Identification of Precipitation Thresholds for Flooding in Basins With Scarce Data: Case Study for José Pereira River, Itajubá/MG

Abstract Extreme events, such as flash floods, have increased globally, causing significant damage. In Brazil, this concern is worsened by climate change, with characteristics such as La Niña in 2022 and El Niño until 2024. Hydrological modeling with geoprocessing is crucial for risk management, requiring precise knowledge about limits on rainfall quantity and duration. This study aims to establish duration limits that trigger floods in the José Pereira river basin, a very representative small river basin in most cities of Brazil. The process begins with the characterization of the basin, considering climate, flood history, topography, soils, land use and subdivision into sub-basins. Parameters such as curve coefficient (CN), concentration time and lag time are calculated for each sub-basin. After characterization, we selected two flood events between 2022 and 2023 to calibrate the HEC-HMS model, using level and specialization data. Adjusted basin data for greater realism. We chose simulated events based on maximum rainfall and return time, using records from the 2013 to 2023 historical series rainfall station, decreasing 40 mm in 24 hours. These events showed technical and distributed rainfall, highlighting the influence of upstream basins on the variation in river levels during flood events. A detailed analysis revealed that the significant occurrence occurs between the first 25% and 50% of the total duration of rainfall caused by overflows or floods. From the simulations, we identified the minimum accumulation of rain necessary to exceed 1.5m in the river. The threshold curve highlighted that, upon reaching 34.60 mm of accumulation, it is crucial to issue an alert, as exceeding this value and the curve, the river is susceptible to overflowing at any time.