Flood prediction in Brazil: A review of trends, gaps and methodological perspectives

Abstract

Flood prediction is crucial in mitigating the adverse effects of extreme hydrometeorological events, as climate change and urban sprawl intensify their frequency and severity. This review systematically synthesizes scientific literature published between 2012 and 2024, mapping methodological trends, regional disparities, and model typologies in Brazilian flood forecasting. The analysis reveals a concentration of studies in the South and Southeast, where robust data infrastructures support the adoption of physical, statistical, machine learning, and hybrid models. Conversely, the North, Northeast, and Midwest regions remain critically underrepresented. Methodological gaps include a predominance of short-term forecasts and the underutilization of climatic and urban variables in long-term modeling. To address these challenges, we advocate for expanding real-time monitoring networks, integrating multisource data, and promoting hybrid and AI-enhanced forecasting techniques. These actions are essential for fostering climate resilience and equitable hydrological risk management across Brazil.