Spatial Relationship Among Forest Landscape Pattern,Meteorology And Aerosol Based On GWR Model In Zhengzhou City

Zhengzhou,Henan Province,is an important transportation hub city in China.The ecological environment and air pollution issue of Zhengzhou have always been the focus of social attention.Forest landscape pattern is an important regulatory factor of atmospheric aerosol pollution.This study will analyze its spatial dependence on aerosol optical depth(AOD)from the perspective of landscape ecology.The AOD of Zhengzhou was selected as the research object,and GF-2 remote sensing image,CLDAS V2.0 meteorological data set and MODIS04₃K aerosol remote sensing data were used as the data sources.Firstly,based on MODIS04₃K remote sensing data after clipping and resampling,the spatial distribution characteristics of quarterly and annual average AOD of Zhengzhou city are quantitatively analyzed,and the spatial autocorrelation of AOD data is studied.Secondly,after the interpretation of GF-2 remote sensing data and the extraction of forest landscape pattern index,the spatial and temporal distribution characteristics of forest landscape pattern index,CLDAS meteorological factor and their spatial autocorrelation were analyzed.Finally,based on the above research data,a global OLS(Ordinary Least Square)model and a GWR(geographical weighted regression)model were constructed with the quarterly and annual average AOD as the dependent variables,and forest landscape pattern index and meteorological elements as explanatory factors,respectively,to analyze the spatial dependence between aerosol and independent variables.The specific research contents and main conclusions are as follows:(1)The quarterly average AOD value of the study area was from strong to weak,which was in the order of summer AOD(0.842)>autumn AOD(0.652)>spring AOD(0.567)>winter AOD(0.407),and the average annual AOD value was 0.617.In 2017,the distribution of quarterly and annual average AOD value showed the characteristics of low in the south,high in the north,low in the West and high in the East.The lowest value area was mainly distributed in mountainous and hilly areas,and the highest value area was mainly distributed in plain and urban settlements.In addition,Moran's I index analysis found that both quarterly and annual average AOD showed significant spatial autocorrelation.(2)In OLS regression,all indexes significantly correlated with quarterly or annual average AOD and passed the collinearity test were TA(Total Landscape Area),NP(number of patches),PD(patch density),LPI(Largest Patch Index),LSI(Landscape Shape Index),COHESION(Patch Cohesion Index),quarterly and annual average ground pressure and 2 m air specific humidity,2 m air temperature and 10 m wind speed except summer.(3)Among meteorological factors,the pressure value is consistent with the trend of geographical elevation,the air specific humidity is smaller in the central and western regions,the temperature is higher in the north of Zhengzhou and the mountain areas have lower temperature,and the low wind speed area appears in the northeast.And among landscape elements,the areas have high value of TA,NP,PD,LSI,COHESION,LPI are gathered in the southwest of Zhengzhou,but the areas with high LPI are slightly scattered than others.In addition,the spatial autocorrelation test shows that the global Moran's I index of all forest landscape pattern factors and meteorological factors is significant,and the spatial autocorrelation of meteorological factors is greater than that of forest landscape pattern factors.Local Moran's I of forestland landscape elements appeared high-high aggregation in the western hilly area,and low-low aggregation in the eastern part to the north.(4)The fitting effect of GWR model on AOD,forestland landscape pattern and meteorology is much better than OLS model.In the GWR model,the fitting effect of the annual average AOD model was higher than that of the quarterly average AOD model,and the fitting effect of the quarterly average AOD model was the best in spring and the worst in winter.(5)All the forestland landscape pattern indexes and meteorological indexes tested by OLS significance were taken into GWR model analysis,and it was found that the influence of all variables on AOD had spatial non-stationarity.In addition,the geographical range of meteorological factors significantly related to AOD is greater than that of forestland landscape pattern.In terms of forestland landscape pattern index,the indexes with the widest influence on the quarterly and annual average AOD are LPI,LPI,NP,TA and COHESION.And in terms of meteorological index,the indexes with the most significant influence on the four seasons and annual average AOD are in the order of surface pressure,surface pressure,10 m wind speed,2 m air temperature and 2 m specific humidity.Based on the above research,the spatial autocorrelation of aerosol and landscape pattern should be considered in the research model construction of aerosol influencing factors,so as to estimate local parameters more accurately.On the basis of the above research,the forest landscape planning of aerosol pollution is considered according to the local situation in different seasons,so as to achieve the accurate mitigation of aerosol pollution in the largest area,and provide reference for the planning of local forest landscape pattern guided by atmospheric aerosol mitigation.