The evaluation of groundwater quality plays an important part in the evaluation of groundwater resources. It analyses the temporal and spatial distributions and utilisation of underground water according to the main components and corresponding water quality standards for underground water. Thereby, it can provide a scientific basis for the development, utilisation, planning, and management of groundwater resources. Set pair analysis (SPA), based on the improved five-element connectivity degree, was used in this research to establish a comprehensive evaluation model of water quality, so as to evaluate the groundwater quality in XuChang, Henan Province, China. Meanwhile, fuzzy evaluation was also used to measure groundwater quality. As demonstrated in the research results, SPA is proven to be convenient and useful with objective and stable results, it therefore is an effective approach with which to evaluate groundwater quality. In addition, the results obtained using SPA matched those from fuzzy comprehensive evaluation; it was concluded, based on the analysis, that the groundwater in XuChang was severely polluted. The groundwater quality at the observation points located in the lower reaches is poorer than that of the upper reaches; hazardous substances permeate underground to pollute shallow groundwater through decomposition and loss due to weathering and rainfall.

INTRODUCTION

Water is essential for life, and the safety of drinking water is directly related to human health (Viala 2008). The evaluation of groundwater quality plays an essential role in the evaluation of groundwater resources. It analyses the temporal and spatial distributions and utilisation of underground water based on the main components and the corresponding water quality standards of underground water. In this way, it can provide a scientific basis for the development, utilisation, planning, and management of groundwater resources (Ranjan et al. 2012).

There are many mathematical models used for water quality evaluation, and among them the frequently used models include: principal component analysis, analytic hierarchy process, fuzzy mathematics methods, and so on (Dahiya et al. 2007; Faruk 2010; Pathak & Hiratsuka 2011; Maiti et al. 2013). Each model has its own advantages and disadvantages (Pathak & Limaye 2011; Khader & McKee 2014). While evaluating groundwater quality, set pair analysis (SPA) firstly performs qualitative analysis of the samples based on a certainty and uncertainty analysis, and describes the connectivity of set pairs simultaneously (Li et al. 2015; Wang et al. 2015). Thereby, the relative proportion accounted for by groundwater quality in each evaluation grade is reflected, and it therefore can produce more reasonable results in such studies (Zou et al. 2013). Meanwhile, as a decision system based on dialectical thinking, it is an effective method for analysing uncertain multi-objective decisions. Based on the principle of SPA, the three-element connectivity of SPA was extended to an improved five-element connectivity degree (IFCD) in this research. Meanwhile, certain and uncertain factors were taken into account to carry out grade evaluation of groundwater quality in XuChang. The groundwater sources of drinking water are limited in XuChang, and garbage, pesticides, and fertilizer pollution pose potential threats to water quality. It is therefore necessary to evaluate the quality of groundwater in the study area.

The principle of SPA

To begin with, SPA qualitatively analyses samples based on a certainty and uncertainty analysis, and then quantitatively evaluates groundwater quality by calculating the connectivity, so as to realize the best use of groundwater resources (Feng et al. 2014; Yang et al. 2014). It is assumed that there are N evaluation indices of groundwater quality, among which, the number of the evaluation indices superior to the required standard is S, while that inferior thereto is P. The number of evaluation indices that are not detected or compared is F. While evaluating groundwater quality by SPA, each index and evaluation standard in the evaluation area is supposed to be combined to construct a set pair (Zou et al. 2013). Then, the expression of their connectivity in this area is given by: 
formula
1
where i and j are the markers of the uncertainty of difference and their degree of opposition.
Assume that a=S/N, b=F/N, and c=P/N, then a, b, and c represent the similarity, the uncertainty of any difference, and the extent thereof, respectively. Then, formula (1) can be simplified to: 
formula
2
where a, b, and c all satisfy the normalisation condition, namely, . In accordance with SPA, the similarity a and the difference c in formula (2) is certain, while the difference b is uncertain. Based on the size relationship between a, b and c, and the aforementioned quantitative analysis, the quality of groundwater can be evaluated. Furthermore, the quantitative relationship between the value of the evaluation indices and the grading standard used for evaluating groundwater quality was analysed. It can be seen that the groundwater in different areas has different qualities due to the different values of the evaluation indices even if they belong to the same grade. Therefore, SPA can further be performed on the grading standard in terms of similarity, difference, and opposition.
According to the selected evaluation indices of groundwater quality, it can be seen that all the evaluation indices belong to the cost index category, i.e. the smaller the better. The evaluation index is given as: 
formula
3
where s1, s2, and s3 represent the threshold values of the evaluation indices, and are regarded as the basis for determining the values of the similarity, difference, and opposition in the expression of the connectivity of SPA; x indicates the actual groundwater quality in each evaluation area (to be estimated), while m denotes the mth sample of groundwater quality that remains to be evaluated, and k refers to the kth evaluation index. Based on the calculated results from formula (1), the quality of the groundwater to be evaluated can be ranked and classified accordingly. The average connectivity μ̄m of sample m was obtained by averaging the results from formula (3). The average connectivity is expressed as: 
formula
4
As for sample m, the values of a, b, and c in μm were compared such that p=max{a, b, c}. Then, the grade of groundwater quality was determined according to the grade of the sample.

SPA based on the improved multi-element connectivity

As an example, the connectivity splits the state space in which the research object lies into three parts. Although this represents high precision for many problems, it has some drawbacks because it is too inaccurate to simply split the state space into three parts. Therefore, the connectivity can be studied at different layers according to different conditions. For example, if b is subdivided at a deeper level, formula (2) can be expanded to: 
formula
5
and when n = 2, formula (5) may be written as: 
formula
6
For convenience, formula (6) can be written as: 
formula
7

Meanwhile: a ∈ [0, 1], b ∈ [0, 1], c ∈ [0, 1], d ∈ [0, 1], and a+b+c+d = 1, i ∈ [0, 1], j ∈ [−1, 0], k = −1; i, j and k are applied as markers when their values are not taken into account. Meanwhile, a, b, c, and d are called the similarity, positive difference, negative difference, and opposition respectively. The aforementioned formula (7) gives the four-element connectivity and, in similar fashion, the n (n= 5, 6)-element connectivity degree can be obtained.

SPA based on the IFCD for evaluating groundwater quality

In accordance with the commonly used Quality Standard for Groundwater (GB/T 14848-93), groundwater quality can be divided into five classes (1 to 5). According to the five grades of single-factor indices classified by the Quality Standard for Groundwater, while evaluating groundwater quality using SPA, each index and evaluation standard in the evaluation area is supposed to be combined to construct a set pair H. The evaluation indices that conform to the I-class standard, and accord or exceed the V-class standard, are the bases for determining the values of the similarity and opposition, respectively: those in classes II, III, and IV are considered as the basis for determining the difference. Thereinto, difference can further be subdivided into the identical difference degree (class II), difference degree (class III) and opposite difference degree (class IV). That is to say, the connectivity of the regional groundwater quality is portrayed using IFCD. The portrayed connectivity is formulated as follows: 
formula
8
where m is the mth water sampling point to be evaluated, while N represents the total number of evaluation indices, S, F, P, Q, and T indicate the numbers of evaluation indices that conform to the standard of each of class I, class II, class III, class IV and class V, respectively.
IFCD is an expansion of the three-element connectivity: in formula (8), a, b, c, d, and e are called the connectivity components, and thereinto, a ∈ [0, 1], b ∈ [0, 1], c ∈[0, 1], d ∈[0, 1], and e ∈ [0, 1]. The coefficients of b, c, and d divide the interval [−1, 1] into three subintervals including i∈ [0.333, 1], j ∈ [−0.333, 0.333], and k ∈ [−1, −0.333] based on the method of equal division. The connectivity coefficients including i, j, k, and l are merely used as markers when their values do not need to be considered. Otherwise, they represent a gain effect and an attenuation effect on a, which fully displays the unity of opposites between the connectivity number and various connectivity components. Based on SPA, groundwater quality can be analysed by analysing the size relationship among a, b, c, d, and e in formula (8). Thereafter, the quantitative relationship between the value of a single evaluation index and the grading standard of groundwater quality was analysed. It can be seen from the analysis that different evaluation indices varied even if they belong to the same grade. Therefore, the membership function of each evaluation index relative to the grading standard of evaluation was constructed to describe its quantitative relationship with the standard similarity, difference, and opposition. The indices in the Quality Standard for Groundwater all belong to the ‘smaller the better’ type. Thereto, the connectivity of each evaluation index relative to the grading standard is: 
formula
9
where s1, s2, s3, and s4 represent the limiting standard values of class I, class II, class III, and class IV, respectively; x indicates the measured index value of the water quality in each sampling point which remains to be evaluated, and m and k denote the mth sampling point to be evaluated and the kth evaluation index, separately. The calculation results in formula (9) were averaged and the average connectivity of each evaluation sample was obtained from: 
formula
10
where n is the number of evaluation indices. After formula (8) was processed using formula (4), the connectivity of set pair μm of the sampling point to be evaluated could be obtained by normalising formula (8). Thereby, the composite water quality grade at the mth sampling point is given by: 
formula
11
where Gm represents the composite water quality grade.

CASE STUDY

XuChang is located in central Henan Province, and to the south of the Yellow River (113°03′ to 114°19′ E and 33°46′ to 34°24′ N). Thirteen representative water quality samples were selected from the groundwater in XuChang for evaluation (Figure 1).
Figure 1

The distribution of sampling points.

Figure 1

The distribution of sampling points.

Thirty-one indices were included in the evaluation of groundwater quality. Meanwhile, by considering the influence of environmental pollution on the groundwater, seven evaluation factors including total hardness, chloride, sulphate, nitrate, nitrite, and solids contents were selected (Table 1). The Quality Standard for Groundwater (GB/T14848-93) was applied as the evaluation standard.

Table 1

Sampling point data

Serial numberSampling pointClSO42−NO3NO2FTotal hardnessSolids
G1 Lizhuangqiao 226.17 243.99 1.826 0.044 0.9 653.5 1,213.5 
G2 Wuwan 101.03 98.463 82.55 0.052 1.24 776.5 937.48 
G3 Wangdian 226.17 92.22 50.25 0.24 0.52 715 922.54 
G4 Fentaicun 727.08 234.39 230.7 0.02 0.58 1,369 2,248.47 
G5 Eastern Guojicun 245.67 276.17 130.25 0.012 0.72 752 1,488.02 
G6 Huangtun 419.02 86.45 232.5 0.064 0.76 1,060.5 1,584.57 
G7 Mapengyang 337.13 210.85 67.48 1.2 0.74 1,107 1,547.64 
G8 Dahuzhuang 38.64 62.92 66.65 0.004 0.58 493 657.81 
G9 Zhusi 288.92 269.55 214.75 0.008 0.64 1,060.5 1,473.25 
G10 Zushimiao 103.87 317.48 1.77 0.001 1.68 289.5 1,197.62 
G11 Guozhuang 204.19 216.62 19.89 0.26 0.82 616.5 1,171.44 
G12 Northern Changqu 57.78 198.84 0.19 0.001 209.5 1,031.92 
G13 Eastern Shenzhuang 57.78 74.45 20.56 0.004 1.04 357.5 654.49 
Serial numberSampling pointClSO42−NO3NO2FTotal hardnessSolids
G1 Lizhuangqiao 226.17 243.99 1.826 0.044 0.9 653.5 1,213.5 
G2 Wuwan 101.03 98.463 82.55 0.052 1.24 776.5 937.48 
G3 Wangdian 226.17 92.22 50.25 0.24 0.52 715 922.54 
G4 Fentaicun 727.08 234.39 230.7 0.02 0.58 1,369 2,248.47 
G5 Eastern Guojicun 245.67 276.17 130.25 0.012 0.72 752 1,488.02 
G6 Huangtun 419.02 86.45 232.5 0.064 0.76 1,060.5 1,584.57 
G7 Mapengyang 337.13 210.85 67.48 1.2 0.74 1,107 1,547.64 
G8 Dahuzhuang 38.64 62.92 66.65 0.004 0.58 493 657.81 
G9 Zhusi 288.92 269.55 214.75 0.008 0.64 1,060.5 1,473.25 
G10 Zushimiao 103.87 317.48 1.77 0.001 1.68 289.5 1,197.62 
G11 Guozhuang 204.19 216.62 19.89 0.26 0.82 616.5 1,171.44 
G12 Northern Changqu 57.78 198.84 0.19 0.001 209.5 1,031.92 
G13 Eastern Shenzhuang 57.78 74.45 20.56 0.004 1.04 357.5 654.49 

Water quality as estimated by IFCD-based SPA

The connectivity of each water quality sampling point

The four boundaries for water quality between classes I to II were regarded as the basis for determining the values of similarity, difference, and opposition. Taking observation point G1 as an example, two evaluation factors including NO3 and Fwere lower than the standard of class I, while there was no evaluation factor ranging between class I and class II. Meanwhile, Cl and SO42− were in the range of class II to class III and NO2 and total dissolved solids ranged between classes III and IV. There was one evaluation index with its total hardness higher than the standard of class IV. Furthermore: 
formula
Then, the connectivity of water sample G1 is given by: 
formula
Similarly, the connectivities of the other 12 sampling points are expressed as: 
formula

Finding the average connectivity

The connectivity of each evaluation index relative to each sample of groundwater quality to be evaluated was calculated using formula (9), and then formula (4) was used to calculate the average connectivity, as shown in Table 2.

Table 2

The connectivity of each sampling point and the mean value

Connectivityμ1
μ2
abcdeabcde
Cl 0.238 0.762 0.490 0.510 
SO42− 0.06 0.94 0.515 0.485 
NO3 
NO2 0.7 0.3 0.6 0.4 
F 0.76 0.24 
Total hardness 
Solids 0.787 0.213 0.125 0.875 
Mean value 0.29 0.043 0.456 0.073 0.14 0.144 0.16 0.319 0.09 0.29 
  μ3μ4
Connectivityabcdeabcde
Cl 0.238 0.762 
SO42− 0.578 0.422 0.156 0.844 
NO3 
NO2 
F 
Total hardness 
Solids 0.155 0.845 
Mean value 0.225 0.116 0.23 0.0 0.43 0.14 0.022 0.263 0.0 0.57 
 μ5μ6
Connectivityabcdeabcde
Cl 0.043 0.957 
SO42− 0.738 0.262 0.636 0.364 
NO3 
NO2 0.8 0.2 0.45 0.55 
F 
Total hardness 
Solids 0.512 0.488 0.415 0.585 
Mean value 0.14 0.12 0.344 0.107 0.29 0.234 0.052 0.124 0.162 0.43 
 μ7μ8
Connectivityabcdeabcde
Cl 0.129 0.871 
SO42− 0.392 0.608 0.871 0.129 
NO3 
NO2 0.667 0.333 
F 
Total hardness 0.57 0.43 
Solids 0.452 0.548 0.684 0.316 
Mean value 0.14 0.056 0.17 0.203 0.43 0.505 0.164 0.127 0.06 0.14 
 μ9μ10
Connectivityabcdeabcde
Cl 0.611 0.389 0.461 0.539 
SO42− 0.805 0.195 0.325 0.675 
NO3 
NO2 0.222 0.778 
F 0.32 0.68 
Total hardness 0.07 0.93 
Solids 0.527 0.473 0.802 0.198 
Mean value 0.175 0.111 0.278 0.151 0.29 0.362 0.21 0.207 0.222 0.0 
 μ11μ12
Connectivityabcdeabcde
Cl 0.458 0.542 0.922 0.078 
SO42− 0.334 0.666 0.512 0.488 
NO3 0.007 0.993 
NO2 
F 
Total hardness 0.603 0.397 
Solids 0.829 0.171 0.968 0.032 
Mean value 0.14 0.114 0.433 0.024 0.29 0.504 0.141 0.208 0.147 0.0 
 μ13
Connectivityabcdeabcde
Cl 0.922 0.078 F 0.96 0.04 
SO42− 0.756 0.244 Total hardness 0.617 0.383 
NO3 0.944 0.056 Solids 0.691 0.309 
NO2 0.842 0.158 Mean value 0.36 0.255 0.371 0.014 0.0 
Connectivityμ1
μ2
abcdeabcde
Cl 0.238 0.762 0.490 0.510 
SO42− 0.06 0.94 0.515 0.485 
NO3 
NO2 0.7 0.3 0.6 0.4 
F 0.76 0.24 
Total hardness 
Solids 0.787 0.213 0.125 0.875 
Mean value 0.29 0.043 0.456 0.073 0.14 0.144 0.16 0.319 0.09 0.29 
  μ3μ4
Connectivityabcdeabcde
Cl 0.238 0.762 
SO42− 0.578 0.422 0.156 0.844 
NO3 
NO2 
F 
Total hardness 
Solids 0.155 0.845 
Mean value 0.225 0.116 0.23 0.0 0.43 0.14 0.022 0.263 0.0 0.57 
 μ5μ6
Connectivityabcdeabcde
Cl 0.043 0.957 
SO42− 0.738 0.262 0.636 0.364 
NO3 
NO2 0.8 0.2 0.45 0.55 
F 
Total hardness 
Solids 0.512 0.488 0.415 0.585 
Mean value 0.14 0.12 0.344 0.107 0.29 0.234 0.052 0.124 0.162 0.43 
 μ7μ8
Connectivityabcdeabcde
Cl 0.129 0.871 
SO42− 0.392 0.608 0.871 0.129 
NO3 
NO2 0.667 0.333 
F 
Total hardness 0.57 0.43 
Solids 0.452 0.548 0.684 0.316 
Mean value 0.14 0.056 0.17 0.203 0.43 0.505 0.164 0.127 0.06 0.14 
 μ9μ10
Connectivityabcdeabcde
Cl 0.611 0.389 0.461 0.539 
SO42− 0.805 0.195 0.325 0.675 
NO3 
NO2 0.222 0.778 
F 0.32 0.68 
Total hardness 0.07 0.93 
Solids 0.527 0.473 0.802 0.198 
Mean value 0.175 0.111 0.278 0.151 0.29 0.362 0.21 0.207 0.222 0.0 
 μ11μ12
Connectivityabcdeabcde
Cl 0.458 0.542 0.922 0.078 
SO42− 0.334 0.666 0.512 0.488 
NO3 0.007 0.993 
NO2 
F 
Total hardness 0.603 0.397 
Solids 0.829 0.171 0.968 0.032 
Mean value 0.14 0.114 0.433 0.024 0.29 0.504 0.141 0.208 0.147 0.0 
 μ13
Connectivityabcdeabcde
Cl 0.922 0.078 F 0.96 0.04 
SO42− 0.756 0.244 Total hardness 0.617 0.383 
NO3 0.944 0.056 Solids 0.691 0.309 
NO2 0.842 0.158 Mean value 0.36 0.255 0.371 0.014 0.0 

The average connectivity was used to process the expression for the connectivity. The coefficients were then normalised to estimate the water quality grade at each observation point. Estimated results are shown in Table 3.

Table 3

Results evaluated by SPA

Serial numberG1G2G3G4G5G6G7G8G9G10G11G12G13
Grade III III III II 
Serial numberG1G2G3G4G5G6G7G8G9G10G11G12G13
Grade III III III II 

Evaluation of water quality results obtained using fuzzy comprehensive evaluation

The main steps in the fuzzy evaluation process are as follows:

  • (1)

    Establishing the factor set: the indices having the greatest influence on water quality were selected to establish the factor set of evaluation indices. Thereinto, the factors affecting water quality varied across different regions.

  • (2)

    Determining the comment set of evaluation: the five classes of water quality were used to represent excellent, fine, good, poor, and extremely poor water quality respectively.

  • (3)

    Establishing the membership function and constructing the fuzzy relationship matrix: the curve of the distribution form of the function was used to solve the membership function. As a general rule, the descending and ascending semi-trapezoid forms were used to obtain the membership degree of the grades at each end of the distribution, while a symmetrical triangular form was used to acquire the degree of intergrade memberships. Then, the fuzzy matrix was determined according to the number of evaluation indices, the grading standard of water quality, and its membership degree in each grade.

  • (4)

    Determining the weights: in the comprehensive evaluation, indices are supposed to have different influences on environmental pollution and public health considering the different levels of each single index. Therefore, their influences on the overall level of pollution were also different. When the environment is influenced by multiple factors, any collaborative and antagonistic effects between factors are supposed to be reflected. The super-weighting method of pollutant concentration can be used here.

  • (5)

    A compound operation was performed on the weight matrix of evaluation indices and the fuzzy relation matrix. In this way, the degree matrix B that the evaluation index subordinates to each evaluation grade was obtained. According to the maximum extent of membership, the grade corresponding to the maximum value was the grade of the evaluation index.

Fuzzy comprehensive evaluation, simply speaking, aims to evaluate the correlation between factors related to the object being evaluated using the principle of fuzzy transformation and the maximum subordination principle. Here, the principle of fuzzy transformation can be represented thus: 
formula
12
where A represents the weight matrix of evaluation factors, while R indicates the fuzzy relationship matrix constituted by the evaluation matrices of various single factors, and B denotes the comprehensive evaluation result obtained through the compound operation of A and R. Since the results obtained by the matrix model varied, the groundwater quality results were determined based on the principle of taking the maximum value. That is to say, the final result was represented by the grade of the maximum value in B. According to measured data in the research area, the weight matrix A can be calculated (Table 4).
Table 4

The weight matrix A

A1 0.1484 0.1600 0.0168 0.1763 0.0945 0.2365 0.1676 
A2 0.0404 0.0394 0.4635 0.1270 0.0794 0.1714 0.0790 
A3 0.0715 0.0292 0.2231 0.4636 0.0263 0.1248 0.0614 
A4 0.1288 0.0415 0.5738 0.0216 0.0164 0.1339 0.0839 
A5 0.0752 0.0845 0.5596 0.0224 0.0353 0.1270 0.0959 
A6 0.0806 0.0166 0.6276 0.0752 0.0234 0.1125 0.0642 
A7 0.0341 0.0213 0.0959 0.7419 0.0120 0.0618 0.0330 
A8 0.0247 0.0402 0.5978 0.0156 0.0593 0.1738 0.0885 
A9 0.0626 0.0584 0.6525 0.0106 0.0222 0.1267 0.0671 
A10 0.0917 0.2802 0.0219 0.0054 0.2373 0.1410 0.2225 
A11 0.0679 0.0721 0.0929 0.5283 0.0437 0.1132 0.0821 
A12 0.0629 0.2165 0.0029 0.0067 0.3485 0.1259 0.2366 
A13 0.0610 0.0786 0.3048 0.0258 0.1758 0.2084 0.1456 
A1 0.1484 0.1600 0.0168 0.1763 0.0945 0.2365 0.1676 
A2 0.0404 0.0394 0.4635 0.1270 0.0794 0.1714 0.0790 
A3 0.0715 0.0292 0.2231 0.4636 0.0263 0.1248 0.0614 
A4 0.1288 0.0415 0.5738 0.0216 0.0164 0.1339 0.0839 
A5 0.0752 0.0845 0.5596 0.0224 0.0353 0.1270 0.0959 
A6 0.0806 0.0166 0.6276 0.0752 0.0234 0.1125 0.0642 
A7 0.0341 0.0213 0.0959 0.7419 0.0120 0.0618 0.0330 
A8 0.0247 0.0402 0.5978 0.0156 0.0593 0.1738 0.0885 
A9 0.0626 0.0584 0.6525 0.0106 0.0222 0.1267 0.0671 
A10 0.0917 0.2802 0.0219 0.0054 0.2373 0.1410 0.2225 
A11 0.0679 0.0721 0.0929 0.5283 0.0437 0.1132 0.0821 
A12 0.0629 0.2165 0.0029 0.0067 0.3485 0.1259 0.2366 
A13 0.0610 0.0786 0.3048 0.0258 0.1758 0.2084 0.1456 

The fuzzy matrix R of each sampling point was calculated, and then the coincidence operation was performed on the weight matrix A and the fuzzy matrix R, and the comprehensive evaluation table for water quality at each sampling point was then found (Table 5).

Table 5

Comprehensive evaluation table of water quality parameters

Sampling pointabcdeGrade
G1 0.111 0.045 0.519 0.089 0.237 III 
G2 0.040 0.050 0.206 0.070 0.635 
G3 0.043 0.039 0.106 0.000 0.812 
G4 0.016 0.006 0.057 0.000 0.920 
G5 0.035 0.021 0.188 0.069 0.687 
G6 0.034 0.006 0.060 0.079 0.821 
G7 0.012 0.008 0.032 0.048 0.900 
G8 0.129 0.071 0.427 0.075 0.298 III 
G9 0.025 0.008 0.121 0.067 0.779 
G10 0.079 0.381 0.345 0.195 0.000 II 
G11 0.044 0.056 0.445 0.114 0.341 III 
G12 0.143 0.466 0.235 0.156 0.000 II 
G13 0.137 0.457 0.381 0.024 0.000 II 
Sampling pointabcdeGrade
G1 0.111 0.045 0.519 0.089 0.237 III 
G2 0.040 0.050 0.206 0.070 0.635 
G3 0.043 0.039 0.106 0.000 0.812 
G4 0.016 0.006 0.057 0.000 0.920 
G5 0.035 0.021 0.188 0.069 0.687 
G6 0.034 0.006 0.060 0.079 0.821 
G7 0.012 0.008 0.032 0.048 0.900 
G8 0.129 0.071 0.427 0.075 0.298 III 
G9 0.025 0.008 0.121 0.067 0.779 
G10 0.079 0.381 0.345 0.195 0.000 II 
G11 0.044 0.056 0.445 0.114 0.341 III 
G12 0.143 0.466 0.235 0.156 0.000 II 
G13 0.137 0.457 0.381 0.024 0.000 II 

ANALYSIS OF THE WATER QUALITY EVALUATION RESULTS

SPA and fuzzy evaluation were used to evaluate the 13 groups of groundwater quality data taken from XuChang. As shown in Table 6, nine sampling points showed the same evaluation result, while four sampling points had different evaluation results (Figure 2). Four groundwater samples were mainly classes I and II, which meet the Drinking Water Quality of China, with nine samples of classes III and V which required pretreatment to make them drinkable. The environment in this water source region is seriously polluted by industrial wastes, slag, domestic waste, and so on which are stored around water sources.
Table 6

Comparison of evaluation results

Serial numberSampling pointSPAFuzzy comprehensive evaluation
G1 Lizhuangqiao III III 
G2 Wuwan 
G3 Wangdian 
G4 Fentaicun 
G5 Eastern Guojicun III 
G6 Huangtun 
G7 Mapengyang 
G8 Dahuzhuang III 
G9 Zhusi 
G10 Zushimiao II 
G11 Guozhuang III III 
G12 Northern Changqu II 
G13 Eastern Shenzhuang II II 
Serial numberSampling pointSPAFuzzy comprehensive evaluation
G1 Lizhuangqiao III III 
G2 Wuwan 
G3 Wangdian 
G4 Fentaicun 
G5 Eastern Guojicun III 
G6 Huangtun 
G7 Mapengyang 
G8 Dahuzhuang III 
G9 Zhusi 
G10 Zushimiao II 
G11 Guozhuang III III 
G12 Northern Changqu II 
G13 Eastern Shenzhuang II II 
Figure 2

Comparison of evaluation results (the red and green characters represent the results obtained by SPA and fuzzy evaluation, respectively).

Figure 2

Comparison of evaluation results (the red and green characters represent the results obtained by SPA and fuzzy evaluation, respectively).

CONCLUSIONS

The use of set pair analyses based on the improved five-element connectivity and fuzzy comprehensive evaluation analysed the groundwater quality in XuChang to be endowed with the following characteristics.

  • (1)

    SPA and fuzzy comprehensive evaluation were used to evaluate groundwater quality in the paper. The groundwater quality of different regions at the same classification may be different because of the different evaluation indicators. The SPA method is simple and practical, and its outcomes are objective and stable. So, it is an effective evaluation method of water quality. The main reason is that the samples were firstly classified preliminarily by calculating the degree of relation between the samples and evaluation indicators. Then the samples were ranked further through similarity, difference and opposition SPA.

  • (2)

    The groundwater quality at the observation points located in the lower reaches is poorer than that of the upper reaches. This is because the surface water in the city is polluted to varying degrees, and there is hydraulic connectivity between the surface water and groundwater. The infiltration of surface water pollutes the shallow groundwater near both banks along the area's rivers. As a result, the surface water in the lower reaches is more heavily polluted than that in the upper reaches, and therefore the groundwater in the lower reaches is of poor quality.

  • (3)

    The groundwater was mainly classes III and V, and this does not meet the Drinking Water Quality of China, which requires pretreatment to make it drinkable. A small number of class I and II samples meet the Drinking Water Quality of China. The environment in this water source region is seriously polluted by industrial waste, slag, domestic waste, and so on which are stored around water sources. As a result, hazardous substances permeate underground to pollute shallow groundwater through decomposition and loss due to weathering and rainfall.

Based on the comprehensive evaluation of the groundwater quality in XuChang, the following suggestions regarding the development, utilisation, and protection thereof are proposed.

(1) Restriction and suspension of the exploitation of the groundwater in urban areas.

To protect groundwater resources in urban areas and improve the quality of drinking water for residents, the municipal government in XuChang issued a document on 26 November 2011: the government would close artisanal and private bores and wells belonging to various categories of businesses and individuals in the planning area. Meanwhile, such wells with excess exploitation of groundwater and drilled bores in mixed layers should be limited in terms of depth or served with a cessation of abstraction notice.

(2) Promoting the policy of water diversion, recharge, and improving water quality.

If pollution sources were eradicated, pollutants could be gradually diluted and purified after a period of time under the influence of the natural supply and motion of groundwater, as well as their adsorption by the local geological formations. However, this self-purification process is slow as groundwater runoffs are obstructed. Therefore, groundwater artificial recharge is adopted so as to greatly accelerate the dilution and purification processes by conducting recharge of surface water that reaches the requisite standard into groundwater. The XuChang Water Conservancy Bureau began to implement an action plane for protecting the groundwater in urban areas in June 2004. As to the specific steps taken, a project diverting water and recharging the groundwater to purify water sources from three vertical and four horizontal routes is planned for excavation in DongCheng District. It is suggested that this be further accelerated in the name of groundwater protection.

ACKNOWLEDGEMENTS

This study was financially supported by the National Natural Science Foundation of PR China (No. 41402225), and the Non-Profit Industry Specific Research Projects of the Ministry of Water Resources, China, Grant (No. 201401041).

REFERENCES

REFERENCES
Dahiya
S.
Singh
B.
Gaur
S.
Garg
V. K.
Kushwaha
H. S.
2007
Analysis of groundwater quality using fuzzy synthetic evaluation
.
Journal of Hazardous Materials
147
(
3
),
938
946
.
Faruk
D. O.
2010
A hybrid neural network and ARIMA model for water quality time series prediction
.
Engineering Applications of Artificial Intelligence
23
(
4
),
586
594
.
Feng
L. H.
Sang
G. S.
Hong
W. H.
2014
Statistical prediction of changes in water resources trends based on set pair analysis
.
Water Resources Management
28
(
6
),
1703
1711
.
Li
L.
Yong
C.
Long-Jun
X.
Li-Qiu
J.
Ning
M.
Ming
L.
2015
An integrated method of set pair analysis and association rule for fault diagnosis of power transformers
.
IEEE Transactions on Dielectrics and Electrical Insulation
22
(
4
),
2368
2378
.
Maiti
S.
Erram
V. C.
Tiwari
R. K.
Kulkarni
U. D.
Sangpal
R. R.
2013
Assessment of groundwater quality: a fusion of geochemical and geophysical information via Bayesian neural networks
.
Environmental Monitoring and Assessment
185
(
4
),
3445
3465
.
Pathak
H.
Limaye
S. N.
2011
Study of seasonal variation in groundwater quality of Sagar city (India) by principal component analysis
.
Journal of Chemistry
8
(
4
),
2000
2009
.
Ranjan
R. K.
Ramanathan
A. L.
Purushothaman
P.
Kumar
A.
2012
Hydrochemical characteristics of groundwater in the plains of Phalgu River in Gaya, Bihar, India
.
Arab J. Geo. Sci.
6
,
3257
3267
.
Wang
M.
Xu
X.
Li
J.
Shen
F.
2015
A novel model of set pair analysis coupled with extenics for evaluation of surrounding rock stability
.
Mathematical Problems in Engineering
2015
,
892549
.
Yang
X. H.
He
J.
Di
C. L.
Li
J. Q.
2014
Vulnerability of assessing water resources by the improved set pair analysis
.
Thermal Science
18
(
5
),
1531
1535
.
Zou
Q.
Zhou
J.
Zhou
C.
Song
L.
Guo
J.
2013
Comprehensive flood risk assessment based on set pair analysis-variable fuzzy sets model and fuzzy AHP
.
Stochastic Environmental Research and Risk Assessment
27
(
2
),
525
546
.