Integrated study for the impacts of air pollution and climate conditions on road traffic accidents

Abstract

Abstract Traffic accidents are among the most frequent incidents in society, directly posing threats to human life and property loss. In 2022, China recorded 256,409 traffic accidents, resulting in direct property losses of 1.2393 billion yuan. Analyzing potential factors affecting road traffic safety is crucial to reduce accidents and subsequent losses. This article focuses on traffic accidents, climate, and air pollution in Chengdu, a megacity in southwestern China. Linear and nonlinear models are used to analyze the impacts of air pollution and climate conditions on road traffic accidents. The linear model examines the linear relationships between the response variable (number of traffic accidents) and various indices initially. After verifying nonlinear dependencies using a novel statistical test, XGBoost and generalized additive models are further utilized to capture the nonlinear relationship between response and covariate. The results indicate that climate indices, such as temperature and precipitation, have varying impacts within different ranges of covariate. Integrating linear and nonlinear models can enhance the interpretation ability of statistical analysis.

Keywords

• Road traffic accidents
• Air pollution
• Climate conditions
• Linear and nonlinear analysis
• XGBoost