Current Events

Join us as we watch the crisis unfolding

March 8, 2006

Fifty years ago today, on the morning of March 8, 1956, M. King Hubbert gave a talk in San Antonio in which he predicted that US oil production would start to decline by the early 1970s. Up until minutes before Hubbert began, executives at the head office of Shell Oil (his employer) were on the phone asking him to cancel the talk. Hubbert's presentation is widely regarded as the opening of the great debate about the finiteness of our oil supply. The original letter inviting Hubbert to give the talk has been posted at

Fifty years ago today . . . M. King Hubbert gave a talk . . . in which he predicted that US oil production would start to decline by the early 1970s.

Hubbert enlisted two senior petroleum geologists (Wallace Pratt and Lewis Weeks) to provide opinions about the total amount of oil that would eventually be produced from the US. The more optimistic opinion, 200 billion barrels, gave a production peak in the early 1970s. The lower estimate, 150 billion barrels, would have yielded a peak in the early 1960s. Most of us give Hubbert the benefit of the doubt and cite the early 1970s peak; the actual year of peak US production was 1970. The 150 billion barrel estimate, however, was a better fit to the data available in 1956.

By 1967, the US exploration and production history was far enough along for Hubbert to dispense with the expert opinions and estimate the size of the resource from the discovery and production data alone. Surprisingly, to this day some critics claim that all Hubbertian estimates require an advance guess at the total oil resource.

It was not until 1982 that Hubbert published a long paper explaining the math behind his analyses. It is typical of Hubbert that we do not know how much of the math he utilized at the beginning and how much he worked up later to justify his earlier efforts.

I was not the first to push Hubbertian methods to examine the world oil peak. Colin Campbell, Jean Laherrere, and several others published pioneering analyses in the late 1990s. My own route was cowardly: Stay very close to Hubbert's original methods to avoid the criticism that, "all those predictions turn out to be wrong." One large thing and one small thing turned up. As explained on page 49 of my book Beyond Oil, Hubbert's image of the discovery curve as a time-shifted version of the production curve cannot be correct. Oilfield discoveries, additions to reserves, and production all start at the same time, but their peaks are about 20 years apart. The small thing was a mathematical convenience. On page 38 of Beyond Oil, three lines of high-school algebra replace pages of calculus in Hubbert's 1982 paper. No longer is there any mathematical mystery behind the basic Hubbert methodology.

Another frequent misinterpretation of the Hubbert approach concerns the mirror symmetry between the upside and downside parts of the bell-shaped curve. Some analysts devoutly wish the decay side to be slower than the growth side. The Hubbert method uses the hypothesis that the ease of finding oil is proportional to the fraction of oil that remains undiscovered. That statement seems to fit for most major oil-producing countries. If it is true, then the mirror symmetry follows automatically.

One would have thought that 50 years would be enough time for us to resolve the major oil-supply questions. An unfortunate contrary reminder came early this month. On March 1, 2006, The New York Times posted on its subscribers-only website (Times Select) an essay by Robert B. Semple Jr. with the conclusion that the peak of world oil production is upon us. (I confess to a bias: Semple quoted my stuff.) Semple is a senior editor at the Times and a Pulitzer prizewinner. On March 2, the Times op-ed page ran an ad from ExxonMobil headed "Peak oil? Contrary to the theory, oil production shows no sign of a peak." That hurt. Was the Semple piece buried to avoid offending ExxonMobil? The operators of one website said to hell with the copyright law; this news threatens to end the Industrial Revolution! The entire Semple essay is posted at The ad is at

I do have an apology to make for a line in my February Current Events comment on this website. After stating that the world oil peak had already occurred on December 16, 2005, I reported that the Bush administration hoped to double the direct solar electric generation from the present one percent to two percent by the year 2025. My fingers got away from me and typed out: "By 2025, we'll be back in the Stone Age." I'm sorry that some readers thought that I actually meant that we would be wearing furs and hunting buffalo with flint spear points. It's called "hyperbole." Nevertheless, I have been looking into acquiring some property on the Arkansas novaculite belt. Great flint.

My conclusion: Hubbert was a great man, but with enough flaws to be really, really interesting.

© 2001—2011 - K. Deffeyes All Rights Reserved.
SD  site by Deffeyes Design