Comparison Of The Performance Of HEC-HMS And RF Models Models In Rainfall Runoff Simulation

Abstract

River discharge prediction is a critical subject in water resource engineering concerns. The HEC-HMS model and random forest (RF) models have been successfully used in discharge forecasting problems. In this research HEC-HMS model and RF were applied for forecast daily discharge for M.H. Halli gauge Station which lies in Karnataka state region in India. In order to calibration and validation of the models, the recorded daily rainfall, temperature and discharge which available for fifteen years (2003–2017) was divided into two sets. The output of the HEC-HMS model was compared with the recorded data of the discharge, the residual time series values, which resulted from the comparison, was solved by using the RF approach. The root means square error (RMSE), mean absolute error (MAE), and coefficient of determination (R2) for M.H. Halli station using HEC-HMS model were found to be more efficient than RF model in forecasting the daily discharge. The current study's findings supported the proposed methodology's applicability for precise discharge modeling.