Net emissions, Keystone XL and refining heavy oil

Last week, in a speech delivered at Georgetown University, U.S. President Obama laid out a plan to reduce greenhouse gas emissions. The plan itself includes a wide range of promised actions, among them a carbon standard for power plants, renewable energy projects on federal lands, and fuel economy standards for heavy duty trucks. In his speech, however, President Obama touched on a specific and controversial question that was not addressed in the published White House plan when he remarked that the Keystone XL oil pipeline would only be approved if it can be demonstrated that it “does not significantly exacerbate the problem of carbon pollution.”

That carefully worded phrase predictably touched off a good deal of discussion in the U.S. and Canada as to what it revealed about the president’s leanings on the question (Michael Babad at Canada’s Globe and Mail reported that many analysts still expect the project to get the go ahead, while David Roberts at Grist described it as ‘a sort of Rorschach blot for the energy world’). But it also serves to put a spotlight on the question of what we know or can reasonably estimate, so far, about what Keystone XL would mean for “carbon pollution.”

The purpose of the proposed Keystone XL pipeline is to improve market access for bituminous oil (bitumen is a very viscous, degraded form of oil) produced from Alberta’s tar sands, and much of the controversy around the project has come because it is widely recognized that both the extraction and refining of bituminous oil are high-carbon-intensity exercises. As described in the State Department’s assessment of it, the pipeline would carry 830,000 barrels per day of oil, largely bitumen-based synthetic crude and “dilbit”, a mixture of raw bitumen and a diluent such as naphtha. Given President Obama’s statement, a key question is whether this increased influx of bituminous material will increase net global greenhouse gas emissions, and by how much. Tar sands might increase net emissions because it is more carbon intensive to extract oil from tar sands than from most other oil resources (for instance many new projects use ‘Steam Assisted Gravity Drainage’ in which large volumes of steam are injected into the ground to soften the bitumen), because some tar sands bitumen goes through energy intensive upgrading processes before being sent to refineries, and because it takes more energy than for lighter oils to refine heavy tar sands crude into gasoline, diesel, and other fuels. 

One part of oil sector emissions comes from refineries, and we can estimate the change that Keystone might cause in those using results from a study we published in February, carried out by the refinery consultancy MathPro. The study analyzed the effects of possible changes in the set of crude oils entering U.S. refineries and the mix of gasoline, diesel, and other products they produce—generally termed the crude oil and refined product “slates” respectively. Crude oil slates are called “heavy” or “light” depending on the proportion of light oil to heavy oil – with tar sands bitumen being at the heaviest end, and fracked tight oil typically at the lightest. 

For our “heavy” slate scenario, we assume that by 2025 the processing of Canadian oil sands crude in the U.S. is about 2.2 million barrels per day higher than in our baseline case – a difference of about two and a half times the capacity of Keystone XL. The Canadian heavy oil displaces from U.S. refineries a combination of Atlantic basin oils (including Venezuelan heavy), and light tight oils from fracking in the U.S. In our central case for the 2025 mix of refined products, this results in an emissions increase of about 13 million tonnes CO2e per annum. Scaled to the size of Keystone XL (which would supply only 40% of the increase we modeled) you could be talking about a U.S. refinery sector emissions increase of the order of 5 million tonnes CO2e per year. That’s like putting another million vehicles on the road (EPA estimates average annual passenger vehicle emissions at 5.1 metric tonnes). 

We should note that this calculation only works if you treat all the oil coming through Keystone XL as additional. The State Department, in contrast, has argued that in the absence of Keystone XL the bitumen and synthetic crude will be transported by train, and will come to U.S. refineries anyway. Rail is a substantially more expensive way of transporting bitumen, and it would need a significant increase in rail traffic to match the volumes that Keystone Xl would be able to move. The Globe and Mail recently reported on an analysis of the situation by Canadian wealth management firm RBC Dominion that found that “a third of oil sands growth, or 450,000 barrels a day, could be put on hold between 2015 and 2017 if TransCanada Corp.’s Keystone XL pipeline is not approved by U.S. President Barack Obama.” If we take this as a reasonable estimate of the marginal impact of a Keystone XL rejection, then half of the refinery emissions increase we calculated above could be realized – i.e., 2.5 million tonnes CO2e per year, equivalent to putting 500,000 more cars on the road. In the longer term, a sustained pattern of regulatory action to discourage high-carbon oil extraction could shift oil industry investment decisions even more than this. 

Just to reiterate, we’re only talking here about part of the carbon impact of Keystone; there’s also the question of how much carbon is emitted when bitumen is extracted from the tar sands by mining or steam injection. The eventual real outcome would depend on a number of specifics: the crudes displaced by bitumen from Alberta, for example, and—most crucially—whether Canada found an alternative outlet for some of those 450,000 barrels a day. So we can expect to see a lot more debate over numbers. Still, this should be a good first approximation of what the Keystone pipeline would mean for U.S. refinery emissions.


GHG emissions