Analysing traffic crashes in Riyadh City using statistical models and geographic information systems

Road safety is a serious societal concern in Riyadh city, Kingdom of Saudi Arabia. Because of the negative impact of traffic crashes which cause losses in the form of deaths, injuries and property damage, in addition to the pain and social tragedy affecting families of the victims, it is important for transport policy makers to reduce their impact and increase safety standards by reducing the severity and frequency of crashes in the city of Riyadh. It is therefore important to fully understand the relationship between traffic crash severity and frequency and their contributing factors so to establish effective safety policies which can be implemented to enhance road safety in Riyadh city. Data used in previous research have only consisted of basic information as there was unavailability of suitable and accurate data in Riyadh and there are very few studies that have undertaken as small area-wide crash analysis in Riyadh using appropriate statistical models. Therefore safety policies are not based on rigorous analyses to identify factors affecting both the severity and the frequency of traffic crashes. This research aims to explore the relationship between traffic crash severity and frequency and their contributing factors by using statistical models and a GIS approach. The analysis is based on the data obtained over a period of five years, namely AH 1425, 1426, 1427, 1428, and 1429 (roughly equivalent to 2004, 2005, 2006, 2007, and 2008). Injury crash severity data were classified into three categories: fatal, serious injury and slight injury. A series of statistical models were employed to investigate the factors that affect both crash severity (i.e. ordered logit and mixed logit models) and area-wide crash frequency (i.e. classical Poisson and negative binomial models). Because of a severe underreporting problem on the slight injury crashes, binary and mixed binary logistic regression models were also estimated for two categories of severity: fatal and serious crashes. The mixed binary logit model and the negative binomial model are found to be the best models for crash severity and crash frequency analyses respectively. The model estimation results suggest that the statistically significant factors in crash severity are the age and nationality of the driver who is at fault, the time period from 16.00 to 19.59, excessive speed, road surface and lighting conditions, number of vehicles involved and number of casualties. Older drivers are associated with a higher probability of having a fatal crash, and, as expected, excessive speeds were consistently associated with fatal crashes in all models. In the area-level crash frequency models, population, percentage of illiterate people, income per capita and income per adult were found to be positively associated with the frequency of both fatal and serious injury crashes whereas all types of land use such as percentages of residential use, transport utilities, and educational use in all models were found to be negatively associated with the frequency of occurrence of crashes. Results suggest that safety strategies aimed at reducing the severity and frequency of traffic crashes in Riyadh city should take into account the structure of the resident population and greater emphasis should be put on native residents and older age groups. Tougher enforcement should be introduced to tackle the issue of excessive speed. This thesis contributes to knowledge in terms of examining and identifying a range of factors affecting traffic crash severity and frequency in Riyadh city.