Estimating And Analysis Of Ground-Level NO₂ Concentrations In The Yangtze River Economic Belt Based On Geographically And Temporally Weighted Regression Model

NO₂ is one of the common atmospheric pollutants,and long-term exposure to high NO₂ can cause adverse effects on human health.In recent years,with the rapid development of China’s economy and further increase in energy consumption caused by industrial production and transportation,China has become one of the most serious NO₂ polluted regions in the world.As one of the three national-level development strategies in China,the Yangtze River Economic Belt,with 21.4%of the country’s land area and more than 45%of the country’s gross domestic product,is a pioneering demonstration belt for the construction of ecological civilization in China.The twenty tenth report proposes to promote green development and harmonious coexistence between human and nature,and it is of great practical significance to study the distribution pattern of ground-level NO₂ concentration in the Yangtze River Economic Belt to promote the green and efficient development of the Yangtze River Economic Belt.Although the traditional ground-based monitoring of ground-level NO₂concentration can accurately monitor the ground-level NO₂ pollution condition around the station,it is difficult to achieve continuous monitoring on a large spatial scale due to the limited distribution of monitoring stations;satellite can realize remote sensing monitoring of tropospheric NO₂ on a large spatial scale,but it cannot provide the ground-level NO₂ concentration directly related to human health.Therefore,fusing satellite remote sensing monitoring surface data with ground-leveal monitoring point data to achieve ground-leveal NO₂ concentration estimation at large spatial scale is important for studying the spatial pattern of NO₂ pollution in the Yangtze River Economic Belt.In this paper,based on ground-level NO₂ concentration data,OMI-L3 level vertical tropospheric column concentrations,and ERA5 meteorological data from December 2019-November 2020,we construct Geographically and temporally weighted regression（GTWR）for ground-level NO₂ concentrations in the Yangtze River Economic Belt at each spatial and temporal scale by season and region,and compare them with Ordinary linear regression models（OLS）,Geographically weighted regression models（GWR）,and Time weighted regression models（TWR）.The results were compared with those of OLS,GWR and TWR models.And based on the spatio-temporal weights calculated by the GTWR model,the spatio-temporal distances between the monitoring points and the estimation points were calculated,and the ground-level NO₂ concentrations at each unknown point were estimated based on the coefficients of the model explanatory variables of the spatio-temporal closest monitoring points,and the IDW interpolation was used to realize the fast mapping of the cryptographic estimation results of ground-level NO₂ concentrations.The estimation results show that compared with OLS,GWR and TWR models,the GTWR model has the highest R² and lowest RMSE,MAE and MAPE at all scales,and the GTWR model has better estimation performance;the seasonal estimation can effectively reduce the interference of temporal heterogeneity and improve the estimation performance of the model,but due to the influence of atmospheric stability and solar radiation,the model shows higher estimation performance in winter and autumn than in spring and summer.However,due to the influence of atmospheric stability and solar radiation,the model shows a higher estimation performance in winter and autumn than in spring and summer;sub-regional estimation can effectively reduce the interference of spatial heterogeneity to the model and improve the estimation performance of the model.On the whole,there is a high degree of similarity in the spatial and temporal distribution of NO₂ concentrations near the ground in the Yangtze River Economic Belt and the vertical column concentrations in the troposphere.In terms of temporal distribution,both show a seasonal trend of highest concentration in winter,followed by spring and autumn,and lowest concentration in summer;in terms of spatial distribution,both show a spatial pattern of high concentrations in Yangtze River Delta,eastern Hubei and Chengdu-Chongqing regions,and gradually decreasing in surrounding cities,but in some low and middle value regions,tropospheric NO₂ concentrations usually show low value aggregation due to limited spatial resolution,while the estimated ground-level NO₂ concentrations can show more details of spatial distribution.can show more details of spatial distribution.Locally,the ground-leveal NO₂ concentration in the upstream area of the Yangtze River Economic Belt shows a distribution trend of high in the east and low in the west,and the overall pattern of spatial distribution is that the high values are centered in the Chengdu-Chongqing area and Panzhihua City and gradually decrease in the surrounding areas;the Ground-leveal NO₂ concentration in the midstream area shows a spatial distribution pattern that the high values are centered in the capital cities of each province and gradually decrease in the surrounding areas;the ground-leveal NO₂concentration in the downstream area shows a spatial distribution pattern that the high values are centered in the Yangtze River Delta city cluster and gradually decrease in the surrounding areas.