1. Introduction
Estuaries are ecotones between terrestrial and marine systems that receive large quantities of nutrients from tributary rivers [
1]. Given their unique characteristics, estuaries, like mangrove wetlands, certainly attract aquatic animals; these aquatic animals congregate in estuaries, which eventually become the habitat of such animals [
2]. Mangrove ecosystems are important inter-tidal estuarine wetlands along the coastlines of tropical and subtropical regions; these ecosystems are closely associated with human activities and are subject to contamination [
3]. However, with the rapid development of economic and human activities, more and more pollutants (e.g., domestic sewage, industrial wastewater, agricultural wastewater, and domestic wastewater) are discharged into seawater, thereby resulting in a series of environmental problems, such as organic over-enrichment, eutrophication, and health-related problems [
4]. Some adverse effects of pollution can be reversed by abatement actions [
4]. However, factors such as the high variability of environmental conditions and the existence of time lags in recovery responses make the evaluation of progress in achieving goals in pollution reduction difficult [
5]. Thus, timely investigations and assessments of coastal water quality are important for the protection of the littoral environment.
Hainan Island is the second largest island of China and has a mangrove area of 4772 hm
2, which accounts for one third of the total mangrove forest areas in China [
3,
6]. A total of 26 mangrove plant species are distributed in Hainan among the 27 species in China [
3,
6,
7]. As the first National Mangrove Reserve of China established in 1986, Dongzhaigang was included in the list of wetlands of international importance in 1992 to protect the mangrove wetland ecosystem [
3,
7]. This estuary and coastal mudflat ecosystem is on the edge of boreal tropics and offers protection of the mangrove wetland; this ecosystem is also an important habitat for wintering birds [
7]. Hence, the mangrove ecosystem has been always protected. However, in recent years, because of the increase in coastal settlement and rapid economic development in Dongzhaigang, tourism and aquaculture have led to a certain degree of pollution. According to research in 2013, hundreds of acres of mangrove of Dongzhaigang died because of water pollution [
8,
9]. This event was mainly caused by production and business activities near the protected areas, especially the intensive breeding of saltwater ducks; thus, the original biological chain was affected, which resulted in eutrophication and the proliferation of gribble [
8,
9]. Moreover, sewage from shrimp ponds and the swine industry in Qienan village in the Meilan district of Haikou City was directly discharged into the protected area such that excessive nitrogen and phosphorus were present in the seawater [
8,
9]. Due to the seriously polluted water in such an important protected area, there is a need to carry out a comprehensive health assessment on the water quality in Dongzhaigang.
In terms of seawater quality assessment, indicators of seawater quality assessment are few [
4,
5]. As for water quality evaluation method, the single factor method is simple and intuitive, so is frequently applied in water quality assessment, while the results of the fuzzy comprehensive evaluation method are more objective and comprehensive [
10,
11]. Researchers such as Aydinol conducted a statistical study and used the comprehensive evaluation method to evaluate seawater quality based on the investigation data around Princes’ Islands’ beaches in Istanbul [
10]. However, there are few indicators except total and fecal coliform concentrations in Aydinol’s study [
10]. In order to evaluate the seawater quality of the region thoroughly, water quality evaluation indicators should be comprehensive. Therefore, fuzzy comprehensive evaluation is clearly a more appropriate method. However, information on the comprehensive assessment of the water quality in the Dongzhaigang mangrove wetland is lacking. Researchers in China, such as Li Pengshan [
11], selected the II-Class of
Water Quality Standards of China (GB3097-1997) [
12], which clearly did not meet the real requirements. On the other hand, Li only used the index of organic pollution and the eutrophication level method to evaluate seawater quality through the organic pollutants and nutrients, which did not give a comprehensive evaluation of the water quality in Dongzhaigang [
11].Given the increasingly serious and complicated pollution situation, a simple study of nutrients or heavy metals in seawater cannot solve practical problems.
Since various factors are important to assess the water quality, determining the weight of each factor has become a vital problem. To solve a variety of optimization problems, including problems in which the objective function is discontinuous, non-differentiable, stochastic, or highly nonlinear, a genetic algorithm can be applied [
13]. The weight of each factor can be determined in a more objective way using a genetic algorithm and programs that simulate the logic of Darwinian selection [
13,
14]. The major advantage of the genetic algorithm is that it can tackle generally-shaped fuzzy membership functions on both sides of the model constraints [
15]. At present, the genetic algorithm is seldom used to assess seawater quality, which is used to calibrate a water quality model [
13]. The genetic algorithm has become the preferred water system design optimization technique for many researchers and practitioners [
14]. Researchers such as Xu used a combined genetic algorithm and fuzzy simulation approach to solve water quality management problems; Van Dijk optimized water distribution systems by using a weighted penalty in the genetic algorithm [
14,
15]. In this paper, the genetic algorithm was used to perform fuzzy comprehensive evaluation of seawater quality in Dongzhaigang.
Various factors need to be considered comprehensively in the latest survey and assessment of water quality in the entire range of Dongzhaigang. Given the results of the survey of the Dongzhaigang water environment in 2013, we evaluated the water quality of the samples based on the nutrients, organic pollutants, and heavy metals that were present. In addition, the genetic algorithm was used to perform fuzzy comprehensive evaluation. In this comprehensive evaluation, the results for organic pollutants, nutrients, and heavy metal pollution can be combined with the fuzzy comprehensive evaluation of water quality. Eventually, this study will provide crucial information on the health of seawater in the Dongzhaigang National Mangrove Nature Reserve, serving as a scientific reference for the protection of the Mangrove Nature Reserve and other similar regions.
The objectives of this study are as follows: (1) to obtain quantitative information on the concentration of nutrients, organic pollutants, and heavy metals in seawater at different sampling sites in the Dongzhaigang mangrove wetland of China; (2) to determine the distribution of these factors and their levels as compared with the criteria; (3) to evaluate the respective levels of organic pollution, eutrophication, and heavy metal pollution; and (4) to assess the water quality of this area by genetic algorithm-based fuzzy comprehensive evaluation.
2. Materials and Methods
2.1. Study Area
The Hainan Dongzhaigang National Mangrove Nature Reserve is located in the northeast of the Meilan district, Haikou City (110°30'–110°37' E, 19°51'–20°01' N). This reserve has most of the typical original natural mangroves in China. Dongzhaigang is a harbor-styled lagoon with a semi-closed shape (
Figure 1), with a length of 50 km; the total area is 3337.6 hm
2. The Dongzhaigang mangrove wetland has a tropical oceanic climate with monsoons, and the annual temperature reaches 23.8 °C; the annual duration of sunshine is up to 2200 h [
7].
Figure 1.
Sampling stations in the study area.
Figure 1.
Sampling stations in the study area.
2.2. Sample Collection and Processing
Surface seawater sampling was conducted by a research ship in Dongzhaigang National Mangrove Nature Reserve. Seawater samples were collected in July 2013, with a total of 13 stations distributed along the coastal area of the study area (
Figure 1). The sampling stations were chosen according to the criterion of even distribution in the entire study area. Samples were selected and collected with a 5-L organic glass surface sampler and transferred into glass bottles or plastic jugs. All sample processing followed
The Specification for Marine Monitoring. Part 3: Sample Collection, Storage and Transportation (GB17378.3-2007) [
16]; some samples that were used to measure DO, chemical oxygen demand (COD), and heavy metals were chemically processed on site by adding specific reagents (manganese chloride, potassium iodide, and nitric acid, respectively). After on-site chemical processing, all samples were stored in a dark place. The samples that were used to measure biochemical oxygen demand were placed into poly foam boxes with ice for cryopreservation. After the research ship docked daily, seawater samples were directly transferred to the laboratory for further experiments.
2.3. Analysis Factors and Methods
These analysis factors were included: NO
2−–N, NH
3–N, NO
3−–N, PO
43−–P, dissolved oxygen (DO), chemical oxygen demand (COD), five-day biological oxygen demand (BOD
5), oil, Si, PO
43−, NO
3−, NO
2−, NH
4+, PO
43−, Hg, total Cr, Cu, As, Zn, Pb, and Cd, for a total of 21 factors. Testing instruments included: an inductively-coupled plasma spectrometer (IRIS Intrepid II XSP), an inductively coupled plasma source mass spectrometer (X-SERIES 2), and a dual-channel atomic fluorescence spectrophotometer (AFS-3000). All experiments followed
The Specification for Marine Monitoring. Part 4: Seawater Analysis (GB17378.4-2007) [
17]: nitrate was tested by reduction through the cadmium column colorimetric method; nitrite was tested by
N-(
l-naphthyl)- ethylenediamine spectrophotometry; ammoniacal nitrogen was tested by hypobromite oxidimetry method; reactive phosphate was tested by phosphorus-molybdenum spectrophotometry; chemical oxygen demand was tested by basic potassium permanganate; BOD
5 was tested by five-day biochemical cultivation method; oil was tested by fluorescence spectrophotometry; silicate was tested by silicon molybdenum blue method; Hg was tested by atomic absorption spectrophotometry; As was tested by hydride generation atomic absorption spectrophotometry; Zn was tested by flame atomic absorption spectrophotometry; and Cr, Pb, Cd, and Cu were tested by non-flame atomic absorption spectrophotometry. The environment temperature was set to 21.7 °C, and the humidity was maintained at 40% throughout the experiment.
2.4. Evaluation Criteria
The Specification for Marine Monitoring. Part 4: Seawater Analysis (GB17378.4-2007) was used as the quality standard for seawater sample testing.
The People’s Republic of China Marine Water Quality Standard (GB 3097-1997) was used for the evaluation criteria in this study (
Table 1) [
12,
17].
2.5. Data Analysis
Data were processed with SPSS 12.0 and Excel 2003 to obtain the range of variation, average value, and the coefficient of variation of each index. ArcGIS 9.3 was applied to analyze the concentration of each index and the significant distribution. Simultaneously, the analysis index was classified into three categories, as follows: organic pollutants, e.g., DO, COD, BOD5, and oil; nutrients, e.g., reactive phosphate, dissolved inorganic nitrogen (DIN),and reactive silicate; and heavy metals, e.g., Hg, Cr, Cu, As, Zn, Pb, and Cd. The seawater quality was evaluated based on these three aspects.
In this study area, the index of organic pollution was used to assess the level of organic pollution, whereas the eutrophication level was applied to analyze the nutritional pollution. The water quality index model was used to analyze the monitoring index and pollution level of heavy metals. Finally, genetic algorithm-based fuzzy comprehensive evaluation was used to comprehensively evaluate the water quality of the whole sea area of Dongzhaigang.
Table 1.
The standard of seawater quality (mg/L).
Table 1.
The standard of seawater quality (mg/L).
Items | I | II | III | IV |
---|
DIN | 0.2 | 0.3 | 0.4 | 0.5 |
PO43−–P | 0.015 | 0.03 | 0.03 | 0.045 |
DO | 6 | 5 | 4 | 3 |
COD | 2 | 3 | 4 | 5 |
BOD5 | 1 | 3 | 4 | 5 |
Oil | 0.05 | 0.05 | 0.3 | 0.5 |
Hg | 0.00005 | 0.0002 | 0.0002 | 0.0005 |
Cr | 0.05 | 0.1 | 0.2 | 0.5 |
Cu | 0.005 | 0.01 | 0.05 | 0.05 |
As | 0.02 | 0.03 | 0.05 | 0.05 |
Zn | 0.02 | 0.05 | 0.1 | 0.5 |
Pb | 0.001 | 0.005 | 0.01 | 0.05 |
Cd | 0.001 | 0.005 | 0.01 | 0.01 |
2.6. Modeling Process of Genetic Algorithm-based Fuzzy Comprehensive Evaluation
To comprehensively evaluate the water quality of the whole sea area of Dongzhaigang, the fuzzy comprehensive evaluation model was constructed based on a genetic algorithm, with the following steps [
13,
14]: (1) establish the factor assembly and remark assembly; (2) confirm the weight with the genetic algorithm; (3) construct the membership matrix; and (4) calculate the comprehensive evaluation vector.
Step 1: Confirm the factor assembly matrix, remark assembly matrix, and orders of magnitude.
In this study, the DIN, PO
43−P, DO, COD, BOD
5, oil, Hg, total Cr, Cu, As, Zn, Pb, and Cd were selected as seawater quality evaluation indexes. These indexes were processed by a dimensionless method to construct the evaluation matrix
C while keeping the changing information of each index value. The index Do was processed by the following standardized formula:
Other indexes were processed by the following standardized formula:
where
xmax(
i) and
xmin(
i) were the maximum and minimum values of index
i in each station and
c(
i,
j) was the standardized evaluation index value.
Second, the judging matrix
B = (
bik)
n × m can confirm the weight of every evaluation index based on
c(
i,
j). The standard deviation was as follows:
where
was used to reflect the influence degree of each index on the comprehensive evaluation and to construct the judging matrix
B = (
bik)
n × m as follows:
where
Smax, and
Smin were the maximum and minimum of {
S(
i)/
i =
i →
n}, respectively. The relative importance parameter value could be expressed as:
where
int was a rounding function.
Step 2: Confirm the weight using the genetic algorithm.
Based on the judging matrix
B = (
bik)
n × m and
S(
i), we wrote the program with Matlab7.0. The weight range was (0,1), the parameter
d was 0.2, the population number was 100, and the hereditary algebra was 100. The weight matrix was
W = (
w1,
w2, …,
wp), where
Step 3: Construct the membership matrix.
The evaluation index was evaluated according to the evaluation scale, and the value of membership was confirmed. The final membership matrix R (ris)n × h could be constructed with the actual situation as the standard by considering the real characteristics of fuzzy pheromones and summarizing the experience of experts.
Step 4: Calculate the comprehensive evaluation vector and quantization value of comprehensive evaluation.
After the weight vector
W and the membership matrix
R were confirmed, the comprehensive evaluation vector was calculated according to the following expression:
The quantization value of remark assembly was defined by V = {vs}, s = 1,2 ,..., n. The remark conclusions are divided into four levels, namely, “good”, “normal”, “poor”, and “very poor”.
The quantization value of comprehensive evaluation was calculated by the following expression:
5. Conclusions
The data in this study suggest that the water monitoring indexes in seawater samples from the Dongzhaigang National Mangrove Nature Reserve more or less exceeded the I-Class seawater standard in July 2013. Among these, DIN standard was the most serious and was below the limits of only the IV-class seawater standard. The PO43−–P, DO, and Pb concentrations were below the limits of the II-Class seawater standard, whereas the oil and Zn concentrations were below the limits of the III-Class seawater standard. Meanwhile, the COD, BOD5, Cr, Hg, Cu, As, and Cd concentrations were below the limits of the I-Class seawater standard.
The index of organic pollution showed that the pollution level in this study area reached level 6 in 2013. Therefore, the sea area under study was seriously polluted. This result was mainly due to the increase in DIN concentration in recent years. The measured eutrophication level showed that the nutritional level was 4; that is, the seawater had a high level of eutrophication. Among all the sampling stations, stations S7 and S9 had relatively higher eutrophication. In addition, NO3−–N was the main form of DIN, thereby reflecting the fact that DIN in seawater followed a trend of excess accumulation. Phosphate was the limiting factor for phytoplankton growth in Dongzhaigang. The water quality index model showed that the sea area in this research was slightly polluted by heavy metals. The average concentrations of the seven metals in seawater were ranked Zn > Cr > Pb > As > Cu > Cd > Hg; Zn and Pb were the vital pollution factors in Dongzhaigang. The spatial variation of heavy metals showed a decreasing trend from the inner to outer parts of Dongzhaigang; the high concentration was probably caused by human activities in the inner part of the region, such as the waste water discharge of the swine industry and aquaculture. Moreover, the genetic algorithm-based fuzzy comprehensive evaluation indicated that the water quality of Dongzhaigang was not healthy.
Overall, the water quality in the inner parts of Dongzhaigang (southwest) is poorer than that of the external parts (northeast). The distribution of the increasing trend spreads from the near-shore to the offshore regions, with low distribution, terrestrial input, coastal river diluted water, and salinity as the main factors. The monitoring indicators of high concentration areas were detected in Beipai village of Yanfeng town, Shangyuan village, and the estuary of Sanjiang River (near stations S9 and S13), which indicated that more frequent and intensive human activities in these regions affected the coastal waters. Therefore, the Dongzhaigang National Mangrove Nature Reserve needs to be further protected by the government. Reasonable limits of human disturbance are necessary to reduce water pollution.