Research Of Remote Sensing Snow Product Assimilation Based On CLDAS And Its Application

Snow cover is an essential component of climate,which is widely distributed in the frozen circle with the most significant seasonal and interannual variations.Because of its high albedo,low roughness and low heat transfer characteristics,the thermal and hydrological effects of snow cover will infulence soil temperature and humidity and surface temperature.There are a wide range of snow cover data sources.The time range of the stations observation is long enough but exist representative problem.The spatial and temporal resolution of remote sensing product is relatively high,but the inversion algorithm is affected by various factors,and the results of the land surface model are uncertain.Data assimilation technology can combine the advantages of these data to generate high-quality snow cover data.Therefore,it is of great significance for improving snow simulation for numerical weather prediction and short-term climate prediction by carrying out land surface data assimilation research based on satellite snow products.This paper firstly analyzes the quality of a variety of satellite-based snow products and the snow cover simulation of Noah land surface model,and determines the observation error and model error;and further assimilate the snow cover fraction with quality control from a microwave radiation imager snow depth product.Based on the method of EnKF+DI assimilation,parameter optimization was carried out to improve the assimilation in China;assimilation of microwave snow depth products was carried out in China and the effects of snow assimilation on soil temperature and humidity and surface temperature was studied.The main conclusions are as follows:(1)In order to verify the precision of retrieval algorithms and provide an objective evidences for climate studies,the precision of snow products is evaluated.To verify the quality of MODIS(Moderate-Resolution Imaging Spectroradiometer)and FY-3(Fengyun-3)satellite remote sensing snow products,and to quantitatively analyze observation errors.The analysis shows that the two products maintain a good spatial-temporal consistency,but may be affected by the difference in the processing of cloud detection,and the snow-free consistency of the two products decreases slightly during the snow melting period.In addition,there were significant interannual,seasonal,and monthly variations in snow cover fraction deviations for the two products.Compared with MODIS,FY-3 and IMS have a high consistency of snow pixels detected during the snow period(Interactive Multi-sensor Snow and Snow Mapping System).The number of snow pixels detected by MODIS is slightly higher than that of FY-3.The deviation and false alarm rate of MODIS and FY-3 are different in the Qinghai-Tibet Plateau.Both of two products performed well in the cropland.(2)Compared the effects of different forcing on the model results,and analyze the effects of the improved precipitation forcing on the snow cover simulation results in China.Determine the error of Noah3.6 snow cover simulation model under different precipitation forcing conditions: CLDAS(CMA Land Data Assistance System)and CLDAS-Prcp.This paper adopts CLDAS and CLDAS-Prcp forcing drives Noah3.6 land surface model to simulate and evaluate snow cover fraction(SCF),snow depth(SD),the snow water equivalent(SWE)in major snow area,such as Northeastern China,Xinjiang and Qinghai-Tibet Plateau.The result shows that CLDAS-Prcp can improve snow simulation in the winter,removes poor snow simulation due to underestimates of precipitation of CLDAS.Result is the most consistent with observations in northeastern China.The improvement of snow water equivalent is most obvious than other snow variables.CLDAS-Prcp not only has better simulation capability,but also slightly accurately reflects the extreme snow events.(3)In order to reduce that influence of unreasonable information(caused by cloud pollution and other factor)in the process of assimilating snow cover fraction,In the process of assimilation,it is proposed to use microwave snow depth products to control the quality of snow cover products to be assimilated(the quality of microwave snow cover products is less affected by weather and cloud).Comparison of assimilation results with MODIS snow cover fraction and snow depth shows that SCFDA?WSD(with control quality of snow depth product)also showed greater improvements during the snow accumulation and snowmelt periods than the SCFDA(snow cover products without snow depth products).The SCFDA?WSD improvements for woodland and shrub land are the most obvious.At different altitudes,the effects of the SCFDA?WSD are basically equivalent,and the deeper the snow depth is,the better the effect.Simultaneously,specific parameters suitable for Noah land surface model,FY-3 snow products and China area are proposed.Based on the method of EnKF(Ensemble Kalman Filter)+DI(Direct Insertion)assimilation,parameter optimization was carried out to improve the snow assimilation in China.The method determines the increment of snow water equivalent in the model according to the average relative error of observation and model,and improves the assimilation effect of snow cover in China.The results of assimilation showed that the improved landcover of the optimization method of EnKF + DI assimilation parameters were mainly concentrated in cropland compared with that of EnKF + DI method.Evaluated by snow depth of in-situ,it is found that the optimization method of EnKF + DI assimilation parameters is more significant in the stations of cropland,grassland,forest,shrubland and snow depth are in the range of 0.05 m to 0.1 m.Compared with other snow cover assimilation schemes,the optimization method based on EnKF + DI assimilation parameters is the most optimal,which is more suitable for solving the problem of snow assimilation in complex underlying surface and terrain areas.(4)Compared with the assimilation of snow cover fraction,the assimilation of snow depth products was carried out in China,and the Long-term snow depth dataset of China was assimilated,and the assimilation of snow depth was controlled with snow cover fraction products.The results of assimilation in Northeast China,Xinjiang and Qinghai-Tibet Plateau region show that the snow depth assimilation can make up for the shortage of precipitation forcing and improve the snow depth simulation.The snow depth assimilation method using snow cover fraction to control the quality of snow depth products can obtain better results in the snow melting period in the northeastern China.(5)The effects of snow cover assimilation and snow depth assimilation on soil moisture,soil temperature and surface temperature were evaluated in spring.The results show that snow cover assimilation and snow depth assimilation have different effects on soil moisture,and the effect of snow depth assimilation on soil moisture is more obvious when precipitation forcing is poor.The improvement effect of snow depth assimilation on the site soil temperature and surface temperature is related to land cover and soil texture.The improvement on the grassland and sandy clay loam station is the most obvious.The effect of snow cover assimilation on soil temperature depends on latitude and snow thickness.The accuracy of soil temperature improved by the thermal and hydrological effects of snow cover.The influence of snow cover assimilation on surface temperature is related to the change of snow cover area.