Road surface conditions must be monitored constantly and classified accurately (icy vs. snowy vs. dry etc.) to give decision-makers the information they need to carry out highway winter maintenance effectively. In the USA and around the world, thousands of automatic road weather stations (RWSs) are installed at roadsides to serve this purpose. Since information on road "slipperiness" is critical for driver safety, the RWS information must be as accurate as possible.

Vaisala has designed and introduced a new model of pavement sensor to market. It measures surface water thickness on the pavement using an optical reflection method. The information on water layer thickness leads to improved accuracy in the measurement of de-icing chemical concentration in the surface solution. In addition, the new optical reflection method enables snow detection based on high reflection value. Both of these factors improve the detection accuracy of pavement conditions.

This paper describes the results of laboratory and field tests that were conducted using the Vaisala pavement sensor, described above, attached to a road weather station. Several different salt concentrations were measured in the laboratory with an accuracy of 10% or better. The water layer thickness measurement results are presented in the range of 0...8 mm.

In the field tests, the surface state reported by the RWS was compared against human observations. The road surface was classified as icy, snowy, frosty, wet (& chemical), moist (& chemical) and dry. The observations were collected over several winter periods in Columbus, Ohio, and in Finland. The Columbus results (total of 496 samples) show 97% agreement with the human observer.