Abstract

Nonpoint source (NPS) pollution has been studied for many years but it exhibits random, widespread, complex uncertainties which make it difficult to manage and control. We employ group decision-making utilizing the fuzzy comprehensive evaluation method (FCE) and the analytic hierarchy process method (AHP) and comparatively evaluate the optimal allocation of NPS pollution control measures. Here, we present the top-three evaluation results ranked as follows: combination of contour tillage and vegetative filter strips (CT & VFS), vegetative filter strips (VFS), and combination of contour tillage and fertilizer reduction and vegetative filter strips (CT & FR & VFS). The fourth, fifth and sixth results by FCE method are chemical fertilizer reduction (FR), returning farmland to forest or pasture (RF), and combination of contour tillage and fertilizer reduction (CT & FR), while the corresponding results by AHP method are returning farmland to forest or pasture (RF), combination of contour tillage and fertilizer reduction (CT & FR), and fertilizer reduction (FR). The seventh results for each of the two methods are contour tillage (CT), which has a positive but limited effect on nutrient loss reduction. Our results provide new underlying insights needed to guide the resonable allocation of NPS pollution control measures.

Introduction

Agricultural activities (e.g. irrigation and drainage, application of fertilizers) are the primary sources of phosphorus and nitrogen loss in the environment (Wang et al., 2010). Phosphorus is a nutrient that occurs in many forms that are bioavailable. It is most often transported to water bodies via soil erosion because many forms of phosphorus tend to be adsorbed to soil particles (Wu et al., 2012a, 2016a). Nitrogen is the other key ingredient in fertilizers. Similar to that of phosphorus in fresh waters, excess amounts of bioavailable nitrogen in aquatic systems lead to eutrophication and algae blooms (Wu et al., 2015). The Chinese Loess Plateau is considered to be one of such regions with the largest soil loss in the world (Jiao et al., 1999). Serious soil and water loss in the Loess hilly and gully region leads to nitrogen and phosphorus loss (Zhang et al., 2004; Li et al., 2008). Sediment-carrying nutrients may enter surface waters from eroding stream banks and also from surface runoff due to improper plant cover on agricultural land (ISU, 2001; Wu et al., 2012b, 2013, 2016b).

NPS pollution has characteristics of wide distribution, dispersion, randomness, uncertainty, potential presence, persistence, and uncontrollability (Loague & Corwin, 2006; Lv et al., 2012; Wu et al., 2014), and it has chronic effects on human health and soil-aquatic degradation (Zhang, 2004; Liu et al., 2009; Wu et al., 2017). Since the 1970s, many academics in the world have done a great deal of research on NPS pollution rules, monitoring means, control strategies, and comprehensive management. Some developed countries have formed a set of measures, and the United States first put forward the concept of ‘best management practices’ (BMPs). BMPs are defined by the United States Environmental Protection Agency as: ‘any methods, measures or operational procedures that can reduce or prevent water pollution, including engineering and non engineering measures or procedures’. BMPs have been paid more and more attention in the control of NPS pollution and are widely used in the world because they are in accordance with the principles of efficiency, economy, and ecology (Lin & Hsieh, 2003). According to the actual situation and local conditions, BMPs have good effects on NPS pollution control by one or several typical technologies. Maryland and Tennessee in the United States are examples of successful NPS pollution control projects by BMPs' application. Since the early 1980s, the study of NPS pollution in China has made some progress in some important lakes and rivers. Generally, the control systems of NPS pollution may include: fertilizer reduction and discharge reduction (Xue et al., 2013), ecological ditches, buffer zones, grass cover, denitrification ditches and wetland-multistage ponds (Shi et al., 2013), reuse nitrogen and phosphorus in animal manures, straws, wastewater from rural domestic treatments, and tail water treatments (Chang et al., 2013), ecological floating-bed technology, aquatic plants' restoration technology, ecological revetment technology (Liu et al., 2013b). They can be summarized as four levels of system control, including source reduce, process retain, nutrient reuse, and ecological restore (Yang et al., 2013). The difficulties of NPS pollution control in China mainly include two aspects: (1) the uncertainty of NPS pollution makes it difficult to evaluate; (2) farmers lack professional knowledge about the ecological damage of NPS pollution caused by agricultural activities (Zhang et al., 2004).

There is an increasing need for improved process-based planning tools to assist watershed managers in the selection and placement of effective BMPs (Brooks et al., 2015). Many existing studies have combined the intelligence algorithms and NPS prediction models to optimize the selection and placement of BMPs in a watershed (Srivastava et al., 2002; Veith et al., 2003; Gitau et al., 2004; Arabi et al., 2006; Maringanti et al., 2009; Shen et al., 2013). For example, linear and dynamic programming optimization of infiltration-based stormwater management BMPs and sediment-trapping BMPs are both compared with genetic algorithm (GA) optimization using a nonlinear distributed model (Limbrunner et al., 2013). A preference-based multi-objective model was designed by modifying the commonly used non-dominated sorting genetic algorithm (NSGA-II) to optimize BMPs at watershed scale (Chen et al., 2015). A Markov-based simulator that has been developed to quantify water quality responses is described, and a new framework is also proposed for the optimal design of BMPs by integrating the Markov approach, SWAT (Soil and Water Assessment Tool) model, and an NSGA-II evolutionary algorithm (Chen et al., 2016). The reduction of high-level NPS pollution discharges in Haean highland agricultural catchment (62.8 km2) was evaluated by applying BMPs of vegetation filter installation (VFS), fertilizer control (FC), and rice straw mulching (RSM) using SWAT (Sun et al., 2016). An interval-fuzzy possibilistic programming (IFPP) method was developed by integrating interval paramenter programming (IPP), fuzzy possibilistic programming (FPP), and a fuzzy expected value equation within a general optimization framework to identify optimal placements for BMPs in a watershed (Dai et al., 2016). In order to facilitate information transfer to stakeholders, Giri et al. (2015) integrated statistical and hydrological models to identify implementation sites (distance to the watershed outlet or the stream order) for agricultural conservation practices. A SWAT model of the Sunrise watershed was constructed to estimate load reductions due to selected BMPs (vegetated filter strips, grassed waterways, and reduction of soil-phosphorus concentrations) and to determine how phosphorus export coefficients scaled with the contributing area (Almendinger & Ulrich, 2017). Four different methods (the efficiency ratio, summation of loads, regression of loads, and frequency of rainfall) were applied and compared to eliminate uncertainties of removal efficiency determination for NPS BMPs (Lee et al., 2012). The effects of increases in effective impervious area (EIA) and the implementation of water quality protection designed detention pond BMPs on storm runoff and stormwater quality were assessed in eight small watersheds in Georgia for the period 2001–2008 (Aulenbach et al., 2017). An integrated approach, based on the Water Erosion Prediction Project model and a pesticide transport model, was presented to identify dominant hydrologic flow paths and critical source areas for a variety of pollutant types, and to compare the relative impacts of BMPs on hydrology, erosion, sediment, and pollutant delivery within different landscapes (Brooks et al., 2015). A total of 171 management-practice combinations that incorporate nutrient management, vegetated filter strips (VFS), and grazing management were evaluated for their performances in improving water quality in a pasture-dominated watershed with dynamic land-use changes during 1992–2007 by using the SWAT model (Chiang et al., 2012). A BMP effectiveness estimator driven by hydrologic soil groups and slope classes was developed and tested for estimating the effectiveness of BMPs in controlling NPS pollution by collecting and analyzing 60 existing BMPs' data from previous studies (Geng et al., 2015). A GA, an evolutionary optimization technique, was coupled with a semi-distributed hydrologic model, SWAT, to find an optimum combination of structural BMPs (detention ponds, parallel terraces, filter strips, grassed waterways, and grade stabilization structures) that meets the treatment goals at a watershed scale (Kaini et al., 2012). The decision support tool (DST) integrates the river basin SWAT model that serves as the NPS pollution estimator into an optimization framework consisting of a multi-objective GA that searches for optimal selection and location of BMPs in the landscape (Panagopoulos et al., 2013). A conceptual framework that relates agricultural BMP effectiveness with dominant hydrological flow paths was proposed and used to analyze plot, field and watershed-scale published studies on BMP effectiveness to develop transferable recommendations for BMP selection and placement at the watershed scale (Rittenburg et al., 2015).

In summary, new modeling techniques and existing or new data analysis are important requirements and opportunities for quantifying effectiveness of BMPs (Liu et al., 2017). Watershed models are generally used to evaluate the effectiveness of BMP performance in improving water quality as the basis for watershed management recommendations (Chiang et al., 2012). However, there are few studies on the application of modern comprehensive evaluation methods for the selection and placement of NPS control measures. Thus, this study focuses on soil and water conservation measures in a typical small watershed, and takes two different, modern comprehensive evaluation methods as the starting point: (1) to evaluate the reduction effects of nutrients' loss for different NPS pollution control measures in the Majiagou River watershed; (2) to determine the best selection and placement pattern of NPS pollution control measures. Results may provide a scientific basis for the development of control strategies to guarantee regional ecological environment security.

Material and methods

Soil conservation measures and NPS pollution control measures

Soil erosion and NPS pollution are inseparable symbiotic phenomena; soil erosion is the major occurrence form especially in sloping farmland NPS pollution of the Loess hilly and gully region in China (Wu et al., 2016c). China's soil and water conservation measures have been carried out for many years and have made significant achievements (Wu et al., 2010; Xie et al., 2010); fruitful experiences on prevention and control of soil loss have been accumulated and massive soil erosion prevention measures have been achieved (Wu et al., 2016d). Because of the comparatively significant correlation between soil erosion and NPS pollution, these measures have positive effects on the control of NPS pollution. On the basis of summarizing the previous research results, Liu et al. (2013a) put forward a classification system of soil and water conservation measures. This classification can be divided into three first-level categories: biological measures, engineering measures, and tillage measures, including 32 second-level types and 59 third-level types. For example, Wang et al. (2017) found that the contour plow was more effective in reducing the magnitude of runoff amount and sediment loss as well as increasing the rainwater infiltration amount compared to the traditional plow. Vegetative filter strips (VFS) implemented downstream to the source of pollution can trap sediments and thus limit sediment export from agricultural fields (Lambrechts et al., 2014). Therefore, the soil and water conservation measures and NPS pollution control measures were combined to determine and analyze NPS pollution control measures suitable for the Majiagou River watershed of the Loess hilly region. Majiagou River watershed (latitude 36°31′–37°19′N, longitude 108°51′–109°26′E), which is located in the western Ansai County of Shaanxi Province, is a first-grade tributary on the right bank of Yanhe River (Figure 1). The total watershed area is 73.83 km2 (Jia et al., 2014). The main channel length is about 17.5 km and the average slope gradient of the main channel is about 6.5‰ (Chen et al., 2011). The average annual precipitation of the Majiagou River watershed is 508 mm, the precipitation in flood season from June to September accounts for 69.5% of the annual precipitation, and it is characterized by heavy rainfall, short duration, and high intensity (Fu et al., 2010). The land-use types in the watershed are mainly dominated by farmland and grassland, of which the sloping farmland area accounts for about 50%, and the grassland area accounts for more than 45% (Figure 1(c)). Majiagou River watershed is one of the most typical soil erosion regions of the Loess Plateau. The implementation of a returning farmland project in 1997 was an important strategy to improve the ecological environment of the Loess Plateau (Teng et al., 2015; Wu et al., 2016e). According to existing research results (Zheng et al., 2005; Wu & Ma, 2015; Ma et al., 2016), the main pollution control measures and their reduction rates are summarized in Table 1.

Table 1.

Main nonpoint source (NPS) pollution control measures and their reduction rate in the Majiagou River watershed.

Control measures TN reduction rate TP reduction rate 
Returning farmland to forest (pasture) 1.03%–5.35% 0.94%–8.09% 
Contour tillage 0.51%–2.77% 0.49%–4.54% 
Fertilizer reduction 0.65%–6.52% 0.01%–2.95% 
Vegetative filter strips with width (0.5 m,1.0 m,1.5 m, and 2.0 m) of Bermuda grass and 7.1 m Panicum virgatum 8%, 42%, 56%, 59%, and 80% 20%, 77%, 86%, 89%, and 78% 
Combination of contour tillage (0.4) and fertilizer reduction (0.6) 0.59%–5.02% 0.2%–3.59% 
Combination of contour tillage (0.3) and vegetative filter strips (0.7) 29.55%–30.23% 54.05%–55.26% 
Combination of contour tillage (0.2), fertilizer reduction (0.4) and vegetative filter strips (0.4) 17.16%–19.96% 30.9%–32.89% 
Control measures TN reduction rate TP reduction rate 
Returning farmland to forest (pasture) 1.03%–5.35% 0.94%–8.09% 
Contour tillage 0.51%–2.77% 0.49%–4.54% 
Fertilizer reduction 0.65%–6.52% 0.01%–2.95% 
Vegetative filter strips with width (0.5 m,1.0 m,1.5 m, and 2.0 m) of Bermuda grass and 7.1 m Panicum virgatum 8%, 42%, 56%, 59%, and 80% 20%, 77%, 86%, 89%, and 78% 
Combination of contour tillage (0.4) and fertilizer reduction (0.6) 0.59%–5.02% 0.2%–3.59% 
Combination of contour tillage (0.3) and vegetative filter strips (0.7) 29.55%–30.23% 54.05%–55.26% 
Combination of contour tillage (0.2), fertilizer reduction (0.4) and vegetative filter strips (0.4) 17.16%–19.96% 30.9%–32.89% 

TN, total nitrogen; TP, total phosphorus.

Fig. 1.

(a) Relative location of China's river system, Yellow River basin and Yanhe River basin; (b) relative location of rivers, Yanhe River basin and Majiagou River watershed, DEM of Yanhe River basin; and (c) reclassified land use types of the Majiagou River watershed.

Fig. 1.

(a) Relative location of China's river system, Yellow River basin and Yanhe River basin; (b) relative location of rivers, Yanhe River basin and Majiagou River watershed, DEM of Yanhe River basin; and (c) reclassified land use types of the Majiagou River watershed.

Methodology

Fuzzy comprehensive evaluation (FCE) method and its application

In 1965, American scientist, Professor L. A. Zadeh of the University of California, Berkeley published an article ‘fuzzy sets’ in the Journal of Information and Control (Zadeh, 1965), and put forward the concept of fuzzy mathematics, which is the theory basis of the FCE method. This method reasonably quantified uncertain and unclear border information by the concept of fuzzy mathematics, and then comprehensively evaluated the membership degree of the evaluated things by means of the fuzzy relation synthesis principle. According to the evaluation results, we can get the rank of the evaluation objects and their specific information. This method has a good result in the evaluation of a large number of uncertain factors; the application scope is wide (Chen & Sun, 2002). Although this method may be affected by subjective factors, it has the advantage of combining quantitative and qualitative research for the fuzzy things under various objective conditions. In this study, the comprehensive index system of NPS pollution control measures in the Majiagou River watershed is determined as shown in Figure 2.

Fig. 2.

Comprehensive evaluation structure of nonpoint source (NPS) pollution control measures in the Majiagou River watershed.

Fig. 2.

Comprehensive evaluation structure of nonpoint source (NPS) pollution control measures in the Majiagou River watershed.

The steps of the FCE method in this study may be summarized as follows. (1) Determine the set of evaluation indicators: the index system involves a total of two levels. The first layer: the total evaluation target U includes two impact indicators, the index set is U = {UA, UB} = {maneuverability, reduction rate}. The second layer: UA = {UA1, UA2} = {acceptability of local residents, difficulty of implementation}, UB = {UB1, UB2} = {the reduction rate of TN, the reduction rate of TP}. (2) Select the evaluation set: this study intended to adopt China's five classification levels: very good (class I), good (class II), general (class III), poor (class IV), and very poor (class V); and so to determine the comments set V = {v1,v2,v3,v4,v5} = {very good, good, general, poor, very poor}. (3) Calculate the membership degree matrix: unify the evaluation indexes and convert them into efficiency indicators, class the grade of TN and TP reduction rate through the quantitative method, and then calculate the membership degree matrix. The corresponding TN and TP reduction rates for different NPS pollution control measures in the Majiagou River watershed are listed in Table 2. (4) Determine the weight vector of the factor set and normalize the evaluation set. (5) Calculate the comprehensive evaluation vector (comprehensive membership degree). (6) Calculate the comprehensive evaluation value according to the principle of maximum membership degree.

Table 2.

The reduction rate level of TN (√) and TP (#) for different nonpoint source (NPS) pollution control measures in the Majiagou River watershed.

Abbreviation Evaluation grades II III IV 
Reduction rate [8%, ∞) [6%, 8%) [4%, 6%) [2%, 4%) (−∞, %] 
RF Returning farmland to forest or pasture  √   
CT Contour tillage   √  
FR Fertilizer reduction  √   
VFS Vegetative filter strips √#     
CT & FR Combination of contour tillage and fertilizer reduction   √  
CT &VFS Combination of contour tillage and vegetative filter strips √#     
CT & FR & VFS Combination of contour tillage, fertilizer reduction and vegetative filter strips √#     
Abbreviation Evaluation grades II III IV 
Reduction rate [8%, ∞) [6%, 8%) [4%, 6%) [2%, 4%) (−∞, %] 
RF Returning farmland to forest or pasture  √   
CT Contour tillage   √  
FR Fertilizer reduction  √   
VFS Vegetative filter strips √#     
CT & FR Combination of contour tillage and fertilizer reduction   √  
CT &VFS Combination of contour tillage and vegetative filter strips √#     
CT & FR & VFS Combination of contour tillage, fertilizer reduction and vegetative filter strips √#     

TN, total nitrogen; TP, total phosphorus.

Analytic hierarchy process method and its application

Analytic hierarchy process (AHP) is a structured technique for organizing and analyzing complex decisions, based on mathematics and psychology. AHP is a hierarchy decision-making analytic method which combines qualitative and quantitative analysis to solve complex problems with multiple objectives. It was developed by Professor Thomas L. Saaty, an operational research expert of the University of Pittsburgh in the early 1970s (Saaty, 2001) and has been extensively used and refined since then. By deeply analyzing the inherent relationship between factors and group decision-making using less quantitative information, the characteristic of AHP may be summarized to provide a simple decision-making method for complex multi-objective problems. Rather than precribing a ‘correct’ decision, the AHP helps decision-makers find one that best suits their goal and their understanding of the problem. This method is especially suitable for occasions where the decision result is difficult to measure directly and accurately. It not only has the advantages of practicality and simplicity, but also has the characteristics of systematicness and reliability. The comprehensive evaluation chart of the AHP method is shown in Figure 3.

Fig. 3.

The comprehensive evaluation chart of nonpoint source (NPS) pollution control measures in Majiagou River watershed. Combination 1: contour tillage and fertilizer reduction (CT & FR); combination 2: contour tillage and vegetative filter strips (CT & VFS); combination 3: contour tillage, fertilizer reduction, and vegetative filter strips (CT & FR & VFS).

Fig. 3.

The comprehensive evaluation chart of nonpoint source (NPS) pollution control measures in Majiagou River watershed. Combination 1: contour tillage and fertilizer reduction (CT & FR); combination 2: contour tillage and vegetative filter strips (CT & VFS); combination 3: contour tillage, fertilizer reduction, and vegetative filter strips (CT & FR & VFS).

The specific calculation steps of the AHP method are as follows. (1) Establish a hierarchical analysis model. (2) Construct conjugate comparison matrix: evaluate four control measures to obtain a comparison matrix with a single standard ‘maneuverability’, ‘reduction rate of TN’, and ‘reduction rate of TP’. (3) The weight vector is calculated by the standard column average method. In addition, the relative importance (the relative weight, also known as the standard eigenvector) of each criterion in the total target should be obtained. (4) The consistency test for comparison matrix. (5) Calculate the combined weight vector which can be used as the quantitative basis for decision-making.

Results and discussion

The evaluation results of the FCE method

Figure 4 shows first-level evaluation result, second-level evaluation results of ‘maneuverability’, and ‘reduction rate’ by the FCE method, respectively. According to the principle of maximum membership degree in the FCE method and different evaluation grades, the first-level evaluation ranks (Figure 4(a)) of each single control measure of NPS pollution and their combinations in the Majiagou River watershed were obtained as follows: first, combination of contour tillage and vegetative filter strips (CT & VFS); second, vegetative filter strips (VFS); third, combination of contour tillage, fertilizer reduction, and vegetative filter strips (CT & FR & VFS); fourth, fertilizer reduction (FR); fifth, returning farmland to forest or pasture (RF); sixth, combination 1: contour tillage and fertilizer reduction (CT & FR); seventh, contour tillage (CT).

Fig. 4.

First-level evaluation results (a), second-level evaluation result of ‘maneuverability’ (RA) (b), and ‘reduction rate’ (RB) (c) for various control measures by the FCE method. RF: returning farmland to forest or pasture; CT: contour tillage; FR: fertilizer reduction; VFS: vegetative filter strips; CT & FR: combination of contour tillage and fertilizer reduction; CT & VFS: combination of contour tillage and vegetative filter strips; CT & FR & VFS: combination of contour tillage, fertilizer reduction, and vegetative filter strips.

Fig. 4.

First-level evaluation results (a), second-level evaluation result of ‘maneuverability’ (RA) (b), and ‘reduction rate’ (RB) (c) for various control measures by the FCE method. RF: returning farmland to forest or pasture; CT: contour tillage; FR: fertilizer reduction; VFS: vegetative filter strips; CT & FR: combination of contour tillage and fertilizer reduction; CT & VFS: combination of contour tillage and vegetative filter strips; CT & FR & VFS: combination of contour tillage, fertilizer reduction, and vegetative filter strips.

For the second-level evaluation results of ‘maneuverability’ (Figure 4(b)), the evaluation ranks of each single control measure of NPS pollution and their combinations are listed as: CT & FR, RF, CT & VFS, VFS, CT, CT & FR & VFS, and FR. For the second-level evaluation results of ‘reduction rate’ (Figure 4(c)), the evaluation ranks are described as: CT & FR & VFS, VFS, CT & VFS, FR, RF, CT & FR, and CT. Based on the second-level evaluation results of ‘maneuverability’ and ‘reduction rate’, we found that there are some differences in priority levels of different control measures. However, if combining the results of the first evaluation level with the second evaluation level, measures of VFS, CT & VFS, and CT & FR & VFS all have a good performance. Thus, the results of the first-level evaluation are basically consistent with the results of the second-level evaluation by the FCE method. Majiagou River watershed is located in the Loess hilly region. Sloping farmland is widely distributed in this watershed. Vegetation filter strips applied on the valley slope and returning farmland measures implemented on steep slopes are important ways to reduce NPS nitrogen and phosphorus loss in the Loess hilly region.

The evaluation results of the AHP method

Through establishing a comparison matrix, the standard eigenvectors of ‘maneuverability’ (0.17), ‘reduction rate of TN’ (0.387), and ‘reduction rate of TP’ (0.476), and comprehensive grades were respectively calculated and are listed in Figure 5. Based on the total scores of each scheme, the comprehensive evaluation ranks of different NPS pollution control measures by the AHP method are as follows: first, combination CT & VFS; second, VFS; third, combination of CT & FR & VFS; fourth, RF; fifth, combination of CT & FR; sixth, FR; seventh, CT. The above results indicate that VFS and its different combinations have good effects on nonpoint source (NPS) pollution control. VFS are areas of either planted or indigenous established vegetation designed to improve the quality of surface runoff (Pan et al., 2017). Studies indicate that they can be incorporated into pastures, grassed waterways, terraces or cropland to remove sediment, nitrogen or phosphorus from runoff (Lobo & Bonilla, 2017). However, the effect of contour tillage is relatively weak. Fang & Sun (2017) also found that measures such as terraced and contour tillage lands are not effective enough to comprehensively control soil erosion. The majority of up/downslope tillage land and steep slopes with gradients above 25% still suffered severe soil loss. Comprehensive soil conservation should be urgently applied to reduce soil erosion and sediment delivery to rivers.

Fig. 5.

The comprehensive grades of different measures and their combinations for a single evaluation standard of AHP method in the Majiagou River watershed. TN: total nitrogen; TP: total phosphorus; RF: returning farmland to forest or pasture; CT: contour tillage; FR: fertilizer reduction; VFS: vegetative filter strips; CT & FR: combination of contour tillage and fertilizer reduction; CT &VFS: combination of contour tillage and vegetative filter strips; CT & FR & VFS: combination of contour tillage, fertilizer reduction and vegetative filter strips.

Fig. 5.

The comprehensive grades of different measures and their combinations for a single evaluation standard of AHP method in the Majiagou River watershed. TN: total nitrogen; TP: total phosphorus; RF: returning farmland to forest or pasture; CT: contour tillage; FR: fertilizer reduction; VFS: vegetative filter strips; CT & FR: combination of contour tillage and fertilizer reduction; CT &VFS: combination of contour tillage and vegetative filter strips; CT & FR & VFS: combination of contour tillage, fertilizer reduction and vegetative filter strips.

Analysis and discussion on results of optimal configuration

The evaluation results were obtained through two kinds of modern comprehensive evaluation methods, namely the FCE method and the AHP method. Comprehensive evaluation results of NPS pollution control measures according to the above two evaluation conclusions of Majiagou River watershed can be drawn as follows: first, combination 2: CT & VFS; second, VFS; third, combination 3: CT & FR & VFS. These results show that the comprehensive control effects of these three kinds of NPS pollution control measures in the Majiagou River watershed are all good. However, the seventh evaluation levels by the two kinds of evaluation method are both contour tillage, which has a positive but limited effect for the reduction of TN and TP loss in the Majiagou River watershed.

In addition, the fourth evaluation result obtained by the FCE method is the reduction of chemical fertilizer, the fifth is the returning farmland to forest or pasture, and the sixth is the combination of contour tillage and chemical fertilizer reduction. Meanwhile, the fourth evaluation result obtained by the AHP method is the returning farmland to forest or pasture, the fifth is the combination of contour tillage and chemical fertilizer reduction, and the sixth is the reduction of chemical fertilizer. Due to the difference between the two methods in principle and the evaluation process, the results of the two judgments are inconsistent to a certain extent. Generally, sloping farmland and desertification land should stop farming in a planned and step-by-step manner, and then forest/pasture planted to recover vegetation according to planting principle and local conditions. Contour tillage is a kind of conservation tillage, also known as cross-slope cultivation, and the most common form is terraced fields. It mainly reduces soil erosion and NPS pollution by retarding the gradient of slope. Chemical fertilizer is one of the main sources of agricultural NPS pollution. As far as the source control idea is concerned, chemical fertilizer reduction is the fundamental measure to reduce NPS pollution. The vegetative filter strips are also known as the vegetation buffer zone and can be understood as a kind of vegetation zone that separates the pollution source from the receiving water bodies. The use of vegetative filter strips is considered to be one of the most effective methods to control NPS pollution.

As there are few studies about the reduction rate of various NPS pollution control measures in the Majiagou River watershed at the current stage, the reduction rates of various NPS pollution measures in Ashe River basin reported by Ma et al. (2016) were applied in this study. Due to differences in various natural conditions and other aspects between Majiagou River watershed and Ashe River basin, the reduction rates of various measures in Ashe River basin have certain differences from the Majiagou River watershed; such reasons may cause inaccurate results. In addition, owing to lack of actual statistical data of the ‘maneuverability’ index for various NPS pollution control measures in the Majiagou River watershed, too much of the evaluation vector in the FCE method is based on personal subjective factors, so there are unreliable factors in the process of judgment. From a methodological point of view, AHP is a subjective judgment method, where subjective factors have great influence on the whole process from the establishment of the hierarchical structure model to the construction of the comparison matrix, which makes the results difficult to accept by all decision-makers. Of course, group judgment method by experts is a way to overcome this shortcoming. As well, the comparison, judgment, and calculation processes of the FCE method are also relatively rough. On the whole, the modern comprehensive evaluation methods have some reliability in the allocation study of NPS pollution control measures.

Suggestions for NPS pollution control in Majiagou River watershed

According to the allocation results of NPS pollution control measures by the FCE method and the AHP method, the combination scheme of contour tillage and vegetative filter strips was suggested as the top preference in the process of watershed NPS pollution control. If there are some difficulties in implementing the combined measures, the measure of vegetative filter strips can also be directly implemented in this watershed. To some extent, the regional NPS is a collection of different point sources. Cutting off the spatial exchange between point source and NPS pollutants may be an important aspect of NPS pollution control, and eliminating pollution connection may become an effective way to solve the NPS pollution problems. The current researches are mainly focused on a certain specific control measure and the corresponding measures are confined in a local range due to watershed conditions. Surface runoff system and groundwater system are often closely related to land ecosystem, and a separate system cannot fundamentally solve the problem of NPS pollution. Therefore, strengthening research on pollution association, exchange pathway and carrying capacity between related systems may acquire the maximum effectiveness of NPS pollution control in a watershed.

Conclusion

In summary, use of modern comprehensive evaluation methods provides a new way of thinking and assessment for optimal allocation studies of NPS pollution control measures. The main NPS pollution control measures suitable for the Majiagou River watershed were determined by combining soil and water conservation measures with NPS pollution control measures. The selection and placement of NPS pollution control measures were comparatively carried out to explore effectiveness of various control measures and their combinations by the FCE method and the AHP method. The evaluation results by two kinds of optimization methods both showed that the combination of contour tillage and vegetative filter strips (CT & VFS) is the best scheme, vegetative filter strips (VFS) is second, and the combination of contour tillage, fertilizer reduction and vegetative filter strips (CT & FR & VFS) is third. The fourth, fifth and sixth comprehensive evaluation results obtained by the FCE method are respectively fertilizer reduction (FR), returning farmland to forest or pasture (RF), and the combination of contour tillage and fertilizer reduction (CT & FR). The corresponding results obtained by the AHP method are respectively returning farmland to forest or pasture (RF), combination of contour tillage and fertilizer reduction (CT & FR), and fertilizer reduction (FR). Both the seventh results by the two methods are contour tillage (CT), which is not recommended to be used independently. The optimal allocation results of NPS pollution control measures by these two methods are feasible. Results may provide policy support for the resonable allocation of NPS pollution control measures in Loess hilly region.

Acknowledgments

Special thanks are given to the anonymous reviewers and the editor for their useful suggestions on the manuscript. This study was supported by the National Natural Science Foundation of China (51679206, 51309194), Tang Scholar (Z111021720), Youth Science and Technology Nova Project in Shaanxi Province (2017KJXX-91), International Science and Technology Cooperation Fund (A213021603), the Fundamental Research Funds for the Central Universities (2452016120), Special Research Foundation for Young Teachers (2452015374), the Doctoral Fund of Ministry of Education of China (20130204120034). This study was also supported by the National Fund for Studying Abroad. Data support came from the Loess Plateau Data Center, National Earth System Science Data Sharing Infrastructure, National Science & Technology Infrastructure of China (http://loess.geodata.cn). There are no conflicts of interest.

References

References
Almendinger
J. E.
&
Ulrich
J. S.
, (
2017
).
Use of SWAT to estimate spatial scaling of phosphorus export coefficients and load reductions due to agricultural BMPs
.
Journal of the American Water Resources Association
53
(
3
),
547
561
.
Arabi
M.
,
Govindaraju
R. S.
&
Hantush
M. M.
, (
2006
).
Cost-effective allocation of watershed management practices using a genetic algorithm
.
Water Resources Research
42
,
1
8
.
Aulenbach
B. T.
,
Landers
M. N.
,
Musser
J. W.
&
Painter
J. A.
, (
2017
).
Effects of impervious area and BMP implementation and design on storm runoff and water quality in eight small watersheds
.
Journal of the American Water Resources Association
53
(
2
),
382
399
.
Brooks
E. S.
,
Saia
S. M.
,
Boll
J.
,
Wetzel
L.
,
Easton
Z. M.
&
Steenhuis
T. S.
, (
2015
).
Assessing BMP effectiveness and guiding BMP planning using process-based modeling
.
Journal of the American Water Resources Association
51
(
2
),
343
358
.
Chang
Z. Z.
,
Huang
H. Y.
,
Jin
H. M.
,
Ma
Y.
,
Ye
X. M.
,
Xue
L. H.
&
Yang
L. Z.
, (
2013
).
Reduce-retain-reuse-restore technology for controlling the agricultural non-point source pollution in countryside in China: reuse of nitrogen and phosphorus in agricultural wastes
.
Journal of Agro-Environment Science
32
(
10
),
1901
1907
.
Chen
Y. H.
&
Sun
C. Y.
, (
2002
).
Further study of validity for the maximum subordination principle
.
Journal of Chongqing Normal University (Natural Science Edition)
19
(
1
),
47
49
.
Chen
F. J.
,
Zhang
X. P.
,
Fu
Y. L.
,
Wu
Y.
,
Wan
L.
&
Liu
G. B.
, (
2011
).
Land use change of Majiagou catchment in Ansai county, Northern Shaanxi Province in recent 20 years
.
Bulletin of Soil and Water Conservation
31
(
1
),
244
248
.
Chen
L.
,
Qiu
J.
,
Wei
G.
&
Shen
Z.
, (
2015
).
A preference-based multi-objective model for the optimization of best management practices
.
Journal of Hydrology
520
,
356
366
.
Chiang
L. C.
,
Chaubey
I.
,
Hong
N. M.
,
Lin
Y. P.
&
Tao
H.
, (
2012
).
Implementation of BMP strategies for adaptation to climate change and land use change in a pasture-dominated watershed
.
International Journal of Environmental Research & Public Health
9
(
10
),
3654
3684
.
Dai
C.
,
Cai
Y. P.
,
Ren
W.
,
Xie
Y. F.
&
Guo
H. C.
, (
2016
).
Identification of optimal placements of best management practices through an interval-fuzzy possibilistic programming model
.
Agricultural Water Management
165
,
108
121
.
Fu
Y. L.
,
Zhang
X. P.
,
Chen
F. J.
,
Wan
L.
,
Wu
Y.
&
Liu
G. B.
, (
2010
).
Land use/cover pattern survey under the background of grain for green policy implementation in the Loess Hilly-Gully Region – a case study of Majiagou Catchment, Ansai County
.
Research of Soil and Water Conservation
17
(
6
),
81
85
.
Giri
S.
,
Nejadhashemi
A. P.
,
Zhang
Z.
&
Woznicki
S. A.
, (
2015
).
Integrating statistical and hydrological models to identify implementation sites for agricultural conservation practices
.
Environmental Modelling & Software
72
(
C
),
327
340
.
Gitau
M. W.
,
Veith
T. L.
&
Gburek
W. J.
, (
2004
).
Farm-level optimization of BMP placement for cost-effective pollution reduction
.
Transactions of the ASAE
47
,
1923
1931
.
Iowa State University (ISU)
(
2001
).
Iowa Fact Sheet: Agriculture and Water Quality
.
October 2001, Document No. EDC232a
.
University Extension
,
Ames, IA
.
Jia
L. Z.
,
Gao
J. E.
,
Zhang
Y. X.
,
Zhang
M. J.
,
Wang
X. W.
&
Li
X. H.
, (
2014
).
Analysis of rainstorm erosion disaster in terrace field in Loess Hilly and Gully region
.
Research of Soil and Water Conservation
21
(
4
),
7
11
.
Jiao
J. Y.
,
Wang
W. Z.
&
Hao
X. P.
, (
1999
).
Precipitation and erosion characteristics of different types of rainfall in the Loess Plateau
.
Journal of Arid Land Resources and Environment
13
(
1
),
34
42
.
Kaini
P.
,
Artita
K.
&
Nicklow
J. W.
, (
2012
).
Optimizing structural best management practices using SWAT and genetic algorithm to improve water quality goals
.
Water Resources Management
26
(
7
),
1827
1845
.
Lee
S.
,
Maniquiz
M. C.
&
Kim
L. H.
, (
2012
).
Appropriate determination method of removal efficiency for nonpoint source best management practices
.
Desalination & Water Treatment
48
(
1–3
),
138
147
.
Li
R.
,
Li
B. C.
&
Yang
W. Z.
, (
2008
).
Research and Prospect of Loess Plateau in China
.
Science Press
,
Beijing
.
Limbrunner
J. F.
,
Vogel
R. M.
,
Chapra
S. C.
&
Kirshen
P. H.
, (
2013
).
Classic optimization techniques applied to stormwater and nonpoint source pollution management at the watershed scale
.
Journal of Water Resources Planning & Management
139
(
5
),
486
491
.
Liu
Z. J.
,
Lu
J. H.
,
Fang
H.
&
Zhang
Y. A.
, (
2009
).
Types, characteristics, sources of non-point source pollution and its control technology
.
Anhui Agricultural Science Bulletin
15
(
5
),
98
101
.
Liu
B. Y.
,
Liu
Y. N.
,
Zhang
K. L.
&
Xie
Y.
, (
2013a
).
Classification for soil conservation practices in China
.
Journal of Soil and Water Conservation
27
(
2
),
80
84
.
Liu
F. X.
,
Song
X. F.
,
Zou
G. Y.
,
Fu
Z. S.
,
Liu
Y. Q.
,
Xue
L. H.
&
Yang
L. Z.
, (
2013b
).
Reduce-retain-reuse-restore technology for controlling the agricultural non-point source pollution in countryside in China: eco-restoration technology
.
Journal of Agro-Environment Science
32
(
11
),
2105
2111
.
Liu
Y.
,
Engel
B. A.
,
Flanagan
D. C.
,
Gitau
M. W.
,
McMillan
S. K.
&
Chaubey
I.
, (
2017
).
A review on effectiveness of best management practices in improving hydrology and water quality: needs and opportunities
.
Science of the Total Environment
601–602
,
580
593
.
Loague
K.
&
Corwin
D. L.
, (
2006
).
Point and Nonpoint Source Pollution. Encyclopedia of Hydrological Sciences
.
John Wiley & Sons
,
New York
.
Lv
X. Z.
,
Yu
X. X.
,
Fan
D. X.
&
Li
Q.
, (
2012
).
Estimation of non-point source pollution loads caused by soil erosion in China
.
Journal of Food Agriculture & Environment
10
(
2
),
1045
1050
.
Ma
F.
,
Jiang
X. F.
,
Wang
L.
,
Li
G. M.
&
Li
Z.
, (
2016
).
Non-point source pollution control of Ashihe Basin based on SWAT Model
.
China Environmental Science
36
(
2
),
610
618
.
Pan
D.
,
Gao
X.
,
Dyck
M.
,
Song
Y.
,
Wu
P.
&
Zhao
X.
, (
2017
).
Dynamics of runoff and sediment trapping performance of vegetative filter strips: run-on experiments and modeling
.
Science of the Total Environment
593–594
,
54
64
.
Panagopoulos
Y.
,
Makropoulos
C.
&
Mimikou
M.
, (
2013
).
Multi-objective optimization for diffuse pollution control at zero cost
.
Soil Use & Management
29
(
s1
),
83
93
.
Rittenburg
R. A.
,
Squires
A. L.
,
Boll
J.
,
Brooks
E. S.
,
Easton
Z. M.
&
Steenhuis
T. S.
, (
2015
).
Agricultural BMP effectiveness and dominant hydrological flow paths: concepts and a review
.
Journal of the American Water Resources Association
51
(
2
),
305
329
.
Saaty
T. L.
, (
2001
).
Fundamentals of Decision Making and Priority Theory with the Analytic Hierarchy Process (AHP)
.
RWS Publications
,
Pittsburgh, PA
.
Shi
W. M.
,
Xue
L. H.
,
Wang
J. G.
,
Liu
F. X.
,
Song
X. F.
&
Yang
L. Z.
, (
2013
).
Reduce-retain-reuse-restore technology for controlling rural non-point source pollution in China: eco-retain technlogy
.
Journal of Agro-Environment Science
32
(
9
),
1697
1704
.
Srivastava
P.
,
Hamlett
J. M.
,
Robillard
P. D.
&
Day
R. L.
, (
2002
).
Watershed optimization of best management practices using AnnAGNPS and a genetic algorithm
.
Water Resources Research
38
,
1
14
.
Sun
S. J.
,
Ahn
S. R.
&
Kim
S. J.
, (
2016
).
Evaluation of executable best management practices in Haean highland agricultural catchment of South Korea using SWAT
.
Agricultural Water Management
180
,
224
234
.
Teng
H.
,
Xia
J. X.
&
Ren
H. T.
, (
2015
).
Study on GTGP effect of soil and water conservation function in Yan river basin in the Loess plateau
.
Journal of Sediment Research
40
(
4
),
14
19
.
Veith
T. L.
,
Wolfe
M. L.
&
Heatwole
C. D.
, (
2003
).
Optimization procedure for cost effective BMP placement at a watershed scale
.
Journal of the American Water Works Association
39
,
1331
1343
.
Wang
Q. J.
,
Wang
H.
&
Guo
T. L
, . (
2010
).
Soil Solute Transport Characteristics and Mathematical Models in Loess Slope
.
Science Press
,
Beijing
.
Wu
L.
&
Ma
X. Y.
, (
2015
).
Research progress in erosion-type nonpoint source pollution process simulation of the Loess Plateau
.
China Sciencepaper
10
(
13
),
1497
1506
.
Wu
Y.
,
Zhang
X. P.
,
Chen
F. J.
,
Wan
L.
,
Fu
Y. L.
&
Liu
G. B.
, (
2010
).
Preliminary investigation on soil erosion and spatial distribution after returning farmland in typical watershed of Northern Shaanxi Province
.
Research of Soil and Water Conservation
17
(
4
),
29
33
.
Wu
L.
,
Long
T. Y.
,
Liu
X.
&
Mmereki
D.
, (
2012b
).
Simulation of soil loss processes based on rainfall runoff and the time factor of governance in the Jialing River Watershed, China
.
Environmental Monitoring and Assessment
184
(
6
),
3731
3748
.
Wu
L.
,
Qi
T.
,
Li
D.
,
Yang
H. J.
,
Liu
G. Q.
,
Ma
X. Y.
&
Gao
J. E.
, (
2015
).
Current status, problems and control strategies of water resources pollution in China
.
Water Policy
17
(
3
),
423
440
.
Wu
L.
,
Liu
X.
&
Ma
X. Y.
, (
2016c
).
Tracking soil erosion changes in an easily-eroded watershed of the Chinese Loess Plateau
.
Polish Journal of Environmental Studies
25
(
1
),
332
344
.
Wu
L.
,
Liu
X.
&
Ma
X. Y.
, (
2016d
).
Impacts of Grain for Green project on spatiotemporal variations of soil erosion in a typical watershed of Chinese Loess hilly and gully region
.
Fresenius Environmental Bulletin
25
(
11
),
4506
4516
.
Xie
H. X.
,
Yang
Q. K.
,
Li
R.
,
Zhang
X. P.
,
Cheng
L.
&
Pang
G. W.
, (
2010
).
Quantitative assessment of impacts on decreasing erosion by soil and water conservation measures in Yanhe River Basin
.
Science of Soil and Water Conservation
8
(
4
),
13
19
.
Xue
L. H.
,
Yang
L. Z.
,
Shi
W. M.
&
Wang
S. Q.
, (
2013
).
Reduce-retain-reuse-restore technology for controlling the agricultural non-point source pollution in countryside in China: source reduction technology
.
Journal of Agro-Environment Science
32
(
5
),
881
888
.
Yang
L. Z.
,
Shi
W. M.
,
Xue
L. H.
,
Song
X. B.
,
Wang
S. Q.
&
Chang
Z. Z.
, (
2013
).
Reduce-retain-reuse-restore technology for controlling the agricultural non-point source pollution in countryside in China: general countermeasures and technologies
.
Journal of Agro-Environment Science
32
(
1
),
1
8
.
Zadeh
L. A.
, (
1965
).
Fuzzy sets
.
Information and Control
8
(
3
),
338
353
.
Zhang
H. Y.
, (
2004
).
Economic Research on the Non-Point Source Pollution in Developed Areas
.
Doctoral dissertation
,
Fudan University
,
Shanghai
.
Zhang
S. G.
,
Zhang
X. H.
&
Wang
Y. L.
, (
2004
).
Reflections on non point source pollution control problem in the Yellow River
.
Soil and Water Conservation in China
8
,
11
12
.
Zheng
T.
,
Mu
H. Z.
,
Huang
H. C.
&
Zhang
C. P.
, (
2005
).
Study progress on nonpoint source pollution
.
Environmental Protection
1
(
2
),
31
34
.