Authors: Debin Lu*, Zhejiang University, Wenze Yue, Zhejiang University, Wanliu Mao, Geomatics Center of Zhejiang, Dongyang Yang, School of Geographic Sciences, East China Normal University, Jianhua Xu, School of Geographic Sciences, East China Normal University
Topics: Geography and Urban Health, Geographic Information Science and Systems, Environment
Keywords: PM2.5; Land use regression; Random forest; Yangtze River Delta
Session Type: Paper
Start / End Time: 9:55 AM / 11:35 AM
Room: Senate Room, Omni, West
Presentation File: No File Uploaded
Limited by data accessibility, few exposure assessment studies of air pollutants have been conducted in China. There is an urgent need to develop models for assessing the region concentration of key air pollutants in Chinese region. In this manuscript, a land use random forest model was established to estimate PM2.5 in Yangtze River Delta in 2015 year. PM2.5 were measured at 115 sampling stations from the urban air quality monitoring data in real-time publishing platform at China Environmental Monitoring Station, 10 different predictors associated with meteorological, land use, topography, traffic, and population density were used to construct LUR and LURF models. Cross validation was used to quantify and compare model performance. The model R2 and the cross-validation (CV) R2 of the LURF model were 0.567 and 0.733, LUR model were 0.406 and 0.603, respectively. The LURF model of PM2.5 outperformed the LUR model in Yangtze River Delta. Our findings suggest that the LURF model may provide a cost-effective method of air pollution exposure assessment and machine learning methods may provide more accurate exposure assessment.