The purpose of this diploma thesis is to investigate the correlation between accident counts and basic meteorological conditions, such as rainfall and temperature. For this reason, analytical daily data for the number of accidents, fatalities and injuries were used. These data were obtained from the database of the Hellenic Statistical Authority (EL.STAT.) for the basin of Athens, the capital city of Greece. Moreover, the corresponding analytical daily data for the precipitation and mean temperature were employed. The later data were obtained from the database of the Hellenic National Meteorological Service (HNMS). For the analysis of time series, a system of SURE models was developed. The results show that the drop of temperature causes a decrease in the number of accidents, while an increase in precipitation causes a decrease in the number of accidents and fatalities as well. Furthermore, the results reveal that temperature and rainfall of the previous day (lagged values) have a significant influence on the number of accidents. Finally, the impact of the variations in temperature or rainfall on the accident counts may be different depending on which month they occur.
|Tags||accident analysis, statistical modelling, weather conditions|