Retrieval Of Aerosol Optical Depth Using Multi-source Remote Sensing Data In Arid Area

Aerosol Optical Depth?AOD?was a variable that can comprehensively assess aerosol loading in the atmosphere and was critical for regional air quality assessment.MODIS aerosol product data was widely used in the study of aerosol properties worldwide due to its long time series,short revisit cycle and reliable data accuracy.However,there was a lack of spatial resolution and lack of verification of data accuracy in arid regions.Therefore,this study used medium resolution data?GF-1,HJ and Landsat series?to obtain high spatial resolution?500 m?AOD data,which can provide auxiliary reference data for ecological construction and oasis regulation in arid areas.This study selected Jinghe County in the dust source area and Wusu City in the downstream as a research area,using common remote sensing data?GF-1,HJ and Landsat?combined with Deep Blue algorithm?DB?and Dark Target algorithm?DT?,analyzes the annual and seasonal spatial distribution characteristics of AOD in the study area.The random forest algorithm was used to analyze the influencing factors of AOD.The research results are as follows:?1?This study found that continental aerosols are more suitable for complex aerosol changes in the study area.Based on DT?R²>0.74?and DB?R²>0.68?,the spatial distribution of AOD in the study area can be well reflected.The DT algorithm presents a slight underestimation,and the DB algorithm was the opposite.Compared with DT,the DB algorithm has a significantly reduced AOD data null value area,which was more suitable for the study area.?2?Estimated AOD data?such as Landsat AOD?can better reflect the spatial distribution and changes of aerosols in key areas,but there are serious data missing conditions due to revisit cycle and data quality.Therefore,the mainstream data of the aerosol research MODIS AOD in the arid area,supplemented by the high spatial resolution AOD data obtained from the Landsat/HJ/GF-1 data,can improve the intra-analytical accuracy of the key areas.?3?The seasonal distribution characteristics of AOD in the study area showed the pattern of AOD_spring>AOD_summer>AOD_autumn.The AOD?0.3521?located in Jinghe County of the dust source area was significantly higher than that of Wusu City?0.2173?.Random forest results showed that temperature,precipitation and potential evapotranspiration were the main influencing factors,and their importance scores were0.263,0.209 and 0.290.