Assessing the impacts of flow diversion terraces on stream water and sediment yields at a watershed level using SWAT model Qi Yang a, *, Fan-Rui Meng a , Zhengyong Zhao a , Thien Lien Chow b , Glenn Benoy c , Herb W. Rees b , Charles P.-A. Bourque a a Faculty of Forestry and Environmental Management at the University of New Brunswick, 28 Dineen Drive, Fredericton, NB E3B 6C2, Canada b Potato Research Centre, Agriculture and Agri-Food Canada, 850 Lincoln Road, Fredericton, NB E3B 4Z7, Canada c Environment Canada and Agriculture and Agri-Food Canada, Potato Research Centre, 850 Lincoln Road, Fredericton, NB E3B 4Z7, Canada 1. Introduction Soil erosion can lead to reduction of soil fertility, loss of nutrients, and declines of crop yields in farmlands. In a review of mechanized agricultural systems in which wheat, corn, soybean, and barley were planted, Bakker et al. (2004, 2005) found that on average, soil erosion reduced crop productivity by about 4% for each 10 cm of soil lost. Concerns about soil loss are particularly important in northwestern New Brunswick that has shallow soils (Langmaid et al., 1980). The intensity of agricultural activities largely determines the magnitude of soil and nutrient (N, P and K) loss to surface water. As a consequence, sediment yields and leaching of pollutants into surface water can lead to degradation of important aquatic habitat, affect recreational uses of water, and introduce toxins into the human food chain (Gitau et al., 2005). For example, elevated concentrations of NO 3 -N have been found in surface waters and groundwater in many agricultural lands (Forrest and Caux, 1988; Chow et al., 1995; Hubbard and Sheridan, 1989). Nagasaka et al. (2005) also found that the concentration of suspended sediments was consistently high in streams that drained from intensively cultivated catchments. In order to restore and maintain the quality of surface water, it is important to limit nutrient loading through soil conservation and water protection practices (Isermann, 1990). The potato growing region of New Brunswick (NB) has been identified as one of the regions in Canada most prone to soil erosion by water (Chow et al., 1999). Average annual soil losses of 17 tonne ha 1 yr 1 were reported by Saini and Grant (1980) for continuous potato planted up-and-down slope. Soil erosion rates found through plot measurements were reported to be from 1.2 to 24.3 tonne ha 1 yr 1 in the same region (Chow et al., 1990). Climatic conditions (especially intensive thunderstorms), shallow Agriculture, Ecosystems and Environment 132 (2009) 23–31 ARTICLE INFO Article history: Received 28 October 2008 Received in revised form 26 December 2008 Accepted 15 February 2009 Available online 27 March 2009 Keywords: Sediment yield Water quality Watershed Flow diversion terraces (FDT) Beneficial Management Practices (BMPs) Soil and Water Assessment Tool (SWAT) ABSTRACT Around the world, soil erosion is a major concern for the sustainability of agricultural systems and a threat to the integrity of aquatic ecosystems. Soil conservation Beneficial Management Practices (BMPs) have been widely used to reduce soil loss from cultivated lands and minimize degradation in water quality. As an effective BMP, flow diversion terraces (FDT) have been widely implemented in Canada and their impacts at field levels have been well documented. The objective of this study was to use the Soil and Water Assessment Tool (SWAT) to assess the efficacy of FDT systems on maintaining surface water quality at the watershed level in the Black Brook Watershed (BBW) in northwestern New Brunswick. The SWAT model was calibrated with three years of data (1992–1994) when limited amount of FDT had been implemented in the watershed. For this period, we found that SWAT performed well in predicting the seasonal variation of water yield (R 2 = 0.91) and moderately well for sediment yield (R 2 = 0.5). For key water variables, the SWAT model captured the variation of soluble phosphorous very well (R 2 = 0.81). However, the SWAT model over-predicted the amount of nitrate loading (R 2 = 0.02), and poorly represented the general trend of nitrate. The calibrated model was used to assess seasonal and annual effects of FDT on water quality for the entire watershed during the period of 1995–2005. The results indicated that FDT implemented in BBW contributed to the reduction of sediment yield by 4 tonne ha 1 yr 1 on average, which represented a reduction of 56%. We also found that the FDT system as a BMP not only reduced the sediment loading at a watershed level, but also reduced water yield during the summer growing seasons. FDT systems contributed to the reduction of water yield by 158 mm yr 1 , which represented a reduction of 20%. ß 2009 Elsevier B.V. All rights reserved. * Corresponding author. Tel.: +1 506 452 6332; fax: +1 506 453 3538. E-mail address: c2g63@unb.ca (Q. Yang). Contents lists available at ScienceDirect Agriculture, Ecosystems and Environment journal homepage: www.elsevier.com/locate/agee 0167-8809/$ – see front matter ß 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.agee.2009.02.012