| Surface albedo is a physical quantity that characterizes the surface's ability to reflect solar radiation.It is defined as the ratio of the surface reflected solar radiation flux to the incident solar radiation flux.High resolution surface albedo remote sensing products are becoming important data sources for regional energy balance and climate change research due to their high spatial resolution.At present,surface albedo retrieval method has three ways:narrow band to broadband transformation;based on remote sensing observation data for phase of BRDF modeling;direct inversion algorithm,these algorithms are difficult to both high time and spatial resolution,and the surface albedo high resolution remote sensing retrieval algorithm and data products are assuming that the surface smooth and flat,The retrieval algorithm applied to flat surface but lack of considering of the surface heterogeneity and complexity of terrain,will be some error.The mountain radiation transfer model considers the real solar radiation transfer process in the complex terrain area,so it can accurately calculate the upstream and downstream radiation of the complex mountain surface,and then accurately simulate the albedo of the mountain surface.In this paper,the direct inversion algorithm of surface albedo is combined with the mountainous radiative transport model,which provides a new idea for high-resolution surface albedo inversion in mountainous areas,and retrieval of high resolution surface albedo products in mountainous areas.However,due to the influence of snow and cloud pollution in underlying area,the time-domain image inversion is discontinuous,and there are many missing values.Therefore,it is impossible to construct spatiotemporal continuous surface albedo products to support the research on surface energy balance in mountainous areas.In order to solve this problem,based on the GF-4 data,use improved direct retrieval algorithm,combined with MODIS BRDF/Albedo product build prior knowledge background fields,then use the ensemble kalman filter method to retrieval mountain surface Albedo,then fill in time and space build the time-space continuum surface Albedo retrieval method,and produced 2016-2017 mountain surface Albedo product.The main innovations of this paper include:1:The topographic effect of mountain surface is considered;2:the ensemble Kalman filter algorithm is used to calculate the albedo products with high spatial and temporal resolution through few high spatial images.The main results are as follows:(1)The improved Angular Bin(AB)algorithm combined with mountain radiation transfer model performs well in surface albedo inversion,which can accurately reflect the characteristics of mountain area and the texture characteristics of rugged terrain in mountain area,is suitable for albedo inversion in mountain area.(2)Although MCD43A3 has a low spatial resolution,it has a high temporal resolution and can provide a continuous trend of surface albedo in the study area under a long time series.we can produce the background field of Kalman filter algorithm based on MCD43A3data set,the results show that the processed MODIS albedo product is greatly optimized in the details of spatial scale,and it is more suitable to be used as background field in ensemble Kalman filtering algorithm.In Zhang Ye and ARou stations,the background field data before and after processing were quantitatively compared with the high-resolution surface albedo.After processing,the RMSE of the background field data at the two stations decreased by0.0139 and 0.0053,respectively.R~2increased by 0.3788 and 0.1439,respectively.(3)The inversion results are in good agreement with the ground measurements,RMSE is less than 0.01 on the flat surface,and RMSE is 0.0163 on the site with higher slope.(4)As a domestic geostationary satellite,GF-4 has a good application potential in surface albedo inversion in rugged terrain in mountain area,and provides a good support for the study of surface albedo with high spatial and temporal resolution in China. |
PDF Full Text Request