ATOLL RESEARCH BULLETIN NO. 566 CAN ORGANISMS ASSOCIATED WITH LIVE SCLERACTINIAN CORALS BE USED AS INDICATORS OF CORAL REEF STATUS? BY PATRICK SCAPS AND VIANNEY DENIS ISSUED BY NATIONAL MUSEUM OF NATURAL HISTORY SMITHSONIAN INSTITUTION WASHINGTON, D.C., U.S.A. NOVEMBER 2008 B Figure 1 . (A ) Hoga and (B) Kaledupa islands in the Tukang Besi Archipelago and (C) location of the survey sites. CAN ORGANISMS ASSOCIATED WITH LIVE SCLERACTINIAN CORALS BE USED AS INDICATORS OF CORAL REEF STATUS? BY PATRICK SCAPS '* AND VIANNEY DENIS 1 * ABSTRACT With the intention of knowing if coral associates can be used as indicators of coral reef status, we studied organisms associated with live scleractinian corals on different sites located around Hoga and Kaledupa islands in the Tukang Besi Archipelago of the south-eastern coast of Sulawesi in the Banda Sea, in Indonesia. Twenty meter long line intercept transects were used to estimate the number of coral colonies infested by coral associates. The number of coral associates found on each coral colony on the transect was recorded and corals were identified to the most precise level. Massive and encrusting coral colonies were also measured in order to estimate the densities of infestation per square meter of coral colonies. To link the assemblages of coral associates observed with the characteristics of the benthic habitats, the coral cover was estimated using a 0.5 m point intercept transect method. Three hundred seventy-six colonies (45%) were infested by a total of 2,815 coral associates. In total, coral associates amounted to 2,062 lithophagid bivalves (73% of total coral infestations), 306 dwelling hermit crabs of the genus Paguritta (10.9%) and 242 vermetid snail Dendropoma maxima (8.6 %). The most infested colonies belonged to the genera Montipora , Pavona and Porites. They represented 33%, 23% and 18% of the total number of colonies infested respectively. The highest densities of infestation were found for the boring Lithophaga spp. for which a density of 88 ind/m2 was noticed in encrusting corals of the genus Pavona. The density of lithophagid bivalves and the number of infested colonies were high in the most impacted site (Sampela) and one of the intermediately impacted site (Pak Kasim’s) whereas they were low in the most pristine site (Kaledupa). The other intermediately impacted site (Buoy 3) had an intermediate number of infestations. Despite the lack of any significant difference in biotic cover between the most pristine site and the intermediately impacted sites, a common gradient tends to emerge based on coral associates. Although having a high biotic cover, Pak Kasim’s suffers from a similar level of infestation as Sampela suggesting process of reef degradation previously experienced by the most impacted site. Our results suggest that coral associates can be used as indicators of coral reef status. Taboratoire de Biologie Animale, Universite des Sciences et Technologies de Lille, 59 655 Villeneuve d’Ascq Cedex, France. §Present Address : Laboratoire d’ECologie MARine (ECOMAR), Universite de la Reunion, 15 Avenue Rene Cassin, BP 7151, 97715 St Denis Messag Cedex 09, France. “"Corresponding author: e-mail: patrick.scaps@univ-lillel.fr Manuscript received 16 October 2008; revised 27 October 2008 2 INTRODUCTION Living scleractinian corals provide microhabitats for a large number of parasitic and commensal coral associates which use the tissue and skeleton of the coral colonies as substrata (Frank et ah, 1995; Floros et ah, 2005). Coral associates are defined as sessile invertebrates that live on or within the coral skeleton (Risk et ah, 2001) and whose apertures open through the living coral tissue (Scott, 1987). Many taxa are involved, including sponges, polychaetes, bivalves, tunicates and hydroids (reviewed in Scott and Risk, 1988). Most of these coral associates stress the coral to some degree, and some of them, particularly some sponges, polychaetes and bivalves can do considerable harm, at least to the skeleton (Sammarco and Risk, 1990; Smith and Harriott, 1988; Floros et al., 2005). Almost all coral associates are filter-feeding heterotrophs, hence would be expected to increase in numbers in water with elevated nutrient concentrations (Risk et al., 2001; Floros et al., 2005). In consequence, as suggested by Risk et al. (2001), the health of a reef may be evaluated by scouring the density of coral associates on massive corals. This is based on the theory that coral associate numbers will increase with organic loading: stressed corals will be less able to protect themselves from settlement and overgrowth (Risk et al., 1993). To date, little attention has been paid to the associates of living corals in Indonesia although this country lies within the triangle of biodiversity harbouring the most biologically diverse coral reefs in the world. So, the goal of this study was to document the community structure of coral asssociates and to link the different assemblages of coral asssociates on reefs around Hoga and Kaledupa islands in Indonesia with the health of these reefs. STUDY SITES AND METHODS This study was conducted on the reefs around the islands of Hoga and Kaledupa in the Tukang Besi Archipelago of the south-eastern coast of Sulawesi in the Banda Sea, in Indonesia (Elliot et al., 2001) (Fig. 1A and IB). This area is considered extremely important for global diversity, evolutionary biology and biogeography. Both islands are located in the Wakatobi Marine National Park (MNP) where a Marine Research Station run by Operation Wallacea is situated (Dioum, 2000). The Wakatobi MNP was established in 1996 and contains approximately 50,000 ha of coral reefs for a total area of about 1.39 million hectares. A Rapid Ecological Assessment conducted in the Wakatobi MNP in May 2003 (Pet-Soede and Erdmann, 2003) recorded 396 species of hermatypic scleractinian corals belonging to 68 genera and 15 families. In addition, 10 species of non- scleractinian or ahermatypic hard coral species and 28 soft coral genera were added to this list. Despite relatively low diversity of habitat type visited, coral species diversity was relatively high. This is an indication of Wakatobi’s position near the center of high coral biodiversity or “Coral Triangle”. The three major causes of degradation recorded in the Wakatobi MNP are: bomb fishing, crown-of-thorns starfish proliferation and bleaching (Turak, 2003). 3 The study took place in July and August 2005 with the help of volunteers and different scientists working for Operation Wallacea. Four sites were studied (Fig. 1C) and were selected with a gradient of degradation according to former investigations led by Operation Wallacea during the previous years (Crabbe and Smith, 2002 ; Crabbe and Smith, 2003) and visual observations conducted in the beginning of the study. Site Descriptions Table 1 provides information of the GPS position and the environmental characteristics of the different selected sites. The first site (Kaledupa, Buoy 1), close to the island of Kaledupa, was considered to be in pristine condition (Crabbe and Smith, 2002) and to have no obvious anthropogenic or sedimentation damage. The second site, Sampela, is located within proximity to the Bajo village of Sampela. This site was considered to be extensively impacted (Crabbe and Smith, 2002). There was, notably, a significantly higher hard corals species richness at Kaledupa than at Sampela. The two other sites, Buoy 3 and Pak Kasim’s, were considered to be intermediately impacted. Buoy 3 is located in a 1 km-long Non Fishing Zone (NFZ). Pak Kasim’s is the closest site from the non-fishing area. Table 1. GPS position and characteristics of the different sites (* means that the difference with the other sites is signifiant, p<0.01). Kaledupa Sampela Pak Kasim’s Buoy 3 Latitude South 05°28'22" 05°29'01" 05°27'569" 05°28'40" Longitude East 123°43'47" 123°45'08" 123°45T79" 123°45'45" Rugosity 0.56 (0.08) N=25 0.73 (0.10)* N=15 0.58(0.15) N=25 0.60 (0.19) N=23 Sedimentation rate Average of means (g d.wt.m' 2 .d' 1 ) 5.21 (1.01) N= 8 20.46 (2.12)* N= 9 7.25 (0.28) N= 6 Light attenuation coefficient (K) 0.16(0.01) N =5 0.24 (0.01) N = 4 0.12(0.01) N = 5 0.13 (0.01) N = 5 The chain method, which gives an indication on the reef complexity, showed a significant difference for the rugosity at Sampela (ANOVA One-Way, F=4.74, p<0.01) (Table 1). This method was performed using a chain laid over the substrata over five replicate sections of each transect. The straight-line distance occupied by the chain was measured, and the rugosity index calculated by dividing the total length of chain by the straight-line distance (McCormick, 1994). The sedimentation rate was measured using sediment traps at three locations at each site (Buoy 3 and Pak Kasim’s are considered as the same site as “Tioga” for the measurements of sedimentation rate). Table 1 shows the average of the means for these locations. An analysis of the variance shows a higher sedimentation rate at Sampela than at the other sites (ANOVA One Way, F=24.2, 4 p<0.01).The measurements of turbidity, corresponding to the amount of suspended sediment and plankton in the water column, was assessed using a Secchi disk and the measurements of the light attenuation coefficient (K), agreed to the results obtained by sediment traps and showed higher turbidity in Sampela. The salinity and the temperature were assessed at two depths (3 and 12 meters) and were relatively constant during the entire study period, whether between the different sites and between the two depths. The seawater temperature varied between 27 and 28°C and the salinity ranged from 32 to 33 %o. The maximal tidal range on Hoga is 2 meters but is typically 1.5 meters. Survey Method After the characterization of the sites, the study consisted of the selection of the macrobioeroder and coral predator organisms that were among the most common in the reefs around the islands. The majority of bioeroders are not immediately visible on the exterior and it has been suggested that their numbers and combined mass equal or exceed that of the surface biota (Kleemann, 2001). However, for reasons of convenience and adoption of a non-destructive method, only their visible parts were taken into account and studied. Moreover, due to the restrictions on coral sampling which do not allow close examination for sponge invasions, sponges boring in live corals were not studied. Observations conducted during the initial dives at the different sites led to a selection of nine groups of macrobioeroder or coral predator organisms. All these groups (see Table 2) are composed of organisms, which either carry out a bioeroder activity or are predators of living scleractinian corals with an activity sometimes considered more similar to grazing than to bioeroding. Due to problems in identifying the animals, apart from a few exceptions, the different groups were not identified to species level. Coral associates can be classified into two groups: endolithic bioeroders living within the coral skeletons and epilithic organisms living and feeding directly on exposed surfaces. Their activity can be of various intensities depending on the group considered. For example, the serpulid annelid Spirobranchus giganteus corniculatus can have a bioerosion rate reaching 1800 g of CaCCf per square meter by year (Glynn, 1997). The bioerosion rate of the boring bivalves of the genus Lithophaga can reach 9000 g CaC0 3 /m 2 /yr with a density sometimes reaching 1879 ind/m 2 (Glynn, 1997). The dwelling hermit crabs of the genus Paguritta can live either in polychaete tubes associated with hard corals (Schumacher, 1977; Miyake, 1978) or other adapted holes, without important bioerosive activity in this case but proof of an active past bioerosion; or in self-created boreholes in living corals (Tewinsohn, 1978) which can lead to significant bioerosion, depending on its density. The number of coral colonies infested by coral associates was estimated along a 20 m long Tine Intercept Transect (TIT) (English et al., 1997) at two different depths (6 m and 12 m). The first transect was laid randomly and the position of the following transects was based on the first one. Each transect was spaced by a gap of approximately 20 meters. For Pak Kasim’s and Buoy 3, three transects were deployed and surveyed at both depths. Only the upper part of the reef of Sampela was assessed because the environment at 12 meters consisted of a sandy slope and was devoid of corals. In consequence, only three 5 transects were surveyed at 6 meters. Moreover, due to the heterogeneity of the habitat in Kaledupa in comparison with the other sites, more replicates were assessed for this site. Thus, five transects were deployed and surveyed at each depth. The number of macrobioeroders and coral predators found on each coral colony on the transect was recorded and corals were identified to the most precise level. Massive and encrusting coral colonies were also measured in order to estimate the densities of infestation per square meter of coral colonies. The infestations were quantified with consideration of the species, and the locations and depths of the sites. Table 2. Major groups of macrobioeroders and coral predators found in the reefs during this study and their corresponding ecology. Group Ecology Mollusca Gastropoda 1. Coralliophila neritoidea (Tamarck, 1816) (Plate 1A) Corallivore 2. Dendropoma maxima (Sowerby, 1825) (Plate ID) Excavater 3. Drupella cornus (Roding, 1798) (Plate IB) Corallivore 4. Serpulorbis grandis (Gray, 1850)(Plate 1C) Excavater Bivalvia 5. Area ventricosa Lamarck, 1819 (Plate 2E ) Borer 6. Lithophaga spp. (Plates 2 A, 2B and 2C) Borers 7. Pedum spondyloideum (Gmelin, 1791) (Plates IF, 2C) Borer Other bivalvia ( Gastrochaena spp., (Plate 2D) Modiolus Borers philippinarum (Plates 1G, 2B, 2C and 2E) , Barbatia foliata (Plate IE). Annelida - polychaeta 8. Spirobranchus giganteus corniculatus (Grube , 1862) Borer and excavater (Plate 2F) Crustacea 9. Paguritta spp. Excavaters Tink with Coral Reef Status The living conditions of macrobioeroder and coral predator organisms as well as the link between their presence and the state of the coral habitat were analysed. Firstly, their distributions were linked with measured environmental parameters and the characteristics of the different studied sites. Then, to link the assemblages of macrobioeroders and coral predators observed with the characteristics of the benthic habitats, the coral cover was estimated using monitoring methods to assess the reef health. To describe the cover of the major functional groups and dominant coral taxa, a 6 Plate 1. Organisms associated with live scleractinian corals. (A.) Coralliophila neritoidea ; (B.) Drupella cornus ; The vermetid gastropods Dendropoma maxima (C.) and Serpulorbis grandis (D); (E.) The bivalve Barbatia foliata\ (F.) The scallop Pedum spondyloideum ; (G.) The mytilid bivalve Modiolus philipp inarum. 7 Plate 2. Organism associated with live scleractinian corals. (A.) Boring’s openings of dumbbell outline at the surface of a coral colonie of the bivalves Lithophaga spp.; (B.) The mussel Modiolus philippinarum and boring’s openings of Lithophaga spp.; (C.) Modiolus philippinarum , Pedum spondyloideum and boring’s openings of lithophagid bivalves; (D.) Boring’s openings with narrow siphon tubes of Gastrochaena spp.; (E.j Area ventricosa and Modiolus philippinarum; (F.) The Christmas tree worm Spirobranchus giganteus corniculatus. 8 0.5 m Point Intercept Transect (PIT) method (English et ah, 1997) was used on the same 20 m-long transect as the macroinvertebrate survey. To support this method, comparisons with datasets obtained from 50 m-long PIT (interval of 0.5 meter) were performed. No significant difference was found, which is a proof of a relative homogeneity of the benthic cover for these stations. Therefore, a 20 m-long PIT could be used in order to maximise the number of replicates. In addition, a dataset on the presence of “keystone species”, such as Acanthaster pi and or Diadema spp., which may have important ecological impacts on the reef was used to complete the analysis. Statistical Analyses Statistical analyses were performed with Minitab for parametrical and non- parametrical statistics. PRIMER v6 (Plymouth Marine Laboratory, Clarke and Warwick, 2001) was used for analysis of community. Cochran tests were used to test for homogeneity of variances before ANOVA. Turkey’s pairwise comparisons were used for post hoc comparisons. ANOSIM were performed to analyse similarities between sites after the ordinations (Multidimensional Scaling, MDS). RESULTS A total of 83 1 scleractinian coral colonies belonging to 39 genera was recorded during this study. A colony was considered infested when at least one of the studied organisms was found in the coral colony. Three hundred seventy-six colonies (45%) were infested by a total of 2,815 coral associates. In total, coral associates amounted to 2,062 lithophagid bivalves (73% of total coral infestations), 306 dwelling hermit crabs of the genus Paguritta{ 10.9%) and 242 vermetid snail Dendropoma maxima (8.6 %) (Fig. 2 a and b). Other coral associates were less numerous and their contribution to total coral infestations were negligible (less than 3%). Infestation Rate by Scleractinian Coral Genus The most infested colonies belonged to the genera Montipora , Pavona and Porites (Fig. 3a). They represented 33%, 23% and 18% of the total number of colonies infested respectively (Fig. 3b). These genera are also those that are the most common on the transects (Fig. 3c); the other genera representing less than 10 % of the total number of infested colonies (Fig. 3 b). The densities of infestation per square meter for massive and encrusting coral colonies are reported in Table 3. The highest densities of infestation were found for the boring Lithophaga spp. for which a density of 88 ind/m 2 was noticed in encrusting corals of the genus Pavona. Densities per square meter and per colony were positively correlated. For each of the studied group of organisms, the same trend was observed. For example, for the lithophagid bivalves, a positive correlation was noticed between the number of organisms per colony and per square meter of colony (R=0.871, p=0.01). 9 a. b Figure 2. (a. ) Number of coral associates recorded during this study ; (b.) Repartition of the different coral associates. Table 3. Density of coral associates for the major form and genera of scleractinian corals infested (ind. m 2 of colony' 1 ). CoralliophUa neritoidea Dendropoma maxima Drupella cornus Serpulorbis grandis Area ventricosa + other bivalvia Lithophaga spp. Pedum spondyloideum Spirobranchus giganteus corniculatus Paguritta spp. Total massive colonies 0,22 0,85 0,06 0,13 0,53 8,73 0,22 0,28 0,66 Total encrusting colonies 0,12 5,00 0,14 0,07 0,19 40,48 0,09 1,27 6,22 Massive Porites 0,84 2,05 0,12 0,00 1,57 6,88 0,72 0,48 2,17 Encrusting Pavona 0,00 5,62 0,00 0,00 0,09 88,10 0,09 1,31 5,62 Encrusting Montipora 0,23 6,04 0,17 0,06 0,35 21,84 0,17 1,86 9,87 Infestation by Site More than half of the coral colonies were infested in Pak Kasim’s (54%) and Sampela (52%) (Table 4) whereas 44% of the coral colonies were infested in Buoy 3 and only 32% in Kaledupa. The distributions of coral associates in the different studied sites are illustrated in Figure 4. All the coral associates were found in Pak Kasim’s and Sampela whereas Drupella cornus and Serpulorbis grandis were absent in Buoy 3 and Sampela respectively (Fig. 4). 10 c. 250 (f) Figure 3. (a.) Number of scleractinian coral colonies infested for the ten most infested genera; (b.) Percentage of the number of colonies infested corresponding to a given genus on the total number of infested colonies; (c.) Number of colonies for the ten most common scleractinian coral genera. Table 4. Total number of colonies recorded, infested, and percentage of infestation by site. Kaledupa Buoy 3 Sampela Pak Kasim's Total of colonies 203 242 126 260 Total of infested colonies 64 106 65 141 % of infested colonies 32 44 52 54 Within the same site, no significant difference was found between the two chosen depths (6 and 12 meters) in terms of the total number of infested colonies and the distribution of the studied organisms. In consequence, no statistical significant difference (ANOVA One-way) for the parameter “depth” was found. A highly significant difference between locations (ANOVA One-Way, F=18.42, p<0.01) was noticed only for the lithophagid bivalves. The distributions of coral associates in Sampela and Pak Kasim’s were significantly different from those in the other sites (Tukey’s pairwise comparisons). However, the difference between locations was significant for p-value <0.1 for the gastropods Coralliophila neritoidea and Drupella cornus and for the dwelling hermit crabs of the genus Paguritta (Fig. 4). Composition of coral associates communities by site (a. Kaledupa b. Buoy 3 c. Sampela d. Pak Kasim’s). 11 25 , 3 % - 6 , 9 % - 3 , 4 % - 5 , 7 % - 6 , 9 % - 2 , 3 % - 0 , 6 % c. b 13 , 6 % d. uLithophaga spp. □ Paguritta spp. □ Dendropoma maxima □ Sptobranchus giganteus comic ulatus ■ Dmpeiia cornus E Area ventiicosa + Other bivalvia □ Coralliophilla neritoidea □ Pedum spondyloideum □ Serpulorbis grandis Figure 4. Composition of coral associates communities by site (a. Kaledupa b. Buoy 3 c. Sampela d. Pak Kasim’s). Community Analysis. Comparison of coral associates communities by Non-metric Multidimensional Scaling (MDS) and analysis of similarity (ANOSIM) indicated a significant difference among sites (ANOSIM One-way, Global R = 0.691, p=0.001) (Table 5). In contrast, no significant difference was found with regards to depth. Table 5. Results of the ANOSIM tests comparing coral associates communities between the different sites. R Statistic Level (%) Global Test 0,691 0,1 Pairwise comparison sites Kaledupa / Sampela 1 0,3 Sampela / Buoy 3 1 1,2 Pak Kasim's / Kaledupa 0,994 0,1 Pak Kasim's / Buoy 3 0,846 0,2 Kaledupa / Buoy 3 0,274 2,6 Pak Kasim's / Sampela 0,191 16,7 12 Examination of the MDS plot (Fig. 5a) and cluster (Fig. 5b) showed a tendency of differentiation between the sites. Furthermore, pairwise comparisons of sites from ANOSIM resulted in R-values indicating important differences between Pak Kasim’s / Kaledupa, Sampela / Kaledupa, Sampela / Buoy 3 (with R-values > 0.9) and between Pak Kasim’s / Buoy 3 (R value >0.8). No difference was found between the other paired sites. However, with regards to their positions on the MDS plot, the stations seem to be positioned along a gradient between two sites: Kaledupa and Sampela. These results were confirmed after superposition of the Euclidian distance calculated and obtained from the cluster on the MDS. We observed two distincts groups of transects on the MDS and cluster, the first one including transects from Sampela and Pak Kasim’s and the second one including transects from Buoy 3 and Kaledupa. Transform. Squme roM Rnanttacf D 3 ESicideai toiler 2 U stress 0 . 0 $ a. SITES A Puk Kusim'S T? Kslf.lllfl."! □ Ssauipdu O Buoy T 2Qh t 5 1 * r"j 0 $. 0 — b. □ □ FA □ A A A A A i TrwifftiTB. Squv* mini Rescmhhncc D 3 5 a:tid«nn distance SITKS A Pak Kasim's y Kaledipa □ Sampela O Bucry i i r X M AVYVVVYVVYVOOOQ ■' ~~ ~ _ ■“ ■' ^ -■ Hi a !0 M M ® M SJ © m ® iP N f-J <*-J tfN ifi- 4CJ T -1 r-i r-j ^ ,- a It It +* Q GQ SR ■f