Nebraska Rail Crossing Safety Research

Safety at the junction of highways and rails has been a concern for a long time. In the United States, more than 97% of these crossings are at-grade, meaning they are at the same level. While trains have the right-of-way, every year a number of crashes occur when highway users fail to yield the right-of-way to trains. Crashes at rail crossings are invariably more severe compared to crashes on the rest of the transportation network due to train involvement. In 2015, the number of crashes reported at rail crossings in the United States was 2,060, resulting in 237 fatalities; fatal crashes were 11.5% of the total reported incidents. Rail crossing safety models have been around since the 1940s. Some of the more prominent models include the Peabody Dimmick Formula, the National Cooperative Highway Research Program (NCHRP) Hazard Index, and the U.S. Department of Transportation (USDOT) Accident Prediction Formula. Based on crashes reported at rail crossings, these models predict future crashes at rail crossings; the results provide a ranking of competing rail crossings for the expenditure of limited safety funds. The Nebraska Department of Transportation (NDOT) currently uses the Nebraska Accident Prediction Model for rail crossings to identify and rank crossings that may need scrutiny and perhaps subsequent safety improvements. Developed by the Midwest Research Institute (under contract to HNTB Corp.) in 1999, this crash prediction model was based on 5-year rail crossing accidents and inventory data from September 1993 through August 1998. It updated the previously used 1973 Nebraska Department of Roads (NDOR) Hazard Index (a modified version of the NCHRP Report 50 Formula). The model over-predicts (about 10%), and results may not be optimal as many changes have occurred in terms of train and motor vehicle traffic, crash trends, and rail crossing inventory information. Other state DOTs (e.g., Iowa DOT) are in the process of updating their rail crossing crash prediction models. For acceptance and adoption by NDOT, the new model must outperform the existing NDOT Nebraska Accident Prediction Model for rail crossings. The proposed research has two objectives: (1) update NDOT’s Nebraska Accident Prediction Model for rail crossings using the latest crash and rail crossing inventory data, and (2) develop guidelines for improving safety (via uniformity of driver expectations) at urban rail crossings that are not Quiet Zones but are in vicinity of other Quiet Zone crossings.