The road weather information systems (RWIS) developed in the United Kingdom in the 1980s have provided highway engineers with reliable road-weather forecasts for almost 20 years. The system architecture provides the highway engineer with site-specific forecasts for normally one sensor site in each climate domain. The highway engineer then has to convert this information into a decision as to which salting routes need to be treated, at what time and with how much salt. This is a difficult decision because the engineer may have anything up to 48 salting routes (as in Devon, U.K.). In Birmingham, U.K., forecast data for one outstation is used to decide which of the 26 routes need salting. This can lead to considerable problems at times of severe winter weather.
Route them up
XRWIS is a simple new intuitive concept that provides a separate road weather forecast for each salting route and gives an initial prediction as to which routes require salting, why they require salting, at what time and with how much salt. An example is given in Figure 1d (red = salt; amber = standby and green = no action). The minimum road surface temperature is given for each salting route together with the optimum time for salting of the red routes. This information can be presented to the highway engineer for all salting routes in a region or split into divisions as required. The engineer can then click on a colored route to see a forecast graph for the coldest point on that route and forecast road temperature and condition as shown in Figures 1a-c. The front-end map can be customized as required by the user as shown in Figure 1d.
On many nights, all the routes will be red and XRWIS will give the time, reason and salting requirements for each route. On other nights all the routes will be green and no action will be necessary. On several nights a combination of colors will appear which will reflect a marginal night signaling the system will need to be monitored through the night. The thresholds for color coding can be agreed with each highway authority when the system is installed and fine-tuned over time using a self-learning neural network approach.
Ideally, route-based forecasts will be driven by a “mesoscale” weather forecast model (currently models typically operate with a 10-km grid). The relevant grid point will be chosen for each salting route or a combination of grid points. This will ensure that the best forecast is made available for each route. If a salting route does not have a road-weather outstation, it is allocated with a virtual forecast site—usually located at the coldest section of road in the route.
The secret to the success of XRWIS for route-based forecasting is the construction of a geographical database for each salting route. A sky-view factor survey (or thermal geomatic survey if it is combined with thermal mapping) is conducted to provide the proportion of visible sky at 20-m intervals along the route. The sky-view factor controls the energy balance of the road surface (1 = exposed site; 0 = sheltered site, e.g., tunnel) and measures the presence of buildings and trees and other topographic features. The database also contains other geographical and road data for each 20-m stretch (latitude, longitude, altitude, slope, aspect, road construction, thermal map residual temperature, land-use and traffic volume)
XRWIS then combines the mesoscale weather forecast with the microclimate of each salting route to predict whether or not the route needs treating. Each night is taken on it’s own merits and there is no need for fixed climate zones and restrictive thermal map types. An energy balance model is run for every 20 meters of road and predicts the road surface temperature and condition every 20 minutes. An additional feature will show a “league table” of salting routes on a given night, ranking the routes from the one that needs salting first to the one that needs salting last (Fig. 2). This thermal ranking of routes is very important and will reflect the geography of each route as well as the predicted weather conditions on each night.
In snow situations XRWIS will show the expected snow accumulation for each salting route and the required timing and amount of salt required. Alerts for the use of snow plows and snow blowers will be given as an additional layer of information.
The new thermal geomatic surveys (TGS) also will enable more efficient salting route optimization and could lead to future developments in selective salting of routes and dynamic routing. The full benefits of the system will be enhanced if salting routes are efficiently optimized and ranked from hot to cold. For example, it may be prudent to assign all bridge decks to a single salting route as these are often the coldest section of the road network. A full summary of the steps required to implement XRWIS is shown in Table 1.
The shared decision
XRWIS presents a new intuitive route-based prediction system that gives the highway engineer all the information required to make the correct salting decisions. The delivery is via the Internet with password protection and the geographical database can be built very quickly and economically to enable the microclimate of each salting route to be compiled. The use of virtual outstations and the latest mesoscale weather forecasting models will ensure more accurate salting decisions, removing the need for subjective thermal map types and climate domains thus reducing both costs and delays as well as improving safety.
The most important potential benefit of XRWIS is that it will lead to more uniform decision making across a region or country. Currently highway engineers make their own decisions as to which routes to salt based on their own RWIS installation and experience. XRWIS will potentially use standards agreed across a region such that every highway engineer in a region would potentially make the same decision for the same weather conditions. The experience of the highway engineers can be built into the algorithms that color-code the salting routes and estimate the time of salting and amount of salt to be used.
It will soon be possible, for example, to simultaneously forecast for all 3,000 salting routes in the U.K. in less than an hour. This information would then be disseminated to the relevant regional highway engineers who would get a picture of the predicted night’s weather across the whole country as well as in their own region. In the event of severe winter weather it would be possible to monitor the forecasts nationwide and share updated information across the country.
Like any new paradigm, XRWIS will continue to use the best parts of the old RWIS paradigm; indeed, all the components will still be there. XRWIS builds on existing techniques to give the highway engineer many more layers of information to help make the correct salting decisions. The intuitive front-end visualization of which routes need to be salted on a given night synthesizes the geography and microclimate of each salting route with the latest mesoscale weather forecasts. XRWIS route-based forecasting provides all the information that the highway engineer needs.
