Abstract—

The objective of this paper is to test the feasibility of using the SWAT model under limited data availability to estimate soil erosion in two adjacent watersheds, namely Mazer (gauged) and El Himer (ungauged) watersheds. In this study, we used the physical proximity approach as one of the regionalization methods while ensuring the similarity between both watersheds. Moreover, the comparison of the most important characteristics that influence runoff production shows that Mazer and El Himer watersheds are nearly similar in terms of soil, land use and all other morphological and physical characteristics. In aiming to achieve the objectives set out in this study, SWAT Model was calibrated and validated on a monthly time step at Mazer watershed using SWAT-CUP (SUFI-2). Results showed a good correlation between the observed and simulated streamflow with a NSE (Nash–Sutcliffe Efficiency) of 0.65, 0.89 and R² of 0.75, 0.95 for calibration and validation, respectively. After the calibration and validation processes in Mazer watersheds, the fitted values for the most sensitive parameters have been applied at El Himer watershed and both models were executed for 5 years to estimate streamflow and soil erosion at Mazer and El Himer watershed. The results showed that all studied subwatersheds present a weak amount of soil erosion rate, with a maximum of 5.20 t/ha/year. Generally, soil erosion in El Himer is slightly high. The average annual values recorded of sediment yield at Mazer and El Himer were 725 and 2991 tons/year, respectively. Moreover, the results obtained can be used in other watersheds with the same characteristics. The interest shown in this type of study stems from our major problem inherent in the developing country, where very few measuring stations are available. It is hoped to demonstrate the utility of regionalization and to encourage modelers to work more on ungauged watersheds.