Introduction:
In many watersheds, particularly in semi-arid regions, the lack of long-term hydrometric data and the complexity of hydrological ecosystems pose major challenges for analyzing basin behavior. The difficulties of obtaining accurate and sufficient hydrological data in these regions necessitate the use of conceptual hydrological models that can simulate runoff processes even with limited or incomplete data. Among such models, IHACRES has gained wide recognition for its simple structure, low data requirements, and reliable performance in diverse climatic conditions. Numerous studies worldwide have demonstrated its effectiveness in simulating rainfall-runoff processes at various temporal scales and under climate change scenarios. Despite IHACRES has been successfully applied in humid, semi-arid, and data-scarce catchments, its performance has not yet been evaluated in the Seyedabad watershed, located in a semi-arid region of Kohgiluyeh and Boyer-Ahmad Province, Iran. This study aimed to calibrate and validate the IHACRES model at daily and monthly scales for this understudied basin. By addressing this research gap, the current study contributed to the localization of hydrological modeling approaches in semi-arid environments of Iran.

Material and Methods:
This study was conducted in the Seyedabad watershed, located in a semi-arid region of Kohgiluyeh and Boyer-Ahmad Province in southwestern Iran. The watershed covers an area of approximately 795 km² and includes the Cham Siah river, a tributary of the Kheirabad river. The basin lies between 30° to 41° N latitude and 43° to 50° E longitude. The Seyedabad watershed is crucial for providing water for drinking, agriculture, and ecological purposes in the region. The semi-arid climate leads to significant variations in rainfall and runoff, making accurate hydrological modeling essential for managing the water resources effectively. Daily and monthly climate data (including precipitation and temperature) were obtained from the Seyedabad meteorological station for the period 2007 to 2017. Streamflow data corresponding to this period were sourced from the Cham Siah hydrometric station. For modeling purposes, the dataset was divided into two segments: calibration and validation. Daily data from 2010 to 2017 and monthly data from March 2008 to August 2013 were used for calibration, while the remaining years were allocated for model validation. To simulate rainfall-runoff processes, the IHACRES hydrological model (version 2.1) was utilized. This model integrates a non-linear module to convert rainfall into effective rainfall and a linear module for streamflow routing. Calibration of model parameters was conducted using a manual trial-and-error approach to optimize performance. Key parameters included the drying rate coefficient (f), temperature modulation coefficient (c), and the catchment drying time constant (τw), along with routing parameters such as τq (quick flow recession constant), τs (slow flow recession constant), and sv (scaling factor). The model’s performance was evaluated using multiple statistical metrics: Nash–Sutcliffe Efficiency (NSE), Coefficient of Determination (R²), Mean Squared Error (MSE), and Root Mean Square Error (RMSE). These indices were calculated separately for daily and monthly time steps to assess the model’s predictive capabilities in varying temporal resolutions. The approach enabled an in-depth understanding of IHACRES performance in semi-arid, data-scarce environments.

Results and Discussion: 
The IHACRES model was successfully calibrated and validated for the Seyedabad watershed at both daily and monthly time scales. The model's performance indicators revealed acceptable accuracy in simulating streamflow, especially considering the semi-arid nature of the basin and limited data availability. At the daily time scale, the model showed moderate accuracy, with a Nash-Sutcliffe Efficiency (NSE) of 0.51 during the calibration period and 0.47 during the validation period. The Coefficient of Determination (R²) values were 0.55 and 0.49, respectively, indicating a fair fit between observed and simulated streamflow. The Root Mean Square Error (RMSE) values were 1.26 m³/s for calibration and 1.38 m³/s for validation, suggesting relatively low error levels. These results indicated that while the model successfully captured the daily dynamics of streamflow, the accuracy was limited by the complexity of the runoff processes and the natural variability of precipitation and temperature in the semi-arid environment. The performance of the model improved significantly at the monthly time scale. The Nash-Sutcliffe Efficiency (NSE) values reached 0.71 during the calibration phase and 0.65 during validation, reflecting better accuracy in simulating the runoff processes over longer time periods. The corresponding Coefficient of Determination (R²) values were 0.72 and 0.67, and the RMSE values were 0.94 m³/s and 1.09 m³/s, respectively. These improved results at the monthly scale suggested that the IHACRES model was more effective at simulating aggregated flow patterns over extended periods, making it particularly suitable for long-term water resource assessments in semi-arid regions. The improved model performance at the monthly level could be attributed to the smoothing effects of aggregation, which reduces the sensitivity to short-term variability in runoff. Graphical comparisons between observed and simulated flows confirmed the model’s ability to reproduce seasonal variations in streamflow and to predict peak discharge events accurately. However, some discrepancies were observed, particularly during low-flow conditions and high-intensity rainfall periods. Despite these differences, the overall performance of the model in simulating streamflow behavior in the Seyedabad watershed was acceptable. These findings demonstrate that the IHACRES model, with its conceptual structure and low data requirements, can be successfully applied to semi-arid catchments with limited hydrological data.


Conclusion:
This study evaluated the performance of the conceptual IHACRES model in simulating the streamflow of the Cham Siah river, located in the Kheyrabad watershed of Kohgiluyeh and Boyer-Ahmad Province, Iran. Daily and monthly rainfall, temperature, and discharge data from the Seyedabad station (2007–2017) were used as model inputs. The results of the calibration and validation phases demonstrated that IHACRES could simulate the hydrological behavior of the Seyedabad watershed with reasonable accuracy. The model performed better at the monthly scale, likely due to the natural smoothing of data and reduced sensitivity to short-term variations in streamflow. The model’s performance at the daily scale was also acceptable, with moderate accuracy in simulating the runoff dynamics of the watershed. Given its simplicity, low data requirements, and efficient performance in semi-arid conditions, IHACRES is a suitable tool for simulating and predicting runoff in data-scarce basins. Its ability to simulate long-term water flow patterns makes it particularly useful for water resource management in semi-arid regions like Iran. The model’s application in combination with climate change scenarios can provide valuable insights into the potential impacts of future climatic changes on runoff and water availability. Further research is recommended to integrate IHACRES with other models, such as SWAT or Artificial Neural Networks (ANN), to enhance its predictive capabilities and improve performance under different environmental conditions.