Is Peak Oil Almost Here?

To a geologist, gauging how much coal the world has left to burn is a fairly straightforward, if daunting, business. Millions upon millions of drill holes have revealed where the coal is. So geologists can just evaluate each seam's quality and the cost of extraction. Add up all the coal worth mining and you've got lots and lots--within the United States, a century or even two of U.S. consumption; globally, 150 years' worth for the world.

To a geologist, gauging how much coal the world has left to burn is a fairly straightforward, if daunting, business. Millions upon millions of drill holes have revealed where the coal is. So geologists can just evaluate each seam's quality and the cost of extraction. Add up all the coal worth mining and you've got lots and lots--within the United States, a century or even two of U.S. consumption; globally, 150 years' worth for the world.

But there's another, emerging approach to assessing coal resources that yields more sobering results. Rather than go into the field, these analysts go to the record books to see how fast miners have been producing coal of late. By fitting curves to that production history, they come up with a number for the total amount of coal that will ever be mined and a date for the greatest production, the time of "peak coal," after which production inevitably declines.


Early results from this curve-fitting analysis of production history show much less coal being mined than geologists ever expected and a peak in coal production looming as early as a decade from now. Curve fitting "is a worthy competitor to a geological estimate" of remaining coal, says David Rutledge of the California Institute of Technology in Pasadena, a nongeologist who has produced such an estimate himself. Geologists beg to differ. "The whole notion of applying statistics to time series [of coal production] is fraught with danger," says energy resource geologist Peter McCabe of the Commonwealth Scientific and Industrial Research Organisation (CSIRO) in North Ryde, Australia. "I think what you see in Rutledge's presentation is a fundamental misunderstanding."

At stake are two central questions: How bad is greenhouse warming likely to get? And when must alternatives to coal come online? By Rutledge's calculations, burning all the coal, oil, and gas humans can get their hands on won't pump enough carbon dioxide into the atmosphere to drive global warming past 3°C, a far less intimidating ultimate warming than the 8°C or 10°C of some scenarios. But engineers developing energy alternatives would have only a few decades to get them in place.

Ringing the bell

The growing popularity of curve fitting to fossil fuel production got its start with geophysicist M. King Hubbert (1903-1989) and his bell-shaped curves. Hubbert--a prominent researcher at Shell Oil and then the U.S. Geological Survey--concluded that the rate at which oil was produced would follow the same bell-shaped curve over time as production from a single field does (see top figure, attached).

Production in Hubbert's scheme gradually accelerates from the bell's left side as drillers tap the most accessible, most easily extracted oil pools at an ever-faster rate. But eventually, no matter how much they drill, it gets harder and harder to suck out the oil because the remaining pools are fewer, smaller, deeper, and more difficult to drain. The earlier acceleration now slows across the crown of the bell until production stops growing, peaking at the bell's top. It then declines until, at the right side, all the oil that will ever be produced has been produced.

In 1956, Hubbert published his curve-fitting prediction that U.S. oil production would peak in the early 1970s. He used the U.S. production history up to that point to set the shape of the curve and an estimate of the amount of oil that would ultimately be produced to set the area encompassed by a complete curve.

U.S. production in fact peaked in 1970, inspiring a generation of oil "peakists" to apply Hubbert's approach to world oil (Science, 21 August 1998, p. 1128). As of 1998, they broadly agreed that world oil would reach its peak by now or at least by the middle of the next decade. In fact, production outside oil-rich OPEC (Organization of the Petroleum Exporting Countries) has not risen since 2004, despite years of encouragement by high oil prices. And total world production has gone nowhere since 2005.

On to coal

In the last couple of years, forecasting coal production by Hubbert's approach has come into vogue, partly because geologists seemed to be having trouble assessing how much minable coal was left. For example, "40% of the world's coal disappeared in 3 years," recalls retired U.S. Geological Survey coal expert Harold Gluskoter. For the World Energy Council's triennial survey of coal resources in 1990, China cut its recoverable coal reserves--the amount of remaining coal geologists believe can be extracted with today's technology at today's prices--to one-sixth of what it had reported in 1987. The coal was mostly still there; the Chinese just decided they could extract only a smaller proportion of it.

Less dramatically, in 2007 a committee of the U.S. National Research Council that Gluskoter served on could not support the long-standing estimate of about 267 billion short tons of recoverable reserves in the United States. Divided by current U.S. production, the old estimate gave the oft-quoted figure of a 250-year supply for the United States. "We probably have 100 years. We don't know how much after that," says Gluskoter.

Reserve estimates around the world were also coming down, and dividing estimates of minable coal by current production says nothing about when coal might peak. To get at both forecasts, analysts tried a variation on curve fitting à la Hubbert. No results have been published in the peer-reviewed literature, but Rutledge's effort has probably gotten the widest exposure ( ).

Rutledge, an electrical engineer, adapted a technique used by geologist Kenneth Deffeyes, professor emeritus at Princeton University, to predict a world oil peak in 2005. (Deffeyes sees his prediction holding up nicely.) In the technique, an as-yet-to-peak production history (see bottom graph, above) is replotted in terms of cumulative production and annual production in such a way as to produce a straight line, at least if production approximates an ideal bell-shaped curve. The straight line intercepts the x-axis at the ultimate production--all the coal that will ever be produced--and the year in which half of the ultimate production will be achieved is the year of peak production.

Rutledge tested the method on regions long past their coal peaks. The coal production of the United Kingdom--once the world's premier energy supplier--peaked in 1913 and is now at about 6% of its peak. A straight line makes "a beautiful fit" to the replotted history, Rutledge told audience members at last December's meeting of the American Geophysical Union. The projected ultimate production in the United Kingdom is about 28 billion metric tons of coal, Rutledge said, which should be reached in 8 years. Geologic estimates made in the 19th century had reserves near 200 billion tons. In fact, the World Energy Council's geologically based reserve estimates--provided by the United Kingdom--stayed at 19th century levels until the 1970s, when they collapsed toward Rutledge's number.

Applied to 14 major coal-producing regions, Rutledge's method gives a world ultimate production of 660 billion metric tons. That's only one-quarter of geologic estimates of ultimate production, he says. And when combined with similar estimates of ultimate production of oil and gas, the total emissions of carbon as the greenhouse gas carbon dioxide till 2100 are smaller than any of the 40 emissions scenarios that climate scientists have been working with for the past 10 years.

As to when coal will peak, Rutledge declines to say, citing the way peak timing varied widely among regions already well past their peak. He will say, however, that in his projection the world will have produced a whopping 90% of its coal by 2069. Physicist Mikael Höök of Uppsala University in Sweden and his colleagues are willing to point to a peak. They have taken a similar approach to Rutledge's but with some reliance on estimated reserves. Still, they see world coal production topping out by 2020, entering a 30-year-long plateau, and then declining.

Outside influences

Geologists and resource economists aren't ready to give up on coal so soon. "The bell-shaped curve is nice to look at after the fact," says Gluskoter, but "I'm not sure how predictive it is." "You can put me in the skeptical camp" as well, says physicist and energy analyst Klaus Lackner of Columbia University. Rutledge "puts too much weight on a simple model," he says. "The world is not a two-parameter curve."

U.K. coal "did not decline because it reached some magic percentage of coal depletion," argues CSIRO's McCabe. "Gradually, [U.K.] demand disappeared. Coal was no longer used for power, steamships, railroads, domestic heating, or iron and steel production. When I was a kid in England many years ago, gas for the home was produced from coal. There are much cheaper alternatives in the U.K. for energy." Natural gas from the North Sea replaced coal gas in the home, he notes. Relatively inexpensive--and cleaner-burning--oil became available.

Does it matter why coal production behaves as it does? responds Rutledge. Whether it's a physical lack of resources or a demand shift to cheaper, more attractive sources, a limit is a limit. Gluskoter and many colleagues are betting on a demand shift. "I believe in technology," he says. But first, technology could actually stretch coal resources. "The resource is there," he notes, "it's perfectly minable. It's just not economical right now." Improved technology will let miners get at thinner, less accessible seams, he says; uneconomic coal might even be turned into gas right in the ground.

Despite any coal added by technology, "we'll stop using coal before we run out of it," says Gluskoter. Just as cleaner, cheaper fuels displaced British coal, less-polluting, less-expensive energy sources will replace coal worldwide, he argues. His bet is on solar energy. Plenty of usable coal will be left in the ground, he adds, but "we're not going to be burning coal in a couple hundred years."