Locating suitable sites for the construction of underground dams using the subsurface flow simulation (SWAT model) and analytical network process (ANP) (case study: Daroongar watershed, Iran)

Abstract

Underground dams are constructions that prevent the natural flow of ground water and create underground water resources. Although lateral-flow volume is one of the main criteria to locate an underground dam, this factor is ignored in the most studies. In this research, the water balance equation was simulated using the SWAT model in the Daroongar watershed (northeastern Iran), and the paths of subsurface water flows were identified. Continuing by giving the weight to an 11-layer analytical network process (ANP), potential suitable locations for the construction of underground dams were determined. With the implementation of the SWAT model in the Daroongar watershed, R2, bR2, and NS (Nash Satklyf) coefficients were 0.77, 0.75, and 0.68, respectively, in the calibration phase; in the validation phase, they were 0.71, 0.67, and 0.61 in the same order. All these coefficients show that the efficiency of the SWAT model in simulating water balance is acceptable. Results of the sensitivity analysis of the SWAT model show 27 effective parameters, where the runoff curve number (CN2) was identified as the most important parameter. The results of the ANP model prove that subsurface flow, weighing 0.131, includes the most weight allocated. Results from the ANP model indicated that this is necessary for determining the extent of subsurface flow using the SWAT model. In addition, the reservoir volume and runoff volume, respectively, with a weight of 0.109 and 0.101 are more important than other factors. As a result, the greater the current subsurface stream, the more important it is in relation with other streams. The results showed that streams with ranks 3 and 4, which are located on slopes less than 15%, are suitable for the construction of underground dams.