Inversion Of Soil Surface Roughness By Combining Optical And SAR Data

The roughness of soil surface affects the reflection and scattering characteristics of the incident electromagnetic wave and controls the energy balance of the bare soil surface,which plays a crucial role in meteorology,hydrology,agriculture,geology and planetary science.At present,the accurate acquisition of soil surface roughness parameters is still one of the difficulties in quantitative remote sensing research.This dissertation is supported by the Basic Work of the Ministry of Science and Technology Project and the National Natural Science Foundation of China and dependent on the Jingyuetan Remote Sensing Experiment Station of the Chinese Academy of Sciences in Changchun.The soil surface roughness inversion was studied by conducting a series of observation experiments.The main research results are as follows:(1)A large number of spectral measurement experiments and soil surface roughness measurement experiments have been carried out.The variation of soil surface reflectance spectral with soil type and soil surface structure were analyzed.A set of effective methods for measuring and calculating soil surface roughness were summarized,and the software of "Soil Surface Roughness Calculating System" have been developed.(2)The soil surface roughness parameters described by radar remote sensing and optical remote sensing are different.In this study,a large number of theoretical simulations were used to reproduce the surface structure of random rough farmland.The effective soil surface roughness factor Zg was used to combine optical remote sensing and radar remote sensing.The expressions between the Zg and the mean slope angleq under different surface distribution functions(gaussian,exponential or fractal)are established.(3)Soil surface roughness(RMSH)and soil moisture(SM)were taken as two independent variables to analyze their role on the reflectance spectral data measured on the bare soil surface in field condition.By a lot of regression analysis,RMSH and SM estimation models based on multi-band reflectance were established.The root-mean-square error(RMSE)between the estimated value and the measured value of RMSH and SM are 0.69 cm and 2.6%,respectively.(4)Based on Beer-Lambert law and Hapke shadow model,the soil moisture and soil surface roughness were introduced into Hapke model.According to the physical meaning of the parameters and the results of global sensitivity analysis,the dimension reduction of Hapke model was realized,and the model is simplified to the expression only about the mean slope angle q and SM.Using multi-band reflectance data,the retrieval of soil surface roughness and soil moistue was realized.The model precision was verified with the Sentinel-2 satellite data.The RMSE values of the mean slope angleq and SM are 3.77°and 4.42%,respectively.(5)The soil surface scattering characteristics under different RMSH,Cl,SM and incidence angles were simulated based on IEM model and Oh empirical model.Based on the simulation database,the expression of the backscattering coefficient of C band with effective roughness factors Zg and SM under VH and VV polarization was established,and the influence of the incidence angle on the fitting coefficient was quantified.(6)By using spectral reflectance data and backscatter coefficient data,and combining with the optical bidirectional reflectance model and the radar backscatter model,the inversion model of soil surface roughness was established.The soil surface roughness,soil moisture and dry soil reflectance can be obtained simultaneously.This study provides a new idea for the future research of optical and radar remote sensing.