An assessment of the quality of real-time mobile pavement temperature observations from snowplows

In an effort to improve upon currently utilized weather and road condition products and to increase the density of weather observations near the surface, the quality of Automatic Vehicle Location (AVL) data supplied by Colorado Department of Transportation (CDOT) snowplows was assessed. These data contained air temperature and pavement temperature observations from an external sensor package and was supplied by a commercial AVL provider.

In order to verify the snow plow observation for accuracy, each individual AVL observation was matched to the closest (spatially and temporally) Road Weather Information System (RWIS) measurement automated weather station measurement. Basic quality control (QC) techniques were then applied to the matching pairs of data, and if either the AVL or the RWIS measurement failed the QC, then the matching pair was not included in the final analysis. The analysis was conducted on a dataset spanning the 2008/2009 winter season (November 1, 2008 – February 27, 2009).

Upon completion of the QC step, basic statistical analyses were performed on the matched observation pairs from the CDOT AVL and the RWIS locations. The results indicate that both air temperature and road temperature measurements by the snowplow fleet sensors correspond fairly well to QC-passed RWIS observations. Although little is known about the maintenance and cleanliness of the snowplow sensors, these results show that mobile fleet observations could indeed prove to be valuable in areas without direct RWIS observations, as well as in areas with high variability between RWIS locations (e.g. variable topography, coastal areas). These results are encouraging because these relatively inexpensive mobile sensors could possibly be used to “fill-in-the-gaps” between RWIS observation locations and to also perhaps aid in real-time thermal mapping of the roadways.