Railroad accident rates for use in rail transportation risk analysis

Anderson, R.T. and C.P.L. Barkan. 2004. Railroad accident rates for use in rail transportation risk analysis. Transportation Research Record: Journal of the Transportation Research Board. 1863: 88-98.


Annual safety statistics published by FRA provide train accident counts for various groupings, such as railroad, accident type, cause, track type and class, train length, and speed. However, hazardous materials transportation risk analysis often requires more detailed accident rate statistics for specific combinations of these groupings. The statistics that are presented enable more precise determination of the probability that Class I and non-Class I railroad freight trains will be involved in an accident on various classes of main-line track. An increase in the overall accident rate from 1997 to 2001 can be largely attributed to the increase in yard accidents. During that time, the main-line derailment rate for Class I freight trains remained nearly constant. Track class-specific derailment rates for Class I main-line freight trains show two orders of magnitude difference between the lowest and highest FRA track classes. Depending on the risk analysis question, accounting for these differences in rates will often be important in developing an accurate estimate of risk over the length of a route or at particular locations along a route. A sensitivity analysis suggests that the distribution of freight train miles by FRA track class may have changed since a study conducted by the Association of American Railroads in the early 1990s. More up-to-date estimates of track class-specific accident rates would require new data on this distribution.