How Safe are Canadian Railroads?

In the aftermath of the Lac-Mégantic disaster, there were several newspaper articles and op-eds discussing the relative safety of transporting crude by rail versus via pipelines. For examples, see this article on the risk of tightened rail regulations, this one on the acceptability of rail, one on the future of oil by rail, and one on relative safety. Many of the articles I found on the safety of pipelines versus rail made statements that were anecdotal in nature; one that did have facts were based on US data.

It is unfortunate that it has taken this disaster to make us consider how safe our transportation infrastructure is. And while it is a tragedy, we shouldn’t let it cloud our thinking. The reality is, rail is going to continue to be used as a transportation mode for crude oil and other dangerous goods. What matters now is what the safest method is, and whether regulations or policies need to be changed to make rail safer. This brings up the question, how safe are Canadian railroads?

To evaluate the safety of railroads in Canada, we want information on how often dangerous goods are transported and how much, number of accidents/incidents, whether product was released, and how much was released. Unfortunately, much of the data I would need is not publicly available, making it rather hard to judge the safety of rail as a transportation mode. The rest of this post is devoted to giving you a sense of what is available and what we can know about the transportation of dangerous goods by rail.

Statistics on rail transportation of dangerous goods are a little hard to come by. The top ten dangerous goods commodities in CANSIM Table 409-0008 are: liquefied petroleum gases, anhydrous ammonia, sulphuric acid (more than 51% acid), fuel oil, molten sulphur, methanol, sodium hydroxide solution, gasoline, chlorine, and styrene monomer (stabilized). Unfortunately, the carloads statistics (Table 404-0002) are not that specific, so I am going to approximate dangerous goods with gasoline and aviation fuel, fuel oils and crude petroleum, gaseous hydrocarbons and sulphuric acid. With the understanding that this is likely to be an underestimate, these commodities accounted for 7.8% of all car loadings in Canada and 7.5% of total tonnage transported.

Turning to accident rates, the number of accidents and number of incidents for federally regulated railroads are available from Table S3 in Appendix A of Transport in Canada, published by Transport Canada. In 2011, there were 1023 accidents and 204 incidents. Of the accidents, 117 involved dangerous goods and 51 incidents involved a dangerous goods leaker. Figure 1 shows the share of accidents involving dangerous goods or fire, and the share of incidents involving dangerous goods. While very few accidents involve fire, on average 13% of accidents involved a dangerous good. And while the share of incidents involving a dangerous goods leakage has been declining over time, it still seems very high. Of course, we don’t know how much was leaked in the case of incidents, and whether there was a leak/release in the case of accidents.

Figure 1: Share of Accidents and Incidents Involving Dangerous Goods or Fire

Unfortunately, that is the only information available from Transport Canada on rail safety – there may be more data, but the link to the statistics page was broken. Statistics Canada has data on “reportable” dangerous goods accidents. Definitions of “reportable rail accident” and “reportable rail incident” are in the Transportation Safety Board Regulations. Essentially, an accident is reportable if the amount released is greater than some lower bound, which is determined by the class of dangerous good, defined in the Transportation of Dangerous Goods Regulations.

CANSIM Table 409-0005 has data on reportable rail accidents by dangerous goods class. Between 2004 and 2011, on average 10% of dangerous goods rail accidents were reportable. So, 10% of 13% of rail accidents involved a release of sufficient magnitude to be considered reportable. A pretty small number. However, there are two things missing from this analysis: how bad the releases were, and what an acceptable risk level is. Transportation of dangerous goods is inherently risky, but how much risk are we willing to bear? This is an important question, especially given how many cities and towns have a rail line running through them.

To answer the question posed by my title, I have no idea. Unfortunately, there is very little data in Canada that can be used to evaluate the safety of railroads. While I am certain the analysts at Transport Canada have better data, and can better inform policy makers, it is concerning that the public has very little ability to evaluate this issue with publicly available data.

**As a caveat, I am not an expert on rail or rail safety. This is really just an exploration of the data that is publicly available to evaluate rail safety.