Research On Soil Moisture Estimation Model Based On Multi-source Information

Soil moisture is the key to the connection between surface water and groundwater.It is an important influencing factor for plant growth and development,and it is the basis for flood forecasting.It is of great significance to improve the precision of soil moisture estimation model.The conventional remote sensing inversion model of soil moisture generally uses remote sensing image information as an independent variable and combines the soil moisture measurement data to fit the model.It can be seen that the conventional remote sensing inversion model ignores the soil self-information factors that have important influence on soil moisture,such as soil texture information and soil moisture parameters,in the selection of independent variables.Therefore,in this study,Zhengzhou City was taken as an example to take the time series(2-,3-,4-,5-,6-,9-,11-,12-month in 2016),based on the time series of soil moisture measurement data(Soil Moisture Observation Data of 0~10cm Sites from 16 meteorological observations in Zhengzhou City)and MODIS Remote Sensing Image Constructed Soil Moisture Estimation Model Based on Conventional Remote Sensing Inversion Model(including Temperature Vegetation Drought Index Model and Apparent Thermal Inertia Model),Based on this,additional soil moisture parameters(soil wilting point,field capacity,available water,saturation,saturated hydraulic conductivity,matric bulk density)and soil texture information(percent of sand,silt,and clay in the soil)As an independent variable,a soil moisture estimation model based on multi-source information was constructed.Finally,the soil moisture measurement data was used as a standard to verify the accuracy of the above two types of soil moisture estimation models.The results of the study are as follows:(1)Among the conventional remote sensing inversion models of soil moisture,the estimation accuracy of soil water storage capacity estimation model based on apparent thermal inertia is not high,while the accuracy of other soil moisture estimation models based on apparent thermal inertia or temperature vegetation drought index is greater than 80%,and the value of RMSE varies from 1.99 to 3.91.(2)Among the soil moisture estimation models based on multi-source information,the accuracy of the soil moisture estimation model with additional soil property information is 90% more or less.And RMSE ranges from 1.29 to 2.48,which is smaller than the corresponding soil moisture estimation model based on conventional remote sensing inversion model.The above results indicate that the conventional remote sensing inversion model of soil moisture can basically satisfy the soil moisture estimation in Zhengzhou City.Soil moisture parameters and soil texture information supplement the important factors affecting soil moisture that are not involved in conventional remote sensing inversion models.The construction of a soil moisture estimation model based on multi-source information can be used to revise the conventional remote sensing inversion model to a certain extent and improve the estimation accuracy of soil moisture.