In this study, with the aim of improving the accuracy of daily runoff simulation in the Ghareso basin located in Kermanshah province, a hybrid modeling framework was developed based on the integration of the IHACRES conceptual model and advanced machine learning algorithms. The data used included precipitation, minimum, maximum, and average temperatures, and river discharge from the Pol-e-Kohne and Qorbaghestan hydrometric stations during the period 1995 to 2023. In the first stage, the IHACRES semi-distributed model was implemented to simulate the precipitation-runoff process and its parameters were optimized using a genetic algorithm. In order to remove short-term fluctuations, a three-day moving average filter was applied to the model output. Then, using the IHACRES output and a set of derived variables including lag features, precipitation and temperature statistics, time and drought indices, an XGBoost type machine learning model was designed. The performance of the models was evaluated with the statistical indices RMSE and NSE in two training and testing periods. The results showed that the basic IHACRES model had an acceptable performance (NSE≈0.44), but by applying the moving average filter and optimizing the parameters, its accuracy increased by an average of 30%. Finally, the combined IHACRES–XGBoost model provided the highest accuracy by achieving an NSE value of more than 0.97 and an RMSE of less than 10. These findings indicate the high efficiency and high potential of the combined models in improving daily runoff forecasting.
bayatikhatibi, M. and kakapor, V. (2025). Improving the hybrid IHACRES–XGBoost model for simulating daily runoff in the Gharesu Basin. Journal of Hydrogeomorphology, (), -. doi: 10.22034/hyd.2025.68164.1800
MLA
bayatikhatibi, M. , and kakapor, V. . "Improving the hybrid IHACRES–XGBoost model for simulating daily runoff in the Gharesu Basin", Journal of Hydrogeomorphology, , , 2025, -. doi: 10.22034/hyd.2025.68164.1800
HARVARD
bayatikhatibi, M., kakapor, V. (2025). 'Improving the hybrid IHACRES–XGBoost model for simulating daily runoff in the Gharesu Basin', Journal of Hydrogeomorphology, (), pp. -. doi: 10.22034/hyd.2025.68164.1800
CHICAGO
M. bayatikhatibi and V. kakapor, "Improving the hybrid IHACRES–XGBoost model for simulating daily runoff in the Gharesu Basin," Journal of Hydrogeomorphology, (2025): -, doi: 10.22034/hyd.2025.68164.1800
VANCOUVER
bayatikhatibi, M., kakapor, V. Improving the hybrid IHACRES–XGBoost model for simulating daily runoff in the Gharesu Basin. Journal of Hydrogeomorphology, 2025; (): -. doi: 10.22034/hyd.2025.68164.1800
)