*2.1. Study Area*

Known for having over 100 lakes, Wuhan city (113◦41 E–115◦05 E; 29◦58 N–31◦22 N) is located in the interior of China, in the eastern part of Jianghan Plain. Located on the northern side of the Northern Tropic, the sub-tropical monsoon humid climate zone has a mean annual temperature of 15.8 ◦C–17.5 ◦C and an annual rainfall of 1150–1450 mm. Wuhan city has 143 lakes and the total area of lakes is 803.2 km2, which is the highest of all Chinese cities. Because of the high ecological value and the large number of lakes, the local economy is increasing rapidly. Additionally, various functions of these lakes play significant roles in the development of the city. But, as a result of the excessive increase of economy and other factors, the number of lakes is rapidly decreasing and some lakes are polluted to a certain degree, bringing adverse effects to sustainable urban development.

### *2.2. Sample Collection and Analysis Methods*

One hundred and forty-three lakes in Wuhan (Figure S1) were divided into large lakes ( ≥20 km2), medium lakes (10 km2–20 km2), and small lakes ( ≤10 km2) with the proportion of 3%, 6%, and 91%, respectively. The large and medium-sized lakes were small in number and important to the development of Wuhan city, so they were all included in the typical lakes. Other typical lakes in the small lake category were selected by combing the region and function further. Based on regionalization, the lakes could be divided into Central District, Dongxihu District, Caidian District, Hannan District, Jiangxia District, Huangpi District, Xinzhou District, Economic Development Zone, and the High and New Technology Development Zone of Donghu lake. The lakes have five main water functions: regulation, irrigation, water supply, aquaculture or planting, landscape or entertainment, and reservation. Most lakes had multiple functions, especially large lakes, while small lakes were relatively simple in function. According to the preliminary classification of region and function, the small lakes were selected comprehensively to contain all city regions and functions. The selection method is presented in Figure S2. Therefore, under the premise of ensuring the results are representative and scientific, a total of 21 typical lakes (Figure 1) were screened and selected from 143 lakes in Wuhan city based on comprehensive consideration of their characteristics of area, function, and region. Figure 1 was made based on our investigation's data and the ArcGIS software version 9.3 (Environmental Systems Research Institute, Redlands, CA, USA) (https://www.arcgis.com/index.html). The 21 selected typical lakes were Tangxun Lake (L1), Niushan Lake (L2), Baoxie Lake(L3), Luhu Lake (L4), Shenshan Lake (L5), Qingling Lake (L6), Huangjia Lake (L7), Yanxi Lake (L8), Wuhu Lake (L9), Houhu Lake (L10), Jingyin Lake (L11), Donghu Lake (L12), Nanhu Lake (L13), Moshui Lake (L14), Tanghu Lake (L15), Lanni Lake (L16), Zhongshan Lake (L17), Guanlian Lake (L18), Tonghu Lake (L19), Xiaozha Lake (L20), and Hougong Lake (L21).

**Figure 1.** The geolocation of the 21 lakes in Wuhan (ArcGIS software version 9.3 (https://www.arcgis. com/index.html)).

The design layout and sampling methods of lake water samples were strictly according to the Chinese Technical Specifications Requirements for Monitoring of Surface Water and Waste Water (HJ91-2002) and the Chinese Water Quality Technical Regulation on the Design of Sampling Programs (HJ495-2009), combined with the area size and hydrological characteristics of the studied lake. All water samples were collected into polytetrafluoroethylene bottles which were rinsed with lake water at least three times before using. Afterwards, water samples were acidified to pH 1–2 with H2SO4 and then kept in thermostats with ice bags, and transferred to the laboratory within 24 h. The water temperature (T), pH, dissolved oxygen (DO), electrical conductivity (EC), and Chl-*a* were measured by a portable water quality analyzer (HQ40d, HACH, Loveland, CO, USA) and the transparency (SD) was measured by the lead method in situ. In the laboratory, the concentration of TP was measured by Molybdenum Antimony Spectrophotometry (GB11893-89, China), the concentration of TN was measured by Alkaline Potassium Persulfate Digestion UV Spectrophotometry (HJ636-2012, China), and the concentration of CODMn was measured by Water Quality—Determination of Permanganate Index (GB11892-89, China). In addition, the total amounts of Cu, Zn, Cr, Cd, Mn, Fe, Ni, and Pb were measured with Atomic Absorption Spectroscopy (AAS ZEEnit 700P, Jena, Germany), the total amount of As and Hg were measured by Atomic Fluorescence Spectrometry (AFS-9730, Haiguang Instrument Co. Ltd., Beijing, China), both of the determinations were according to Water Quality—Digestion of Total Metals-Nitric Acid Digestion Method (HJ677–2013, China) and Water Quality—Determination of Mercury, Arsenic, Selenium, Bismuth and Antimony—Atomic Fluorescence Spectrometry (HJ694-2014, China).

To ensure the accuracy and reliability of analysis, parallel samples and blank samples were used to analyze error and at least 10% of each batch of samples was taken as parallel samples. If the relative deviation of the results was less than 20%, the analysis results were considered reliable. The analysis results of blank samples should be lower than the detection limits, so as to eliminate the pollution that might generate between processing procedures and determinations. The standard curve would be drawn for each sample analysis, and the correlation coefficient of the standard curve was not below 0.995.

### *2.3. Comprehensive Trophic Level Index (TLI) Method*

TLI is one of the comprehensive eutrophication evaluation methods taking chlorophyll a (Chl-*a*), total phosphorus (TP), total nitrogen (TN), transparency (SD), and permanganate index (CODMn) as the evaluation indicators [28,29]. It was widely used in trophic state assessment of lakes and rivers due to its diversity and applicability in evaluation indicators [30,31]. TLI takes Chl-*a* as the benchmark parameter, obtaining the corresponding weights of all parameters depending on the correlation degree between the benchmark parameter and the other parameters, and then obtains the TLI by a weighted algorithm [29]. The TLI model is as follows [32]:

$$TLI(\Sigma) = \sum\_{j=1}^{n} \mathcal{W}\_{\dot{j}} \times TLI(\dot{j}) \tag{1}$$

where *TLI*(Σ) is comprehensive trophic level index. *Wj* represents the corresponding weight of parameter *j. TLI*(Σ) represents trophic state index of parameter *j. n* is the numbers of evaluation parameters.

With Chl-*a* as the benchmark parameter, the normalized correlation weight of parameter *j* is as follows [32]:

$$\mathcal{W}\_{\text{j}} = r\_{\text{ij}}^2 / \sum\_{\text{j=1}}^n r\_{\text{ij}}^2 \tag{2}$$

where *Wj* represents the corresponding weight of parameter *j. rij* represents the correlation coefficient between benchmark parameter and parameter *j. n* is the numbers of evaluation parameters. Based on a eutrophication survey of Chinese lakes, the correlation coefficients *rij* between the benchmark parameter (Chl-*a*) and other parameters are *<sup>r</sup>*Chl-*a* = 1, *r*TN = 0.82, *r*TP = 0.84, *r*SD = 0.83 and *rCODMn* = 0.83 [28,32].

Trophic level indexes of each parameter are calculated as Equations (3)–(7).

$$TLI(\text{Chl} - \text{u}) = 10 \times (2.5 + 1.086 \ln \rho\_{\text{Chl} - \text{u}}) \tag{3}$$

$$TLI(TP) = 10 \times (9.436 + 1.624 \ln \rho\_{TP}) \tag{4}$$

$$TLI(TN) = (10 \times 5.453 + 1.694 \ln \rho\_{TN})\tag{5}$$

$$TLI(SD) = 10 \times (5.118 - 1.94 \ln \rho\_{SD}) \tag{6}$$

$$TLI(COD\_{\rm MN}) = 10 \times \left(0.109 + 2.66 \ln \rho\_{\rm COD\_{\rm MN}}\right) \tag{7}$$

where *ρ*Chl*-a* represents concentration of Chl*-a* (mg/m3), and *ρTP ρTN ρCODMn* represent concentrations of *TP*, *TN*, and *CODMn* (mg/L), respectively. *ρSD* represents transparency (m). The trophic state of the lakes is graded using continuous numbers from 0 to 100, as shown in Table 1 [28,32].

**Table 1.** Classification of eutrophication levels.


### *2.4. Health Risk for Heavy Metals in Lakes*

Health risk assessment is described as processes used to estimate event probability and probable degree of adverse health effects over a specific period [33–35]. Risk level of environmental pollutants to human beings depends on the body's exposure dose to the pollutants and the toxicity of the pollutants. There are two main pathways for human exposure to trace elements in water: ingestion and dermal absorption, ignoring exposure via inhalation [35,36]. The exposure dose can be calculated by Equations (8) and (9) [37,38].

$$ADD\_{\text{ing}} = \frac{C\_W \times IR \times EF \times ED}{BW \times AT} \tag{8}$$

where *ADDing* (μg/(kg·day)) represents the exposure dose through ingestion. In this study, the ingestion mainly refers to the intake through water from studied lakes. *Cw* is the mean concentration of trace element in water (μg/L). *IR* is the intake rate of water, including direct drinking rate and indirect drinking rate (L/day). *EF* is the exposure frequency to pollutants (day/year). *ED* is the exposure duration, and it means the length of time over which contact with the contaminant lasts (year). *BW* represents the body weight (kg). *AT* is the average time (day). For carcinogenic risk, *AT* is the average life expectancy of people [37,38].

$$ADD\_{drm} = \frac{\mathbb{C}\_W \times \mathbb{S}A \times \mathbb{K}\_P \times ET \times EF \times ED \times 10^{-3}}{BW \times AT} \tag{9}$$

where *ADDderm* (μg/(kg·day)) represents the exposure dose through dermal absorption. *SA* is the exposure area of skin (cm2). *Kp* is the dermal permeability coefficient of pollutants in water (cm/h), in this study, 0.001 cm/h for Cu, Cd, and As, 0.0001 cm/h for Pb, 0.002 cm/h for Cr, and 0.0006 cm/h for Zn [14,35], and *ET* is the exposure time (h/day). In this study, *ET* is 0.6 h/day. For the meanings of *Cw*, *EF*, *ED*, *BW*, and *AT*, please refer to Equation (8).

The health risks caused by environmental pollutants can be divided into carcinogenic risk and non-carcinogenic risks according to their properties. In general, carcinogens are of greater risk than non-carcinogens. Therefore, the risk of cancer caused by lake water is used as the assessment medium of lake health risk.

Carcinogenic risk is the product of daily exposure dose and cancer slope factor, which is shown in Equation (10). Under the assumption that there is no antagonism and synergism between pollutants, the integrated carcinogenic risk can also be identified as the sum of carcinogenic risks exposure by various pollutants via different pathways as shown in Equation (11). The EPA believes that carcinogenic risk value of human being is acceptable within 1 × <sup>10</sup>−4, while the maximum acceptable risk value recommended by International Commission on Radiological Protection (ICRP) is 5 × 10−<sup>5</sup> [14]. The significant difference between the two evaluation standards may mislead the decision makers in their final judgment. Furthermore, it should be noted that there is currently no official and uniform standard of acceptable risk value in China and many developing countries, which may lead to uncertainty and incomparability among different decision-makers. Therefore, risk classification was carried out in this study in order to make the evaluation results clearer and more intelligible. Risk levels were rated as five levels based on the Delphi method, assessment criteria of USEPA and ICRP, as well as existing research (Table 2) [38].

$$CR\_i = ADD\_i \times CSF\_i \tag{10}$$

$$CR = \sum\_{i} CR\_{i} \tag{11}$$

where *CRi* is the carcinogenic risk of trace elements through ingestion or dermal absorption, dimensionless. *ADDi* (μg/(kg·day)) is the daily exposure dose of carcinogenic pollutants. *CSFi* (kg·day/μg) is the cancer slope factor of carcinogenic pollutants. *CR* is the sum of *CRi. i* is the pathways of exposure. *n* is the kinds of trace elements.

**Table 2.** Levels and values of assessment standards.


*2.5. Fuzzy Comprehensive Lake Health Assessment Method (FCLHAM)*

As one of the most important human habitats, lakes provide a variety of service functions, and their health status is closely related to the survival and development of human beings. How to comprehensively and scientifically evaluate the health status of lakes is becoming an important concern in the field of environmental science and ecology. And it has extremely important application value for the monitoring and managemen<sup>t</sup> of lakes. Therefore, a fuzzy assessment method needs to be developed to efficiently identify comprehensive lake health states. Based on *TLI*, Health Risk Assessment framework for heavy metals, and fuzzy theory, it is of significance to explore a novel assessment method synthetically considering heavy metals' health risk, eutrophication risk, and fuzziness of the assessment system. Based on the fuzzy comprehensive evaluation theory [39,40], the comprehensive lake health state was defined as follows:

$$Risk = f(Risk\_E, Risk\_H) \tag{12}$$

where *RiskE* represents the eutrophication risk of studied lakes, which is assessed by TLI. *RiskH* represents the health risk for the heavy metals of studied lakes. And *f* represents the comprehensive lake health state calculation functions.

Fuzzy language recognition theory in fuzzy mathematics was used to identify the risk in this model. The comprehensive lake health state can be calculated as follows:

$$\text{Risk} = \stackrel{\sim}{\mathbf{C}} \cdot \stackrel{\sim}{R} = (\mathbb{C}\_1, \mathbb{C}\_2) \cdot \left( \begin{array}{ccccc} A\_1 & A\_2 & A\_3 & A\_4 \\ B\_1 & B\_2 & B\_3 & B\_4 \end{array} \right) \tag{13}$$

where *C* · *R* characterize the *f* in Equation (13). *C* is the weight values of *RiskE* and *RiskH. C1* and *C*2 was determined as 0.4 and 0.6 by the Delphi method, which indicated that the risk of lakes depends on the health risk of heavy metals more than the eutrophication risk by expert advices [41,42]. ∼ *R* is membership matrix for levels of *RiskE* and *RiskH*. *A*1, *A*2, *A*3, *A*4, and *A*5 represent membership degrees of five levels of *RiskE* (Table 1), and *B*1, *B*2, *B*3, *B*4, and *B*5 represent membership degrees of five levels of *RiskH* (Table 2).

Therefore, the comprehensive lake health state can be represented as a matrix with one row and five columns. The calculated comprehensive lake health state were divided into five levels as follows: (1) level I, low risk; (2) level II, moderate risk; (3) level III, considerable risk; (4) level IV, high risk; (5) level V, very high risk. The membership degree of each assessment factor plays a key role in the fuzzy comprehensive risk assessment, which is the basis of the foundation of comprehensive fuzzy assessment. According to Tables 1 and 2, the membership function of *RiskE* and *RiskH* was established, and the membership degree of each level can be calculated by the following formulas [39,40]:

(1) *RiskE*

∼

∼

$$u\_1(r) = \begin{cases} 1, r \in [0, 30) \\ (50 - r)/20, r \in [30, 50) \\ 0, r \in [50, +\infty) \end{cases} \tag{14}$$

$$\mu\_2(r) = \begin{cases} \ 0, r \in [0, 30) \text{ or } [60, +\infty) \\ \ (r - 30)/20, r \in [30, 50) \\ \ (60 - r)/10, r \in [50, 60) \end{cases} \tag{15}$$

$$u\_3(r) = \begin{cases} \ 0, r \in [0, 50) \text{or} [70, +\infty) \\ \ (r - 50) / 10, r \in [50, 60) \\ \ (70 - r) / 10, r \in [60, 70) \end{cases} \tag{16}$$

$$u\_4(r) = \begin{cases} 0, r \in [0, 60) \\ (r - 60)/10, r \in [60, 70) \\ 0, r \in [70, +\infty) \end{cases} \tag{17}$$

$$u\_5(r) = \begin{cases} \ 0, r \in [0, 70) \\ \ (r - 70) / 30, r \in [70, 100) \\ \ 1, r \in [100, +\infty) \end{cases} \tag{18}$$

(2) *RiskH*

$$u\_1(r) = \begin{cases} 1, r \in \left[0, 10^{-6}\right) \\ \left(10^{-5} - r\right) / \left(10^{-5} - 10^{-6}\right), r \in \left[10^{-6}, 10^{-5}\right) \\ 0, r \in \left[10^{-5}, +\infty\right) \end{cases} \tag{19}$$

$$u\_2(r) = \begin{cases} 0, r \in \left[0, 10^{-6}\right) \text{ or } \left[5 \times 10^{-5}, +\infty\right) \\ \left(r - 10^{-6}\right) / \left(10^{-5} - 10^{-6}\right), r \in \left[10^{-6}, 10^{-5}\right) \\ \left(5 \times 10^{-5} - r\right) / \left(4 \times 10^{-5}\right), r \in \left[10^{-5}, 5 \times 10^{-5}\right) \end{cases} \tag{20}$$

$$u\_3(r) = \begin{cases} 0, r \in \left[0, 10^{-5}\right) \text{ or } \left[10^{-4}, +\infty\right) \\ \left(r - 10^{-5}\right) / \left(4 \times 10^{-5}\right), r \in \left[10^{-5}, 5 \times 10^{-5}\right) \\ \left(10^{-4} - r\right) / \left(10^{-4} - 5 \times 10^{-5}\right), r \in \left[5 \times 10^{-5}, 10^{-4}\right) \end{cases} \tag{21}$$

### *Int. J. Environ. Res. Public Health* **2018**, *15*, 2617

$$u\_4(r) = \begin{cases} 0, r \in \left[0, 5 \times 10^{-5}\right) \\ \left(r - 5 \times 10^{-5}\right) / \left(10^{-4} - 5 \times 10^{-5}\right), r \in \left[5 \times 10^{-5}, 10^{-4}\right) \\ 0, r \in \left[10^{-4}, +\infty\right) \end{cases} \tag{22}$$

$$u\_{\mathbb{S}}(r) = \begin{cases} \ 0, r \in \left[0, 10^{-4}\right) \\ \ 1, r \in \left[10^{-4}, +\infty\right) \end{cases} \tag{23}$$

### **3. Results and Discussion**

### *3.1. Basic Parameters and Trace Element Concentrations in Surface Water from Studied Lakes*

Table 3 provides a statistical summary of the water quality parameters measured in the 21 lakes in Wuhan. The pH of lakes was basically between 9 and 10, and the highest was in L16, the lowest in L1. According to the five levels of standard limited values stipulated in the Chinese Environmental Quality Standard for Surface Water (GB3838-2002), the concentrations of Chl-*a* and EC in lakes were within their target water quality standard. With the exception of L1 and L3, which were slightly below the target, the DO of the other 19 lakes met the target water quality standards. However, the concentrations of TN and TP did not reach the standard. The TP concentrations of 18 of the 21 lakes did not reach the target water quality. There are four lakes (L10; L13; L15, L21) with higher TP concentrations than the Class V water quality standards, far from reaching the target water quality standards of GB3838-2002. The TN concentrations ranged from 2.16 to 5.48 mg·L−1, with the highest value in L15 and lowest in L4. Not only did all the studied lakes not meet their target water quality standards, but they also did not reach the limit of the Class V water quality standard of GB3838-2002, indicating that 21 lakes have been heavily polluted with nitrogen and have shown significant eutrophication pollution characteristics. In addition, the SD values of 16 of the 21 lakes did not reach the target water quality standard limit, and the CODMn concentrations of 12 lakes did not reach the target water quality standard limit, which also showed the serious pollution of lake water.

The detected heavy metal concentrations of 21 lakes in Wuhan are listed in Table 4. Cr was not detected in water samples. The concentration range of As is 1.237–12.148 <sup>μ</sup>g·L−1. Except for L20, the concentrations of As in the studied lakes was within the permissible limits of USEPA, WHO, and Chinese Ministry of Health (2007). The concentrations of Cd were all within the permissible limits of China, WHO, and USEPA. The results indicated that the concentration of carcinogenic heavy metals in lakes was not very high, and other related risks of lake health need to be further explored.

**Table 3.** Basic parameters and target water quality of 21 lakes in Wuhan.



**Table 4.** The concentrations of heavy metals in 21 lakes in Wuhan.

aWHO, 2008; b USEPA, 2009; c Chinese Ministry of Health, 2007.

### *3.2. Eutrophication State Analysis of the Studied Lakes*

The comprehensive trophic level index (TLI) of 21 lakes in Wuhan are calculated and shown in Table S1. L2 and L5 had relatively better water quality with TLI values of 49.14 and 49.44 (corresponding to the medium trophic condition). Unfortunately, most of the studied lakes presented eutrophication to different extents. Particularly, the TLI of lake L15 was higher than the other lakes, which indicated the most severe eutrophication status. The TLI values in the lakes L3, L4, L7, L8, L9, L11, L12, and L19 varied from 49.14 to 72.96, indicating light eutrophication. Furthermore, the other 10 lakes reached medium eutrophication states, decreasing in the order of L13 > L10 > L17 > L6 > L14 > L16 > L1 > L21 > L20 > L18. The main exceeding standard factors of each lake include TN, TP, SD, and CODMn. This result accords with the data listed in Table 3, which indicates that the cause of eutrophication in lakes is the result of the combined effect of reducing substances and nutrients in the water [14,43].

### *3.3. Health Risk Assessment for Heavy Metals in the Studied Lakes*

As the results show in Table S2, the values of CRCr and CRCd in all lakes were both below <sup>10</sup>−6, indicating that there was no carcinogenic risk of Cr and Cd. However, there was a certain carcinogenic risk of As because the CRAs values of 21 lakes exceeded <sup>10</sup>−6, decreasing in the order of: medium risk (L20 > L3 > L13 > L6 > L10 > L16) > low risk (L15 > L21 > L8 > L12 > L4 > L14 > L1 > L2 > L11 > L17 > L7 > L9 > L18 > L19 > L5). The maximum and minimum values of CRAs were 2.44 × 10−<sup>5</sup> and 2.48 × <sup>10</sup>−6, which were 24.4 times and 2.5 times than the lowest risk limit. Moreover, Table S2 indicated that the risk levels of lakes L3, L6, L10, L13, L16, and L20 were Grade III, which indicates that the pollution of carcinogenic heavy metals in these lakes should be given certain attention by relevant local departments. The risk level of the remaining lakes was Grade II, indicating that carcinogenic heavy metal pollution was not very serious, and not a current health risk concern.

### *3.4. Results of the Fuzzy Comprehensive Lake Health Assessment Method (FCLHAM)*

To sum up, we can see that some differences surely existed between the results of eutrophication and health risk assessment, which may confuse the decision-maker because these methods unilaterally focus on evaluating the eutrophication level or the health risks of heavy metals of the lakes. For example, the eutrophication level of L15 is very high and belongs to the severely eutrophic category. However, the result of health risk assessment of L15 is low risk. FCLHAM assigns weights to the results of TLI and CR in order to evaluate comprehensive lake health states more comprehensively and accurately. Then, comprehensive lake health state can be calculated. According to *ui*(*r*) arithmetic calculation, the assessment matrix is shown in Table 5. The five numbers contained in the assessment matrix represent the degree to which Risk belongs to the level represented by *ui*(*r*). According to the maximum membership principle, the closer the membership of *ui*(*r*) to 1, the higher the degree to which Risk belongs to the level represented by *ui*(*r*). As is shown in Table 5, average comprehensive lake health state decreased in the sequence of L20 (considerate risk level) > L1~L17, L19, L21 (moderate risk level) > L18 (low risk level). It indicated that L20 needs increased human, material, and financial resource investment and to be given priority regarding its governance. L1~L17, L19, and L21 need more attention, governance, and oversight to maintain their current state. Although L5's risk level was close to the low risk level, its low risk level (0.508) and moderate risk level (0.492) were too close to each other. Under the principle of maximum risk protection, L5 was determined as moderate risk. If the absolute difference between the memberships of two adjacent risk levels is less than 10%, the final risk level can be determined as the higher level.



Through comparative analysis, we can see: (i) only the heavy eutrophication of L15 also had a high health risk for heavy metals, with a level II risk assessment, indicating that the risk should be noticed. Comparing the results of FCLHAM, L15 belongs to the level II (moderate risk) category, which should attract attention. (ii) All 10 lakes at moderate eutrophication and 8 lakes in mild eutrophication were all rated at level II or level III, while the corresponding results of the FCLHAM levels are also level II or level III, except for L18, whose level was I (low risk). (iii) All three assessment results of L2 and L5, which were in the medium level of nutrition, were level II. That also provided further proof of the reliability of FCLHAM. The first two methods of L5 were level II, and the results of FCLHAM was level

I. The examples of L5 and L18 illustrated the more hierarchical and scientific results of FCLHAM than the other two evaluation methods, which makes up for the deficiency of the deterministic assessment.

### *3.5. Classification and Control Countermeasures of Studied Lakes*

It is understood that lakes in Wuhan are currently subject to cross-sectoral managemen<sup>t</sup> by functional classification. With the exception of the water sector, other departments, such as fishery, transportation, environmental protection, health, land, forestry, tourism, health, and other related departments, have a certain managemen<sup>t</sup> function for lakes. However, the pollution treatment of lakes depends on the managemen<sup>t</sup> and promotion of the environmental protection department, so it is necessary to consider the lakes' function and nutritional status in the lake classification. Based on the results of FCLHAM, the 21 studied lakes were classified into three categories: general management, enhanced management, and priority managemen<sup>t</sup> (Figure 2). Figure 2 was made based on the results of FCLHAM with ArcGIS software version 9.3 (https://www.arcgis.com/index.html). The low risk lake (L18) which was rated level I corresponds to common management; enhanced managemen<sup>t</sup> corresponds to the moderate risk lakes (L1–L17, L19, L21); and priority managemen<sup>t</sup> corresponds to the considerable risk lake (L20).

According to the characteristics of each kind of lake, we give countermeasures of lake management, specifically: for the common managemen<sup>t</sup> lakes, its eutrophication level and health risk for heavy metal are both low, so at present, it does not need a grea<sup>t</sup> deal of manpower and material resources for remediation, but observation and supervision policies should be implemented so that the lake stays in good condition. The enhanced managemen<sup>t</sup> lakes, whose risk is in the middle, should be given attention and some treatment measures should be taken. The priority managemen<sup>t</sup> lakes, whose levels of eutrophication and the health risks for heavy metals are both high, require greater human, material, and financial resource expenditure for remediation to prevent threats to the physical health of the surrounding residents. In addition, the corresponding policies and the necessary source investigation should also be carried out to protect the lake after treatment.

**Figure 2.** Management classification chart of 21 investigated lakes (ArcGIS software version 9.3 (https://www.arcgis.com/index.html)).

If the 143 lakes in Wuhan were classified by the "area-region-function" classification, they would be assigned to the same category as the representative lakes of the same type. At this point, we will attribute all of Wuhan's lakes to the three types of common managemen<sup>t</sup> lakes, enhanced managemen<sup>t</sup> lakes, and priority managemen<sup>t</sup> lakes. Depending on the characteristics of each type, a targeted approach to different types of managemen<sup>t</sup> for each type of lake is a more efficient way to manage many of Wuhan's lakes. Based on FCLHAM, a novel hierarchical managemen<sup>t</sup> system for urban lake health based on lake characteristics classification was obtained (Figure 3). Here, we have tested and verified the rationality, efficiency, and science of this novel hierarchical managemen<sup>t</sup> system for innovative lake managemen<sup>t</sup> of Wuhan. Therefore, this managemen<sup>t</sup> mode can serve as an effective reference for the environmental managemen<sup>t</sup> of urban lakes both at home and abroad, to manage urban lake health hierarchically and efficiently.

**Figure 3.** Workflow of the established hierarchical managemen<sup>t</sup> system.
