Estimating the water balance of the Kowsar Dam watershed using the SWAT hydrological model and satellite data

Estimating the water balance is a key factor in water resource management and environmental planning. The water balance refers to the equilibrium between water inputs and outputs in a specific area, including factors such as precipitation, evaporation, runoff, and water withdrawal. The most important elements of the water balance from the perspective of water resource management are precipitation, evapotranspiration, surface runoff, groundwater flow, and lateral flow. A correct understanding of this balance can help optimize water resource use, predict climate change, and assess human impacts on ecosystems. Given the increasing population and the growing need for water, accurate estimation of the water balance has become crucial for informed management decisions and sustainable development in the agricultural, industrial, and urban sectors. Therefore, dynamic and coherent planning, along with the implementation of appropriate management and conservation measures, is essential for the optimal use of the country's water and soil resources. In this context, employing mathematical models and field data can enhance the analysis of water resource status and future planning. Today, hydrological models are utilized to study and plan for the sustainable and effective comprehensive management of watersheds. An example of a physically based hydrological model is SWAT, which simulates large-scale processes and monitors them based on the characteristics of the watershed and its climatic conditions.

Materials and Methods

In this study, to model the hydrological conditions of the Kowsar Dam basin using the SWAT model, we first imported a digital elevation model with a resolution of 30 meters into the model software environment (ArcSWAT). The output location was then specified, and the watershed boundary was established. Next, we overlaid land use, soil, and slope class maps to obtain hydrological response units (HRUs) for the region. At this stage, the basin was divided into 35 sub-basins and 184 HRUs. To run the model, we utilized daily climatic data from the meteorological stations, which included precipitation, maximum and minimum temperatures, and relative humidity. The model was calibrated using the SUFI2 program, based on the data from 2007 to 2019. Initially, to identify the parameters influencing runoff in the region, a sensitivity analysis was conducted using the One Parameter at a Time (OAT) method, which helped identify the sensitive parameters for model calibration. By implementing the SUFI2 algorithm, we determined the optimal values for these sensitive parameters. Model validation was carried out using the modified parameter values obtained during the calibration stage. To evaluate model performance during both the calibration and validation phases, we used the coefficient of determination (R²) and the Nash-Sutcliffe coefficient (NS).

Results and Discussion

The results showed that the SWAT model simulated the water balance components of the Kowsar Dam watershed with acceptable accuracy. For this reason, the values of the R² and NSE indices were relatively high. Based on the results, it was also determined that in the studied basin, most of the precipitation occurred in the fall and winter seasons, with the maximum occurring in November (Aban) and the lowest precipitation in June (Khordad). The amount of surface runoff in November (Aban) gradually begins with the onset of autumn and winter precipitation, so that the highest amounts of surface runoff were observed in November (Aban) and February (Bahman). The temporal changes in base flow throughout the year showed that its highest amount was related to late winter and its lowest amount was related to October (Mehr). Actual evapotranspiration gradually increases from November (Aban) with the onset of precipitation, so that from late winter, as the weather warms up, the actual evapotranspiration rate increases and reaches its maximum in May (Ardibehesht). In relation to potential evapotranspiration, unlike actual evapotranspiration, this parameter will increase with increasing temperature and decreasing precipitation. The lowest potential evapotranspiration rate is in January (Day) due to the sharp decrease in temperature in this month, and the highest rate is in early summer, that is, July (Tir).

Conclusion

By implementing the SWAT model in the Kowsar Dam basin, we were able to simulate the monthly flow for the studied period. Statistical comparisons of this modeling demonstrated acceptable results. The comparison of the simulated and observed hydrographs showed a strong correlation according to the Nash-Sutcliffe criterion. Therefore, we can conclude that the SWAT physical model performs acceptably in the Kowsar Dam basin, based on the simulation results. By comparing the appearance and statistics of the observed hydrograph with those of the simulated hydrograph, we found a high similarity between the two during the study period. There is good agreement between the hydrographs regarding important characteristics such as peak discharge, runoff volume, and time to reach peak discharge. Overall, the results indicate that the SWAT model has the ability and acceptable accuracy to simulate the monthly runoff discharge of the Kowsar Dam watershed. In this study, the model's calibration and validation results showed its efficiency in estimating the water balance in the Kowsar Dam watershed. The final results showed that on average, about 51 percent of precipitation enters the atmosphere as evaporation and transpiration, about 21 percent as surface runoff, 5 percent as lateral flow, and 16 percent as return flow directly into waterways. In total, about 26 percent of water enters the soil layers and aquifer. The results indicate the effectiveness of the SWAT model in simulating the water balance of the Kowsar Dam watershed.