Intra-colony spatial variance of oxyregulation and hypoxic thresholds for key Acropora coral species

Abstract

Oxygen (O₂) availability is essential for healthy coral reef functioning, yet how continued loss of dissolved O₂ via ocean deoxygenation impacts performance of reef building corals remains unclear. Here, we examine how intra-colony spatial geometry of important Great Barrier Reef (GBR) coral species Acropora may influence variation in hypoxic thresholds for upregulation, to better understand capacity to tolerate future reductions in O₂ availability. We first evaluate the application of more streamlined models used to parameterise Hypoxia Response Curve data, models that have been used historically to identify variable oxyregulatory capacity. Using closed-system respirometry to analyse O₂ drawdown rate, we show that a two-parameter model returns similar outputs as previous 12th-order models for descriptive statistics such as the average oxyregulation capacity (T_pos) and the ambient O₂ level at which the coral exerts maximum regulation effort (P_cmax), for diverse Acropora species. Following an experiment to evaluate whether stress induced by coral fragmentation for respirometry affected O₂ drawdown rate, we subsequently identify differences in hypoxic response for the interior and exterior colony locations for the species Acropora abrotanoides, Acropora cf. microphthalma and Acropora elseyi. Average regulation capacity across species was greater (0.78-1.03 ± SE 0.08) at the colony interior compared with exterior (0.60-0.85 ± SE 0.08). Moreover, P_cmax occurred at relatively low pO₂ of <30% (±1.24; SE) air saturation for all species, across the colony. When compared against ambient O₂ availability, these factors corresponded to differences in mean intra-colony oxyregulation, suggesting that lower variation in dissolved O₂ corresponds with higher capacity for oxyregulation. Collectively, our data show that intra-colony spatial variation affects coral oxyregulation hypoxic thresholds, potentially driving differences in Acropora oxyregulatory capacity.

Keywords: Climate change; Coral oxyregulator; Coral reefs; Hypoxia Response Curves; Hypoxic tolerance; Ocean deoxygenation.

FIGURE 1

(a) Schematic diagram outlining points of extraction for hypoxia response curve (HRC) parameters (T_pos, P_cmax and P_cmin) using the rho(pO₂) equation (Equation 1), in addition to (b–h) models of varying polynomial degrees (1–12th order) fit to replicates of HRC data sets of model species Acropora kenti (n = 3) (from Hughes, Alexander, et al., 2022). Replicates indicated by colour, and individual model fits (per rep) by lines.

FIGURE 2

(a) and (b) Acropora hyacinthus hypoxia response curve (HRC) data set Rep 2 and Rep 3, fit with 2‐parameter Michaelis–Menten (MM) function (yellow line), and 12th‐order constrained dodecic model (green line). Reps 2 and 3 also annotated on (g) P_cmax scatterplot | (c–e) Extracted parameters: T_pos (relative units), P_cmax and P_cmin (% air saturation) from HRC data as analysed by Hughes, Alexander, et al. (2022) i.e., ‘Original’ (solid colour), and re‐analysed here, that is, ‘Re‐fit’ (striped colour) using the MM model, of Acropora species (n = 3): A. hyacinthus, A. intermedia, and A. kenti. Error bars are calculated standard error. | (f–h) correlation analyses of the same extracted data: T_pos, P_cmax, and P_cmin, from original fits (y‐axes), and from re‐fit curves using MM model (x‐axes). Linear R ² values calculated across averages (shown as the solid black line), and Pearson's Correlation Coefficients (not shown): T_pos = 0.19, P_cmax = 0.49, and P_cmin = 0.12, and the significance of the correlation coefficients (where p‐value < .05 is considered significant, shown). Shading indicates the 90% confidence interval.

FIGURE 3

Comparison of the extracted parameters: (a) T_pos (relative units), (b) P_cmax, and (c) P_cmin (% air saturation) from hypoxia response curve data of replicates (n = 6) of Acropora loripes. Plots compare extracted parameters from fragments clipped to the reef substrate for 7 days using the CoralClip® (solid colour), and extractions from samples freshly fragged (striped colour), fit with the selected Michaelis–Menten model. Error bars are calculated standard error.

FIGURE 4

Dissolved oxygen (DO) content (mg O₂ L⁻¹) displayed on the left y‐axis was measured at the interior (yellow) and exterior (blue) sections of the three branching Acropora colonies; panels (a) A. abrotanoides, (b) A. cf. microphthalma, and (c) A. elseyi, respectively. Ambient water temperature (orange, °C) on the inside right y‐axis, was also measured by the loggers deployed in the sample colonies, and is coupled with tidal height (dark shaded area with black outline, m) on the outside right y‐axis. Daily light cycles are represented by vertical shading and signified by the sun and moon symbols, over time on the x‐axis (date). Note that vertical dashed lines indicate fragment sampling times for each Acropora colony, carried out over the experimental period (as summarised in Table 1), whilst the horizontal dashed line indicates 100% air sat (i.e., at ~7.67 mg O₂ L⁻¹).

FIGURE 5

(a–c) Comparison of the extracted parameters: T_pos (relative units), P_cmax, and P_cmin (% air saturation) from HRC data of replicates (n = 6) of Acropora species: A. abrotanoides, A. cf. microphthalma, and A. elseyi. Plots compare extracted parameters from fragments taken from the interior section of the colony (solid colour), and fragments from the exterior thickets (striped colour), fit with the selected MM model. Error bars are calculated standard error. Note that significance levels (where p‐values < .05 considered significant) are identified by a bold * above relevant data. (d) Correlation analysis between mean interior and exterior delta (Δ) O₂ logger data (DO, mg L⁻¹) and Δ T_pos data (relative units), between species. Linear R ² was calculated across means (shown as the solid black line), as well as Pearson's Correlation Coefficient = 0.83, and the significance of the correlation coefficient (also shown). (e) Additional correlation analysis between Δ O₂ logger range (max‐min), and Δ T_pos data, between species, with Pearson's Correlation Coefficient = 0.54, and linear R ² and significance of correlation coefficient (shown). Shading is 90% confidence interval.