The Retrieval Of Snow Depth Based On Multi-source Remote Sensing Data

As a major element in the cryosphere,snow cover is closely related to climate change,and the feedback effect of snow cover on climate is particularly significant.Owing to its high albedo,snow on the Earth may significantly affect the surface radiation budget,subsequently our weather and climate.The Intergovernmental Panel on Climate Change(IPCC)stated in its fifth assessment report:From 1901 to 2012,the global average surface temperature has accelerated.China is the country with the most widespread snow cover in the northern hemisphere,with global warming,the dynamic response of snow cover to climate change has become one of the important factors that affecting our social economy and natural ecosystems.Among the many characteristics of snow cover,Snow Depth(SD)is an important input parameter for many hydrological and climate models.Therefore,accurate snow depth information is crucial for studying the response of snow cover to climate change,hydrological cycle,and ecosystem changes etc.Traditionally,we used to measure snow depth directly,but it would be more difficult in mountainous or isolated areas.With the development of remote sensing technology,large-scale snow depth information may be easily obtained.Currently,passive microwave remote sensing is the most efficient way to monitor snow depth on global and regional scales,however its resolution is relatively coarse.Combining the characteristics of the two types of data,the fusion method of multi-source remote sensing data can take advantage of the advantages of the two sensors,thereby improving the temporal and spatial resolution and accuracy of snow depth remote sensing inversion,which is also an important input factor for watershed hydrological and climate models.Therefore,this research comprehensively uses multi-source remote sensing data,using the latest MODIS daily snow area ratio data SSEmod-FSC,AMSR 2 L1B brightness temperature data,and develops a spatial dynamic snow depth inversion based on multi-source data fusion.Algorithm(Spatial Dynamic Down-scaling,SDD),and generated a daily snow depth dataset with a spatial resolution of 500 m in China during the 2017-2018 snow season.In addition,the snow depth estimation model of lidar altimetry technology at the regional scale is already very mature.It can easily and efficiently obtain high-precision snow depth information at the regional scale.However,due to its high economic cost,it is difficult to achieve large-scale snow depth estimation.Snow depth detection.Fortunately,ICESat-2(Ice,Cloud and land Elevation Satellite,ICESat-2)can measure the surface elevation,and obtain the snow depth by monitoring the changes in the elevation before and after snowfall.This is a large range of snow depth.Monitoring provides a new method for monitoring snow depth.Therefore,this study tried to use the surface elevation data of the laser altimetry radar satellite ICESat-2(ATL08),using the satellite trajectory intersection self-elevation method to obtain snow depth information during the snowy and snowless periods,and explored the energy of the ICESat-2satellite.Whether it has the potential to provide reliable snow depth information.The results indicated the following:(1)The result of the ideal snow depth inversion is provided in China by using the spatial dynamic down-scaling snow depth inversion algorithm(SDD).In terms of spatial distribution of the depth of seasonal snow,the blocky distribution structure of snow is obviously alleviated,which more closely reflects the actual snow depth distribution.In terms of the accuracy the depth of seasonal snow,using the snow depth data and snow survey data from the weather stations in the three major snow-covered areas in China to verify that the inversion snow depth and the actual snow depth are highly consistent,and the coefficient(R~2)of determination between the measured and estimated snow depth is 0.74,and the root mean square error(RMSE)is 2.9 cm.(2)The improvement of snow depth inversion accuracy depends on the accuracy of the two types of remote sensing data.In the deep snow area,AMSR 2 brightness temperature data played a major role.In the shallower snow area that can be identified by the two types of data,the combined effect of the two types of data improves the accuracy of snow depth and avoids the phenomenon of overestimation of snow depth.In areas where AMSR 2 does not recognize snow,SSEmod-FSC data plays a major role,so the missed recognition error of snow is reduced,and the underestimation of snow depth in shallow snow areas is avoided.(3)ICESat-2 has excellent potential to provide reliable snow depth information.A survey in the Altay region of northern Xinjiang found that the snow depth obtained by using ICESat-2ATL08 surface elevation data is in good agreement with the snow depth of the weather station(R2=0.83),and the root mean square error(RMSE=4.16 cm)and mean absolute error are both low(MAE=3.37 cm).However,as the amount of data accumulates,the satellite-borne laser altimetry radar data will provide a new and effective method for obtaining large-scale snow depth information.