The following methodology applies to this dataset in addition to other datasets published in Edmunds et al. (2019).

Methodology:

Overview

Back reef communities were assembled in four flumes, with each randomly assigned to pCO2 treatments targeting ambient (400 μatm), 700 μatm, 1000 μatm, and 1300 μatm pCO2 to approximate atmospheric pCO2 projected for ~ 2140 under representative concentration pathways (RCP) 2.6, 4.5, 6.0 and 8.5, respectively. Treatments were maintained for one year from November 2015, and actual pCO2 treatments differed from target values. Each flume consisted of a working section that was 5.0 m long, 30 cm wide and filled to ~ 30-cm depth with ~ 500 L of seawater that was circulated and refreshed with sand-filtered (pore size ~ 450–550 µm) seawater from Cook’s Bay (-17.491, -149.826, 14-m depth) at ~ 5 L min-1.

Planar growth and community structure were measured because they are used in ecological analyses of coral reefs, and we reasoned they would sharpen the ability to interpret the ecological implications of the physiological impacts of OA on calcification. We anticipated that the community response to OA would include reduced linear extension, impaired planar growth of tissue and skeleton, and increase partial mortality (as in Dove et al. 2013). The mean linear extension expected for the corals in the present study (Porites rus = 15.2 ± 5.7 mm y-1, massive Porites = 10.0 ± 0.6 mm y-1, Montipora = 27.7 + 3.0 mm y-1, and Pocillopora verrucosa = 24.7 ± 2.4 mm y-1 [https://coraltraits.org/, accessed 8 October 2018]) were expected to create annual changes in planar area of 52 cm2 (with mean initial size of 69 cm2), 32 cm2 (with mean initial size of 68 cm2), 106 cm2 (with mean initial size of 70 cm2), and 150 cm2 (with mean initial size of 218 cm2), respectively, in the ambient flume. To evaluate the precision of the photographic method, 10 independent images of mounding and branching corals in the flumes were recorded, and were processed to provide replicate determinations of organism size (i.e., planar area). These images showed that the standard deviations of mean area determinations were 2.3% for massive Porites, and 3.8% for Pocillopora verrucosa). Based on these measures of precision, there would be a 75% chance of detecting annual growth of 0.6 cm2 for massive Porites and 4.8 cm2 for Pocillopora verrucosa, which represent reasonable estimates for the growth of these corals in our flumes. Given effect sizes ranging from 21.1% for Lithophyllum to 10.2% for massive Porites upon exposure to 1067 µatm pCO2 (Comeau et al. 2014), an effect of pCO2 on growth in the present study would be detectable for Montipora, while smaller effects of pCO2 for other taxa might be prone to Type II errors in detection (i.e., they might not be detected when present).

Back reef communities were assembled to correspond to the mean percent cover of the major space holders in this habitat in 2013 (data archived in Edmunds 2015). The Back reef community source was latitude: -17.481, and longitude: -149.836 ± 4 km from this point along the north shore. The communities began with ~ 25% coral cover, with 11% massive Porites spp., 7% Porites rus, 4% Montipora spp., 3% Pocillopora spp., and ~ 7% crustose coralline algae (CCA), consisting of 4% Porolithon onkodes and 3% Lithophyllum kotschyanum. Coral rubble (~ 1-cm diameter) was added to ~ 5% cover, and the remainder of the benthic surface was sand. Analyses of community structure focused on the central, 2.4-m long portion of this community where corals and CCA were secured to a plastic-coated, metal grid (5 × 5 cm mesh) and represented the “fixed” community. Securing organisms to the grid was critical to reduce parallax errors in photography, to allow the organisms to grow and interact as they extended over the year experiment, and to allow ecologically meaningful analysis of community structure using photographs.

The central section of each flume included a 2.4-m long sediment box that extended the width of the flume, and contained 30-cm depth of sediment. The sediment box was flanked by ~ 2.6 m of the fiberglass floor of the flume, along which 0.8 m was occupied by the same benthic community, but with corals and CCA resting on the bottom (i.e., “unfixed”). Members of the fixed community were buoyant weighed at the start and end of the year to measure Net changes in mass (Gnet), but otherwise were left in place. Members of the unfixed community were removed monthly to measure buoyant weight to calculate Gnet (described below). The unfixed portion of the community allowed monthly resolution of Gnet, but the necessity for removal from (and return to) the flume to measure Gnet resulted in relocation error that negated their use in photographic measurement of community structure. In addition to the coral, sand, CCA, and rubble, the flumes were augmented with holothurians (~ 8-cm long, Holothuria spp.), and macroalgae (Turbinaria ornata and Halimeda minima) to approximate the cover of these algae in the back reef in 2013 (~ 4–5%).

Corals, CCA, and rubble were collected from ~ 2-m depth in the back reef, and were attached with epoxy (Z-Spar A788) to plastic bases. Sediments were collected in the same location, and were placed into boxes that were buried in situ, flush with the sediment for 3 d to promote stratification, and then installed in each flume. Back reef communities were constructed in the flumes on 12 November 2015, and were maintained under ambient conditions until 17 November 2015, when pCO2 treatments began in three flumes, with levels increased to target values over 24 h. Throughout the experiment, the flumes were cleaned of algal turf that grew on the walls of the flumes as well as exposed plastic and the metal grid on the floor of the flume. Turfs were not removed from natural surfaces (i.e., coral bases and rubble) with the rationale that they are a normal component of back reef communities.

Physical and chemical parameters:

Seawater was circulated at ~ 0.1 m s-1 using a pump (W. Lim Wave II 373 J s-1), and flow speeds were measured across the working sections using a Nortek Vectrino Acoustic Doppler Velocimeter. This flow speed was relevant for the back reef of Mo'orea. The flumes were exposed to sunlight that was shaded to a photon flux density (PFD) of photosynthetically active radiation (PAR) approximating 2-m depth in the back reef. Light was measured using cosine-corrected sensors (Odyssey, Dataflow Systems Ltd, New Zealand) that were calibrated with a LI-COR meter (LI-1400, Li-COR Biosciences, Lincoln, NE) attached to a 2π sensor (LI 192A). Maximum daily PFD varied by day and season from 364–1,831 μmol quanta m-2 s-1. Temperatures were regulated close to the mean monthly temperature in the back reef that increased from ~ 27.8°C in December 2015, to ~29.3°C in April 2016, and back to ~ 27.4 °C in November 2016.

Seawater carbonate chemistry was uncontrolled in one flume (ambient), and in the three others, seawater pH was controlled through the addition of CO2 gas (using solenoids controlled with an Aquacontroller, Neptune Systems, USA) to approximate pCO2 targets. A diurnal upward adjustment of ~ 0.1 pH was applied to the treatments to simulate natural variation in seawater pCO2 in the back reef. The ambient flume also maintained a diurnal variation in pCO2 with a nighttime pH ~ 0.1 lower than daytime. Ambient air was bubbled into all flumes.

PAR and temperature (Hobo Pro v2 [± 0.2°C], Onset Computer Corp., MA, USA) were recorded, and pH was measured daily (at various times of day) on the total hydrogen ion scale (pHT). Temperature and pH were used to adjust the thermostat and pH-set points close to values that were calculated (using seacarb) to correspond to target treatments of 400 µatm, 700 µatm, 1000 µatm, and 1300 µatm (~ 8.04, ~7.81, ~7.70 and ~7.65, respectively). Seawater carbonate chemistry (pH and AT) and salinity were measured at 14:00 hrs and 20:00 hrs weekly. A conductivity meter (Thermo Scientific, Orionstar A212, Waltham, MA, USA) was used to measure salinity. The remaining parameters of the seawater carbonate system were calculated from temperature, salinity, pHT, and AT, using the R package seacarb. Calculations were made using the carbonic acid dissociation constants, the KSO4 concentration for the bisulfate ion, and the Kf constant.

pHT was measured using a DG 115-SC electrode (Mettler Toledo, Columbus, OH, USA) that was calibrated with a TRIS buffers. AT was measured using open-cell, acidimetric titration (SOP 3b [Dickson et al. 2007]) using certified titrant with a titrator (T50 with a DG 115-SC electrode, Mettler Toledo). The accuracy and precision of measurements were determined using reference materials (from A. Dickson, Scripps Institution of Oceanography, CA, USA), against which measured values of AT maintained an accuracy of 1.7 ± 0.3 μmol kg-1 (n = 15) and precision of 1.8 ± 0.1 μmol kg-1 (n = 475).

Response variables:

Net changes in mass (Gnet) of corals and CCA was measured using buoyant weight (± 1 mg) by month (unfixed) or year (fixed community). Buoyant weight was converted to dry weight of CaCO3 using empirical seawater density (~1.02278 g cm-3) and the density of pure aragonite (2.93 g cm-3, corals) and pure calcite (2.71 g cm-3, CCA). Gnet in each month was expressed as the percentage change in mass relative to the initial mass in November 2015. As the area of tissue changed throughout the experiment through growth and partial mortality, “growth” could not be expressed on an area-normalized scale.

Community structure was quantified using planar photographs recorded in ambient light using a GoPro Hero 4 camera (12 MP, 3-mm focal length). The camera was moved along the flume to record the community in the working section using ~ 16 frames sampling-1.

Photographs were analyzed using ImageJ software, in which the planar area of living tissue on corals and CCA was quantified by outlining organisms and scaling the image using the metal grid as a reference. Size (cm2) was expressed as a percentage of the area (240 × 30 = 7200 cm2) occupied by the fixed members of the community. The summed area of community members was used to determine overall cover of the benthic community, and changes in area were used to quantify growth. Where organisms died, their area was set to zero.

Data parameter "Actual" are data as obtained with missing values. NMDS = data processed for NMDS analyses with missing values replaced as described in Edmunds et al. (2019).

See Edmunds et al. (2019) for analyses that used these data.