Enhancing Flood Forecasting Resilience with Interpretable Models ARIMA-iGARCH for Discharge Dynamics and Fuzzy Logic for Peak Prediction

A complementary modeling framework: ARIMA-iGARCH: captures linear structure + volatility in variance, providing highly accurate modeling of overall discharge fluctuations Fuzzy Logic: captures nonlinear patterns and uncertainty, with a strong focus on precise peak detection Rather than competing, the models address different aspects of flood dynamics: one ensures temporal accuracy and variance representation; the other enhances extreme event (peak) prediction Both models are applied to short-term hourly discharge data from Schwarzbach River (Hesse), Nauheim gauge, demonstrating their combined potential for robust and interpretable flood forecasting.