Effects of sub-seabed CO2 leakage: Short- and medium-term responses of benthic macrofaunal assemblages

Marine Pollution Bulletin 128 (2018) 519–526

Contents lists available at ScienceDirect

Marine Pollution Bulletin journal homepage: www.elsevier.com/locate/marpolbul

Effects of sub-seabed CO2 leakage: Short- and medium-term responses of benthic macrofaunal assemblages

T

⁎

T. Amaroa,b,c, , I. Bertoccic, A.M. Queirosd, E. Rastellic,e, G. Borgersenb, M. Brkljacicb, J. Nunesd, K. Sorensenb, R. Danovaroc,e, S. Widdicombed a

Hellenic Center for Marine Research (HCMR), 710 03 Heraklion, Crete, Greece Norwegian Institute for Water Research, Oslo, Norway c Stazione Zoologica Anton Dohrn, Villa Comunale, Naples, Italy d Plymouth Marine Laboratory, Prospect Place, West Hoe, PL1 3DH, Plymouth, UK e Department of Life and Environmental Sciences, Polytechnic University of Marche, Ancona, Italy b

A R T I C L E I N F O

A B S T R A C T

Keywords: CCS CO2 Macrobenthos Impacts

The continued rise in atmospheric carbon dioxide (CO2) levels is driving climate change and temperature shifts at a global scale. CO2 Capture and Storage (CCS) technologies have been suggested as a feasible option for reducing CO2 emissions and mitigating their effects. However, before CCS can be employed at an industrial scale, any environmental risks associated with this activity should be identified and quantified. Significant leakage of CO2 from CCS reservoirs and pipelines is considered to be unlikely, however direct and/or indirect effects of CO2 leakage on marine life and ecosystem functioning must be assessed, with particular consideration given to spatial (e.g. distance from the source) and temporal (e.g. duration) scales at which leakage impacts could occur. In the current mesocosm experiment we tested the potential effects of CO2 leakage on macrobenthic assemblages by exposing infaunal sediment communities to different levels of CO2 concentration (400, 1000, 2000, 10,000 and 20,000 ppm CO2), simulating a gradient of distance from a hypothetic leakage, over short-term (a few weeks) and medium-term (several months). A significant impact on community structure, abundance and species richness of macrofauna was observed in the short-term exposure. Individual taxa showed idiosyncratic responses to acidification. We conclude that the main impact of CO2 leakage on macrofaunal assemblages occurs almost exclusively at the higher CO2 concentration and over short time periods, tending to fade and disappear at increasing distance and exposure time. Although under the cautious perspective required by the possible context-dependency of the present findings, this study contributes to the cost-benefit analysis (environmental risk versus the achievement of the intended objectives) of CCS strategies.

1. Introduction The accelerating rise in atmospheric carbon dioxide (CO2) levels (IPCC, 2013) is causing ocean warming and acidification at unprecedented rates, posing critical threats to single species, habitats, oceanic regions and overall global ecosystem functioning (Caldeira and Wickett, 2003; Feely et al., 2004; Hale et al., 2011; Mora et al., 2013, Cerrano et al., 2013; Meadows et al., 2015; Gattuso et al., 2015). As a direct consequence, it is urgently needed to identify suitable options for reducing/mitigating CO2 emissions (McCormack et al., 2016). One particularly promising technology involves capturing CO2 from point source effluents (mostly, energy generation plants), then transporting it as a supercritical liquid to be stored in deep porous geological rock formations, such as saline aquifers or existing hydrocarbon reservoirs

⁎

(Gibbins et al. 2006; Holloway 2007). This process is defined as CO2 Capture and Storage (CCS). In Europe and North America the technical feasibility of CCS approaches has been already demonstrated. For example, at the Sleipner West gas field in the Norwegian sector of the North Sea, CCS has been operational since 2000 with approximately 1 million tons of CO2 pumped into the storage reservoir every year (Paulley et al., 2012, Jones et al., 2015). However, as with almost any other human activity, this technology is not risk-free in terms of posing potential environmental hazards (reviewed by Damen et al. 2006). Before industrial scale CCS activities become widely accepted and implemented these risks need to be identified and quantified. Perhaps the greatest environmental risk associated with CCS is that of CO2 leaking into the marine environment either during transport, sequestration or from the geological storage reservoir itself. Whilst current evidence

Corresponding author at: Hellenic Center for Marine Research (HCMR), 710 03 Heraklion, Crete, Greece. E-mail address: [email protected] (T. Amaro).

https://doi.org/10.1016/j.marpolbul.2018.01.068 Received 18 January 2017; Received in revised form 30 January 2018; Accepted 31 January 2018 0025-326X/ © 2018 Published by Elsevier Ltd.

Marine Pollution Bulletin 128 (2018) 519–526

T. Amaro et al.

2. Material and methods

suggests that leakage from CCS reservoirs would be extremely unlikely it is not impossible (Blackford et al., 2009, 2014). Given that any major increase in seawater CO2 concentrations, and the associated changes in carbonate chemistry, has the potential to considerably impact marine life and ecosystem functions, assessing the biological and ecological effects of CO2 leakage is essential to support environmental risks assessments required by CCS operations (Widdicombe et al., 2013; Jones et al., 2015). This is especially relevant for benthic assemblages living in the immediate proximity of any potential leak, since they would be exposed to relatively large and rapid changes in carbonate chemistry, in both the sediment pore waters and the overlying seawater (Lichtschlag et al., 2014; Queiros et al., 2014). The exposure to a range of CO2 concentrations has been tested on a variety of marine organisms, as well as on some biogeochemical processes and ecosystem functions (Widdicombe et al., 2013, 2015; Laverock et al., 2013; Tait et al., 2014; Rastelli et al., 2015). It has also been demonstrated that the impact of elevated CO2 on marine organisms depends on both the severity and the duration of the exposure (Blackford et al., 2013). In general, it is hypothesized that a CCS leakage is immediately associated with a localized acute exposure to harmful high CO2 conditions whose effects are likely to get attenuated at increasing distance from the source. Moreover, more prolonged leakage or persisting influences of temporary seepage of any level could represent chronic stressed conditions to the surrounding abiotic and biological environment (Jones et al., 2015). Whilst previous studies have started to provide a better understanding of the potential impacts of CCS leakage on specific benthic organisms (e.g. Widdicombe & Needham 2007), our knowledge of the possible effects at the community level remains limited (Widdicombe et al. 2015). In addition, the mechanisms underlying such changes are still largely unknown, as well as the difference between direct and indirect effects of increasing CO2 leakages on the macrofaunal community. It has been reported, however, that low-pH levels predicted by realistic scenarios of CCS leakage might severely reduce the prokaryotic-mediated processes (Rastelli et al. 2015), while acidified conditions could favor blooms of benthic microbial primary producers including cyanobacteria and diatoms (Tait et al., 2015). Notably, the exposure to high CO2 levels can alter microbial-mediated processes able to affect the quality and quantity of the sedimentary organic matter (OM) (Rastelli et al., 2015). Since the availability of OM is a key driver of the abundance, distribution and biodiversity of benthic fauna (Fabiano and Pusceddu, 1998, Pusceddu et al., 2009), the effects of changes in this variable due to CCS leakage might indirectly propagate to associated macrofaunal assemblages. Full community level effects of CO2 leakages can only be unequivocally assessed using simulated leakage experiments conducted in the field (e.g. Blackford et al. 2014) or from studying actual leakage events or accidents in areas where data on the response variables of interest are available before and after the event. Both options are normally unavailable either due to the lack of data or to logistic, financial and/or ethical constraints. Performing manipulative experiments in mesocosms can be a feasible alternative especially when the results are used to inform ecosystem level models. A strength of an experimental approach is that by exposing initially comparable assemblages to different levels of CO2 concentration (such as ‘naturally’ occurring along a gradient of distance from a supposed leakage) under controlled conditions allows testing for their relative effects in an unconfounded way. In this study, we performed a mesocosm experiment to test the potential impact of CO2-enriched (from 400 ppm to 20,000 ppm) seawater plumes on the abundance and diversity of soft-bottom macrofauna. Specifically, we tested the null hypotheses that (i) the whole structure (taxon composition and relative abundance), richness, total abundance of the macrofaunal assemblages and the abundance of individual taxa, did not differ depending on the CO2 concentration; (ii) such a lack of differences was consistent between a few weeks (shortterm) and some months (medium-term) of continued exposure.

2.1. Collection of sediment samples and associated fauna and mesocosm setup Using a KC Denmark box corer, intact sediment samples containing natural infaunal assemblages were collected during the 3rd week of August 2012 from randomly selected points located some meters apart from each other at the outer Oslofjord (59°49.4788′ N, 10°58.8595′ E), Norway, at 100 m water depth. Each box corer was equipped with an inner liner, which allowed the sediments and the overlying water to be retrieved with minimal disturbance. A total of 46 independent liners (0.09 m2 each, with average sediment penetration of ~40 cm) were collected and transferred immediately to the benthic mesocosm systems at the Marine Research Station, Norwegian Institute of Water Research, Solbergstrand, Norway. During transportation, all liners were shaded and continuously covered with seawater to prevent desiccation and minimise temperature changes. The experimental system was set up according to Widdicombe et al. (2009), as described in detail elsewhere (Queiros et al., 2015, Rastelli et al., 2015). Briefly, all liners were placed in an aquarium in a flowthrough holding basin filled with seawater to a depth of 1 m (mesocosm) and supplied continuously with unfiltered natural seawater at a flow rate of 120 ml/min from a pipeline situated at 60 m depth in the adjacent fjord. All liners were maintained in these conditions for two weeks prior to the beginning of the experiment to allow the fauna, microbes and geochemical processes to acclimatize to mesocosm conditions. 2.2. Preliminary survey To guarantee that the randomly assigned experimental levels of CO2 were not confounded by initial differences between replicate cores in terms of hosted macrofaunal assemblages would have required us to compare macrofauna among all (allocated) treatments before manipulation. Unfortunately, the needed destructive sampling made such an option impossible. Alternatively, a total of 6 liners were chosen at random from the 46 liners initially collected and these 6 were randomly allocated to one of two groups of three. These were then compared (by means of one-way PERMANOVA, see Supplement S1) for the structure of macrofauna, under the hypothesis that the lack of significant difference between one group and the other could provide information (not exhaustive, but relevant) to assume that significant differences were not likely to exist also among the sets of replicate liners allocated at random to the experimental levels. In addition, data on the sediment grain size, estimated by laser analysis at the beginning of the experiment, were available for one liner per each of the total five experimental conditions (e.g. McCave, 2013). For each liner, macrofaunal assemblages were sampled, after the acclimation period, by sieving all the sediment over a 500 μm mesh, with the residue from each sample being fixed in 10% buffered formalin until further processing. In the laboratory, the fauna was extracted from the residue under a binocular microscope and all specimens were sorted into major taxa and then identified to species level whenever possible. Species (or higher taxa) abundance was determined in each replicate and expressed as the total number of individuals per m2 of sampled area. 2.3. Experimental setup and sampling The 40 liners remaining after the preliminary survey were randomly allocated in equal numbers (Berge, 1990) to each of five CO2 treatments: 400 (control), 1000, 2000, 5000, and 20,000 ppm, with two sampling times (2 weeks, 20 weeks). These levels were consistent with those specifically tested by Rastelli et al. (2015) and Queiros et al. 520

Marine Pollution Bulletin 128 (2018) 519–526

T. Amaro et al.

3. Results

(2015). Seawater acidification was achieved as described by Widdicombe et al. (2009). Briefly, CO2 gas passed through a 450 l reservoir tanks filled with natural seawater. Using an automated feedback relay system (Walchem), the CO2 flux into the reservoir tanks was regulated in order to maintain the required pH level. The reservoir tanks were continuously supplied with natural seawater (pH ∼ 8.1). To assess short-term (a few weeks scale) and medium-term (several months) effects of CO2 exposure, sampling took place after 2 weeks exposure (T1) and again, on a different set of replicate liners not previously sampled, after 20 weeks exposure (T2). At each time, macrofaunal assemblages were sampled as described for the preliminary survey. A procedural control involving the flux of air only with no CO2 enrichment could not be established due to logistic constraints. Therefore, the present experimental setup cannot separate the actual intended effects of CO2 treatments from the possible influence of the physical disturbance by the manipulated flux of gas per se. However, the fact that the used experimental device was analogous for all experimental units and conditions, allowed to test for the relative effects of the treatments in an unconfounded way. In each of the liners and header tanks, seawater temperature, salinity, oxygen concentration and pH were monitored a total of 31 times (at 2 to 6 days intervals) during the course of the experiment (October 2012 to February 2013) using macroprobes.

3.1. Preliminary survey and effectiveness of experimental treatments The PERMANOVA performed on six liners before the start of the experiment did not detect any significant differences in the structure of macrofaunal assemblages between replicates belonging to each of two randomly established groups (MS = 2717.6, pseudo-F1,4 = 2.3, p > 0.1, full details are reported in Appendix S1). The chosen levels of CO2 concentration were capable of producing clear differences in pH between treatments (Appendix S2; see also Queiros et al. 2015, Rastelli et al., 2015). On the contrary, temperature, salinity and O2 values were maintained considerably constant and comparable across all liners independently of the treatment (Appendix S2; see also Queiros et al. 2015, Rastelli et al., 2015). Analogously, the sediment grain size was very similar (mean ± SE = 10.37 ± 0.49 μm) among the five (one per experimental condition) liners examined before the start of the experiment. 3.2. Macrofaunal abundance, diversity and community structure responses to increasing CO2 A total of 173 macrofaunal species or higher taxa (102 Annelida, 27 Mollusca, 23 Arthropoda, 7 Sipuncula, 7 Echinodermata, 6 Cnidaria, 1 Nemertea and 1 Hemichordata – Appendix S3) were identified in the experiment and used for PERMANOVA on the whole assemblage structure. Macrofaunal assemblages changed between experimental conditions depending on time, irrespectively of the square root or the presence/absence transformation (Table 1). At 2 weeks of exposure, pairwise tests indicated a significant difference between the control and all treatments, but the 20,000 ppm CO2 treatment. At 20 weeks, the only significant difference was between the control and the highest CO2 treatment (Table 1 and Fig. 1 A, B). However, both MDS ordination plots based on square root- and presence/absence-transformed data did not show a clear separation between centroids corresponding to each treatment and exposure time, while the dispersion of the points representing assemblages exposed to the 20 weeks exposure was clearly larger than that of points corresponding to assemblages exposed to the

2.4. Statistical analyses Permutational multivariate analysis of variance (PERMANOVA) was used to test for the null hypothesis of no differences in the macrofaunal community structure among experimental CO2 treatments and for their consistency independently of the exposure time (Anderson, 2001). The analysis was based on Bray-Curtis dissimilarities, calculated from the whole matrix of square root-transformed (to reduce the weight of the most abundant taxa) abundance data, and a two-way model including the crossed factors ‘Time’ (random, two levels: short- vs. medium term exposure; note that treating this factor as random was driven by the fact that we did not intend to examine differences precisely between the 2 and the 20 weeks exposure, but only to test for the consistency of the effects of acidification treatments between ‘a few weeks’ and a ‘some months’ exposure, denominated as short- and medium-term, respectively) and ‘Treatment’ (fixed, five levels: 400, 1000, 2000, 10,000 and 20,000 ppm CO2). The four liners allocated to each combination of factors provided the replicates for the analysis. Since the Bray-Curtis measure combines differences in both the identity and the relative abundance of taxa between samples, the same analysis was repeated twice using, as the original input data matrix, either square roottransformed abundances, or presence/absence data (Clarke & Green, 1988). The PERMDISP test was used to assess whether multivariate differences among groups were due to differences in the dispersion rather than in the location of centroids (Anderson, 2006). Multivariate patterns were illustrated by non-metric multidimensional scaling (nMDS) ordination based on Bray–Curtis dissimilarities calculated on both square-root and presence-absence data. The same model of analysis, but based on Euclidean distances between samples, was used to test for responses to experimental treatments of the total abundance, total richness of taxa and the abundance of individual conspicuous (the most common in all treatments) macrofaunal taxa. When relevant, post-hoc comparisons between levels of the CO2 treatment were performed with paired t-tests. All analyses were carried out using the PRIMER 6.0 & PERMANOVA+ β 3 package (Anderson et al., 2008).

Table 1 Permutational multivariate analysis of variance (PERMANOVA) and post-hoc pair-wise tests examining short- and medium term differences in macrofaunal assemblages in each experimental condition (400, 1000, 2000, 5000, 20,000 ppm CO2). Significant results are reported in bold. Analysis based on square-root transformed Bray-Curtis dissimilarities (A) and presence-absence matrix of data (B). Source of variation

df

MS

pseudo-F

p

# unique perm.

A CO2 treatment = C 4 1314.0 0.9 0.628 998 Time 1 5156.3 5.6 0.001 997 C × Time 4 1433.3 1.6 0.012 998 Residual 30 917.1 Pairwise-test Short term (2 weeks): 400 ≠ 1000 = 2000 = 5000; no alternative to the null hypothesis for the other comparisons Medium term (20 weeks): 400 ≠ 20,000 ppm; no alternative to the null hypothesis for the other comparisons B CO2 treatment = C 4 1342.1 1.5 0.010 998 Time 1 4663.4 5.3 0.001 998 C × Time 4 1364.8 1.6 0.006 998 Residual 30 875.9 Pairwise-test Short term (2 weeks): 400 ≠ 1000 = 2000 = 5000; no alternative to the null hypothesis for the other comparisons Medium term (20 weeks): 400 ≠ 20,000 ppm; no alternative to the null hypothesis for the other comparisons

521

Marine Pollution Bulletin 128 (2018) 519–526

T. Amaro et al.

A)

Table 2 Permutational multivariate analysis of variance (PERMANOVA) and post-hoc pair-wise tests examining short- and medium term differences in total richness (A) and total abundances of individuals (B) in each experimental condition (400, 1000, 2000, 5000, 20,000 ppm CO2). Significant results are reported in bold. Source of variation

Stress: 0.15

B)

df

MS

pseudo-F

p

# unique perm.

A CO2 treatment = C 4 197.2 1.4 0.340 998 Time 1 48.4 1.1 0.300 976 C × Time 4 136.5 2.9 0.035 999 Residual 30 45.9 Pairwise-test Short term (2 weeks): 400 > 1000 = 2000 = 5000 ppm, no alternative to the null hypothesis for the other comparisons Medium term (20 weeks): 5000 = 400 = 1000 = 2000 = 20,000 ppm

1000 ppm - 2 weeks 2000 ppm - 2 weeks 5000 ppm - 2 weeks 20000 ppm - 2 weeks 400 ppm (control) - 2 weeks 1000 ppm - 20 weeks 2000 ppm - 20 weeks 5000 ppm - 20 weeks 20000 ppm - 20 weeks 400 ppm (control) - 20 weeks

B CO2 treatment = C 4 16,820 0.7 0.603 999 Time 1 3294.2 0.6 0.444 996 C x Time 4 22,908 4.0 0.010 999 Residual 30 5691 Pairwise-test Short term (2 weeks): 400 > 2000 = 5000; no alternative to the null hypothesis for the other comparisons Medium term (20 weeks): 5000 = 1000 = 20,000 = 400 = 2000 ppm

A) Total richness

60

1000 ppm 2000 ppm 5000 ppm 20000 ppm 400 ppm (control)

# of taxa / 0.09 m

2

50

Stress: 0.19 Fig. 1. nMDS ordination for the Bray–Curtis similarity from A) square root transformed macrofauna species abundance data at each CO2 treatment (2 weeks; 20 weeks), B) presence/absence transformed macrofaunal species abundance data at each CO2 treatment (2 weeks; 20 weeks).

40 30 20 10 0

shorter exposure (Fig. 1 A and PERMDISP: F = 62.1, p = 0.001; Fig. 1 B and PERMDISP: F = 55.4, p = 0.001). Both total richness of taxa and total abundance of individuals differed among treatments depending on the exposure time (Table 2). Specifically, 2 weeks after the start of the experiment, the control hosted a larger number of taxa than all treatments, but the 20,000 ppm, while no significant differences were displayed at 20 weeks of exposure (Table 2 and Fig. 2 A). The total number of individuals per m2 at 2 weeks was also higher in the control than in any treatment, but significant differences were detected only relative to the 2000 and 5000 ppm CO2 concentrations. Analogously to richness, all significant differences disappeared after 20 weeks exposure (Table 2 and Fig. 2 B). The analysis performed on the abundance of the 8 most common macrofaunal taxa were tested between the different experimental conditions over time. Specifically, after two weeks of exposure, the abundance of the polychaete Heteromastus filiformis was higher, although not significantly, in the control than in all CO2 treatments. At 20 weeks, this species was less abundant in the control than in the two highest concentrations (Table 3 and Fig. 3 A). Another polychaete, Prionospio cirrifera, was, on the other hand, significantly more abundant in the control than in the 2000 ppm treatment after two weeks of exposure, while no significant differences in the abundance of this species occurred at twenty weeks of exposure (Table 3 and Fig. 3 B). Three taxa, namely the Nemertea, and the polychaetes Paradoneis eliasoni/lyra and Paramphinome jeffreysii differed significantly between

2 weeks

20 weeks B) Total abundance

600

# of individuals / 0.09 m

2

500 400 300 200 100 0

2 weeks

20 weeks

Fig. 2. The effects of time and pH on the mean (+SE, n = 4) macrofaunal A) species richness at each CO2 treatment and exposure time (2 weeks; 20 weeks), B) mean total number of individuals (+SE, n = 4) at each CO2 treatment and exposure time (2 weeks; 20 weeks).

522

Marine Pollution Bulletin 128 (2018) 519–526

T. Amaro et al.

Table 3 Permutational multivariate analysis of variance (PERMANOVA) and post-hoc tests (when relevant) examining short- (2 weeks exposure) and medium (20 weeks exposure) term differences in the abundance of macrofaunal taxa in each experimental condition (400, 1000, 2000, 5000, 20,000 ppm CO2). Significant results are reported in bold. Source of variation

Heteromastus filiformis df

MS

Prionospio cirrifera

pseudo-F

4 2850.0 0.6 CO2 treatment = C Time 1 325.9 0.2 C × Time 4 4876.9 2.7 Residual 30 1797.7 Post-hoc test 2 wk.: 400 = 20,000 = 5000 = 1000 = 2000 20 wk.: 5000 = 20,000 > 400

Source of variation

CO2 treatment = C Time C × Time Residual

Nemertea

p

MS

0.688 0.673 0.047

789.2 0.4 0.814 1626.6 3.6 0.059 1956.3 4.3 0.007 453.2 35.1 Post-hoc test 2 wk.: 400 > 2000 20 wk.: 400 = 1000 = 2000 = 20,000 = 5000

Paramphinome jeffreysii

pseudo-F

p

Adontorhina similis

Paradoneis eliasoni/lyra

MS

pseudo-F

p

MS

pseudo-F

p

105.2 545.6 43.5

2.4 15.5 1.2 51.4

0.214 0.001 0.300

30.9 717.3 18.5

1.7 14.0 0.4

0.300 0.001 0.837

Chaetozone sp.

Thyasira equalis

df

MS

pseudo-F

p

MS

pseudo-F

p

MS

pseudo-F

p

MS

pseudo-F

p

4 1 4 30

57.0 614.7 31.9 82.8

1.8 7.4 0.4

0.300 0.011 0.848

123.4 42.7 71.9 76.1

1.7 0.6 0.9

0.295 0.452 0.457

110.7 1.9 53.7 87.0

2.1 0.02 0.6

0.252 0.893 0.652

77.0 529.9 77.0 24.9

1.0 21.3 3.1

0.483 0.001 0.022

conditions may have led to behavioural and/or metabolic changes, ultimately leading to mortality and, consequently, to changes in the whole benthic community composition. Notably, previous studies have also reported a rapid, negative impact on macrofaunal diversity and structure from a controlled experimental release of CO2 from below the seafloor. Even though this response only became evident five weeks after the start of the release it took several weeks for the within sediment porewater pH to drop significantly. This was due to natural chemical buffering processes which affected the carbonate dynamics (Lichtschlag et al., 2014; Taylor et al., 2015; Widdicombe et al., 2015). In the present study, it seems that there was less potential for the sediment buffer the changes in seawater chemistry and the impacts on infauna occurred rapidly. This highlights the importance of understanding how the different chemical and biological characteristics of different sediments will affect the speed of impact following a CO2 leak. The results observed after medium-term exposure (20 weeks) in the current study would suggest that the only impacts of prolonged CO2 exposure were observed in the highest treatment level (20,000 ppm). This is in contradiction to previous mesocosm studies that have shown significant impacts of CO2 exposure to persist over many weeks (e.g. Widdicombe et al. 2009). However, it should be noted that in the current study the similarity observed between the control treatments and the majority of CO2 exposure treatments after 20 weeks was not due to any recovery in the fauna of the CO2 treatments but due to a decrease in the abundance and diversity of the fauna in the control treatments. So it could be hypothesized that the similarity of assemblages exposed to almost all treatments after the twenty weeks exposure was due, at least in part, to the negative effect of holding this particular fauna under mesocosm conditions. Such negative impacts could result from limiting food availability (e.g. Guppy and Withers, 1999), which, once maintained over or occurred after a relatively long period, could have exerted a negative influence on macrofaunal assemblages able to mask any concomitant effect of CO2. Unfortunately, we do not have empirical data suitable to unambiguously support or discard this hypothesis. Other stressful environmental variables, such as temperature, desiccation, anoxia and hypersalinity, which are capable of inducing drastic reductions in metabolic rates of almost all animal taxa (Guppy and Withers, 1999), could also have, in principle, occurred in mesocosms and played a role in the present findings. The continuous supply of mesocosms with new water from the adjacent fjord, however, suggests that reaching drastically limiting conditions of such variables was also unlikely during the experiment. In addition, the previous

times of exposure, irrespectively of the CO2 concentration (Table 3), with the first two taxa being, on average, less abundant after 2 than after 20 weeks of exposure, and the other species displaying the opposite pattern (Fig. 3 C, D and E). Of the remaining three taxa, the bivalve Adontorhina similis and the polychaete Chaetozone sp. were not significantly affected by any CO2 treatment applied over any time (Table 3 and Fig. 3 F and G), while the bivalve Thyasira equalis was comparably abundant in each CO2 treatment at 2 weeks exposure and completely absent at 20 weeks exposure (Table 3 and Fig. 3 H). 4. Discussion The present study was designed to investigate the impact on the abundance and diversity of benthic macrofaunal assemblages of exposure to a plume of CO2-enriched seawater that could result from CO2 leakages from sub-seabed CCS. Results indicated that over a short-term period (2 weeks), the macrofaunal assemblage structure was significantly affected by all experimental levels of increased CO2, with the only exception of the highest concentration. Conversely, after 20 weeks of exposure, the only significant difference was between the control assemblages and those subject to the highest CO2 concentration. Rapid impacts on macrofauna community structure, diversity and abundance following short-term exposure to elevated CO2, similar to that seen in the current experiment, has been reported from a number of previous mesocosm studies (e.g. Widdicombe et al., 2009, Meadows et al., 2015). In this study, after 2 weeks of exposure, the control treatments hosted a larger number of taxa than all treatments, but the 20,000 ppm. Similarly, the total abundance was larger in the control than in any treatment, but significant differences occurred only relative to the 2000 and 5000 ppm CO2 concentrations. The apparent lack of impact in the 20,000 ppm treatments is perhaps surprising. However, one explanation could be that many of the organisms in this treatment had actually died as a result of this extremely high CO2 exposure but their bodies had not had time to decay, especially if the microbial decomposition was also inhibited by the low pH, and these organisms were then falsely counted as living in the subsequent analysis. Another possibility is that under very extreme CO2 exposure organisms go into a severe state of metabolic depression that maintains them for a limited period of time before they inevitably die. The rapid response of all CO2 enriched treatments, except the 20,000 ppm treatment discussed above, indicates that the most sensitive species were likely affected negatively by even relatively low CO2 treatments injected just for 2 weeks. Such 523

Marine Pollution Bulletin 128 (2018) 519–526

T. Amaro et al.

A) Heteromastus filiformis

1000 ppm 2000 ppm 5000 ppm 20000 ppm 400 ppm (control)

B) Prionospio cirrifera 200

150

150

100

100

50

50

0

0

# individuals / 0.09 m

2

200

C) Nemertea

D) Paradoneis eliasoni / lyra 200

150

150

100

100

50

50

0

0

# individuals / 0.09 m

2

200

E) Paramphinome jeffreysii

F) Adontorhina similis 200

150

150

100

100

50

50

0

0

# individuals / 0.09 m

2

200

G) Chaetozone sp.

H) Thyasira equalis 200

150

150

100

100

50

50

# individuals / 0.09 m

2

200

0

0

2 weeks

2 weeks

20 weeks

20 weeks

Fig. 3. The effects of time and pH on the mean (+SE, n = 4) number of individuals/0.09 m ) of the most abundant taxa (A – H. filiformis, B – P. cirrifera, C-Nemertea, D – P. eliasoni/lycra, E – P. jeffreysii, F – A. similis, G – Chaetozone sp., H – T. equalis) from the pooled samples at each CO2 treatment (2 weeks; 20 weeks). 2

was less suitable for mesocosm experimentation than that which was used in the previous studies. In the current experiment the sediment was collected from an area twice as deep than the area used for collection of materials in Widdicombe et al. (2009); 100 m compared with

mesocosm experiments of Widdicombe et al. (2009) used similar conditions as used in the current study and saw no evidence of detrimental mesocosm impacts during a 20 week experiment. It is most likely therefore that the sediment or community selected for this experiment

524

Marine Pollution Bulletin 128 (2018) 519–526

T. Amaro et al.

specimens/species from surrounding non-impacted systems (Danovaro 2010; Widdicombe et al., 2015). Moreover, present findings cannot obviously provide any unambiguous information to derive expectations on possible responses, even of the same assemblages, to longer-term (> 20 weeks) exposure, or on responses to any exposure of assemblages dominated by other groups of organisms, such as echinoderms (Spicer et al. 1988, Spicer 1995, Kroeker et al., 2013) and more calcified taxa. For example, it is reasonable to assume that calcifying organisms would be particularly sensitive to increases of CO2 and consequent reductions of pH, which could eventually lead to critical loss of their fitness and survival rates through the allocation of more energy to ion removal processes to detriment of other important physiological processes (Pörtner, 2008; Wood et al., 2008). In this context, other studies carried out in natural acidic shallow vents (e.g. Rodolfo-Metalpa et al. 2011; Gambi et al., 2016; Kamenos et al., 2016) indicated that even CO2 increases considerably less than in the present experiment can determine profound changes in exposed benthic assemblages. None of such studies, however, are fully comparable to the present one. Specifically, Rodolfo-Metalpa et al. (2011) focused on calcifying organisms, which, instead, were almost not represented in present macrofaunal assemblages. Both Gambi et al. (2016) and Kamenos et al. (2016) examined the distribution and diversity of benthic organisms (coralline algae and polychaetes) at increasing distance and increasing pH from natural vents, thus along natural gradients of CO2 concentrations to which such organisms were adapted for a much longer time compared to the temporal scale of our experiment.

50 m. Given that many potential CCS sites are located in deep water, it may be that the value of mesocosm experiments may be limited to assessing short term exposures and that there is a greater need for developing in-situ experimental procedures in these areas. Despite the issues associated with mesocosm effects, it was clear that over a longer exposure, the communities in all the CO2 enriched treatments, except 20,000 ppm, converged as the hardiest and most resistant species persisted (the number of taxa that were absent in all treatments almost doubled from the 2 weeks to the 20 weeks time). What actually constitutes a resistant species to elevated levels of CO2 will depend on the specific metabolic and physiological adaptations of macrofaunal organisms (see Widdicombe and Spicer, 2008). This resistance, therefore, is largely variable among taxa, both in terms of overall extent and underlying mechanisms (Lessin et al., 2016). Echinoderms show very little compensation for hypercapnia-related disturbance (Spicer et al., 1988, Spicer 1995, Kroeker et al., 2013). Calcifying organisms need to increase pH (by active removal of H+ ions) in order to maintain the formation of biogenic structures where needed (Widdicombe et al., 2015). Other organisms prefer to suppress metabolism by shutting down various cellular processes (Guppy and Withers, 1999, Widdicombe et al., 2009). In this study, the exposure to the most extreme CO2-leakage scenario, in a medium-term (20 weeks), was tolerated by highly resistant taxa, such as borrowing polychaete worms from the family Capitellidae (to which H. filiformis belongs). In fact, this taxon is described as opportunistic able to dominate macrofaunal invertebrate communities under perturbed conditions, likely due to its short generation time and direct development which can allow their efficient use of the habitat (Pearson and Rosenberg, 1978; Berge, 1990; Preckler, 2015) and increases in biomass after the elimination of more sensitive species (Lessin et al., 2016). This ability has been also explored for restoring polluted sediments by adding bioturbating species of capitellid polychaetes (Chareonpanich et al. 1994, Ueda et al. 1994). Therefore, the fact that we did not observe a reduced abundance of H. filiformis in the most acidified treatments compared with the control is consistent with its known disturbance-resistant trait as a capitellid polychaete. At the same time, this species was the most abundant in the examined macrofaunal samples, hence its response to experimental treatments could have driven, at least for a considerable part, that of the whole structure of assemblages. It is worth noting, however, that the convergence between more acidified and control assemblages could have been also modulated by the temporal variability (obviously due to processes other than changes in CO2 inputs) of the latter ones, which can be as large as that driven by the increased CO2 inputs (Widdicombe et al., 2015). In principle, temporal fluctuations in patterns of abundance and diversity of macrofaunal assemblages in the control could have made them similar, even just by chance, to the treated ones during the experiment. A large and significant temporal variability, irrespectively of CO2 treatments, was confirmed by several conspicuous taxa here examined. Although the results of the current mesocosm study can provide crucial information on actual cause-effect relationships between CO2 enrichment and macrofaunal responses, any attempt to extrapolate them to predicting the ecological and biological consequences of possible field leaks should be made with caution. The main reasons for this include: a) the mesocosms being a confined system, which does not allow an organism to escape or relocate to avoid unfavourable conditions, potentially leading to overestimate mortality rates over scales larger than the experimental one; b) the likelihood that the response is specific for the examined assemblage, which, in spite of even analogous main traits, would not be necessarily the same as another one from a different location and/or time; c) possible different buffering effects due to the different mineralogy of the sediments, with special focus on carbonate content; d) the potential change between mesocosms and field biological responses due to the drastic difference in the depth (hydrostatic pressure) to which the system is exposed; e) the lower resilience or recovery of communities due to the lack of immigration of

5. Conclusion In spite of all the above listed factors and processes which are likely to jeopardise the accuracy of extrapolations of mesocosm findings to real circumstances, the structure of the present experiment was suitable to examine the relative responses of macrofaunal assemblages and individual taxa to increased CO2 inputs in an unconfounded way. As such, present findings, although not guaranteeing that field responses to possible future leakages associated to CCS strategies will be exactly the same, reasonably suggest that the main significant impact of such events on macrofaunal assemblages would occur close to the hypothetical source of CO2 and would occur rapidly over short time periods (< 2 weeks). Additional experiments are needed to understand the mechanisms responsible for the present findings and their possible consistency under field conditions, but this controlled study provides a relevant contribution to the debate on the cost-benefit balance (environmental risk vs. intended goals) of CCS technologies. Supplementary data to this article can be found online at https:// doi.org/10.1016/j.marpolbul.2018.01.068. Acknowledgements This research was conducted as part of the European Community's Seventh Framework Programme FP7/2007-2013 for the project Subseabed CO2 storage: impact on marine ecosystems (ECO2), grant agreements N. 265847, DEVelopment of innovative tools for understanding marine biodiversity and assessing good Environmental Status (DEVOTES), grant agreement no. 308392 and FME Success. TA was partially supported by Marie Curie Actions through the project CEFMED (project number 327488). We are grateful to Oddbjorn Petersen, Per Ivar Johannessen, Morten Schaanning personnel at the Marine Research Station (Solbergstrand, Norway) of the Norwegian Institute of Water Research (NIVA, Oslo, Norway) and at Plymouth Marine Laboratory for support and advice during the ECO2 mesocosm experiments. Dr Mats Walday at NIVA, Dr Andrew Sweetman from Heriot Watt University of Edinburgh and Dr Sarah Dashfield and Dr Carolyn Harris from Plymouth Marine Laboratory are also thanked for support during the organization and analyses for this work. Additionally, we would like to thank Dr Victor Quintino for the useful comments to an earlier version 525

Marine Pollution Bulletin 128 (2018) 519–526

T. Amaro et al.

of the manuscript.

Lessin, G., Artioli, Y., Queiros, A.M., Widdicombe, S., Blackford, J.C., 2016. Modelling impacts and recovery in benthic communities exposed to localised high CO2. Mar. Pollut. Bull. 109 (1), 267–280. Lichtschlag, A., James, R.H., Stahl, H., Connelly, D., 2014. Effect of a controlled subseabed release of CO2 on the biogeochemistry of shallow marine sediments, their pore waters, and the overlying water. Int. J. Greenh. G. Con. 38, 80–92. McCave, I.N., 2013. Evaluation of a laser-diffraction-size analyzer for use with natural sediments. J. Sediment. Res. 56, 561–564. McCormack, C.G., 2016. Key impacts of climate engineering on biodiversity and ecosystems, with priorities for future research. J. Integr. Environ. Sci. 13 (2-4), 103–128. Meadows, A.S., Ingels, J., Widdicombe, S., Hale, R., Rundle, S.D., 2015. Effects of elevated CO2 and temperature on an intertidal meiobenthic community. J. Exp. Mar. Biol. Ecol. 469, 44–56. Mora, C., Wei, C.-L., Rollo, A., Amaro, T., Baco, A.R., Billett, D., Bopp, L., Chen, Q., Collier, M., Danovaro, R., Gooday, A.J., Grupe, B.M., Halloran, P.R., Ingels, J., Jones, D.O.B., Levin, L.A., Nakano, H., Norling, K., Ramirez-Llodra, E., Rex, M., Ruhl, H.A., Smith, C.R., Sweetman, A.K., Thurber, A.R., Tjiputra, J.F., Usseglio, P., Watling, L., Wu, T., Yasuhara, M., 2013. Biotic and human vulnerability to projected changes in ocean biogeochemistry over the 21st century. PLoS Biol. 11 (10), e1001682. Paulley, A., Maul, P.R., Metcalfe, R., 2012. Research into impacts and safety in CO2 storage: Scenarios for potential impacts from hypothetical leakage from geological storage facilities for carbon dioxide. In: Quintessa Limited (Ed.), RISCS Deliverable D1.3, 60 pp. Pearson, T.H., Rosenberg, R., 1978. Macrobenthic succession in relation to organic enrichment and pollution of the marine environment. Oceanogr. Mar. Biol. Annu. Rev. 16, 229–311. Pörtner, H.-O., 2008. Ecosystem effects of ocean acidification in times of ocean warming: a physiologist's view. Mar. Ecol. Prog. Ser. 373, 203–217. Preckler, C.A., 2015. Biodiversity and Chemical Interactions in Antartic Benthic Communities of Deception Island (South Shetland Islands). PhD thesis. 314 pp. Barcelona University. Pusceddu, A., Dell'Anno, A., Fabiano, M., Danovaro, R., 2009. Quantity and bioavailability of sediment organic matter as signatures of benthic trophic status. Mar. Ecol. Prog. Ser. 375, 41–52. Queiros, A.M., Taylor, P., Cowles, A., Reynolds, A., Widdicombe, S., Stahl, H., 2014. Optical assessment of impact and recovery of sedimentary pH profiles in ocean acidification and carbon capture and storage research. Int. J. Greenh. G. Con. 38, 110–120. Queiros, A.M., Norling, K., Amaro, T., Nunes, J., Cummings, D., Yakushev, E., Sorensen, K., Harris, C., Woodward, M., Danovaro, R., Rastelli, E., Alve, E., DeVittor, C., Karuza, A., Cibic, T., Monti, M., Ingrosso, G., Fornasaro, D., Beaubien, S.E., Guilini, K., Vanreusel, A., Molari, M., Boetius, A., Ramette, A., Wenzhöfer, F., de Beer, D., Weber, M., Grünke, S., Bigalke, N., Widdicombe, S., 2015. Potential Impact of CCS Leakage on Marine Communities. Ed. Plymouth Marine Laboratory. 101 pp. Plymouth, UK. . Rastelli, E., Corinaldesi, C., Dell'Anno, A., Amaro, T., Queiros, A.M., Widdicombe, S., Danovaro, R., 2015. Impact of CO2 leakage from sub-seabed carbon dioxide capture and storage (CCS) reservoirs on benthic virus-prokaryote interactions and functions. Front. Microbiol. 6, 935. http://dx.doi.org/10.3389/fmicb.2015.00935. Rodolfo-Metalpa, R., Houlbrèque, F., Tambutté, É., Boisson, F., Baggini, C., Patti, F.P., Jeffree, R., Fine, M., Foggo, A., Gattuso, J.-P., Hall-Spencer, J.M., 2011. Coral and mollusc resistance to ocean acidification adversely affected by warming. Nat. Clim. Chang. 1, 308–312. Spicer, J.I., 1995. Oxygen and acid–base status of the sea urchin Psammechinus miliaris during environmental hypoxia. Mar. Biol. 124, 71–76. Spicer, J.I., Taylor, A.C., Hill, A.D., 1988. Acid–base status in the sea urchins Psammechinus miliaris and Echinus esculentus (Echinodermata: Echinoidea) during emersion. Mar. Biol. 99, 527–534. Tait, K., Stahl, H., Taylor, P., Widdicombe, S., 2014. Rapid response of the active microbial community to CO2 exposure from a controlled sub-seabed CO2 leak in Ardmucknish Bay (Oban, Scotland). Int. J. Greenh. G. Con. 38, 171–181. Tait, K., Beesley, A., Findlay, H.S., McNeill, C.L., Widdicombe, S., 2015. Elevated CO2 induces a bloom of microphytobenthos within a shell gravel mesocosm. FEMS. http://dx.doi.org/10.1093/femsec/fiv092. Taylor, P., Stahl, H., Vardy, M.E., Bull, J.M., Akhurst, M., Hauton, C., James, R.H., Lichtschlag, A., Long, D., Aleynik, D., Toberman, M., Naylor, M., Connelly, D., Smith, D., Sayer, M.D.J., Widdicombe, S., Wright, I.C., Blackford, J., 2015. A novel subseabed CO2 release experiment in forming monitoring and impact assessment for geological carbon storage. Int. J. Greenh. G. Con. 38, 3–17. Ueda, N., Tsutsumi, H., Yamada, M., Takeuchi, R., Kido, K., 1994. Recovery of the marine bottom environment of a Japanese bay. Mar. Pollut. Bull. 28, 676–682. Widdicombe, S., Needham, H.R., 2007. Impact of CO2 induced seawater acidification on the burrowing activity of Nereis virens and sediment nutrient flux. Mar. Ecol. Prog. Ser. 341, 111–122. Widdicombe, S., Spicer, J.I., 2008. Predicting the impact of ocean acidification on benthic biodiversity: what can animal physiology tell us? J. Exp. Mar. Biol. Ecol. 366 (1–2), 187–197. Widdicombe, S., Dashfield, S.L., McNeill, C.L., Needham, H.R., Beesley, A., McEvoy, A., Øxnevad, S., Clarke, K.R., Berge, J.A., 2009. Effects of CO2 induced seawater acidification on in faunal diversity and sediment nutriente fluxes. Mar. Ecol. Prog. Ser. 379, 59–75. Widdicombe, S., Blackford, J.C., Spicer, J.I., 2013. Assessing the environmental consequences of CO2 leakage from geological CCS: generating evidence to support environmental risk assessment. Mar. Pollut. Bull. 73, 399–401. Widdicombe, S., McNeill, C.L., Stahl, H., Taylor, P., Queiros, A.M., Nunes, J., Tait, K., 2015. Impact of sub-seabed CO2 leakage on macrobenthic community structure and diversity. Int. J. Greenh. G. Con. 38, 182–192. Wood, H.L., Spicer, J.I., Widdicombe, S., 2008. Ocean acidification may increase calcification rates—but at a cost. Proc. R. Soc. B 275, 1767–1773.

References Anderson, M.J., 2001. A new method for non-parametric multivariate analysis of variance. Austral. Ecol. 26, 32–46. Anderson, M.J., 2006. Distance-based tests for homogeneity of multivariate dispersions. Biometrics 62, 245–253. Anderson, M.J., Gorley, R.N., Clarke, K.R., 2008. PERMANOVA + for PRIMER: Guide to Software and Statistical Methods. PRIMER-E (Plymouth). Berge, J.A., 1990. Macrofauna recolonization of subtidal sediments. Experimental studies on defaunated sediment contaminated with crude oil in two Norwegian fjords with unequal eutrophication status. I. Community responses. Mar. Ecol. Prog. Ser. 66, 103–115. Blackford, J.C., Jones, N., Proctor, R., Holt, J., Widdicombe, S., Lowe, D., Rees, A., 2009. An initial assessment of the potential environmental impact of CO2 escape from marine carbon capture and storage systems. Proc. Inst. Mech. Eng. Part A – J. Power Energy 223, 269–280. Blackford, J.C., Hattam, C., Widdicombe, S., Burnside, N., Naylor, M., Kirk, K., Maul, P., Wright, I., Gluyas, J., Mathias, S., 2013. CO2 leakage from geological storage facilities: environmental, societal and economic impacts, monitoring and research strategies. In: Geological Storage of Carbon Dioxide (CO2): Geoscience, Technologies, Environmental Aspects and Legal Frameworks. Woodhead publishing, Swaston, Cambridge, GB, pp. 149–178. Blackford, J.C., Stahl, H., Bull, J.M., Bergès, B.J.P., Cevatoglu, M., Lichtschlag, A., Connelly, D., James, R.H., Kita, J., Long, D., Naylor, M., Shitashima, K., Smith, D., Taylor, P., Wright, I., Akhurst, M., Chen, B., Gernon, T.M., Hauton, C., Hayashi, M., Kaieda, H., Leighton, T.G., Sato, T., Sayer, M.D.J., Suzumura, M., Tait, K., Vardy, M.E., White, P.R., Widdicombe, S., 2014. Detection and impacts of leakage from subseafloor deep geological carbon dioxide storage. Nat. Clim. Chang. 4, 1011–1016. Caldeira, K., Wickett, M., 2003. Anthropogenic carbon and ocean pH. Nature 425, 365. Cerrano, C., Cardini, U., Bianchelli, S., Corinaldesi, C., Pusceddu, A., Danovaro, R., 2013. Red coral extinction risk enhanced by ocean acidification. Sci. Rep. 3. Chareonpanich, C., Tsutsumi, H., Montani, S., 1994. Efficiency of the decomposition of organic matter, loaded on the sediment, as a result of the biological activity of Capitella sp. I. Mar. Pollut. Bull. 28, 314–318. Clarke, K.R., Green, R.H., 1988. Statistical design and analysis for a ‘ biological effects ’ study. Mar. Ecol. Prog. Ser. 46, 213–226. Damen, K., Faaij, A., Turkenburg, W., 2006. Health, safety and environmental risks of underground CO2 storage — overview of mechanisms and current knowledge. Clim. Chang. 74 (1), 289–318. Danovaro, R., 2010. In: Taylor & Francis (Ed.), Methods for the Study of Deep-sea Sediments, Their Functioning and Biodiversity. CRC Press, London 456 pp. Fabiano, M., Pusceddu, A., 1998. Total and hydrolizable particulate organic matter (carbohydrates, proteins and lipids) at a coastal station in Terra Nova Bay (Ross Sea, Antarctica). Polar Biol. 19, 125–132. Feely, R.A., Sabine, C.L., Lee, K., Berelson, W., Kleypas, J., Fabry, V.J., Millero, F.J., 2004. Impact of anthropogenic CO2 on the CaCO3 system in the oceans. Science 305, 362–366. Gambi, M.C., Musco, L., Giangrande, A., Badalamenti, F., Micheli, F., Kroeker, K.J., 2016. Distribution and functional traits of polychaetes in a CO2 vent system: winners and losers among closely related species. Mar. Ecol. Prog. Ser. 550, 121–134. Gattuso, J.P., Magnan, A., Billé, R., Cheung, W.W., Howes, E.L., Joos, F., Allemand, D., Bopp, L., Cooley, S.R., Eakin, C.M., Hoegh-Guldberg, O., Kelly, R.P., Pörtner, H.O., Rogers, A.D., Baxter, J.M., Laffoley, D., Osborn, D., Rankovic, A., Rochette, J., Sumaila, U.R., Treyer, S., Turley, C., 2015. Contrasting futures for ocean and society from different anthropogenic CO2 emissions scenarios. Science 349, aac4722. Gibbins, J., Haszeldine, S., Holloway, S., Pearce, J., Oakey, J., Shackley, S., Turley, C., 2006. Scope for future CO2 emission reductions from electricity generation through the deployment of carbon capture and storage technologies. In: Schnellhuber, H.J. (Ed.), Avoiding Dangerous Climate Change. Cambridge University Press, pp. 379–384. Guppy, M., Withers, P., 1999. Metabolic depression in animals: physiological perspectives and biochemical generalizations. Biol. Rev. Camb. Philos. Soc. 74, 1–40. Hale, R., Calosi, P., McNeill, L., Mieszkowska, N., Widdicombe, S., 2011. Predicted levels of future ocean acidification and temperature rise could alter community structure and biodiversity in marine benthic communities. Oikos 120, 661–674. Holloway, S., 2007. Carbon dioxide capture and geological storage. Philos. Trans. R. Soc. A Math. Phys. Eng. Sci. 365 (1853), 1095–1107. IPCC, 2013. In: Stocker, T.F. (Ed.), Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA 1535 pp. 2013. Jones, D.G., Beaubien, S.E., Blackford, J.C., Foekema, E.M., Lions, J., De Vittor, C., West, J.M., Widdicombe, S., Hauton, C., Queiros, A., 2015. Developments since 2005 in understanding potential environmental impacts of CO2 leakage from geological storage. International Journal of Greenhouse Gas Control 40, 350–377. Kamenos, N.A., Perna, G., Gambi, M.C., Micheli, F., Kroeker, K.J., 2016. Coralline algae in a naturally acidified ecosystem persist by maintaining control of skeletal mineralogy and size. Proc. R. Soc. B 283, 20161159. Kroeker, K., Kordas Rebecca Crim, R., Hendriks, I., Ramajo, L., Singh, G., Duarte, C., Gattuso, J.-P., 2013. Impacts of ocean acidification on marine organisms: quantifying sensitivities and interaction with warming. Glob. Chang. Biol. 19, 1884–1896. Laverock, B., Kitidis, V., Tait, K., Gilbert, J.A., Osborn, M., Widdicombe, S., 2013. Bioturbation activity determines the response of coastal benthic nitrogen-cycling microorganisms to ocean acidification. Phil. Trans. of the Royal Soc. of London Series B.-Biol. Sci. 368 (1627), 1–13.

526