2-1 Introduction
Scleractinian corals are framework builders on coral reef ecosystems, they play a key role on ecological functioning of coral reefs and thus of particular importance to study the composition and structure of coral communities. Quantitative studies of coral communities started in 1970s (e.g., Loya 1972; Glynn 1976; Goodwin et al. 1976; Bak and Engel 1979; Bouchon 1981). These studies mainly focused on description of species composition, zonation, diversity pattern and possible factors affecting the observed distributions and structure of communities. According to the distribution of scleractinian species, coral communities exhibit a distinct biogeographic pattern of reduced biodiversity with increasing latitude (Stehli and Wells 1971; Veron 1995, 2000).
In a small scale of 400 km around Taiwan, well-developed coral reefs are found in southern Taiwan and non-reefal coral communities exit in northern Taiwan (Chen 1999).
Previous studies also suggested the factors controlling the global distribution of coral reefs included salinity, temperature, light, nutrients, wave exposure and hydrodynamic
coral communities usually have lower species diversity and coral cover (Crossland 1988;
Veron 1995), although higher species diversity and coral cover might occur in some specific cases (Harriott et al. 1999; Harriott and Banks 2002; Perry 2003). Coral species compositions in subtropical coral communities are also different from those of tropical coral reefs. van Woesik and Done (1997) showed that the abundances of massive
Porites and branching Acropora declined with increasing latitude, while foliaceous or
encrusting species such as Turbinaria and Montipora species increased with latitudes along the eastern coast of Australia. The absence of the dominant coral species of tropical reefs at high latitudes suggests that they are unable to establish and survive at sub-optimal environmental conditions such as cooler temperatures, lower light levels, and higher sedimentation rates (Perry and Larcombe 2003).
A variety of sampling methods have been applied to quantitative studies of coral communities including line-intercept transect, line-point transect, video transect, mapped quadrats, and photo quadrats (Loya 1978; Hill and Wilkinson 2004; Leujak and Ormond 2007). Line intercept transect is one of the earliest quantitative technique, in which a transect line is laid close to the bottom and benthic cover is calculated as the fraction of the total length of line (Loya 1978). Line-point transect is a recommended quicker method modified from line-intercept transect and now mainly used by monitoring programs such as Reef Check and Reef Keeper (Hodgson 1999; Hill and
Wilkinson 2004). Video transect method is a belt transect filmed by video cameras which can provide a permanent record of benthic organisms. This method has become increasingly popular due to the improvement of digital camera and camcorder (Brown et al. 2004; Houk and van Voesik 2006). Quadrat method has been used for monitoring reef communities since 1970s. Different assessment techniques such as detailed mapping of quadrats, point-sampling or outlining benthos of photo quadrat images can be used in determining the cover of different benthos inside the quadrat (Leujak and Ormond 2007). Although the performance of different coral community methods have been compared in many studies, a clear conclusion remains unavailable to date (Weinberg 1981; Ohlhorst et al. 1988; Done 1995; Lam et al. 2006; Nadon and Stirling 2006). However, when information of numbers and sizes of colonies or processes causing change is required, use of photo-quadrats will be preferred to line transect methods (Leujak and Ormond 2007), especially in a low coral coverage area (Fung 2007).
Yenliao Bay (25° 2' N, 121° 56' E) is a semi-enclosed bay located on the
Yenliao Bay, and a total of 106 scleractinians species in 43 genera and 14 families were recorded. The most specious families were Faviidae, Agariciidae and Pectiniidae and the growth forms were mostly massive or foliaceous. Although branching Acropora and
Pocillopora species were rare, this area still had well-developed coral communities in
northern Taiwan. However, no quantitative study on coral communities has been conducted in this area.
The aims of this study are to examine the community structure of scleractinian corals in northern Taiwan using both quadrats and quadrats along line transect methods, and to evaluate the application of these methods on studying coral communities.
2-2 Materials and methods
Study area
Yenliao Bay (25° 2' N, 121° 56' E) is a semi-enclosed bay located on the northeastern coast of Taiwan (Fig. 2.1). The bay is approximately 3 km long (north to south) and 1 km wide (east to west), and there are three rocky reefs (designated as A, B, and C) sited at depths between 0 and 15 m. These reefs are extensions of shale and sandstone formations from land which provide hard substrate for corals and other benthic organisms. Major environmental conditions include a low sea temperature (<
18°C) in winter, heavy sedimentation from runoffs of adjacent rivers during rainy
season, and physical and biological erosion. Consequently, coral reefs in Yenliao Bay are poorly developed and coral communities are characterized by depauperate scleractinian fauna (Yang and Dai, 1982).
Sampling
Belt quadrats
On each of the rocky reefs (A, B, C, Fig. 2.1), three to four permanent quadrats (each 1 x 10 m) were established on a relatively flat area (ca. 200 m2) in 2003 (Fig. 2.2).
These quadrats were photographed twice annually from August 2003 to September 2009 using a 1 × 1 m frame and a digital camera (Nikon D70 with a 12-24 mm lens in a Subal underwater housing) from above. The images of a quadrat were first processed in Adobe Photoshop and combined as a complete image of the quadrat in ArcGis. All visible scleractinian coral colonies on the images were identified to species. All coral colonies were mapped and their projected areas were measured using the ArcGis.
at three depths (5, 10, 15 m), except Reef A where the survey were conducted at 10 and 15 m in depth since its shallowest part was about 10 m. Along each transect, 10 quadarts (each 100 cm × 70 cm) were randomly placed and all benthic organisms underneath the quadrats were photographed by the same equipment. In total, 240 quadarts with 168 m2 in area were sampled from three reefs.
The images were first processed in Abobe Photoshop to reduce distortion. Then were analyzed with CPCe (Coral Point Count with Excel extensions, Kohler and Gill 2006) to quantify the proportion of benthic organisms and substrate types. Thirty random selected points were automatically spread on each image and codes corresponding to the recognized objects in the image were denoted. All corals were identified to species and the coverage of each coral species was calculated.
Data analysis
Both coral cover and colony number in August 2003 were used to reveal the structure of coral community. Community indices including the Shannon diversity index, Simpson index, and Species evenness index were also calculated. Cluster analysis and non-metric multidimentional scaling (nMDS) analysis of Bray-Curtis similarity measures (PRIMER, version 6, Clarke and Warwick 2001) using the coverage of each genera were employed to reveal the relationship of survey sites by the two
methods (belt quadrats in 2009, quadrats along line transect in 2009-2010). Analysis of similarity (ANOSIM) was applied to determine if there was any significant difference in species compositions using the two methods. Similarity Percentage (SIMPER, Primer, v.
6) analysis was performed to identify the coral genera that contributed to the similarities and difference between the two methods.
2-3 Results
In total, 1404 coral colonies of 69 scleractinian species in 26 genera and 8 families were recorded at the study sites in August 2003 (Table 2.1, 2.2). These corals cover 16.2% of the reef surface. The most specious family was Faviidae (31 species) and they accounted for 44.9% of the total number of species, followed by Acroporidae (17 species, 24.6%) and Agariciidae (7 species, 10.1%). Moreover, Lobophyllidae (5 species, 7.2%), Fungiide (4 species, 5.8%), Poritide (2 species, 2.9%), Pocilloporidae (2 species, 2.9%), and Plesiastreidae (1 species, 1.4%) were also recorded.
In terms of both colony number and coverage (Table 2.2), Faviidae was the most
Stylophora (11.2%), and Cyphastrea (7.1%). In terms of coral cover, the top 5 genera
were Favites (23.1%), Montipora (22.9%), Favia (11.4%), Cyphastrea (6.1%), and
Platygyra (6.1%).
Size-frequency distribution of all coral colonies in August 2003 are shown in Fig.
2.3. The log-transformed data of colony number showed major classes of colonies were intermediate size (32cm2~128cm2). Colonies of the largest size (> 2048cm2) and smallest size (< 4cm2) classes were relatively rare.
Comparisons of quantitative methods
The species compositions of coral community were similar using the two methods (Table 2.3). The most abundant families revealed by both methods were Faviidae (Belt quadrats: 49.2%, Quadrats along transect: 54.1%). However, coral cover, species number, and diversity indices were higher in quadrats along line transect method than in permanent belt quadrats (Table 2.4).
Although the result of nMDS showed a grouping pattern between Sites (Fig. 2.4), the result of ANOSIM suggested that the two survey methods were not separated with significant differences (R = 0.259, P = 0.3). In addition, the results of cluster analysis and nMDS grouping showed that the surveyed sites of the same methods tended to form a cluster while those at Site C were independent groups (Fig. 2.4, 2.5).
The results of SIMPER analysis showed that average similarities were 62.11%
within belt quadrats method and 67.27% within quadrats along line transect method (Table 2.5a). In belt quadrats method, Favites, Favia, Cyphastrea, Porites, Montastrea, and Platygyra were the most contributive genera. In quadrats along line transect method,
Favites, Favia, Goniastrea, Cyphastrea, Porites, and Platygyra were the most
contributive genera. The average dissimilarity was 39.26% between belt quadrats and quadrats along line transect method (Table 2.5b).
2-4 Discussion
Community structure
This study showed that the abundance and cover (1404 colonies/100m2, 16.8%
cover) of scleractinian corals at Yenliao Bay were comparable with those of other high-latitude or subtropical reefs. For example, in the Solitary Islands Marine Reserve, eastern Australia, coral cover ranged from 8.5 to 50.9% (Harriott et al. 1994). In Amakusa, southern Japan, coral colonies and cover were 2471 colonies/100 m2 and
Acroporidae, Agariciidae, and Lobophylliidae. Faviid species of massive or encrusting forms was the most abundant ones (31 spp. from 69 in total). The species composition in Yenliao Bay was similar to other high latitude coral communities (Veron and Done 1979; Harriott et al. 1994; Nozawa et al. 2008; Wicks et al. 2010). For example, Faviidae was the most specious family (22 spp. from 54) in Amakusa, southwestern Japan. In Kermadec Islands, New Zealand, The most common species were two Faviidae species i.e., Montastraea curta, Hydnophora pilosa and an encrusting
Montipora sp. In contrast to the scleractinian assemblages in southern Taiwan (Dai
1993), the abundance of main reef builders such as branching Acropora or Pocillopora species was unusually low. This may be the main reason for poor reef formation in study area. About 30 species of Acropora and 9 species of Pocillopora were recorded in southern Taiwan, while only A. solitaryensis, A. valida, P. damicornis, and Stylophora
pistillata were frequently found in Yenliao Bay. Moreover, low seawater temperature in
winter (< 18℃) is suggested to be the main factor that limit the growth rates and reef-building activities of hermatypic corals in this area (Yang and Dai 1982).
Previous studies indicated that environments with high wave energy favor compact growth forms of corals (Veron 1993; Hughes and Connell 1999). In this study, Faviidae was the most abundant family in Yenliao Bay. This is probably related to the morphological differences of coral species due to selective effect of strong waves
generated by occasional typhoons and seasonal monsoon that caused greater mortality to foliaceous or branching corals (Hughes and Connell 1999). Such a selective effect on coral species with different growth forms generally agrees with the observations in high latitude areas worldwide. In Amakusa, Japan, massive or encrusting forms (e.g., Faviidae) were the dominant morphologies of scleractinian species (Nozawa et al. 2008).
In the Kermadec, Julian Rocks and Cook Island, the low abundance of branching species (e.g., acroporids, pocilloporids) was suggested to be caused by high degrees of wave exposure (Wicks et al. 2010). At Lord Howe Island, Australia, reef-building species were able to establish assemblages, but only in lagoonal areas; they were uncommon on the seaward slopes where reef accretion capacity was very limited (Harriott et al. 1995). In addition, low water temperature in winter, strong monsoon, and high water turbidity may inhibit the recovery of laminar or branching species after typhoon disturbances. As a consequence, massive and encrusting species may become the dominant morphologies of coral communities in the study area.
to assess the accuracy of the monitoring methods (Weinberg 1981; Ohlhorst et al. 1988;
Lam et. al. 2006). The use of quadrat method is recommended for general monitoring program as it provides the most reliable estimation of the coral cover under most circumstances (Weinberg 1981; Fung 2007; Leujak and Ormond 2007). In this study, the differences between percentage of coral cover, number of species, and diversity indices by the two methods were similar. However, these parameters were all higher in quadrat along line transect method than those in permanent belt quadrat method (Table.
2.4). This may be due to the total area surveyed in quadrat along line transect method (126 m2) was larger than that in permanent belt quadrat method (100 m2).
According to the results of multivariate comparisons on coral community structure, there were no significant differences in species composition between permanent belt quadrat method and quadrat along line transect method (P = 0.3, ANOSIM). The result of ANOSIM also showed that the two methods were not separated (R = 0.259). High similarity of dominant genera was found within methods as well as low dissimilarity between methods (SIMPER, Table 2.5). The within group similarity (62.22 and 67.27%) was only higher than the between group similarity (60.74%). This gave additional evidence that samples generated from the 2 methods were not clearly separated from each other.
Based on the results of nMDS ordination plot and cluster analysis (Figs. 2.4, 2.5),
the community structure revealed by these methods showed higher similarity at Sites A and B. The separation of samples at Site C was probably related to the spatial heterogeneity of coral distribution. The results indicate the permanent belt quadrat method is appropriate for monitoring the dynamics of coral communities in Yenliao Bay.
Table 2.1 A list of scleractinian species recorded at survey sites in Yenliao Bay in August 2003.
Family Species Family Species Family Species
Acroporidae Montipora aequituberculata L. yabei F. flexuosa
M. efflorescens Pachyseris speciosa F. halicora
M. monasteriata Psammocora profundacella G. pectinata
M. spongodes Lobophylliidae Echinophyllia aspera G. retiformis
M. tuberculosa E. echinoporoides Cyphastrea chalcidicum
M. turgescens Lobophyllia hemprichii C. microphthalma
M. venosa Oxypora lacera C. serailia
Acropora solitaryensis Symphyllia agaricia Echinopora lamellosa
A. valida Faviidae Favia favus Montastera valenciennesi
Agariciidae Pavona explanulata F. rotundata Mycedium elephantotus
P. varians F. speciosa Plesiastreidae Plesiastrea versipora
P. venosa F. veroni
Leptoseris explanata Favites abdita Total: 8 families, 26 genera, 69 species
L. mycetoseroides F. chinensis
Table 2.2 The order of scleractinian genera by colony number and cover recorded in the study area in August 2003. Data in parentheses are percentages.
Genus No. of colonies Rank Genus Total cover (cm2) Rank
Table 2.3 A list of scleractinian species recorded at survey sites by two methods.
Family Species Belt quadrats Quadrats along line transect
Acroporidae Montipora aequituberculata X X
M. foveolata X X
Acropora solitaryensis X X
A. valida X X
Agariciidae Pavona explanulata X
P. varians X
P. venosa X
Leptoseris explanata X X
L. mycetoseroides X X
Pachyseris speciosa X
Euphyllidae Euphyllia ancora X
Pocilloporidae Pocillopora damicornis X X
Stylophora pistillata X X
Fungiidae Lithophyllon undulatum X
Leptastrea pruinosa X X
L. purpurea X X
Psammocora profundacella X X
P. superficialis X
Coscinaraea columma X
Lobophylliidae Lobophyllia hemprichii X
Symphyllia agaricia X
S. radians X
Echinophyllia aspera X X
E. echinata X
E. echinoporoides X
Oxypora lacera X
Faviidae Favia favus X X
F. laxa X X
Barabattoia amicorum X X
Favites abdita X X
Cyphastrea chalcidicum X X
C. microphthalma X X
C. serailia X X
M. valenciennesi X X
Echinopora gemmacea X
E. lamellosa X X
Table 2.4 Coral community parameters and diversity indices of two survey methods at study sites.
Surveys of permanent belt quadrats (PBQ) was conducted in September 2009, and those of quadrats along line transect (QLT) was in August 2009-April 2010.
Parameters/Diversity Indices PBQ QLT
Coral cover 9.34% 9.93%
Species no. 59 61
Shannon Index 3.19 3.66
Species Evenness 0.78 0.89
Simpson Index 0.93 0.97
Table 2.5 Results of SIMPER analysis comparing the coral community structure based on genera percentage cover data from two survey methods. (a) Similarity within group, (b) dissimilarity between groups. %: percentage of contribution.
(a)
Belt quadrat Quadrat along
transect
Coral genus % Coral genus %
Favites
15.01Favites
13.57Favia
12.64Favia
13.25Cyphastrea
10.21Goniastrea
12.13Porites
9.85Cyphastrea
11.19Montastrea
7.74Porites
7.63Platygyra
6.91Platygyra
7.15Similarity within
group 62.11 Similarity within
group 67.27
Fig. 2.1 Map of the study area showing three rocky reefs (A, B, C) and three study sites (blank squares).
Fig. 2.2 Layout and locations of 10 permanent quadrats on three reefs in Yenliao Bay, northern Taiwan.
0 50 100 150 200 250 300 350 400
1 2 4 8 16 32 64 128 256 512 1024 2048 4096
Colony size (cm2, log2 scale)
Number of colonies
Fig. 2.3 Frequency distribution of colony sizes of total coral colonies recorded in August 2003.
Fig. 2.4 nMDS ordination plot comparing the coral community structures using percentage cover of coral genera by quadrat along transect method (Line) and belt quadrat method (Quadrat), A, B and C denote sampling Sites A, B, C.
Fig. 2.5 Dendrogram showing the clustering of coral communities using percentage cover of coral genera by quadrat along transect method (Line) and belt quadrat method (Quadrat), A, B and C denote sampling Sites A, B, C.