Investigating the effect of estimating urban air pollution considering transportation infrastructure layouts

Abstract

Transportation infrastructure layouts have significant impacts on urban air quality. This study develops a street-scale estimation model involving transportation infrastructure layouts based on machine learning to tackle the practical challenge in many cities of lacking dense air quality monitoring networks, and then investigates the effects of such layouts on both air pollution modeling and actual pollution levels. The results demonstrate that introducing transportation infrastructure layout variables into conventional models can significantly improve their general accuracy and stability, and particularly address their typical shortcomings in peak time predictions. Such variables together contribute an average of 28.63 % of model results, indicating their significance and non-negligible roles. To mitigate urban air pollution, constructing transportation infrastructure that encourages green travel over private vehicles is generally effective, yet the increased buses, subways, and charging stations in certain areas may lead to adverse effects. The improved petrol station layout and parking infrastructure utilization may also work.