Soil Moisture Retrieval Based On Sentinel-1A And Landsat-8

Soil moisture,an important component of the earth's ecosystem,is an important foundation for water-heat transfer and energy exchange between terrestrial and atmospheric systems as well as the key bond between surface and groundwater circulation and the carbon cycle between lands.As one of the most commonly used surface model parameters,soil moisture plays an important role in the analysis of global water cycle regulations,establishment of watershed hydrological models,monitoring of crop growth,and drought assessment.Compared with the traditional method of soil moisture acquisition,remote sensing has become the main method for soil moisture information acquisition due to its advantages of strong macroscopicity,fast timeliness,large area dynamic monitoring,and other advantages.In recent decades,scholars have developed hundreds of soil moisture retrieval algorithms for different sensor types.Among them,the active microwave collaborative optical data retrieval algorithm for soil moisture has the characteristics of high timeliness,strong penetration,not affected by the weather,the weak effect of vegetation,etc.It has become an important method for the retrieval of surface soil moisture in vegetation cover areas.At the same time,the optical remote sensing algorithm for retrieval of soil moisture by constructing feature space with different vegetation indices has become one of the commonly used methods.This paper selects Yingke Oasis agricultural irrigation district in Zhangye area of the middle reaches of the Heihe River basin as the study area.Based on the active microwave remote sensing Sentinel-1A Sentinel data and the optical remote sensing Landsat-8 data,combined with the measured data,the two kinds of soil moisture retrieval algorithm are determined by comparing different input parameters.The comparative analysis of soil moisture retrieval algorithms was conducted.The main conclusions are as follows:(1)The Ts-NDVI and Ts-MSAVI feature spaces constructed with NDVI and MSAVI indices and surface temperature all meet the conditions of the temperature vegetation dryness index method.The correlation between the temperature vegetation dryness index of the two feature spaces and measured data shows that The Ts-MSAVI feature space is higher than the Ts-NDVI feature space,and the MTVDI index computed by Ts-MSAVI can better retrieve soil moisture status.(2)The vegetation water content calculated by NDWI index is higher than that of NDVI index.The former can effectively remove the influence of vegetation on the backscatter coefficient.Combined with the Oh model,the final soil moisture retrieval results obtained by the look-up table algorithm show that the soil water obtained by retrieval of vegetation water content of NDWI index is closer to the actual situation and the retrieval accuracy is higher than that obtained by retrieval of NDVI vegetation water content.(3)Through comparing and assessing the accuracy of MTVDI index soil moisture retrieval algorithm and NDWI index water cloud model soil moisture retrieval algorithm,the result shows that the NDWI has high correlation coefficient and low root mean square error.The retrieval of soil moisture by the active microwave cooperative optical data algorithm can better reflect the surface soil moisture status in the study area.