Journal of Watershed Management Research

Extended Abstract
Background: Water resource management in semi-arid regions faces major challenges due to environmental conditions, such as flash floods and intense surface runoff. These areas are prone to destructive flooding because of high-intensity convective rainfall, uneven precipitation distribution, high evapotranspiration, sparse vegetation, and fragile soils. Accurate modeling of hydrological processes—particularly peak discharge and runoff volume—is essential for designing effective management measures, such as flood control, runoff regulation, and vegetation restoration. Hydrological models, especially HEC-HMS, are widely used for simulating rainfall–runoff processes due to their ability to integrate climatic, hydrometric, and watershed characteristics. Parameter sensitivity analysis further helps identify influential factors, reduce uncertainty, and improve prediction accuracy. In Iran’s semi-arid regions, flash floods frequently cause severe damage. This study examines how vegetation restoration influences the sensitivity of HEC-HMS parameters in two adjacent sub-watersheds of the Gonbad paired watershed in Hamadan Province—one control (without restoration) and one treated (with restoration). The main goal is to analyze how restoration affects parameters controlling peak discharge and runoff volume. The study involves simulating rainfall–runoff processes, calibrating and validating the model with observed data, and performing sensitivity analysis on key parameters, such as curve number (CN), maximum retention (S), lag time, channel slope, and Manning’s roughness coefficient. By comparing results between the two basins, the research evaluates how vegetation restoration alters model performance and parameter responsiveness, providing insights for improved water resource management and flood mitigation in semi-arid environments.
Method: This study was conducted in the paired watershed of Gonbad, located in Hamadan Province. Two adjacent sub-watersheds with similar physiographic and climatic characteristics were selected, one as a control watershed (without any restoration interventions) and the other as a restored watershed (subjected to vegetation restoration activities). The restoration operations in the restored watershed included planting drought-resistant plant species and implementing watershed management practices. The HEC-HMS hydrological model was employed to simulate the rainfall–runoff process. The data required for model simulation included meteorological data (precipitation, temperature, etc.), hydrometric data (streamflow), soil characteristics, and watershed physical attributes. Ten years of observed rainfall and runoff data were used for model calibration and validation, and various statistical indicators, including the Nash–Sutcliffe Efficiency (NSE) coefficient and the Root Mean Square Error (RMSE), were applied to evaluate model performance. Sensitivity analysis of the model parameters was carried out using a percentage variation method (±30%) for key parameters, such as the Curve Number (CN), maximum retention (S), lag time, channel slope, and Manning’s roughness coefficient.
Results: The vegetation restoration significantly affected soil properties, runoff simulation accuracy, and parameter sensitivity in the studied watersheds.
Changes in soil properties: Measurements indicated that both initial and final soil infiltration rates were significantly higher in the restored watershed than in the control watershed. Specifically, initial infiltration rates increased from a range of 3.82 to 4.76 mm/h in the control watershed to a range of 130.2 to 144.4 mm/h in the restored watershed. Similarly, final infiltration rates rose from 4.2 to 6.8 mm/h in the control watershed to 10.8 to 15.4 mm/h in the restored watershed. Furthermore, the restored watershed showed approximately 20% higher silt content, an average of 20% lower bulk density, and about 35% higher soil porosity than the control watershed. Soil saturation moisture and field capacity also increased in the restored watershed by approximately 30% and over 10%, respectively. These changes indicate an overall improvement in soil quality as a result of the restoration activities.
Model simulation results: The HEC-HMS model demonstrated acceptable performance in simulating watershed outflows, with NSE values exceeding 0.6 and RMSE values below 0.5. The match between observed and simulated hydrographs was stronger in the restored watershed, particularly at the peak discharge segment. Simulation accuracy in the restored watershed was higher than in the control watershed, consistent with findings from previous studies. Deviations in the hydrograph, especially during peak and recession phases, were more pronounced in the control watershed, likely due to insufficient vegetation cover and increased surface runoff.
Sensitivity analysis results: The sensitivity analysis showed that the maximum retention (S) and Curve Number (CN) parameters had the greatest influence on peak discharge. Specifically, a 30% increase in CN led to an approximate 25% increase in peak discharge in the control watershed, while this increase was about 15% in the restored watershed. Likewise, a 30% increase in maximum retention resulted in a reduction of runoff volume by nearly 35% in the control watershed, whereas the reduction was limited to around 20% in the restored watershed. These findings suggest that the model’s sensitivity to these parameters was lower in the restored watershed, which could be attributed to improved soil infiltration, greater water storage capacity, and enhanced hydrological resilience. Other parameters, such as channel slope, lag time, and Manning’s roughness coefficient, also affected peak discharge; however, their impact diminished at higher levels of variation. For example, a 30% increase in Manning’s coefficient in the control watershed resulted in only about a 5% reduction in peak discharge.
Conclusion: The results of this study indicate that vegetation restoration can have a substantial impact on improving soil properties, enhancing runoff simulation accuracy, and reducing the sensitivity of hydrological models to parameter variations. Improved soil infiltration capacity and water storage potential in the restored watershed played a key role in reducing surface runoff and increasing the reliability of hydrological predictions. These findings highlight the importance of vegetation restoration as a practical and effective strategy for watershed management. They can be applied to the design and implementation of water resources management and watershed rehabilitation programs in similar semi-arid regions.