FIRE RISK ALERT SYSTEM IN THE PANTANAL WITH FORECASTS UP TO 3 DAYS IN THE FUTURE

Abstract

In the Pantanal, located in the states of Mato Grosso do Sul and Mato Grosso, several fires occur every year. Prolonged drought, dry organic matter, and sparks or fire can cause fires of varying magnitude. To aid in firefighting in the Pantanal, a web-based fire prediction system called Saripan was developed. This system previously predicted the fire risk for each municipality in the Pantanal region up to the present day. However, in its new version, the system can now predict fire risk up to three days in the future with a high probability of accuracy. This is thanks to a system that predicts climate data using a machine learning algorithm based on at least 20 years of daily historical climate data. With this data, the system learned how climate data varies over time for the municipalities in the Pantanal region. Thus, using predicted climate data and used to calculate fire risk, the Saripan system can predict fire risk for up to three days in the future in any municipality in the Pantanal. This paper describes this system, Saripan, which predicts fire risk for up to three days from the current day, with a good accuracy rate of around 90%.