Weather can have a significant impact on rural highway operations. For example, highway agencies in the United States spend more than $2 billion per year on winter maintenance activities. Adverse winter weather, along with other weather challenges such as fog, rain, and wind, can significantly increase crash frequency on otherwise benign roadways. Because of the significant variability in weather across years during the same season, it is difficult to quantify the effect that weather may have, positively or negatively, on highway operations. It is also difficult to “correct” for the effects of weather when identifying the advantages or disadvantages of various operational and safety improvements. A weather index for roadways would help in such situations. Although a weather severity index for roadways has not been developed, severe weather indices have been developed for application in many other fields, such as wildlife ecology, icebreaking on the Great Lakes, and degree days for household utilities. In every case, these indices tend to have higher values with more extreme (e.g. colder or hotter) weather. For surface transportation, extreme conditions may not necessarily be the best measure of weather severity. Moreover, while daily or monthly averages may be adequate for other applications, it is not clear whether this is true for surface transportation. The purpose of this study is to research and develop a weather severity index that would be appropriate for roadways. The desired attributes of the index are as follows.§ The index would directly correlate with winter maintenance costs. For example, a 10 percent increase in the index value might equate to a 10 percent increase in maintenance costs.§ The index would correlate with the relative hazards of driving conditions experienced by motorists, including snow, blowing snow, ice, frost, rain, fog and wind. This correlation could be shown by the relative accident frequencies on the same segment of roadway during years of significantly differing weather conditions.§ The index would use data that is already available, such as through the National Weather Service (NWS) or RWIS stations.§ The index would be easy to use and interpret.