Friday, March 30, 2012

The Composition of US Oil Production

The above graphs shows US liquid fuel production, 1980-2011, broken out into its major components:
  • Crude plus condensate (actual oil coming out of the ground)
  • Natural Gas Processing Liquids (ethane, propane, butane, etc extracted from natural gas production).
  • Other liquids (mainly biofuels - almost all corn ethanol)
  • Volume changes from cracking heavy oil in US refineries - including both domestic and imported heavy oil.
You can see that at this point a large fraction of the liquid fuel numbers are not actually crude coming out of the ground.

This next picture shows the same data but putting the non C&C components onto the other axis so that they are not stacked on top of the C&C line:

This makes it clear that most of the production resurgence of recent years has come from the non-crude components. Biofuels have made the largest contribution, but NGLs and refinery gains have contributed also.


JoulesBurn said...

I know this argument isn't new, but I find "all liquids" to be a rather useless aggregate. Look at the input to US refineries and see how little NGL goes into it. The fact that some propane is used for transport is basically noise in the background. Crude demand is driven by gasoline and diesel. And while there is still some consumption of heating oil, the majority of the transition from that goes to natural gas, not propane. NGL is mostly petrochemical feedstock.

A useful aggregate is when the individual components share the same properties relative to use. This of course depends on what expects to learn from the aggregate metric.

Kenneth D. Worth said...

Another issue would be those elusive refinery gains. I mean, energy is consumed to turn a dense energy source into a less dense energy source that takes up more space, i.e. there are more barrels after the transition. Net energy is lost.

On the other hand, of course, if what goes in is residual from the first refining process, otherwise useless, and what comes out is gasoline, perhaps it makes sense to count refinery gains in the total fuels number.

Makes more sense than NGL's, at any rate, IMHO.

A Quaker in a Strange Land said...

One word. Wow.

Mike Aucott said...

Interesting and informative post Stuart.

Given that refinery gain is from total crude processed, shouldn't this portion be reduced to represent only that portion derived from U.S. crude?

The net energy available in these fuels is important, but, as has been discussed, these data are not readily available. It's possible to make some guesses, however, of how this chart would look if we could see the net energy:

* The other portion would be about 1/3 as big, since the EROI of ethanol is about 1.5.

* The refinery gain portion would essentially go away, since the sum total of energy available in crude is not increased, and the crude portion would be reduced by about 10% to account for the portion of crude that is consumed in the process of making the less dense fuels that provide the refinery gain.

* The crude portion would be reduced further, perhaps by about 5% in the recent years, since the EROI of crude produced today in the U.S. is probably about 20:1 at best. In the early years of the period, the reduction would be less, since the EROI was probably in the range of 30 or 40 to 1 in the 80s.

* The NGLs portion would be reduced too to account for EROI, but probably by less recently than formerly, since - it appears - the EROI of shale gas is pretty good.

Stephen B. said...

I'm with Greg....Wow.

That second graph is really a stunner. The proportions of our liquid fuel flows have made a remarkable transformation over the years.

That said, Mike Aucott raises some good points about including refinery gains from refining imported oil, along with essential questions of EROEI. With Mike's observations in mind, I think we can gather that the overall BTU content of US oil/liquid fuel production coming out of the ground, is quite a bit lower than what a simple measurement of US liquid fuel production statistics over time alone would hint at. This is probably another reason that "increased" US fuel production hasn't done as much to dampen pump prices as some might have hoped. (The other reason, of course, being increased world demand for liquid fuels overall.)

This is another excellent, informative post Stuart. Thanks!

Stuart Staniford said...

It's important to be really clear on that second graph that the C&C component is *behind* the others, not stacked on top. (The other three are stacked on each other). For a regular stacked area graph, see the first one.

@Joules - yes, all liquids is somewhat unsatisfactory. However, it's widely available from multiple sources and there's something to be said from that. And for certain purposes it's more meaningful than just looking at C&C. Eg as far as satisfying demand goes, ethanol clearly does count as we do in fact burn it in our cars. Refinery gains to some extent represent the fact that heavier crudes have more energy content than lighter crudes. The NGLs are the most dubious component - but even there if we weren't using them for petrochemicals, we'd probably be using a crude-based feedstock instead, so there is some substituability.

All-in-all I think it's best to look at all-liquids, but also maintain some awareness of how it breaks down and how the subcomponents are trending.