Analysis Of Traffic Accidents Hot Spot Areas Using Spatial Statistics And Geostatistical Methods Of Geographic Information System: A Case Study Of Gibe Bridge To Sekoru Town Road Segment

Abstract

Geographic Information System (GIS) technology has been a popular tool for identifying hotspots in highways and analyzing traffic accident data. Many institutions and researchers are using GIS for accident analysis. The geometry of Gibe bridge to Sekoru town highway segment (total of 64.4 km) has many curves. 50.71% of the road segment is dominated by curves with design speed 35-40 km/hr and an average grade of 6% from the whole segment. Thus, this worst road geometric condition leads to high amount of traffic accident problems in the study area. The traffic accidents occurred along this segment need to be analyzed with GIS tools. Therefore, the main objective of this study is to analyze traffic accident hotspots along the road segment from Gibe bridge to Sekoru town using GIS tools. The study identified sites of accident hotspots and major causes of traffic accidents along the road segment depending on 5 years’ (2013–2017) of property damage only (PDO), slight injuries, serious injuries and fatal accident data obtained from Jimma Zone, Sekoru and Yem Districts Police Offices. The hot spot areas are analyzed using spatial statistics and geostatistical methods. The spatial statistical analysis includes Getis-Ord Gi*, Anselin Local Moran Index, Moran Index, and hotspot optimization. The geostatistical approaches also contains inverse distance weighting, empirical Bayesian kriging and kernel smoothing density method with supportive methods such as the geographical weighting matrix, exploratory regression and ordinary least square. As observed from the GIS spatial auto correlation analysis results for accident data in the study area, the GiPValue of Shen Debitu curves (around Abelti) and Kumbi comes P < 0.05 and P < 0.1 which is in between 90 and 95% confidence level (Gi_Bin) with GiZScore values > 1.96 and > 1.65, respectively. The GiPValue of Natri, western Saja, Simini and Birilea river (around eastern part of Sekoru town also comes P < 0.1 which is 90% confidence level (Gi_Bin) with GiZScore values > 1.65. in addition, the Kernel Smoothing Density estimation, Inverse Distance Weighting and Empirical Bayesian Kriging method (geostatistical method) analysis method also have shown that shen debitu curves (around Abelti), Kumbi, Natri, western Saja, Simini and Birilea river (around eastern part of Sekoru town) were hotspot areas. High accident zones were concentrated in Abelti (Shen Debitu), Kumbi, Natri, Saja and Sekoru cluster due to high speed, many number of curves, teenage drivers, night driving, design defects, and improper sight distance. This research recommends redesigning, reconstructing curves across the segment, limiting the maximum speed and developing road infrastructure along the Gibe bridge to Sekoru town road segment would decrease the occurrence of traffic accidents in the identified hotspot areas. It shall be also appropriate to use GIS tools in identifying traffic accident hotspot areas, if applied in the road network of Ethiopia.