SISALERT - A Methodological Aproach for Decision Making On Plant Disease Management.

Disease forecasting has become an established component of quantitative epidemiology and is a rapidly developing area in disease management. Currently, a great number of plant disease models are available, however few have really been used in decision making. However, plant disease models can have a more practical application by taking the advantage of combining on-line weather data, weather forecast, computational techniques and biophysical models. This work has as objective to present a novel approach for the development of a decision support system based on existing knowledge and up to date computational technologies aiming better disease management. The methodology proposed was mainly based on design pattern MVC (Model-Vision-Control), programming layers and object oriented languages. Specific servers were designed for different tasks. These was conceived as a modular and adaptable platform. The pathosystem Gibberella-Wheat was used in this work due to its economic importance and great losses observed in epidemics years. Its was implemented and validated in an important wheat growing region in Brazil, known as the Campos Gerais of Paraná. The ABC Foundation, a farmer cooperative branch was the institution responsible for the system. The test enclosed 38 cities and a network of ten automatic meteorological stations. The platform was projected to be accessed of several forms, however in this test was used processing in batch (server side) and Web interface (client side). The platform revealed to functionally efficient and potentially applicable, resulting in a tool of great utility for use in decision aid. As result in 2008 and 2009, it was expanded to other wheat growing regions in the country. The approach used also was exported to other crops, as for example, apples.