Geostatistical Spatial Decision-Making for Identifying Road Hazardous Road Segments in Rural Areas

Abstract

Road safety has now become a major concern in many countries. In Algeria, as elsewhere, many factors are involved in the phenomena of road accidents. In this field of research, our study aims to spatial identification along a road network the hazardous road segments which present a dangerousness index. We proceeded the use of geostatistical approach, in particular global G(d) statistic and local Gi*(d) statistic of Getis-Ord. However, in an operational framework, this approach is highly dependent on hypothesis on the process of the measurement of the spatial proximity based on the distances between the BSUs. Thus, the interpretation of empirical results based on the optimal choice of the weighting structure and the critical distance. Methodologically, this work allowed to test an approach for assess the safety levels experimented in the case of road network specifically higher accidents risk in the Wilaya of Mascara (north west of Algeria). The first results confirm the validity of the method and allow a recoverable mapping by managers in the choice of measures of preventive management.