Spatial analysis of traffic accidents using gis. The case of banda aceh, indonesia

Resumen

As in most countries in the world, road crashes are a major issue in Indonesia. Currently, they cost approximately 2.8% to 3.1% of Indonesia’s GDP. In the province of Aceh, located in Northern Sumatra, the number of traffic crashes recorded on the national road system has been continuously rising since 2001. In 2014 and 2015, the number of crashes and associated fatalities in the province were higher than the national average. It has been an Indonesian government priority to reduce the number of crashes on its roads. Most methodologies developed to identify concentrations of road crashes (hot spots or hot zones) do not investigate the local factors involved. Efficiently reducing the number of accidents requires more information than simply knowing where they occur. Indeed, only by understanding what local factors contribute most strongly to road crashes will decision-makers be able to develop effective countermeasures. This thesis develops a methodology to locate road crashes and factors contributing to them. In total, two highways representing more than 1,050 km of road are analysed. First, the differences between identifying hot spots and identifying hot zones are analysed. The results show that the most dangerous hot spots identified are predominantly located inside hot zones. In addition, these hot spots and hot zones are located in urban areas and primarily on the east road. The latter is the busiest highway in the province. The most densely populated districts also appear to present the highest number of hot spots and hot zones. Next, using a Bayesian spatial approach, factors contributing to road crashes are identified and safety performance functions are developed. Amongst the variables analysed, the AADT (Annual average daily traffic) coefficient returned positively significant values in all scenarios (total crashes, fatal crashes, major injury crashes, minor injury crashes and PDO (property damage only) crashes). The coefficients for land use and horizontal alignment were negatively significant in the total crashes scenario but positively significant in the fatal crashes scenario. This indicated that the total crash rate was lower on windy and meandering roads in rural and urban-rural environments, whereas the fatal crash rate was higher in the same environmental conditions. Lastly, a series of countermeasures are suggested based on the results in the analyses. These countermeasures, tailored to the local context, consider the dangerousness and features of hot zones. For example, it was found that urban areas are characterised by high crash rates. Suggested countermeasures include implementing more frequent speed controls and traffic patrols to encourage road users to comply with traffic regulations. In rural areas, it is suggested that road safety measures should target areas identified with high crash rates. Additionally, while sections of highways in rural areas record lower daily traffic volumes, medical services tend to be less accessible. The proposed two-step methodology can be run with limited high-quality data and information and provides results authorities can use to start implementing appropriate countermeasures. This aspect of methodology is particularly important in Indonesia, as in the broader South-East Asian region, where data availability and data access remain a challenge. Overall, this study provides transportation agencies of the province with specific information on where improvements should be implemented to increase road safety conditions.