Improvement and Validation of Deep Soil Moisture Estimation Model

Soil moisture is an important variable in the water cycle of farmland. Obtaining regional soil moisture quickly and accurately is of great significance for monitoring the occurrence and development of agricultural drought, crop growth status, and so on. Taking the winter wheat planting area of Yongnian District in Hebei Province as the research area, the field soil moisture monitoring experiment was during the critical growth period of winter wheat. The inversion of surface soil moisture in the study area was carried out in conjunction with Sentinel-1A SAR data and Landsat 8 OLI data. At the same time, based on the change characteristics of soil moisture in different soil layers, combined with the inflection point of soil moisture change curve, the original single and universal deep soil moisture estimation model was improved, and a more accurate segmented deep soil moisture estimation model at the surface scale was constructed. A method for quickly and efficiently estimating a large area of deep soil moisture in the region through remote sensing data was proposed. The results indicate that compared with the single and universal deep soil moisture estimation model, the correlation between the estimated and measured values of soil volumetric moisture content obtained through the established segmented soil moisture estimation model at the surface scale is significantly increased, indicating a significant improvement in the estimation accuracy of the established model. Using the proposed method, which first combines multi-source remote sensing data and uses the established surface soil moisture inversion model to invert the surface soil moisture in the region, and then uses the established segmented soil moisture estimation model at the surface scale to estimate the deep soil moisture, can quickly and efficiently obtain a large area of deep soil moisture in the region. The research results can not only provide data support for local water resource scheduling, irrigation management and other work, but also provide reference for using remote sensing data to quickly and efficiently invert large-area deep soil moisture in the region.