Geomatics inspire winter maintenance revolution

Dr. Lee Chapman, Technical Director, Entice Technology Ltd, Birmingham, UK
Dr. John E. Thornes, Reader in Applied Meteorology, University of Birmingham, UK

Countries with a marginal winter climate often face a tough decision of whether or not to treat roads for ice on a particular night. Traditionally this is achieved by using a network of road surface temperature sensors to forecast, model and monitor changes in road conditions. However, on marginal nights where the temperatures fluctuate around freezing, the presence of a few weather outstations is of limited benefit to the highway engineer. The data collected is only representative of the site under study and hence assumptions have to be made for the sections of road between sensor sites.

Figure 1: Variation in residual road surface temperature surveyed by thermal mapping for various levels of atmospheric stability.

Over the last couple of decades, interpolations between outstations have been made using thermal mapping surveys. By using a vehicle-mounted infrared thermometer, road surface temperatures (RST) are measured at a set spatial resolution across the road network. The magnitude of temperature variation across an area is dependent on atmospheric stability, but the actual pattern of RST variation remains similar on a nightly basis (Figure 1). For example, sections of road under bridges and in urban areas are always the warmest sections of road, regardless of weather conditions.

Thermal mapping techniques have remained largely unchanged over the past 20 years, but are now starting to show their age. Despite being costly and difficult to produce, thermal maps only provide a basic prediction of what the minimum temperature will be on the road network for five surveys at three predefined levels of atmospheric stability.

GIS, GPS and Ice-Miser
The University of Birmingham, UK, is developing a new generation of ice prediction techniques called Ice-Miser. The inspiration for a new system is a result of the proliferation of commercial "off the shelf" geomatics technology; in particular Geographical Information Systems (GIS) and Global Positioning Systems (GPS). The collection and display of spatial data has never been easier and a synergy of geomatic techniques has enabled massive innovation in road ice prediction.

Figure 2: Fish-eye image showing a typical urban canyon. The sky-view factor is calculated by the ratio of sky pixels to non-sky pixels.
Ice-Miser takes a completely novel approach to road ice prediction. Instead of modelling road condition at a single site and interpolating temperatures by thermal maps, Ice-Miser models surface conditions at thousands of sites around the road network (typically every 20m of road). This is achieved by considering the influence of site-specific geographical parameters on the climatology of the road. Data is collected by a single survey of the sky-view factor (a measure of the degree of sky obstruction by buildings and trees: Figure 2) along a road network. This is then combined with land use and elevation data in a GIS to produce a high-resolution geographical parameter database (Figure 3). Hence, data can be collected on a single daytime survey run—massively cheaper than the currently-specified five nocturnal thermal mapping surveys.

Figure 3: GIS screenshot showing altitude, land use and predicted road surface temperatures around a study-route in Birmingham, UK.
The geographical data is combined with meteorological data in a forecast model to predict road condition at typical spatial and temporal resolutions of 20 metres and 20 minutes respectively. The output is displayed as a colour-coded map of road temperature where it is disseminated over the Internet to the end-user (Figure 4).

Trials are constantly underway at the University of Birmingham, and early results show that Ice-Miser can consistently forecast spatial variations of RST. In a recent study, 72 percent of the variations in RST around a survey route could be explained. Also, depending on the accuracy of the meteorological forecast data, up to 95 percent of predictions were correct to within 1°C; impressive as variations of up to 13°C were present on the survey route shown in Figure 3. Away from the UK, trials are ongoing in many other countries including Japan, Slovenia, Austria and Poland.

Figure 4: Screenshot of Ice-Miser results displayed on the Internet GIS during a trial in Slovenia.

The use of GPS and GIS technologies in Ice-Miser provide a 21st Century alternative to traditional thermal mapping techniques. The benefits of this new approach include:

  • Dynamic forecast system allows users to identify exactly where the road needs treating at a continuous range of atmospheric stability.

  • Times are shown of when each section of road needs to be treated by enabling salting route optimisation strategies.

  • Cheap to survey and install—no specialist software or training is required and an operational system can be set up in a matter of days.

  • Easily integrated with winter maintenance fleet-fitted GPS units for GPS monitoring and exception reports.

  • Easily incorporated and validated with existing thermal maps and sensors to ensure best value.

  • A solid, dynamic base which will continue to improve with advances in technology.

Based at the University of Birmingham, Entice Technology provides advice on a wide range of meteorological problems and services, including thermal surveying, environmental monitoring, salting route optimization, and ice prediction system training. For more information visit, or to reach Lee Chapman or John Thornes send e-mail to