By Joseph Steenbergen
Summary
Meteo Consult is involved in a project called “Prevention of slippery roads based on local meteorological, thermal mapping and GPS data”. The aim of this project is to make road slipperiness prevention more effective. Meteo Consult contributes to this goal by providing (and improving the method for) a road service temperature (RST) forecast.
One of the parameters of the RST forecasting model is the Sky View Factor (SVF). This parameter is a measure of the degree to which the sky is obscured by the surroundings (trees, buildings etc.) for a given point.
So far, the SVF was obtained by using a car equipped with a camera and a fisheye lens, making photographs pointed at the zenith. These photographs were than analyzed and the SVF computed. This is a very time consuming and therefore expensive method. This caused Meteo Consult to search for alternative methods.
Inspired by the developments in the GIS sector, the decision was made to explore the possibilities of computing the SVF using a GIS based on geodata. Preliminary research pointed out that a GIS and geodata can indeed be used for the SVF computation, although several hurdles still had to be taken.
One of the major problems was the correct modeling of trees. In winter, deciduous trees generally do not carry leaves. Using the GIS method, the landscape is modeled using a raster dataset. This limits the amount of information that can be held in the dataset. Complex 3D structures such as trees cannot be modeled. Only 1 parameter (in this case terrain height) can be modeled in this raster.
A solution for this problem was found. The 2 major types of objects blocking the sky view are ‘hard objects’ such as buildings, hills, mountains etcetera and ‘soft objects’ like for instance trees. By computing the SVF in 2 ways, once based on only ‘hard objects’ and once based on both the hard objects and the ‘soft objects’, it is possible to determine what fraction of obscured sky is purely obstructed by soft objects. By applying a correction factor to this fraction, compensation for the fraction of sky which is visible through a bare tree is possible.
The application of this method yielded results which in many cases approached the reference SVF closely, but in other cases under or overestimated the reference SVF. The main cause for this fact is the lack of detailed and comprehensive information with regard to the location, shape and dimensions of objects blocking the sky view.
The conclusion of this research was that GIS could very well replace the photographical method. However, the quality of the results depends mainly on the quality of the input data. Given the current developments in GIS and geodata, one can expect these problems to be solved in the future.
Taking into account the developments in 3D GIS, one can expect the current use of rasters to be replaced with 3D vector datasets. This will increase the accuracy with which the geometry of the environment can be modeled, which in turn will benefit to the quality of the GIS SVF.