Soil Moisture Retrieval And Its Spatial Character Analysis In Bare Random Surface

The spatial distribution of surface soil moisture content (~ 5cm deep) is very important in the application of Hydrology, Climate, Vegetable Ecology, etc, and often plays as input parameters of hydrological model, climate models and ecological model. Therefore, it is an urgent need to get the space distribution information of soil moisture in big range surface. The traditional measuring methods, such as optical remote sensing and passive microwave remote sensing, have certain restrictions in obtaining soil moisture.Recent studies show that active microwave remote sensing can compensate for the insufficient of the optical and passive microwave remote sensing in soil moisture monitoring for the basin scale, and provides a new method and means on soil moisture monitoring application.Numerous studies have shown that, C band may achieve a good accuracy in estimating 0~5cm layer soil moisture, so now Radarsat-2 SAR which provide multi-polarization, multi-angle and C band SAR data, has great potential in monitoring surface soil moisture. The purpose of this study is using C band SAR data to retrieval soil moisture and analysis its spatial character in bare random rough surface.In this study, firstly, acquire the surface parameter of the study area, and through AIEM model to simulation the relationship between soil water content, surface roughness and backscatter coefficient of C band SAR signal in bare surface. Then set up the experience model using nonlinear regression method, and propose soil moisture inversion method according to the actual situation of the study area. Secondly, verify the inversion accuracy using measured data, the correlation of the inversion results and the measured data is good, the correlation coefficient is 91.29%, it shows that the inversion results can reflect the spatial distribution of soil moisture trend of study area correctly. Finally, analyze the spatial character of soil moisture of the study area through analyzing the effect of terrain, vegetation index on soil moisture and the interaction between the sample points. The results showed that the influence of the vegetation on soil moisture space distribution is quite obvious but the terrain has less effect on it, and the human impact factor still can not be ignored; Study area has medium partial to strong variation, and the pattern variation has strong space dependency; Soil moisture of study area has significant positive spatial autocorrelation, and has obvious characteristics of spatial agglomeration, but its level is not particularly high.