Environmental stability and phenotypic plasticity benefit the cold-water coral *Desmophyllum dianthus* in an acidified fjord

Cold-water corals (CWCs) are key ecosystem engineers in deep and cold marine environments, yet face rapid anthropogenic changes such as warming, acidification, and deoxygenation. Laboratory studies have largely examined CWC responses under constant conditions, potentially misrepresenting performance in the naturally variable habitats where CWCs live. Environmental variability in temperature, salinity, oxygen, and pH occurs seasonally to daily in CWC habitats, but in situ physiological datasets remain scarce. This study investigates how natural environmental variability and contrasting carbonate chemistry influence the fitness, acclimatization potential, and potential local adaptation of the cosmopolitan CWC Desmophyllum dianthus within Chile’s stratified Comau Fjord. The central questions were: (1) How do D. dianthus calcification, respiration, and tissue traits vary across horizontal (head-to-mouth) and vertical (shallow-to-deep) environmental gradients and seasons? (2) Do corals show local adaptation or rapid acclimatization when reciprocally transplanted? (3) Which environmental drivers (including variability) best explain fitness proxies?

Prior work has emphasized carbonate chemistry, temperature, salinity, oxygen, food availability, and hydrodynamics as key controls on CWC distribution and physiology. Many studies used constant laboratory conditions, whereas in situ records show high-frequency fluctuations driven by tides, internal waves, and advection. Organisms from variable environments can exhibit enhanced tolerance, and tropical corals sometimes show greater stress resistance in variable regimes; however, responses are context- and trait-dependent, and few reciprocal transplants have been conducted with CWCs. Some CWC reefs occur near or below the aragonite saturation horizon, and laboratory and field studies show CWCs can calcify under aragonite undersaturation. Yet, the extent to which natural variability modulates CWC fitness remains unresolved, motivating reciprocal transplantation combined with long-term in situ environmental monitoring.

Study site: Comau Fjord, northern Chilean Patagonia (~45 km long, max depth ~500 m), with strong tidal range (up to 7.5 m) and pronounced stratification due to freshwater inputs. Deep marine layers are relatively low in oxygen and pH. D. dianthus occurs unusually shallow (to ~15 m) and abundantly below ~25 m. Environmental monitoring: Temperature loggers (Tidbit v2; 15-min intervals) at each coral station recorded September 2016–August 2017. CTD profiles (SBE 19plus V2 with SBE 43 O2 sensor) were taken once per season (January, May, August 2017). An additional CTD (AML plus X) at 25 m at station X logged temperature and salinity every 30 min. Discrete seasonal water samples (near experimental corals) measured total alkalinity (TA), dissolved inorganic carbon (DIC), and nutrients. Carbonate system parameters (pH, pCO2, Ωarag, [CO3 2−]) were calculated with CO2SYS using standard dissociation constants. Experimental design and transplantation: Seven stations were established: six shallow (~20 m; A–F spanning head to mouth) and one deep (E deep, ~300 m). A year-long in situ assessment (September 2016–August 2017) combined with reciprocal transplantation was performed. Transplants were done across the strongest contrasts: A↔F (horizontal gradient) and Es↔Ed (vertical gradient). Stations B, C, and D corals were reinstalled only at native sites. In total, 392 corals were used. For repeated physiological measurements (experimental corals), 68 native and 38 novel (cross-transplanted) individuals were tracked. Coral handling used plates and screws affixed to fjord walls; deep plates were mounted on a rack lowered to 300 m via pulley. Corals were collected by SCUBA (shallow) or ROV (deep). Prior to reinstallation, shallow corals with bare skeleton were trimmed to reduce bioeroded portions; all corals were glued to labeled screws. A calcein mark defined the start of the experiment. Physiological measurements: Calcification rates were assessed seasonally (after ~4, 8, 11 months: January, May, August 2017) via buoyant weight, converted to skeletal dry mass and normalized to tissue-covered surface area (mg CaCO3 cm−2 d−1). Respiration was measured by 6-h dark, closed-cell incubations (800 mL vials with 100 µm filtered seawater; magnetic stirring) at standardized temperatures representative of shallow stations at each season (winter ~11.8 °C, autumn ~12.2 °C, summer ~14.2 °C) to compare metabolic potential between stations (not in situ temperatures). Oxygen consumption was corrected for controls and normalized to tissue-covered area (µmol O2 cm−2 d−1). Tissue-covered surface area was measured at experiment end (digital caliper with geometric approximation for trumpet-shaped calyx); for deep corals, seasonal scaled image analyses (ImageJ) captured strong area increase. Comparison with mass-normalized calcification confirmed minimal bias. Environmental variability metric: Long-term temperature records were decomposed to derive diurnal anomaly-based mean seasonal/annual variability as a proxy for variability in co-varying parameters (pH, Ωarag, salinity, oxygen). Statistics: Linear mixed-effects models (lmer, R/lme4) assessed effects of depth, season, station, and transplantation on calcification and respiration; season, station, and station×transplant were fixed factors; coral ID was a random factor. Separate models examined (i) native shallow corals across the horizontal gradient and seasons, and (ii) native and novel corals at A, F, Es, and Ed to test transplantation and depth effects. Post hoc contrasts used lsmeans. Multifactor linear models with AIC-based model selection (AICcmodavg) tested relationships between calcification and environmental predictors (mean seasonal temperature, its variability, pH, Ωarag, salinity, oxygen). Assumptions were checked via residual diagnostics.

• Environmental regimes: Shallow (20 m) waters exhibited higher temperatures (annual mean 12.5 ± 0.9 °C) and strong high-frequency fluctuations (daily swings up to 3.7 °C; summer peaks to 16.6 °C). Deep (300 m) waters were cooler and more stable (mean 11.4 ± 0.2 °C), with higher salinity and lower oxygen, pH, and Ωarag than shallow waters, following similar seasonal patterns.
• Depth effect: Both calcification and respiration were significantly higher at 300 m than at 20 m (LMM; Es–Ed p < 0.001). Deep corals occurred under lowest Ωarag (<1) and T (<12 °C) yet showed the highest growth and fitness.
• Horizontal (shallow) gradient: Calcification was higher at the mouth (station F) than head (A) (LMM p < 0.001). Respiration was generally higher toward the mouth (A–F p = 0.011) and peaked at station C (e.g., A–C p < 0.001).
• Seasonality: Calcification was lower in winter (August) than summer (January) and autumn (May) (Jan–Aug and May–Aug p < 0.001). Respiration was higher in summer versus autumn and winter (Jan–May p = 0.007; Jan–Aug p = 0.038).
• Transplant response (phenotypic plasticity): Novel (cross-transplanted) corals rapidly adjusted calcification and respiration to local conditions. No significant transplant effect on calcification (LMM transplant p = 0.257) or respiration (p = 0.562). Deep-transplanted corals expanded tissue surface and biomass similarly to native deep corals and developed trumpet-shaped calyces.
• Drivers of calcification: Although calcification correlated with pH and Ωarag patterns, the highest rates occurred under aragonite undersaturation. Multifactor modeling identified mean seasonal temperature and mean seasonal temperature variability as the strongest predictors, explaining 55% of variance (adjusted R² = 0.555). Temperature variability correlated negatively with calcification; the deep station (lowest variability) had the highest growth.
• Energetics and allocation: Deep corals exhibited ~2.3-fold higher calcification and ~1.7-fold higher respiration than shallow corals and invested ~6.7 times more into tissue biomass, suggesting greater energy availability and/or allocation to somatic growth at depth. Shallow corals often had low tissue coverage and more endolithic infestation, implying additional stress and energetic trade-offs.

The reciprocal transplant across sharp horizontal and vertical gradients demonstrates that D. dianthus possesses high phenotypic plasticity: transplanted corals rapidly acclimatized, matching the physiological performance of natives at each site. Contrary to expectations that higher temperatures would enhance calcification, the fittest phenotype occurred in deep waters with aragonite undersaturation but stable conditions, while shallow, highly variable environments were associated with reduced performance. This indicates that environmental variability, rather than carbonate saturation state per se, can be the dominant constraint on CWC calcification in macrotidal systems. Potential mechanisms include energetic costs of coping with frequent fluctuations, altered feeding behavior during thermal/salinity swings, shifts in zooplankton availability or composition, and interactions with bioerosion/infestation. The findings reconcile observations of thriving CWC banks below the aragonite saturation horizon and emphasize that live CWCs can maintain calcification under low Ωarag if energetic conditions are favorable. Incorporating environmental variability into physiological studies and predictive frameworks is critical for assessing CWC resilience to future ocean change.

This study shows that Desmophyllum dianthus in Comau Fjord achieves highest fitness under environmentally stable, aragonite-undersaturated deep conditions and that corals exhibit rapid acclimatization across contrasting habitats, evidencing strong phenotypic plasticity. Environmental variability, especially short-term thermal variability, negatively correlates with calcification and can outweigh the direct effect of carbonate chemistry on performance. The work underscores the need to account for natural high-frequency variability when predicting CWC responses to climate change. Future research should: (1) disentangle direct versus indirect effects of variability (e.g., food availability, community interactions); (2) evaluate trait-specific acclimatization/adaptation (including reproductive cycles, biochemical and transcriptomic responses); (3) expand multifactor experiments incorporating realistic environmental fluctuations; and (4) broaden spatial replication of deep and shallow habitats to generalize patterns.

• Respiration was measured at standardized, not in situ, temperatures, which may over- or underestimate deep coral metabolic rates relative to natural conditions.
• Food availability, zooplankton composition, and potential swarming/micronekton contributions at coral scales were not directly quantified; energetic interpretations are partly inferential.
• Bioerosion and endolithic infestation differed between depths; although skeletal trimming was applied to shallow corals, residual effects may persist.
• Deep environment was represented by a single station; broader replication would strengthen generality.
• Some tissue surface area estimates differ in method between shallow (end-of-study caliper) and deep (seasonal image analysis), potentially introducing minor biases.
• Only certain station pairs were reciprocally transplanted; other combinations were not tested.
• The study duration (one year) may not capture longer-term trait adjustments (e.g., reproductive investment) or delayed responses.
• Co-variation among environmental parameters limits attribution of causality despite multifactor modeling.