Inversion Of Soil Moisture In Arid Area Based On C-and L-Band SAR Images

As an important critical element of the soil and an important stress factor for arid ecosystems,soil moisture plays an important role in global and regional hydrology,water resources meteorological and climatic processes.The inversion of surface soil moisture in arid areas is not only of great significance for understanding the surface conditions,energy conversion,ecological environment and climate change,but also can provide valuable information for drought monitoring,land degradation and ecological environmental protection in arid areas.The rapid development of remote sensing technology provides great potential of multi-temporal and large-scale soil moisture inversion.Owing to its high resolution,high sensitivity to soil moisture,and distinctive advantage to observe under all-day and all-weather condition,,active microwave remote sensing has been widely used in climate and environmental monitoring,particularly in soil moisture monitoring and surface parameter inversion in arid areas.Different radar systems have different incident wavelengths(frequencies),and under different wavelengths,the penetrating capabilities and the responses to surface parameters will be different,thus differentiating their ability to climate and ecological environment monitoring.Among them,C-band(4~8GHz)and L-band(1~2GHz)are commonly used bands for soil moisture inversion.Explore the moisture inversion effects of different bands(frequency)in arid areas can help to understand the scattering characteristics of the target features.Meanwhile,it will improve our comprehension to under the band have certain reference value for understanding the remote sensing the inversion mechanism of soil moisture.In this paper,Juyanze,which is located to the southeast of Ejinaqi,Inner Mongolia,is used as the research area.Based on the measured surface parameters,soil moisture was retrieved based on RADARSAT-2 and ALOS-2/POLSAR full polarization SAR data.Firstly,the backscattering characteristics of C-band and L-band bare surface are simulated by AIEM model respectively,and the response relationship of backscattering coefficient to different radar parameters and surface parameters is analyzed.Secondly,the conversion relationship between Fresnel reflection coefficient and soil moisture,and the response relationship between the combined roughness parameters and backscattering coefficient are discussed.Finally,the empirical model is established based on the relationship between combined roughness,Fresnel reflection coefficient and backscatter coefficient.Combined with the empirical relationship between soil moisture and Fresnel reflection coefficient,soil moisture in the study area was retrieved and,and comparative analysis of the inversion effects of different bands was carried out.The main conclusions are as follows:(1)The backscattering coefficient increases linearly with the increase of radar incident frequency,soil moisture content and root mean square height,and decreases linearly with the increase of radar incident angle and correlation length.The effect of incident angle and soil moisture content on the backscattering coefficient is stable.The L-band has a smaller back-scattering coefficient than the C-band under the same surface parameters.(2)The effect of root mean square height on backscattering coefficient is more significant than the correlation length.For C-and L-band,the effect of root mean square height and correlation length on backscattering coefficient is also different.Application scope in the performance is also different.For the study area,the C-band roughness parameter is applicable to the root mean square height of less than 1.7cm and the correlation length is greater than 8cm;the L-band roughness parameter is applicable to the root mean square height of less than 2.1cm and the correlation length of more than 6cm.(3)There is a good quadratic function relationship between soil moisture content and Fresnel reflection coefficient.The combined roughness and Fresnel reflection coefficient have a good logarithmic relationship with the backscatter coefficient.Based on the above relationship,an empirical model was built to invert the soil moisture.It was verified that R~2 is0.7749,RMSE is 5.87%and MAE is 4.03%in C-band,while in L-band R~2 is 0.7809,RMSE is 5.23%and MAE is 3.64%,respectively.(4)According to the inversion results,,the soil moisture content in most areas of Juyanze is below 10%,which is generally low,but the soil moisture content of Swan Lake and East and West Juyan paleo-lake core areas is significantly higher than that of other areas.When monitoring soil moisture,the C-band is more susceptible to surface roughness and cover,while the L-band is more susceptible to deeper soil layers.