| The Hetao irrigation area is an important grain production area in China,where the serious problem of soil salinization has become an important factor limiting the development of local agricultural production.Timely and accurate monitoring of soil salinity dynamics in irrigation areas is of great importance.Currently,research on monitoring soil salinity using remote sensing has made some progress,but the accuracy can be limited by vegetation cover and soil moisture-salinity coexistence.Therefore,it is necessary to study the interaction mechanism between soil moisture,salinity,and spectrum to provide strategies to improve the monitoring accuracy of soil salinity.In this paper,the soil samples collected from April to October 2019 in Shahao Channels of Hetao irrigation area were used as soil moisture and salinity data,the Sentinel-2 satellite acquired simultaneously were used as spectral data.According to different surface cover conditions,the models for quantifying the spectrum response to soil moisture and salinity were constructed separately to analyze variation in spectral reflectance and explore the mechanism of soil moisture and salinity on spectrum,so as to achieve accurate monitoring of soil salinity under different different vegetation coverage based on remote sensing technology.The research results provide a theoretical basis and technical means for salinity monitoring in irrigation areas using optical satellite remote sensing.It is hoped that the above research can contribute to the development of precision agriculture.The main findings of the study are as follows:(1)There exists an interactive effect between soil moisture and salinity that can affect variation in soil spectrum.Compared with theoretical model for quantifying the spectrum response to soil moisture and salinity,interaction model can more accurately estimate the spectral reflectance under different moisture-salinity conditions,with correlation coefficients improving from0.14～0.44 to 0.29～0.59 and root mean square errors decreasing from 0.032～0.082 to0.029～0.068,which indicating that the interaction effect helps to improve the accuracy of soil moisture and salinity monitoring when using satellite remote sensing.(2)Soil moisture and salinity have promoting and inhibiting effects on vegetation spectrum,respectively,and can affect vegetation spectra in an interactive manner,causing spectral reflectance to vary at the same soil moisture(salinity)content.Among the components of vegetation spectra,soil moisture is the primary factor affecting the variation of vegetation spectra.Effect of vegetation cover degree on the mechanism of action between soil water salinity and vegetation spectrum.With the increase of vegetation cover coverage,the soil depth range corresponding to the soil water and salt variables in the ORVS model was deeper,the influence of soil water and salt on the changing distribution pattern of vegetation spectra weakened,the proportion of soil water in the spectral composition decreased,and the proportion of soil salt in the spectral composition increased.(3)Accurate monitoring of soil salinity under vegetation cover conditions can be achieved by identifying depth and eliminating moisture interference,when using the inversion method based on decision tree model and machine learning techniques.The coefficient of determination,root mean square error,relative distribution error,and relative analysis error between the measured and simulated values were 0.77,9.43 m S/cm,2.65,and 8.99,respectively.The results of the study provide a way to monitor the dynamics of soil salinity during crop fertility in the irrigation area. |
PDF Full Text Request