Spatiotemporal Simulation Of Air Pollution With Synergistic Effects Of Multiple Geographic Processes

Environmentalpollution, especially air pollution has sounded the alarm for human survival, and brought new challenges for the sustainable development of environment. In order to control the air pollution effectively, specific measures should be taken according to spatiotemporal characteristic of air pollution. The spatiotemporal simulation of air pollution, provided the spatiotemporal variation of air pollution, hence it can help to prevent and control air pollution more effectively. Currently, the spatiotemporal simulation of air pollution has focused on the analysis of the spatiotemporal characteristics of a specific region and the forecasting analysis by using the existing air pollutants data. Moreover, the static factors are used for the spatiotemporal analysis with dynamic observed air pollutants data, which reduces the accuracy of the spatiotemporal simulation. Therefore, dynamic geographic processes were used as dynamic factors and their synergistic effects were considered for the air pollution simulation in this paper. The simulation method of this paper can solve the problems of incomprehensive factors and the single time scale of the spatiotemporal simulation of air pollution. This method combines the temporal analysis and the spatial analysis of air pollution and built a integration way for the spatiotemporal simulation of air pollution, thus it provided reference for spatiotemporal simulation of other geographic events.With "air pollution factor evaluation and analysis—multiple factors' geographical process simulation--air pollution spatiotemporal simulation" as the main line, method for spatiotemporal simulation of air pollution, with the synergistic effects of multiple geography processes, was constructed. As influential factors, various geographical processes were considered, and their collaborative roles for air pollution simulation were showed at the simulation method. The main work of this paper can be summarized as the following aspects:Combined with the space and time analysis, the main influencing factors of air pollution in Beijing were evaluated. The grey correlation model was used as method for time analysis of the influential factors of air pollution. The rank of the different factors' effects were produced; geographically weighted regression model was used as method for effects' spatial variation of these influential factors. The results of the time analysis show that the economics, the grass land and water land of the land use, the precipitation and the mean temperature of meteorological factors are the top rank of the influential factors. Spatial analysis showed that terrain, precipitation and forest land is negatively related to air pollution, while residential areas, temperature, and population are active related to the air pollution. Meanwhile, population, forest land, topography, residential areas and precipitation are the top rank of the influential factors. Finally, combining the time and space analysis, the final influential factors were determined, which were the natural, economic, and human aspects as three major elements, meteorological factor, topographic factor, land use, economic factor, population factor, and traffic factor as six main indicators.Simulation methods of the multiple geographic processes of the influential factors of air pollution, were studied in this paper. Among these influential factors, terrain is taken as the static factor, because it changes slower with time. For the traffic factor, the multiple periodic traffic data are obtained from Open Street Map. Therefore, meteorological factor, land use, population and economic factor are the main four geographic processes and their simulation methods were proposed in this paper. Meteorological process was simulated by the way of combining the optimal sequence analysis and optimal spatial interpolation method; land use variation process was simulated by the method of Markov cellular automata(CA-Markov); a novel spatiotemporal simulation method for population and economy were proposed based on adaptive spatialization model of population (or economy) (AP(E)SM). The population (or economy) simulation method, i.e., the AP(E)SM, showed an effective solution and solved the spatial scale constraints for the model outputs.With multiple geographical processes'synergistic effects, the air pollution spatiotemporal simulation method were proposed in this paper. This spatiotemporal model was constructed based on optimal timing simulation and spatialization model of air pollution. According to the situations of whether the monitoring air pollutants have the spatial autocorrelation characteristic, two different methods were proposed. If there is no spatial autocorrelation characteristic for the monitored air pollutants, a self-adaptive revised land use regression model (SALUR) model is used, which is an improved method and solve the shortcoming of traditional LUR model with low explanation rate at the sampling region; in view of the air pollutants has spatial autocorrelation, geographically weighted land use regression model with the region partitioned by a Voronoi graph (Voronoi-GWLUR) is used. In this way, it can represent the spatial variations of the influence of different influential factors. The SALUR model can improve the accuracy of traditional LUR model, with the RMSE error from 20.643?g/m3 to 17.443?g/m3. Accuracy of Voronoi-GWLUR reached 4.351?g/m3.Experimental analysis for the spatiotemporal simulation method of air pollution was conducted in this paper. Based on the case of PM2.5 air pollution in Beijing during year 2014 to 2016, the method was validated from three aspects:accuracy analysis, model evaluation and multiple geographic processes' synergistic effect analysis. The accuracy analysis results showed that the average error is within 0.2?g/m3, the root mean square error is less than 9?g/m3, the error rate is less than 8%, which verified the reliability of the model. The spatial distribution of accuracy showed that the accuracy is higher at the center of Beijing than the surrounding areas. The accuracy distribution characteristic is consistent with the fact that the monitoring sites are denser in the center than the surrounding areas. Thus it showed the performance of the model in this paper is accurate and reasonable. The spatiotemporal analysis results showed the spatial characteristics of the PM2.5 concentration was low in northern areas, especially the northwestern areas, and high in the southern and central regions, and the temporal characteristics of the concentrations are increase year by year. The spatial variation analysis results were consistent with the fact that the northern Beij ing is mountainous and contains few people; thus, transportation is less than in the other parts, resulting in low air pollution. Thus the results show the simulation results are reliable. The temporal analysis showed the PM2.5 concentration is decreased from 2014 to 2016. Temporal characteristics of PM2.5 pollution are consistent with the government's increasingly intensified protection measures. An analysis of the synergistic effects of different influential factors i.e., the geographical process, is providing decision supports for air pollution prevention and control at different regions during different periods.