How much crude is moving by rail in Canada, of what kind, and where is it going? The statistics, cited in most reporting on oil-by-rail, claim that over the past five years, oil shipments by rail across Canada have increased by a huge proportion, as much as 28,000 per cent, typically rising from 500 to 140,000 carloads a year between 2008 and 2013.
Given the limited evidence, however, all the claims are misleading. The 2008 number may be right, but the 2013 number is certainly too high. It includes other petroleum products besides crude, as well as both heavy crude moving from Alberta to U.S. refineries and highly volatile light crude from the Bakken field in North Dakota moving to Atlantic Canadian refineries, principally the Irving facility in St. John, New Brunswick.
Governments are now scrambling to fix the holes in the regulatory framework revealed by the accident in Lac-Mégantic. Better regulations depend on better data.
Small amounts of crude have always moved by rail, but the failure of pipeline capacity to keep up with the rapid increase in oil production is behind rising shipments of crude by rail, with predictions of more to come. Reliable data from the Association of American Railroads (AAR) suggest that U.S. crude shipments increased from less than 10,000 car loads a year in 2008 to a likely total of about 400,000 carloads in 2013, as shown in chart in the attached infographic. Most of the increase is light crude moving from the Bakken field, but some is Alberta crude largely carried by CN – perhaps 30,000 carloads in 2012 according to National Energy Board estimates.
The AAR data include the U.S. operations of Canadian railroads (CP is a major shipper of Bakken crude), and shipments exported to Canada, but AAR does not know how much crude is moving by rail in Canada. So what do we know about the Canadian movements, which, as shown in the chart, have also increased since 2011?
In its time series on monthly railway carloadings, Statistics Canada lumps "fuel oils and crude petroleum" together. In the 12 months to July, 2013, there were 148,785 such carloads, close to the number quoted in news reports, but, in 2008, there were 64,368 such carloads. That time series is based on question 28 in a monthly survey that aggregates two items: 1) fuel oil; and, 2) crude petroleum, including Naphthenate, the diluent added to oil sands crude to allow it to flow in pipelines. We can be sure, therefore, that shipments of crude alone in 2013 were less than 148,000 car loads, but it may be safe to infer that the increase since 2008 – approximately 85,000 carloads – is largely shipment of crude oil by rail. That number is somewhat less than the widely cited 140,000 carloads, but it is still a significant increase.
The next question is about the kind of crude being shipped, and where it is going. Statistics Canada uses the same aggregation in its Annual Rail Commodity Origin and Destination Survey, which is published with a two-year time lag. In 2011, the last year available, most of the roughly 65,000 carloads of "crude petroleum and fuel oil" was shipped within Canada, and no Bakken crude moved to Atlantic Canadian refineries. When 2012 data becomes available later this year, it will probably show that Bakken crude is now moving in significant volumes to Atlantic Canada, or so we infer from the monthly Refinery Supply of Crude Oil and Equivalents survey that shows that Atlantic Canada refineries began receiving U.S. crude only in June, 2012. That time series will not indicate how much of that crude would have come via largely U.S. routes, including by transhipment on coastal barges.
We do know, however, that that crude oil from the Bakken Shale in North Dakota may be more flammable than other types of crude, according to a U.S. safety alert issued on January 2, 2014. Regulators in Canada are therefore challenged on two fronts: they have limited knowledge of the properties of Bakken crude and limited knowledge of how much Bakken crude is moving along thousands of kilometers of our railroads.
If Canadians are to better understand the regulatory implications of this new pipeline on wheels, we need better numbers, and we need them sooner. That's why Statistics Canada needs to revise its questionnaire to split crude oil from fuel oils. And it needs to analyze and release its data sooner.
Musfiq Islam is an MPA candidate and Robert Wolfe is a professor in the School of Policy Studies at Queen's University in Kingston.