2026/01/01 by Yan Zhao, Trevor Philip, Jason Brider +2 · 1 voice
Environmental Science · #Soil Moisture and Remote Sensing #Remote Sensing in Agriculture #Soil Geostatistics and Mapping
paper · doi:10.1093/insilicoplants/diag017
openalex publication_date 2026/01/01 · openalex created_date 2026/07/15 · openalex updated_date 2026/07/28
Abstract Informed decision-making ahead of planting is essential for minimizing sorghum production risk, particularly under increasing climate variability. Antecedent soil moisture status during the fallow period is a key factor shaping preseason management, influencing sowing time, cultivar selection, and nitrogen application. This study investigates the potential of using remote sensing-derived indictors to estimate absolute gravimetric soil moisture (GSM; 0–90 cm) and derive relative, decision-relevant moisture indicators for presowing assessment at field scale across sorghum-growing regions in Australia. High temporal and spatial resolution satellite observations from Sentinel-1 backscatter and Sentinel-2 shortwave infrared bands were used to derive twelve moisture-sensitive variables. These variables were aggregated over accumulation windows ranging from 30 to 270 days prior to prediction dates, capturing recent wetting and drying dynamics. GSM from 98 soil cores across north-eastern Australia, primarily on Vertosols, was used to train machine learning models. Linear regression, random forest, XGBoost and CatBoost models were evaluated using different cross validation strategies. The random forest model achieved the best performance (R2 = 0.93, root mean squared error = 3.05%) and was applied to estimate GSM across fields for mapping and benchmarking. The model was implemented to (1) estimate GSM at target presowing date, (2) generate 5-day interval predictions during the 3-month fallow period, and (3) benchmark current conditions against historical fallow periods. This framework provides a practical tool for assessing presowing soil moisture, enabling context-aware interpretation of field conditions and supporting improved decision-making, risk management, and climate resilience in dryland farming systems.