Seasonal disease risk prediction using disease epidemiological models and seasonal forecasts has been actively sought over the last decades, as it has been believed to be a key component in the disease early warning system for the pre-season planning of local or national level disease control. We conducted a retrospective study using the wheat blast outbreaks in Bangladesh, which occurred for the first time in Asia in 2016, to study a what-if scenario that if there was seasonal disease risk prediction at that time, the epidemics could be prevented or reduced through prediction-based interventions. Two factors govern the answer: the seasonal disease risk prediction is accurate enough to use, and there are effective and realistic control measures to be used upon the prediction. In this study, we focused on the former. To simulate the wheat blast risk and wheat yield in the target region, a high-resolution climate reanalysis product and spatiotemporally downscaled seasonal climate forecasts from eight global climate models were used as inputs for both models. The calibrated wheat blast model successfully simulated the spatial pattern of disease epidemics during the 2014–2018 seasons and was subsequently used to generate seasonal wheat blast risk prediction before each winter season starts. The predictability of the resulting predictions was evaluated against observation-based model simulations. The potential value of utilizing the seasonal wheat blast risk prediction was examined by comparing actual yields resulting from the risk-averse (proactive) and risk-disregarding (conservative) decisions. Overall, our results from this retrospective study showed the feasibility of seasonal forecast-based early warning system for the pre-season strategic interventions of forecasted wheat blast in Bangladesh.
The development of an early warning system for climate-sensitive disease risk with a focus on dengue epidemics in Southeast Brazil
