Answers are grounded in this paper and cite their sources.
Despite uniform climate and input management, sugarcane production in southwestern Iran exhibits substantial yield variability a persistent yield gap that challenges the predictive capacity of precision agriculture. This study developed an integrated remote sensing and machine learning framework to diagnose the primary drivers of this variability across 336 fields. Daily Crop Water Stress Index (CWSI) maps were derived from fused Landsat-MODIS thermal data and combined with field-level datasets, including soil salinity, water table depth, soil texture, crop age, and crop variety. A Random Forest model (70:30 training–testing split) was utilized to quantify predictor importance, evaluate stress interactions, and predict sugarcane yield. The model achieved high predictive accuracy (R ^2 = 0.93, RMSE= 5.38 t.ha −1 ). Feature importance analysis identified water stress (CWSI) as the dominant driver of yield variability (40.6%), followed by soil salinity (27.1%) and crop age (9.0%). Furthermore, strong synergistic interactions were observed between water and salinity stresses, leading to disproportionate yield losses that accelerated markedly when high salinity coincided with elevated CWSI, particularly in ratoon fields. These findings demonstrate that integrating remote sensing with machine learning can uncover hidden subsurface stressors and support dynamic, zone-based precision management. The proposed framework offers a scalable diagnostic tool for optimizing water use, mitigating salinity impacts, and enhancing agricultural sustainability in water-limited environments.
No takes yet. Share an insight, caveat, or question.
Motlagh et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: