Using The Gaofen-4 Geostationary Satellite To Retrieve Aerosols Over Land With High Spatiotemporal Resolution

GaoFen-4(GF4)is China's first Optical remote sensing geostationary satellite which could observe the Earth's surface every 20 seconds.The GF4 is equipped with a 50 m optical gaze camera which has red,green,blue and other wavelengths that are sensitive to Aerosol Optical Depth(AOD).In this paper,according to the characteristics of GF4's observation and data bands,an aerosol retrieval algorithm for high spatio-temporal resolution was constructed.The main research content and conclusions are as follows:In the high spatial-temporal resolution aerosol retrieval process of GF4,considering the improvement of the SNR of radiation data requires joint retrieval of pixels within a certain size window.In this process pixels that are not easily recognized such as broken clouds,cloud edges and thin clouds will introduce errors and reduce the accuracy of retrieval.Therefore,accurate cloud mask data is a key in aerosol retrieval.The high spatial variability in the process of cloud's formation,especially on satellite images with high spatial resolution,clouds show a higher spatial variability than other land cover types such as vegetation,soil,water,and aerosols.Based on the characteristics of cloud higher spatial variability,this paper joined the clouds 'spatial variability,spectral reflectivity of clouds and the HOT thin clouds detection method,focused on the identification of broken clouds,cloud edges and thin clouds which have great interference with aerosol retrieval,which achieves a good result.Accurate surface reflectance estimation in the process of ground-atmosphere is a key to aerosol retrieval.The GF4 has no fixed shooting mode and cannot regularly cover the same area.This results in difficulties in correctly constructing time-series imagery for surface reflectance estimation.By using the initial surface reflectance library synthesized as the iterative initial value of the time-series method,the time-series imagery can be correctly constructed for surface reflectance estimation.Then,following the assumption: the AOD varies quickly with time,but slowly with the location,and the surface reflectance varies quickly with location but slowly with time.The AOD and surface reflectance that minimize the changes in the surface reflectivity of the two images are the true AOD and surface reflectance to be retrieved.Eventually,produces AOD products with a time resolution of 1 minute and a spatial resolution of 800 m.In order to verify the accuracy of the aerosol products obtained by the above method,the data from the North China area from June to October 2016 were used for retrieval tests.The derived results agreed reasonably well with the moderate resolution imaging spectroradiometer collection 6.0 aerosol product,with a slop of regression equation,1.09,and a correlation coefficient of R =0.893.The derived results also agreed reasonably well with the measured data form AERONER and SONET,which have high accuracy,with a slop of regression equation,0.797,and a correlation coefficient of R = 0.794.The atmospheric distribution and changes in some parts of the North China Plain were analyzed using the aerosol products of GF-4.The results showed that it is feasible to use the time-series imaging method to conduct high spatiotemporal resolution aerosol inversion by using GF-4's data,which can be used for detecting changes in air pollution.