**3. Results**

After the preparation of requisite input file, including different selected parameters, the model was applied to all selected sites for the evaluation of the effectiveness of soil conservation structures. For this purpose, the model was first run without soil conservation structures, and then conversation structures were modeled to see their effectiveness. The model was also applied for the above-mentioned four different scenarios related to field practices being adopted by the farmers in the area.

The modeled period was from 2009 to 2011 for Catchment-25. The runoff and sediment yield data collected during 2009–2010 were used for model calibration, while the 2011 data were used for validation. Some of the appropriate parameters were adjusted (Table 2) until the predicted runoff and sediment yield approximately matched the measured ones at the outlet (Figure 6). To determine the most sensitive parameters for model calibration, the sensitivity analysis was performed in the ArcSWAT interface using five parameters for sediment yield (Table 2): USLE practice factor (PUSLE), USLE conservation practice factor (CUSLE), USLE soil erodibility factor (KUSLE), the linear parameter for calculating the maximum amount of sediment that can be re-entrained during channel sediment routing (SPCON), and the exponent parameter for calculating sediment re-entrained in channel sediment routing (SPEXP). The PUSLE factor was found to be the most sensitive parameter during model calibration using sensitivity analysis. Moreover, the obvious correspondence (coefficient of determination (R2) = 0.80 and Nash–Sutcliffe efficiency (NSE) = 0.70) of the hydrographs of the observed and simulated surface runoff and sediment yield indicated that the SWAT is capable of simulating the hydrological regime of small watersheds in the Pothwar region (Figure 6).

## *3.1. Model Application without Conservation Structures*

After separately setting up the SWAT model for each watershed, the model simulation was performed with the default set of parameters in the default setting. Then, the soil erosion parameters (Table 2) were used for sediment yield simulation in each watershed. The modeled period was from 2010 to 2015. We estimated that all the watersheds generated a maximum sediment yield in 2010, while a minimum sediment yield was simulated in 2012. This indicated that the sediment yield is a direct function of runoff and rainfall intensity. In 2010, Khaliq Gully model estimation was 59.3 t ha<sup>−</sup>1, while in 2012, it was 2.3 t ha<sup>−</sup>1. Similarly, the Ashraf Gully, Khokar Bala, Chak Khushi, Dhoke Dhamal, Dhoke Hafiz Abad, and Khandoya watershed models produced annual sediment yields of 25, 37.6, 1.6, 15.3, 32.3, and 45.9 t ha<sup>−</sup>1, respectively, in 2010 (Table 3).

## *3.2. Model Application with Conservation Structures*

After the model application without conservation structures with calibrated soil erosion parameters, the model was applied to the small watersheds using soil and water conservation structures. The model setting was done in accordance with the location of conservation structures for the correct delineations of sub-basins. The intervention of the soil and water conservation structures was made by modifying the surface runoff and sediment yield parameters, as given in Table 2. The SWAT provides various options to consider soil and water conservation structure impacts [59] including: (i) surface runoff may be modified through the adjustment of the runoff ratio (curve number) and/or the consideration of a micro-pond (pothole) at the related HRU level, which also impacts the soil erosion, and (ii) impacts on the sediment yield levels may be modified via the adjustment of the support P-factor and/or the slope length and steepness factor (LS) of the MUSLE [60]. The ideal factors that describe the effect of stone bunds are the USLE support P-factor, the curve number, and the SLSUBBSN.

128

**Table 2.** Soil erosion parameters used during the model's application without conservation structures (pre-condition) also used for Catchment-25 calibration. Post-condition parameters represent the conservation structures. SPEXP: exponential channel sediment routing factor; SPCON: linear channel sediment routing factor; USLE\_C: conservation practice factor; USLE\_K: soil erodibility factor; SLSUBBSN: the average slope length for the sub-basin; HRU\_SLP: average slope steepness of the hydrological response unit (HRU) input tables; CN: curve number.


These small watersheds already have existing soil and water conservation structures for the control of soil erosion. The crests of the structures play a major role in reducing the flow velocity and sediment deposition (erosion reduction) due to ponding upstream of the structures, whereas the downstream sections of the structures prevent channel or gully development. The topography of the region consists of permanent gullies where farmers use these gullies for the cultivation of crops. The farmers manage the gullies in a terraced land use system by making field boundary bunds, as shown in Figure 7 for the example of the Khokar Bala site. During the monsoon season, heavy rainstorms cause the shear failure of terrace edges (field bunds) due to the heavy surface runoff. This problem creates a loss of soil and damage to the crops. To reduce this problem, soil and water conservation structures have been installed to retain water in the terrace up to a certain rainfall amount (without overflowing the terrace) and then to divert the excess rainfall in a non-erosive way. These structures appear as a type of stone bund.

**Figure 7.** Permanent gullies for the cultivation of crops; an example of Khokar Bala site.

#### *3.3. Soil Erosion Estimation and E*ff*ect of Conservation Structures*

The sediment yield results were compared under each condition, as shown in Table 3, by modifying the SWAT parameters representing the conservation structures. The six parameters were modified according to the slope characteristics of the small watersheds and field conditions, in addition to being modified according to the terraced and contoured section of the SWAT user's manual [59] and a literature review [61–64]. Soil and water conservation structures, such as stone bunds, act as vital measures in the reduction of flow velocity, surface runoff, soil erosion, and slope length in a watershed system [65]. Suitable parameters that signify the effect and importance of loose stone structures are the SLSUBBSN, land management practice parameter (USLE\_P), and the CN2 for rainfall–runoff conversion [61].

The impact of stone bund soil and water conservation structures was simulated through the reduction of the CN2 for surface runoff ratio modification, as well as the adjustment of the P-factor to account for trapped sediments at the stone bunds. Table 3 presents a significant sediment yield reduction achieved by incorporating the parameter values recommended for stone structures. The average annual sediment yield reduction varied from 40% to 98%; the Khokar Bala site showed the maximum reduction. The average five-year sediment yield reduction engendered by structures at various sites varied from 54% to 98%, and these results are relatively comparable to the findings of various studies [61,63,66].

Betrie et al. indicated that 6–69% sediment reductions in the Upper Blue Nile River basin were caused by stone bunds [61]. A field-scale study in the northern part of Ethiopia by Gebremichael et al. indicated a 68% sediment yield reduction was engendered by stone bunds [66]. In addition, Herweg and Ludi conducted a study at plot scale in the Eritrean highlands and Ethiopia, and they reported 72–100% sediment yield reductions engendered by stone bunds [63]. Based on the plot experiments carried out in 2013, stone bund structures were found to reduce surface runoff by approximately 60–80% and sediment yield between 40% and 80% [67]. This is consistent with other plot experimental findings reported by Adimassu et al., where stone bunds were found to reduce the sediment yield by roughly 50% compared to untreated plots [68]. The effect of conservation structures on sediment yield reduction was elucidated by Oweis and Ashraf in the Dhrabi watershed, and it was found that the average soil loss rates in 2009 without and with structures were calculated were 47 and 37.98 t ha−<sup>1</sup> year<sup>−</sup>1, respectively, with a 20% reduction. However, the maximum soil loss rates without and with structures were 2716.17 and 1731 t ha−<sup>1</sup> year<sup>−</sup>1, respectively, with a 37% reduction [69].

The large variation in sediment reduction with conservation structures was observed due to the watershed topography and the numbers of soil and water conservation structures. For example, the Khokar Bala site showed the maximum 98% reduction because this site has a 90% area at a 0–10% slope (Table 3) and a total of 13 soil and water conservation structures. Based on the field observation findings: (i) The conservation structures require regular maintenance because non-meshing can cause stones to slide, which may lead to the displacement of the whole structure, and (ii) the structures were not designed according to the hydraulic characteristics of the surface flow. Downstream damage of the structures was common due to the non-availability of downstream energy dissipation arrangements.


**Table 3.** The effect of stone structures on the sediment yield reduction.
