Spatiotemporal dynamic interpolation simulation and prediction method of fine particulate matter based on multi-source pollution model

Abstract

Taking the air pollution monitoring data of 34 air monitoring stations in Beijing from February 8 to February 9, 2020, as an example. A spatiotemporal dynamic interpolation model of PM2.5 based on a multi-source pollution model was established. Based on the hourly spatiotemporal data of the day, the dispersion and attenuation of non-point source pollution in Beijing were interpolated. An improved hybrid genetic algorithm was used to solve the parameters of the air pollution model. The spatiotemporal Kriging model was used to predict the PM2.5 concentration diffusion on an hourly scale. The data of this area were analyzed quantitatively and qualitatively. The prediction data based on the spatiotemporal data before the current time was verified by the actual monitoring data. The results show that the model and method constructed in this paper could simulate and predict PM2.5 concentration on an hourly scale well, which could provide a good reference for the analysis, simulation, and prediction of air pollution.