92nd American Meteorological Society Annual Meeting (January 22-26, 2012)

Wednesday, 25 January 2012
Assessing the Significance of Pavement Composition Variability on Gridded Road Weather Forecasts
Crystal Burghardt, NCAR, Boulder, CO; and S. Linden and M. Chapman

Model of the Environment and Temperature of Roads (METRo) is a numerical model distributed by Environment Canada to forecast road conditions and has been used on an operational basis since 1999. METRo obtains surface temperature observations from Road Weather Information Systems (RWIS) stations, which are similar to other common fixed weather stations, but are generally deployed along roadways. METRo uses these observations, as well as an atmospheric forecast and unique road composition information, to produce road weather forecasts which are specific to that point. Because the model will likely produce gridded forecasts in the future, it is important to determine the potential impacts of expected differences in pavement composition under each grid box. The model could, if necessary, generate a pavement forecast with uncertainty information that is based on the variability of pavement characteristics. Mean absolute errors were computed and analyzed and time series plots were generated for each change in parameter value to determine the sensitivity of the model to variations in road composition (e.g. pavement type, pavement thickness).

Three road segments in Colorado (two 2-layer segments and one 3-layer segment), each with different compositions, were selected for analysis on three different dates under varying weather conditions. METRo appears to be slightly sensitive to changes in pavement layer thicknesses, but in most cases, this sensitivity is negligible. The model does appear to be significantly sensitive to changes in the pavement's surface material (e.g. concrete and asphalt). As a result, changes in surface material should be considered when providing a generic gridded pavement temperature forecast to the end-user. A more robust study incorporating a larger data set from Minnesota will be analyzed using similar techniques and these results will also be included.

Supplementary URL: