Abstract:

Rainfall leads to significant impacts on urban road traffic, such as traffic congestion and traffic accidents, resulting in detrimental effects on traffic efficiency. Exploring the changes in road traffic efficiency and their influencing factors in rainy weather is of great significance for urban traffic planning, management, and traffic information release. Due to the complexity of urban road network and diversity of transportation demands, rainfall, on one hand, causes flooding on the road and reduces visibility, both of which reduce the driving speed; on the other hand, rainfall also reduces unnecessary trips, resulting in reduced traffic flow, which increases road speed. However, most previous studies focused on changes in the average traffic efficiency of all roads and ignored the differences among different road types. Moreover, there is a lack of related literature analyzing the influencing factors of road speed changes on rainy days. Therefore, this study uses the rainfall monitoring data and large-scale taxi GPS data of Shenzhen as the data source to analyze the differences in speed change on large-scale urban roads and their influencing factors. First, the percentage difference in road speed between sunny days and rainy days was calculated. It was found that rainfall did not always reduce road speeds: approximately 30% of the road speeds in the study area increased. Then, a multivariate linear regression model was applied to analyze the different effects of certain influencing factors, such as rainfall intensity, traffic volume, road grade, and the speed on sunny days, on two types of roads (i.e., speed decrease and speed increase). The results show that for the roads with reduced speed, the rainfall intensity, traffic volume, speed on sunny days, and coefficient of vehicle speed variation on rainy days have significant negative correlations with the speed decline rate. For example, the heavier the rainfall is, the faster the road speed drops. The number of bus stops has a positive correlation with the speed decline rate. For roads with increased speed, the rainfall intensity and coefficient of vehicle speed variation on rainy days have significant positive correlations with the speed increase rate, while the speed on sunny days is negatively related to the speed increase rate. For the two types of roads, the road grade is not a significant influencing factor. This research expands our understanding of the impact of rainfall on urban road traffic efficiency and can provide useful insights for the operation and evaluation of various road facilities in urban areas.

Key words: rainfall weather, urban traffic, road speed, multiple linear regression, Shenzhen city