Gains and losses of coral skeletal porosity changes with ocean acidification acclimation

Abstract

Ocean acidification is predicted to impact ecosystems reliant on calcifying organisms, potentially reducing the socioeconomic benefits these habitats provide. Here we investigate the acclimation potential of stony corals living along a pH gradient caused by a Mediterranean CO₂ vent that serves as a natural long-term experimental setting. We show that in response to reduced skeletal mineralization at lower pH, corals increase their skeletal macroporosity (features >10 μm) in order to maintain constant linear extension rate, an important criterion for reproductive output. At the nanoscale, the coral skeleton's structural features are not altered. However, higher skeletal porosity, and reduced bulk density and stiffness may contribute to reduce population density and increase damage susceptibility under low pH conditions. Based on these observations, the almost universally employed measure of coral biomineralization, the rate of linear extension, might not be a reliable metric for assessing coral health and resilience in a warming and acidifying ocean.

ARTICLE

Received 30 Dec 2014 | Accepted 10 Jun 2015 | Published 17 Jul 2015

DOI: 10.1038/ncomms8785 OPEN

Gains and losses of coral skeletal porosity changes with ocean acidication acclimation

Paola Fantazzini1,2, Stefano Mengoli3, Luca Pasquini1, Villiam Bortolotti4, Leonardo Brizi1,2, Manuel Mariani1,2, Matteo Di Giosia5, Simona Fermani5, Bruno Capaccioni6, Erik Caroselli7, Fiorella Prada7, Francesco Zaccanti7, Oren Levy8, Zvy Dubinsky8, Jaap A. Kaandorp9, Pirom Konglerd9, Jrg U. Hammel10, Yannicke Dauphin11, Jean-Pierre Cuif11, James C. Weaver12, Katharina E. Fabricius13, Wolfgang Wagermaier14, Peter Fratzl14, Giuseppe Falini5 & Stefano Goffredo7

Ocean acidication is predicted to impact ecosystems reliant on calcifying organisms, potentially reducing the socioeconomic benets these habitats provide. Here we investigate the acclimation potential of stony corals living along a pH gradient caused by a Mediterranean CO2 vent that serves as a natural long-term experimental setting. We show that in response to reduced skeletal mineralization at lower pH, corals increase their skeletal macroporosity (features 410 mm) in order to maintain constant linear extension rate, an important criterion for reproductive output. At the nanoscale, the coral skeletons structural features are not altered. However, higher skeletal porosity, and reduced bulk density and stiffness may contribute to reduce population density and increase damage susceptibility under low pH conditions. Based on these observations, the almost universally employed measure of coral biomineralization, the rate of linear extension, might not be a reliable metric for assessing coral health and resilience in a warming and acidifying ocean.

1 Department of Physics and Astronomy, University of Bologna, Viale Berti Pichat 6/2, 40127 Bologna, Italy. 2 Centro Enrico Fermi, Piazza del Viminale 1, 00184 Rome, Italy. 3 Department of Management, University of Bologna, Via Capo di Lucca 34, 40126 Bologna, Italy. 4 Department of Civil, Chemical, Environmental, and Materials Engineering, University of Bologna, Via Terracini 28, 40131 Bologna, Italy. 5 Department of Chemistry G. Ciamician, University of Bologna, Via F. Selmi 2, 40126 Bologna, Italy. 6 Department of Biological, Geological and Environmental Sciences, Section of Geology, University of Bologna, Piazza di Porta S. Donato 1, 40126 Bologna, Italy. 7 Marine Science Group, Department of Biological, Geological and Environmental Sciences, Section of Biology, University of Bologna, Via F. Selmi 3, 40126 Bologna, Italy. 8 The Mina and Everard Goodman Faculty of Life Sciences, Bar-Ilan University, Ramat-Gan 52900, Israel. 9 Section Computational Science, Faculty of Science, University of Amsterdam, Science Park 904, room C3.147, 1090 GE Amsterdam, The Netherlands. 10 Institute of Materials Research, Helmholtz-Zentrum Geesthacht, Outstation at DESY, Building 25c Notkestr. 85, D-22607 Hamburg, Germany.

11 Micropalontologie, UFR TEB Universit P. & M. Curie, 75252 Paris, France. 12 Wyss Institute for Biologically Inspired Engineering at Harvard University,60 Oxford Street, Cambridge, Massachusetts 02138, USA. 13 Australian Institute of Marine Science, PMB 3, Townsville, 4810 Queensland, Australia.

14 Department of Biomaterials, Max Planck Institute of Colloids and Interfaces, 14424 Potsdam, Germany. Correspondence and requests for materials should be addressed to P.F. (email: mailto:[email protected]

Web End [email protected] ) or to G.F. (email: mailto:[email protected]

Web End [email protected] ) or to S.G. (email: mailto:[email protected]

Web End [email protected] ).

NATURE COMMUNICATIONS | 6:7785 | DOI: 10.1038/ncomms8785 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 1

& 2015 Macmillan Publishers Limited. All rights reserved.

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8785

Climate change is among the biggest threats to the health of marine ecosystems. Rising atmospheric carbon dioxide partial pressure (pCO2)1 increases ocean surface pCO2

due to CO2 diffusion across the air-water interface, leading to ocean acidication (OA)2. Since global warming and OA are coupled and are predicted to act synergistically3,4, the future health of marine ecosystems and their corresponding long-term economic impacts on human coastal populations remain uncertain57. It is therefore of great interest to understand how increasing atmospheric CO2 concentrations will affect these marine habitats and the species that inhabit them.

Since the early 1800s, ocean pH has decreased from 8.2 by ca. 0.1 U8 and, if CO2 emissions continue at their current rates, the average sea surface pH is predicted to drop to 7.8 by the year 2100 (ref. 1). The Mediterranean Sea, with its closed circulation patterns and limited water exchange with the adjacent Atlantic Ocean, has already undergone a larger decrease in surface pH compared with the global average9, making it an ideal site for OA studies1013.

Near Panarea Island, off the southwestern coast of Italy, lies a series of active volcanic vents in the seabed releasing CO2 emissions that acidify the surrounding seawater, making this location an ideal natural laboratory for OA studies. The underwater CO2 emissions generate a stable pH gradient with levels matching several Intergovernmental Panel on Climate Change (IPCC) sea surface pH predictions associated with different atmospheric CO2 emission scenarios for the end of the century1.

The present study investigates the effects of environmental pH on skeletal structures and growth at multiple length scales in the

solitary scleractinian coral Balanophyllia europaea living along the pH gradient. We studied 74 corals of similar age (mean age of 12 years) that had spent their lives at the CO2-pH gradient. Using a combination of scanning electron microscopy (SEM), atomic force microscopy (AFM), small-angle X-ray scattering (SAXS), micro computed tomography (mCT), nanoindentation, hydrostatic weight measurement and time-domain nuclear magnetic resonance (TD-NMR), we document the skeletal mass, bulk volume, pore volume, porosity, biomineral density, bulk density, hardness, stiffness (ratio between elastic stress and strain), biometry data, net calcication rate and linear extension rate for each coral. Weight measurements combined with TD-NMR data represent a non-destructive technique for quantifying skeletal porosity over length scales spanning from tens of nanometres to tens of micrometres14,15, whereas mCT analysis permits a detailed large-scale quantitative 3D analysis of skeletal architecture, including the interseptal regions.

We show that in response to depressed calcication at lower pH, corals increase their skeletal porosity maintaining constant linear extension rate, which is important for reaching critical size at sexual maturity. However, higher skeletal porosity and reduced bulk density and stiffness may contribute to reduced mechanical strength, increasing damage susceptibility, which could result in increased mortality in an acidic environment.

ResultsStudy site and seawater carbonate chemistry. The four sites around the main vent are reported in Fig. 1. Site 1 (S1) has an

a

c

pH (total scale)

9

7

5

3

1

9

7

5

3

1

Italy

Panarea

arag

Km

1

d

0

38 39

38 38

pH TS 7.4 arag 1.4

pH TS 7.7 arag 2.3

pH TS 7.9 arag 2.5

0 2

2

4

pH TS 8.1 arag 3.4

Panarea

Basiluzzo

Depth (m)

6

8

Elevation line

15 03 15 04 15 05 15 06

Panarelli 40 30

10

3 2 1

(site number)

b

1 km

12

4

0 5 10 15 20 25 30 36

Distance from crater centre (m)

Lisca Bianca

e

f

Bottaro

Dattilo

10 20

Lisca Nera

CO2vent

Expanded Contracted

Figure 1 | Study location. Located off the southwestern coast of Italy (a), near Panarea Island (b), there are underwater volcanic vents releasing persistent gaseous emissions (9899% CO2 without instrumentally detectable toxic compounds), resulting in a stable pH gradient. Four sites at various distances from the primary vent were initially selected for study. No temperature difference exists among the four study sites throughout the year. (c) Ranges of measured pHTS (number of observations [n] 103110 per site) and Oarag (n 96104 per site) values at the four sites, showing consistent increases in

both pHTS and Oarag from the vent crater to its periphery. (d) Bathymetric prole of the sites with associated mean pHTS and Oarag values. No corals were found at Site 4, characterized by the lowest pH (mean pHTS 7.4). Living specimens of Balanophyllia europaea, photographed at night with expanded tentacles (e) and during the day with contracted tentacles (f); marker 5 mm.

2 NATURE COMMUNICATIONS | 6:7785 | DOI: 10.1038/ncomms8785 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

& 2015 Macmillan Publishers Limited. All rights reserved.

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8785 ARTICLE

average total scale pH (pHTS) of 8.1, equivalent to the average surface pH of modern oceans. S2s average pHTS of 7.9 aligns

with IPCCs predictions of a conservative CO2 emissions scenario (Representative Concentration Pathway (RCP6.0)), and the average pHTS of 7.7 for S3 ts the predictions of the business-as-usual CO2 emissions scenario (RCP8.5). Since no corals were found at S4 (within the vent crater, pHTS 7.4), only S13, which had growing coral populations and pHTS values ranging from 8.1 to 7.7, were included in the present study. Of the measured parameters at the three sites along the pH gradient (pHTS, total alkalinity, temperature and salinity), only pHTS differed signicantly across sites (N 103110; Po0.001,

KruskalWallis w2-test). The changing pH was accompanied by signicant shifts in carbonatebicarbonate equilibria, with aragonite saturation (Oarag) decreasing by 430% from S1 to S3 (Fig. 1, Supplementary Table 1, Supplementary Note 1).

Multi-scale analysis of coral skeletal properties. Combined results of SEM, mCT, AFM and TD-NMR skeletal analyses of corals growing at each study site revealed the detailed, multi-scale structural organization of the skeletons of B. europaea (Fig. 2, Supplementary Figs 13).

At the macroscale (relating to feature sizes 410 mm), linear extension rate (averaging ca. 1 mm per year) did not vary among sites (Supplementary Table 2), whereas net calcication rate (N 44; Po0.01, robust t-statistics test) and skeletal bulk

density (N 44, Po0.001, robust t-statistics test) signicantly

declined (Fig. 3, Supplementary Table 3) and skeletal porosity and macroscale porosity increased (N 44; Po0.001, robust

t-statistics test, Fig. 3, Supplementary Fig. 1, and Supplementary Table 3). The differences between S1 and S3 were ca. 18%

for net calcication rate, ca. 7% for bulk density, ca. 21% for

porosity and ca. 30% for macroscale porosity (Supplementary

Table 2). The corallite interseptal volume fraction from mCT (Supplementary Fig. 4) showed a difference among sites (N 30;

Po0.05, F and KruskalWallis w2-tests, Supplementary Table 4) but no signicant dependence on pH (Supplementary Fig. 5). The biometric data for the corallites (Supplementary Table 5) did not vary among sites.

At the micro/macroscale, skeletal stiffness signicantly declined with decreasing pH (n 122; Po0.001, robust

t-statistics test, Fig. 3, Supplementary Fig. 6, Supplementary Tables 4 and 6), whereas skeletal microscale porosity did not vary among sites (Supplementary Table 2 and Supplementary Fig. 1).

At smaller length scales (at the micro and nanoscales), SEM and AFM showed that the organization of the aragonite bre bundles (Fig. 2c,g,k) and the morphology and packing of the constituent mineral grains (Fig. 2d,h,l) appeared similar among corals from the three different sites, conrming that the basic biomineralization products were not affected by reduced pH13 (Fig. 2 and Supplementary Fig. 2). Skeletal biomineral hardness did not change among sites (Supplementary Fig. 7, Supplementary Table 4). Also, skeletal biomineral density values were similar across sites and consistent with those measured in previous studies16 (Supplementary Table 2). Similar results were obtained from both SAXS and TD-NMR analyses, which revealed that nanoscale porosity did not change signicantly with changing pH (Supplementary Figs 1e and 810 and Supplementary Tables 2 and 4).

The principal component analysis identied three major components: growth, skeletal porosity and biomineral density (Supplementary Tables 7 and 8). Only the rst two components depended signicantly on pH (Supplementary Table 9).

Macroscale Microscale Nanoscale

a b c d

Site 1 (pH 8.1)

Site 2 (pH 7.9)

Site 3 (pH 7.7)

e

f

g

h

i j k l

Figure 2 | Skeletal morphology of Balanophyllia europaea growing under different pH conditions from the macroscale to the nanoscale. Each row in the gure corresponds to a different study site and sample age is 911 years. Images are representative of all observed skeletons. (a,e,i) Low magnication SEM images of coral skeletons, marker 5 mm. (b,f,j) Internal sections of corallites from mCT images, marker 5 mm. (c,g,k) SEM images of entire skeletal bres from fractured septae, marker 10 mm. (d,h,l) AFM images of mineral grains on the skeletal bre surfaces, marker 50 nm.

NATURE COMMUNICATIONS | 6:7785 | DOI: 10.1038/ncomms8785 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 3

& 2015 Macmillan Publishers Limited. All rights reserved.

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8785

In summary, skeletal nano and microstructural features, linear extension rate, interseptal volume fraction and corallite biometry of B. europaea did not change signicantly with decreasing pH, despite a clear reduction in net calcication rates. This reduction in net calcication rate was accompanied by an increase in skeletal porosity (Fig. 3f, N 44; Po0.001, robust t-statistics test)

and a consequent decrease in skeletal bulk density and stiffness.

DiscussionResults of the present study complement previous research onB. europaea at this same vent site, which revealed no changes in skeletal calcium carbonate polymorph, organic matrix content, aragonite bre thickness and skeletal hardness in corals growing along the pH gradient13. There was, however, a signicant reduction in population density along the pH gradient, decreasing by a factor of 3 with increasing proximity to the vent crater centre (that is, from S1 to S3)13.

Figure 4 summarizes these results at the ocean, population, macro, micro and nanoscales for B. europaea. At the macroscale, increasing acidity was associated with a reduction in net calcication rate and a parallel increase in skeletal porosity, coupled with a decrease in skeletal bulk density. Linear extension rate and corallite shape (biometry and interseptal volume fraction) did not depend on pH, probably as a result of the compensation of reduced net calcication rate by increased skeletal porosity. At the micro/macroscale, the declining skeletal stiffness with decreasing pH could be coupled to an increased volume fraction of pores having a size comparable to the indentation area (that is, at the border between the micro and macroscales). At the nanoscale, porosity, biomineral hardness and density were not signicantly affected by pH. These results, bolstered by qualitative SEM and AFM analyses, suggest that the building blocks produced by the biomineralization process are substantially unaffected, but the increase in skeletal porosity is both a gain and a loss for the coral. In fact, in an acidic

a b

3.0

2.5

2.0

1.5

1.0

45

40

75%ile 75%ile

75%ile

75%ile

Net calcification rate

(mg mm2 per year)

*

**

Mean Mean

Mean

Mean

25%ile 25%ile

Bulk density (mg mm3 )

2.2

2.1

2.0

1.9

1.8

1.7

1.6

1.5

1.4

***

***

7.7 7.8 7.9 8.0 8.1 7.7 7.8 7.9 8.0 8.1

7.7 7.8 7.9 8.0 8.1

pHTS

pHTS

pHTS

pHTS

c

d

40

TD-NMR macroscale porosity (%)

***

***

***

35

Porosity (%)

30

35

***

75%ile

25

15

Mean

30

**

25%ile

25%ile

20

20

25%ile

25

7.7 7.8 7.9 8.0 8.1

8.1

pHTS

e f

Stiffness (GPa)

90 50

80

70

60

50

***

Porosity (%)

45

40

35

30

25

***

***

7.7 7.8 7.9 8.0 1.0 1.5 2.0 2.5 3.0

Net calcification rate (mg mm2 per year)

Figure 3 | Scatterplots of skeletal parameters, and correlation analysis between porosity and net calcication rate. Site 1 blue, Site 2 green, Site

3 red. Straight lines represent the best-t linear regression (mean, solid line), 25% quantile and 75% quantile (dashed lines). (ae) Skeletal parameters

(y-axes) plotted against pHTS. (f) Scatterplot of porosity (PA) versus net calcication rate in corals from Sites 1 to 3. For ad,f N 44;

for e n 122. ***Po0.001; **Po0.01; *Po0.05, robust t-statistics test.

4 NATURE COMMUNICATIONS | 6:7785 | DOI: 10.1038/ncomms8785 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

& 2015 Macmillan Publishers Limited. All rights reserved.

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8785 ARTICLE

pH

Ocean scale

(km)

***

Population scale(m)

Population density

NS NS

NS

Corallite biometry

** *** *** ***

***

Net calcification rate

Linear extension rate

Interseptal volume fraction (CT)

Bulk density (weight)

Porosity (weight)

Macro porosity (TD-NMR)

Macroscale (10 m 20 mm)

Skeletal stiffness

NS

Micro porosity (TD-NMR)

Micro structure (SEM)

Microscale (100 nm 10 m)

Nanoscale (<100 nm )

NS

NS

NS NS

Porosity (SAXS)

Nano porosity (TD-NMR)

Biomineral hardness

Biomineral density (weight)

Nano structure (AFM)

Figure 4 | Summary of responses in Balanophyllia europaea skeletal parameters as a function of pH from the ocean to the nanoscale. The signicant decline in population density with pH was measured in previously published research13. Net calcication rate and bulk density decrease with decreasing pH, whereas porosity (PA) increases, preserving the linear extension rate and corallite shape (biometry and interseptal volume fraction as measured by mCT do not correlate with pH). The increase in porosity is associated with a decrease in skeletal stiffness. At the nanoscale, the building blocks (the fundamental structural components of the coral skeleton, aragonite bre bundles and their constituent mineral grains) produced by the biomineralization process are substantially unaffected by increased acidity. Green boxes denote parameters found to have a direct relationship with pH, blue boxes denote parameters that have an inverse relationship and red boxes denote parameters found to have no relationship with pH. The signicances of the regression of each parameter (dependent variable) with pH (independent variable) are indicated; ***Po0.001; **Po0.01; NS indicates no signicance. Micro and nanoscale structure observations by SEM and AFM represent qualitative data.

environment, where the net calcication is depressed, enhanced macroporosity keeps linear extension rate constant, potentially meeting functional reproductive needs (for example, the ability to reach critical size at sexual maturity); however, it also reduces the mechanical strength of the skeletons, increasing damage susceptibility, which could result in increased mortality and the observed population density decline13.

While the results reported here for B. europaea may not be representative of the generalized response of all coral species to OA, they are consistent with eld observations made on other reef-building scleractinians. For example, while maintaining constant skeletal linear extension, decreased rates of calcication and losses in bulk skeletal density as a function of reduced aragonite saturation have been observed in Montasraea annularis17 and Porites astreoides18. While low aragonite saturation as a sole driver for the observed reduction in coral calcication has been discussed19, our conclusions regarding a balance between reduced net calcication rate and increased macroporosity to maintain constant linear extension can explain the outcomes of those studies17,18. In fact bulk density depends both on biomineral density and porosity. Our multi-scale analysis shows that all the skeletal features at the nano and microscales,

including biomineral density, are unchanged. The decrease of bulk density with decreasing pH depends on the increase of macroporosity, leaving the linear extension rate constant.

Our ndings, together with the well-described detrimental effects of heat stress on the scleractinian zooxanthellate coralB. europaea16,2023, provide several independent and consistent clues regarding the sensitivity of this species to global climate change predicted for the coming decades. Together with results from previous studies24, we demonstrate that the almost universally employed measure of coral biomineralization, namely the rate of linear extension, might not be a reliable metric for assessing coral health and resilience in a warming and acidifying ocean. Indeed, although the corals ability to maintain linear extension rate and gross skeletal morphology under conditions of decreasing oceanic pH could allow it to reach sexual maturity, it could reduce skeletal resistance to environmental challenges, affecting the long-term survivability of the species.

Methods

Study site. Off the southwestern coast of Italy, near the island of Panarea, an area delimited by the islets of Dattilo, Bottaro, Lisca Nera and Lisca Bianca (Fig. 1) is characterized by underwater volcanic CO2 vents. The main vent, a crater measuring 20 m 14 m and B10-m deep, generates a persistent column of CO2 bubbles that

rise from the seabed to the sea surface. In this hydrothermally stable setting with ambient temperature, CO2 emissions establish a pH gradient that extends B34 m from the centre of this crater to its periphery13. Three sites along this pH gradient were chosen for study. Distances (d) from the main crater centre and mean pHTS

values of the three sites from which corals were collected are: S1 (the control site), d 34 m, pHTS 8.07; S2, d 13 m, pHTS 7.87; S3, d 9 m, pHTS 7.74. Water

depth varies from 11.6 m within the crater to 8.8 m at the crater edge (S2) to 9.2 m at the outer margins (S1), where the local pH matches that of the surrounding seawater. The study site has stable hydrothermalchemical properties25.

The corals. B. europaea (Fig. 1) is a solitary zooxanthellate scleractinian coral endemic to the Mediterranean Sea found at depths ranging from 0 to 50 m26. Specimens of B. europaea were randomly collected by SCUBA diving at the three study sites along the pH gradient (26 from S1, 26 from S2 and 22 from S3) between November 2010 and May 2013. This sample size was chosen to limit damage on the natural population, which signicantly diminishes in the most acidic sites13, and was considered suitable for properly describing the skeletal properties, as shown in previous studies13,14. Biometric data were recorded for the specimens (that is, width-to-length, width-to-height and height-to-length ratios, Supplementary Table 5). Nanoindentation, hydrostatic weight measurement, SEM, AFM, SAXS, mCT and

TD-NMR analyses were performed on a subsample of specimens from each site.

Coral sample cleaning. Coral skeletons were submerged in 1% solution of sodium hypochlorite for 3 days to dissolve polyp tissue. After washing with deionized water and drying at 50 C for 34 days, each coral was examined under a binocular microscope to remove fragments of sediment, rock and encrusting organisms16.

Weight measurements. The buoyant method, usually applied for corals2729, was used to measure the total volume occupied by the coral skeleton (called bulk volume, VB), the bulk density (db, ratio of the mass to VB), the biomineral density (dr, ratio of the biomineral mass to biomineral volume, excluding pore volume connected to the external surface, also called real density or micro-density) and the apparent porosity (or effective or connected porosity, PA)30 (ratio of the pore volume connected to the external surface (VA) to VB). This method is based on the Archimedean principle applied to a specimen after full saturation with the same uid in which it is submerged (water in our case). It is worth to underline that the pore volume (VA) measured by the buoyant method is only the volume of the pores that can be saturated with water, that is, connected with the external surface. Pores inside the biomineral that are not connected to the external surface (occluded pores) are not measured. An estimate of the occluded porosity gave a negligible maximum value of B3% (in porosity units)

(Supplementary Note 2), which was homogeneous among sites. For coherence with coral literature, for the apparent porosity we use the term porosity.

The skeletons were weighed with a precision balance to determine the dry mass (m) and then placed inside a dryer chamber and evacuated with a rotary mechanical pump down to a vacuum of 10 2 mbar. After 6 hours, water was gently introduced to fully saturate the samples. The pump was switched off,the chamber was vented to the ambient atmosphere, and the masses of the fully water-saturated samples (ms) determined. With a hydrostatic balance, the masses of saturated samples fully immersed in water (mh) were determined. The skeletal parameters were calculated (Supplementary Table 2) by means of the following operational denitions: rw water density, VA (ms m)/rw; VB (ms mh)/rw;

PA VA/VB (ms m)/(ms mh); db m/VB; dr m/(VB VA).

NATURE COMMUNICATIONS | 6:7785 | DOI: 10.1038/ncomms8785 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 5

& 2015 Macmillan Publishers Limited. All rights reserved.

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8785

After weight measurement, the fully saturated samples were removed from the water container and rapidly placed on wet paper to remove excess water on the external surface. Each specimen was then placed in the bottom of a glass tube, which was then sealed for TD-NMR measurements.

TD-NMR method and parameter denitions. This technique (Supplementary Note 3) was applied to obtain skeletal pore-size distributions by the analysis of quasi-continuous distributions of the transverse relaxation time T2 (ref. 14) by means of the algorithm UPEN31, implemented in the UpenWin software32. Supplementary Figure 1a shows examples of T2 distributions for specimens of

B. europaea. In all cases, the slope of the distribution presented a strong increase at a specic T2 value, the cutoff, revealing a sharp boundary between two distinct pore classes. This point was chosen as the boundary between smaller (shorter T2)

and larger (longer T2) pores. The ratio of the area under the distribution for T2 larger than the cutoff to the total area under the distribution, called macroscale pore volume fraction (Supplementary Fig. 1b), indicates the fraction as a percentage of the pore volume in the macroscale. On the basis of the comparison with mercury intrusion porosimetry data, the sizes of pores in the macroscale pore volume fraction, corresponding to the major fraction of pores, are 41020 mm.

The tail in the distribution is about 3 orders of magnitude long and should correspond to smaller pore sizes up to tens of nanometres14. The distribution with T2 o3 ms was chosen to identify the signal corresponding to water in the smallest pores or size-scale, which we refer to here as nanoscale pore volume fraction. The remaining fraction of the distribution is called microscale pore volume fraction. The nano, micro and macroscale pore volume fractions multiplied by the skeletal porosity (Supplementary Table 2) produced the TD-NMR nano, micro and macroscale porosities, respectively.

TD-NMR apparatus and data acquisition. A homebuilt relaxometre basedon a 0.2 T permanent magnet operating at 8 MHz was used for acquisition of the transverse relaxation curve of the 1H nuclei of water molecules saturatingthe cleaned coral skeletons. The relaxometer was equipped with a coil B2 cm in diameter to analyse the entire coral and a Spinmaster console (Stelar, Mede, Pavia, Italy) for automatic pulse sequence transmission and data acquisition. The trans-verse relaxation data were acquired by using the CarrPurcellMeiboomGill sequence33 with a 200 ms echo time. Sixteen corals from S1, 16 corals from S2 and 12 corals from S3 were analysed.

SEM images. To investigate the meso and macroscale organization of the aragonite bre bundles, SEM analyses were performed. The cleaned coral skeletons were mounted (uncoated) to conductive carbon tape and examined with a Tescan Vega3 GMU (Czech Republic) variable pressure scanning electron microscope. Three corals per site were analysed.

Evaluation of net calcication rate. The age of each coral was determined based on growth band analysis via computerized tomography22,34. The linear extension rate was obtained by dividing the length of each sample (maximum axis of the oral disc) by its age. The net calcication rate was calculated as: net calcicationrate bulk density linear extension rate21,35 (Supplementary Table 2).

lCT analysis. To compute the corallite interseptal volume fraction (ratio of the total pore space volume inside the corallite to the total volume occupied by the corallite itself), mCT analyses were acquired and the convex hull approach was followed. mCT scans of coral skeletons were performed with a GE phoenix X-ray Nanotom S. The voxel sizes in the scans varied from 2.91 to 13.25 mm depending on the actual size of the investigated coral sample. All image processing steps in the analysis were performed with the open source software FIJI36. Supplementary Fig. 3 shows different views of the 3D volume renderings obtained from three corals of the same age from each of the three study sites. Ten corals from each site were analysed.

Nanoindentation. To analyse mechanical properties of the coral skeletons, skeletal stiffness (Youngs modulus) and hardness were determined by OliverPharr analysis of load-depth curves37, using a Nanoindentation Tester, model NHT-TTX by CSM Instruments, Switzerland, equipped with a Berkovich diamond tip (opening angle 142.3). On each sample, a minimum of 10 indentation tests were carried

out, in the oral region, with a minimum distance of 30 mm between two tests (the total number of tests for each site is reported in Supplementary Table 4). To avoid inuence from the embedding resin, a minimum distance of 100 mm from the skeleton/resin boundaries was always maintained.

SAXS analysis. To obtain information on skeletal porosity at the nanoscale, SAXS analyses (Supplementary Note 4) were performed on three corals per site. Sections were mounted onto a laboratory SAXS apparatus with a rotating anode X-ray generator (Bruker, AXS, Karlsruhe, Germany) operating with a copper anode. Light microscopy images of the three specimens are shown in Supplementary Fig. 8 together with X-ray radiographies taken with the SAXS instrument.

AFM analysis. To study the morphology and packing of skeletal mineral grains at the nanoscale, AFM images were acquired. Samples were lightly polished using diamond paste, cleaned with milli-Q water and observed using a Veeco AFM Dimension 3100 Nanoscope III, Plainview, NY. The probe consisted of a cantilever with an integrated silicon nitride tip. Samples were imaged at room temperature and in air using tapping mode phase contrast imaging.

Statistical analysis. Statistical analyses (analysis of variance, multivariate, principal components and quantile analysis) were performed using the Statistical Package STATA 9.0 (StataCorp LP). To test the signicance of the differences among sites, parametric (F) or non-parametric (KruskalWallis w2) tests were run. Non-parametric tests were performed for data that did not assume normal distributions. Multivariate regression analyses were performed to investigate the relationships between one dependent and one or several independent variables, using ordinary least squares robust to outliers. The model is described by the equation (1):

yi a X

M

j1

bjxi;j ei 1

where the index i refers to the i-observation, xj is an independent variable, yi is the value of the dependent variable and ei is the corresponding error. The constants a, bj (j 1, M) are the best-t parameters, to be determined by the best t. Quantile

analysis was performed to study the previous relationships for homogeneous groups of values of the dependent variable. This analysis was used to give a more comprehensive picture of the effect of the independent variable (pH) on the dependent variables of Fig. 3 and Supplementary Fig. 1, as it can show different effects of the independent variable in different ranges of the dependent variable.

References

1. Stocker, T. F. et al.in Climate Change 2013: The Physical Science Basis (Cambridge Univ. Press, 2013).

2. Sabine, C. L. et al. The oceanic sink for anthropogenic CO2. Science 305,

367371 (2004).

3. Six, K. D. et al. Global warming amplied by reduced sulphur uxes as a result of ocean acidication. Nat. Clim. Change 3, 975978 (2013).4. Rodolfo-Metalpa, R. et al. Coral and mollusc resistance to ocean acidication adversely affected by warming. Nat. Clim. Change 1, 308312 (2011).

5. Hoegh-Guldberg, O. et al. Coral reefs under rapid climate change and ocean acidication. Science 318, 17371742 (2007).

6. Dove, S. et al. Future reef decalcication under a business-as-usual CO2 emission scenario. Proc. Natl Acad. Sci. USA 110, 1534215347 (2013).

7. Pandol, J. M., Connolly, S. R., Marshall, D. J. & Cohen, A. L. Projecting coral reef futures under global warming and ocean acidication. Science 333, 418422 (2011).

8. Orr, J. C. et al. Anthropogenic ocean acidication over the twenty-rst century and its impact on calcifying organisms. Nature 437, 681686 (2005).

9. Touratier, F. & Goyet, C. Impact of the Eastern Mediterranean transient on the distribution of anthropogenic CO2 and rst estimate of acidication for the

Mediterranean Sea. Deep-Sea Res. I 58, 115 (2011).10. Yilmaz, A. et al. Impact of acidication on biological, chemical and physical systems in the Mediterranean and Black Sea Mediterranean. Mediterranean Sci. Committee (CIESM), Monograph Ser. 36, 124 (2008).

11. Calvo, E. et al. Effects of climate change on Mediterranean marine ecosystems: the case of the Catalan Sea. Climate Res. 50, 129 (2011).

12. Ziveri, P. Research turns to acidication and warming in the Mediterranean Sea, IMBER (Integrated Marine Biogeochemistry and Ecosystem Research). http://www.imber.info/index.php/News/Newsletters/Issue-n-20-May-2012

Web End =http://www.imber.info/index.php/News/Newsletters/Issue-n-20-May-2012 (2012).

13. Goffredo, S. et al. Biomineralization control related to population density under ocean acidication. Nat. Clim. Change 4, 593597 (2014).

14. Fantazzini, P. et al. Time-Domain NMR study of Mediterranean scleractinian corals reveals skeletal-porosity sensitivity to environmental changes. Environ. Sci. Technol. 47, 1267912686 (2013).

15. Laine, J., Labady, M., Albornoz, A. & Yunes, S. Porosities and pore sizes in coralline calcium carbonate. Mater. Charact. 59, 15221525 (2008).

16. Caroselli, E. et al. Environmental implications of skeletal micro-density and porosity variation in two scleractinian corals. Zoology 114, 255264 (2011).

17. Carricart-Ganivet, J. P. Sea surface temperature and the growth of the West Atlantic reef-building coral Montastraea annularis. J. Exp. Marine Biol. Ecol. 302, 249260 (2004).

18. Crook, E. D., Cohen, A. L., Rebolledo-Vieyra, M., Hernandez, L. & Paytan, A. Reduced calcication and lack of acclimatization by coral colonies growing in areas of persistent natural acidication. Proc. Natl Acad. Sci. USA 110, 1104411049 (2013).

19. Iglesias-Prieto, R., Galindo-Martinez, C. T., Enriquez, S. & Carricart-Ganivet, J.P. Attributing reductions in coral calcication to the saturation state of

6 NATURE COMMUNICATIONS | 6:7785 | DOI: 10.1038/ncomms8785 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

& 2015 Macmillan Publishers Limited. All rights reserved.

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8785 ARTICLE

aragonite, comments on the effects of persistent naturals acidication. Proc. Natl Acad. Sci. USA 111, E300E301 (2014).20. Goffredo, S., Caroselli, E., Mattioli, G., Pignotti, E. & Zaccanti, F. Variation in biometry and demography of solitary corals with environmental factors in the Mediterranean Sea. Mar. Biol. 152, 351361 (2007).

21. Goffredo, S. et al. Inferred level of calcication decreases along an increasing temperature gradient in a Mediterranean endemic coral. Limnol. Oceanogr. 54, 930937 (2009).

22. Goffredo, S., Caroselli, E., Mattioli, G., Pignotti, E. & Zaccanti, F. Relationships between growth, population structure and sea surface temperature in the temperate solitary coral Balanophyllia europaea (Scleractinia, Dendrophylliidae). Coral Reefs 27, 623632 (2008).

23. Airi, V. et al. Reproductive efciency of a Mediterranean endemic zooxanthellate coral decreases with increasing temperature along a wide latitudinal gradient. PLoS ONE 9, e91792 (2014).

24. Wooldridge, S. A. Assessing coral health and resilience in a warming ocean: Why looks can be deceptive. Bioessays 36, 10411049 (2014).

25. Capaccioni, B., Tassi, F., Vaselli, O., Tedesco, D. & Poreda, R. Submarinegas burst at Panarea Island (southern Italy) on 3 November 2002: A magmatic versus hydrothermal episode. J. Geophys. Res. 112, B05201 (2007).

26. Zibrowius, H. Les scleractiniaires de la Mditerrane et delAtlantique nord-oriental. Mem. Inst. Oceanogr 11, 1284 (1980).

27. Jokiel, R. L., Maragos, J. E. & Franzisket, L. In: Coral Reef Research Methods. (eds Stoddart, D. R. & Johannes, R. E.) 5 529542 (UNESCO, 1978).

28. Davies, P. S. Short-term growth measurements of corals using an accurate buoyant weighing technique. Mar. Biol. 101, 389395 (1989).

29. Walsh, S.-J. in Proceedings of 12th International Coral Reef Symposium, Cairns, Australia, 913 July 2012.

30. Manger, G. E. Porosity and bulk density of sedimentary rocks. Geol. Survey Bull. 1144E (1963).

31. Borgia, G. C., Brown, R. J. S. & Fantazzini, P. Examples of marginal resolution of NMR relaxation peaks using UPEN and diagnostics. Magn. Reson. Imaging 19, 473475 (2001).

32. Bortolotti, V., Brown, R. J. S. & Fantazzini, P. UpenWin: a Software to Invert Multi-Exponential Decay Data; http://www.unibo.it/PortaleEn/Research/Services&plus;for&plus;companies/UpenWin.htm

Web End =http://wwwunibo.it/PortaleEn/Research/ http://www.unibo.it/PortaleEn/Research/Services&plus;for&plus;companies/UpenWin.htm

Web End =Services http://www.unibo.it/PortaleEn/Research/Services&plus;for&plus;companies/UpenWin.htm

Web End =for http://www.unibo.it/PortaleEn/Research/Services&plus;for&plus;companies/UpenWin.htm

Web End =companies/UpenWin.htm (2009).

33. Cowan, B. Nuclear Magnetic Resonance and Relaxation (Cambridge Univ. Press, 1997).

34. Goffredo, S., Mattioli, G. & Zaccanti, F. Growth and population dynamics model of the Mediterranean solitary coral Balanophyllia europaea (Scleractinia Dendrophylliidae). Coral Reefs 23, 433443 (2004).

35. Lough, J. M. & Barnes, D. J. Environmental controls on growth and of the massive coral Porites. J. Exp. Mar. Biol. Ecol. 245, 225243 (2000).

36. Schindelin, J. et al. Fiji: an open-source platform for biological-image analysis. Nat. Methods 9, 676682 (2012).

37. Oliver, W. C. & Pharr, G. M. An improved technique for determining hardness and elastic modulus using load and displacement sensing indentation experiments. J. Mater. Res. 7, 15641583 (1992).

Acknowledgements

The research leading to these results has received funding from the European Research Council under the European Unions Seventh Framework Programme (FP7/20072013)/ ERC grant agreement n 249930CoralWarm: Corals and global warming: the Mediterranean versus the Red Sea. Bartolo Basile, Francesco Sesso and Eolo Sub diving center assisted in the eld. Francesca Gizzi and Giorgia Polimeni helped during preparation and participated in eld surveys. The Scientic Diving School collaborated with the underwater activities. We are grateful to Ingrid Zenke (MPI Potsdam) for help with the SAXS measurements, Michela Reggi for sample preparation, Gianni Neto and Francesco Sesso for the images in the Panarea site.

Author contributions

S.G., G.F., P.F., L.P. and Z.D. conceived and designed the research. E.C., F.P., S.G. and B.C. collected the samples and performed the diving eld work. L.B., M.M., M.D.G., E.C., F.P., P.K., J.U.H., Y.D., J-P.C., J.C.W., W.W. and G.F. performed the laboratory analyses. P.F., S.M., L.P., V.B., L.B., M.M., E.C., F.P., J.A.K., P.K., J.U.H., Y.D., J-P.C., J.C.W., W.W., P.F., G.F. and S.G. analysed the data. All authors wrote the manuscript and participated in the scientic discussion.

Additional information

Supplementary Information accompanies this paper at http://www.nature.com/naturecommunications

Web End =http://www.nature.com/ http://www.nature.com/naturecommunications

Web End =naturecommunications

Competing nancial interests: The authors declare no competing nancial interests.

Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/

Web End =http://npg.nature.com/ http://npg.nature.com/reprintsandpermissions/

Web End =reprintsandpermissions/

How to cite this article: Fantazzini, P. et al. Gains and losses of coral skeletal porosity changes with ocean acidication acclimation. Nat. Commun. 6:7785 doi: 10.1038/ncomms8785 (2015).

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the articles Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Web End =http://creativecommons.org/licenses/by/4.0/

NATURE COMMUNICATIONS | 6:7785 | DOI: 10.1038/ncomms8785 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 7

& 2015 Macmillan Publishers Limited. All rights reserved.