A weather-forecast driven early warning system for wheat blast disease: User-centered design, validation, and scaling in Bangladesh and Brazil.

We describe the user-centered design (UCD) of a numerical weather-forecast-driven early warning system (EWS) as a climate service for managing wheat blast, a fungal disease capable of causing complete crop yield losses that is strongly dependent on weather conditions. Our mixed-methods process was guided by stakeholder input on the design, testing, and refinement of the EWS from agricultural extension organizations, meteorological departments, and farmers' groups in Bangladesh and Brazil, where concerns about blast disease risks are high. The UCD process led to a wheat blast disease prediction model, server systems, and user-facing enhancements, including an open-source dashboard (https://beattheblastews.net/) that displays historical, real-time, and forecasted weather data, along with geographically explicit disease predictions, to support informed decision-making on wheat blast management. We describe the back- and front-end design of the dashboard, which supports disease risk forecasting, hindcasting, and the dissemination of early warning advisories co-designed with user organizations. We validated the EWS through comparisons with field observations in both countries. Model results generally agreed with disease incidence records, and model hindcasting confirmed alignment with disease outbreak patterns in Bangladesh and Brazil. Collaboration between agricultural research, meteorological and extension organizations in developing and supplying weather forecasts, disease management advisories, and early warning systems-along with presenting hindcast validation results to stakeholders-led to the formal endorsement of the EWS in both countries. This process also enabled the registration and training of over 14,500 extension officers, lead farmers, and farmers' cooperative members who now receive advisories via email, SMS, agro-meteorological bulletins, smartphone applications, WhatsApp and social media messages. These tools support them in interpreting and sharing wheat blast early warnings with farmers to improve disease preparadness and management actions in both countries.