Research On Methods For Soil Moisture Retrieval In Prairies Areas Based On Multi-frequency And Multi-polarization SAR Data

Soil moisture plays an important role in the global and regional hydrologicial and meteorological processes,which is considerded as an indispensable state variable in Earth science researches.Especially in arid,semi-arid,and alpine prairie areas where the natural environment is rigorous,soil moisture has become the most important factor for influencing the vegetation phenology.The development of remote sensing technology makes it possible to monitor the change of soil moisture at different temporal and spatial resolutions.Due to its high sensitivity to soil moisture and all-weather and all-time work capability,microwave remote sensing has been widely used for soil moisture retrieval under bare soil surface and vegetated areas.However,scattering contributions induced by vegetation and surface roughness reduce the sensitivity of microwave signal to soil moisture,which increases the complexity and difficulty for soil moisture retrieval.Therefore,simultaneously eliminating the influence of vegetation and surface roughness has become the critical technical problems for establishing the methods for soil moisture retrieval in prairie areas.Based on synthetic aperture radar(SAR)data,optical remote sensing data,and ground measured datasets as important data sources,three methods for soil moisture retrieval are constructed by coupling the soil dielectric constant models,bare surface scattering models,and vegetation scattering models,which are tested and validated in the Wutumeiren prairie,Ruoergai prairie,and Qinghai Lake watershed.The main works of this dissertation are summaried as following:(1)One method is developed to estimate soil moisture in prairie areas based on advanced integral equation model(AIEM)and ratio method,which is depended on the measured roughness parameters used to parameterize AIEM.It makes full use of ground measured roughness parameters to parameterize AIEM,which helps to establish the relationship between the backscattering coefficients of bare soil surface and soil moisture and calibrate the unknowns in the ratio method.In the established method for soil moisture retrieval,AIEM is served to simulate the backscattering coefficients of the bare soil surface,and the ratio method is applied to separate the scattering contribution of vegetation.Four different vegetation parameters including leaf area index(LAI),vegetation water content(VWC),normalized difference vegetation index(NDVI),and enhanced vegetation index(EVI)are used to parameterize the ratio method,respectively.The results compared with ground measured soil moisture show that the developed method can be used for soil moisture retrieval in prairie areas.Simultaneously,it is found that LAI is more suitable for parameterization of vegetation compared with other vegetation parameters in the Wutumeiren prairie,and LAI,NDVI,and EVI all can be used for characterizing the scattering mechanisms of grass in Qinghai Lake watershed.(2)The quantitative relationship between the dielectric constant and observed co-polarized backscattering coefficients is established,which is used to retrieve soil moisture in prairie areas.This method fully utilized the multi-polarimetric SAR to eliminate the dependence on surface roughness parameters.The objective of this method is to consider how to make full use of multi-polarimetric SAR to realize the retrieval of soil moisture without the priori konwlege of surface roughness parameters.Firstly,Dubois model is simplified and transformed to establish the relationship between the dielectric constant and backscattering coefficients of bare soil surface.Then the ratio method and water cloud model(WCM)are both used to eliminate the scattering contribution of vegetation,while the scattering mechanism of vegetation is characterized by LAI,VWC,NDVI,and EVI,respectively.The results compared with ground measurements show that the proposed method could effectively solve the problem of surface roughness parameters and vegetation.Analyzed the vegetation parameterization in these two study areas,the LAI and EVI are suggested for the Wutumeiren and Ruoergai prairies,respectively.This method does not depend on any surface roughness parameters,which greatly improves the applicability.(3)Based on AIEM,ratio method,and effective roughness parameters,a method suitable for soil moisture retrieval in prairie areas is established.The objective of this method is to realize the soil moisture retrieval while taking into account the surface roughness parameters but not depending on the ground measured surface roughness parameters.In the constructed method,AIEM is used for simulating the backscattering coefficients of bare soil surface with the given roughness parameetrs used as the input parameters.The ratio method is responsible for separating the scattering contribution of vegetation,and four vegetation parameters including LAI,VWC,NDVI,and EVI are used to characterize the scattering mechanism of vegetation.The experiments results show that the proposed method can be used for estimating the soil moisture in prairie areas,and the retrieval accuracy is obviously improved.For better parameterization of the ratio method,LAI and EVI are recommended to describe the scattering mechanisms of vegetation in the Wutumeiren and Ruoergai prairies,respectively.The LAI,NDVI,and EVI can be used for characterizing the scattering contribution of vegetation in the Qinghai Lake watershed.This method takes into account the contribution of surface roughness parameter to the backscattering coefficient.However,at the same time the involvement of effective roughness parameters eliminates the dependence on the measured roughness paraemters,which greatly improves the applicability.(4)Based on the polarimetric feature parameters extracted from full-polarized Radarsat-2 data and multiple linear regression equations,we explored the feasibility of polarimetric feature parameters for estimating soil moisture in prairie areas.It tries to solve for the soil moisture from the polarimetric feature parameters while keeping the contributions of vegetation and surface roughness.The polarimetric feature parameters considered includes the Cloude decomposition parameters(polarimetric entropy,scattering alpha,and ansitropy),three eigenvalues,combination parameters of three eigenvalues(single bounce eigenvalue relative difference,double bounce eigenvalue relative difference,and radar vegetation index),and Freeman decomposition parameters(single-boucne scattering,double-bounce scattering,and volume scattering).Experiments condutcted in the Wutumeiren and Ruoergai prairies have validated the usefulness of polarimetric feature paramters for helping soil moisture retrieval.The most challenging problem for soil moisture retrieval in vegetated area is how to eliminate the influence of vegetation and roughnes parameters simultaneously.In this dissertation,three methods based on soil dielectric constant models,bare surface scattering models,and vegetation scattering models are constructed to estimate soil moisture in prairie areas.The breakthrough of these theories and methods will provide new theoretical and methodological support for soil moisture retrieveal in prairie areas.