ALBERT

Description: 〈p〉Publication date: Available online 4 October 2019〈/p〉 〈p〉〈b〉Source:〈/b〉 Transportation Research Part D: Transport and Environment〈/p〉 〈p〉Author(s): Suhong Zhou, Rongping Lin〈/p〉 〈div xml:lang="en"〉 〈h5〉Abstract〈/h5〉 〈div〉〈p〉Research on the relationship between built environment and PM2.5 has attracted notable attention during the past decades. However, previous studies were less to test the spatial-temporal heterogeneity of on-road PM2.5 and its related factors at micro scale. To this end, collecting high-resolution PM2.5 data by mobile monitoring along different roads in Guangzhou, China, this paper explored the spatial-temporal heterogeneity of the relationship between built environment and on-road PM2.5 during the morning (7–9 am) and evening (7–9 pm) rush hours. Semi-variogram method and geographically weighted regression (GWR) model were utilized to reveal the non-stationarity associations among the large spatial dataset. In terms of temporal heterogeneity, the results showed that the spatial independent radii of on-road PM2.5 were 17 m and 21 m for morning and evening rush hours respectively. The aggregated median value of PM2.5 in the morning rush hours was 34.95 μg/m〈sup〉3〈/sup〉, while the evening was up to 55.49 μg/m〈sup〉3〈/sup〉. There were more significant factors of street conditions impact on on-road PM2.5 in the morning while more significant factors of land use and centrality that reflecting the cumulative effect of daily human activities with smaller buffer thresholds in the evening. In terms of spatial heterogeneity, GWR models achieved much better performance than the global ones of multivariate regression models with lower AICc, RMSE and higher adjusted 〈em〉R〈sup〉2〈/sup〉〈/em〉, explaining 10–69% of variance across different roads and rush hours. There was a high degree of spatial heterogeneity that the leading factors were different along various roads on rush hours. The results indicated that the policies and interventions should be more targeted to improve the on-road air environment and reduce personal exposure according to the spatial-temporal geographical context. It can be adopted to provide more realistic and practical guides for urban planning and environmental pollution control.〈/p〉〈/div〉 〈/div〉