Accuracy Verification Of A New Remote Sensing Soil Moisture Product And Its Application In Surface Evapotranspiration Calculation

Soil moisture is the key factor of hydrological process,bioecological process and biogeochemical process.The long-term observation of regional and global soil moisture is very important for the study of drought and flood disaster monitoring,water resources management and crop yield prediction.Microwave remote sensing has become the most effective means to continuously monitor soil moisture on a large scale,and a large number of microwave remote sensing water products have been released by institutions at home and abroad.At present,the authenticity test of quantitative remote sensing products generally has the problems of global practicability and comprehensive verification,which seriously restricts the popularization and application of remote sensing products.With the continuous improvement of the accuracy requirements of quantitative remote sensing products for business and engineering mobilization,the authenticity inspection of remote sensing products has become extremely important.In previous studies,the authenticity test of remote sensing soil moisture products is mainly aimed at the evaluation of remote sensing soil moisture data in a short time span in a specific region.For the lack of verification and analysis in a long time span around the world.This paper selects three satellite soil moisture products: ESA soil moisture and ocean salinity satellite soil product（SMOS）,NASA standard algorithm soil moisture product（JAXA AMSR2）and NASA soil moisture product（SMAP）,and uses the measured data of FLUXNET2015 data set and ISMN data set,The research focuses on the authenticity verification of soil moisture products from microwave remote sensing satellites.When testing and analyzing the authenticity and error sources of three soil moisture products based on traditional methods,first count the time matching sites between FLUXNET data set and SMOS,SMAP and AMSR2 products,then make spatial maps all over the world,and determine the research area;Then,based on the traditional statistical analysis methods,some networks in FLUXNET2015 dataset and ISMN are compared with SMOS,SMAP and AMSR2 products respectively,and the accuracy is verified and analyzed by using the traditional statistical methods（mean deviation,root mean square error,unbiased root mean square error and time series correlation coefficient）,so as to provide a more comprehensive description of product performance.In order to study the significance of remote sensing soil moisture products in the calculation of surface evapotranspiration,after verifying the remote sensing soil moisture products,the US-Var station in the FLUXNET 2015 data set is selected.Firstly,the parameter evaluation model is input with the FLUXNET measured data,and then the SMOS and AMSR2 data are used as the remote sensing data into the evapotranspiration estimation model based on stability to calculate the evapotranspiration and evaluate the calculation results of different remote sensing products.The research results of this paper include the following aspects:In the verification of soil moisture of SMOS and AMSR2 products based on FLUXNET2015 data set:（1）the inversion accuracy of soil moisture products of SMOS and AMSR2 products in the area covered by low and sparse vegetation is higher than that in the area covered by dense vegetation,and the inversion accuracy is low under the surface covered with coniferous forest;（2）SMOS products mostly underestimate the measured value,while amsr2 products mostly overestimate the measured value;（3）On the whole,the inversion value of SMOS products has a certain "dry value deviation" from the measured value.In the soil moisture verification of SMAP and AMSR2 products based on ISMN International Soil Moisture Network:（1）the inversion accuracy of soil moisture products of SMAP and AMSR2 products in the area covered by low and sparse vegetation is higher than that in the area covered by dense vegetation,and the inversion accuracy is lower when the surface is covered by wet underground;（2）The remote sensing soil moisture products underestimate the measured value in the period of more and frequent water,and overestimate the measured value in the period of less and infrequent precipitation;（3）The accuracy of SMAP products is higher than AMSR2 products,and the data values of SMAP products are mostly lower than AMSR2 products.The parameter optimization method based on soil moisture and temperature stability has been well verified at the US-Var site in the southeast of Sacramento,USA,and the root mean square error of evapotranspiration is less than 25W/m^2;When moisture data is brought into the evapotranspiration model,more accurate soil moisture products can be used to calculate more accurate evapotranspiration.