Answers are grounded in this paper and cite their sources.
Accurate yield prediction under climate-induced stress is crucial for sustainable crop production. Existing models often rely on vegetation indices (VIs) derived from discrete growth stages which fail to capture the entire crop phenology and thus limit their predictive reliability particularly for broad-scale, satellite-driven agricultural monitoring. This study introduces a novel approach that integrates UAV-based multispectral imaging with double logistic function fitting to track the overall wheat phenology and predict grain yield under ‘heat’ (H) and ‘heat+drought’ (HD) induced stress in wheat. Vegetation indices (NDVI, NDRE, and gNDVI) were estimated throughout the growing season and fitted with a double logistic function to generate smooth phenological curves. From these curves, key parameters representing green-up, peak vegetation and senescence phases were extracted and used as predictors in multiple linear regression models. Among the indices, gNDVI achieved the highest yield prediction accuracy for the full season (R² = 0.96), followed by NDRE (R² = 0.95) and NDVI (R² = 0.88). For phase-specific predictions, gNDVI also performed well with R² values of 0.88 during the green-up phase and 0.91 during senescence. The study also revealed significant varietal differences in yield performance under stresses (H and HD). BWMRI1 was invariably more proficient than other varieties and yielded lowest yield loss caused by H stress (3.71%) and HD stress (58.13%) compared to yields recorded during the optimum growing season. The findings demonstrate that modeling the entire phenological trajectory rather than relying on isolated growth phases substantially enhances yield prediction.
No takes yet. Share an insight, caveat, or question.
Tuhin et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: