Road safety and land use : different methods in detecting hazardous road locations

Road traffic collisions are recognized as a serious public health issue worldwide, with the identification of hazardous road locations being regarded as one key component in road safety improvement. This study seeks to undertake a systematic comparative analysis between existing hot spot and hot zone methodologies, focusing in particular on the spatial patterns emerging from road crashes within Hong Kong. This analysis initially explored a junction hot spot methodology. Various traffic volume characteristics are incorporated as risk exposure variables. Detection methodologies, based on equal length road segments, were subsequently designed and tested. By analysing the crash count, crash rate and the hazardous road locations identified, the associations that emerged between speed limits, land use and road safety are presented. This research focuses in particular on the relationship between speed limits and their effects on different roads, within the same study period. When reviewing land use, this study applied a systematically defined land use category, which not only takes commercial and residential lands into consideration, but all land usage types, as well as their combined effects. The empirical Bayes (EB) method was used to identify hazardous road locations, as compared to the crash count method and excess crash count method. The results show that the spatial patterns of hot spots and hot zones differ significantly, especially when taking urban or suburban variations into consideration. Hot spots appear to be predominately concentrated in main urban centers (downtown and town centers in the suburb), but seem dispersed when analyzed at the local scale. In contrast, hot zones appear to be mainly dispersed throughout the city, but seem concentrated within local regions. Their distributions are also associated with environmental factors such as speed limits and land type usage. Generally, roads with higher speed limits are associated with lower crash counts and crash rates. Road segments containing more than one speed limit appear to be more dangerous than other route sections. The speed limit impact is not as sensitive when considering crashes which occur during peak/non-peak hours or weekdays/weekends. However, it does significantly influence the severity levels of crashes which take place during various degrees of congestion and under different weather conditions, including wet roads. The land use impact on road safety is quite complex, when a systematic land use category is applied. The impact of different land uses not only relates to the collision frequency, but also varies according to the urban/suburban context, the hot spot/hot zone methodology and a combination of all these factors. Thus, more efforts are needed to evaluate the effect of land use on the identification of hazardous road locations. It is anticipated that the conclusions drawn from this thesis will contribute to gaining a more comprehensive understanding of the methodological approaches used in the identification of hazardous road locations, in addition to determining the environmental factors which impact upon road safety. The ultimate goal is to assist policymakers and engineers to ensure that roads become safer.published_or_final_versionGeographyDoctoralDoctor of Philosoph