ABSTRACT
This study compares land use regression (LUR) and geographically weighted regression (GWR) models in PM2.5 concentration mapping over California (USA). Results show that R2 values of LUR model are 0.78, 0.74 and 0.74 times lower than those of GWR model at annual, seasonal and monthly scales. Relative errors are 2.18, 1.79 and 1.60 times higher, and root mean square error (RMSE) are 1.48, 1.32 and 1.28 higher. Furthermore, performance difference is significant under polluted conditions, but is minor under clean conditions. It demonstrates that LUR model is effective under low concentration at short time scales.
Acknowledgments
This work was supported in part by the National Key Research and Development Program [grant numbers 2016YFC0206205], the National Natural Science Foundation of China [grant numbers 41871317, 41501034] and the Natural Science Foundation of Hunan Province [grant numbers 2018JJ2498]. We highly appreciate the anonymous editor and reviewers for their valuable suggestions and comments.
Disclosure statement
No potential conflict of interest was reported by the authors.