Peak Oil – When Will We Run Out? (November 2008 Edition)

(11/16/2008) - Ever since the use of oil as a major transportation fuel began in the early twentieth century, we have known that we will some day run out.  This was clear after observing the initial surge of production wells offered and their subsequent deterioration. 

Predicting the demise of oil has been a long-standing American tradition.  Leonardo Maugeri, senior vice president of the Italian oil company Eni, writes in his book “The Age of Oil” that a World War One era senate investigation concluded that oil production in the United States has already peaked and that supplies would be depleted within 25 years.  In 1919, the head of the U.S. Geological Survey predicted the end of American oil production would occur by 1928.

Of course, none of this happened.  These estimates were made before the science of oil exploration has come into its own.  But not all such predicti predictions.  Author Kenneth S. Deffeyes writes that when M. King Hubbert made his 1956 prediction that oil production in the U.S. would peak between 1965 and 1970, many people ridiculed him for it.  Looking back, we know that U.S. production peaked in 1970.

Lucky for us, increases in worldwide production more than compensated for sinking American production.  And when it costs $2 a barrel for a supertanker to move oil anywhere in the world, it doesn't make much of a difference at the pump whether the oil was drilled in Saudi Arabia or Texas. 

But we know this can only go on for so long.

Peak oil is not just America’s story.  It’s also the story of Mexico, Norway, India, Egypt, Venezuela, Libya, Germany, Oman, France, Indonesia, Syria, and New Zealand production.  If we reach peak oil in country after country, then we will reach a worldwide peak oil. 

The Question is: WHEN?

There are many, many different ways that peak oil is estimated.  The simplest method is probably that employed by M. King Hubbert.  Unlike other methods, which look at factors including estimated reserves, geopolitical problems, and economic factors, Hubbert’s analysis is based on a simple method, as Deffeyes describes.  First, make a graph.  Let the X axis equal the total amount of oil pumped out of the ground to date.  Make the Y axis equal to P/Q, where P equals the production for a single year and Q equals the cumulative production for all years prior to that year.  Put it all together, and you get this for United States production from 1900-2007:

(Sorry, but I couldn't get the X-axis to work correctly.  The end point represents about 195 billion barrels of production)

And from 1945-2007:

Deffeyes then adds in a linear trend line to this second graph:

By looking at where this trend line intersects the X-axis,  Deffeyes estimates total U.S. production, and how much we will produce every year from here on out. 

While this model has proven to be fairly accurate and provides a good base for analysis, it consistently underestimates future oil production.  Hubbert considered his 1970 estimate very optimistic, and considered 1965 as the peak for U.S. production to be much more likely.  Hubbert predicted in 1969 that world oil production would peak between 1995 and 2000.  Using Hubbert’s methods, Deffeyes predicted that worldwide oil production would peak in 2005 (the year he was writing).  This has proven to be at least three years too pessimistic, as production in the first half of 2008 is the highest on record. 

The mistake I think that Hubbert’s model makes is that it uses a linear line of best fit.  Deffeyes basically ignores all the U.S. data before 1945 because of its more sporadic nature.  Looking at the first graph, I think it is is obvious that P/Q does not form a straight line – it forms an exponential line, one that approaches zero but does not meet it.

The most important thing Hubbert’s methods teach us is that peak oil is so much about geopolitics or the price of oil.  These sort of events create small variances, but do not change long-term trends.

What are some other estimates for global peak oil?  Lets start with the U.S. Federal Government.  The Energy Information Administration (EIA) provides us with all the Fed’s official energy statistics.  Their historical numbers are accepted by everyone I’ve read as accurate.  Lets look at their estimates: 

The EIA presents five different scenarios a reference case, and cases for high and low economic growth and oil price scenarios.  In their reference case, worldwide liquids production (including biofuels, shale oils, etc.) grows from 84.3 million barrels a day (mbd) in 2005 to 112.5 per day in 2030, for a cumulative annual growth rate of 1.2%.  Projections based on economic growth rates and oil prices follow a similar pattern of growth, but to different degrees.

Looking at some of their country by country estimates reveal that these numbers are very inflated.  American oil production has dropped at an average annual rate of 1.74% since 1970.  While this hasn't been completely steady, the drop has remained fairly consistent.  Since 1985, oil production has dropped every single year except one.  Yet the EIA estimates, for no real reason that I know of, that American production of conventional liquids (as opposed to "unconventional" liquids like ethanol) is going to begin to rise again, from 7.9 mbd in 2005 to 8.9 mbd in 2020, before sinking again to 8.2 mbd in 2030, for a cumulative growth rate of 0.1%.

Their analysis shows similar patterns of expected production for other countries.  Venezuela, whose production has been falling since 1997, has a comeback beginning in 2015, when the EIA predicts it will produce 1.7 mbd before rising to 2.4 mbd in 2030.  Egypt, the U.K., and other countries experience similar trends in the EIA’s projections.

All told, I’d say that the EIA’s estimates are the best case scenario, and probably too optimistic.  Its always possible that we will discover lots of new petroleum resources to mine, enough to compensate for the losses in existing wells.  However, an increase in discoveries would buck a 40+ year old trend:

(Source: The Oil Drum)

Most other peak oil estimates I’ve seen fall somewhere between the day they were written and 2030.  A recent study (LINK) by the U.K.-based Industrial Task Force on Peak Oil and Energy Security thinks peak oil will arrive between 2011-2013.

One of the reasons why it is hard to get a comfortable grasp on when peak oil might arrive is because the available data is not good.  Part of this is due to the fact that geology is still an inexact science.  Many of the world’s largest producers, especially the members of OPEC, jealously guard their reserves from 3rd party inspections, leaving the rest of the world to rely upon their declared reserves: 

(Source – Wikipedia) Mrd Baril = Billions of barrels

Two things about this graph jump out as fishy.  The first and most striking is the manner in which almost every country has, at some point, suddenly increased their reserves by upwards of 100%.  Why this occurred, mostly in the late 1980’s, is a mystery.  There wasn’t a miraculous new method of exploration that might have uncovered previously unknown reserves.  

The second troubling part of the graph is just how consistent the estimated reserves have been if you don’t include the sudden jumps.  Most of these nations are pumping billions of barrels from the ground every year.  It is extremely unlikely that every OPEC country would consistently find new reserves every year that just happen to be almost equal to the amount of oil extracted that same year.  For the sake of comparison, here is a look at the history of U.S. oil reserves:

I believe the spike in reserves that occurred around 1970 reflected the discovery of the Prudhoe Bay oil field in Alaska.  Even with this discovery included, the long term trends of U.S. oil reserves is clear, and similar trends are found in many other countries.  Only in the members of OPEC do you find flat lines. 

Another idea to consider is whether or not oil production will flatten, steadily decline, or crash after peak oil has been reached.  My general inclination is toward a steady decline, as that is what we’ve seen in the United States.  However, this is no real reason, just an anecdote.

I’d love to hear your opinion.  Do you think the oil peak is at hand, or do we have awhile yet?  If the price of oil gets high enough, will market incentives be strong enough to bring about the development of a way to use a carbon-based alternative like shale oil?

Part 1 of the Peak Oil Series - What is Peak Oil?
Part 3 of the Peak Oil Series - Review: The Age of Oil

Further Reading

---The Oil Drum's Peak Oil Update
---Our Finite World (I found this after I had already launched my site) - A collection of Peak Oil related articles by gailtheactuary, an editor at TOD.
---Cambridge Energy Research Associates (CERA) - Peak Oil skeptics.  See their rebuttal of peak oil theory (not eventual peak oil, but peak oil within the next 40 years).  TOD's reply to CERA's rebuttal.