From climate forecast skill to farm profit: A framework to value seasonal climate forecasts for maize production in Ethiopia

Seasonal climate forecasts can support climate-smart agronomic decisions, but their practical value depends on whether they improve farm-level outcomes relative to traditional practices or simple climatology-based strategies. This progress report documents an end-to-end framework that links (i) probabilistic forecast skill, (ii) forecast-informed decision optimization, and (iii) economic and risk metrics for maize production in Ethiopia. The core idea is to treat climate forecasts not as passive information products, but as decision inputs that parameterize agronomic choices such as planting date, cultivar maturity class, and fertilizer rates. We operatically implement this by combining forecast probabilities (monthly terciles) with a yield response model, then optimizing management choices under forecast and under climatology. The difference in expected profit and downside risk between these strategies defines the value of the forecast. Preliminary results indicate that forecast skill and economic value vary across months and locations. Decision changes are most frequent when forecast probabilities are confident and when the yield response surface is steep with respect to rainfall timing and input rates. Importantly, the valuation framework also quantifies risk reduction (e.g., lower probability of large losses, improved lower-tail outcomes) which matters for smallholders and for advisory design. This report is intended to provide a transparent technical basis for the ongoing analysis. The next phase will strengthen calibration, expand sensitivity analysis, and refine cost and price assumptions in partnership with stakeholders.