Inversion Of Soil Moisture In Plateau Wetland Based On SAR And Optical Data

Soil moisture is an important variable affecting the circulation of terrestrial water resources,and is the basic condition for crop growth.It is also an important parameter for crop growth monitoring,yield estimation and land drought monitoring.Optical remote sensing is vulnerable to data and other factors,especially for high-altitude plateaus,it is difficult to obtain long-term sequence optical data,so it is difficult to effectively monitor soil moisture.Microwave remote sensing is less affected by environmental factors,the ability to observe all-day,the ability to penetrate the ground and the sensitivity to soil moisture change can make up for the shortcomings of optical remote sensing,and is widely used in soil moisture research.This paper takes the Dashanbao Plateau Wetland of Zhaotong City,Yunnan Province as the research area,and uses the C-band Sentinel-1 radar data from 2015to 2017 with the Sentinel-2,Landsat 8 OLI optical data and surface measured data to carry out the soil surface layer(0-10cm)water inversion.The study area data is tailored based on land use data,and the areas where the land use types are marsh wetlands,grasslands,and reservoir wetlands are retained.The Normalized Vegetation Index(NDVI)calculated from optical data divides the marsh wetland and grassland parts during the three-year period into bare or sparse vegetation cover periods and high vegetation cover periods.Based on the analysis of soil and vegetation scattering models,based on the advanced integral equation model(AIEM)simulation analysis of radar backscatter coefficient and radar parameters,surface roughness,soil water content,through the construction of the same pole.The influence of the regression equation between different angles of backscattering coefficient and the combination of roughness parameters on the radar backscattering coefficient is established,and the soil moisture inversion model suitable for bare or sparse vegetation cover is established.Based on the water cloud model,the backscattering of vegetation canopy is obtained,and the backscattering of vegetation is removed from the radar backscattering,the disturbance of vegetation on soil moisture inversion is weakened,and the soil moisture inversion model suitable for high vegetation cover period is established.The inversion model is applied.Conduct soil water inversion in the study area,map soil moisture distribution,and based on measured data.The accuracy of the inversion results was verified.Utilize 36-scenes Sentinel-1 data to extract the monthly water area of the reservoir wetland in three years based on the support vector machine(SVM)method.Statistics and analysis of the change law of reservoir wetland area.The main findings of this paper are as follows:(1)According to the two vegetation cover periods,the soil moisture inversion model is established.The backscatter coefficient values simulated by the AIEM model are compared with the backward scattering coefficient values simulated by the inversion model.The correlation coefficient(R~2)is 0.92.The correlation between the two is good,indicating that the inversion model has certain feasibility.Furthermore,the inversion model is developed by Taylor series,so that the model can perform soil moisture inversion without the measured surface roughness parameters,and improve the accuracy of model inversion.(2)Based on the constructed inversion model and surface measured data,the spatial distribution of soil water content in Dashanbao Wetland in March 2015,October 2016,and September 2017 was obtained,the overall performance was that the soil was in a humid state,the soil moisture content in the marsh wetland distribution area is higher than that in the grassland,and with the increase of distance from the marsh wetland,the soil water content gradually decreases.From the overall trend of returning farmland wetland near the sample point,the soil water content increased with the increase of the returning farmland,and the performance increased first,then decreased and then increased,which was unstable;the soil moisture inversion value and measured in 2015-2017 were compared and analyzed.It is found that the comprehensive utilization of radar and optical data can obtain better inversion results during bare or sparse vegetation coverage and high vegetation coverage.R~2 is 0.72,0.87,0.94 respectively,and root mean square error(RMSE)is0.199 cm~3/cm~3,0.054 cm~3/cm~3,0.015 cm~3/cm~3 respectively.(3)Through the monthly variation time series analysis of the reservoir wetland,it is found that the water area of the Dashanbao Reservoir wetland increases from July to October,and the area decreases from November to June,and the wetland area of the Dashanbao Reservoir changes seasonally.Consistent with regional climate change;the three-year water area generally showed an upward trend,but the specific performance was that the first point was to fall after 2016,and there was a positive correlation with the inversion of regional soil water content,indicating that the reservoir wetland has soil moisture.Has a certain impact.The soil moisture inversion results of the Dashanbao Plateau wetland based on SAR and optical data are basically consistent with the actual situation,indicating that the method of inversion of soil moisture in the plateau wetland using SAR and optical data is feasible,which can provide support and basis for the restoration,protection and rational development and utilization of plateau wetlands.