Wednesday, April 11, 2007

What was Hubbert Trying to Do?

One thing we must remember in assessing Hubbert's original predictions is that they were not made in a vacuum. Alongside those predictions, Hubbert had data that indicated the maximum estimated URR (he used at least two values - 150 GB low and 200 GB high), the average annual rate of growth of consumption, and the total oil consumed by the date from which he was writing. Now what Hubbert did was extremely simple. He applied that rate of growth continuously to our oil consumption until the sum of prior oil consumed plus growing consumption equaled approximately half of the estimated URR. Using that date, he then predicted peak production in the US based upon that date. He repeated this exercise for a low URR estimate and a high URR estimate, giving him a bracket and then he made his prediction.

This is fairly simple stuff when you have an initial URR estimate and an average rate of growth that you do not expect to change drastically.

But what about Hubbert Linearization, as it is sometimes called? Well, this appears to be the result of a misunderstanding by Professor Deffeyes as to what Hubbert was trying to accomplish as well as what he actually did accomplish. Hubbert did not try to predict URR. Deffeyes does. This is one of Deffeyes' errors. An HL plot is not useful until it falls within the range of predicted high/low estimated URR. Thus, the URR estimate from an HL plot is not useful unless you have an estimated URR to bound the plot. In other words, it is my belief that all that an HL plot can do is confirm a prior estimate but it can't act predictively by itself.

Further, as Robert Rapier demonstrated, an HL plot can give lots of false positives, even if it is in the bounded range for URR. One thing that an HL plot does appear to do is confirm peak after the fact. HL gives hints about possible peaks, but by itself I don't consider it definitive. For instance, there are lots of people who believe we are at peak right now or very close (within a few years). Almost none of those people rely solely on an HL plot to arrive at that conclusion. Instead there is a large body of evidence, of which the global HL plot is just one piece. None of the pieces in isolation is "proof" and the entire body of evidence is circumstantial. But it is very strong. We are seeing Jeffrey Brown's "export land" model play out before our very eyes. Global discovery data is in a hard 40 year down trend. Reserve growth is starting to become reserve shrinkage as major reserves are reevaluated and restated, usually downwards. Despite record prices, production remains flat for nearly three years now, and indeed, peak C&C production was 23 months ago in May 2005. There is much much more as well and it all points to us being at peak or very very close, within a few million barrels per day tops.

So why is all of this important? Many of us believe it is important because so few people realize how integrated oil is into modern life and what a declining oil supply means. Those of us who do believe in peak oil also believe that our best chance to mitigate the bad effects is right now, not later. In short, many of us now speaking out do not want "mad max". No, we want to avoid it. But if society keeps its head buried in the sand, we may get mad max and even more.

No comments: