines. Extracting oil from these sands about 40 Gbo a year) by 2020. research by $1 billion over the next five and shales creates air pollution. The The switch from growth to decline in years. That is a small step in the right Orinoco sludge contains heavy metals oil production will thus almost certain- direction, one that must be followed by and sulfur that must be removed. So ly create economic and political tension. giant leaps from the private sector. governments may restrict these indus- Unless alternatives to crude oil quickly The world is not running out of oil— tries from growing as fast as they could. prove themselves, the market share of at least not yet. What our society does In view of these potential obstacles, our the OPEC states in the Middle East will face, and soon, is the end of the abun- skeptical estimate is that only 700 Gbo rise rapidly. Within two years, these na- dant and cheap oil on which all indus- will be produced from unconventional tions’ share of the global oil business trial nations depend. SA reserves over the next 60 years. will pass 30 percent, nearing the level reached during the oil-price shocks of On the Down Side the 1970s. By 2010 their share will quite probably hit 50 percent. eanwhile global demand for oil is The world could thus see radical in- The Authors M currently rising at more than 2 creases in oil prices. That alone might COLIN J. CAMPBELL and JEAN H. LA- percent a year. Since 1985, energy use is be sufficient to curb demand, flattening HERRÈRE have each worked in the oil in- up about 30 percent in Latin America, production for perhaps 10 years. (De- dustry for more than 40 years. After com- 40 percent in Africa and 50 percent in mand fell more than 10 percent after pleting his Ph.D. in geology at the University of Oxford, Campbell worked for Texaco as Asia. The Energy Information Adminis- the 1979 shock and took 17 years to an exploration geologist and then at Amoco tration forecasts that worldwide de- recover.) But by 2010 or so, many Mid- as chief geologist for Ecuador. His decade- dle Eastern nations will themselves be long study of global oil-production trends past the midpoint. World production has led to two books and numerous papers. 800 will then have to fall. Laherrère’s early work on seismic refraction With sufficient preparation, however, surveys contributed to the discovery of Africa’s largest oil field. At Total, a French 700 the transition to the post-oil economy

VES oil company, he supervised exploration tech- 600 need not be traumatic. If advanced niques worldwide. Both Campbell and La- SAUDI ARABIA

OIL AND GAS JOURNAL methods of producing liquid fuels from herrère are currently associated with Petro- 500 CE: natural gas can be made profitable and consultants in Geneva. 400 IRAN scaled up quickly, gas could become the CE; SOUR A next source of transportation fuel [see Further Reading TED OIL RESER 300 U.A.E. “Liquid Fuels from Natural Gas,” by URIE GR Updated Hubbert Curves Analyze A 200 IRAQ L Safaa A. Fouda, on page 92]. Safer nu- (BILLIONS OF BARRELS) World Oil Supply. L. F. Ivanhoe in World REPOR 100 KUWAIT clear power, cheaper renewable energy, Oil, Vol. 217, No. 11, pages 91–94; No- and oil conservation programs could all vember 1996. VENEZUELA 0 help postpone the inevitable decline of The Coming Oil Crisis. Colin J. Camp- 1980 1985 1990 1995 conventional oil. bell. Multi-Science Publishing and Petro- consultants, Brentwood, England, 1997. SUSPICIOUS JUMP in reserves reported by Countries should begin planning and Oil Back on the Global Agenda. Craig six OPEC members added 300 billion bar- investing now. In November a panel of Bond Hatfield in Nature, Vol. 387, page rels of oil to official reserve tallies yet fol- energy experts appointed by President 121; May 8, 1997. lowed no major discovery of new fields. Bill Clinton strongly urged the adminis-