Modeling Highway Safety and Simulation in Rainy Weather

In risk analysis of highway safety, severe crashes with injuries and fatalities are more emphasized because of their significant human and economic loss potential. In Wisconsin from 1999 to 2006, 4,919 crashes with injuries and fatalities occurred on highways in all adverse weather conditions: rain, fog, snow, sleet, and wind (3). Of these crashes, 1,805, approximately 37%, were crashes with injuries and fatalities occurring in rainy weather. This proportion is the highest for all adverse weather conditions. During the same period, 899 multivehicle crashes occurred specifically on Wisconsin Interstate highways in rainy weather; the multivehicle crash frequency is approximately 1.5 times that of the single-vehicle crash frequency. Moreover, the proportion of severe multivehicle crash occurrences is 1.4 times more than the proportion of the severe single-vehicle crash occurrences. Regardless of the number of crashes in rainy weather, comparatively few studies have been conducted in a disaggregated fashion to assess rainfall-derived factor effects on highway safety, particularly crash severity. Therefore, the objectives for this research were to develop a novel methodology for microscopic weather data estimation and to apply this methodology to statistical modeling of highway safety and weather-sensitive parameter adjustments during traffic simulation.