Weather Codes and Travel Behavior

Abstract

In this report an analysis is presented on the impacts of weather codes on travel behavior. Loop detector data of seven segments in different provinces in the Netherlands are analyzed with regression models with autoregressive errors. Weather codes were found to be significantly influencing travel demand. Particularly codes orange for slipperiness and snow, and codes red for slipperiness, snow and wind yield significant results for most road segments. Furthermore, some trip rescheduling behavior was observed. The unreliability of the previous weather code was found to reduce the impacts for the next weather codes. The analysis of Twitter data was not useful to confirm hypotheses on the incentives for changing travel behavior. As this study is the first revealed preference study into the effects of weather codes, lots of research gaps remain. Secondary roads were not included in the study, which might be affected in other ways than highways. Furthermore, interaction effects between time and weather effects might explain some of the unexplained variability of the model results.