1 2 3 4 5 6
Effects of sub-seabed CO2 leakage: short- and medium-term responses of benthic macrofaunal 7 assemblages 8 9 10 11 12 13 14
Amaro, T.1,2,3*, Queiros, A.M.4,Rastelli, E.3,5, Borgersen, G.2, Brkljacic, M.2, Nunes, J.4, Bertocci, I. 3, , Sorensen, 15 K.2, Danovaro, R.3,5, Widdicombe, S.4 16 17 18 19 20 21 22 23 24 25 26 27 28
*Corresponding author: teresa.amaro@szn.it
29
1Hellenic Center for Marine Research (HCMR), 710 03 Heraklion, Crete, Greece. 30
2 Norwegian Institute for Water Research, Oslo, Norway. 31
3 Stazione Zoologica Anton Dohrn, Villa Comunale, Naples, Italy 32
4 Plymouth Marine Laboratory, Prospect Place, West Hoe, PL1 3DH, Plymouth, UK
33
5 Department of Life and Environmental Sciences, Polytechnic University of Marche, Ancona, 34
Italy
35
Formattato: Italiano (Italia) Formattato: Italiano (Italia)
Formattato: Inglese (Regno Unito) Formattato: Inglese (Regno Unito)
ABSTRACT 36
The continued rise in atmospheric carbon dioxide (CO2) levels is driving climate change and 37
temperature shifts at a global scale. CO2 Capture and Storage (CCS) technologies have been 38
suggested as a feasible option for reducing CO2 emissions and mitigating their effects. 39
However, before CCS can be employed at an industrial scale, any environmental risks
40
associated with this activity should be identified and quantified. Significant leakage of CO2 41
from CCS reservoirs and pipelines is considered to be unlikely, however direct and/or indirect
42
effects of CO2 leakage on marine life and ecosystem functioning must be assessed, with 43
particular consideration given to spatial (e.g. distance from the source) and temporal (e.g.
44
duration) scales at which leakage impacts could occur. In the current mesocosm experiment we
45
tested the potential effects of CO2 leakage on macrobenthic assemblages by exposing infaunal 46
sediment communities to different levels of CO2 concentration (400, 1000, 2000, 10000 and 47
20000 ppm CO2), simulating a gradient of distance from a hypothetic leakage, over short-term 48
(a few weeks) and medium-term (several months). A significant impact on community
49
structure, abundance and species richness of macrofauna was observed in the short-term
50
exposure. Individual taxa showed idiosyncratic responses to acidification. We conclude that the
51
main impact of CO2 leakage on macrofaunal assemblages occurs almost exclusively at the 52
higher CO2 concentration and over short time periods, tending to fade and disappear at 53
increasing distance and exposure time. Although under the cautious perspective required by the 54
possible context-dependency of the present findings, thisThe results of the present study
55
contributes to the cost-benefit analysis (environmental risk versus the achievement of the
56
intended objectives) of CCS strategies.
58
INTRODUCTION 59
The accelerating rise in atmospheric carbon dioxide (CO2) levels (IPCC, 2013) is 60
causing ocean warming and acidification at unprecedented rates, posing critical threats to
61
single species, habitats, oceanic regions and overall global ecosystem functioning (Caldeira
62
and Wickett, 2003; Feely et al., 2004; Hale et al., 2011; Mora et al., 2013, Cerrano et al., 2013;
63
Meadows et al., 2015; Gattuso et al., 2015). As a direct consequence, it is urgently needed to
64
identify suitable options for reducing/mitigating CO2 emissions (McCormack et al., 2016). One 65
particularly promising technology involves capturing CO2 from point source effluents (mostly, 66
energy generation plants), then transporting it as a supercritical liquid to be stored in deep
67
porous geological rock formations, such as saline aquifers or existing hydrocarbon reservoirs
68
(Gibbins et al. 2006; Holloway 2007). This process is defined as CO2 Capture and Storage 69
(CCS). In Europe and North America the technical feasibility of CCS approaches has been
70
already demonstrated. For example, at the Sleipner West gas field in the Norwegian sector of
71
the North Sea, CCS has been operational since 2000 with approximately 1 million tons of CO2 72
pumped into the storage reservoir every year (Paulley et al., 2012, Jones et al., 2015).
73
However, as with almost any other human activity, this technology is not risk-free in terms of
74
posing potential environmental hazards (reviewed by Damen et al. 2006). Before industrial
75
scale CCS activities become widely accepted and implemented these risks need to be identified
76
and quantified. Perhaps the greatest environmental risk associated with CCS is that of CO2 77
leaking into the marine environment either during transport, sequestration or from the
78
geological storage reservoir itself. Whilst current evidence suggests that leakage from CCS
79
reservoirs would be extremely unlikely it is not impossible (Blackford et al., 2009; 2014).
80
Given that any major increase in seawater CO2 concentrations , and the associated changes in 81
carbonate chemistry, has the potential to considerably impact marine life and ecosystem
82
functions, assessing the biological and ecological effects of CO2 leakage is essential to support 83
environmental risks assessments required by CCS operations (Widdicombe et al., 2013; Jones
84
et al., 2015). This is especially relevant for benthic assemblages living in the immediate
85
proximity of any potential leak, since they would be exposed to relatively large and rapid
86
changes in carbonate chemistry, in both the sediment pore waters and the overlying seawater
87
(Lichtschlag et al., 2014; Queiros et al., 2014).
88
The exposure to a range of CO2 concentrations has been tested on a variety of marine 89
organisms, as well as on some biogeochemical processes and ecosystem functions
90
(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 92
both the severity and the duration of the exposure (Blackford et al., 2013). In general, it is
93
hypothesized that a CCS leakage is immediately associated with a localized acute exposure to
94
harmful high CO2 conditions whose effects are likely to get attenuated at increasing distance 95
from the source. Moreover, more prolonged leakage or persisting influences of temporary
96
seepage of any level could represent chronic stressed conditions to the surrounding abiotic and
97
biological environment (Jones et al., 2015).
98
Whilst previous studies have started to provide a better understanding of the potential
99
impacts of CCS leakage on specific benthic organisms (e.g. Widdicombe & Needham 2007),
100
our knowledge of the possible effects at the community level remains limited (Widdicombe et
101
al. 2015). In addition, the mechanisms underlying such changes are still largely unknown, as
102
well as the difference between direct and indirect effects of increasing CO2 leakages on the 103
macrofaunal community. It has been reported, however, that low-pH levels predicted by
104
realistic scenarios of CCS leakage might severely reduce the prokaryotic-mediated processes
105
(Rastelli et al. 2015), while acidified conditions could favor blooms of benthic microbial
106
primary producers including cyanobacteria and diatoms (Tait et al., 2015). Notably, the
107
exposure to high CO2 levels can alter microbial-mediated processes able to affect the quality 108
and quantity of the sedimentary organic matter (OM) (Rastelli et al., 2015). Since the
109
availability of OM is a key driver of the abundance, distribution and biodiversity of benthic
110
fauna (Fabiano and Pusceddu, 1998, Pusceddu et al., 2009), the effects of changes in this
111
variable due to CCS leakage might indirectly propagate to associated macrofaunal
112
assemblages.
113
Full community level effects of CO2 leakages can only be unequivocally assessed 114
using simulated leakage experiments conducted in the field (e.g. Blackford et al. 2014) or from
115
studying actual leakage events or accidents in areas where data on the response variables of
116
interest are available before and after the event. Both options are normally unavailable either
117
due to the lack of data or to logistic, financial and/or ethical constraints. Performing
118
manipulative experiments in mesocosms can be a feasible alternative especially when the
119
results are used to inform ecosystem level models. A strength of an experimental approach is
120
that by exposing initially comparable assemblages to different levels of CO2 concentration 121
(such as ‘naturally’ occurring along a gradient of distance from a supposed leakage) under
122
controlled conditions allows testing for their relative effects in an unconfounded way. 123
In this study, we performed a mesocosm experiment to test the potential impact of
124
CO2-enriched (from 400 ppm to 20000 ppm) seawater plumes on the abundance and diversity
of soft-bottom macrofauna. Specifically, we tested the null hypotheses that (i) the whole
126
structure (taxon composition and relative abundance), richness, total abundance of the
127
macrofaunal assemblages and the abundance of individual taxa, did not differed depending on
128
the CO2 concentration; (ii) such a lack of differences was consistentchanged between a few 129
weeks (short-term) and some months (medium. term) of continued exposure.
130 131
MATERIAL AND METHODS 132
Collection of sediment samples and associated fauna and mesocosm setup 133
Using a KC Denmark boxcorer, intact sediment samples containing natural infaunal
134
assemblages were collected during the 3rd week of August 2012 from randomly selected points 135
located some meters apart from each other at the outer Oslofjord (59°49.4788’ N, 10°58.8595’
136
E), Norway, at 100 m water depth. Each boxcorer was equipped with an inner liner, which
137
allowed the sediments and the overlying water to be retrieved with minimal disturbance.
138
A total of 46 independent liners (0.09 m2 each, with average sediment penetration of 139
~40 cm) were collected and transferred immediately to the benthic mesocosm systems at the
140
Marine Research Station, Norwegian Institute of Water Research, Solbergstrand, Norway.
141
During transportation, all liners were shaded and continuously covered with seawater to
142
prevent desiccation and minimise temperature changes.
143
The experimental system was set up according to Widdicombe et al. (2009), as
144
described in detail elsewhere (Queiros et al., 2015, Rastelli et al., 2015). Briefly, all liners were
145
placed in an aquarium in a flow-through holding basin filled with seawater to a depth of 1 m
146
(mesocosm) and supplied continuously with unfiltered natural seawater at a flow rate of 120
147
ml/min from a pipeline situated at 60 m depth in the adjacent fjord. All liners were
148
maintained in these conditions for two weeks prior to the beginning of the experiment to allow
149
the fauna, microbes and geochemical processes to acclimatize to mesocosm conditions.
150 151
Preliminary survey 152
To guarantee that the randomly assigned experimental levels of CO2 were not 153
confounded by initial differences between replicate cores in terms of hosted macrofaunal
154
assemblages would have required us to compare macrofauna among all (allocated) treatments
155
before manipulation. Unfortunately, the needed destructive sampling made such an option
156
impossible. Alternatively, a total of 6 liners were chosen at random from the 46 liners initially
157
collected and these 6 were randomly allocated to one of two groups of three. These were then
158
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
160
the other could provide information (not exhaustive, but relevant) to assume that significant
161
differences were not likely to exist also among the sets of replicate liners allocated at random to
162
the experimental levels. In addition, data on the sediment grain size, estimated by laser analysis 163
at the beginning of the experiment, were available for was also estimated by laser analysis one 164
liner per each of the total five experimental conditions (e.g. McCave, 2013).
165
For each liner, macrofaunal assemblages were sampled, after the acclimation period, by
166
sieving all the sediment over a 500 μm mesh, with the residue from each sample being fixed in
167
10% buffered formalin until further processing. In the laboratory, the fauna was extracted from
168
the residue under a binocular microscope and all specimens were sorted into major taxa and
169
then identified to species level whenever possible. Species (or higher taxa) abundance was
170
determined in each replicate and expressed as the total number of individuals per m2 of 171
sampled area.
172 173
Experimental setup and sampling 174
The 40 liners remaining after the preliminary survey were randomly allocated in equal
175
numbers (4) to each of five CO2 treatments: 400 (control), 1000, 2000, 5000, and 20000 ppm, 176
with two sampling times (2 weeks, 20 weeks). These levels were consistent with those
177
specifically tested by Rastelli et al. (2015) and Queiros et al. (2015). Seawater acidification
178
was achieved as described by Widdicombe et al. (2009). Briefly, CO2 gas passed through a 179
450 L reservoir tanks filled with natural seawater. Using an automated feedback relay system
180
(Walchem), the CO2 flux into the reservoir tanks was regulated in order to maintain the 181
required pH level. The reservoir tanks were continuously supplied with natural seawater (pH∼
182
8.1).
183
To assess short-term (a few weeks scale) and medium-term (several months) effects of
184
CO2 exposure, sampling took place after 2 weeks exposure (T1) and again, on a different set of 185
replicate liners not previously sampled, after 20 weeks exposure (T2). At each time,
186
macrofaunal assemblages were sampled as described for the preliminary survey.
187
A procedural control involving the flux of air only with no CO2 enrichment could not 188
be established due to logistic constraints. Therefore, the present experimental setup cannot 189
separate the actual intended effects of CO2 treatments from the possible influence of the 190
physical disturbance by the manipulated flux of gas per se. However, the fact that the used 191
experimental device was analogous for all experimental units and conditions allowed to test for 192
the relative effects of the treatments in an unconfounded way. 193
Formattato: Pedice
Formattato: Tipo di carattere:
In each of the liners and header tanks, seawater temperature, salinity, oxygen
194
concentration and pH were monitored a total of 31 times (at 2 to 6 days intervals) during the
195
course of the experiment (October 2012 to February 2013) using macroprobes.
196 197
Statistical analyses 198
Permutational multivariate analysis of variance (PERMANOVA) was used to test for
199
the null hypothesis of no differences in the macrofaunal community structure among
200
experimental CO2 treatments and for their consistency independently ofpossible variation 201
depending on the exposure time (Anderson, 2001). The analysis was based on Bray-Curtis
202
square-root transformed dissimilarities, calculated from the whole matrix of square root-203
transformed (to reduce the weight of the most abundant taxa) abundance data, and a two-way
204
model including the crossed factors ‘Time’ (random, two levels: short- vs. medium term
205
exposure; note that treating this factor as random was driven by the fact that we did not intend
206
to examine differences precisely between the 2 and the 20 weeks exposure, but only to test for
207
the consistency of the effects of acidification treatments between ‘a few weeks’ and a ‘some
208
months’ exposure, denominated as short- and medium-term, respectively) and ‘Treatment’
209
(fixed, five levels: 400, 1000, 2000, 10000 and 20000 ppm CO2). The four liners allocated to 210
each combination of factors provided the replicates for the analysis. Since the Bray-Curtis
211
measure combines differences in both the identity and the relative abundance of taxa between
212
samples, the same analysis was repeated twice using, as the original input data matrix, either
213
square root-transformedraw abundances, or presence/absence data (Clarke & Green, 1988).
214
The PERMDISP test was used to assess whether multivariate differences among groups were 215
due to differences in the dispersion rather than in the location of centroids (Anderson, 2006). 216
Multivariate patterns were illustrated by non-metric multidimensional scaling (nMDS)
217
ordination based on Bray–Curtis dissimilarities calculated on both square-root and
presence-218
absence data.
219
The same model of analysis, but based on Euclidean distances between samples, was
220
used to test for responses to experimental treatments of the total abundance, total richness of
221
taxa and the abundance of individual conspicuous (the most common in all treatments)
222
macrofaunal taxa.
223
When relevant, post-hoc comparisons between levels of the CO2 treatment were 224
performed with paired t-tests. All analyses were carried out using the PRIMER 6.0 &
225
PERMANOVA+ β 3 package (Anderson et al., 2008).
226 227
RESULTS 228
Preliminary survey and effectiveness of experimental treatments 229
The PERMANOVA performed on six liners before the start of the experiment did not
230
detect any significant differences in the structure of macrofaunal assemblages between
231
replicates belonging to each of two randomly established groups (MS = 2717.6, pseudo-F1,4 = 232
2.3, p>0.1, full details are reported in Appendix S1).
233
The chosen levels of CO2 concentration were capable of producing clear differences in 234
pH between treatments (Appendix S2; see also Queiros et al. 2015, Rastelli et al., 2015). On
235
the contrary, temperature, salinity and O2 values were maintained considerably constant and 236
comparable across all liners independently of the treatment (Appendix S2; see also Queiros et
237
al. 2015, Rastelli et al., 2015). Analogously, the sediment grain size was very similar (mean +/-
238
SE = 10.37 +/- 0.49 μm) among the five (one per experimental condition)all sampled liners
239
examined before the start of the experiment.
240 241
Macrofaunal abundance, diversity and community structure responses to increasing CO2 242
A total of 180 173 macrofaunal species or higher taxa (1026 Annelida, 34 27 Mollusca,
243
followed by 15 23 Arthropoda, 7 Sipuncula, 67 Echinodermata, 6 and Cnidaria, 1 Nemertea 244
and 1 Hemichordata - Appendix S3) were identified in the experiment and used for
245
PERMANOVA on the whole assemblage structure.
246
Macrofaunal assemblages changed between experimental conditions depending on
247
time, irrespectively of the square root or the presence/absence transformation (Table 1). At 2
248
weeks of exposure, pairwise tests indicated a significant difference between the control and all
249
treatments, but the 20000 ppm CO2 treatment. At 20 weeks, the only significant difference was 250
between the control and the highest CO2 treatment (Table 1 and Fig. 1 A, B). However, both 251
MDS ordination plots based on square root- and presence/absence-transformed data did not
252
show a clear separation between centroids corresponding to each treatment and exposure time, 253
while the dispersion of the points representing assemblages exposed to the 20 weeks exposure 254
was clearly larger than that of points corresponding to assemblages exposed to the shorter 255
exposure (Fig. 1 A and PERMDISP: F = 62.1, p = 0.001; Fig. 1 B and PERMDISP: F = 55.4, p 256
= 0.001).
257
Both total richness of taxa and total abundance of individuals differed among
258
treatments depending on the exposure time (Table 2). Specifically, 2 weeks after the start of the
259
experiment, the control hosted a larger number of taxa than all treatments, but the 20000 ppm,
260
Formattato: Pedice
Formattato: Non Evidenziato Formattato: Non Evidenziato Formattato: Non Evidenziato Formattato: Non Evidenziato Formattato: Non Evidenziato Formattato: Non Evidenziato
while no significant differences were displayed at 20 weeks of exposure (Table 2 and Fig. 2
261
A). The total number of individuals per m2 at 2 weeks was also higher in the control than in 262
any treatment, but significant differences were detected only relative to the 2000 and 5000 ppm
263
CO2 concentrations. Analogously to richness, all significant differences disappeared after 20 264
weeks exposure (Table 2 and Fig. 2 B).
265
The analysis performed on the abundance of the 8 most common macrofaunal taxa
266
were tested between the different experimental conditions over time. Specifically, after two
267
weeks of exposure, the abundance of the polychaete Heteromastus filiformis was higher, 268
although not significantly, higher in the control than in all CO2 treatments. At 20 weeks, this 269
species was less abundant in the control than in the two highest concentrations (Table 3 and
270
Fig. 3 A). Another polychaete, Prionospio cirrifera, was, on the other hand, significantly
271
moreless abundant in the control than in either the 20000 ppm treatment or all treatments after
272
two and twenty weeks of exposure, while no significant differences in the abundance of this 273
species occurred at twenty weeks of exposure respectively (Table 3 and Fig. 3 B).
274
Three taxa, namely the Nemertea, and the polychaetes Paradoneis eliasoni/lyra and
275
Paramphinome jeffreysii differed significantly between times of exposure, irrespectively of the 276
CO2 concentration (Table 3), with the first two taxaon being, on average, lessmore abundant 277
after 2 than after 20 weeks of exposure, and the other two species displaying the opposite
278
pattern (Fig. 3 C, D and E).
279
Of tThe remaining three taxa,(the polychaete Chaetozone sp. and the bivalves Thyasira 280
equalis and Adontorhina similis and the polychaete Chaetozone sp. were not significantly 281
affected by any CO2 treatment applied over any time (Table 3 and Fig. 3 F and. G and H), 282
while the bivalve Thyasira equalis was comparably abundant in each CO2 treatment at 2 283
weeks exposure and completely absent at 20 weeks exposure (Table 3 and Fig. 3 H).
284 285
DISCUSSION 286
The present study was designed to investigate the impact on the abundance and
287
diversity of benthic macrofaunal assemblages of exposure to a plume of CO2-enriched 288
seawater that could result from CO2 leakages from sub-seabed CCS. Results indicated that over 289
a short-term period (2 weeks), the macrofaunal assemblage structure was significantly affected
290
by all experimental levels of increased CO2, with the only exception of the highest 291
concentration. Conversely, after 20 weeks of exposure, the only significant difference was
292
between the control assemblages and those subject to the highest CO2 concentration. 293
Rapid impacts on macrofauna community structure, diversity and abundance following
294
short-term exposure to elevated CO2, similar to that seen in the current experiment, has been 295
reported from a number of previous mesocosm studies (e.g. Widdicombe et al., 2009,
296
Meadows et al., 2015). In this study, after 2 weeks of exposure, the control treatments hosted a
297
larger number of taxa than all treatments, but the 20000 ppm. Similarly, the total abundance
298
was larger in the control than in any treatment, but significant differences occurred only
299
relative to the 2000 and 5000 ppm CO2 concentrations. The apparent lack of impact in the 300
20,000 ppm treatments is perhaps surprising. However, one explanation could be that many of
301
the organisms in this treatment had actually died as a result of this extremely high CO2 302
exposure but their bodies had not had time to decay, especially if the microbial decomposition
303
was also inhibited by the low pH, and these organisms were then falsely counted as living in
304
the subsequent analysis. Another possibility is that under very extreme CO2 exposure 305
organisms go into a severe state of metabolic depression that maintains them for a limited
306
period of time before they inevitably die. The rapid response of all CO2 enriched treatments, 307
except the 20,000 ppm treatment discussed above, indicates that the most sensitive species
308
were likely affected negatively by even relatively low CO2 treatments injected just for 2 weeks. 309
Such conditions may have led to behavioural and/or metabolic changes, ultimately leading to
310
mortality and, consequently, to changes in the whole benthic community composition.
311
Notably, previous studies have also reported a rapid, negative impact on macrofaunal
312
diversity and structure from a controlled experimental release of CO2 from below the seafloor . 313
Even though this response only became evident five weeks after the start of the release it took
314
several weeks for the within sediment porewater pH to drop significantly. This was due to
315
natural chemical buffering processes which affected the carbonate dynamics (Lichtschlag et al.,
316
2014; Taylor et al., 2015; Widdicombe et al., 2015). In the present study, it seems that there
317
was less potential for the sediment buffer the changes in seawater chemistry and the impacts on
318
infauna occurred rapidly. This highlights the importance of understanding how the different
319
chemical and biological characteristics of different sediments will affect the speed of impact
320
following a CO2 leak. 321
The results observed after medium-term exposure (20 weeks) in the current study
322
would suggest that the only impacts of prolonged CO2 exposure were observed in the highest 323
treatment level (20,000 ppm). This is in contradiction to previous mesocosm studies that have
324
shown significant impacts of CO2 exposure to persist over many weeks (e.g. Widdicombe et al 325
2009). However, it should be noted that in the current study the similarity observed between
326
the control treatments and the majority of CO2 exposure treatments after 20 weeks was not due 327
to any recovery in the fauna of the CO2 treatments but due to a decrease in the abundance and 328
diversity of the fauna in the control treatments. So it could be hypothesized that the similarity
329
of assemblages exposed to almost all treatments after the twenty weeks exposure was due, at
330
least in part, to the negative effect of holding this particular fauna under mesocosm conditions.
331
Such negative impacts could result from limiting food availability (e.g. Guppy and Withers,
332
1999), which, once maintained over or occurred after a relatively long period, could have
333
exerted a negative influence on macrofaunal assemblages able to mask any concomitant effect
334
of CO2. Unfortunately, we do not have empirical data suitable to unambiguously support or 335
discard this hypothesis. Other stressful environmental variables, such as temperature,
336
desiccation, anoxia and hypersalinity, which are capable of inducing drastic reductions in
337
metabolic rates of almost all animal taxa (Guppy and Withers, 1999), could also have, in
338
principle, occurred in mesocosms and played a role in the present findings. The continuous
339
supply of mesocosms with new water from the adjacent fjord, however, suggests that reaching
340
drastically limiting conditions of such variables was also unlikely during the experiment. In
341
addition, the previous mesocosm experiments of Widdicombe et al (2009) used similar
342
conditions as used in the current study and saw no evidence of detrimental mesocosm impacts
343
during a 20 week experiment. It is most likely therefore that the sediment or community
344
selected for this experiment was less suitable for mesocosm experimentation than that which
345
was used in the previous studies. In the current experiment the sediment was collected from an
346
area twice as deep than the area used for collection of materials in Widdicombe et al (2009);
347
100m compared with 50m. Given that many potential CCS sites are located in deep water, it
348
may be that the value of mesocosm experiments may be limited to assessing short term
349
exposures and that there is a greater need for developing in-situ experimental procedures in
350
these areas.
351
Despite the issues associated with mesocosm effects, it was clear that over a longer
352
exposure, the communities in all the CO2 enriched treatments, except 20,000 ppm, converged 353
as the hardiest and most resistant species persisted (the number of taxa that were absent in all 354
treatments almost doubled from the 2 weeks to the 20 weeks time). What actually constitutes a
355
resistant species to elevated levels of CO2 will depend on the specific metabolic and 356
physiological adaptations of macrofaunal organisms (see Widdicombe and Spicer, 2008). This
357
resistance, therefore, is largely variable among taxa, both in terms of overall extent and
358
underlying mechanisms (Lessin et al., 2016). Echinoderms show very little compensation for
359
hypercapnia-related disturbance (Spicer et al., 1988, Spicer 1995, Kroeker et al., 2013).
360
Calcifying organisms need to increase pH (by active removal of H+ ions) in order to maintain 361
the formation of biogenic structures where needed (Widdicombe et al., 2015). Other organisms
362
prefer to suppress metabolism by shutting down various cellular processes (Guppy and
363
Withers, 1999, Widdicombe et al., 2009). In this study, the exposure to the most extreme CO2 -364
leakage scenario, in a medium-term (20 weeks), was tolerated by highly resistant taxa, such as
365
borrowing polychaete worms from the family Capitellidae (to which H. filiformis belongs). In
366
fact, this taxon is described as opportunistic able to dominate macrofaunal invertebrate
367
communities under perturbed conditions, likely due to its short generation time and direct
368
development which can allow their efficient use of the habitat (Pearson and Rosenberg, 1978;
369
Berge, 1990; Preckler, 2015) and increases in biomass after the elimination of more sensitive
370
species (Lessin et al., 2016). Analogously, the spionid polychaete P. cirrifera, which was here 371
also more abundant in the most acidified treatment compared to the control, especially after 372
short-term exposure, was described among the dominant species in polluted areas (Shen et al., 373
2010). This ability has been also explored for restoring polluted sediments by adding
374
bioturbating species of capitellid polychaetes (Chareonpanich et al. 1994, Ueda et al. 1994).
375
Therefore, the fact that we did not observe a reduced abundance of H. filiformis in the most
376
acidified treatments compared with the control is consistent with its known
disturbance-377
resistant trait as a capitellid polychaete. At the same time, this species was the most abundant
378
in the examined macrofaunal samples, hence its response to experimental treatments could
379
have driven, at least for a considerable part, that of the whole structure of assemblages. It is
380
worth noting, however, that the convergence between more acidified and control assemblages
381
could have been also modulated by the temporal variability (obviously due to processes other
382
than changes in CO2 inputs) of the latter ones, which can be as large as that driven by the 383
increased CO2 inputs (Widdicombe et al., 2015). In principle, temporal fluctuations in patterns 384
of abundance and diversity of meiofaunal macrofaunal assemblages in the control could have
385
made them similar, even just by chance, to the treated ones during the experiment. A large and
386
significant temporal variability, irrespectively of CO2 treatments, was confirmed by several 387
conspicuous taxa here examined.
388
Although the results of the current mesocosm study can provide crucial information on
389
actual cause-effect relationships between CO2 enrichment and macrofaunal responses, any 390
attempt to extrapolate them to predicting the ecological and biological consequences of
391
possible field leaks should be made with caution. The main reasons for this include: a) the
392
mesocosms being a confined system, which does not allow an organism to escape or relocate to
393
avoid unfavourable conditions, potentially leading to overestimate mortality rates over scales
394
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
396
the same as another one from a different location and/or time; c) possible different buffering
397
effects due to the different mineralogy of the sediments, with special focus on carbonate
398
content; d) the potential change between mesocosms and field biological responses due to the
399
drastic difference in the depth (hydrostatic pressure) to which the system is exposed; e) the
400
lower resilience or recovery of communities due to the lack of immigration of
401
specimens/species from surrounding non-impacted systems (Danovaro 2010; Widdicombe et
402
al., 2015). Moreover, present findings cannot obviously provide any unambiguous information
403
to derive expectations on possible responses, even of the same assemblages, to longer-term (>
404
20 weeks) exposure, or on responses to any exposure of assemblages dominated by other
405
groups of organisms, such as echinoderms (Spicer et al. 1988, Spicer 1995, Kroeker et al.,
406
2013) and more calcified taxa. For example, it is reasonable to assume that calcifying
407
organisms would be particularly sensitive to increases of CO2 and consequent reductions of pH, 408
which could eventually lead to critical loss of their fitness and survival rates through the
409
allocation of more energy to ion removal processes to detriment of other important
410
physiological processes (Pörtner, 2008; Wood et al., 2008). In this context, other studies
411
carried out in natural acidic shallow vents (e.g. Rodolfo-Metalpa et al. 2011; Gambi et al.,
412
2016; Kamenos et al., 2016) indicated that even CO2 increases considerably less than in the 413
present experiment can determine profound changes in exposed benthic assemblages. None of
414
such studies, however, are fully comparable to the present one. Specifically, Rodolfo-Metalpa
415
et al. (2011) focused on calcifying organisms, which, instead, were almost not represented in
416
present meiofaunal macrofaunal assemblages. Both Gambi et al. (2016) and Kamenos et al.
417
(2016) examined the distribution and diversity of benthic organisms (coralline algae and
418
polychaetes) at increasing distance and increasing pH from natural vents, thus along natural
419
gradients of CO2 concentrations to which such organisms were adapted for a much longer time 420
compared to the temporal scale of our experiment.
421 422
CONCLUSION 423
In spite of all the above listed factors and processes which are likely to jeopardise the
424
accuracy of extrapolations of mesocosm findings to real circumstances, the structure of the
425
present experiment was suitable to examine the relative responses of macrofaunal assemblages
426
and individual taxa to increased CO2 inputs in an unconfounded way. As such, present findings, 427
although not guaranteeing that field responses to possible future leakages associated to CCS
428
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 430
would occur rapidly over short time periods (<2 weeks).. Additional experiments are needed to
431
understand the mechanisms responsible for the present findings and their possible consistency
432
under field conditions, but this controlled study provides a relevant contribution to the debate
433
on the cost-benefit balance (environmental risk vs. intended goals) of CCS technologies.
434 435
ACKNOWLEDGEMENTS 436
437
This research was conducted as part of the European Community’s Seventh Framework
438
Programme FP7/2007-2013 for the project Sub-seabed CO2 storage: impact on marine
439
ecosystems (ECO2), grant agreements N. 265847, DEVelopment Of innovative Tools for 440
understanding marine biodiversity and assessing good Environmental Status (DEVOTES),
441
grant agreement no. 308392 and FME Success. TA was partially supported by Marie Curie
442
Actions through the project CEFMED (project number 327488). We are grateful to Oddbjorn
443
Petersen, Per Ivar Johannessen, Morten Schaanning personnel at the Marine Research Station
444
(Solbergstrand, Norway) of the Norwegian Institute of Water Research (NIVA, Oslo, Norway)
445
and at Plymouth Marine Laboratory for support and advice during the ECO2 mesocosm 446
experiments. Dr Mats Walday at NIVA, Dr Andrew Sweetman from Heriot Watt University of
447
Edinburgh and Dr Sarah Dashfield and Dr Carolyn Harris from Plymouth Marine Laboratory
448
are also thanked for support during the organization and analyses for this work.
449
450
REFERENCES 451
1. Anderson, M. J. A new method for non-parametric multivariate analysis of variance.
452
Austral. Ecol. 26, 32–46 (2001). 453
454
2. Anderson M.J. 2006. Distance-based tests for homogeneity of multivariate dispersions. 455
Biometrics 62: 245-253. 456
457
32. Anderson, M. J., Gorley, R. N. & Clarke, K. R. PERMANOVA + for PRIMER: Guide to
458
Software and Statistical Methods. PRIMER-E (Plymouth, 2008). 459
460
43. Berge, J.A. Macrofauna recolonization of subtidal sediments. Experimental studies on
461
defaunated sediment contaminated with crude oil in two Norwegian fjords with unequal
462
eutrophication status. I. Community responses. Mar. Ecol. Prog. Ser. 66, 103-115 (1990).
463 464
54. Blackford, J. C., Jones, N., Proctor, R., Holt, J., Widdicombe, S., Lowe, D. & Rees, A. An
465
initial assessment of the potential environmental impact of CO2 escape from marine carbon
466
capture and storage systems. Proc. Inst. Mech. Eng. Part A – J. Power Energy 223, 269–280
467
(2009).
468 469
Formattato: Tipo di carattere:
Times New Roman, 12 pt
Formattato: Tipo di carattere:
65. Blackford, J. C., Hattam, C., Widdicombe, S., Burnside, N., Naylor, M., Kirk, K., Maul, P.
470
& Wright, I. CO2 leakage from geological storage facilities: environmental, societal and 471
economic impacts, monitoring and research strategies. In Geological storage of carbon dioxide
472
(CO2): Geoscience, technologies, environmental aspects and legal frameworks (ed. Gluyas, J.
473
& Mathias, S.) 149-178 (Swaston, Cambridge, GB, Woodhead publishing, 2013).
474 475
76. Blackford, J. C., Stahl, H., Bull, J. M. , Bergès, B. J. P., Cevatoglu, M., Lichtschlag, A.,
476
Connelly, D., James, R. H. , Kita, J., Long, D., Naylor, M., Shitashima, K., Smith, D., Taylor,
477
P., Wright, I., Akhurst, M., Chen, B., Gernon, T. M., Hauton, C., Hayashi, M., Kaieda, H.,
478
Leighton, T. G., Sato, T., Sayer, M. D. J., Suzumura, M., Tait, K.,Vardy, M. E., White, P. R. &
479
Widdicombe, S. Detection and impacts of leakage from sub-seafloor deep geological carbon
480
dioxide storage. Nat. Clim. Change 4, 1011–1016 (2014).
481 482
87. Caldeira, K. & Wickett., M. Anthropogenic carbon and ocean pH. Nature 425:365 (2003).
483 484
98. Chareonpanich, C., Tsutsumi, H. & Montani, S. Efficiency of the decomposition of organic
485
matter, loaded on the sediment, as a result of the biological activity of Capitella sp. I.. Mar.
486
Pollut. Bull. 28, 314–318 (1994).
487 488
109. Cerrano, C., Cardini, U., Bianchelli, S., Corinaldesi, C., Pusceddu, R Danovaro. A. Red
489
coral extinction risk enhanced by ocean acidification. Scientific reports 3 (2013).
490 491
101. Clarke, K. R. & Green, R. H. Statistical design and analysis for a ‘ biological effects ’
492
study. Mar. Ecol. Prog. Ser. 46, 213 – 226 (1988).
493 494
112. Damen , K., Faaij, A. & Turkenburg, W. Health, Safety and Environmental Risks of
495
Underground Co2 Storage – Overview of Mechanisms and Current Knowledge. Climatic
496
Change 74(1), 289-318 (2006).
497 498
123. Danovaro, R. Methods for the study of deep-sea sediments, their functioning and
499
biodiversity. (ed. Taylor & Francis) 456 pp (CRC Press, London, 2010). 500
501
134. Fabiano, M. & Pusceddu, A. Total and hydrolizable particulate organic matter
502
(carbohydrates, proteins and lipids) at a coastal station in Terra Nova Bay (Ross Sea,
503
Antarctica). Polar Biol. 19, 125–132 (1998).
504 505
145. Feely, R. A., Sabine, C. L., Lee, K., Berelson ,W., Kleypas, J., Fabry V.J. & Millero, F. J.
506
Impact of anthropogenic CO2 on the CaCO3 system in the oceans. Science 305, 362–366.
507
(2004).
508 509
156. Gambi, M. C., Musco, L., Giangrande, A., Badalamenti, F., Micheli, F. & Kroeker, K. J.
510
Distribution and functional traits of polychaetes in a CO2 vent system: winners and losers
511
among closely related species. Mar. Ecol. Prog. Ser. 550, 121-134 (2016).
512 513
167. Gattuso, J. P., Magnan, A., Billé, R., Cheung, W.W., Howes, E. L., Joos, F., Allemand,
514
D., Bopp, L., Cooley, S. R., Eakin, C. M., Hoegh-Guldberg, O., Kelly, R. P., Pörtner, H. O.,
515
Rogers, A. D., Baxter, J. M., Laffoley, D., Osborn, D., Rankovic, A., Rochette, J., Sumaila, U.
516
R., Treyer, S., & Turley, C. Contrasting futures for ocean and society from different
517
anthropogenic CO2 emissions scenarios. Science 349, aac4722 (2015). 518
519
178. Gibbins, J., Haszeldine, S., Holloway, S., Pearce, J., Oakey, J., Shackley, S. & Turley, C.,
Scope for future CO2 emission reductions from electricity generation through the deployment 521
of carbon capture and storage technologies in Avoiding Dangerous Climate Change (ed.
522
Schnellhuber, H.J.) 379–384 (Cambridge University Press, 2006). 523
524
189. Guppy, M. & Withers, P. Metabolic depression in animals: physiological perspectives and
525
biochemical generalizations. Biol. Rev. Camb. Philos. Soc. 74, 1–40 (1999).
526 527
1920. Hale, R., Calosi, P., McNeill, L., Mieszkowska, N. & Widdicombe, S. Predicted levels
528
of future ocean acidification and temperature rise could alter community structure and
529
biodiversity in marine benthic communities. Oikos 120, 661–674 (2011).
530 531
201. Holloway, S. Carbon dioxide capture and geological storage. Philos. Trans. R. Soc. A
532
Math. Phys. Eng. Sci. 365 (1853), 1095–1107 (2007). 533
534
212. IPCC, 2013: Climate Change 2013: The Physical Science Basis. Contribution of Working
535
Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change 536
(eds. Stocker, T.F., et al. ). 1535 pp (Cambridge University Press, Cambridge, United Kingdom
537
and New York, NY, USA, 2013).
538 539
223. Jones, D. G., Beaubien, S. E., Blackford, J. C., Foekema, E. M., Lions, J., De Vittor, C.,
540
West, J.M., Widdicombe, S., Hauton, C & Queiros, A. Developments since 2005 in
541
understanding potential environmental impacts of CO2 leakage from geological storage. 542
International Journal of Greenhouse Gas Control 40, 350-377 (2015). 543
234. Kamenos, N. A., Perna, G., Gambi, M. C., Micheli, F. & Kroeker, K. J. Coralline algae in
544
a naturally acidified ecosystem persist by maintaining control of skeletal mineralogy and size.
545
Proc. Royal Soc. B 283, 20161159 (2016). 546
245. Kroeker, K., Kordas Rebecca Crim, R., Hendriks, I., Ramajo, L., Singh, G., Duarte, C. &
547
Gattuso, J.-P. Impacts of ocean acidification on marine organisms: quantifying sensitivities and
548
interaction with warming. Glob. Change Biol. 19, 1884–1896 (2013).
549 550
256. Laverock, B., Kitidis, V; Tait, K; Gilbert, JA; Osborn, M. & Widdicombe, S. Bioturbation
551
activity determines the response of coastal benthic nitrogen-cycling microorganisms to ocean
552
acidification. Phil. Trans. of the Royal Soc. of London Series B.-Biol. Sci. 368 (1627), 1-13
553
(2013).
554 555
267. Lessin, G., Artioli, Y., Queiros, A., Widdicombe, S. & Blackford, J.C. Modelling impacts
556
and recovery in benthic communities exposed to localized high CO2. Mar. Poll. Bull. (in press)
557
(2016).
558 559
278. Lichtschlag, A., James, R. H., Stahl, H., & Connelly, D. Effect of a controlled subseabed
560
release of CO2 on the biogeochemistry of shallow marine sediments, their pore waters, and the
561
overlying water. Int. J. Greenh. G. Con. 38, 80-92 (2014).
562 563
289. McCave, I. N. Evaluation of a laser-diffraction-size analyzer for use with natural
564
sediments. J. Sedim. Res. 56, 561-564 (2013).
565 566
2930. McCormack, C.G. Key impacts of climate engineering on biodiversity and ecosystems,
567
with priorities for future research. Journal of Int. Env. Sci. (in press). (2016).
568 569
Formattato: Inglese (Regno Unito) Formattato: Inglese (Regno Unito)
301. Meadows, A.S., Ingels, J., Widdicombe, S., Hale, R. & Rundle, S. D. Effects of elevated
570
CO2 and temperature on an intertidal meiobenthic community. Journal of Exp. Mar. Biol. and 571
Ecol. 469, 44-56 (2015). 572
573
313. Mora, C., Wei, C.-L., Rollo, A., Amaro, T., Baco, A.R. Billett, D. Bopp, L., Chen, Q.,
574
Collier, M., Danovaro, R., Gooday, A.J., Grupe, B.M., Halloran, P.R., Ingels, J., Jones,
575
D.O.B., Levin, L.A., Nakano, H., Norling, K., Ramirez-Llodra, E., Rex, M., Ruhl, H.A., Smith,
576
C.R., Sweetman, A.K., Thurber, A.R., Tjiputra, J.F., Usseglio, P., Watling, L., Wu, T. &
577
Yasuhara, M. Biotic and Human Vulnerability to Projected Changes in Ocean Biogeochemistry
578
over the 21st Century. PLoS Biol 11(10), e1001682 (2013).
579 580
323. Paulley, A., Maul, P.R. & Metcalfe, R. Research into impacts and safety in CO2 storage:
581
Scenarios for Potential Impacts from Hypothetical Leakage from Geological Storage Facilities
582
for Carbon Dioxide in RISCS Deliverable D1.3 (ed. Quintessa Limited) 60 pp (2012).
583 584
343. Pearson, T. H. & Rosenberg, R. Macrobenthic succession in relation to organic
585
enrichment and pollution of the marine environment. Oceanogr. Mar. Biol. Ann. Rev. 16, 229–
586
311 (1978).
587 588
345. Pörtner, H-O. Ecosystem effects of ocean acidification in times of ocean warming: a
589
physiologist’s view. Mar. Ecol. Prog. Ser. 373, 203–217 (2008).
590 591
356. Preckler, C.A. Biodiversity and chemical interactions in Antartic benthic communities of
592
Deception Island (South Shetland Islands). PhD thesis. 314 pp (Barcelona university, 2015). 593
594
367. Pusceddu, A., Dell’Anno, A., Fabiano, M. & Danovaro, R. Quantity and bioavailability of
595
sediment organic matter as signatures of benthic trophic status. Mar. Ecol. Progr. Ser. 375,
596
41-52 (2009).
597 598
378. Queiros, A. M.,Taylor P., Cowles A., Reynolds A., Widdicombe S. & Stahl H. Optical
599
assessment of impact and recovery of sedimentary pH profiles in ocean acidification and
600
carbon capture and storage research. Int. J. Greenh. G. Con. 38, 110-120 (2014).
601 602
389. Queiros, A.M. Norling, K., Amaro, T., Nunes, J., Cummings, D., Yakushev, E., Sorensen,
603
K., Harris, C., Woodward, M., Danovaro, R., Rastelli, E., Alve, E., DeVittor, C., Karuza, A.,
604
Cibic, T., Monti, M., Ingrosso, G., Fornasaro, D., Beaubien, S. E., Guilini, K., Vanreusel, A.,
605
Molari, M., Boetius, A., Ramette, A., Wenzhöfer, F., deBeer, D.,Weber, M., Grünke, S.,
606
Bigalke, N. & Widdicombe, S. Potential Impact of CCS Leakage on Marine Communities. (Ed.
607
Plymouth Marine Laboratory) 101 pp. (Plymouth, UK, 2015).
608 609
3940. Rastelli, E., Corinaldesi, C., Dell’Anno, A., Amaro, T., Queiros, A. M., Widdicombe, S.
610
& Danovaro, R. Impact of CO2 leakage from sub-seabed carbon dioxide capture and storage 611
(CCS) reservoirs on benthic virus-prokaryote interactions and functions. Fronteirs in
612
Microbiology 6, 935. doi:10.3389/fmicb.2015.00935 (2015). 613
614
401. Rodolfo-Metalpa, R., Houlbrèque, F., Tambutté, É., Boisson, F., Baggini, C., Patti, F. P.,
615
Jeffree, R., Fine, M., Foggo, A., Gattuso, J.-P & Hall-Spencer, J. M. Coral and mollusc
616
resistance to ocean acidification adversely affected by warming. Nat. Clim. Change 1, 308-312
617
(2011).
618 619
41. Shen, P.-P, Zhou, H. & Gu, J.-D. Patterns of polychaete communities in relation to 620
environmental perturbations in a subtropical wetland of Hong Kong. Journal of the 621
Marine Biological Association of the United Kingdom 90, 923-932 (2010). 622
623
42. Spicer, J. I, Taylor, A. C. & Hill, A. D. Acid–base status in the sea urchins Psammechinus
624
miliaris and Echinus esculentus (Echinodermata: Echinoidea) during emersion. Mar. Biol. 99,
625
527–534 (1988).
626 627
43. Spicer, J. I. Oxygen and acid–base status of the sea urchin Psammechinus miliaris during
628
environmental hypoxia. Mar. Biol. 124, 71–76 (1995).
629 630
44. Tait, K., Stahl, H., Taylor, P., & Widdicombe, S. Rapid response of the active microbial
631
community to CO2 exposure from a controlled sub-seabed CO2 leak in Ardmucknish Bay
632
(Oban, Scotland). Int. J. Greenh. G. Con. 38, 171-181 (2014).
633 634
45. Tait, K., Beesley, A., Findlay, H.S., McNeill, C.L. & Widdicombe, S. Elevated CO2 635
induces a bloom of microphytobenthos within a shell gravel mesocosm. FEMS
636
10.1093/femsec/fiv092 (2015).
637 638
46. Taylor, P., Stahl, H., Vardy, M. E., Bull, J. M., Akhurst, M., Hauton, C., James, R.H.,
639
Lichtschlag, A., Long, D., Aleynik, D., Toberman, M., Naylor, M., Connelly, D., Smith, D.,
640
Sayer, M.D.J., Widdicombe, S., Wright, I.C. & Blackford, J. A novel sub-seabed CO2 release 641
experiment in forming monitoring and impact assessment for geological carbon storage. Int. J.
642
Greenh. G. Con. 38, 3–17. (2015). 643
644
47. Ueda, N., Tsutsumi, H., Yamada, M., Takeuchi, R. & Kido, K. Recovery of the marine
645
bottom environment of a Japanese bay. Mar. Pollut. Bull. 28, 676–682 (1994).
646 647
48. Widdicombe, S. & Needham, H. R. Impact of CO2 induced seawater acidification on the 648
burrowing activity of Nereis virens and sediment nutrient flux. Mar. Ecol. Prog. Ser. 341,
111-649
122 (2007).
650 651
49. Widdicombe, S. Dashfield, S. L., McNeill, C. L., Needham, H. R., Beesley, A., McEvoy,
652
A., Øxnevad, S., Clarke, K. R. & Berge, J. A. Effects of CO2 induced seawater acidification on 653
in faunal diversity and sediment nutriente fluxes. Mar. Ecol. Prog. Ser. 379, 59–75 (2009).
654 655
50. Widdicombe, S., Blackford, J. C. & Spicer, J. I. Assessing the environmental consequences
656
of CO2 leakage from geological CCS: generating evidence to support environmental risk
657
assessment. Mar. Pollut. Bull.73, 399–401 (2013).
658 659
51. Widdicombe, S., McNeill, C. L., Stahl, H., Taylor, P., Queiros, A. M., Nunes, J., & Tait, K.
660
Impact of sub-seabed CO2 leakage on macrobenthic community structure and diversity. Int. J. 661
Greenh. G. Con. 38, 182-192 (2015). 662
663
52. Widdicombe, S., Spicer, J. I. Predicting the impact of ocean acidification on benthic
664
biodiversity: What can animal physiology tell us? J Exp Mar Biol Ecol 366(1-2),
187-665
197(2008).
666 667
53. Wood, H. L., Spicer, J. I. & Widdicombe, S. Ocean acidification may increase calcification
668
rates—but at a cost. Proc. R. Soc. B. 275, 1767–1773 (2008).