Integrating Seasonal Forecasts with Process-Based Crop Modeling for Responsive Adaptation to Food Risks in Sub-Saharan Africa

Understanding the impacts of climate variability on crop yields is critical for food security, particularly in Sub-Saharan Africa, where rainfed agriculture dominates and is highly sensitive to climatic changes. While process-based crop models are commonly used with long-term climate scenarios to inform transformative adaptation, integrating long-range seasonal forecasts offers an opportunity to inform short-term, responsive adaptation strategies. This study uses the LPJmL process-based hydrology-crop model with SEAS5 seasonal hindcasts as climatic inputs (temperature, precipitation, and radiation) to evaluate the skill of seasonal forecasts in predicting crop yields at lead times of one to seven months for major crops in three countries in Sub-Saharan Africa: Ghana (West Africa), Kenya (East Africa), and Zimbabwe (Southern Africa). We validate the results against the WFDE5 dataset and observed weather station data from national meteorological agencies. We calibrate LPJmL with sub-national yield data to ensure local relevance and accuracy. We use performance metrics, including cumulative probability distributions and Ranked Probability Skill Scores, to evaluate forecast reliability. By capturing interannual and intraseasonal variability, this seasonal yield forecasting can serve as an early warning system to support a range of short-term response strategies, such as agricultural measures (adjusting sowing dates, early harvest due to extreme weather events, and fertilizer application) and broader strategies that include market interventions, cash transfers, food reserve management, and food assistance programs. This study advances the integration of seasonal forecasts into process-based crop models and the use of yield forecasts for responsive adaptation strategies for food security in Sub-Saharan Africa.