A literature review: rainfall thresholds as flash flood monitoring for an early warning system

Flash floods, increasingly common and impactful, can be better managed through effective early warning systems. This study reviews methods for determining rainfall thresholds using empirical, hydrological, and machine learning approaches. Empirical methods, based on historical data, identify patterns between rainfall and floods but struggle with spatial and temporal variations. Hydrological methods, which consider watershed characteristics, offer greater accuracy but require detailed data. Machine learning, with its capacity for real-time, adaptive analysis of big data, shows promise for improving predictions. Integrating these approaches can enhance early warning systems, making them more effective in the face of climate change and significantly reducing the impact of flash floods.