A few weeks ago I put up a post that summarized some results from playing around with a simple depletion model for oil production from a shale formation. I'm not the only one who plays with toys like that, of course. On January 10th of this year, the North Dakota Department of Mineral Resources made a presentation to the North Dakota House Appropriations Committee (PDF). I used to work for the Colorado state legislature's Joint Budget Committee -- trust me when I say that the department did their best to give the Appropriations Committee an accurate picture of how things were most likely to unfold in the Bakken Shale oil play in western North Dakota in the near future. Legislators, particularly those that decide on executive-branch spending levels, dislike getting bad information and seem to remember forever that you once gave them bad information.
Here's the slide that summarizes the expected outcome, used in the sense of what is most likely to occur. Once drilling was ramped up to its full level, the shape of the curve is a close match to those generated by my toy model. This isn't surprising; there seems to be good agreement among geologists about how fractured shale wells behave. Production peaks at 850,000 barrels per day in about 2015, then begins a steady decline. This behavior reflects an assumption (which appears on the slide preceding the one shown) of a constant 2,000 new wells per year. Other scenarios in the presentation vary the number of new wells per year, with optimistic forecasts having more and pessimistic ones less. All of them have an important characteristic in common: the peak of North Dakota production happens sometime in the next several years.
More interesting is the effect that is left out in this chart. Down at the bottom there's a statement that 40,000 more (than exist today) wells are possible in the "thermal mature area." Thermally mature is jargon that means "there's oil there." At 2,000 wells per year, though, it only takes 20 years to drill 40,000 more wells. That chore will be done by 2033 or so. What's missing in this chart, that extends all the way to 2055, is any change in production when the drilling stops. My toy depletion model, and assorted professional ones, show what happens when there are no more reasonable places to drill: exponential decline in production. I can understand why the department chose not to include that in their model. It's a scary looking collapse. The Committee isn't going to make decisions about today's funding based on what's going to happen in 20-30 years. Sometimes politicians engage in a "kill the messenger" response to bad news. Still, it would have been nice to include, so the politicians understand that the end to this particular party happens within many of their lifetimes.