Modeling road surface temperature from air temperature and geographical parameters-Implication for the application of floating car data in a road weather forecast model

Abstract

Precise forecasts of road surface temperature (RST) and road conditions allow winter roads to be maintained efficiently. The upcoming "big data" application known as "floating car data" (FCD) provides the opportunity to improve road weather forecasts with measurements of air temperature Ta from in-car sensors. The research thus far with regard to thermal mapping has mainly focused on clear and calm nights, which occur rarely and during low traffic intensity. It is expected that more than 99% of the FCD will be collected during conditions other than clear and calm nights. Utilizing 32 runs of thermal mapping and controlled Ta surveys carried out on mostly busy roads over one winter season, it was possible to simulate the use of Ta and geographical parameters to reflect the variation of RST. The results show that the examined route had several repeatable thermal fingerprints during times of relatively high traffic intensity and with different weather patterns. The measurement time, real-time weather pattern, and previous weather patterns influenced the spatial pattern of thermal fingerprints. The influence of urban density and altitude on RST can be partly seen in their relationship with Ta, whereas the influence of shading and sky-view factor was only seen for RST. The regression models with Ta included explained up to 82% of the RST distribution and outperformed models that are based only on the geographical parameters by as much as 30%. The performance of the models denotes the possible utility of Ta from FCD, but further investigation is needed before moving from controlled Ta measurements to Ta from FCD.