The Reversal Method Study To Moisture Content And Salinity Of Soda Saline-alkaline Soil By Integrating Optics And Microwave Remote Sensing

Soil salinization is one of the serious problems in the world which effects theeco-environment and sustainable agricultural development in arid and semi-arid areas.Western Jilin province is one of the three major regions of soda salt-alkalinized soil inthe world, and the soil salinization severely affects the ecological balance of theregion. Using remote sensing technology to rapidly and accurately monitor the changeprocess of soil salinization has great significance on the reasonable exploitation andutilization of saline-alkali land.Da’an sodic land experiment station of Northeast institute of geography andagroecology,Chinese academy of sciences carry out the research of soda alkali-salineland improvement for a long time, and studies show that soil with different alkalineproperties grown with different vegetation types, in addition, when the salinizationdegree is high, there is no vegetation growth on. The optical remote sensingtechnology can effectively monitor bare saline-alkali land and vegetation cover landwhile unable to monitor the plant height and vegetation growth effectively which cannot identify the saline-alkali land meticulously, especially the monitoring of waterand salt content change. Salinized soil with a certain water content have superiorelectrical conductivity which result to strong sensitivity between its dielectric constantand the characteristics of microwave backscattering and radiation, further more, thethe application of microwave remote sensing monitoring of salinized soil is feasible.However, signal of microwave remote sensing is often affected by factors such asvegetation and surface roughness which increases the complexity and uncertainty ofsaline-alkali land feature recognition. This paper is carried out under the supported of the Innovation Program ofChinese Academy of Sciences“the mechanism research on soil moisture and snowparameter high-precision inversion from passive microwave remote sensing(KZCX2-YW-340)” and also JingYue remote sensing experiment station and Da’ansodic land experiment station. Through effective selection of research area, and theLandsat8OLI and RADARSAT-2C band full polarization SLC image as the datasource and at the same time on the basis of ground-based and satellite-basedsynchronous observation experiment, the water and salt content inversion methodstudy of soda alkali-saline land in Western Jilin province were carried out. The mainresearch results in the following aspects:(1) Based on the SAR polarization decomposition theory, the microwave remotesensing backscattering characteristics of saline-alkali land were extracted in the studyarea which can effectively identify the mild saline-alkali land and improve theaccuracy of identification of saline-alkali soil.(2)Using the scattering entropy parameter derived by Cloude-Pottierdecomposition, this paper discusses the scattering random characters of saline-alkalisoil. The scattering entropy of saline-alkali soil is higher than cropland and waterbody which means that the scattering of saline-alkali soil is arbitrary distribution. Thehigh scattering entropy suggested that it is infeasible to carry out monitoring ofsaline-alkali soil by microwave remote sensing through establishing deterministicfunction relationship of radar scattering coefficient and dielectric constant ofsaline-alkali soil.(3)Using the variation coefficient (ratio of standard deviation and mean value) ofthree scattering mechanism parameters derived by Freeman-Durden decomposition,this paper shows that the saline-alkali soil has a relatively high discrete degree, whichpresents the large variable ranges of full polarization statistical information with25*25pixels. At the same time, this paper also illuminates that the saline-alkali soil is a mixed pixel with the randomness of grassland cover, the randomness of grasslandgrowth,the randomness of salinity content, and the the randomness of bladedirection,all of which result to the scattering randomness of a radar backscatteringecho.(4)According to the frequency response difference of saline-alkali soil dielectricconstant, the three parameters soil complex dielectric constant (real and imaginarypart) and surface root-mean-square hight are simultaneously retrieved by Levernberg-Marquardt algorithm from dual-frequency ground-based microwave brightnesstemperature observation data.(5)Combined full polarization radar and NDVI data, the experience retrievalformula of salt and water content was established, and the results RMSE of watercontent is3.169%and the RMSE of salt content is3.175g/Kg.(6)The semi-empirical retrieval program of saline-alkali soil water and saltcontent were proposed based on theoretical knowledge and simulation database, andhas carried on the preliminary implementation.First of all, establish the simulationdatabase by using AIEM model, and then through the parameter sensitivity analysis,the relationship between combined roughness, complex dielectric constant andHH\VV backscattering coefficient were proposed. Secondly, combined withwater-cloud model, the vegetation effect was removal. At last, the dielectric constantamplitude and combined roughness were simultaneously inversed.Thesis research work provides a new way for monitoring the characteristics ofsaline-alkali land by remote sensing technique.Further more, this work laid afoundation for the project approval and research of the national natural sciencefoundation item “The Reversal Method Study to Moisture Content and Salinity ofSoda Saline-alkaline Soil by Integrating Optics and Microwave Remote Sensing”.