Enhanced L-MEB Model for Soil Moisture Retrieval Over Soybean Fields During the Growing Season

Soybean, a pivotal global source of oil and protein, exhibits heightened sensitivity to soil moisture (SM) conditions throughout its growth cycle. Accurate monitoring of SM in soybean fields during the growing season is indispensable for optimizing yields and forecasting sustainable agricultural practices. By leveraging advancements in remote sensing technology, passive microwave SM retrieval has emerged as a crucial tool for large-scale precision agriculture and enduring environmental monitoring. Challenges in the L-band microwave emission of the biosphere (L-MEB) model, particularly in the computation of vegetation transmissivity, may, however, compromise the accuracy of SM retrieval. In this study, we improved the Beer-Lambert law to more accurately quantify the attenuation effect of the vegetation layer on microwave signals, aiming to ameliorate the inherent limitations in the L-MEB model. The proposed SM retrieval method, primarily validated in soybean fields, was also subjected to supplementary experiments in canola and wheat fields to further assess its effectiveness and generalizability. The proposed method integrates passive microwave and optical data, demonstrating a substantial improvement in accuracy. Experimental results reveal that our enhanced method significantly outperforms the L-MEB model in soybean fields: Pearson correlation coefficients of SM, derived using vegetation water content (VWC) and leaf area index (LAI), are 0.712 and 0.692, respectively. Root-mean-square errors (RMSEs) have, furthermore, decreased to 0.056 m3/m3 and 0.050 m3/m3, a reduction of 39.78% and 19.35%, respectively. In canola and wheat fields, the method exhibited an approximate 10% enhancement in retrieval accuracy. This advancement not only furnishes novel technical support for water management in soybean cultivation but also contributes theoretical and technical insights to the domain of passive microwave SM retrieval.