Geographically Weighted Quantile Regression Model And Its Application

With the development of economy and technology,geographic information system and remote sensing technology have made great progress,and more and more spatial data have emerged.Expanding the scope and strength of current modeling,developing and applying new modeling methods to mine useful information in spatial data has very important scientific significance for the study of spatial phenomena.This thesis established a Quantile Regression(QR)model to explore the impact of atmospheric pollutants on air quality in Hohhot,Inner Mongolia Autonomous Region,and applied the Geographically Weighted Regression(GWR)model and Geographically Weighted Quantile Regression(GWQR)model to discussing the influencing factors of air quality in the Yellow River Basin.The research in this thesis is helpful for relevant departments to adapt to local conditions and take effective measures to control air pollution.The research is mainly from the following aspects:First,explore the relationship between the concentration of air pollutants and the Air Quality Index(AQI)in Hohhot,and analyze the time changes of AQI and the concentration of air pollutants.The QR model was used to analyze the relationship between the AQI of Hohhot from 2014 to 2019 and the concentration of six air pollutants in the same period.The results show that the air quality in Hohhot has obvious seasonal variation characteristics,that is,the overall performance is the worst in winter and better in summer.Quantile regression results show that at the significance level of=0.01,PM_2.5,PM₁₀,SO₂,and O₃significantly affect AQI at all quantiles,and the effects of air pollutants on different levels of AQI are significantly different.Air pollutants in Hohhot mainly come from multiple sources such as fuel combustion,and emissions from industrial production and transportation processes.Second,65 cities in the Yellow River Basin are selected as the study area,and the GWR model is used to analyze the impact on AQI from meteorological factors and socio-economic factors.The results show that the rainfall has a significant impact on the AQI of some cities in the upper reaches of the Yellow River Basin.The impact of temperature is significant in three-quarters of the study area,and both factors show negative effects,indicating that more rainfall or higher temperature will help the diffusion of pollutants and improve air quality.The proportion of the secondary industry only has a significant impact in a few cities in the middle and lower reaches of the Yellow River Basin,and some are positive and some are negative,indicating that the development of industrialization has aggravated air pollution,and the economic development brought by industrial development may in turn promote air treatment to improve air quality.Significantly affected areas of the total population are concentrated in the upper and middle reaches of the Yellow River Basin,both of which are positively affected,indicating that an increase in population will lead to poor air quality.Third,establish a GWQR model to analyze the relationship between AQI and data on rainfall,temperature,the proportion of secondary industry,the total population and the proportion of urban population in 65 cities in the Yellow River Basin.The results show that,with different quantile levels(?=0.10,0.25,0.50,0.75,0.90),the relationship between the respective variables and AQI is not only spatially,but also different in the distribution of dependent variables.Specifically,at different quantile levels of low,medium,and high(?=0.10,0.50,0.90),the regression coefficients of rainfall are all negative,and the regression coefficients of temperature are positive except for a few cities,and all other cities are negative.The influence of the proportion of the secondary industry on the AQI varies in different geographical locations.The influence of the total population on the high quantile level of AQI is spatially stable.When the impact of the urban population ratio on the AQI is at the low and high quintiles,as the geographic location changes from the southwest to the northeast,the coefficients are estimated to gradually decrease from a positive impact,and finally to a negative impact.When the median value of the quantile is taken,the estimated value of the coefficient is only negatively affected in such as Jinan,Zibo and Binzhou in Shandong Province.The content of this thesis has certain practical value for studying air quality in other regions and other environmental science related issues.