Agricultural production is increasingly exposed to hydroclimatic extremes, including drought, extreme heat, heavy precipitation, flooding, frost, and water scarcity. Yet detecting a hydroclimatic extreme alone does not establish when and where crops will experience damaging stress, how severe that stress will become, or whether it will lead to substantial yield losses or crop failure.
The growing availability of Earth-observation, meteorological, reanalysis, and seasonal forecast data, together with advances in crop and hydrological modelling, statistics, and artificial intelligence, provides new opportunities to complement retrospective monitoring with crop-impact prediction and early warning. Realising this potential requires robust quantification of predictive uncertainty, rigorous evaluation of model performance during extreme events, and assessment of how well prediction systems generalise across years, crop types, environments, and regions.
Cross-regional comparisons can therefore help identify key controls on crop vulnerability and resilience while improving the transferability and interpretation of crop-risk prediction systems.
Another important challenge is predicting whether, and to what extent, crops can recover following hydroclimatic stress and how this recovery affects final yield.
We invite contributions addressing, but not limited to:
- crop responses to individual, sequential, and compound hydroclimatic extremes, including drought, extreme heat, frost, heavy precipitation, flooding, and water scarcity;
- early and in-season prediction of crop stress, yield anomalies, yield loss, and crop-failure risk;
- Earth-observation indicators for detecting and anticipating crop stress, including vegetation condition, evapotranspiration, soil moisture, and canopy temperature;
- advances in data-driven, process-based, physics-informed, and hybrid prediction frameworks that draw on statistical modelling, AI, and crop, hydrological, and land-surface models;
- probabilistic prediction, uncertainty quantification, forecast calibration, and reliability;
- crop recovery following hydroclimatic stress and differentiation between recoverable and yield-damaging stress;
- identification of intervention windows and decision-relevant thresholds for agricultural early warning.
The session aims to advance robust and transferable approaches for predicting crop stress and yield risk and translating these predictions into actionable agricultural early warning.