Citation

Holbrook LF. Transp. Res. Rec. 1977; 623: 29-39.

Copyright

(Copyright © 1977, Transportation Research Board, National Research Council, National Academy of Sciences USA, Publisher SAGE Publishing)

DOI

unavailable

PMID

unavailable

Abstract

Both urban and rural state trunkline intersections are examined with regard to their wet accident percentages. The examination first takes account of the estimated percentage of highway surface wet time for each month. Because precipitation data are available only for designed time intervals, a method is developed to convert these data into percent wet time - a factor necessary in assessing wet surface exposure at intersections. Using this conversion method the precipitation data from 120 of Michigan's Weather stations are transformed to give a month by month wetness profile for the entire state for the years 1963 to 1974. The range in monthly wetness for this period is from less than 1 percent to more than 25 percent. This potential 25 to 1 ratio is very influential in wet accident incidence and should be taken into account before other variables are examined. Nearly 40,000 accidents occurring at over 2,000 intersection locations for which a skid coefficient value was available were tabulated to provide wet accident percentages. These data together with the location's wet time percentage, as estimated from the nearest weather station, provide an opportunity to statistically fit a wet accident model for the variables included. The fit is satisfactory and suggests an accelerating function for skid coefficient. For all levels of wetness, a skid coefficient less than about 30 is accompanied by an acceleration increase in wet accident percentages; although the actual shape of the curve depends on wet time. The model appears useful in designing cost-effective intersection resurfacing plans which minimize wet accidend occurrence.