Black Sigatoka (Mycosphaerella fijiensis) threatens the banana trade in most of world's producing areas, damaging the production and bringing significant financial loss. It is necessary to study the plant's susceptibility in the various developmental stages and climatic conditions that favor the disease occurrence. This paper presents a probabilistic model based on polynomial functions for estimating the risk of Black Sigatoka occurrence. A case study was developed in a commercial banana plantation located in Jacupiranga, Vale do Ribeira, SP, Brazil, considering the weekly monitoring of the disease's state evolution (EE), time series of meteorological data and remote sensing data. Georeferenced risk maps were prepared for different dates. We obtained a model to estimate the evolution of the disease from satellite imagery, with a coefficient of determination of 0.9. This methodology was developed for the detection of times and locations that have favorable conditions to the occurrence of Black Sigatoka and can be applied, with appropriate adjustments for different locations, to assess the risk of disease occurrence in other areas of banana production.
Mychosphaerella fijiensis; forecasting; modeling
