Water Discharge Peak Estimation Based on HEC-HMS and Predicted Curve Numbers for Flood Forecast in the River Brembo (Northern Italy)

Abstract

This paper proposes a novel methodology for peak flow estimation. This methodology uses single-event hydrological modeling based on the software HEC-HMS and curve numbers (CNs) estimated as a function of antecedent and current weather variables and is applied to the river Brembo case study in Northern Italy. By using rainfall, weather, and water discharge data collected over an eleven-year-long period, from 2013 to 2023, HEC-HMS is first used to optimize the CN values at two cross sections in the Brembo basin, in an attempt to reproduce the flood peak in numerous single rain events. Then, regression equations are constructed to express CN as a function of current event rainfall depth and antecedent rainfall depth and temperature, as explicative variables for current soil conditions. The good predictive performance of HEC-HMS based on CN values estimated through the regression equations (for the peak flow at the two cross sections, a mean absolute percentage error [MAPE] of 0.26 and 0.29, respectively, in calibration, and 0.33 and 0.45, respectively, in validation; and an index of agreement [d] of 0.84 and 0.92, respectively, in calibration, and 0.86 and 0.88, respectively, in validation) makes the modeling tool constructed in the paper efficient and effective for potential early-warning applications.