The objective of this paper is the correlation of the number of road accidents and of basic meteorological conditions, such as rainfall and temperature. For this reason, detailed daily data of road accidents and people killed and injured in road accidents from the EL.STAT. database for the basin of Athens were used, as well as the corresponding analytical data for daily precipitation and average daily temperature from the Hellenic National Meteorological Service database. For the analysis of time series, a system of SURE models was developed (Seemingly Unrelated Regression Equations) with simultaneous correction of the error for serial autocorrelation of first grade. The results show that a decrease in temperature causes a decrease in the number of road accidents, while an increase in rainfall causes a decrease in the number of accidents and in people killed in these accidents. Moreover, the results reveal that temperature and average rainfall of the previous day (lagged variables) have a significant influence on the number of accidents, while the phenomenon of the influence of weather on road accidents presents strong evidence of non-linearity.
|Tags||accident analysis, weather|