Research On Soil Moisture Passive Microwave Remote Sensing Inversion Method In Northeast Of China

Soil moisture in the surface of earth is only a small part of the world's totalamount of water, but it is link of surface water and ground water. Its change can affectthe distribution and evaporation process of surface energy, and control regional waterand energy cycle, thus it has great significance for research on hydro-meteorologicaland ecological environment. In addition, soil moisture determines available water ofcrop. Either low or high soil moisture will affect the growth of crops, and evendroughts and floods. So its temporal and spatial dynamic monitoring can serveprecision agriculture, drought early warning and crop yield estimation.Microwave can penetrate low vegetation canopy and detect shallow surface soilproperties. Besides, the complex relative permittivity of soil in the microwave domainhas high sensitivity to its moisture, thus microwave remote sensing has become themost effective technology to large scale soil moisture retrieval. This paper issupported by the CAS Knowledge Innovation Program of "the mechanism researchon soil moisture and snow parameter high-precision inversion from passivemicrowave remote sensing("KZCX2-YW-340). In this paper, we details the domesticand foreign progress of soil moisture inversion from passive microwave remotesensing and concises its theoretical approachs. Combined with ground-based passivemicrowave remote sensing experiment and electromagnetic wave propagation theory,we carried out in-depth research on microwave radiation model of soil coverd byvegetation and focused on soil moisture inversion from passive microwave remotesensing in the corn farmland and saline-alkali soil of Northeast China. The innovativeresearch of this paper is in the following aspects:(1) According to the relations of Fourier transform pair between spatialautocorrelation function and power spectral density, a method is proposed for surfacerough parameter estimation using pencil-beam antenna radiometer. The spatial autocorrelation function of rough surface is achieved from surface power spectraldensity measured by pencil-beam radiometer, and roughness parameter (rms heightand autocorrelation length of soil surface) is calculated from surface spatial auto corr-elation function.(2) According to the energy conservation law, microwave radiation model ofperiodic surface is established by the discrete method. The emissivity simulationresult of this model reveals that microwave radiation of periodic surface hasdirectional characteristic and the emissivity difference with a flat surface is relatedwith the ridge structure parameters (ridge height, ridge cycle). Soil moisture inversionresult from grounded and multi-angle brightness temperature confirms that the ridgestructure will affect inversion accuracy, and periodic surface radiation model willimprove soil moisture retrieval accuracy.(3) Soil moisture inversion below the corn canopy is implemented based on thelong-time series grounded remote sensing experimental data and zero-order radiationtransfer model. Compared the accuracy of four different soil moisture inversion mode,found that even in the region covered by vegetation canopy, periodic surface modelcan also improve soil moisture retrieval accuracy, and soil moisture retrieval accuracycan be improved when the relationship between vegetation optical thickness andvegetation water content was the second power function.(4) Followed on the frequency response difference of saline-alkali soil permittivity,soil moisture and salt content of saline-alkali soil are extracted from dual-frequencymicrowave brightness temperature data, the average absolute error of soil moisture isabout0.009cm3/cm3, and the average absolute error of soil salt content is about3g/kg, and the error range about0.5-12g/kg, and this confirms the feasibility andeffectiveness of saline-alkali soil parameter extraction from passive microwaveremote sensing data.(5) Combined with the Monte Carlo method and the method of moments, thesurface backscattering coefficient of cropland is simulated and the result indicates thatsurface backscattering coefficient increases with the number and size of clods, and itsangular dependence is the opposite. Based on double-layer canopy radiation transfermodel, backscattering coefficient of crop canopy is simulated, and found thatbackscattering coefficient increases with the increasing canopy water content, butdecreases with the increase of surface roughness.