Journal of Watershed Management Research

Introduction and Objectives:
Water resource management in semi-arid regions faces numerous challenges, including the occurrence of flash floods and excessive surface runoff. These regions, due to their unique climatic and geographical characteristics—such as high-intensity rainfall with uneven temporal and spatial distribution, high evapotranspiration, sparse vegetation cover, and vulnerable soils—exhibit a high potential for runoff generation and destructive floods. Accurate assessment and modeling of hydrological processes, particularly the prediction of peak discharge and runoff volume, are of paramount importance for designing and implementing effective management measures, including flood control structures, runoff management, and vegetation restoration. Hydrological models are powerful tools for simulating these processes, and the HEC-HMS model is one of the most widely used in this field. This model, with its ability to integrate climatic, hydrometric, and watershed data, enables precise simulation of rainfall-runoff processes. Sensitivity analysis of model parameters is an effective method for evaluating the impact of input parameter changes on model outputs and identifying key parameters. This analysis helps reduce uncertainties and enhance prediction accuracy. In semi-arid regions of Iran, flash floods have caused significant damage to infrastructure and natural resources. This study investigates the impact of vegetation restoration on the sensitivity of HEC-HMS model parameters in two adjacent watersheds in the Gonbad paired watersheds, Hamedan—one serving as a control watershed (without restoration practices) and the other as a restored watershed (with vegetation restoration practices). The study aims to improve the understanding of hydrological processes and provide solutions for optimal water resource management in similar regions. The primary objective of this study is to analyze the sensitivity of parameters affecting peak discharge and runoff volume in two adjacent watersheds in the Gonbad paired watersheds, Hamedan, using the HEC-HMS model. The study evaluates the impact of vegetation restoration practices on improving runoff simulation and reducing the sensitivity of hydrological parameters. Specific objectives include simulating the rainfall-runoff process in both watersheds, calibrating and validating the model, analyzing the sensitivity of key parameters such as Curve Number (CN), maximum storage (S), lag time, channel slope, and Manning’s roughness coefficient, and comparing the simulation and sensitivity analysis results between the two watersheds to assess the effect of restoration practices on model accuracy and parameter sensitivity. The study seeks to provide practical outcomes for designing water resource management programs and mitigating flood impacts in semi-arid regions.
Materials and Methods: This research was conducted in the Gonbad Hamedan paired watershed. Two adjacent watersheds with similar characteristics, one as a control watershed and the other as a restored watershed, were selected. Vegetation restoration operations in the restored watershed included planting drought-resistant plant species and implementing watershed management practices. The HEC-HMS hydrological model was used to simulate the rainfall-runoff process. The required data for the model included meteorological data, hydrometric data, soil characteristics, and watershed characteristics. Ten years of rainfall and discharge data were used for model calibration and validation, and various statistical criteria, including the Nash-Sutcliffe Efficiency (NSE) and Root Mean Square Error (RMSE), were used. For parameter sensitivity analysis, the percentage change method (±30%) was used for parameters such as Curve Number (CN), maximum storage (S), lag time, channel slope, and Manning's roughness coefficient.
Findings: The results of this research showed that vegetation restoration operations have had a significant impact on soil characteristics, runoff simulation, and parameter sensitivity.
Changes in Soil Characteristics: The initial and final soil infiltration rates in the restored watershed significantly increased (initial infiltration from 4.76-3.82 to 130.2-144.4 mm/h and final infiltration from 2.4-8.6 to 10.8-15.4 mm/h). The percentage of silt increased (about 20%), the bulk density decreased (on average 20%), and the soil porosity increased (about 35%). Soil saturation moisture and field capacity also increased in the restored watershed (saturation moisture about 30% and field capacity more than 10%). These changes indicate an improvement in soil quality due to restoration operations.
Model Simulation Results: The HEC-HMS model was able to simulate the outflow of the watersheds with acceptable accuracy (NSE above 0.6 and RMSE less than 0.5). The agreement between the simulated and observed data in the restored watershed, especially in the peak flow section, was greater. The simulation accuracy in the restored watershed was higher than in the control watershed, which is consistent with the findings of previous studies. Hydrograph deviations in the control watershed, especially in the peak flow and recession sections, were greater, which may be due to the lack of vegetation cover and increased surface runoff.
Sensitivity Analysis Results: The maximum storage and CN parameters had the greatest impact on peak discharge. Increasing the maximum storage caused a decrease, and increasing the CN caused an increase in peak discharge. Maximum storage had the greatest impact on runoff volume. The sensitivity of peak discharge and runoff volume to parameter changes was reduced in the restored watershed compared to the control watershed. For example, the effect of changes in maximum storage and CN was stronger in the control watershed. This indicates that restoration operations have reduced sensitivity to parameter changes, which is due to improved infiltration and water storage characteristics and increased hydrological resilience. Changes in channel slope, lag time, and Manning's roughness coefficient also had significant effects on peak discharge, but these effects gradually decreased with increasing percentage changes.
Conclusion: The results of this research show that vegetation restoration operations can have a significant impact on improving soil characteristics, increasing the accuracy of runoff simulation, and reducing the sensitivity of hydrological models to parameter changes. Improving infiltration and water storage capacity in the restored watershed plays an important role in reducing runoff and increasing prediction accuracy. These findings can be used in the design and implementation of water resource management and watershed restoration programs in similar regions.