Research On Operational Algorithm Of Dynamic Snow Depth Inversion Based On Passive Microwave Remote Sensing Data In Northeast China

Snow cover plays an important role in the global hydrological cycle and climate system,it is the main recharge source of rivers and groundwater.The high reflectivity and the adiabatic effect significantly affect the global energy interaction.Snow depth is an important parameter in climate and hydrological model,therefore,accurate retrieval of snow depth is very important for understanding the climate system and water resources management in cold regions.At present,the most commonly used snow depth inversion method is operational algorithm,however,the existing operational algorithms assume that the snow characteristics are fixed throughout the winter,and are affected by vegetation,especially forests,which often lead to large errors in the inversion results.Based on the "Snow Characteristics and Distribution Survey in China" project,the field observation experiments on snow characteristics were carried out in Northeast China during 2017-2018.According to the snow characteristics data,the applicability of the existing operational algorithms were verified in Northeast China,and the main factors affecting the inversion of snow depth were analyzed.On this basis,a dynamic snow depth inversion algorithm considering the evolution of snow characteristics and forest attenuation is proposed,and verified in a long time series.In addition,the "Snow Depth Dataset Production Platform" software was established for the operational production of daily snow depth datasets in Northeast China.The specific research work is summarized as follows:(1).Verification and Analysis of Snow Depth Inversion Algorithm in Northeast China Based on Snow Observation Data.Firstly,based on "Snow Characteristics and Distribution Survey in China" project,the field observation experiments were carried out in December 2017-March 2018.It was found that the snow cover characteristics in Northeast China were spatiotemporal heterogeneity.In addition,the observed snow depth in the filed experiments were taken as the actual snow depth,and the applicability of Chang algorithm,Che's products,and AMSR2 algorithm in Northeast China was evaluated,and their precision was analyzed based on the underlying surface types and snow falling time.The results shown that: Che product had the highest inversion accuracy,and its RMSE and Bias were 8.24 cm and-2.27 cm.For Chang algorithm,its RMSE,Bias and R were 17.08 cm,7.97 cm and 0.19 respectively,and the AMSR 2's RMSE,Bias and R were 0.42 were 22.01 cm,15.44 cm and 0.04,respectively.(2)Research and Validation of Dynamic Snow Depth Retrieval Algorithm Based on Passive Microwave Remote Sensing Data in Northeast China.The chapter discusses the influence of the change of snow characteristics on the different of brightness temperature between 18 GHz and 36 GHz based on MEMLS model.It could find that the change of snow density and snow particle size would cause a greater impact on the brightness temperature.The traditional empirical algorithm(based on brightness temperature difference)assumed that snow density and snow particle size was fixed,which would lead to large errors in actual snow depth inversion,Based on the above research,this study used the snow characteristics data observed in the field as the input parameters of MEMLS model,the dynamic coefficient that could reflect the change of snow characteristics was determined,then a dynamic snow depth inversion algorithm considering the evolution of snow characteristics was constructed.Considering the influence of forest on brightness temperature,the algorithm was optimized by introducing forest attenuation factor in forest areas.Finally,a dynamic inversion algorithm considering snow evolution and forest attenuation was obtained.The novel algorithm was verified based on meteorological stations in Northeast China.The results show that the inversion results are in good agreement with the measured snow depth,with R of 0.61,the Bias and RMSE of 1.07 cm and 7.79 cm,respectively.Compared with Che product,Chang algorithm and AMSR2 product,the novel algorithm had higher accuracy and is more suitable for snow depth inversion in Northeast China.In addition,the "Snow Depth Dataset Production Platform" software was established for the operational production of daily snow depth datasets in Northeast China.