Fish Diversity and Fish Assemblage Structure in Seagrass Meadows at Sikao Bay , Trang Province , Thailand

We investigated the relationship between environmental variables as descriptors of the fish community, in terms of species diversity indices and fish assemblage patterns, in seagrass meadows at Sikao Bay, Trang Province, Thailand. Fish data and water quality parameters were collected from January to December 2012. A total of 10,596 fish specimens of 97 taxa in 48 families from four stations, Kham Bay (7 ̊30'9.21''N, 99 ̊18'7.67''E), Boonkong Bay (7 ̊31'2.49''N, 99 ̊17'40.09''E), Ban Pak Klong (7 ̊36'17.67''N, 99 ̊16'32.89''E) and Sai Cape (7 ̊38'30.19''N, 99 ̊14'49.85''E), were collected using gillnets of three different mesh sizes. The three most diverse families were Leiognathidae, Carangidae and Tetraodontidae, respectively. The dominant species were Atherinomorus duodecimalis, Sillago sihama and Pelates quadrilineatus. Specimens were highly abundant in July and less so in January and species richness was high in July and less so in March. Fish assemblages were classified into two patterns and the average of the Shannon index was 2.7. The environmental parameters in each month were analyzed by one-way ANOVA which did not show significant difference (P > 0.05) of pH, orthophosphate, wind speed and rainfall. Fish diversity and assemblage, and environmental parameters, were categorized into four groups. These could be promoted to local fisheries so that conservation programs cpuld be set up to ensure the ecological sustainability of seagrass meadows.


Introduction
Seagrass meadows are considered a valuable component of coastal ecosystems, as they are an important habitat for numerous fish, invertebrates and other animals, perform different ecological functions and act as environmental reserves.They supply a permanent habitat to fulfil the life cycles of aquatic animals, especially as a temporary nursery area for juvenile stages [1] [2], and as feeding areas for all life stages [3] [4].Seagrass meadows also provide benefits to humans in terms of protein source security, as well as occupation for coastal populations [5] [6].Seagrass beds have high primary productivity; the habitat provides suitable protective cover for many species of marine fish and invertebrates and is an important tropical marine [7].Variations in environmental variables are the major factors that govern marine fish communities in terms of species richness and distribution of individual species.They are important in terms of fish biology and ecology and related studies of life history and population dynamics.Understanding these relationships is tremendously important to support fisheries management, economics, social aspects and environmental protection.Seagrass meadows in particular are very sensitive ecosystems with a huge diversity of species and related assemblage patterns.In Thailand, huge seagrass meadows of 12 species extending to 18,986 hectares (ha) can be found.The biggest seagrass meadows in Thailand are located in Trang Province (3,435 ha) with 11 species, where Sikao Bay has 350 ha of seagrass meadows.However, fish diversity and assemblage in the seagrass meadows of Sikao Bay have not been studied thoroughly, and environmental factors in particular have not been considered recently.Our study examines fish diversity and assemblages, and relationships between species richness and environmental parameters.The results can be used in the future as a basis for comparison with other seagrass meadows in South East Asia and other tropical marine systems.They also support aquatic resources conservation strategies, which should be prepared in the near future to protect this sensitive seagrass meadow ecosystem.

Fish Sampling
Fish sampling was carried out during the night at four sites on Sikao Bay from January to December 2012.Fishing was done with three experimental gillnets of different mesh sizes (2, 3.5 and 5 cm), 180-m long and 1.5 m deep, connected into one net for fish sampling.Specimens were fixed in 10% formalin for 1 month and then changed to ethanol 30%, 50% for a month.Finally, they were preserved in 70% ethanol and re-checked and identified taxonomically to species level level [8]- [12] at the Maejo Aquatic Resources Natural Museum (MARNM), Faculty of Fisheries Technology and Aquatic Resources, Maejo University, Chiangmai, Thailand.

Environmental Parameter Sampling
The physico-chemical water quality parameters measured in situ using a YSI® 556 multi-Probe System (MPS) made in the USA were conductivity (Con), total dissolved solid (Tds), salinity (Sal), dissolved oxygen (Dis), pH (pH), water temperature (Tem) and transparency (Tra).Other parameters were sampled for laboratory analyses using the standard methods: ammonia (Amm), nitrite (Nti), nitrate (Nta), orthophosphate (Ort) and chlorophyll a (Chla) [13].Data were collected in Sikao Bay at the four fish sampling sites on a monthly basis from January to December 2012.Meteorological parameters data were maximum air temperature (Mat), minimum air temperature (Mit), average air temperature (Aat), wind direction (Wid), wind speed (Wis) and rainfall (Rai).These data were provided by the Thai Meteorological Department [14].The highest water level (Hwa) and lowest water level (Lwa) were provided by the Hydrographic Department, Royal Thai Navy [15].

Data Analyses
The physico-chemical water quality parameters, meteorological parameters, fish species (Fsp) and fish numbers (Fnu) were analyzed using Pearson rank correlation tests and one-way ANOVA (Tukey honestly significant difference test, P = 0.05) conducted with the SPSS © software package.The data set is presented as the percentage of occurrence frequency (% OF), species richness and Shannon diversity index [16], to examine the relationships between species richness and environmental parameters using Canonical Correspondence Analysis (CCA) and Principal Component Analysis (PCA), an ordination technique designed for direct analysis of relationships between multivariate ecological data [17].All statistical analyses were performed using R-statistical software, packages "stats" [18] [19].
The average Shannon-Wiener index was 2.7.

Environmental Parameters
The 20 environmental parameters, fish species and fish numbers were analyzed by Pearson rank correlation tests.
The results showed significance of conductivity, ammonia, fish numbers, fish species and highest water level (P < 0.05).The results analyzed by one-way ANOVA, the environmental parameters measured in each month did not show significant difference (P > 0.05) of pH, orthophosphate (Ort), wind speed (Wis) and rainfall (Rai).Four seagrass meadows and environmental parameters showed no significant relationships (P > 0.05).The environmental parameters measured in dry seasons and rainy seasons showed a significant difference (P < 0.05) of water temperature (Tem), wind direction (Wid), lowest water level (Lwa) and highest water level (Hwa).

Prediction of Species Richness and Environmental Parameters
Ninety-seven fish species and 20 environmental variables from four sampling sites were log-transformed and fed to a CART model as a response variable to be predicted using fish species and environmental variables.The major factor in physico-chemical and meteorological parameters was Chlorophyll a.The CART model using the tree "pruning" process and optimal tree selection favored three parameters including Chlorophyll a (Chl), lowest water level (Lwa) and total dissolved solid (Tds).Chlorophyll a was used in both the first and the second splits.This model showed that if chlorophyll a levels were higher than 0.5 µg/l, species diversity was high with about 26 species (Figure 4).
Further results then appear in a cluster dendrogram, summarizing the similarity of species in terms of environmental parameters (plotted by CCA) and showing the composition at the sampling site throughout the year (Figure 5(d)), which can also be classified into two main groups.The first group (A) represents the similarity in environmental variables at the end of the dry season (March and April) and at the beginning and central part of the rainy season (May, June, July, August and September).The second group (B) represents the similarity in environmental variables at the end of the rainy season (October, November and December) and at the beginning of the dry season (January and February).

Discussion
In this study, we found fish diversity in seagrass meadows in Sikao Bay, Trang Province, Thailand, was 97 taxa in 48 families relative with seagrass meadows, which is similar to the findings, [20] who found 81 fish species and to [21], who found 87 species of juveniles in their study of fish assemblages in Caribbean seagrass beds.[22] reported that 42 fish species, 36 species and 27 species were collected from seagrass.[23], reported 37 fish species belonging to 22 families when studying differences in fish assemblage structures between fragmented and continuous seagrass beds in Trang, Southern Thailand.[24], reported 36species of fish from 24 families from spatial patterns in fish herbivory in a temperate Australian seagrass meadow; and [25] noted 35 families.[26] found 65 taxa in 35 families of fish in seagrass beds at Kham Bay, Trang Province, Thailand; [27] identified 70 taxa from diverse species of fish in seagrass beds at Sai Cape, Trang Province, Thailand; and [28] were reported 62 taxa in 35 families of fish in seagrass beds at Ban Pak Klong, Trang Province, Thailand; thier fish diversity were less than in this study.[29] used poison as a sampling technique and found 189 species in 46 families, mainly from Thalassia.[30] found 249 fish species in 62 families from seine net catches; their fish diversity was greater than in this study.
In another study [31] found similar results, as the most abundant fish species were from the families Labridae, Siganidae, Atherinidae, Pomacentridae and Nemipteridae, with variations between the study sites.[23], reported fish species such as Clupeidae, Plotosidae, Leiognathidae, Sillaginidae, Lutjanidae, Lethrinidae and Gerreidae in  continuous and fragmented beds at the study sites.[32] found that permanent residents were defined by the presence of all life history stages within the habitat.Some common species (Cheilio inermis, Halichoeres argus, Halichoeres chloropterus, Pentapodus trivittatus, Apogon margaritiphorus) were found regularly as both adults and juveniles in the seagrass beds.Other species of reef-associated families that might use adjacent seagrass beds as nurseries were found exclusively as juveniles, such as Chaetodontidae, Haemulidae and Ephippidae.
Another finding was that environmental parameters are important factors for fish distribution patterns.This was proved by CCA, which showed the relation with species richness and indicated patterns in seagrass meadows.The major factors important for fish assemblage were Chl a, lowest water level (Lwa) and total dissolved solid (Tds), in relation with season change, fish abundance, fish species and fish species richness.These results are similar with [33] where spatial and seasonal variations of fish assemblages in mangrove creek systems in Zanzibar (Tanzania) were investigated.
Overall, 97 fish species and 20 environmental variables are analyzed by CCA, classifying the relationships of fish assemblage structure and environmental parameters into four groups.Two cluster dendrograms summarize the similarity of species richness with environmental parameters, and of environmental parameters with composition in the sampling site throughout the year, respectively.Our study has clearly improved knowledge of fish diversity and the relationship of fish assemblage with environmental parameters in seagrass meadows, and this is important for understanding the functioning and integrity of these ecosystems.Our results can be passed on to local fishermen and responsible stakeholders to encourage them to begin a Thai seagrass meadows conservation program for ecological sustainability in an area with growing tourism.

Figure 2 .
Figure 2. Number of dominant species and percentage of occurrence frequence of fishes in seagrass areas at Sikao Bay, Trang Province, Thailand.

Figure 3 .
Figure 3.The abundant species (a) and species richness (b) of fishes in seagrass meadowns at Sikao Bay, Trang Province, Thailand.

Figure 4 .
Figure 4. CART model to predict species richness in seagrass meadowns at Sikao Bay, Trang Province, Thailand.

Figure 5 .
Figure 5. CCA ordination with species richness related to environmental factors in seagrass meadowns at Sikao Bay, Trang Province, Thailand.(a) Environmental variables loading to CCA axex; (b) PCA analysis of fish assemblage; (c) dendrogram of fish assemblage and; (d) dendrogram of environmental parameters and fish assemblage (RKB = Rain season of Kham Bay, RBB = Rain seasion of Boonkong Bay, RPK = Rain season of Ban Pak Klong, RSC = Rain season of Sai Cape), (DKB = Dry season of Kham Bay, DBB = Dry seasion of Boonkong Bay, DPK = Dry season of Ban Pak Klong, DSC = Dry season of Sai Cape) and (1-12 = Jan.-Dec.).

Table 1 .
Fish species found in the seagrass medowsin Sikao Bay, Trang Province, Thailand.