Adapting disease forecasting models to coarser scales: Global potato late blight prediction
Many predictive models of plant disease rely upon fine-scale weather data collected in hourly increments, or finer. This data requirement is a major constraint when applying disease prediction models in areas of the world where hourly weather data are unreliable or unavailable. In response to the need to apply predictive models where only coarse resolution weather data are available, we developed a framework to adapt an existing potato late blight forecast model, SimCast. We envision this type of coarser resolution model being useful in long term decision making rather than for within growing season. For long term modeling we may be satisfied with being able to estimate the magnitude of upward and downward trends.