Japan has many coastal carbon dioxide seeps as it is one of the most volcanically active parts of the world. These shallow seeps do not have the spectacular aggregations of specialist fauna seen in deep-sea vent systems but they do have gradients in seawater carbonate chemistry that are useful as natural analogues of the effects of ocean acidification on marine biodiversity, ecosystem function and fisheries. Here, we compare macroinvertebrate diversity and abundance on rocky habitats at ambient (mean ≤ 410 µatm) and high (mean 971–1484 µatm) levels of seawater pCO2 in the warm-temperate region of Japan, avoiding areas with toxic sulphur or warm-water conditions. We show that although 70% of intertidal taxa and 40% of shallow subtidal taxa were able to tolerate the high CO2 conditions, there was a marked reduction in the abundance of corals, bivalves and gastropods in acidified conditions. A narrower range of filter feeders, grazers, detritivores, scavengers and carnivores were present at high CO2 resulting in a simplified coastal system that was unable to retain the high standing stocks of marine carbon biomass found in ambient conditions. It is clear that cuts in CO2 emissions would reduce the risks of climate change and ocean acidification impacts on marine biodiversity, shellfish production and biomass in the rocky coastal shores of this region.
Since their discovery in 1977, the diversity and abundance of marine invertebrates around deep-sea (> 200 m) hydrothermal vents has received much more scientific attention than their shallow-water counterparts. Deep-sea vents have obligate animal communities that are reliant upon chemosynthetic processes whereas shallow-water seeps have an abundance of photoautotrophs and few obligate fauna (Tarasov, 2006). Marine CO2 seeps occur where plate tectonics create volcanoes; Japan has many as it sits on the ‘Ring of Fire’ that extends around the landward rim of the Pacific from New Zealand to Chile.
One of the best studied shallow CO2 seeps in the NW Pacific ocean basin is on Yankicha Island (47°31′N, 152°26′E), part of the Kuril archipelago north of Hokkaido. Here, high biomass communities of anemones, holothurians and sea urchins (up to several kg m – 2) benefit from elevated chemosynthetic and photosynthetic primary production in an enclosed coastal lagoon (Zhirmunsky and Tarasov, 1990). Guishan Island, off Taiwan (24°50′N, 121°57′E), is another well-studied shallow NW Pacific CO2 seep, but it is so strongly sulphurous that zooplankton are killed when they are brought into the seep area by currents (Hu et al., 2012). Thus, the marine zoology of hydrothermal systems is highly site-specific.
Research into shallow-water seeps increased once it was realised that these systems can have gradients in seawater pCO2 that provide natural analogues for the future effects of ocean acidification (Hall-Spencer et al., 2008). Ocean acidification is happening due to the uptake of human CO2 emissions by surface seawater from the Earth's atmosphere. Some shallow-water seeps simulate this process by raising pCO2 and bicarbonate levels in seawater but lowering seawater carbonate saturation and pH. Studies of the effects of ocean acidification using CO2 seeps require persistent gradients in seawater carbonate chemistry without toxic levels of metals and sulphur that would otherwise confound the experiments and observations (Pichler et al., 2019; Aiuppa et al., 2021; Blain et al., 2021).
In the NW Pacific, research using CO2 seeps to study ocean acidification began at Iwotorishima Island (27°52′N, 128°14′E) in tropical Japan (Inoue et al., 2013). That study showed that soft corals replaced hard corals and the diversity of benthic invertebrates fell as CO2 levels increased. Enochs et al. (2015) then showed that as CO2 levels increased at a tropical seep site off Maug Island (47°31′N, 152°49′E) the diversity and abundance of calcified algae and scleractinians decreased, being replaced by fleshy macroalgae. This natural analogue approach has been applied to other acidified reef systems, such as at a lagoon in Nikko Bay, Palau (7°20′N, 134°29′E) where the phenotypic plasticity and adaptive potential of tropical corals can be assessed because genetic flow from outside the lagoon is restricted (Golbuu et al., 2016; Kurihara et al., 2021). Studies of genetic adaptation to ocean acidification and warming are difficult at most CO2 seeps because these sites are easily colonised by larvae that drift in from outside the acidified area (Allen et al., 2016).
Agostini et al. (2015) began studies of warm-temperate CO2 seeps at a temperate/tropical biogeographic boundary, paving the way for investigations into range shifts and the balance between temperate kelp forest and tropical coral communities. They selected Shikine Island (34°19′N, 139°12′E), 150 km southwest of Tokyo, as this island has areas with increased levels of pCO2 in the seawater but the same temperature, salinity, dissolved oxygen, total alkalinity, nutrients, wave exposure, current strength, substratum type, and depth as nearby reference sites (Agostini et al., 2018, 2021). Knowledge of the types of organisms that live around Shikine Island CO2 seeps has been growing steadily over the past 5 years but is scattered in the published literature.
The present paper provides a checklist of benthic macroinvertebrate diversity around warm-temperate CO2 seeps off Shikine Island. This list reflects the authors' collective zoological knowledge of the area and draws extensively on surveys by Agostini et al. (2018), Harvey et al. (2019) and Tomatsuri and Kon (2019). Here, we assess how the diversity, abundance, and function of macrobenthic fauna change from ambient to high CO2 areas. Our focus on invertebrates fills a gap, following-up on similar studies of Shikine Island CO2 gradients that focused on microbial biofilms (Kerfahi et al., 2020), algae (Harvey et al., 2021a, b; Wada et al., 2021), and fish (Cattano et al., 2020).
MATERIALS AND METHODS
Shikine is a volcanic island south-east of the Izu Peninsula in Japan. It has many hydrothermal CO2 seeps, several of which have been surveyed from the shore or using RV Tsukuba II since 2014 (Fig. 1). Agostini et al. (2018), Harvey et al. (2019), and Tomatsuri and Kon (2019) describe the methods used for measuring pH, temperature, salinity and total alkalinity and associated environmental factors at our study sites through in situ measurements and/or discrete sampling. The diversity and abundance of macroinvertebrates were surveyed outside the influence of the seeps where seawater pCO2 was ≤ 410 µatm (‘ambient’ sites) and closer to the seeps where pCO2 in seawater was high (‘high CO2’ sites).
Agostini et al. (2018) measured the percent cover of encrusting fauna in 25 × 25 cm quadrats and counted other sessile macroinvertebrates in 50 × 50 cm quadrats by snorkelling, when the tide was in. In both cases, 10–15 replicate quadrats were deployed at least a metre apart on steeply sloping rock faces at nine ambient sites and three high CO2 sites. Tomatsuri and Kon (2019) surveyed the diversity and abundance of hermit crabs and live molluscs at three ambient sites and two high CO2 sites using ten 50 × 50 cm quadrats per site when the tide was out. Their previously unpublished live mollusc data are included here. We collated these data and standardised them to numbers m–2.
Agostini et al. (2018) assessed rocky reef community structure 3–6 m below Chart Datum by SCUBA diving using haphazardly placed 50 × 50 cm digital photoquadrats (n = 8–11) at four ambient sites and three high-CO2 sites. Photographs were analysed using ImageJ (Developed at the National Institutes of Health and the Laboratory for Optical and Computational Instrumentation (University of Wisconsin), released in 1997) by overlaying 64 points on a grid, and recording the organisms at each point. Harvey et al. (2019) sampled two sites for macroinvertebrates, one with mean pCO2 410 ± SD 73 µatm characterised by seaweeds and the other with 971 ± SD 258 µatm and characterised by a diatom mat. Fauna were collected using a SCUBA diver operated airlift to dislodge and lift samples into a 400 µm mesh net for later analysis from four 25 cm diameter circular plots at each site. Samples were fixed in 70% ethanol prior to sorting and identification under a dissecting microscope, and abundance counted.
Seawater carbonate chemistry at intertidal and subtidal sites used in three studies off Shikine Island, Japan in ambient conditions and at high CO2. Mean (± SD) pHT, temperature, salinity, and total alkalinity (TA) are measured values. Seawater pCO2, dissolved inorganic carbon and the saturation states for calcite (Ωcalcite) and aragonite (Ωaragonite) are mean values (± SD) calculated using the carbonate chemistry system analysis program CO2SYS.
We assessed sponge diversity on 25 September 2019, 29 October 2019 and 16 December 2019. On each of these days two SCUBA dives were carried out on rocky seabed at 3–6 m below Chart Datum along 10 m long, 4 m wide transect lines (n = 3), one in a high CO2 zone and one in an ambient CO2 zone. To minimize other factors that influence sponge distribution, such as light exposure, depth and inclination of the rock, the two sites selected had similar topographic features. Each sponge found on each transect was photographed before a small sample was removed using a knife and placed in a ziplock bag and later put in 70% ethanol. Spicule analysis followed methods in Rützler (1978); the samples were hand-cut, put in nitric acid, then washed twice with 70% ethanol before mounting on slides using clear nail varnish. Following Hooper (2000), hand-cut sections were prepared to analyse skeleton morphology, where possible comparing ectosome (tangential) and choanosome (transversal) sides, before being fixed on microscope slides as above. Photographs of the sponge spicules and skeletal structure were taken using a microscope-mounted camera and eyepiece graticule for scale. Identification to genus was made using keys in Hooper and Van Soest (2002) and to species using the World Porifera Database (available from www.marinespecies.org/porifera).
The authors collectively spent hundreds of hours at the seeps carrying out tasks such as helping media teams film the research programme (e.g., https://8billionangels.org/). This allowed incidental observations of bacterial mats/sulphurous conditions at the seeps as well as noting which rocky shore organisms can tolerate high CO2 conditions and which were only seen far from the seeps. Our quantitative data on abundance were standardised to numbers m–2 whereas qualitative ad hoc observations were recorded as presence/absence.
Differences in seawater carbonate chemistry between ambient (< 410 µatm pCO2) and high CO2 conditions at Shikine Island hydrothermal seeps were reported in detail by Agostini et al. (2018), Harvey et al. (2019) and Tomatsuri and Kon (2019) and are synthesised here in Table 1. In brief, whilst there were no significant differences in salinity or total alkalinity between sites, the mean values of seawater pH fell from ≥ 8.0 in ambient conditions to ≤ 7.8 near the seeps. At ambient sites the mean aragonite saturation of seawater was > 2.5 (with no episodes of aragonite undersaturation) but this fell to means as low as 1.4 in our high CO2 sites where there were frequent episodes of aragonite undersaturation.
In ambient conditions thick biogenic carbonate crusts formed by coralline algae, serpulids, barnacles, oysters and mussels characterised bedrock on the lower shore of Shikine Island. Sponges, polychaetes and sipunculids bored into these biogenic carbonate crusts on the low shore but this species-rich habitat was lost in high CO2 areas. Figure 2 shows seven examples of macroinvertebrate taxa that were found living on bedrock in the intertidal where the seawater had high levels of pCO2. The acidified habitats were less complex than in ambient conditions, with a clear reduction in calcareous organisms and an increase in the amount of bare rock and turf algae (see Agostini et al., 2018 for quantitative data on these changes). Of all the intertidal taxa recorded, only the sea anemone Actinia equina was more common in the high pCO2 sites than the ambient sites (Agostini et al., 2018).
Table 2 lists the intertidal benthic macroinvertebrates recorded, arranged by phylum, class, genus and species. The table includes notes on whether these animals have a vagile or sessile lifestyle and on their feeding types. This natural history information was obtained from Nishimura (1992), Hayashi (2006), Imahara (2011) and Okutani (2017). In the high CO2 areas 30% fewer taxa were recorded than in ambient conditions. This reduction affected both vagile and sessile taxa and reduced the diversity of suspension feeders, grazers, carnivores and scavengers (Fig. 3). The abundance of calcifying fauna (e.g., the barnacle Chthamalus challengeri, the limpet Siphonaria japonica and the gastropods Monoplex parthenopeus and Nerita albicilla) decreased in the high CO2 areas where rock was mainly covered in fleshy algal turf or biofilm.
Presence (dot) or abundance (number m–2) of intertidal species at ambient and high mean levels of seawater pCO2 at Shikine island, Japan. 1: Agostini et al., 2018; 2: Harvey et al., 2019; 3: Tomatsuri and Kon, 2019; 4: Harvey et al., 2018; u.d.: unpublished data. Notes indicate sessile vs. vagile along with feeding type (SF = suspension feeders, D = detritivores, G = grazers, C = carnivores and S = scavengers); they were all epifaunal.
Sponge percentage cover decreased and the average number of calcifying organisms fell from several hundred individuals per m2 in ambient conditions to fewer than 40 individuals per m2 at high CO2. The abundance of large barnacles (> 1 cm in diameter, including Capitulum mitella, Megabalanus volcano and Megabalanus rosa) fell from over 60 per m2 in ambient conditions to around 20 individuals per m2 at high CO2 before petering out to zero individuals in the highest CO2 areas (Agostini et al., 2018). Mussels were also reduced in abundance, from around 28 per m2 in ambient conditions to 8 per m2 in the high CO2 areas (Agostini et al., 2018).
There was also a reduction in abundance of intertidal molluscs and hermit crabs from around 175 individuals per m2 in ambient conditions to around 60 per m2 at high CO2 and an associated reduction in the diversity of feeding types (Fig. 4). Hermit crab species richness was lower in acidified areas, although Pagurus filholi was abundant in all areas. Hermit crabs in the high CO2 areas had a lower diversity of carried shells and there were fewer empty gastropod shells available to them in the high CO2 areas (see Tomatsuri and Kon, 2019 for these datasets).
Figure 5 shows six examples of macroinvertebrates found living at high CO2 sites in shallow subtidal waters off Shikine Island. This figure illustrates the acidified rocky habitats in which the organisms were found, which were less complex than the ambient conditions with a notable reduction in the amount of crustose coralline algae and scleractinian corals and an increase in benthic diatom mats and turf algae at high CO2. Table 3 lists the benthic invertebrates recorded in ambient and high CO2 conditions, assigned to lifestyle and feeding groups (using the same reference sources as for Table 1) with quantitative density data converted to numbers per m2.
Presence (dot) or abundance (number m–2) of subtidal species at ambient and high mean levels of seawater pCO2 at Shikine island, Japan. 1: Agostini et al., 2018; 2: Harvey et al., 2019; 3: Tomatsuri and Kon, 2019; 4: Harvey et al., 2018; new: new data. Notes on sessile vs. vagile and epifauna vs. infauna provided along with feeding type (SF = suspension feeders, D = detritivores, G = grazers, C = carnivores and S = scavengers).
There was a 60% decrease in macrofaunal biodiversity from ambient to high CO2 conditions, although the surviving communities retained examples of suspension feeders, detritivores, grazers, carnivores and scavengers. Many benthic taxa were only found in ambient conditions, including species of sponge, polychaete, mollusc, arthropod and echinoderm (Table 3, Fig. 4). There were significantly reduced abundances of calcifying invertebrate macrofauna at high CO2. Scleractinian corals were common in ambient conditions (∼11% cover) but comprised < 1% cover at high CO2 sites. Small gastropods were abundant in ambient conditions but not at high CO2 sites and the shells of large marine snails showed obvious signs of external dissolution by corrosive seawater (e.g., Charonia lampas, Fig. 5).
Air-lift samples confirmed lower species richness at high CO2 sites, and in these samples there was an absence of calcified bivalve molluscs, decapods, mysids and echinoderms. There was about a 50% reduction in the numbers of individuals collected at high CO2 (Fig. 6). Tanaids, which are less calcified, became the most abundant taxon, comprising on average ∼50% of individuals in the elevated pCO2 conditions but only ∼10% in the ambient air-lift samples (Harvey et al., 2019).
The studies synthesised here focused on investigating the biological effects of ocean acidification and so avoided areas that had extremely high levels of CO2 and were characterised by bacterial mats supporting very low levels of invertebrate diversity, as reported at many shallow-water hydrothermal seeps worldwide (Aiuppa et al., 2021; Blain et al., 2021). At our high CO2 sites the seawater was not heated or sulphurous and this allowed observations of the long-term effects of ocean acidification.
We found that most (70%) of the intertidal warm-temperate epifauna recorded could survive in high CO2 conditions. This resilience reflects the fact that intertidal organisms have evolved to cope with extreme fluctuations in physicochemical conditions, including hypercapnia (Inaba and Hall-Spencer, 2020). However, there were large reductions in the abundance and size of calcareous organisms in the intertidal. Body size reductions have been noted as a way in which organisms cope with seawater acidification (Harvey et al., 2018). Large oysters (Saccostrea spp.), mussels (Mytilsepta virgata), limpets (Siphonaria spp.) and barnacles (e.g., Capitulum mitella and Megabalanus spp.) were abundant in ambient conditions but were sparsely distributed and smaller at high CO2. This caused a loss in habitat complexity and a reduction in filter feeding and grazing functions (Agostini et al., 2018), which amount to a loss in ecosystem services such as shellfish productivity and maintenance of good water quality (Lemasson et al., 2017). The diversity and abundance of shells available to hermit crabs and the abundance of these crabs were all significantly reduced at high CO2 (Tomatsuri and Kon, 2019). An increased abundance of anemones in acidified conditions ties in with similar observations in the Mediterranean and may relate to a decrease in competition with calcareous animals for living space (Suggett et al., 2012).
Most taxa listed in Table 3 are epifauna because we carried out our surveys on bedrock. A wider range of habitats have been studied at CO2 seeps in Ischia (Italy), including artificial substrata, and for this reason the range of taxa found able to cope with acidified conditions is large (we list 69 resistant taxa compared with around 250 resistant invertebrate species listed by Foo et al., 2018). Deployment of scouring pads has worked well for sampling a diverse range of fauna at CO2 seeps (Cigliano et al., 2010; Allen et al., 2017) and settlement tiles are a convenient way of collecting bryozoans, serpulids and tunicates (Kroeker et al., 2011; Harvey et al., 2021). Care is needed when interpreting such studies since ocean acidification decreases habitat complexity (Cattano et al., 2020; Sunday et al., 2017) an aspect that can be obscured when using artificial substrata.
Subtidal sponges and scleractinian corals were looked at in detail, reflecting the expertise of Belfiore and Agostini, but for some taxa our list of shallow sublittoral macroinvertebrate fauna will be incomplete. Bringing in experts with taxonomic knowledge of polychaetes, hydroids, bryozoans and ascidians would help fill this knowledge gap. Nevertheless, our faunal inventory provides insights into how ocean acidification can alter marine ecosystems.
The sublittoral invertebrate macrofauna showed a 60% loss of biodiversity at high CO2 levels, with the loss of scleractinian corals and many other calcified organisms. This exceptionally high susceptibility of shallow sublittoral rocky reef communities to ocean acidification appears to be linked to the high exposure of Shikine Island to wave action. During relatively calm spring and summer months, high CO2 levels allow turf algae to rapidly grow and dominate rocky habitats near to the seeps. During each autumn typhoon season, very rough seas hit our study sites. Rough seas decimated the rocky reef communities at the acidified sites, removing most of the algal biomass, but the typhoons had much less effect at the sites with ambient pCO2 levels in seawater. The high typhoon resilience of marine communities growing at ambient pCO2 levels meant that they could store more carbon biomass and retain a higher diversity than marine communities growing at the acidified sites (Harvey et al., 2019, 2021a, b; Wada et al., 2021).
Research at other CO2 seeps
Similar patterns to those seen at Shikine Island occur at seeps on tropical coral reefs in the NW Pacific ocean basin, with habitat degradation and loss of biodiversity in acidified areas (Fabricius et al., 2014; Enochs et al., 2015). These systems have proven very useful for showing which organisms can survive high CO2 conditions and how ecosystem services are affected (Foo et al., 2018; Hall-Spencer and Harvey, 2019).
Work on the ecosystem effects of ocean acidification began at CO2 seeps in Ischia, Italy, where shore access is easy and the gas bubbles out of the seafloor at ambient temperature with no influence of toxic metals or sulphur (Foo et al., 2018). The high CO2 areas have 30% fewer invertebrate taxa and lower animal biomass, although the number of individuals does not differ due to a higher number of small-bodied, tolerant species at the high CO2 sites (Kroeker et al., 2011). Generalists and small-bodied animals dominate at several Mediterranean CO2 seep sites, resulting in food web simplification (Vizzini et al., 2017). Similarly, in Papua New Guinea, reductions in invertebrate diversity were observed with increasing CO2 (Fabricius et al., 2011). A 3-fold decrease in zooplankton biomass was noted in high CO2 areas near Papua New Guinea seeps compared to ambient nearby sites (Smith et al., 2016). The low biodiversity observed at high CO2 is linked to the loss of habitat-forming corals (Fabricius et al., 2014; Sunday et al., 2017).
In a similar study to ours, using quadrats to quantify benthic diversity and abundance, Teixido et al. (2018) recorded 24 macroinvertebrate species in ambient conditions near shallow water seeps in Ischia, Italy but only 13 species at high CO2. Teixdo et al. (2018) explain that ocean acidification caused a loss in functional diversity as, for example, scleractinian corals, molluscs, and sea urchins were present in ambient conditions but absent from their high CO2 quadrats. We recorded similar major reductions in biodiversity and function in warm-temperate Japan with, for example, only four sponge taxa able to live at high CO2 compared with 14 in ambient conditions, one anthozoan species at high CO2 compared with 12 in ambient conditions, and five mollusc species at high CO2 compared with 19 in ambient conditions. All feeding types were still represented at high CO2 but a narrower range of filter feeders, grazers, detritivores, scavengers, and carnivores were present, resulting in a simplified coastal system that was unable to retain the high standing stocks of marine carbon biomass found in ambient conditions. Collectively, these results support the hypothesis that as seawater pCO2 levels increase there is a loss in macroinvertebrate diversity and a proliferation of tolerant species that alter ecosystem function along the warm temperate coast of Japan, to the detriment of shellfish fisheries.
Comparisons of intertidal and subtidal rocky reef communities living under ambient and high CO2 conditions show that ocean acidification is a threat to many marine organisms, driving fundamental shifts in coastal marine ecosystems towards simplified communities with reduced capacity to store carbon. Although intertidal macroinvertebrates retained 70% of their surrounding diversity at high CO2, they decreased in abundance and so their provision of ecosystem services, such as shellfish production and water filtration, was reduced. The shallow sublittoral communities retained only 40% of their diversity, with a proliferation of small opportunist taxa such as tanaeids at high CO2. Our observations suggest that cuts in CO2 emissions will reduce the risks of ocean acidification, helping to ensure that the coastlines of Japan remain highly biodiverse, resilient, and able to provide ecosystem services such as fisheries, good water quality, and carbon storage.
This work would not have been possible without the crew of RV Tsukuba II and the technical staff of Shimoda Marine Research Center, University of Tsukuba. Our research was funded by the ‘International Educational and Research Laboratory Program’, University of Tsukuba, by the Japan Society for the Promotion of Science KAKENHI Grant in Aids (Grant No: 17K17621, 17K17622 and 18H02219), the Ministry of Environment, Government of Japan (Suishinhi: 4RF-1701) and Core-to-Core Program (Grant No: JPJSCCA20210006). The Nii-jima Fishery Cooperatives Shikine-jima Branch kindly permitted access to field sites. Callum Hudson helped improve a draft.
The authors have no relevant financial interests that may influence the interpretation of our results.
SA initiated studies at the Shikine island. GB collected the sponge data. JMH-S prepared the first draft, and all authors interpreted the data and co-wrote the manuscript.