A New Evaluation Method for Groundwater Quality Applied in Guangzhou Region, China: Using Fuzzy Method Combining Toxicity Index Fan Liu1,2, Guanxing Huang1,3*, Jichao Sun1, Jihong Jing1, Ying Zhang1
ABSTRACT: Groundwater quality assessment is essential for drinking from a security point of view. In this paper, a new evaluation method called toxicity combined fuzzy evaluation (TCFE) has been put forward, which is based on the fuzzy synthetic evaluation (FSE) method and the toxicity data from Agency for Toxic Substances and Disease Registry. The comparison of TCFE and FSE in the groundwater quality assessment of Guangzhou region also has been done. The assessment results are divided into 5 water quality levels; level I is the best while level V is the worst. Results indicate that the proportion of level I, level II, and level III used by the FSE method was 69.33% in total. By contrast, this proportion rose to 81.33% after applying the TCFE method. In addition, 66.7% of level IV samples in the FSE method became level I (50%), level II (25%), and level III (25%) in the TCFE method and 29.41% of level V samples became level I (50%) and level III (50%). This trend was caused by the weight change after the combination of toxicity index. By analyzing the changes of different indicators’ weight, it could be concluded that the better-changed samples mainly exceeded the corresponding standards of regular indicators and the deteriorated samples mainly exceeded the corresponding standards of toxic indicators. The comparison between the two results revealed that the TCFE method could represent the health implications of toxic indicators reasonably. As a result, the TCFE method is more scientific in view of drinking safety. Water Environ. Res., 88, 99 (2016). KEYWORDS: groundwater quality, toxicity, fuzzy, assessment, Guangzhou region. doi:10.2175/106143015X14362865227832
Introduction As an important resource of drinking water, groundwater is critical to human. The quality of groundwater for drinking purpose will directly affect human health. Consequently, groundwater quality assessment is essential for drinking safety. The guidelines published by World Health Organization (WHO) are commonly used for drinking water assessment (WHO, 2011). The quality standards for groundwater and drinking water in China, which are stricter than or equal to the values of WHO, 1
Institute of Hydrogeology and Environmental Geology, Chinese Academy of Geological Sciences, Shijiazhuang 050061, China.
2
China University of Geosciences (Beijing), Beijing, 100083, China.
3
Hebei Key Laboratory of Groundwater Remediation, Shijiazhuang 050061, China.
* Corresponding author: E-mail:
[email protected]; Tel.: þ86-311-67598553.
February 2016
are widely used for water quality evaluation (Huang et al., 2013; MHPRC, 2006; MLRPRC, 2008). Water quality assessment is mostly based on hydrochemical analysis (Zhang et al., 2012). Apart from some regular naturally occurring chemicals, many anthropogenic toxic contaminations are of increasingly concern in health safety. As a result, distinguishing the regular indicators and toxicity indicators is a necessary prerequisite for the consideration of health safety because the health implications of them are so different. In the Priority List of Hazardous Substances published by Agency for Toxic Substances and Disease Registry (ATSDR), toxicity points are given to each chemical substance (ATSDR, 2013), from which a corresponding toxicity index can be derived. The combination of the toxicity index and assessment methods highlights the effect of toxic substances on water quality for drinking purposes. The problems of water quality management are characterized by imprecision in objectives and water quality standards (Mujumdar and Sasikumar, 2002). The application of fuzzy mathematics creates a more objective and accurate evaluation method. Fuzzy synthetic evaluation (FSE) classifies samples for known standards, and the classification is determined by a matrix operation of the weighted vector with the fuzzy evaluation matrix (Lu et al., 1999). However, in the FSE method, the regular indicators and toxic indicators are not differentiated. The different health implications of regular indicators and toxic indicators still have not been reflected. To address this issue, the toxicity index is integrated into the weighted vector in the calculation process. By coding this algorithm, the toxicity index and fuzzy method are combined. This new method for water quality assessment is called the toxicity combined fuzzy evaluation (TCFE) method. Guangzhou of the Guangdong Province, China, is an industrial city. The rapid industrialization and urbanization process has inevitably led to water quality degradation. Surface water is the primary source of drinking water there. The improper discharge of wastewater into surface water is exacerbating this situation. Despite the adequate quantity of surface water, in some areas, it is not ideal for drinking according to the standards for drinking water quality in China. As a result, the groundwater resource, which is characterized by abundant quantity and a certain contamination blocking capacity, becomes a resource with widespread concern. In fact, parts of urban residents in Guangzhou depend on groundwater for drinking and domestic purposes (Sun et al., 2009). In this highly developed industrial city, great importance should be 99
Liu et al.
attached to the toxic chemicals in groundwater quality assessment, especially for the groundwater used for drinking purposes. Assuming that there are equivalent amounts of regular and toxic chemicals, which are exceeding the standards to the same extent, the impact of these toxic chemicals on human health is far beyond the regular components. However, the study on groundwater quality assessment of Guangzhou region with toxic effects taken into account has not yet been reported. The main objective of this paper was to evaluate the groundwater quality in Guangzhou region mainly based on China’s national standards for groundwater (MLRPRC, 2008), applying the FSE and the TCFE methods. By comparing the two results, the superiority and practicality of the TCFE method were demonstrated. Meanwhile, the reasons for their difference were clarified from a computational view. The assessment results will be the guidance for the control of some certain toxic chemicals in groundwater, and be beneficial to improve groundwater management in Guangzhou region. Methodology Study Area. The study area, Guangzhou region (112857 0 to 114803 0 E, 22834 0 to 23856 0 N), is located in the north-central part of the Pearl River Delta, covering an area of 7434.4 km2. The resident population of the area was 10.1 million in 2007 (Sun et al., 2009). Characterized by a subtropical monsoon climate consisting of a wet season (May through October) and a dry season (November through April), this region has the typical relative humidity of 77%, and the typical annual rainfall ranges from 1673 to 1909 mm (Duzgoren-Aydin, 2007). North of the study area is hilly covered with forests, and the southern area is mainly alluvial plain. Surface elevation of Guangzhou region ranges from below sea level to approximately 1210 m. The study area belongs to the Pearl River basin. The Pearl River is China’s second-ranked river in terms of water discharge and the nation’s third-ranked river in terms of sediment load (Liu et al., 2014). Pearl River system flows throughout the study area into the estuary of South China Sea. Water Sampling. The sampling plan was designed for groundwater pollution investigation. As a result, the sampling sites were strongly associated to the anthropogenic activities. In the sampling work in July 2006, 75 groundwater samples were collected in the domestic or abandoned wells in Guangzhou region. The locations of sampling points are shown in Figure 1. The investigation before sampling included field measurements of pH, electrical conductivity, dissolved oxygen, redox potential, and temperature analyzed by a WTW Multi 340i/SET multiparameter instrument (WTWGmbH, Weilheim, Germany). The sampling work was done at a fixed depth of 500 mm below the water table by using a stainless steel sampler. Before being stored in polythene cans, the water samples were filtered through 0.45lm filter membranes. Then, the samples were stored in portable cooler box below 4 8C and transported to an analytical laboratory. Analytical Techniques. The 75 groundwater samples were all analyzed at the Groundwater Mineral Water and Environmental Monitoring Center of the Institute of Hydrogeology and Environmental Geology, Chinese Academy of Geological Sciences. The hydrochemical indicators comprise Kþ, Naþ, Ca2þ, 2þ 2þ 2þ 2þ 2þ 3þ 2þ 2þ NHþ 4 , Mg , Cu , Zn , Ba , Be , Al , Se, Pb, Hg , Cd , Cr, 2 Ni, Mo, Co, Fe, Mn, As, F , Cl , I , NO3 , NO2 , SO4 , PO3 4 , 100
total hardness, and total dissolved solids (TDS). The detailed testing methods of each item are illustrated in Table 1. For all groundwater samples, the relative error is less than 65%. The descriptive statistics of the groundwater hydrochemical data are summarized in Table 2. Water Quality Standards Used in Assessment. According to the standards for groundwater quality in China, the water quality condition is divided into five levels (Table 3). Water quality of levels I and II are suitable for a variety of purposes. Level III is suitable for drinking, irrigation, and most industry. Level IV is suitable for irrigation and some industry. Only after being treated, can this water be consumed. Level V is not suitable for drinking. After the combination of toxic index, toxicity is highlighted. The water quality levels from I to V also represent the degree of toxic hazards to health, which is significant for water quality assessment. Water quality of levels I and II are also suitable for a variety of purposes. Level III is suitable for drinking and irrigation with weak toxicity. Level IV is suitable for irrigation with moderate toxicity and it should also be treated before drinking. Level V has with strong toxicity and not suitable for drinking. The standards used in quality assessment should be modified to fit the regional conditions. In the Guangzhou region, the groundwater is generally acidic; the highest value of pH is 7.37 and the lowest is 4.16 (Table 2). As a result, for pH standards, the limitation of alkaline is lifted. When the parameters to be evaluated are not included in the standards for groundwater quality in China or the standards are not presented in the form of five different levels for some indicators, level III is specified based on the known limit in other standards. The level III value of electrical conductivity was specified by Directive, E C (1998). The level III value of Ca, Mg, and PO4 were specified by Federal Republic of Germany (1990). The missing levels are defined by approximate multiples of the existing levels. For example, according to the groundwater quality standards of China, the pH value of I, II, and III levels are all 6.8. The missing pH levels of I and II are specified as 7 and 6.8 on the basis of the known levels. Similarly, the levels of Naþ, Ca2þ, Mg2þ, NHþ 4, Be2þ, I, Se, As, Mn, and TDS are specified according to this role. Toxicity Index. The toxic points of these specific parameters are given by referencing the Priority List of Hazardous Substances published by ATSDR (ATSDR, 2013). The parameters, which are not listed in this priority list, are assigned with toxic values of zero, such as pH, electrical conductivity, total hardness, and TDS (Table 4). Toxic index is deduced from toxic points by normalizing them to the range from 1 to 10. Fuzzy Membership Function. Introduced by Zadeh in 1965, the fuzzy set theory has been widely applied in assessment processes in imprecise environments throughout the world (Dahiya et al., 2007; Zadeh, 1965). The the FSE method is frequently used in the field of water quality assessment, which includes river water, reservoir water, and groundwater (Chang et al., 2001; Lu and Lo, 2002; Singh et al., 2008). The FSE method allows integration of information of various parameters into the evaluation process (Dahiya et al., 2007). The fuzzy membership functions can properly describe the main source of uncertainties involving a large-scale complex decisionmaking process (Chang et al., 2001). The liner membership functions are used to simplify the assessment process (Zhang et Water Environment Research, Volume 88, Number 2
Liu et al.
Figure 1—Locations of study area and sampling sites.
Table 1—Groundwater testing methods. Test theory / instrument WTW Multi 340i/SET multiparameter instrument (Germany) Spectrophotometry (Perkin-Elmer Lambda 35, USA) Inductively coupled plasma mass spectrometry (ICP-MS) (Agilent 7500ce ICP-MS, Tokyo, Japan) Gravimetric Ethylenediaminetetraacetic acid (EDTA) titration
February 2016
Test item pH, electrical conductivity, dissolved oxygen, redox potential 2 3 þ NO 3 , NO2 , SO4 , Cl , PO4 , NH4 , F , I Kþ, Naþ, Ca2þ, Mg2þ, Fe, Mn, Pb, Zn2þ, Cu2þ, Cd2þ, As, Hg2þ, Se, Al3þ, Cr, Ba2þ, Be2þ, Co, Mo, Ni TDS Total hardness
101
Liu et al.
Table 2—Descriptive statistics of groundwater hydrochemical data. Parameters
Mean
Min
Max
S.D.
pH 6.065 4.160 7.370 0.683 EC 430.360 41 2600 381.976 Eh 53.659 32 332 60.672 DO 3.398 0.78 7.83 1.543 23.723 1.780 127.900 23.101 Naþ Ca2þ 44.709 2.400 145.681 34.190 Mg2þ 4.473 0.240 22.363 3.967 NHþ 0.423 BDL 13.500 1.712 4 Fe 1.013 BDL 41.100 4.935 Cl 36.389 2.230 177.070 33.498 SO2 28.479 BDL 138.359 30.285 4 F 0.151 BDL 1.654 0.208 32.007 0.250 185.710 35.323 NO 3 TDS 318.037 35.150 988.170 230.710 Total hardness 129.992 15.010 422.200 98.089 Be2þ 0.00068 BDL 0.03800 0.00441 I 0.050 BDL 1.190 0.157 Zn2þ 0.042 BDL 0.180 0.040 Se 0.00043 BDL 0.00320 0.00076 Cu2þ 0.002 BDL 0.036 0.004 As 0.005 BDL 0.130 0.016 Hg2þ 0.00007 BDL 0.00095 0.00015 Cd2þ 0.00004 BDL 0.00050 0.00010 0.457 BDL 5.130 1.011 PO3 4 Al3þ 0.050 BDL 1.100 0.139 Ba2þ 0.076 0.001 0.480 0.083 Cr 0.002 BDL 0.025 0.005 Pb 0.003 BDL 0.050 0.006 Co 0.00069 BDL 0.01100 0.00160 Mo 0.001 BDL 0.008 0.002 Mn 0.178 BDL 2.840 0.436 Ni 0.003 BDL 0.011 0.003 NO 0.121 BDL 4.160 0.542 2
% C.V. 0.113 0.888 1.131 0.454 0.974 0.765 0.887 4.047 4.871 0.921 1.063 1.384 1.104 0.725 0.755 6.50931 3.147 0.939 1.75966 2.138 3.555 2.19774 2.21148 2.213 2.757 1.092 2.233 1.987 2.31148 1.613 2.453 1.233 4.482
Table 3—Water quality standard used in this assessment.
Al3þ As Ba2þ Be2þ Ca2þ Cd Cl Co Cr Cu2þ EC F Fe Hg2þ I Mg2þ Mn Mo Naþ NHþ 4 Ni NO 2 NO 3 Pb pH PO3 4 Se SO2 4 TDS Total hardness Zn2þ
for water quality II to IV ( j ¼ 2–4): 8 0; ðCi , Sij1 Þ > > > C S i ij1 > > ; ðSij1 , Ci , Sij Þ < rij ¼ Sij Sij1 > > > Sijþ1 Ci > > : Sijþ1 Sij ; ðSij , Ci , Sijþ1 Þ for water quality V ( j ¼ 5): 102
II
III
IV
V
0.005 0.001 0.01 0.001 100 0.0001 50 0.005 0.01 0.01 600 0.2 0.1 0.00005 0.05 10 0.01 0.001 50 0.01 0.005 0.001 2 0 7 1.5 0.001 50 300 150 0.05
0.05 0.005 0.1 0.0015 200 0.001 150 0.05 0.05 0.05 1200 0.5 0.2 0.0005 0.1 20 0.05 0.01 100 0.05 0.01 0.01 5 0.005 6.8 3 0.005 150 500 300 0.5
0.2 0.01 0.7 0.002 400 0.005 250 0.05 0.1 1 2500 1 0.3 0.001 0.2 50 0.1 0.07 200 0.5 0.02 0.02 20 0.01 6.5 6.7 0.01 250 1000 450 1
0.5 0.05 4 0.05 800 0.01 350 1 0.2 1.5 5000 1.5 2 0.001 1 200 1.5 0.5 300 1 0.1 0.1 30 0.05 5.5 15 0.1 350 2000 650 5
.0.5 .0.05 .4 .0.05 .800 .0.01 .350 .1 .0.2 .1.5 .5000 .1.5 .2 .0.001 .1 .200 .1.5 .0.5 .300 .1 .0.1 .0.1 .30 .0.05 ,5.5 .15 .0.1 .350 .2000 .650 .5
Units: Ion concentration, TDS, and total hardness (mg/L); pH (standard units).
Units: Ion concentration (mg/L); pH (standard units); EC ¼ electrical conductivity (lS/cm); Eh ¼ redox potential (mV); DO ¼ dissolved oxygen (mg/L); TDS (mg/L). BDL ¼ below detected level; S.D. ¼ standard deviation; C.V.(%) ¼ coefficient variation.
al., 2012). In this study, there are 5 water quality levels according to the standards (Table 3). The fuzzy membership function is expressed in 5 water quality levels, for water quality I (j ¼ 1): 8 1; ðCi , Sij Þ > > > < Sijþ1 Ci ; ðSij , Ci , Sijþ1 Þ ð1Þ rij ¼ Sijþ1 Sij > > > : 0; ðC . S Þ i ijþ1
I
8 0; ðCi , Sij1 Þ > > > < Ci Sij1 ; ðSij1 , Ci , Sij Þ rij ¼ S Sij1 > > ij > : 1; ðC . S Þ i ij
where rij stands for the fuzzy membership of indicator i to water level j, Sij stands for the ranges of each water quality levels according to the standards (Table 3), and Ci stands for the analytic value of water quality indicator i in one water sample. For the conditions when Cij ¼ Sij, the membership function is
Table 4—Toxicity index used in this assessment (EC ¼ electrical conductivity). Items
ð2Þ
ð3Þ
pH, EC, Naþ, Ca2þ, Mg2þ, Fe, Cl, SO2 4 , TDS, total hardness, I Cu, Al, PO3 4 , Cr, Mn F, Zn, Ba, Mo, NO 3 Se, Ni, NHþ 4 , NO2 Be, Cd, Pb, Co As, Hg
Toxic points
Toxic index
0 10 53 178 400 600
1 1.15 1.795 3.67 7 10
Water Environment Research, Volume 88, Number 2
Liu et al.
Figure 2—Analytical procedure for the toxicity combined fuzzy assessment model. expressed as rij ¼ 0.5, rijþ1 ¼ 0.5. The triangular membership function curves are well depicted in Figure 2. The five points on the x-axis represents the five water quality levels of each indicator. The fuzzy membership matrix R is expressed as 2 3 r11 r12 r13 r14 r15 6 r21 r22 r23 r24 r25 7 6 7 6 7 R ¼ 6 r31 r32 r33 r34 r35 7 ð4Þ 6 .. .. 7 .. .. .. 4 . . 5 . . . ri1 ri2 ri3 ri4 ri5 where i stands for water quality indicators. The weight of each water quality indicator is expressed as n Ci R Ci ð5Þ ai ¼ i¼1 Si Si where ai is the weight of indicator i after normalizing to (0,1), n is the number of all water quality indicators, and Si is the arithmetic mean value of indicator i of 5 levels. This weight calculation approach highlights the influence of a certain indicator’s quality level to this indicator’s weight. The weight matrix A is expressed as A ¼ ½ a1
a2
a3
. . . ai
ð6Þ
where i is the number of indicators. The fuzzy matrix B is expressed as B¼AR
ð7Þ
The assessment results for each sample can be determined by matrix B. For each sample, of all the 5 quality levels, the one with the maximum membership value is thought to be the water quality level of this sample. Weight Function of the TCFE Method. The function of the FSE method was given. The main basis of its weight function was the relative degree of change between the analytic data and the corresponding water quality standards. At present, in water quality standards, regular indicators emphasized particularly the taste characteristics of water quality, whereas the toxic indicators emphasized health implication. Under the same water February 2016
quality standards, the health implication of regular indicators was far less than that of toxic indicators. Therefore, in terms of drinking safety, the weight of water quality indicators should be reassigned based on their toxicity, instead of fully applying the weight function of the FSE method. The differentiation between the TCFE method and conventional FSE method is weight calculation. Based on the weight function of the FSE method, the TCFE method is combined with the toxic index of each indicator (Table 4). The weight function is expressed as n Ci Ci ai0 ¼ Ti 3 ð8Þ R Ti 3 i¼1 Si Si 0
where Ti stands for the toxic index of indicator i, and ai stands for the weight value of indicator i in the TCFE method. Accordingly, weight matrix in the TCFE method is expressed as A 0 ¼ ½ a10
a20
a30
. . . ai0
ð9Þ
The fuzzy matrix is expressed as B0 ¼ A0 R
ð10Þ
The detailed calculation process was depicted in Figure 3. The two evaluation programs were coded in the internal macro language. The assessment results of the FSE and TCFE methods are shown in Figure 4. Imprecise linguistic descriptions could be readily achieved on the general water quality conditions. However, precise decisions could not be easily determined when the many existing parameters were taken into account. This study in the Guangzhou region covered 75 groundwater samples and 31 indicators. As a result, the assessment results by using conventional FSE and TCFE methods were of high practical value. Results and Discussion Assessment results of the FSE method and the TCFE method are illustrated in Figure 4. Compared to the results of the conventional FSE method, there existed changes of all quality levels in the TCFE method. Specifically, for 75 groundwater samples, the proportion of level I was 38.7%, level II was 13.33%, level III was 17.33%, level IV was 8%, and level V was 22.67% in 103
Liu et al.
Figure 3—Triangular membership function curves of each indicator. applying the FSE method. While in the TCFE method, the proportions changed accordingly (Figure 4). The proportion of level I was 40%, level II was 25.33%, level III was 16%, level IV was 2.67%, and level V was 16%. In addition, there was significant reduction in the proportion of level IV and level V in the TCFE method. Significant increase was also observed in the proportion of level II, while the variation degree of level I and level III was indistinctive. Intuitive conclusions drawn from a simple contrast between the two methods in Figure 4 were that the results in the TCFE method appeared more optimistic. A
detailed comparison is depicted in Figure 5. The upward and downward arrows illustrate the direction and degree of level changes between water quality and toxicity. The upward arrows represent worse changes in water quality levels in the TCFE method, while the downward arrows represent better changes. It was clear that 15 samples became better, 8 samples deteriorated, and 50 samples remained constant. Concrete analysis was done on the variation of water quality in levels. Among the 29 samples of level I in the FSE method, 7 samples deteriorated when applying the TCFE method, with a
Figure 4—Comparison of results between fuzzy method and the TCFE method. 104
Water Environment Research, Volume 88, Number 2
Liu et al.
Figure 5—Detailed comparison of level changes for each water sample. variation of 24.41%. Meanwhile, 8 samples from other levels were changed to level I. Among the 10 samples of level II in the FSE method, 1 sample deteriorated, with a variation of 10%, 10 samples from other levels were changed to level II. Among the 13 samples of level III in the FSE method, 8 samples deteriorated, with a variation of 61.54%, 7 samples from other levels were changed to level III. Among the 6 samples of level IV in the FSE method, 4 samples deteriorated, with a variation of 66.67%. Among the 17 samples of level V in the FSE method, 5 samples deteriorated, with a variation of 38.46%. To clarify the causes and mechanisms of variation, the rate of variations in each level are shown in Figure 6, from which the trend of the overall water quality levels after applying the TCFE method could be derived. Positive values presented better changes and negative values presented worse. The proportion of samples deteriorating from better quality levels (level I and level II) was less than the proportion of samples getting better from moderate or poor levels (level III, level IV, and level V). This trend led to the overall improvement of assessment results in the TCFE method. Notably, the 2 level IV samples and 12 level V samples in the TCFE method were among the 17 level IV samples and 6 level V samples in the FSE method, respectively. There were no newly converted samples of level IV and level V in the TCFE method. This phenomenon suggested that samples of level IV and level V in the FSE method were not all toxic to humans. A percentage (66.7% of level IV samples and 29.41% of level V samples) turned better in the TCFE method. It could be concluded that the better-changed samples mainly exceeded the corresponding standards of regular indicators. Take, for instance, the samples of
G15 and G39, which were characterized by the largest degree of change. Because every indicator had a different weight value in each sample, in the weight calculation process of the FSE method, the weights of pH value were 50.23 and 54.16%, which were too high for this regular indicator with low toxicity (Table 4). Meanwhile, the pH values of the 2 samples were 5.27 and 4.73, which were lower than the standards of level V. These two factors resulted in the level V assessment results of G15 and G39 samples under the circumstances that there were no other indicators exceeding the corresponding standards obviously. However, after the application of the TCFE method, the weights of pH value for G15 and G39 were reduced to 20.96 and 30.04%. Coupled with the good quality conditions of other indicators, the assessment results of the 2 samples were changed to level I. For other better-changed samples, they mainly exceeded the corresponding standards of regular indicators (pH, Fe, and total hardness) or weak toxicity indicators, like NO 3 . In the FSE method, the proportion of level I, level II, and level III was 69.33% altogether, which also represented for potable water level. After applying the TCFE method, this proportion rose to 81.33%. In some specific areas, the assessment results of samples turned bad in the TCFE method, for the reasons that toxicity of indicators was emphasized in determining the water quality level. If indicators with strong toxicity exceeded the standards of level V, the results were bound to turn bad. Among the 8 deteriorated samples, G10 and G47 were characterized by the maximum variation. For the sample of G10, the relatively high concentrations of As (0.0091 mg/L) was the main inducements of water quality level change in the TCFE method. The weight of
Figure 6—The rate of variations in each level. February 2016
105
Liu et al.
As in G10 was changed from 6.51% in the FSE method to 34.30% in the TCFE method. For G47, the main inducement is Pb (0.01 mg/L). Similarly, the weight of Pb in G47 changed from 11.28 to 37.96%. Other indicators in G10 and G47 did not exceed the corresponding standards obviously. The quantitative toxicity data of As and Pb were 10 and 7 (Table 3); they were highly toxic indicators. It could be concluded that the deteriorated samples mainly exceeded the corresponding standards of toxic indicators. The increase of their weights in the TCFE method once again suggested that the toxicity of indicators was emphasized. In addition, it is known that the substances with toxic effects were dose related. The current groundwater standards lack detailed dosage instructions. The maximum intake or the longest duration under the acceptable daily intake should be appendixes. Further research is needed on these objectives. A long-term and effective groundwater management approach could be proposed based on the study, which in turn would be benefit for the endemic disease control. Conclusions The groundwater quality assessment of Guangzhou region consisted of 31 indicators and 75 samples. Compared with the FSE method, the character of the TCFE method was demonstrated. In the FSE method, the proportion of level I, level II, and level III was 69.33% altogether. After the application of the TCFE method, this proportion rose to 81.33%. In addition, 66.7% of level IV samples in the FSE method became level I (50%), level II (25%), and level III (25%) in the TCFE method and 29.41% of level V samples became level I (50%) and level III (50%). This trend was caused by the weight change after the combination of toxicity index. By analyzing the changes of different indicators’ weight, it could be concluded that the better-changed samples mainly exceeded the corresponding standards of regular indicators and the deteriorated samples mainly exceeded the corresponding standards of toxic indicators. The effects of regular indicators were weakened and health implications of toxic indicators were emphasized in the TCFE method. As a result, the TCFE method could be better representation in management decision-making. The corresponding targeted treatment scheme also could be determined more effectively. Acknowledgments This research was supported by the China Geological Survey Grant (1212011121167; 1212011220982) and the Basic Scientific Study Fund from the Institute of Hydrogeology and Environmental Geology, Chinese Academy of Geological Sciences (SK201410). Submitted for publication June 9, 2014; revised manuscript submitted September 7, 2014; accepted for publication October 14, 2014.
106
References ATSDR (2013) Priority List of Hazardous Substances. http://www.atsdr. cdc.gov/spl/resources/ATSDR_2013_SPL_Detailed_Data_Table.pdf (accessed Oct 14, 2014). Chang, N-B.; Chen, H-W.; Ning, S-K. (2001) Identification of River Water Quality Using the Fuzzy Synthetic Evaluation Approach. J. Environ. Manage., 63 (3), 293–305. Dahiya, S.; Singh, B.; Gaur, S.; Garg, V. K.; Kushwaha, H. S. (2007) Analysis of Groundwater Quality Using Fuzzy Synthetic Evaluation. J. Hazard. Mater., 147 (3), 938–946. Directive, E C (1998) 98/83/EC: The Quality of Water Intended for Human Consumption. Off. J. Eur. Communities: Legis., 330, May 12. Duzgoren-Aydin, N. S. (2007) Sources and Characteristics of Lead Pollution in the Urban Environment of Guangzhou. Sci. Total Environ., 385 (1), 182–195. Federal Republic of Germany (1990) Trinkwasserverordnung (Drinking Water Regulations) from 5 December 1990. Bundesgesetzblatt, I:2612–2629. Huang, G.; Chen, Z.; Sun, J. (2013) Water Quality Assessment and Hydrochemical Characteristics of Shallow Groundwater in Eastern Chancheng District, Foshan, China. Water Environ. Res., 85 (4), 354–362. Liu, F.; Yuan, L.; Yang, Q.; Ou, S.; Xie, L.; Cui, X. (2014) Hydrological Responses to the Combined Influence of Diverse Human Activities in the Pearl River Delta, China. Catena, 113, 41–55. Lu, R-S.; Lo, S-L. (2002) Diagnosing Reservoir Water Quality Using SelfOrganizing Maps and Fuzzy Theory. Water Res., 36 (9), 2265–2274. Lu, R-S.; Lo, S-L.; Hu, J-Y. (1999) Analysis of Reservoir Water Quality Using Fuzzy Synthetic Evaluation. Stoch. Environ. Res. Risk Assess., 13 (5), 327–336. Ministry of Health of the People’s Republic of China (MHPRC) (2006) Standards for Drinking Water Quality; Standards Press of China: Beijing, China. Ministry of Land and Resources of the People’s Republic of China (MLRPRC) (2008) Standards for Groundwater Quality; Exposure Draft: Beijing, China. Mujumdar, P.; Sasikumar, K. (2002) A Fuzzy Risk Approach for Seasonal Water Quality Management of a River System. Water Resour. Res., 38 (1), 1004. Singh, B.; Dahiya, S.; Jain. S.; Garg, V. K.; Kushwaha, H. S. (2008) Use of Fuzzy Synthetic Evaluation for Assessment of Groundwater Quality for Drinking Usage: A Case Study of Southern Haryana, India. Environ. Geol., 54 (2), 249–255. Sun, J. C.; Jing, J. H.; Huang, G. X.; Liu, J. T.; Chen, X.; Zhang, Y. X. (2009) Report on the Investigation and Assessment of Groundwater Contamination in the Pearl River Delta Area; Institute of Hydrogeology and Environmental Geology, Chinese Academy of Geological Sciences: Shijiazhuang, China (in Chinese). World Health Organization (2011) Guidelines for Drinking-water Quality, 4th ed.; World Health Organization: Geneva, Switzerland. Zadeh, L. A. (1965) Fuzzy Sets. Inf. Control, 8 (3), 338–353. Zhang, B.; Song, X.; Zhang, Y.; Han, D.; Tang, C.; Yu, Y.; Ma, Y. (2012) Hydrochemical Characteristics and Water Quality Assessment of Surface Water and Groundwater in Songnen Plain, Northeast China. Water Res., 46 (8), 2737–2748.
Water Environment Research, Volume 88, Number 2
