Predicting Local Disease Risk Indicators with Multi-Scale Weather, Land & Crop Data

Fungal and bacterial diseases in crops are so strongly dependent on specific environmental conditions that the risk of disease can be predicted from microclimate data such as temperature, relative humidity, solar radiation, wind speed, and surface wetness duration (SWD). Despite the evidence that disease risk modeling provides guidance for efficient spray utilization with no added risk of yield loss and is an improvement over calendar-based spray programs, these systems have failed to achieve widespread use. The inconvenience and expense of local monitoring of key variables such as SWD are factors preventing these models from being widely used. This proposal is for the development of web-based Geographical Information System (GIS) tools that ultimately can be used to map disease risk at fine spatial and temporal scales. The Phase I project is a proof-of-concept that involves the development of two Artificial Neural Networks (ANNs) to estimate key environmental disease risk factors at local scales from local and regional GIS information, weather station data, and site specific sensing data. An extensive dataset collected at Geneva, New York, as well as regional geographic and weather data, will be used to conduct and assess the research.