Emission Estimation Model And Spatial Analysis Of Gaseous Air Pollutants

Gaseous air pollutants have a significant impact on the environment,especially on the global climate.From a long-term perspective,the impact will be very serious.Therefore,this thesis will focus on three kinds of major gaseous air pollutants(sulfur dioxide,nitrogen oxides and carbon monoxide)emissions,and establish emission estimation models and explore their spatial characteristics respectively.The main contents are divided into the following chapters:The first chapter is an introduction,in which the research status of the estimation model of gaseous air pollutants and spatial statistical analysis is reviewed,and thus the research topic of the thesis is founded.The significance of the research is to use the statistical model to establish a high-resolution emission estimation model based on the effective information disclosed by the current Chinese government statistics department,and to use the spatial analysis methods to explore the spatial characteristics of the gaseous air pollutants emissions in a certain area.The second chapter is about the theoretical methods of the emission estimation regression model,and the evaluation criterion of the estimation regression model.In this chapter,we first elaborate the knowledge of Lasso and weighted least squares.Then we introduce two different model evaluation criteria--root-mean-square error and residual plot respectively.The third chapter is mainly about the theoretical knowledge of spatial statistical analysis,including spatial clustering and spatial autocorrelation.The part of spatial autocorrelation includes the definition of spatial weight,global spatial autocorrelation and local spatial autocorrelation.The section of global autocorrelation introduces two commonly used measurement methods: Moran’s I and Geary’s C,and illustrates how to calculate them and how to carry out a test through them.The section of local spatial autocorrelation introduces the local indicators of spatial correlation and Moran scatter charts,and details the computational methods and their relationships.In the fourth chapter,we use the modeling method introduced in the second chapter to establish a regression estimation model of the gaseous air pollutant emission based on the data collected in Handan.Firstly,the Lasso method is employed to select predictors;on the basis of results of Lasso,we establish another linear regression model;finally,the weighted least squares regression is used to obtain the final estimation model,based on a double regression model.The estimated model obtained by these three regressions has good estimation effect and statistical properties.The fifth chapter is to use the spatial statistical analysis method introduced in chapter three to analyze the data of gaseous air pollutants collected in Tangshan.The results are divided into three parts: spatial clustering,global spatial statistics and local spatial statistics.Then we further analyze these results and obtain the spatial characteristics of these three gaseous air pollutants in Tangshan.Concluding remarks are given in the sixth chapter.