Integration of seasonal precipitation forecast information into local-level agricultural decision-making using an agent-based model to support community adaptation

Accessibility and variability of water resources can have profound impacts on social, political, and economic security. In regions with pronounced climate variability (e.g., seasonal and inter-annual variability in precipitation), seasonal climate forecasts issued in advance may enhance sectoral planning and management decisions to benefit vulnerable communities. Yet, as the development and communication of seasonal climate forecasts continue to advance, integration of forecasts into decision-making remains sparse. This work investigates the integration of a locally-tailored seasonal precipitation forecast into agricultural decision-making using a simple agent-based model designed to resemble a stylized local Ethiopian community, to understand factors that may influence adoption. We do not make claims on representativeness, yet results from our model indicate that forecasts improve gross benefit to farmer agents across different climate series, with potential for improved profit, yields, and nutritional outcomes. Accuracy of a seasonal forecast seems to correlate with increased adoption and therefore benefit; yet, the sequence of precipitation conditions, risk preference and heuristics for building trust nuance this relationship. Further, similar to well-established literature in economics and sociology, our stylized model suggests that community-level social dynamics (e.g., peer interaction, sensing others' trust in the forecast, and the ability to learn from peers) seem to have a large impact on patterns of forecast adoption. Ultimately, if the motivation for seasonal forecast development is to enhance water and food security for adaptation to climate variability in vulnerable regions, then interdisciplinary collaborations that connect local-scale forecasts with public engagement and attention to community-level social dynamics are critical.