Dynamic successive assessment method of water environment carrying capacity and its application

Ecological Indicators 52 (2015) 134–146

Contents lists available at ScienceDirect

Ecological Indicators journal homepage: www.elsevier.com/locate/ecolind

Original Research Paper

Dynamic successive assessment method of water environment carrying capacity and its application Tianxiang Wang 1 , Shiguo Xu ∗ Faculty of Infrastructure Engineering, Dalian University of Technology, Dalian 116024, China

a r t i c l e

i n f o

Article history: Received 28 May 2014 Received in revised form 18 September 2014 Accepted 1 December 2014 Keywords: Water environment carrying capacity Temporal and spatial variability Dynamic Assessment method Class

a b s t r a c t Water environment carrying capacity (WECC) is an important foundation of sustainable socioeconomic development and may be affected by many factors such as water resources, water quality, economy, population and environmental protection. This article focuses on the temporal and spatial variability of WECC to explore a method of dynamic successive assessment. First, the Pressure-State-Response (PSR) framework is used to develop a systematic and causal indicator system representing the three aspects of water environment pressure carrying capacity (WEPCC), water environment state carrying capacity (WESCC) and water environment response carrying capacity (WERCC). The Variable Fuzzy Pattern Recognition (VFPR) model and an analytic hierarchy process (AHP) model are combined to successively and dynamically assess WEPCC, WESCC and WERCC, and after that the weighting method is used to calculate WECC. Furthermore, WECC is divided into 27 classes on the basis of WEPCC, WESCC and WERCC contributions. The proposed method is applied to the dynamic successive assessment of WECCs in China, including inter-province comparisons. The results show that the dynamic successive WECC assessment method is reasonable, and it can be used not only to accurately understand the changes of WECC through time but also to distinguish qualitative differences masked by similar WECC values. © 2014 Elsevier Ltd. All rights reserved.

1. Introduction The World Commission on Environment and Development (WCED) (1987) put forth the concept of sustainable development: to satisfy current needs without compromising the ability of future generations to satisfy their own needs. Since then, the harmonious development of society, economy and the environment has been a key issue for regional sustainable development (Arrow et al., 1995; Bouwer, 2002; DuC´ and Urbaniec, 2012). The concept of carrying capacity originated from ecology and has given rise to a series of concepts and measures now used in sustainability assessments, such as water environment carrying capacity (Zhu et al., 2010; Na and Wang, 2011), water resources carrying capacity (Feng and Huang, 2008; Li and Jin, 2009; Meng et al., 2009), soil carrying capacity (Johnson et al., 2011) and population carrying capacity (Shi et al., 2013). The assessment of WECC is an important method to research sustainable development of social economy and environment (Lu et al., 2011). At present, water environment crises have

∗ Corresponding author. Tel.: +86 013841125031; fax: +86 0411 84707680. E-mail addresses: [email protected] (T. Wang), [email protected] (S. Xu). 1 Tel: +86 015382172713; fax: +86 0411 84707680. http://dx.doi.org/10.1016/j.ecolind.2014.12.002 1470-160X/© 2014 Elsevier Ltd. All rights reserved.

become a significant issue for social development because both socioeconomic development and improvement of the human living environment require the quality of the water environment to be improved. Since the 1990s, researchers have focused on studying the relation between the water environment and social economy. It is noted that WECC, which consists of two aspects, i.e., water quality and quantity (Li et al., 2011), exhibits spatio-temporal variability due to both social development and environmental change (Li et al., 2011; Na and Wang, 2011). This coupling of human and natural systems makes WECC outcomes complicated and uncertain (Huang and Qin, 2008). Moreover, it has the character of threshold and variation (Na and Wang, 2011; Gao et al., 2012). There is ongoing debate about how to define WECC. Majority of researchers think that WECC should be defined in terms of the capacity to support socioeconomic development and that its concept should comprise many aspects, such as water resources, water quality, economy, population and environmental protection (Guo and Tang, 1995; Tang et al., 1997; Chen et al., 2000; Na and Wang, 2011). Others think of WECC more narrowly, in sole terms of the processes and capacities of aquatic systems (Gao et al., 2012). Previous studies have indicated that dynamic assessment has the advantages of succession and accuracy (Feng et al., 2010). However, most current studies on WECC are inconsecutive and partial assessments (Feng and Huang, 2008; Li and Jin, 2009; Na and Wang,

T. Wang, S. Xu / Ecological Indicators 52 (2015) 134–146

2011). In reality, WECC is dynamically affected by the coupling of socioeconomic changes with changes in the water environment. In addition, the traditional assessments of WECC are powerless to interpret similar values of WECC in different times or areas, which may be characterized by different economic levels, development models and other factors. Although inconsecutive or partial assessments contribute to understanding WECC, it is difficult for them to reflect trends over time and differences between regions. Therefore, it is urgent to explore a method of dynamically and successively assessing WECC to identify its temporal trends and regional differences. In this context, WECC mapping is a more efficient method for analyzing changes of WECC because it contains more information. Fortunately, geostatistical methods have the advantage of producing maps (El-Fadel et al., 2014) and are convenient for processing the data (Abdideh and Ghasemi, 2014). These tools contribute to the analysis of temporal and spatial variability and have been applied to many issues, such as analyzing the evaluation results of the crop production system in the 31 provinces of China (Tao et al., 2013), and studying the environmental carrying capacity of the Bohai sea rim area in China (Lu et al., 2011). The objective of this paper is to explore the method of dynamic successive assessment and its potential application for the study of the temporal and spatial variability of WECC, which can be used not only to accurately understand the changes of WECC in various areas but also to give more informative interpretations of WECC values that happen to be similar.

2. Dynamic successive assessment method of WECC 2.1. Developing indicator system for WECC using the PSR framework Developing an indicator system is an important step in WECC assessment. The PSR framework shows causal relationships between pressure, state and response indicators (OECD, 1998), and because it systematically represents important indicators of sustainable development in a causal manner (Wang et al., 2010), it has been widely used in various types of assessments, e.g., of water resources’ carrying capacity, environmental impact and sustainable development. This paper considers WECC in terms of the capacity to support socioeconomic development and as a coupled human-natural system. Rapid socioeconomic development will increase water environment pressure (WEP) and degrade water environment state (WES). In turn, these changes will restrict socioeconomic development. Under such circumstances, implementing water environment response (WER) measures to reduce WEP and improve WES will enhance WECC, which forms a virtuous cycle that promotes socioeconomic development, and vice versa. Furthermore, WEP represents the direct factors that degrade WECC, mainly water consumption and pollutant discharge resulting from population increase and socioeconomic development (Feng et al., 2009). WES is the core of WECC and represents the potential for water quality and quantity to support socioeconomic development (Zhou et al., 2011). Forest coverage is an important factor influencing the cycling and purification of water (Stasik et al., 2011). The exploitation and utilization ratio of water resources can indicate the degree to which water resources are consumed and exploited. Rivers not only are important water resources but also greatly influence water quality and quantity in lakes and reservoirs (Wu et al., 2012). Rainfall is uncertain and is significantly correlated with water quantity (Huang and Qin, 2008; Feng et al., 2009). WECC also can be improved by WER in two types of ways. One such way is to improve WES directly by increasing investment in environmental protection and the ratio of ecological water consumption (Zhou et al., 2011). The other is to improve WES indirectly by reducing

135

WEP through strengthening scientific research, promoting adjustments in industrial systems and practices, and decreasing industrial and agricultural water consumption (Zhou et al., 2011). 17 indicators were selected for the construction of the gross WECC index; each is closely related to WECC and has previously been used in the literature (Duan et al., 2009; Feng et al., 2009; Zhou et al., 2011). These indicators are further divided into three subsystems, i.e., WEPCC, WESCC and WERCC, following the PSR framework. The details are shown in Table 1. For each indicator, five grades are developed to judge the level of carrying capacity based on literature. Grades 1–2 of the carrying capacity are at a fine level, grades 2–3 of the carrying capacity are at an acceptable level and grades 3–5 of the carrying capacity are at a poor level.

2.2. Assessment method based on VFPR and AHP model The assessment of WECC can be regarded as the problem of grading each sample with respect to every indicator. The process of comparing the sample indicators with indicator standards has an imprecise character, so the Variable Fuzzy Pattern Recognition (VFPR) model is a better choice for the dynamic successive assessment of WECC. VFPR theory was presented by Professor Chen and is developed from the theory of variable fuzzy sets (Chen and Guo, 2006; Chen, 2009). This theory grades samples by calculating a synthetic relative membership degree in each grade for each sample. This process is more reliable than a definite assignment of grade (Zhou et al., 2009; Wang et al., 2011; Ke and Zhou, 2013). VFPR has been successfully and widely applied to many different problems, such as water resources evaluation (Duan et al., 2009), water renewal assessment (Chen and Guo, 2006), and groundwater evaluation (Zhou et al., 2009). This paper explores a dynamic successive assessment method of the WECC based on VFPR model and AHP model. In the first step, Eqs. (1) and (2) are used to normalize (rij , shj ) the indicators (xij ) and standards (yhj ) so as to remove the influence of inverse indices and different dimensions respectively.

⎧ 0 ⎪ ⎨ rij =

ycj − xij

y −y ⎪ ⎩ cj 1j 1

⎧ 0 ⎪ ⎨ shj =

ycj − yhj

y −y ⎪ ⎩ cj 1j 1

xij ≤ ycj (positive index), xij ≥ycj (inverse index) positive index or inverse index

(1)

xij ≥y1j (positive index), xij ≤ y1j (inverse index)

yhj = ycj , positive index or inverse index positive index or inverse index

(2)

yhj = y1j , positive index or inverse index

where xij is the value of indicator j of the sample i, i is the number of samples and j is the number of indicators; yhj is the value that defines standard h of indicator j, where h = 1, 2 . . ., c, c representing the highest grade of standard; rij and shj are the results of normalization of the indicators (xij ) and standards (yhj ), respectively; the positive indices (X3, X7, X8, X9, X10, X12, X13, X14 and X17) are those that are positively correlated with carrying capacity; the inverse indices (X1, X2, X4, X5, X6, X11, X15 and X16) are those that are negatively correlated with carrying capacity. In the second step, the judgment matrices used in the AHP (Singh et al., 2006; Hosseini and Kaneko, 2011) are defined in accordance with the relative importance of the different indicators. The local weights of the indicators are then obtained by calculating the eigenvalues and eigenvectors of the judgment matrices. In the third step, Eq. (3) is used to calculate the synthetic relative membership degree for sample i belonging to standard h. Eq. (3) has

136

T. Wang, S. Xu / Ecological Indicators 52 (2015) 134–146

Table 1 Indicator system and standard for WECC. Indicator system

Grades

Subsystems

Indicators

WEPCC (Pressure)

10 Population density (PER/km2 , X1) 200 Water consumption per capita (m3 /PER, X2) 50,000 Per capita Gross Domestic Product (GDP) (Yuan/PER, X3) 1.69 Ammoniacal nitrogen discharge intensity of economic output (t/108 Yuan, X4) 8 Chemical oxygen demand discharge intensity of economic output (t/10 Yuan, X5) 17 3 4 Water consumption intensity of GDP (m /10 Yuan, X6) 80

1

WESCC (State)

Forest coverage (%, X7) Annual precipitation (mm, X8) Proportion of lengths of unpolluted stretches of rivers (%, X9) Water resources per capita (m3 /PER, X10) Exploitation and utilization ratio of water resources (%, X11)

WERCC (Response)

Ratio of investment in Research & Development to GDP (%, X12) Ratio of tertiary industry output to GDP (%, X13) Ratio of investment in environmental pollution control to GDP (%, X14) Water consumption intensity of industrial output (m3 /104 Yuan, X15) Water consumption intensity of agricultural output (m3 /104 Yuan, X16) Ratio of ecological water consumption (%, X17)

four variants, corresponding to choices of a and p; therefore, four results are calculated for each sample.

uhi =

⎧ 0 ⎪ ⎪ ⎪ ⎪ ⎪ ⎨

, 1 ≤ h ≤ ai orc≥h≥bi 1



p a



⎪ p  ω r − s ⎪ j ij hj j=1 ⎪ ⎪

p  ⎪ m ⎩ ωj rij − skj

bi

k=ai

 m

, ai ≤ h ≤ bi

(3)

j=1

where uhi is the synthetic relative membership degree for sample i belonging to standard h; k is the interval (ai , bi ) to which sample i belongs; the ai and bi are obtained by comparing rij with shj , with ai being the minimum level of sample i, and bi being the maximum level of sample i; m is the total number of indicators; ωj is the weight of the indicator j, which is determined by the judgment matrices in the AHP model; a is the optimization criteria parameter, a = 1(linear), a = 2(nonlinear); p is the distance parameters, p = 1(Hamming distance), p = 2(Euclidean distance). In the fourth step, Eq. (4) is used to calculate the characteristic value H of the sample i based on the third step, then use the average value as the assessment result. H=

c 

2

40 1600 80 3000 10 4 70 2.05 25 600 5

3

100 400 35,000 2.88 26 110 30 800 70 2200 40 3 55 1.64 45 800 3

200 600 21,000 3.5 36 250

4

5

400 800 7000 3.81 43 600

600 850 4000 4.7 50 700

25 600 60 1700 50

20 400 50 1000 60

2 40 1.36 70 1200 2

1 30 1.13 110 1500 1

10 200 30 500 65 0.5 25 0.66 150 2000 0.5

assessment with a further classification to study the temporal and spatial variability of WECC. This paper classifies WECC according to the levels of the WEPCC, WESCC and WERCC, with these levels as defined in the previous sections of this paper. All WECCs are divided into 27 classes generated by permuting the WEPCC, WESCC and WERCC to accurately distinguish differences between samples with similar WECC values. The details of the WECC classes and their corresponding limit values can be seen in Table 2. In summary, the dynamic successive assessment method of WECC includes three processes: developing an indicator system for WECC in accordance with a PSR framework, using the method to assess WECC dynamically and classifying every WECC on the basis of its WEPCC, WESCC and WERCC components. In addition, to verify the reasonableness of this model, the results of assessments obtained by the proposed method are compared with those from the fuzzy comprehensive assessment model (Wang et al., 2011) and the original data. The details pertinent to this study are described in the following paragraphs.

3. Method application 3.1. Study site and data

uhi h

(4)

h=1

where h is the grade of standard, with h = 1, 2. . . c, and c is the highest grade of standard; H is the carrying capacity of the sample i. In this way, the WEPCC, WERCC, WESCC can be dynamically and successively calculated by the proposed method, after that this paper uses weighting method to obtain the WECC. 2.3. Classification of WECC WECC is a measure of a coupled human–natural system, affected by both natural resources and human activities. Its variation is driven by the interacting influences of WEPCC, WESCC and WERCC. Thus, WECCs of similar value or level in different times or areas might belong to different classes owing to the differences in economy, environment and policy. The traditional assessment of WECC only yields an isolated value of the comprehensive assessment and often ignores the differences of economy, environment and policy between samples. This paper suggests combining the results of

China is one of 13 countries with the highest water scarcity, and some provinces are troubled by water pollution (Daily and Ehrlich, 1996). Notable recent incidents include nitrobenzene leakage into the Songhua River in 2005 and the blue-algae bloom of Lake Taihu in 2007. Although China is currently the second-largest economy of the world, its water resources per capita are less than a quarter of the world average. Water environment issues, especially water shortages, are starting to limit rapid socioeconomic development (Daily and Ehrlich, 1996). Therefore, it is urgent and imperative to study the supporting capacity of China’s water environments. Previous studies of WECC were isolated and inconsecutive (Li and Jin, 2009; Na and Wang, 2011); while they have guided sustainable development at a local scale, they have also lacked the systematic, dynamic analysis of temporal and spatial variability of the WECC at the local and global scales. This paper uses China as an example to explore the presently described method. The data of China (2003–2010) and 31 provinces (2003–2008) were collected, except for Hong Kong, Taiwan and Macao. The data were mainly from national bureau of statistics of China and water resources bulletin of provinces.

T. Wang, S. Xu / Ecological Indicators 52 (2015) 134–146

137

Table 2 Classes of WECC. Class

Level of carrying capacity

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

Limit value of WECC

WEPCC

WESCC

WERCC

Fine level Acceptable level Fine level Fine level Acceptable level Acceptable level Fine level Acceptable level Poor level Fine level Fine level Poor level Acceptable level Poor level Fine level Acceptable level Fine level Poor level Poor level Fine level Acceptable level Poor level Acceptable level Poor level Poor level Acceptable level Poor level

Fine level Fine level Fine level Acceptable level Fine level Acceptable level Acceptable level Acceptable level Fine level Fine level Poor level Fine level Fine level Acceptable level Acceptable level Poor level Poor level Fine level Poor level Poor level Acceptable level Acceptable level Poor level Acceptable level Poor level Poor level Poor level

Fine level Fine level Acceptable level Fine level Acceptable level Fine level Acceptable level Acceptable level Fine level Poor level Fine level Acceptable level Poor level Fine level Poor level Fine level Acceptable level Poor level Fine level Poor level Poor level Acceptable level Acceptable level Poor level Acceptable level Poor level Poor level

3.2. Calculation process The indicator system of WECC is developed in accordance with a PSR framework, and the process has been described in section 2.1. To illustrate the application of the method, the following paragraphs work through an example in detail. Step 1: Eqs. (1) and (2) are used to normalize (rij , shj ) the indicators (xij ) and standards (yhj ). As mentioned above, the gross WECC index incorporates contributions from three subsystems. This paper severally assesses WEPCC, WESCC and WERCC to obtain the WECC of the samples. The values of WECC indicators of China in 2003 are expressed as xij . xij = [134.61 412.90 10542.00 9.54 98.14 391.72 × 16.55 638.00 62.60 2131.30 19.37 1.30 32.28 × 1.20 190.55 1990.38 1.49]



r1×6 = 0.789

0.672

0.142

0.000



0.000

0.497



⎡ 1.000 1.000 1.000 1.000 1.000 1.000 ⎤ ⎢ 0.847 0.692 0.674 0.605 0.727 0.952 ⎥ s5×6 = ⎢ 0.678 0.385 0.370 0.399 0.424 0.726 ⎥ ⎣ ⎦ 0.339 0.000

0.077 0.000

0.065 0.000

0.296 0.000

0.212 0.000

0.161 0.000

2 2.25 2.25 2.5 2.5 2.75 2.75 3 2.75 2.75 3.5 3 3 3.25 3.25 3.75 3.75 3.5 4.25 4.25 3.5 3.5 4 4 4.25 4.5 5

Step 2: the AHP model is used to obtain weights of the indicators. The judgment matrices and weights are shown in Table 3. Step 3: Eq. (3) is used to calculate the synthetic relative membership degree. In the previous step, r1×6 , s5×6 is obtained and it is easy to find that m is 6 for the WEPCC of sample 1. Then, the interval of standard (a1 , b1 ) that sample 1 belongs to could be obtained. The detailed process is as follows. s31 = 0.678 ≤ r11 = 0.798 ≤ s21 = 0.847, here a11 is 2, b11 is 3; s32 = 0.38 ≤ r12 = 0.672 ≤ s22 = 0.692, here a12 is 2, b12 is 3; s43 = 0.065 ≤ r13 = 0.142 ≤ s33 = 0.370, here a13 is 3, b13 is 4; s54 = 0.000 ≤ r14 = 0.000 ≤ s44 = 0.296, here a14 is 4, b14 is 5; s55 = 0.000 ≤ r15 = 0.000 ≤ s45 = 0.212, here a15 is 4, b15 is 5; s46 = 0.161 ≤ r16 = 0.497 ≤ s36 = 0.726, here a16 is 3, b16 is 4; a1 = min {a1j } = 2, j = 1, 2. . . 6 b1 = max {b1j } = 5, j = 1, 2. . . 6 ω is determined by the AHP model, and the details are shown in Table 3 (Sub-criteria 1). ω = 0.125

where i = 1; j = 1, 2. . . 17. Note that the j = 1, 2. . . 6 are used to assess WEPCC, j = 7, 8. . . 11 are used to assess WESCC and the remainder of the indicators are used to assess WERCC. Because the processes of computing WEPCC, WESCC and WERCC are similar, only the process of computing WEPCC will be illustrated. For the WEPCC of sample 1, yhj is the value of standard h of indicator j and could be found in Table 1, where h = 1, 2 . . ., 5 and j = 1, 2. . ., 6. Based on Eqs. (1) and (2), the results of the normalization process are as follows.

0 0.5 0.5 1 1 1.5 1.5 2 0.75 0.75 1.5 1.25 1.25 1.75 1.75 2 2 1.5 2.25 2.25 2.25 2.25 2.5 2.5 2.75 2.75 3

0.25

0.125

0.125

0.125

0.25



After the matrices r1×6 , s5×6 and ω as well as a1 , b1 are obtained, the uh1 are calculated for varying a and p. and the results are shown in Table 4. Step 4: Eq. (4) is used to calculate the characteristic value H of the sample 1 based on the third step, then the average value is used as the assessment result. As seen in Table 5, the average value (3.34) of the four patterns is the final level of WEPCC for China in 2003. Similarly, the above steps are repeated to obtain WESCC (2.62) and WERCC (3.99). After calculations of WEPCC, WESCC and WERCC, the weighting method is used to obtain the characteristic value of WECC. The equation is as follows. H(WECC) = 0.25H(WEPCC) + 0.5H(WESCC) + 0.25H(WERCC) (5) Here, H (WEPCC) is the assessment result of WEPCC, H (WESCC) is the assessment result of WESCC, H (WERCC) is the assessment result of WERCC and H (WECC) is the assessment result of WECC.

138

T. Wang, S. Xu / Ecological Indicators 52 (2015) 134–146

Table 3 Judgment matrices and weights.

3.3. Method verification

Criteria

Local weights

WEPCC

WESCC

WERCC

WEPCC 1 0.5 1 0.25 2 1 2 0.5 WESCC 1 0.5 1 0.25 WERCC The largest eigenvalue of matrix(Criteria) is 3; the consistency ratio is 0 < 0.1 Sub-criteria 1(WEPCC) X1

X2

Local weights

X3

X4

X5

X6

X1 1 0.5 1 1 1 0.5 0.125 2 1 2 2 2 1 0.25 X2 1 0.5 1 1 1 0.5 0.125 X3 1 0.5 1 1 1 0.5 0.125 X4 1 0.5 1 1 1 0.5 0.125 X5 2 1 2 2 2 1 0.25 X6 The largest eigenvalue of matrix(WEPCC) is 6; the consistency ratio is 0 < 0.1 Sub-criteria 2(WESCC) X7

X8

Local weights

X9

X10

X11

X7 1 0.5 0.5 0.33 0.5 0.098 X8 2 1 1 0.5 1 0.184 X9 2 1 1 0.5 1 0.184 X10 3 2 2 1 2 0.349 2 1 1 0.5 1 0.184 X11 The largest eigenvalue of matrix(WESCC) is 5.0076; the consistency ratio is 0.0017 < 0.1

Sub-criteria 3(WERCC) X13

X12

Local weights

X14

X15

X16

X17

X12 1 0.5 0.5 1 1 0.25 0.091 2 1 1 2 2 0.5 0.182 X13 X14 2 1 1 2 2 0.5 0.182 1 0.5 0.5 1 1 0.25 0.091 X15 X16 1 0.5 0.5 1 1 0.25 0.091 X17 4 2 2 4 4 1 0.364 The largest eigenvalue of matrix(WERCC) is 6; the consistency ratio is 0 < 0.1

The weight analysis is shown in Table 3 (Criteria). So the WECC of China in 2003 is obtained (3.14). In addition, it can be found that the WECC of China in 2003 belongs to class 24 based on Table 2 because the WEPCC, WESCC and WERCC are 3.34 (poor level), 2.62 (acceptable level) and 3.99 (poor level), respectively. In this way, the WEPCC, WERCC, WESCC and WECC of China and each of its 31 provinces can be dynamically and accurately assessed by the proposed method (Fig. 1).

Table 4 Synthetic relative membership degree of WEPCC of China in 2003. Grade

1 2 3 4 5

Synthetic relative membership degree a = 1; p = 1

a = 1; p = 2

a = 2; p = 1

a = 2; p = 2

0.000 0.239 0.313 0.237 0.210

0.000 0.220 0.339 0.246 0.195

0.000 0.224 0.384 0.220 0.173

0.000 0.185 0.439 0.231 0.145

Table 5 Result of WEPCC of China in 2003. Sample

1

Characteristic value (H) a = 1; p = 1

a = 1; p = 2

a = 2; p = 1

a = 2; p = 2

Average value

3.42

3.34

3.23

3.35

3.34

The results of assessment can be seen in Figs. 2 and 3 and Appendix A. It is easily seen that the results from the proposed method and the fuzzy comprehensive assessment (FCA) method are similar. However, the traditional fuzzy comprehensive assessment method treats the samples as point forms and leads to many indicators with values of 0, which might miss information from samples (Wang et al., 2011). Fortunately, the proposed method treats the samples as having a continuous degree of membership, which retains the information of samples and uses the average value of the four patterns as the final result. This combination of a linear model with a nonlinear model not only avoids the unstable evaluation results caused by single model but also can reflect the differences between regions. Moreover, the original data also prove the reasonableness of the proposed method. Fig. 2 shows that the WECC of China overall is at an acceptable level and is improving with time, from 3.14 in 2003 to 2.45 in 2010. Because the WECC includes WEPCC, WESCC and WERCC, this paper next verifies the reasonableness of each of the three component indices calculated by the proposed method. The WEPCC of China improves from 3.38 in 2003 to 2.24 in 2010, an improvement from a poor level to an acceptable level. These data show that the pressures of population increase and socioeconomic development have a tendency to reduce WEPCC. The original data show that the population density, water consumption per capita and per capita GDP increased by 3%, 8% and 64.8%, respectively, and the ammoniacal nitrogen discharge intensity of economic output, chemical oxygen demand discharge intensity of economic output and water consumption intensity of GDP was reduced by 64.8%, 68.6%, 68.5%, respectively. Although the population density and water consumption per capita increased slightly, pollutant discharge intensity and water consumption intensity decreased significantly, and the per capita GDP also increased, which enhanced the efficacy of environmental governance. The original data show that the WEPCC has improved too. The WESCC is at an acceptable level on the whole, fluctuating between 2.82 and 2.35 from 2003 to 2010. The analysis of the original data finds that the forest coverage and the water resources per capita increased by 18.7% and 7.7%, respectively, during that time. The exploitation and utilization ratio of water resources stabilized at approximately 19.5% and the proportion of lengths of unpolluted stretches of rivers stabilized at approximately 60%, but the annual precipitation was random and ranged from 601 mm to 695 mm. These changes of the indicators lead to the fluctuation of the WESCC. Although the WERCC of China is mainly at a poor level, it improved from 3.99 in 2003 to 2.84 in 2010. The analysis of original data shows that the ratio of investment in Research & Development to GDP, the ratio of tertiary industry output to GDP, the ratio of investment in environmental pollution control to GDP and the ratio of ecological water consumption increased by 25.7%, 24.8%, 27.6%, 24.8%, respectively, and the water consumption intensities of industry and agriculture were reduced by 59.2% and 54.2%, respectively. These original data show that the WERCC improves. Similarly, the changes of WECC at the province level are consistent with the original data. In addition, the change of WECC is closely related to its class, which can be clearly illustrated by the different change histories of Beijing and Heilongjiang. Fig. 3 shows how the WECC of Beijing (3.10) and that of Heilongjiang (3.32) were similarly poor in 2003. However, Beijing is relatively socioeconomically developed and scarce in natural resources compared with Heilongjiang, and their development models and degrees are also different. Beijing is a developed city where environment protection gets more attention in social development, while Heilongjiang is a

T. Wang, S. Xu / Ecological Indicators 52 (2015) 134–146

139

Fig. 1. Study area and locations of the 31 provinces.

developing province whose socioeconomic development comes at the expense of the environment. Those differences lead to the WECC of Beijing improving to an acceptable level (2.82) and that of Heilongjiang degenerating to a poor level (3.66) over the study period. So WECCs should be classified to distinguish its differences. As seen in Fig. 3, WECC of Beijing improves from class 23 to 11 with the improvement of WEPCC and WERCC, and that of Heilongjiang degenerates from class 24 to class 27 due to the poor WEPCC and WERCC. As mentioned above, the dynamic successive WECC assessment method is reasonable and has an advantage of being able to discriminate differences when the values of WECC are similar.

4. Discussion 4.1. Dynamic change and classes of the global WECC It can be concluded from the above section that the WECC of China overall is at an acceptable level and is improving over time (Fig. 2). The driving factors of improvement will be analyzed. The WECC in China belonged to class 24 in 2003–2007 because of the poor WEPCC and WERCC. At that time, socioeconomic development was more important than environmental protection. China has since made an effort to expand tertiary industry and decrease the water consumption intensities of industry and agriculture. Thus, the WECC of China in 2008–2009 improved to class 21. By 2010, as the proportion of GDP invested in Research & Development further increased to 1.65%, the WECC of China had improved to class 8, which shows how environmental protection would gradually play a more important role in sustainable development.

4.2. Dynamic change and classes of local WECC

Fig. 2. Dynamic changes of the WEPCC, WESCC, WERCC and WECC in China (WECCFCA is obtained by the FCA method, the others are obtained by the proposed method).

Fig. 3 and Appendix A show that from 2003 to 2008, the WECC of the provinces are between 2 and 4.5 and have spatial variation. Over the study period, most of the provinces show improvement in WECC, which is the same trend exhibited by the entirety of China. Overall, the WECC values of provinces improve from 3–3.5 to 2.5–3. Generally, the WECC in southern China is better than that in northern China. The best annual average of WECC is 2.44 in Yunnan which is at an acceptable level, and correspondingly the lowest is 4.36 in Ningxia which is at a poor level. The other interesting concern of the WECC is the greatest progress and decline which can be found in Fig. 3. The greatest progress belongs to Fujian (2.94–2.2), of which the carrying capacity is much closer to grade 2 in 2008. Conversely,

140

T. Wang, S. Xu / Ecological Indicators 52 (2015) 134–146

Fig. 3. Temporal and spatial distributions of WECC, WEPCC, WESCC, WERCC and classes of WECC in 31 provinces.

the greatest decline belongs to Heilongjiang (3.32–3.66), of which the carrying capacity is closer to grade 4 in 2008. The WECC classes of the provinces are shown in Fig. 3. In 2003, classes 18 and 24 were most common in provinces. Between 2004 and 2006, classes 18 and 26 were dominant. Between 2007 and 2008, classes 13, 18, 24, 26 were dominant. The changes of WECC of China can be divided into four types during this time, i.e., improvement, degeneration, fluctuation and stabilization. 11

provinces showed improvement from 2003 to 2008: Fujian and Yunnan improved from class 18 to 13 due to improvements in WEPCC, Zhejiang improved from class 22 to 6 due to improvements in WEPCC and WERCC, Beijing improved from class 23 to 11 due to improvements in WEPCC and WERCC, Tianjin improved from class 23 to 17 owing to improvements in WEPCC and WERCC, Guangdong and Guizhou improved from class 24 to 13 because of improvements in WEPCC and WESCC, Jiangsu and Shanghai

T. Wang, S. Xu / Ecological Indicators 52 (2015) 134–146

141

Fig. 4. Uncertainty analysis showing the method results calculated using the proposed method (gray bars, AHP-weights), the mean from the Monte Carlo simulation (square black dot), 95% confidence interval (error bars), and using the equally weighted alternative method (circular black dot, equal weights). ,  and N are parameters of Monte Carlo model.

improved from class 27 to 26 due to improvements in WEPCC, and Shandong and Liaoning improved from class 26 to 23 owing to improvements in WERCC. Over the same time period, three provinces showed WECC degeneration. Hubei degenerated from class 18 to 24 because of worsening WEPCC and WERCC, and for the same reason, Heilongjiang degenerated from class 24 to class 27. Shaanxi province degenerated from class 21 to class 26 because of worse WEPCC. The fluctuation type contains four provinces, which are Chongqing, Jilin, Inner Mongolia and Anhui; the WECC of those areas change on a small scale as a result of the variation of rainfall and the interaction of the WEPCC, WESCC, and WERCC. The stable type contains the remaining provinces. Their WECC classes do not obviously change over the study period. On the whole, abundant natural resources contribute to improving WECC, and on the contrary, natural resource shortages could impede its improvement. At present, improving WECCs in China mainly reflect efforts to optimize water consumption and reducing pollutant discharge. In the future, direct environmental protection deserves more consideration.

successive assessment model is used to assess the WECC of China in 2015. The result shows that the WECC (2.38) is at an acceptable level and belongs to class 7; the WEPCC (1.94), WESCC (2.58) and WERCC (2.40) are at an acceptable level. All of the carrying capacities in 2015 are better than those of 2010. There are some environmental problems associated with the processes of development, and most countries around the world reduce their Science & Technology expenditures because of the global economic slowdown. Against this background, China still increases investment in Science & Technology, environmental protection and the ratio of ecological water consumption. These measures lead to further improvement the WECC in China. In the future, adjustments of industrial systems and practices should be promoted, and a model of socioeconomic development principally founded on environmental protection and technology should be established in order to achieve sustainable development of the economy and the water environment.

4.3. Prediction of WECC in China

Many uncertainties exist in the assessment of WECC. This paper analyzes the uncertainty of the assessment results attributable to two types of choices made during model construction, i.e., different weights and random values of indicators. First, the versions

This paper predicts the WECC of China in 2015 based on the Chinese national government’s Twelfth Five-Year Plan. The dynamic

4.4. Uncertainty analysis

142

T. Wang, S. Xu / Ecological Indicators 52 (2015) 134–146

with unequal weights (AHP-weights) and equal weights were both calculated. Fig. 4 and Appendix B–E show that the WECC results of two versions show the same change trends but are numerically different. Different versions even lead to different judgments; e.g., the WECC of Yunnan in 2003 (AHP-weights) is at an acceptable level (2.62), while the same situation is evaluated by the equal-weight version of the model to be at a poor level (3.16). A further example is that the WECC of Inner Mongolia in 2003 (AHP-weights) belongs to class 24 and under equal weights belongs to class 27, which is due to neglecting the different importance of indicators (Feng et al., 2010). Choices about how to weight the indicators can significantly influence the results of assessment. In general, a situationally appropriate method should be used to determine the weights for different issues. In addition, the values of indicators inevitably have some random errors from the processes of monitoring and quantification (Sills et al., 2013; Ke and Zhou, 2013). A Monte Carlo method was used to simulate such random errors. The text assumes the probability distributions of values of indicators follow Gaussian distributions, which are the most suitable for unknown real-valued distributions and has been successfully used in the literatures (Niero et al., 2014; Chai and Chen, 2013). The Monte Carlo method is used for a sensitivity analysis and the model parameters are as follows: the average  is Xij ; standard deviations  are 0.1 Xij , 0.3 Xij , 0.5 Xij , 0.9 Xij respectively to analyze influence of different degree of deviation; running times N is 1000. Then, the mean and 95% confidence interval of the Xij from Monte Carlo simulation are obtained. The results of assessment can be seen in Fig. 4 and Appendix B–E. It is easy to find that the actual and the simulated results are similar and both located in the 95% confidence interval. However, there are large uncertainties and deviations, and the higher the variance is, the more significant the deviation becomes. In fact, the probability distributions of each indicator are different, which should be researched in future. 4.5. Analysis of application of the method To study the temporal and spatial variability of WECC, this paper explored a dynamic successive assessment method that combined VFPR and AHP model to evaluate indicators chosen under a PSR framework and used the geostatistics method to produce WECC map. In general, the assessment process includes three principal steps, i.e., developing the indicator system and determining the relevant standards, using the model for assessment, and showing and analyzing the assessment results. In the first phase, because many useful indicators mutually influence each other, the PSR model is chosen to develop the indicator system, which not only can systematically indicate the causality of the indicators and contribute to further classifications of WECC but also avoids the randomness of other common methods. Standards for each indicator are determined by references, which can reasonably distinguish carrying capacity level. The standards used in this study were closely related to WECC and were previously used in the literature for the evaluation of the water environment. In the second phase, VFPR is used to assess WECC due to the fuzziness of evaluation, and the weight of

indicators is calculated by an AHP. The VFPR model reserves information of original data by taking continuous membership degree forms instead of point forms of traditional method to normalize the variables (xij ) and improve the reasonability of results by considering the membership of each indicator to analyze synthetic relative membership degree of samples. Moreover, changing the model parameters (a,p) and using the average value of these four model results as the final result enhance the reliability of the assessment. Furthermore, the method can distinguish qualitative differences between samples with similar WECC values by comparing the levels of WEPCC, WESCC and WERCC and using them to classify the sample WECC. In the third phase, because large volumes of assessment data are hard to analyze, the geostatistics method are used to process the data and produce interpretive maps. The WECC map contains much information, including the temporal and spatial distributions of WECC, WEPCC, WESCC and WERCC as well as WECC classes, which help to show the changes and trends of WECC in China. In addition, the influence of different weights and random values of indicators were considered. On the whole, the method can reasonably quantify WECC and sensitively reflects the changing of weights and values of indicators. 5. Conclusion This paper explored a dynamic successive assessment method to study the temporal and spatial variability of WECC. First, the PSR framework is used to develop a systematic and causal indicator system comprising three aspects, including WEPCC, WESCC and WERCC. Second, an AHP is used to calculate the weights of indicators. After that, VFPR model is used to dynamically and successively assess WEPCC, WESCC, and WERCC, from which WECC is obtained by weighting method. Furthermore, all WECCs are divided into 27 classes by judging the levels of WEPCC, WESCC and WERCC. This study selects 17 indicators for the assessment and takes China as an example. The temporal and spatial variability of WECC in China shows that the potential exists for socioeconomic and waterenvironmental co-development to be harmonious and sustainable because socioeconomic development will strengthen environmental requirements and raise the conservation aspirations of human society, even though development will consume water resources and increase pollution loads. It is concluded that the method is reasonable and it can be used not only to accurately understand the change of WECC in single areas over time but also to distinguish the differences between areas with similar WECC values. This paper believes that the dynamic successive assessment method of WECC contributes to the study of the relationship between socioeconomic development and the water environment, and it also could provide a reference point for policymaking as well as further, similar studies. Acknowledgments The authors would like to thank anonymous reviewers. This work was supported by the National Natural Science Foundation of China (Grant No. 51327004 and No. 51279022)

T. Wang, S. Xu / Ecological Indicators 52 (2015) 134–146

143

Appendix A. Detail results of assessment of WECC in China (2003–2008). Area

2003 (1)

Anhui Beijing Chongqing Fujian Gansu Guangdong Guangxi Guizhou Hebei Henan Hubei Hunan Hainan Heilongjiang Inner Mongolia Jiangsu Jiangxi Jilin Liaoning Ningxia Hui Qinghai Shaanxi Shanxi Shanghai Shandong Sichuan Tianjin Xinjiang Xizang Yunnan Zhejiang Coefficient of Variation(%)

2004 (2)

(3)

(1)

2005 (2)

(3)

2006

(1)

(2)

(3)

(1)

2007 (2)

(3)

(1)

2008 (2)

(3)

(1)

(2)

(3)

3.11 3.10 2.72 2.94 3.65 2.85 2.56 3.09 3.78 3.50 2.93 2.93 2.52 3.32 3.44 3.49 2.57 3.09 3.60 4.40 3.07 2.93 3.76 3.82 3.59 2.64 3.62 2.98 3.03 2.62 2.85

3.17 2.98 2.67 2.83 3.81 2.79 2.64 3.11 4.06 3.46 2.90 2.91 2.55 3.23 3.54 3.60 2.57 3.15 3.75 4.56 2.93 2.96 3.85 3.78 3.56 2.53 3.48 2.92 2.89 2.66 2.73

3.35 2.84 2.93 3.27 3.71 3.20 3.13 3.48 3.70 3.57 3.28 3.37 3.03 3.57 3.62 3.69 3.14 3.23 3.37 4.16 3.49 3.05 3.66 3.77 3.50 3.03 3.41 3.33 3.36 3.16 3.07

3.61 3.25 2.73 2.91 4.08 3.11 2.72 2.86 3.83 3.68 3.08 2.72 2.98 3.52 3.46 3.90 2.79 3.21 3.35 4.40 3.04 3.35 4.04 3.66 3.60 2.33 3.74 3.06 2.93 2.49 2.79

3.62 3.01 2.70 2.90 4.36 3.03 2.72 2.79 4.01 3.61 3.00 2.77 2.89 3.55 3.61 3.85 2.78 3.22 3.40 4.57 2.91 3.39 4.04 3.69 3.64 2.41 3.53 3.09 2.84 2.55 2.67

3.63 2.97 2.92 3.21 4.04 3.35 3.27 3.28 3.64 3.60 3.29 3.16 3.25 3.65 3.57 4.01 3.25 3.27 3.14 4.14 3.42 3.20 3.79 3.61 3.44 2.75 3.51 3.36 3.26 3.01 3.00

3.40 3.04 2.79 2.35 3.42 2.74 2.85 2.87 3.76 3.47 3.08 2.55 2.52 3.42 3.18 3.85 2.53 2.83 3.28 4.39 3.10 3.21 4.08 3.68 3.45 2.46 3.57 3.02 2.94 2.44 2.51

3.38 2.92 2.83 2.43 3.43 2.64 2.76 2.78 3.97 3.45 3.01 2.61 2.59 3.38 3.32 3.87 2.55 2.87 3.20 4.53 2.92 3.24 4.14 3.68 3.41 2.40 3.36 3.02 2.87 2.49 2.37

3.49 2.73 2.97 2.84 3.58 3.01 3.39 3.21 3.58 3.43 3.29 3.07 2.96 3.65 3.36 3.87 3.10 3.06 3.13 4.13 3.53 3.15 3.82 3.58 3.25 2.86 3.32 3.33 3.28 2.91 2.78

3.52 2.92 2.88 2.31 3.80 2.42 2.72 2.91 3.89 3.74 3.23 2.69 2.49 3.54 3.19 3.80 2.50 3.16 3.40 4.38 3.12 3.38 3.97 3.61 3.68 2.61 3.76 3.07 3.08 2.43 2.74

3.45 2.76 2.91 2.41 3.97 2.38 2.69 2.92 4.10 3.79 3.27 2.73 2.60 3.55 3.28 3.80 2.53 3.16 3.49 4.48 3.01 3.45 4.02 3.60 3.61 2.59 3.50 3.11 3.01 2.52 2.43

3.60 2.60 2.98 2.78 3.84 2.76 3.30 3.22 3.60 3.59 3.34 3.14 2.90 3.70 3.33 3.84 3.05 3.20 3.17 4.09 3.55 3.20 3.71 3.50 3.38 2.86 3.47 3.37 3.44 2.86 2.89

3.19 2.85 2.57 2.33 3.53 2.85 2.68 2.79 3.90 3.58 2.91 2.72 2.36 3.77 3.47 3.59 2.67 2.96 3.34 4.32 2.97 3.17 3.79 3.57 3.45 2.48 3.72 3.08 2.94 2.50 2.46

3.26 2.69 2.58 2.34 3.59 2.57 2.63 2.83 3.98 3.49 2.94 2.76 2.46 3.87 3.56 3.51 2.68 3.04 3.50 4.39 2.81 3.11 3.71 3.56 3.21 2.46 3.42 3.08 2.81 2.46 2.27

3.33 2.50 2.85 2.68 3.59 2.96 3.21 3.15 3.58 3.46 3.14 3.05 2.79 3.77 3.43 3.59 3.13 3.01 3.09 4.03 3.41 3.01 3.52 3.39 3.16 2.77 3.37 3.37 3.27 2.91 2.57

3.21 2.82 2.69 2.20 3.62 2.21 2.50 2.60 3.45 3.61 2.84 2.46 2.31 3.66 3.03 3.59 2.48 2.89 3.24 4.26 2.75 3.24 3.88 3.55 3.43 2.22 3.55 3.11 2.91 2.16 2.45

3.18 2.72 2.61 2.25 3.87 2.14 2.46 2.70 3.58 3.50 2.88 2.56 2.43 3.83 2.98 3.55 2.52 2.93 3.28 4.31 2.63 3.23 3.75 3.48 3.24 2.29 3.22 3.26 2.76 2.21 2.11

3.34 2.49 2.88 2.50 3.63 2.44 3.05 3.04 3.19 3.44 3.06 2.84 2.73 3.63 3.09 3.44 2.98 2.90 2.99 3.96 3.16 3.02 3.58 3.29 3.10 2.58 3.23 3.37 3.24 2.60 2.45

14.34

16.02

8.57

15.46

16.78

9.74

16.47

17.39

10.29

16.98

17.72

11.02

16.65

17.44

11.05

18.89

19.33

12.46

Notes: (1) are results of the proposed method(AHP-weight), (2) are results of the FCA method(AHP-weight), (3) are results of the proposed method(equal weight). Appendix B. Detail results of assessment of Monte Carlo simulation (average  is Xij ; standard deviation  is 0.1 Xij ). Area

Anhui Beijing Chongqing Fujian Gansu Guangdong Guangxi Guizhou Hebei Henan Hubei Hunan Hainan Heilongjiang Inner Mongolia Jiangsu Jiangxi Jilin Liaoning Ningxia Hui Qinghai Shaanxi Shanxi Shanghai Shandong Sichuan

2003

2004

2005

2006

2007

2008

Mean

Confidence interval

Mean

Confidence interval

Mean

Confidence interval

Mean

Confidence interval

Mean

Confidence interval

Mean

Confidence interval

3.11 3.10 2.72 2.94 3.64 2.85 2.56 3.09 3.78 3.51 2.93 2.92 2.52 3.33 3.44 3.49 2.57 3.09 3.60 4.41 3.06 2.93 3.76 3.82 3.59 2.64

3.72 3.28 3.17 3.61 3.77 3.51 3.64 3.71 4.00 3.73 3.64 3.69 3.58 3.78 3.93 4.16 3.62 3.47 3.70 4.33 3.59 3.32 3.79 4.07 3.84 3.34

3.61 3.25 2.73 2.91 4.08 3.11 2.72 2.86 3.83 3.68 3.08 2.72 2.98 3.53 3.46 3.90 2.79 3.21 3.35 4.40 3.03 3.35 4.04 3.66 3.60 2.33

3.92 3.31 3.18 3.58 4.19 3.62 3.64 3.59 3.97 3.90 3.62 3.68 3.63 3.92 3.99 4.22 3.65 3.41 3.48 4.33 3.54 3.43 4.10 4.07 4.04 3.30

3.40 3.04 2.79 2.35 3.42 2.74 2.85 2.87 3.76 3.47 3.08 2.54 2.52 3.42 3.19 3.85 2.53 2.83 3.28 4.39 3.10 3.21 4.08 3.68 3.45 2.46

3.75 3.23 3.25 3.47 3.56 3.45 3.65 3.57 3.80 3.58 3.57 3.61 3.60 3.86 3.68 4.27 3.61 3.42 3.47 4.32 3.65 3.42 4.07 4.15 3.78 3.28

3.52 2.92 2.88 2.31 3.80 2.42 2.72 2.91 3.89 3.74 3.23 2.69 2.48 3.54 3.19 3.80 2.50 3.16 3.41 4.38 3.12 3.38 3.96 3.61 3.68 2.61

3.85 3.04 3.23 3.48 3.91 3.41 3.64 3.53 4.01 4.00 3.72 3.63 3.60 3.96 3.62 4.25 3.59 3.46 3.57 4.31 3.68 3.45 4.03 4.12 3.89 3.25

3.19 2.85 2.57 2.33 3.62 2.85 2.67 2.79 3.90 3.59 2.91 2.72 2.36 3.77 3.47 3.53 2.66 2.95 3.34 4.32 2.97 3.17 3.79 3.58 3.45 2.47

3.60 3.02 3.25 3.35 3.78 3.37 3.60 3.49 4.00 3.64 3.51 3.57 3.45 4.01 3.81 4.05 3.59 3.32 3.48 4.26 3.53 3.31 3.84 4.14 3.67 3.21

3.22 2.83 2.69 2.20 3.61 2.20 2.50 2.60 3.45 3.61 2.81 2.46 2.31 3.66 3.03 3.59 2.48 2.89 3.23 4.26 2.75 3.24 3.88 3.55 3.43 2.22

3.55 3.08 3.20 3.35 3.82 3.26 3.54 3.47 3.73 3.82 3.50 3.49 3.42 3.87 3.49 4.04 3.58 3.25 3.36 4.23 3.29 3.32 3.88 4.08 3.66 3.17

2.60 3.03 2.25 2.33 3.22 2.22 2.21 2.56 3.54 3.28 2.37 2.30 2.22 2.92 2.99 3.21 2.24 2.63 3.42 4.02 2.76 2.28 3.63 3.48 3.35 2.43

3.32 3.12 2.25 2.26 3.74 2.52 2.25 2.36 3.65 3.46 2.40 2.28 2.29 3.01 3.01 3.48 2.28 2.86 3.21 4.03 2.78 3.15 3.89 3.34 3.44 2.15

3.01 3.03 2.34 2.16 2.97 2.19 2.25 2.36 3.68 3.33 2.37 2.23 2.23 2.96 2.84 3.50 2.24 2.53 3.06 4.03 2.80 2.88 4.01 3.36 3.28 2.21

3.02 2.78 2.50 2.18 3.44 2.13 2.29 2.42 3.82 3.56 2.72 2.36 2.24 3.05 2.77 3.46 2.23 2.82 3.25 4.05 2.85 3.23 3.87 3.31 3.51 2.24

2.85 2.76 2.31 2.14 3.19 2.11 2.18 2.52 3.85 3.39 2.36 2.24 2.18 3.16 3.01 3.20 2.25 2.64 3.23 4.02 2.74 2.96 3.63 3.33 3.26 2.25

2.81 2.77 2.27 2.11 3.39 2.00 2.12 2.45 3.36 3.43 2.36 2.20 2.17 3.13 2.63 3.29 2.23 2.61 3.15 4.02 2.62 3.12 3.80 3.30 3.27 2.13

144

T. Wang, S. Xu / Ecological Indicators 52 (2015) 134–146

Area

Tianjin Xinjiang Xizang Yunnan Zhejiang

2003

2004

2005

2006

2007

2008

Mean

Confidence interval

Mean

Confidence interval

Mean

Confidence interval

Mean

Confidence interval

Mean

Confidence interval

Mean

Confidence interval

3.62 2.98 3.03 2.62 2.85

3.84 3.40 3.50 3.66 3.18

3.75 3.06 2.93 2.49 2.79

3.86 3.61 3.44 3.54 3.19

3.57 3.02 2.94 2.44 2.51

3.82 3.44 3.44 3.45 3.12

3.76 3.07 3.08 2.43 2.74

3.85 3.51 3.50 3.44 3.14

3.72 3.08 2.94 2.50 2.46

3.80 3.52 3.44 3.38 3.07

3.55 3.11 2.91 2.16 2.45

3.69 3.55 3.43 3.15 2.91

3.44 2.55 2.60 2.28 2.35

3.62 2.63 2.56 2.19 2.28

3.42 2.61 2.56 2.17 2.07

3.73 2.66 2.69 2.17 2.19

3.66 2.69 2.60 2.15 2.10

3.45 2.76 2.56 1.98 1.96

Appendix C. Detail results of assessment of Monte Carlo simulation (average  is Xij ; standard deviation  is 0.3 Xij ). Area

Anhui Beijing Chongqing Fujian Gansu Guangdong Guangxi Guizhou Hebei Henan Hubei Hunan Hainan Heilongjiang Inner Mongolia Jiangsu Jiangxi Jilin Liaoning Ningxia Hui Qinghai Shaanxi Shanxi Shanghai Shandong Sichuan Tianjin Xinjiang Xizang Yunnan Zhejiang

2003

2004

2005

2006

2007

2008

Mean

Confidence interval

Mean

Confidence interval

Mean

Confidence interval

Mean

Confidence interval

Mean

Confidence interval

Mean

Confidence interval

3.11 3.11 2.72 2.94 3.64 2.85 2.56 3.09 3.78 3.51 2.94 2.93 2.52 3.33 3.44 3.48 2.57 3.09 3.59 4.40 3.07 2.93 3.75 3.82 3.58 2.64 3.62 2.98 3.03 2.62 2.85

4.04 3.75 3.63 4.01 4.43 3.95 3.85 3.93 4.50 4.02 4.05 4.03 3.89 3.92 4.12 4.30 3.98 3.90 4.19 4.36 3.86 3.67 4.16 4.28 4.19 3.71 4.00 4.02 3.79 3.92 3.57

3.62 3.26 2.72 2.91 4.09 3.12 2.72 2.85 3.84 3.68 3.07 2.75 2.99 3.51 3.46 3.91 2.79 3.21 3.36 4.40 3.03 3.34 4.08 3.66 3.60 2.44 3.74 3.06 2.94 2.49 2.78

4.46 3.73 3.61 3.99 4.51 4.04 3.89 3.87 4.40 4.37 4.06 3.98 3.98 4.15 4.15 4.39 4.01 3.81 4.03 4.36 3.81 3.72 4.24 4.24 4.20 3.71 3.98 4.05 3.67 3.88 3.48

3.40 3.03 2.79 2.35 3.42 2.74 2.84 2.87 3.75 3.47 3.09 2.56 2.52 3.42 3.18 3.86 2.52 2.83 3.29 4.40 3.10 3.22 4.08 3.68 3.45 2.46 3.57 3.01 2.94 2.44 2.51

4.13 3.65 3.65 3.83 4.12 3.85 3.86 3.91 4.29 4.00 4.02 3.96 3.96 4.04 3.89 4.49 3.96 3.82 3.90 4.36 3.86 3.65 4.21 4.26 4.13 3.77 3.96 4.04 3.67 3.89 3.39

3.52 2.92 2.88 2.31 3.80 2.54 2.72 2.90 3.88 3.74 3.23 2.78 2.48 3.54 3.19 3.81 2.50 3.16 3.40 4.38 3.13 3.38 3.97 3.61 3.69 2.61 3.76 3.07 3.08 2.43 2.74

4.40 3.52 3.63 3.88 4.25 3.79 3.87 3.90 4.33 4.32 4.23 3.97 3.99 4.19 4.00 4.44 3.98 3.88 4.01 4.35 3.94 3.73 4.19 4.21 4.13 3.76 3.98 4.07 3.84 3.90 3.39

3.19 2.85 2.58 2.33 3.62 2.85 2.67 2.79 3.90 3.58 2.91 2.71 2.35 3.76 3.47 3.59 2.67 2.96 3.34 4.32 2.97 3.17 3.78 3.54 3.46 2.47 3.72 3.08 2.94 2.45 2.46

4.10 3.50 3.64 3.81 4.25 3.73 3.83 3.88 4.24 4.05 4.08 3.97 3.89 4.31 4.11 4.24 3.99 3.74 3.99 4.30 3.84 3.58 4.06 4.22 4.03 3.68 3.92 4.04 3.75 3.82 3.31

3.21 2.82 2.68 2.21 3.61 2.20 2.50 2.59 3.44 3.61 2.85 2.46 2.31 3.67 3.02 3.59 2.48 2.89 3.24 4.26 2.75 3.24 3.88 3.55 3.43 2.21 3.54 3.11 2.91 2.16 2.45

4.10 3.55 3.61 3.77 4.27 3.61 3.78 3.92 4.09 4.03 4.00 3.91 3.92 4.23 3.87 4.18 4.00 3.67 3.93 4.27 3.60 3.59 3.99 4.19 3.95 3.67 3.83 4.03 3.69 3.63 3.24

2.17 2.51 2.07 2.04 2.64 2.00 1.91 2.14 3.13 2.67 2.10 1.97 1.98 2.46 2.68 2.52 2.00 1.95 2.77 3.22 2.33 1.95 3.03 2.94 2.89 2.00 2.93 2.28 2.23 2.02 1.76

2.48 2.64 2.09 2.03 2.79 2.03 1.94 2.08 3.13 3.01 2.06 1.98 2.04 2.62 2.68 2.90 2.04 2.16 2.51 3.31 2.33 2.44 3.30 2.99 3.13 1.97 2.96 2.41 2.20 2.03 1.82

2.40 2.55 2.05 1.97 2.39 1.99 1.93 2.10 3.08 2.76 2.07 1.97 2.02 2.57 2.61 2.70 2.02 2.07 2.54 3.26 2.33 2.53 3.52 3.00 2.93 1.99 2.90 2.45 2.19 2.02 1.68

2.23 2.32 2.12 1.99 2.56 1.95 1.93 2.09 3.15 3.10 2.11 2.00 2.03 2.68 2.49 2.68 2.04 2.30 2.59 3.23 2.50 2.51 3.27 2.98 3.12 2.00 3.01 2.54 2.38 2.03 1.88

2.17 2.30 2.10 1.95 2.41 1.94 1.90 2.11 3.11 2.83 2.07 2.03 1.99 2.64 2.50 2.38 2.05 1.97 2.53 3.19 2.32 2.42 3.13 2.99 2.84 2.00 2.94 2.40 2.16 1.97 1.78

2.14 2.44 2.04 1.94 2.55 1.88 1.87 2.08 3.02 2.93 2.06 2.01 2.00 2.49 2.33 2.64 2.04 1.96 2.48 3.23 2.30 2.44 3.24 2.96 2.87 2.01 2.93 2.58 2.15 1.85 1.64

Appendix D. Detail results of assessment of Monte Carlo simulation (average  is Xij ; standard deviation  is 0.5 Xij ). Area

Anhui Beijing Chongqing Fujian Gansu Guangdong Guangxi Guizhou Hebei Henan Hubei Hunan Hainan Heilongjiang Inner Mongolia Jiangsu Jiangxi Jilin Liaoning Ningxia Hui

2003

2004

2005

2006

2007

2008

Mean

Confidence interval

Mean

Confidence interval

Mean

Confidence interval

Mean

Confidence interval

Mean

Confidence interval

Mean

Confidence interval

3.12 3.10 2.71 2.93 3.64 2.84 2.56 3.10 3.78 3.50 2.92 2.93 2.51 3.32 3.43 3.49 2.57 3.08 3.60 4.41

4.21 3.99 3.89 4.30 4.63 4.34 4.04 4.16 4.68 4.24 4.22 4.22 4.09 4.18 4.40 4.32 4.19 4.27 4.52 4.39

3.60 3.25 2.72 2.92 4.07 3.13 2.68 2.87 3.83 3.68 3.03 2.72 2.98 3.53 3.47 3.91 2.79 3.21 3.35 4.41

4.70 3.98 3.90 4.31 4.63 4.65 4.11 4.12 4.56 4.58 4.30 4.21 4.33 4.51 4.45 4.40 4.24 4.11 4.43 4.40

3.39 3.13 2.78 2.34 3.42 2.75 2.74 2.92 3.76 3.45 3.12 2.54 2.53 3.43 3.19 3.85 2.54 2.84 3.29 4.40

4.39 3.91 3.97 4.05 4.37 4.24 4.07 4.15 4.53 4.45 4.25 4.21 4.20 4.41 4.20 4.51 4.17 4.02 4.29 4.40

3.52 2.90 2.88 2.31 3.80 2.41 2.73 2.92 3.89 3.74 3.27 2.70 2.48 3.53 3.19 3.79 2.51 3.17 3.40 4.37

4.70 3.78 3.98 4.15 4.34 4.14 4.09 4.14 4.55 4.59 4.65 4.19 4.25 4.63 4.17 4.50 4.18 4.20 4.35 4.38

3.20 2.85 2.59 2.33 3.53 2.84 2.78 2.79 3.91 3.58 2.91 2.71 2.36 3.76 3.47 3.53 2.66 2.96 3.38 4.32

4.40 3.75 3.92 4.17 4.34 4.21 4.07 4.13 4.41 4.37 4.30 4.22 4.10 4.64 4.18 4.35 4.23 4.21 4.34 4.34

3.22 2.82 2.68 2.20 3.62 2.20 2.45 2.58 3.47 3.60 2.84 2.45 2.31 3.65 3.00 3.59 2.48 2.90 3.22 4.26

4.45 3.79 3.94 4.17 4.39 3.94 3.95 4.15 4.23 4.23 4.28 4.13 4.16 4.50 4.14 4.32 4.21 4.14 4.22 4.31

1.94 2.46 1.91 1.94 2.26 1.89 1.87 2.01 2.64 2.47 1.93 1.92 1.87 2.03 2.48 2.27 1.93 1.85 2.26 3.03

2.06 2.56 1.90 1.92 2.43 1.89 1.89 1.94 2.62 2.42 1.93 1.91 1.92 2.38 2.40 2.62 1.94 1.73 2.05 3.18

1.94 2.45 1.88 1.89 2.13 1.88 1.88 1.94 2.80 2.43 1.92 1.90 1.92 2.22 2.28 2.41 1.93 1.84 2.30 3.22

1.95 2.17 1.87 1.90 2.21 1.86 1.88 1.92 2.94 2.44 1.92 1.90 1.92 2.44 1.90 2.30 1.95 1.94 2.25 3.09

1.93 2.17 1.94 1.86 2.09 1.85 1.84 1.92 2.95 2.38 1.93 1.91 1.88 2.04 1.90 2.12 1.94 1.82 2.19 3.03

1.90 2.19 1.89 1.86 2.12 1.82 1.82 1.93 2.52 2.42 1.93 1.90 1.90 1.93 1.91 2.23 1.94 1.82 1.87 3.06

T. Wang, S. Xu / Ecological Indicators 52 (2015) 134–146

Area

Qinghai Shaanxi Shanxi Shanghai Shandong Sichuan Tianjin Xinjiang Xizang Yunnan Zhejiang

2003

2004

2005

145

2006

2007

2008

Mean

Confidence interval

Mean

Confidence interval

Mean

Confidence interval

Mean

Confidence interval

Mean

Confidence interval

Mean

Confidence interval

3.06 2.93 3.75 3.83 3.58 2.64 3.61 2.98 3.02 2.62 2.83

4.02 3.95 4.54 4.44 4.57 3.99 4.16 4.15 3.92 4.10 4.14

3.03 3.35 4.04 3.67 3.61 2.32 3.74 3.06 2.94 2.49 2.79

3.91 4.08 4.47 4.39 4.41 3.97 4.13 4.19 3.81 4.10 3.98

3.09 3.21 4.08 3.69 3.44 2.47 3.57 3.02 3.00 2.45 2.50

4.01 4.00 4.43 4.39 4.29 3.98 4.09 4.18 3.84 4.08 3.72

3.12 3.38 3.97 3.62 3.68 2.62 3.76 3.07 3.08 2.43 2.62

4.04 4.21 4.36 4.33 4.27 4.04 4.12 4.22 3.93 4.10 3.70

2.93 3.18 3.79 3.52 3.44 2.49 3.72 3.07 2.99 2.46 2.47

3.91 3.89 4.20 4.32 4.16 4.02 4.07 4.22 3.88 4.00 3.63

2.76 3.24 3.88 3.56 3.42 2.22 3.55 3.12 2.96 2.16 2.45

3.67 3.99 4.11 4.26 4.09 4.01 3.97 4.22 3.87 3.83 3.51

2.24 1.85 2.56 2.88 2.57 1.89 2.82 1.88 2.18 1.92 1.58

2.21 2.00 2.87 2.95 2.66 1.87 2.87 1.98 2.13 1.92 1.62

2.25 2.32 2.94 2.97 2.46 1.87 2.84 2.01 2.12 1.91 1.55

2.27 2.04 2.89 2.95 2.89 1.89 2.91 2.19 2.16 1.91 1.62

2.23 1.98 2.76 2.94 2.23 1.89 2.83 2.05 2.10 1.86 1.58

2.08 2.01 2.81 2.88 2.34 1.90 2.79 2.32 2.09 1.77 1.38

Appendix E. Detail results of assessment of Monte Carlo simulation (average  is Xij ; standard deviation  is 0.9 Xij ). Area

Anhui Beijing Chongqing Fujian Gansu Guangdong Guangxi Guizhou Hebei Henan Hubei Hunan Hainan Heilongjiang Inner Mongolia Jiangsu Jiangxi Jilin Liaoning Ningxia Hui Qinghai Shaanxi Shanxi Shanghai Shandong Sichuan Tianjin Xinjiang Xizang Yunnan Zhejiang

2003

2004

2005

2006

2007

2008

Mean

Confidence interval

Mean

Confidence interval

Mean

Confidence interval

Mean

Confidence interval

Mean

Confidence interval

Mean

Confidence interval

3.08 3.14 2.72 2.93 3.67 2.82 2.57 3.05 3.77 3.49 2.96 2.96 2.52 3.33 3.43 3.48 2.57 3.17 3.60 4.41 3.06 2.94 3.75 3.80 3.56 2.63 3.60 2.98 2.98 2.61 2.85

4.29 4.13 3.98 4.53 4.64 4.61 4.11 4.18 4.70 4.66 4.35 4.34 4.17 4.53 4.54 4.39 4.22 4.58 4.54 4.43 4.02 4.04 4.61 4.51 4.59 4.02 4.27 4.18 3.94 4.17 4.29

3.62 3.27 2.74 2.89 4.06 3.12 2.74 2.81 3.83 3.68 3.14 2.81 2.99 3.56 3.45 3.90 2.79 3.20 3.36 4.40 3.04 3.34 4.05 3.67 3.61 2.34 3.73 3.07 2.93 2.49 2.72

4.73 4.13 3.99 4.67 4.64 4.67 4.19 4.15 4.63 4.61 4.71 4.35 4.71 4.59 4.51 4.42 4.35 4.46 4.51 4.43 3.91 4.42 4.62 4.56 4.55 4.02 4.23 4.22 3.90 4.16 4.30

3.42 3.13 2.80 2.29 3.40 2.75 2.84 2.87 3.77 3.48 3.09 2.56 2.54 3.40 3.17 3.85 2.54 2.94 3.30 4.39 3.10 3.22 4.07 3.67 3.46 2.43 3.58 3.02 2.94 2.44 2.50

4.71 4.03 4.02 4.08 4.41 4.63 4.18 4.20 4.57 4.59 4.67 4.41 4.26 4.57 4.24 4.51 4.20 4.12 4.51 4.44 4.01 4.08 4.65 4.51 4.46 4.01 4.13 4.23 3.90 4.14 4.00

3.48 2.92 2.86 2.31 3.78 2.58 2.72 2.89 3.98 3.71 3.21 2.81 2.46 3.54 3.18 3.75 2.49 3.17 3.39 4.38 3.12 3.38 3.95 3.63 3.74 2.61 3.76 3.06 3.09 2.44 2.75

4.73 3.89 4.12 4.18 4.39 4.27 4.17 4.19 4.65 4.60 4.69 4.29 4.37 4.70 4.20 4.51 4.20 4.56 4.49 4.42 4.04 4.62 4.58 4.50 4.44 4.09 4.17 4.26 3.92 4.16 4.08

3.18 2.87 2.59 2.33 3.53 2.71 2.66 2.79 3.90 3.58 2.88 2.85 2.36 3.75 3.48 3.53 2.67 2.94 3.35 4.32 2.97 3.18 3.78 3.58 3.47 2.58 3.71 3.09 2.94 2.52 2.48

4.64 3.80 4.01 4.29 4.39 4.64 4.32 4.17 4.54 4.50 4.66 4.48 4.16 4.71 4.30 4.38 4.40 4.54 4.53 4.39 3.91 4.30 4.46 4.48 4.30 4.04 4.11 4.28 3.88 4.07 4.12

3.22 2.84 2.64 2.20 3.61 2.20 2.50 2.61 3.43 3.59 2.82 2.50 2.31 3.67 3.03 3.59 2.47 2.90 3.22 4.25 2.75 3.24 3.88 3.54 3.43 2.22 3.55 3.12 2.92 2.15 2.46

4.54 3.85 4.10 4.28 4.43 4.30 4.02 4.18 4.48 4.39 4.68 4.31 4.18 4.56 4.21 4.43 4.27 4.55 4.49 4.38 3.68 4.54 4.29 4.37 4.20 4.09 4.05 4.27 3.87 3.97 4.04

1.89 2.25 1.83 1.90 2.16 1.83 1.77 1.85 1.96 2.13 1.86 1.83 1.86 1.83 2.08 2.02 1.90 1.82 1.83 2.52 2.23 1.83 2.49 2.78 2.48 1.84 2.67 1.85 2.10 1.89 1.53

1.90 2.17 1.83 1.88 2.21 1.86 1.77 1.89 1.91 1.91 1.90 1.84 1.89 1.88 1.88 1.96 1.90 1.73 1.81 2.85 2.21 1.83 2.47 2.81 2.46 1.86 2.70 1.89 2.05 1.90 1.49

References Arrow, K., Bolin, B., Costanza, R., Dasgupta, P., Folke, C., Holling, C.S., Jansson, B.O., Levin, S., Mäler, K.G., Perrings, C., Pimentel, D., 1995. Economic growth, carrying capacity and the environment. Science 268 (28), 520–521. Abdideh, M., Ghasemi, A., 2014. A comparison of various statistical and geostatistical methods in estimating the geomechanical properties of reservoir rocks. Petrol. Sci. Technol. 32 (9), 1058–1064. Bouwer, H., 2002. Integrated water management for the 21st century: problems and solutions. J. Irrig. Drain. Eng. ASCE 128, 193–202. Chen, C.H., Wu, R.S., Liaw, S.L., Sue, W.R., Chiou, I.J., 2000. A study of water-land environment carrying capacity for a river basin. Water Sci. Technol. 42 (2), 389–396. Chen, S.Y., Guo, Y., 2006. Variable fuzzy sets and its application in comprehensive risk evaluation for flood-control engineering system. Fuzzy Optim. Decis. Mak. 5, 153–162. Chen, S.Y., 2009. Theory and Model of Variable Fuzzy Sets and its Application. Dalian University of Technology Press, Dalian (in Chinese). Chai, H.X., Chen, W., 2013. Uncertainty analysis by Monte Carlo simulation in a life cycle assessment of water-saving project in green buildings. Inform. Technol. J. 12 (13), 2593–2598. Daily, G.C., Ehrlich, P.R., 1996. Socioeconomic equity, sustain ability and earth’s carrying capacity. Ecol. Appl. 16, 991–1001.

1.88 2.09 1.83 1.85 2.01 1.84 1.76 1.90 2.16 1.90 1.88 1.86 1.91 1.83 1.64 2.14 1.89 1.80 1.80 2.96 2.24 1.88 2.40 2.86 2.36 1.86 2.71 1.94 2.04 1.88 1.50

1.89 1.87 1.82 1.89 2.01 1.82 1.83 1.89 2.49 1.89 1.89 1.87 1.89 2.00 1.56 2.02 1.89 1.80 1.85 2.72 2.25 1.84 2.48 2.77 2.44 1.86 2.88 1.99 2.09 1.88 1.57

1.86 1.86 1.82 1.83 2.07 1.82 1.79 1.88 2.37 1.85 1.90 1.88 1.87 1.89 1.61 1.92 1.87 1.76 1.73 2.75 2.20 1.80 2.17 2.67 1.86 1.85 2.76 1.99 2.04 1.83 1.54

1.82 1.41 1.81 1.82 2.10 1.79 1.77 1.89 1.77 1.82 1.89 1.87 1.88 1.85 1.54 2.00 1.86 1.75 1.68 2.78 2.04 1.72 2.40 2.47 1.81 1.85 2.60 1.97 2.03 1.73 1.34

Duan, C.Q., Liu, C.M., Cao, L.L., Chen, X.N., Qiu, L., 2009. Evaluation of water resources carrying capacity by a variable fuzzy set method. J. Beijing Norm. Univ. (Nat. Sci.) 45, 582–584 (in Chinese). ´ N., Urbaniec, K., 2012. 6th Dubrovnik Conference on Sustainable Development DuC, of Energy, Water and Environment systems. J. Clean. Prod. 24, 202–203. El-Fadel, M., Tomaszkiewicz, M., Adra, Y., Sadek, S., Abou Najm, M., 2014. GIS-based assessment for the development of a groundwater quality index towards sustainable aquifer management. Water Resour. Manag. 28 (11), 3471–3487. Feng, L.H., Huang, C.F., 2008. A risk assessment model of water shortage based on information diffusion technology and its application in analyzing carrying capacity of water resources. Water Resour. Manag. 621 (22), 621–633. Feng, L.H., Zhang, X.C., Gao, Y.L., 2009. Research on the risk of water shortages and the carrying capacity of water resources in Yiwu, China. Hum. Ecol. Risk Assess. 15 (4), 714–726. Feng, F., Xu, S.G., Liu, J.W., Liu, D., Wu, B., 2010. Comprehensive benefit of flood resources utilization through dynamic successive fuzzy evaluation model: a case study. Sci. China Technol. Sci. 53, 529–538. Guo, H.C., Tang, J.W., 1995. A strategy for sustainable development of urban water environment and socio-economy. Acta Sci. Circumst. 15 (3), 363–369 (in Chinese). Gao, P., Song, Y.S., Chao, Y., 2012. Water function zoning and water environment capacity analysis on surface water in Jiamusi urban area. Proc. Eng. 28, 458–463.

146

T. Wang, S. Xu / Ecological Indicators 52 (2015) 134–146

Huang, G.H., Qin, X.S., 2008. Environmental systems analysis under uncertainty. Civ. Eng. Environ. Syst. 25 (2), 77–80. Hosseini, H.M., Kaneko, S., 2011. Dynamic sustainability assessment of countries at the macro level: a principal component analysis. Ecol. Indic. 11, 811–823. Johnson, S., Wang, G.X., Howard, H.D., Anderson, A.B., 2011. Identification of superfluous roads in terms of sustainable military land carrying capacity and environment. J. Terramech. 48, 97–104. Ke, L.N., Zhou, H.C., 2013. Model for the assessment of seawater environmental quality based on multiobjective variable fuzzy set theory. Abstr. Appl. Anal. 4, 1–9. Li, G., Jin, C.J., 2009. Fuzzy comprehensive evaluation for carrying capacity of regional water resources. Water Resour. Manag. 23, 2505–2513. Lu, L., Yi, l., Chen, J.N., Zhang, T.Z., Zeng, S.Y., 2011. Comparative analysis of environmental carrying capacity of the Bohai Sea Rim area in China. J. Environ. Monit. 13, 3178–3184. Li, X., Shi, J.P., Cao, H., 2011. Water environment carrying capacity of Erhai Lake based on index system and analytic hierarchy process. Acta Sci. Circumst. 31, 1338–1344 (in Chinese). Meng, L.H., Chen, Y.N., Li, W.H., Zhao, R.F., 2009. Fuzzy comprehensive evaluation model for water resources carrying capacity in Tarim River basin Xinjiang, China. Chin. Geogr. Sci. 19, 89–95. Na, L., Wang, L.C., 2011. Study on regional water environment carrying capacity during the process of urbanization: a case in Jining city. Adv. Mater. Res. 281, 253. Niero, M., Pizzol, M., Bruun, H.G., Thomsen, M., 2014. Comparative life cycle assessment of waste water treatment in Denmark including sensitivity and uncertainty analysis. J. Clean Prod. 68, 25–35. OECD, 1998. Towards Sustainable Development Environmental Indicators: Environmental Indicators. OECD Publishing. Singh, R.K., Mutry, H.R., Gupta, S.K., Dikshit, A.K., 2006. Development of composite sustainability performance index for steel industry. Ecol. Indic. 7 (3), 565–588.

Stasik, R., Czeslaw, S., Mariusz, K., Daniel, L., 2011. Evaluation of water resource in a small forest catchment in wielkopolska region. ROCZ. OCHR. SR. 13, 1679–1696. Shi, Y.S., Wang, H.F., Yin, C.Y., 2013. Evaluation method of urban land population carrying capacity based on GIS: a case of Shanghai, China. Comput. Environ. Urban Syst. 39, 27–38. Sills, D.L., Paramita, V., Franke, M.J., Johnson, M.C., Akabas, T.m., Greene, C.H., Tester, J.W., 2013. Quantitative uncertainty analysis of life cycle assessment for algal biofuel production. Environ. Sci. Technol. 47, 687–694. Tang, J.W., Guo, H.C., Ye, W.H., 1997. Environmental bearing capacity and its application on environmental planning. China Environ. Sci. 17 (1), 6–9 (in Chinese). Tao, J., Fu, M.C., Zheng, X.Q., Zhang, J.J., Zhang, D.X., 2013. Provincial level-based emergy evaluation of crop production system and development modes in China. Ecol. Indic. 29, 325–338. Wang, Q.S., Yuan, X.L., Zhang, J., Mu, R.M., Yang, H.C., Ma, C.Y., 2010. Key evaluation framework for the impacts of urbanization on air environment: a case study. Ecol. Indic. 24, 266–272. Wang, Y.K., Wang, D., Wu, J., 2011. A variable fuzzy set assessment model for water shortage risk: two case studies from china. Hum. Ecol. Risk. Assess. 17, 631–645. Wu, J.L., Zeng, H., Hong, Y., Ma, L., Xu, L.S., Qin, B.Q., 2012. Water and sediment quality in lakes along the middle and lower reaches of the Yangtze River, China. Water Resour. Manag. 26 (12), 3601–3618. Zhou, H.C., Wang, G., Yang, l.q., 2009. A multi-objective fuzzy pattern recognition model for assessing groundwater vulnerability based on the DRASTIC system. Hydrol. Sci. J. 44 (4), 611–618. Zhu, Y.H., Drake, S., Lü, H.S., Xia, J., 2010. Analysis of temporal and spatial differences in eco-environmental carrying capacity related to water in the Haihe River Basins, China. Water Resour. Manag. 24, 1089–1105. Zhou, H.C., Chai, Z.P., He, B., 2011. Study on water resources carrying capacity in Dalian based on variable fuzzy recognition model. S. N. Water Div. Water Sci. Technol. 9, 46–50 (in Chinese).