[karimwassef]

Abstract

We studied the removal of dissolved organic carbon (DOC) by coral cavities of 50-250 dm3 at a depth range of 5-17 m along the coral reefs of Curaçao, Netherlands Antilles, and the Berau area, East Kalimantan, Indonesia. We found significantly lower DOC concentrations in cavity water compared with ambient reef water. On average, DOC concentrations in cavity water were 15.1 ± 6.0 µmol L-1 (Curaçao) and 4.0 ± 2.4 µmol L-1 (Berau) lower than in reef water. When the cavities were closed, DOC concentrations in the cavities declined by 22 ± 8% and 11 ± 4% in Curaçao and Berau, respectively, within 30 min. This corresponded to average DOC removal rates per cavity surface area of 342 ± 82 mmol C m-2 d-1 in Curaçao and 90 ± 45 mmol C m-2 d-1 in Berau. Bioassays showed that bacterioplankton are not responsible for this DOC removal by coral cavities. DOC fluxes exceeded bacterioplankton carbon (BC) fluxes into cavities by two orders of magnitude. On average, BC fluxes per cavity surface area were 3.6 ± 1.3 mmol C m-2 d-1 (Curaçao) and 1.9 ± 1.3 mmol C m-2 d-1 (Berau area). The net DOC removal per square meter of cryptic surface likely exceeded the gross primary production per square meter of planar reef area. We conclude that coral cavities and their biota are net sinks of DOC and play an important role in the energy budget of coral reefs.

Introduction

Coral cavities are among the largest and least-known habitats in coral reef environments. Their total volume comprises up to two-thirds of the reef volume (Garret et al. 1971; Ginsburg 1983) and their inner surface represents 60-75% of the total available surface of the reef (e.g. Jackson et al. 1971; Logan et al. 1984; Scheffers 2005). Yet, hardly anything is known about the ecological role of this cryptic habitat in the carbon cycling on the reef. The relatively sheltered cryptic habitat is inhabited by a high abundance of different organisms, called cavity dwellers or coelobites (Ginsburg and Schroeder 1973). The biomass of this cryptofaunal community might exceed that of the reef surface (Hutchings 1974; Brock and Brock 1977; Meesters et al. 1991) and the encrusting biota can cover more than 93% of the available hard substrate (Richter and Wunsch 1999; Richter et al. 2001; Scheffers 2005). As a consequence, the competition for space is high in coral cavities (Jackson et al. 1971; Buss 1979; Buss and Jackson 1979). Heterotrophic organisms generally dominate the coelobite community due to low light conditions in the cavities. Two-thirds of the cavity walls are inhabited by suspension feeders (sponges, tunicates, bryozoans, bivalves, and polychaetes), with sponges usually dominating this group. Approximately one-third of the cavity walls consists of calcareous algae (e.g. Vasseur 1974; Gili and Coma 1998; Wunsch et al. 2002).

The large area of the cryptic habitat and the high cover of encrusting organisms provide a potentially important interface in the exchange of material between the cavities and the overlying water column. Several studies showed a depletion of phyto-, nano-, pico-, and bacterioplankton in waters overlying coral reefs (e.g. Ayukai 1995; Yahel et al. 1998; Van Duyl et al. 2002). Gast et al. (1998) were the first to describe bacterioplankton depletion and accumulation of inorganic nutrients in coral crevices of Curaçao compared with overlying reef water. On the reefs along Curaçao, bacterial abundance was usually lower inside cavities than outside and the inorganic nutrient concentrations often differed from that in overlying water. Scheffers et al. (2004) and Van Duyl et al. (2006) reported bacterial removal rates by coral cavities of Curaçao of, on average, 3 mmol C m-2 d-1. A net influx of chlorophyll a was found in framework cavities in the Red Sea at an estimated phytoplankton removal rate in cavities of 75 mmol C m-2 d-1 (Richter et al. 2001).

Phytoplankton and bacterioplankton make up the major part of the particulate organic matter (POM; see List of Abbreviations) in the oligotrophic reef waters. However, by far the largest component (>97%) of organic matter is dissolved organic matter (DOM) (Benner 2002), which, in turn, is the largest carbon standing stock in the oceans (Martin and Fitzwater 1992). This fraction is operationally defined as the organic carbon passing through a fine filter, typically GF/F (Benner 2002; Carlson 2002). Dissolved organic carbon (DOC) is composed of a small labile and a much larger refractory fraction, which is not readily available to bacteria (Carlson 2002). DOC levels are usually enhanced over coral reefs and in lagoons compared to ocean surface waters (e.g. Johannes
1967; Ducklow 1990; Torréton et al. 1997), indicating that in reef waters, the production of DOC exceeds its loss (Van Duyl and Gast 2001). Sources of DOC in coral reefs are, for example, the release of DOC by benthic algae as a function of photosynthesis (Mague et al. 1980; Zlotnik and Dubinsky 1989), and release of DOC by corals through mucus production (Johannes 1967; Richman et al. 1975) or as free amino acids (Schlichter and Liebezeit 1991). The main consumers of DOC are heterotrophic bacteria (Fenchel 1988) mediating the flux of DOC through the microbial loop (Azam et al. 1983). However, DOC may also be a potential food source for marine benthic invertebrates, as has been suggested already since the end of the nineteenth century (reviewed by Jørgensen 1976). Reiswig (1981) found a discrepancy between the supply and demand of carbon in benthic suspension feeders.

However, although DOC has been suggested as the missing carbon (Reiswig 1981; Gili and Coma 1998), not much is known on the uptake of DOC by the benthic community. Yahel et al. (2003) were the first to show extensive feeding on bulk DOC by the sponge Theonella swinhoei. Although the potential importance of DOC in the carbon budgets of coral cavities and for the cryptofauna has been discussed (Richter et al. 2001; Yahel et al. 2003; Van Duyl et al. 2006), DOC removal by the crytofauna has not been determined. Our main question in this study is: Are coral reef cavities net sinks of DOC? If so, is DOC quantitatively an important food source for cryptic coral reef habitats? To answer these questions, we carried out a series of measurements and experiments. We compared the concentration of DOC and bacterial abundance between cavity water and overlying reef water in 19 cavities on the fringing reefs of Curaçao and 21 cavities on different types of reefs in the Berau area, East Kalimantan, Indonesia. To determine DOC uptake rates, we closed cavities in both areas and followed the removal of DOC over time and compared that with the removal of bacterioplankton carbon (BC). We estimated fluxes of DOC and BC for the cryptic habitat. The biodegradability of DOC was determined in a series of bioassays with ambient reef water bacterioplankton as key DOC consumers.

Discussion

This is the first study on the flux of dissolved organic carbon (DOC) in coral cavities, and two independent methods indicated that they are sinks of DOC. DOC removal rates in coral cavities were on average two orders of magnitude higher than bacterioplankton removal rates. All our results strongly support a net influx of DOC and bacterioplankton into cavities. Whereas Richter et al. (2001) suggested phytoplankton and, indirectly, bacterioplankton as the most important organic matter sources for biota in cryptic habitats on coral reefs, we, however, find that the removal of DOC in cavities is more likely to be one to two orders of magnitude larger than these particulate sources. The concentration of DOC in cavity water compared to reef water is 15% and 6% lower than the bulk DOC concentration in ambient reef water in Curaçao and in Berau, respectively (Table 3.2). Depletion of bacterioplankton (29% at Curaçao and 38% at Berau) is higher than of DOC. Bacterioplankton is apparently more efficiently removed from cavity water than bulk DOC in coral cavities. These findings may not be surprising because the cryptic biota is dominated by suspension feeders, particularly by sponges (e.g. Vasseur 1977; Richter and Wunsch 1999; Wunsch et al. 2002). Sponges are very efficient filter feeders, especially in feeding on particles smaller than 2 µm, like bacterioplankton (e.g. Reiswig 1974a; Pile et al. 1996; Kötter and Pernthaler 2002). Smaller differences between cavity and reef water DOC concentrations in Berau as compared to those in Curaçao could be explained by shorter residence times of water in the cavities of the Berau area. This argument is supported by the observation that the coral cavities in Berau had more openings to the ambient reef water compared with those in Curaçao, allowing enhanced water exchange. It is also possible that DOC is less efficiently removed from the water in the cavities in Berau as compared to the cavities in Curaçao. The ratio DOC:bacterioplankton carbon in reef water is comparable in both regions, i.e. 33.3 and 34.5 for Curaçao and Berau, respectively, but not in cavity water, i.e. 38.9 for Curaçao and 50.7 for Berau. There is no clear discrimination in DOC and bacterioplankton carbon uptake in cavities of Curaçao. Cavities in Berau, however, remove relatively more bacterial carbon than DOC. The composition of DOC in Berau could be less favourable in terms of utilisation by cryptic organisms, and this could explain the positive discrimination for bacterioplankton in cavities of the Berau area.

DOC fluxes in cavities found in this study are high and unequaled in the literature, with estimated removal rates per cryptic surface area of 342 ± 82 mmol C m-2 d-1 (range: 151-493) in Curaçao and 90 ± 45 mmol C m-2 d-1 (range 33-167) in Berau. After closure of cavities organic carbon is clearly removed in two major fractions, a fast removable fraction Cf of 28 ± 12 µmol L-1 (Curaçao)
and 12 ± 4 µmol L-1 (Berau), and a slow removal fraction Cs of 84 µmol ± 14 µmol L-1 (Curaçao) and 65 ± 4 µmol (Berau) (Figs. 3.2 and 3.4). Therefore, fluxes based on a linear model are grave underestimations and are presented here only as the most conservative values of DOC fluxes into coral cavities. Another approach to determine fluxes of matter into coral cavities is given by Van Duyl et al. (2006). They calculated fluxes in an open system, based on the relation between differences in concentration of matter in cavity water and overlying reef water and the water exchange coefficient of a series of cavities at different current velocities along the reef bottom. They found an average water exchange coefficient of 0.0041 s-1, which is equivalent to an average residence time of water in coral cavities in Curaçao of 4.07 minutes. This approach rules out any possible closure effects. The average difference between cavity and reef water DOC concentration in this study is 14.8 µmol L-1. Using the average water exchange coefficient of 0.0041 s-1, the weighed average volume (148 dm3) and total cavity surface area (TSA; 2.11 m2), the DOC flux into cavities is 367 mmol C m-2 d-1, which is surprisingly close to the average DOC flux calculated with a 2G-model, namely 342 mmol C m-2 d-1. This implies that DOC fluxes based on the 2G-model are reliable. Bacterioplankton abundance declines exponentially in closed cavities.

This supports our assumption that the suspension or filter feeding activity by the cryptofauna is not arrested nor inhibited by closure of the cavities. Bacterial carbon uptake rates by coral reef cavities (Table 3.3) closely resemble those reported in literature. Scheffers et al. (2004) found significant bacterioplankton depletion within cavities on Curaçao of on average 2.5 mmol C m-2 d-1, whereas Van Duyl et al. (2006) found a bacterial carbon flux into cavities in the same area of 3.8 mmol C m-2 d-1. Ayukai (1995) found bacterioplankton carbon retention rates on the Great Barrier Reef of on average 2.2 mmol C m-2 d-1. The average bacterioplankton carbon removal in cavities in this study is 3.6 ± 1.3 (range: 1.8-6.1) and 1.9 ± 1.3 (range: 0.6-3.9) mmol C m-2 d-1 for cavities on Curaçao and Berau, respectively. This represents only 1-2% of the DOC removal. It is evident that DOC is quantitatively a far more important organic 50 carbon source for coral cavities than bacterioplankton in Curaçao as well as in Berau.

In our bioassays in the reef water of Curaçao we recorded an uptake of 20 µmol L-1, or 19%, of DOC by bacterioplankton in 20 days. We consider this fraction to be the average readily available part of the DOC. Because this percentage is close to the average DOC concentration reduction in closed cavities, it is tempting to suggest that the depletion in DOC concentration in CW as compared to RW was due to removal of labile DOC. The most likely candidate to remove labile DOC from cavity water is bacterioplankton. In the bioassays, it takes bacterioplankton 20 days to take up 19% of the total DOC, while cavities (with a 10-fold lower abundance of bacterioplankton as compared to the bacterial abundance in the bioassays) remove the same amount of DOC within 30 min. In addition, the residence time of water in our coral cavities is more in the range of minutes than days (Van Duyl et al. 2006). It is, therefore, unlikely that bacterioplankton is responsible for the DOC depletion in coral cavities. Uptake of DOC by coral cavities appears to have been an overlooked general function in coral reef ecology. Considering the sheer size of the cryptic habitat and the significance compared to other sources of organic matter, DOC may be a key factor in the carbon and energy budget on coral reefs. Net influx of DOC into cavities is shown in a wide variety of cavities in two distinct coral reef regions, i.e. an Atlantic and Indo-Pacific region.

Bulk DOC uptake rates by coral cavities vary in time and between cavities sampled in different areas. Neither the concentration of bulk DOC at the start of an experiment, nor cavity geometry accounts significantly for the variation in DOC fluxes. The difference in DOC flux size between the sampled areas can be explained by differences in composition and quality of DOC. The composition
of the dissolved organic matter (DOM) pool is very diverse with a size range of low-molecular-weight organic molecules like amino acids, to highmolecular-weight molecules (e.g. mucus, polysaccharides), to minute particles like viruses and colloids. At least 10% of oceanic DOM is colloidal material (>95% consists of non-living particles) in the size range 0.4-1.0 µm that easily
passes the pores of the GF/F filters commonly employed in the separation of DOM and particulate organic matter (POM) and a significant part still passes through the pores of the 0.2 µm filters that we used (Koike et al. 1990). It could well be that the cryptofauna mainly takes up colloidal material in this size range. Sponges can take up minute particles like viruses (Hadas et al. 2006) and 0.1 µm-sized beads from ambient water (Leys and Eerkes-Medrano 2006).

Pile et al. (1996, 1997) showed that the sponges used in their studies did not show selective feeding on any component of the plankton community. They suggested that the composition of the plankton community and the variability in the water column can affect sponge nutrition. The opportunistic feeding of sponges is further strengthened by Ribes et al. (1999) who argued that the composition of the ingested carbon by Dysidea avara mainly varied according to the availability of the different prey types in the water column. In this respect, DOM should also be taken into account as a food source available for the cryptic biota, which is dominated by suspension feeders. There is evidence of extensive DOM feeding by suspension feeders. DOC intake can explain up to 50% of the carbon demand of zebra mussels (Roditi et al. 2000; Baines et al. 2005). Yahel et al. (2003) showed evidence of extensive in situ DOC feeding, representing more than 90% of the total carbon intake, by the marine sponge Theonella swinhoei.

The variation in DOC removal rates between cavities may be attributed to differences in cryptofaunal composition. Interestingly, we know from previous data that each cavity investigated on Curaçao has its unique cryptofaunal composition (Scheffers 2005). Yet, there is no significant difference in carbon fluxes between cavities on Curaçao, and the proportion of the main functional groups that might influence variation in DOC removal rates, like sponges, (calcareous) algae, ascidians, bryozoans, and polychaetes, is relatively constant in coral cavities on Curaçao.

Depending on reef zone, the cryptic surface may range from less than 1× to 8× the planar reef area (Richter et al. 2001; Scheffers 2005). Hatcher (1997b) reviewed the importance of regenerative spaces in reefs for the carbon budget of coral reefs. His gross primary production rates of entire reefs, where back reefs were the most productive reef zones (Hatcher 1990), may, however, be insufficient (200-500 mmol C m2 planar reef d-1) to meet the organic carbon demands of cryptic biota. We measured removal rates of 1,000 mmol C m-2 planar reef d-1, assuming an average cryptic surface of 2.8 m2 m-2 planar reef for the entire Curaçaoan reef (Scheffers 2005), omitting the particulate organic carbon removal by cryptic habitats and omitting the DOC consumption by benthos of the open reef. So, removal of DOC by cryptic habitats alone is already two times more than the gross primary production. Where is all this carbon coming from? Because bulk DOC concentrations are usually higher in reef overlying waters than in adjacent ocean (Ducklow 1990; Torréton et al. 1997; Van Duyl and Gast 2001), the carbon budget is unlikely to be matched by net import of bulk DOC from the ocean to the reef, unless organic carbon is actively taken up against a concentration gradient. This implies that DOC production by reefs and reef overlying waters, and possibly DOC supply from land-based sources is probably larger than currently anticipated. A net input of external particulate organic matter (POM) to the reef, for instance by trapping of oceanic plankton and other particles by the reef (Hamner and Wolanski
1988; Richter et al. 2001), may possibly result in extra DOC supply via the benthic food web. This, however, may not be sufficient to cover the gap between gross primary production and consumption. Coral mucus, a part of colloidal DOC, has been suggested to be an important carrier of energy to the benthic food chains of the reef (Wild et al. 2004). Therefore, we hypothesise that the bulk DOC production by noncryptic reef communities is significantly higher than presently assumed. It is evident that the high DOC removal rates we measured in cryptic habitats of coral reefs in Curaçao and Berau influence our present understanding of energy budgets of coral reefs.