A free-market energy blog
Random header image... Refresh for more!

The Undulating Oil Plateau: Peak without Decline

For some peak oil advocates who are nervous about the idea of a post-apocalyptic vision of society, it has become popular to argue for a peak and plateau rather than a peak and decline of 3–5% per year, as some of the original work postulates. This seems more palatable than calling for a global upheaval, Hollywood notwithstanding.

The original peak and decline scenario was based on the bell curve popularized by M. King Hubbert. A number have disputed the shape of the curve, arguing for a Gaussian curve instead, for example. But they are avoiding the basic question of causality. The appearance of a bell curve appears to be more coincidence than anything else, since it is not often replicated in reality. The 1998 Scientific American article, “The End of Cheap Oil,” by Colin Campbell and Jean Laherrere, contained the laughable figure of several stylized oil fields’ production curves surmounted by a bell curve and the assertion that the one aggregated to the other.

More recently, some peak oil advocates have ‘modeled’ national production as following a rise, plateau, decline shape that, in the most general sense, is accurate, but again assumes that all nations follow a fairly similar path, which implicitly assumes that geology determines that path. In fact, in many cases the level of the peak and subsequent production patterns are due more to fiscal terms than geology, as can be seen by several countries where changing government policies led to a reversal of the decline, such as in Argentina or Venezuela.

The CERA work does not fall into this pattern, however, not just because the author(s) are not in the peak oil camp in any form. Rather, they have attempted to produce a long-term oil supply forecast, out to 2030, and discuss what might occur afterwards. And to paraphrase what Samuel Johnson said about a dog walking on its hind legs: one is less inclined to criticize the performance than to marvel at its being accomplished at all.

Historically, long-term oil supply forecasts have proven useful for only five to ten years at best, and often much less than that. As I have shown in various papers (see below), the difficulties of predicting not just political and fiscal changes but technological progress has made it extremely difficult for even the best constructed models to develop valid forecasts. Peak oil models have proven the worst, but econometric models have fared only a little better.

The primary conclusion from forecasts like that produced by CERA is that the authors believe that growth can be seen for the next two decades, and beyond that they are less certain. An undulating plateau in oil production is possible, but would almost certainly result from the availability of substitute supplies or more efficient equipment.

Similarly, when industry officials predict growth in production to, say, 100 million barrels per day (mb/d) and then a peak, what they are really saying is that they can see growth of 15 mb/d but are uncertain beyond that.  The interpretation of this as a precise prediction of a peak is nonsense.


Lynch, Michael. “Forecasting Oil Supply:  Theory and Practice,” Quarterly Review of Economics and Finance, July 2002.

Lynch, Michael. “The Analysis and Forecasting of Petroleum Supply:  Sources of Error and Bias,” Energy Watchers VII, www.iceed.org.


1 Chris { 12.03.09 at 12:24 pm }

What person would call his son “King”?

2 Don Jackson { 04.18.10 at 12:00 am }

I think this sums up how seriously to take this article:

Leave a Comment