| Abstract: | Many studies have attempted to model the sophisticated influence of traffic emissions on air pollution, but most models only calculate the contribution of traffic emissions near monitoring sites. It is difficult to observe the near-surface dynamics such as wind, rain, and human activities and precisely distinguish traffic emissions. These obstacles make model simulation very expensive in practice. The regional distribution patterns that can help adjust policies and actions taken remain unknown. Therefore, this article proposes a grid-oriented geostatistics-based approach to overcome these obstacles. We chose central Beijing as the study area. An experiment implemented the approach on data collected from Global Positioning System navigation software, car rental companies, and meteorology stations. The results suggest that the northwest area of Beijing has high traffic-related air pollution (TRAP) and the southeast area has low TRAP. Unlike modeling-based methods, this work uses geostatistical methods to directly study the spatiotemporal connections between traffic and PM2.5 (particulate matter with diameter less than 2.5?μm) from the phenomenon. The calculation is conducted under no hypotheses and has little risk of producing results contradictory to facts. This work provides a reference for future TRAP research on directly learning from the phenomenon and assists decision makers with seamless spatiotemporal heat maps of TRAP distribution. Key Words: Beijing, geospatial statistics, PM2.5, spatial correlation, traffic-related air pollution. |
