Statistical distribution of speeds on German motorways.

Figures on the overall mileage under defined conditions are more or less available for the motorway networks of most countries. However, the proportion of mileage performed in different speed classes is usually not known; this is regrettable because speed has a very significant impact on traffic safety and on air pollution. Therefore, the objective of the present paper is to establish the speed distribution of all passenger cars for a given year (1987) on the entire network of West-German motorways. More precisely, speed distribution here is understood to be the proportion of mileage in the speed classes from 0 to 60 km/h, from 60 to 70 km/h and so on in classes of 10 km/h width up to the class from 170 to 180 km/h as well as an overflow class above 180 km/h. The method consisted of three steps. In the first step the traffic volumes were determined for each direction and each section of the German motorway network as well as for each hour of the year. To this end, data of traffic counts were used. In the second step various speed-flow relationships were developed from specific microscopic measurements. These relationships were linked to the volumes in order to yield mean speeds for each part of the network and each hour of the year. In the third step, the speeds were distributed about their mean values and finally the proportions in each speed class were aggregated in order to obtain values that are valid for the overall network. The different steps were then combined to yield the overall speed distribution. The results show that 15 % of the passenger car mileage in 1987 was achieved while the speeds were above 145 km/h and that 31 % were above 130 km/m. The mean speed turned out to be 117,2 km/h. Since these figures refer to a complete year, the seasonal variations that are inherent in the speed distribution were additionally investigated. Detailed data were available from five permanent measurement stations. A statistical model was developed from which some temporal components of variation could be extracted. (A)