Effects of ocean acidification on sponge communities

Authors


Abstract

The effects of ocean acidification on lower invertebrates such as sponges may be pronounced because of their low capacity for acid–base regulation. However, so far, most studies have focused on calcifiers. We present the first study of the effects of ocean acidification on the Porifera. Sponge species composition and cover along pH gradients at CO2 vents off Ischia (Tyrrhenian Sea, Italy) was measured at sites with normal pH (8.1–8.2), lowered pH (mean 7.8–7.9, min 7.4–7.5) and extremely low pH (6.6). There was a strong correlation between pH and both sponge cover and species composition. Crambe crambe was the only species present in any abundance in the areas with mean pH 6.6, seven species were present at mean pH 7.8–7.9 and four species (Phorbas tenacior, Petrosia ficiformis, Chondrilla nucula and Hemimycale columella) were restricted to sites with normal pH. Sponge percentage cover decreased significantly from normal to acidified sites. No significant effect of increasing CO2 levels and decreasing pH was found on spicule form in Crambe crambe. This study indicates that increasing CO2 concentrations will likely affect sponge community composition as some demosponge species appear to be more vulnerable than others. Further research into the mechanisms by which acidification affects sponges would be useful in predicting likely effects on sessile marine communities.

Introduction

Atmospheric CO2 concentration is currently approximately 383 parts per million by volume (ppmv), the highest level for at least 650,000 years and, due to anthropogenic influences, is projected to increase by 0.5% per year throughout the 21st century (Guinotte & Fabry 2008). The oceans are a major sink for CO2 and have absorbed approximately half of all anthropogenic emissions (Sabine et al. 2004). Since CO2 dissolves in seawater, it shifts the carbonate system towards higher concentrations of CO2, bicarbonate (math formula) and protons H+, therefore decreasing the seawater pH and the carbonate ions used by calcifiers to build their shells and skeletons. Since the early 1900s, a 30% increase in H+ concentration of surface waters has resulted in a pH decrease of 0.1 units, and a further drop in surface seawater pH of up to 0.5 units has been projected by 2100 (Caldeira & Wickett 2005).

The adverse effect of seawater acidified with CO2 has been most apparent on marine calcifiers, such as coralline algae (Kuffner et al. 2008; Martin et al. 2008), foraminiferans (Moy et al. 2009; Dias et al. 2010), corals (e.g. Silverman et al. 2009), echinoderms (e.g. Michaelidis et al. 2005) and molluscs (Gazeau et al. 2007). Among these organisms, rates of calcification have been predicted to fall by up to 60% within this century, depending on the physiology of the species and their mineralogy (Kleypas et al. 2006). However, although it is in calcifiers that the effects are most obvious, studies using high CO2 levels indicate that unifying principles define sensitivity to CO2 in both calcifying and non-calcifying animals (Pörtner 2008). The effects of ocean acidification on lower invertebrates such as sponges are likely to be pronounced because of their low capacity for acid–base regulation (Pörtner 2008). As yet no studies have examined the effect of ocean acidification on the Porifera.

Recently, CO2 vents have been used as natural laboratories to advance our understanding of ocean acidification at the ecosystem level (Hall-Spencer et al. 2008; Cigliano et al. 2010; Porzio et al. 2011). While most vents emit hot gases, including toxic sulphur compounds (Dando et al. 1999), vents off Ischia island near the volcano of Vesuvius in Italy eject >90% CO2 at ambient temperature without toxic sulphur (Hall-Spencer et al. 2008), acidifying coastal seawater on the north and south sides of Castello Aragonese for centuries, maybe millennia (Tedesco 1996). Hall-Spencer et al. (2008) studied species community composition of the area along gradients of normal pH (8.1–8.2), lowered pH (mean 7.8–7.9, min 7.4–7.5) to extremely low pH 6.6. They recorded 64 species of plant, alga and animal, many of which were resilient to the acidified conditions. Within sponges, eight genera were found with an apparent decrease in their distribution from normal to extremely low pH values (Hall-Spencer et al. 2008, supporting information). However, no sampling or full quantitative analysis of sponge species composition was undertaken. Although other studies have examined the biodiversity of sponges associated with hydrothermal vent sites (Morri et al. 1999; Pansini et al. 2000), none has looked at the effect of small-scale variations in pH on sponge community composition.

The sponge fauna of Ischia is relatively well studied, with some 86 species reported from the area (Sara 1959). Both demosponges, which possess skeletons of siliceous spicules, and calcarea, which have calcium carbonate spicules, are present. Here we present the results of a complete survey of the sponge fauna at CO2 vents, taking account of both qualitative (species composition) and quantitative (percentage cover of the substratum) aspects of the sponge community composition along pH gradients.

Methods

Study site and seawater carbonate measurements

Sponge community composition was measured during May 2008 on the north and south sides of the Castello Aragonese (40°043.849′ N, 13°57.089′ E) off Ischia in Italy (Fig. 1). Six sites were identified along gradients of increasing pCO2 to the north and south of the Castello: two normal pH stations (N1 and S1; Fig. 1); two stations that had reductions in mean pH of 0.2–0.4 units (N2 and S2; Fig. 1) and two stations with reductions in mean pH of 0.6–1.5 units (N3 and S3; Fig. 1) which are more representative of the localised effects to be expected from deliberate CO2 sequestration rather than from global ocean acidification. Sites were chosen on the basis of the carbonate chemistry of the vents, mapped out by Hall-Spencer et al. (2008), Rodolfo-Metalpa et al. (2010) and Cigliano et al. (2010), who repeatedly measured pHT (in total scale), total alkalinity (TA), temperature and salinity. Water samples were collected at 1 m depth using glass bottles, and the pHT was measured immediately using a metre accurate to 0.01 pH units (Metrohm 826 pH mobile, Metrohm, Herisau, Switzerland) calibrated using TRIS/HCl and 2-aminopyridine/HCl buffer solutions (DOE 1994). Seawater samples were then passed through 0.45-μm pore size filters (GF/F Whatman, Maidstone, Kent, UK) and poisoned with 0.05 ml of 50% HgCl2 (AnalaR: Merck, Darmstadt, Germany) to avoid biological alteration, and stored in the dark at 4 °C. Three replicate 20-ml sub-samples were analysed at 25 °C using a titration system composed of a pH meter with an ORION pH electrode (see Hall-Spencer et al. 2008 for further details). Parameters of the carbonate system [pCO2, math formula, math formula, and saturation state of calcite (Ωcalcite)] were calculated from pHT, TA, temperature and salinity (38‰) using the free-access CO2 Systat package(Systat Software Inc., Chicago, Illinois, USA) (Table 1).

Table 1. Summary of pHT (in total scale), and carbonate chemistry data from 2007 to 2008 at CO2 vents off Ischia (from Hall-Spencer et al. 2008; Cigliano et al. 2010; Rodolfo-Metalpa et al. 2010). Water samples were taken from rocky shore zones N1–N3 and S1–S3 NB: min pH had the max pCO2 levels and vice versa
SitepHTpCO2 (μatm)math formula (μmol·kg−1)math formula (μmol·kg−1)ΩcalcΩarag
N1
Min7.9628619018104.482.92
Mean (n = 18)8.1335326319495.933.86
Max8.1955237028306.584.28
N2
Min7.493097018501.551.00
Mean (n = 24)7.8585415921273.712.41
Max8.17190827024506.204.03
N3
Min6.31654520900.110.07
Mean (n = 23)7.1375534823921.120.73
Max7.9029,88517024803.892.53
S1
Min8.0930424018305.613.62
Mean (n = 12)8.1334725819206.023.91
Max8.1739528020006.504.24
S2
Min7.273145018401.110.73
Mean (n = 23)7.79111914821493.492.27
Max8.16327827024706.354.14
S3
Min6.07886322200.070.04
Mean (n = 35)6.6818,3132124620.460.30
Max7.7951,99714024903.272.13
Figure 1.

Map of CO2 vent sites north and south of Castello Aragonese, off Ischia Island, Italy. Mean surface pH is shown at 35-m-wide stations.

Sponge community composition

Quantitative and qualitative measurements of sponge community composition were assessed at each of the six study sites (N1–N3, S1–S3). At each site, 10 quadrats of 30 cm2 were sampled, deployed haphazardly on very steep or vertical bedrock faces in a depth range of 1–3 m. Each quadrat was photographed with a Nikon D70 camera (Nikon Inc., Tokyo, Japan.) in an Ikelite housing using an 18–70 mm zoom lens for quantitative sponges cover. The number of species in each quadrat was counted and a small sample of each species was scraped off from the rock using a knife and placed in a ziplock bag. Samples were preserved in 70% denatured ethanol and examined microscopically to confirm identification (see Picton & Goodwin 2007 for detailed methodology).

Change in species composition along a pH gradient (7.09–8.14 pH) was examined on a 240-m transect. Three 30-cm2 quadrats were haphazardly deployed at 25 stations spread at 10-m intervals. Quadrats were deployed on very steep or vertical bedrock faces between 1 and 3 m in depth and the composition of sponge communities was measured as described above. Only sponges on the north side of the Castello Aragonese were quantified in this way. Percentage sponge cover per square meter was calculated using IMAGEJ (Rasband 1997–2008). Percentage data were arcsine-square root-transformed. Differences in percentage cover were tested with the Kruskal–Wallis test.

Ten choanosomal style spicules from five samples of Crambe crambe were measured for each of the southern sites (S1–S3). The aim of this was to examine any effects of increasing CO2 and decreasing pH on spicule form. Only C. crambe was examined as it was the only species present at all three sites. The species has siliceous spicules.

Results

There was a significant difference in total percentage sponge cover on both south sites (Kruskall–Wallis H = 16.250, P < 0.0003) and north sites (Kruskall–Wallis H = 16.979, P < 0.0002). The difference was most marked between the highly acidified sites (N3, S3) and the control sites (N1, S1) (Figs 2 and 3).

Figure 2.

Sponge percentage cover per square meter (mean + SE, n = 10) at sampling stations along the rocky cliff of Castello Aragonese. North stations: N1 (mean pH 8.14), N2 (mean pH 7.87) N3 (mean pH 7.09). South stations: S1 (mean pH 8.14), S2 (mean pH 7.83), S3 (mean pH 6.57).

Figure 3.

Representative 25-cm2 sponge cover quadrats from north side transect. (A) Quadrat 1, N3 (Microciona cf. toxitenuis). (B) Quadrat 9 N2 (Crambe crambe, Ircinia variabilis, Chondrosia reiniformis, Phorbas tenacior). (C) Quadrat 9, N1 (Crambe crambe, Phorbas tenacior, Chondrosia reiniformis).

In total 14 sponge species were recorded from the study area (Table 2). Species composition varied between sites. The only species present in any abundance in the extremely acidified sites (S3, N3) was Crambe crambe, although small patches of Microciona cf. toxitenuis and Spirastrella cuncatrix were also present. A greater variety of sponges was present at sites of intermediate pH (mean 7.8–7.9). Four species (Phorbas tenacior, Petrosia ficiformis, Chondrilla nucula and Hemimycale columella) were restricted to the normal pH sites (N1, S1).

Table 2. Sponge species recorded on the north (N1–N3) and south side of Castello Aragonese along gradients of pH. Mean percentage cover/m2 (n = 10)Thumbnail image of

Species composition on the north side of Castello Aragonese varied along the transect with very little sponge cover in the most acidified areas (Fig. 4). Crambe crambe was the most abundant species and was the only species present in the most acidified areas (Table 2).

Figure 4.

Mean cover of sponge species (n = 3) from quadrats taken at 10-m intervals along a transect along the north wall of Castello Aragonese. The transect included sampling stations N1 (mean pH 8.14), N2 (mean pH 7.87) and N3 (mean pH 7.09).

There was no significant difference in mean spicule morphology or length (Kruskall–Wallis H = 3.86, P < 0.145) between sites for Crambe crambe.

Discussion

Whereas vertebrates and higher invertebrates have some capacity for acid–base regulation, sponges lack intercellular junctions and other ultrastructural connections which allow the epithelial layers of other animals to act as barriers (Pörtner 2008). Their interstitial fluid is very similar to that of their environment (Ruppert & Barnes 1994). Consequently they are unable to regulate for changes in acidity and may be particularly vulnerable to alterations in environmental conditions.

Sponge community species composition and abundance has been shown to be affected by substratum inclination (Preciado & Maldonado 2005), exposure to light (Jokiel 1980), degree of siltation (Brien et al. 1973; Bell & Barnes 2000b), exposure to predators (Dunlap & Pawlik 1996) and algal abundance (Bell & Barnes 2000a, 2002). Differences in these environmental factors were controlled for as far as possible by sampling from similar habitats at all sites (vertical rock faces between 1 and 3 m in depth). However, orientation of rock face and therefore exposure to light could not be controlled for. Species abundance and sponge cover were slightly lower on the south sampling sites, possibly due to the greater light exposure. Despite this, a decrease in sponge abundance with increased pH was observed on both south and north sites, suggesting that this is the dominant factor affecting sponge community composition in the study area. There has been little research into the effects of pH on sponge metabolism but there is some evidence that it may affect reproductive success and regeneration. Changes in environmental pH have been shown to affect the calcium-induced cell aggregation of sponge cells, possibly by altering the level of intracellular Ca2+ or by affecting cell permeability (Phillip 1997). Re-agreggation is thought to be one process involved in sponge regeneration (Bergquist 1978). It is possible that the difference in species composition between sites reflects differential effects on re-agreggation mechanisms and consequently potential to recover from disturbance.

Exposure to low pH (5.8–6.5) has been found to affect hatching rate and hatchability of gemmules in the freshwater sponge Ephydatia mülleri (Benfey & Reiswig 1982). Larval phases of marine organisms may be particularly vulnerable to the effects of even small alterations in pH, with increases resulting in larval morality and developmental abnormalities (Dupont et al. 2008) and a significant effect of pH on settlement has been shown for other groups at the study site (Cigliano et al. 2010). A variety of developmental forms was represented by the species present. Crambe crambe, the dominant species at the acidified sites, reproduces sexually with internal fertilisation, lecithotrophic larvae of the parenchymella type are released, spending 24–72 h in the plankton before settlement (Uriz et al. 1998). A range of reproductive modes was present in those species only found at the normal pH sites, including both direct development with no larval phase in Petrosia ficiformis (Gil 2007) and development via a non-tufted parenchymella larvae Phorbas ficiticius (Mariani et al. 2006). Consequently, it does not seem that mode of development can account for the difference in distribution between species; other factors may be affecting larval settlement efficiency.

Crambe crambe is one of the most widespread littoral sponges in the Mediterranean (Uriz et al. 1992). It is an opportunistic species with high reproductive output, and larval swimming behaviour which results in wide dispersal (Uriz et al. 1998). A trade-off between defence and growth rate has been found to occur, with allocation to chemical and physical defences being negatively correlated with reproduction and growth (Becerro et al. 1997) and spatial competition being a key factor. The absence of competitors in the acidified sites may favour C. crambe by enabling it to devote more resources to growth; it was the only species to decrease in abundance at the normal pH sites. There was no noticeable effect of acidity on spicule formation in C. crambe. Spicule formation has been found to be affected by other environmental factors such as water temperature and silica concentration (Elvin 1971; Bavastrello et al. 1993). During the late Permian mass extinction, possibly linked to elevated CO2 levels, in South China there was a 88–92% decrease in spicule diversity: the surviving spicule forms were smooth and simple (oxeas, strongyles, oxy-orthpentactines and oxy-orthohexactines; Liu et al. 2008), mirroring the trend for simplified skeletons in other groups (Knoll et al. 1996). It is possible that CO2 affects silica formation, preventing the formation of complex spicules and consequently preventing species with more complex skeletons from colonising the vent sites. Experimental studies would be useful in determining whether this is the case.

Other studies on the vent sites have noted a shift from typical rocky shore communities with abundant calcareous organisms at the control sites to communities lacking scleractinian corals with significant reductions in sea urchin and coralline algal abundance (Hall-Spencer et al. 2008). Although total algal cover remained relatively consistent throughout the sites, there was a shift from calcareous algal dominance to dominance by non-calcareous algae at the low pH sites (for example Caulerpa, Cladophora, Asparagopsis, Dictyota and Sargassum; Hall-Spencer et al. 2008; Porzio et al. 2011). However, many of the sites at which no sponge cover was recorded lacked macroorganisms, consequently greater competition from non-calcareous than calcareous algae is unlikely to be a controlling factor.

Studies of other hydrothermal vent communities in the Mediterranean have indicated that although sponge species biodiversity can be higher at vent than non-vent sites (Morri et al. 1999; Pansini et al. 2000), proximity to vents had no influence on sponge percentage cover (Pansini et al. 2000). In these studies no direct measurements of pH or attempts to correlate sponge biodiversity with pH were made. The hydrothermal vent fluids were recorded as having pH varying from pH 5.2 to 7.1 (Dando et al. 1999) but when diluted by the surrounding seawater the pH is likely to be higher. The study sites were large, encompassing a depth range of several metres and consequently even those in closest proximity to the vents will have encompassed a range of pH conditions.

Elevated CO2 levels have been proposed as a causative factor for the late Permian mass extinction (Knoll et al. 1996; Pörtner et al. 2004). The extinction was particularly pronounced among reef-building metazoans such as calcareous sponges and rugose corals (Pruss & Bottjer 2005) and other organisms with carbonate skeletons (Knoll 2003; Dias et al. 2010). Disassociation of skeletal elements makes the study of siliceous species more difficult, but in South China extinction rates of 88–92% of spicule forms have been shown at the Permian–Triassic boundary, indicating extinction of the majority of siliceous sponge species, including all hexactinellids and species with tetraxons (Liu et al. 2008). Although the presence of a calcified skeleton is recognised as a factor in sensitivity to elevated CO2 (Knoll et al. 1996; Pörtner et al. 2004), other aspects, also found in siliceous sponges, such as low-metabolic rate, absence of a circulatory system and elaborated gas-permeable surfaces, may also result in increased vulnerability (Knoll et al. 1996). Sponges are important functional components of marine ecosystems; playing roles in substrate modification and bentho-pelagic coupling, as well as having a large number of associations with species from other phyla (Bell 2008). Consequently, a future CO2 rise may have significant ecological effects on sponges. Our study indicates that some demosponge species are likely to be more vulnerable than others and further research into the underlying mechanisms would be useful in predicting likely outcomes.

Acknowledgements

We thank members of the Stazione Zoologica Anton Dohrn Laboratory for their kind support, particularly Marie Cristina Gambi for help with specialist literature searches. J.H.-S. was funded by EU FP7 MedSeA grant 265103 and the Save Our Seas Foundation. Claire Goodwin was funded by Esmée Fairbairn Foundation grant 07-2472.

Ancillary