Office of Operations
21st Century Operations Using 21st Century Technologies

Regional Assessment of Weather Impacts on Freight

Chapter 4. Conclusion

The analysis presented in this report indicates that, overall, weather events have a significant negative impact on traffic speeds—and, therefore, the freight industry—when analyzed at the regional level. In this report's study areas alone, decreased traffic speeds due to the weather events analyzed are estimated to cost the freight industry 3.8 million dollars per year. This estimate is relative to the average cost of trucking, which includes both good and bad weather. To put this estimate in some context, the roadways covered in this study constituted roughly 1,500 miles, or just under one percent of the National Highway System (NHS). Approximately 44 percent of all traffic moves on the NHS so this analysis does not represent delays that occur on regional and local highways.

The regional analysis allowed for a more detailed investigation of how the impacts of weather on freight vary by weather event, highway type, time of day, and region size. The key takeaways from the regional analysis are listed below, followed by a discussion of important considerations and limitations of the analysis and directions for further research. All key takeaways are overall findings from the analysis of all 13 regional study areas.

  • Weather events that fall into the categories of Ice and Snow, Fog, Flood, Wind, Rain, and Extreme Temperature were, together, associated with the vast majority of traffic speed decreases during weather events, as well as costs to the freight industry from weather-related delay.
  • Ice and Snow events were associated with over half of all lost time due to decreased traffic speeds during weather events and are the most costly for the freight industry (costing over 25 dollars per segment hour and over 25 cents per truck per segment).
  • Weather events exert the largest negative impacts on traffic speeds between hour 0 and hour 1. However, small decreases in speed are also seen in the hours leading up to a weather event, and moderate decreases in speed are still seen up to four hours after the event.
  • Throughout the day, irregular flow highways (highways that experience morning and evening rush hours) suffer more than even flow highways in terms of loss of speed during weather events.
  • Time of day matters—all highway types suffer more in terms of loss of speed during weather events that occur during morning and evening rush hour periods.
  • Highways in smaller regions (where region size is based on population size and economic intensity) suffer less than highways in medium and large regions in terms of loss of speed during weather events.

While the analysis was able to detect these trends looking across the 13 regions, it is important to understand the limitations of the methodology employed and the data used. The association of weather events with traffic speeds based on county means that it is not certain that traffic speeds on a given roadway were always directly impacted by the associated weather event. The eight-hour timeframe applied to each weather event (four hours before and after hour 0) also limits the analysis as storms vary in length, but it was necessary to establish a common timeframe in order to have consistency in the analysis. Ideally, the analysis would track each storm individually for the time that it occurred, but the resources required for such a detailed analysis were not adequate for this effort. Nonetheless, given the sample size amassed using these assumptions, the research team was able to determine the significant trends listed above.

The analysis was based upon weather events reported in the National Storm Events Database. These are storms that are significant by several criteria, including having an impact on the State’s commerce. While these are low-frequency, high-impact events, there are many smaller events that are not part of this analysis. These smaller events may not generate significant delays individually but are far more numerous in total and may account for a significant amount of delay. It should also be noted that catastrophic events that shut down roads entirely are not included in the analysis since no traffic was flowing. While events that close highways altogether are very rare, the amount of delay they incur for the freight industry could be significant.

It appears on the surface that this analysis would lead to a somewhat lower estimate of national weather-related freight delay than the $8-9 billion estimated in the previous study. However, some of the caveats noted above, including the limited sample of events and the use of average hourly traffic speeds, probably bias the analysis toward a lower estimate. Significant variability in freight flow estimates is another factor that introduces uncertainty, although the direction of this bias is not known. In summary is not possible to scale the regional delay estimate in this study up to a national estimate. However, the research in this study could be used to develop a national estimate with greater confidence than that of the previous study. A next step would be to match the delay factors by type of event and match them with the weather exposure factors developed in Task 8 of this study. A set of national weather zones could be developed and the delay and exposure factors applied to highways within that zone. A reasonable approach would be to start with the Interstate system and eventually expand to the National Highway System.

Additional future research could try to hone in on smaller geographic areas. This would allow for the use of radar data to precisely track the time and locations of weather events. Additionally, the lack of robust data on trucking volumes and the value of cargo presented challenges in crafting the methodology for this analysis. While research will have to rely on models and gross estimates for the foreseeable future, the further development of trucking data will greatly aid future efforts to continue this research.

Office of Operations