Seasonal Upwelling Conditions Modulate the Calciﬁcation Response of a Tropical Scleractinian Coral

: Natural processes such as upwelling of deeper-water masses change the physical-chemical conditions of the water column creating localized ocean acidiﬁcation events that can have an impact on the natural communities. This study was performed in a coral reef system of an archetypical bay within the Tayrona National Natural Park (PNNT) (Colombia), and aimed to quantify net calciﬁcation rates of a foundational coral species within a temporal context (6 months) taking into account the dynamics of seasonal upwelling that inﬂuence the study area. Net calciﬁcation rates of coral fragments were obtained in situ by the alkalinity anomaly technique in short-term incubations (~2.5 h). We found a signiﬁcant effect of the upwelling on net calciﬁcation rates ( Gnet ) ( p < 0.05) with an 42% increase in CaCO 3 accretion compared to non-upwelling season. We found an increase in total alkalinity (A T ) and dissolved inorganic carbon (DIC) with decreased aragonite saturation ( Ω ara ) for the upwelling months, indicating an inﬂuence of the Subtropical Under Water mass (SAW) in the PNNT coral community. Signiﬁcant negative correlations between net calciﬁcation with temperature and Ω ara , which indicates a positive response of M. auretenra with the upwelling conditions, thus, acting as “enhancer” of resilience for coral calciﬁcation.


Introduction
Coral reefs are one of the most vulnerable marine ecosystems to anthropogenic impacts, despite being classified as a key ecosystem [1]. With calcification rates in the order of 2-10 kg CaCO 3 m −2 y −1 [2], coral reefs constitute one of the most important bioconstructions in the world, product of the net accumulation of calcium carbonate (CaCO 3 ) of organisms such as corals and calcareous algae that are fundamental of a coral reef ecosystem [2]. The rapid increase in atmospheric CO 2 concentrations is changing the conditions in the seawater chemistry with the decrease in pH [3] and the saturation states of calcium carbonate, such as aragonite (Ω ara ), which is important for calcification in different coral species [4][5][6]. It has been estimated that by the end of the 21st century, the aragonite saturation state will shoal, exposing coral reefs around the world to corrosive waters [7,8], which increases the erosion rates, significantly altering the natural balance between accretion (calcification) and dissolution of the biogenic structures [9]. Studies from global calcification models under climate change scenarios forecast a decrease in the calcification rates of coral reefs of up to 150% by the year 2100 [2,9], with important consequences for ecosystem functioning [5,7,10].
Calcification is an important physiological process in coral reef ecosystems, and depends on species-specific functional identity, which include metabolic activity and coral morphology [11]. The formation of new calcium carbonate crystals (e.g., aragonite) occurs in a controlled environment within the organic matrix or extracellular calcification medium (ECM) [12]. To promote calcification, corals are able to up-regulate the internal carbonate chemistry within the ECM to maintain a pH and CaCO 3 saturation state elevated above the external environment [13][14][15]. Corals might also utilize bicarbonate substrate instead of carbonate as a precursor for CaCO 3 crystal formation as another physiological strategy that helps boots calcification in times of stressful conditions [16]. Studies have indicated that the variability in the response and the effects of ocean acidification on marine invertebrates depends on the physiological function analyzed [17]. For example, growth and calcification appears to be one of the most affected processes across organisms, and corals have showed to be among the most vulnerable due to their response to changes in the seawater carbonate chemistry [4,[18][19][20].
Negative effects of low Ω ara on calcification have been well documented for shallowwater (<20 m) [6,19,21] and cold-water corals [22,23], despite evidence of adaptation of some coral species [24][25][26]. Although most of the experimental work has been performed in controlled laboratory conditions, studies in CO 2 vents zones [24,27], tidal inshore reefs [19,25] and upwelling zones [28][29][30] provide good examples of natural laboratories for evaluating the impact of multiple co-varying factors, which will help understand the response in hypothetically future scenarios. For example, within upwelling areas, the carbonate chemistry, such as pH and aragonite saturation states, changes in response to oceanographic conditions [29][30][31][32]. In fact, large upwelling areas (Eastern Pacific) currently create ocean acidification (OA) scenarios with changes in pH of −0.1 and Ω ara of −0.4 [33,34]. Moreover, there is evidence that some upwelling conditions in the Tropical Eastern Pacific (TEP) could be limiting reef development by increasing bioerosion rates and decreasing coral growth and calcification [34][35][36].
The shoreline of the southwestern Caribbean experiences variable and climatic and oceanographic conditions due to localized coastal upwelling [37][38][39]. Within the northern coast of the Colombian Caribbean, the Guajira upwelling system occurs seasonally between January and April due to the strengthening of the trade winds, where waters from~50 to 100 m come to the surface with a decrease in mean temperature and pH [40][41][42]. Nutrients have been reported as low, compared to other upwelling systems [40,42]. These changes in environmental parameters are creating unique conditions in terms of water chemistry that could be affecting the coral communities that growth there. Therefore, this study aimed to answer two key questions: (1) How does carbonate chemistry vary in a reef area subjected to seasonal upwelling and (2) what is the physiological response of scleractinian corals to changes in the physical-chemical dynamics of the water column? To answer these questions, we generated a baseline of the carbonate system in the shallow reef area (13 m), as well as net calcification rates through a 6-month period in a keystone coral species in the study area.

Materials and Methods
Study area: This study was performed in Chengue bay within the marine protected area of the Tayrona National Natural Park (PNNT), Colombia-southwestern Caribbean Sea (11 • 20 N, 74 • 08 W) ( Figure 1). The area is a small exemplary bay inlet (~3.3 km 2 ), relatively narrow, that has a diverse community of corals with coastal lagoons, marine seagrasses dominated by Thalassia testudinum and mangrove forest composed mainly of Rhizophora mangle [43]. The total percentage of live coral is about 26%, with 33% of algal cover, which has remained stable for the last decade [44]. Coral growth is restricted to a narrow band normally not exceeding 30 m. The wave protected area of the bay is dominated by species such as Orbicella spp., Montastraea cavernosa, Pseudodiploria strigosa and Porites spp., as well as extensive mono specific patches of Madracis auretenra. The wave exposed areas are dominated mainly by Acropora palmata stands [45,46]. Porites spp., as well as extensive mono specific patches of Madracis auretenra. The wave exposed areas are dominated mainly by Acropora palmata stands [45,46]. The area is characterized by contrasting climatic and oceanographic conditions, mainly influenced by the northeast trade winds that cause a dry season between December and April when the winds are strong, alternated with a minor rainy season between the months of April to May and a major one between September and November [47]. In the dry season, the PNNT is influenced by an upwelling process with waters from the Subtropical Under Water (SUW) coming from depths of approximately 50 to 100 m [41,42]. This oceanographic process creates relatively cold waters (~23 °C), with increased salinity (36 psu) and lower pH (~7.9) [47,48].
Species description: Madracis auretenra [49] is a common zooxanthellate coral from coral reefs of the greater Caribbean Sea [50]. This species grows as dense aggregations of long, thin, dendrite-like branches, which can cover extensive areas on the substrate, such as in Santa Marta and Tayrona Park area, or grow in the form of isolated semi-hemispheric colonies [49]. This species has a wide bathymetric distribution that ranges from 2 to 60 m deep [49]. In Chengue bay, it is found distributed in the internal protected areas at a depth between 5 and 13 m [46], and its growth covers a large portion of the benthic substrate, forming structures of several square meters with almost 100% coverage (Figures S1 and S2).
In situ experimental set-up: Calcification rates of coral fragments were quantified with the alkalinity anomaly technique using custom-made incubation acrylic chambers (volume ~800 mL. Table S1). These chambers were placed in situ and secured to polyvinyl plastic quadrants (40 × 30 cm) suspended in the water column at 70 cm from the substrate with the help of foams placed on the sides of the structure, providing positive buoyancy. On each corner of the quadrant, a plastic chain attached to a lead weight kept the structure The area is characterized by contrasting climatic and oceanographic conditions, mainly influenced by the northeast trade winds that cause a dry season between December and April when the winds are strong, alternated with a minor rainy season between the months of April to May and a major one between September and November [47]. In the dry season, the PNNT is influenced by an upwelling process with waters from the Subtropical Under Water (SUW) coming from depths of approximately 50 to 100 m [41,42]. This oceanographic process creates relatively cold waters (~23 • C), with increased salinity (36 psu) and lower pH (~7.9) [47,48].
Species description: Madracis auretenra [49] is a common zooxanthellate coral from coral reefs of the greater Caribbean Sea [50]. This species grows as dense aggregations of long, thin, dendrite-like branches, which can cover extensive areas on the substrate, such as in Santa Marta and Tayrona Park area, or grow in the form of isolated semi-hemispheric colonies [49]. This species has a wide bathymetric distribution that ranges from 2 to 60 m deep [49]. In Chengue bay, it is found distributed in the internal protected areas at a depth between 5 and 13 m [46], and its growth covers a large portion of the benthic substrate, forming structures of several square meters with almost 100% coverage ( Figures S1 and S2).
In situ experimental set-up: Calcification rates of coral fragments were quantified with the alkalinity anomaly technique using custom-made incubation acrylic chambers (volumẽ 800 mL. Table S1). These chambers were placed in situ and secured to polyvinyl plastic quadrants (40 × 30 cm) suspended in the water column at 70 cm from the substrate with the help of foams placed on the sides of the structure, providing positive buoyancy. On each corner of the quadrant, a plastic chain attached to a lead weight kept the structure fixed on the substrate, preventing its displacement, but allowing movement from the current as "pendulant table-type" design ( Figure S1). This allowed natural circulation within the chambers, in order to avoid the formation of diffusion boundary layers around the organisms and to provide the homogeneous distribution [51,52]. The structure was located on a sandy bottom adjacent to the coral patch of M. auretenra at a depth of 13 m.
Carbonate chemistry analysis: Using SCUBA, discrete water samples were taken with Niskin bottles (5 L) just above the coral patch where the fragments were collected and near to the experimental set-up. Salinity and temperature data were taken at the time of water sampling with a WTW-3630 IDS conductivity-temperature probe (salinity accuracy ± 0.1; temperature accuracy ± 0.1 • C, resolution 0.1 • C). These samples were used as an initial time measurement (t i ) of total alkalinity (A T ) and dissolved inorganic carbon (DIC) in order to 1) generate the baseline of the water chemistry and 2) to obtain calcification rates of the fragments incubated inside the acrylic chambers. Water samples were collected following standard protocols [53] in 450 mL borosilicate bottles, sealed with Apiezon ® grease and fixed with 100 µL of saturated mercuric chloride solution (HgCl 2 ) to prevent changes in carbonate chemistry due to biological activity [53] and stored in dark at~21 • C for later analysis for A T and DIC, which were performed within two to four weeks of sample collection.
A T was determined by open cell potentiometric titration following standard protocols [54] using hydrochloric acid as titrant (0.1 mol L −1 HCl buffered in 0.6 mol L −1 NaCl, CRM Dickson Lab, Scripps Institution of Oceanography, La Jolla, California, USA). For DIC analysis, we used an automated AIRICA/Li-Cor 7000 analyzer from Marianda ® following standard protocols [54]. The carbonate system was derived from total alkalinity (A T ) and dissolved inorganic carbon (DIC) using the Excel-CO2SYS program [55], taking as parameters the dissociation constant for carbonic acid from Mehrbach et al. [56], refit by Dickson & Millero [57], the constants K1 and K2 from Leuker et al. [58], KHSO 4 by Dickson [59] and Boron by Lee et al. [60]. These two parameters were taken as input together with temperature and salinity data to characterize pH T (total scale), partial pressure of CO 2 in the system (µatm pCO 2 ), bicarbonate (HCO 3 -), carbonate concentration (CO 3 2− ) and aragonite saturation states (Ω ara ). To determine the accuracy of the analysis, the samples were verified with certified reference material (CRM) for A T and DIC, which were <0.5% estimated error (A T : ±10 µmol kg −1 SW and DIC: ±3 µmol kg −1 SW) (Batch 171 and 190, CRM Dickson Lab, Scripps Institution of Oceanography, La Jolla, CA, USA).
Net calcification rates: To derive calcification rates from total alkalinity [61,62], four (4) sets of measurements were performed in situ from December 2020 to May 2021 (Tables 1 and S2). Each set was composed of two (2) field-work days with a total of six (6) incubations performed per day in the following way: day 1: incubations for M. auretenra; day 2: incubations with skeletal fragments (coral skeletons) that were carried out to determine potential dissolution rates of coral fragments only on February and March 2020, which accounts for the transition to upwelling and upwelling months. Fragments of M. auretenra with an average volume of 39 ± 7 cm 3 were collected on the reef between 9 and 12 m depth, and brought to the surface where they were fixed to acrylic bases with epoxy putty. The fragments on acrylic bases were then taken to the site of the experimental set-up and randomly introduced in each chamber to carry out the subsequent controlled incubations. In average, each incubation trial lasted 2.48 ± 0.43 h (Table S1). Temperature was measured in situ at 1-min intervals with a HOBO ® sensor (Onset ® ) pendant UA-002 for the length of the trials (precision 0.14 • C, accuracy ± 0.53 • C). For each set of measurements, a control chamber without coral fragment was also incubated in order to determine possible changes in the water chemistry not related to calcification processes (e.g., microbial activity). These control values were within the range of −5 and 23 µmol kg −1 of seawater for total alkalinity, with respect to environmental samples, which were subtracted from the A T of the chambers with coral fragments. Following the incubation time, the chambers were retrieved from the experimental setup and brought to the surface one by one in the same order they were closed for subsequent sample collection (water and coral fragment). Once on surface, each chamber was opened and measurements of surface temperature and salinity were obtained using a WTW-3630 IDS multiparameter probe. Subsequently, the water was transferred to borosilicate bottles and fixed with HgCl 2 for later analysis of A T and DIC in the laboratory (see methods in carbonate chemistry analysis). Calcification rates (Gnet) were quantified based on changes in total alkalinity (A T ) before and after an incubation period [61,62] using the formula: where ∆AT is the change in total alkalinity over the course of incubation, ρ is the density of water (kg L −1 ), V is the chamber volume (liters), t is the incubation time (hours) and SA is the surface area of the coral fragments. Positive values of Gnet indicate net calcification while negative values indicate net dissolution. Net dissolution rates were measured for the transition pre-upwelling (February) and upwelling (March) using a total o 10 incubations with bare coral skeletons of the studied species with an average incubation time of 3.5 ± 0.20 h. Statistical analysis: To test for the calcification response of M. auretenra in the different environmental conditions caused by the seasonal upwelling, we used the standardized calcification rates (µmol CaCO 3 m −2 h −1 ) as the response (dependent) variable and seasons were used as independent variables using one-factor ANOVA model. The responses of individual coral fragments were grouped within each seasonal period as follow: nonupwelling-transition pre-upwelling-upwelling-transition post-upwelling. A Shapiro-Wilk test was used to test for the assumptions of normality (p = 0.112), and Levene's test for homoscedasticity (p = 0.104), which validated the assumptions of the ANOVA model. Further, to investigate the relationship between temperature and carbonate chemistry on the calcification response of M. auretenra, we calculated the Pearson correlation and performed regression analyses using the least-square method. For these analyses, assumption for normality of standardized residuals was checked on a histogram plot, as well as on a regression vs. the standardized predicted value to check for autocorrelation of residuals and heteroscedasticity. Results were reported as mean ± standard deviation (±s.d.) and for all analyses, p < 0.05 was considered statistically significant. All statistical tests were run using R (v.4.0.3) and SPSS v.22.

Environmental Conditions and Seawater Carbonate Chemistry at the Sampling Site
Temperature and salinity showed variations between the sampling months. We observed a decrease in the mean temperature from 28.27 ± 0.11 • C for the month of December (non-upwelling) to 24.75 ± 0.17 • C corresponding to March 2021 (upwelling) (Table 1, Figure 2). This decrease in temperature was coincident with increase in wind velocity within the area due to the north-west trade winds (Figure 3). Likewise, we observed an increase in salinity from 33.49 ± 0.21 in December to around 36 for the following months (Table 1).

Environmental Conditions and Seawater Carbonate Chemistry at the Sampling Site
Temperature and salinity showed variations between the sampling months. We observed a decrease in the mean temperature from 28.27 ± 0.11 °C for the month of December (non-upwelling) to 24.75 ± 0.17 °C corresponding to March 2021 (upwelling) (Table 1, Figure 2). This decrease in temperature was coincident with increase in wind velocity within the area due to the north-west trade winds (Figure 3). Likewise, we observed an increase in salinity from 33.49 ± 0.21 in December to around 36 for the following months (Table 1).  Seawater chemistry values showed variation during the sampling months, particularly for total alkalinity and DIC that increased approximately 2.5% during the upwelling season from 2371 ± 8 to 2419 ± 5 µmol kg −1 SW and 2063 ± 29 to 2124 ± 17 µmol kg −1 SW, respectively. Significant correlations were found between temperature and AT (Pearson: −0.88, R 2 : 0.78, p < 0.01), temperature and DIC (Pearson: −0.78, R 2 : 0.61, p < 0.01), and temperature and bicarbonate ion (HCO3 -) (Pearson: −0.71, R 2 : 0.50, p < 0.05). Although, we did not find any significant correlations between temperature and the other seawater chemistry parameters, we observed that average Ωara decreased from 3.82 ± 0.24 to 3.34 ± 0.23

Environmental Conditions and Seawater Carbonate Chemistry at the Sampling Site
Temperature and salinity showed variations between the sampling months. We observed a decrease in the mean temperature from 28.27 ± 0.11 °C for the month of December (non-upwelling) to 24.75 ± 0.17 °C corresponding to March 2021 (upwelling) (Table 1, Figure 2). This decrease in temperature was coincident with increase in wind velocity within the area due to the north-west trade winds (Figure 3). Likewise, we observed an increase in salinity from 33.49 ± 0.21 in December to around 36 for the following months (Table 1).  Seawater chemistry values showed variation during the sampling months, particularly for total alkalinity and DIC that increased approximately 2.5% during the upwelling season from 2371 ± 8 to 2419 ± 5 µmol kg −1 SW and 2063 ± 29 to 2124 ± 17 µmol kg −1 SW, respectively. Significant correlations were found between temperature and AT (Pearson: −0.88, R 2 : 0.78, p < 0.01), temperature and DIC (Pearson: −0.78, R 2 : 0.61, p < 0.01), and temperature and bicarbonate ion (HCO3 -) (Pearson: −0.71, R 2 : 0.50, p < 0.05). Although, we did not find any significant correlations between temperature and the other seawater chemistry parameters, we observed that average Ωara decreased from 3.82 ± 0.24 to 3.34 ± 0.23 Seawater chemistry values showed variation during the sampling months, particularly for total alkalinity and DIC that increased approximately 2.5% during the upwelling season from 2371 ± 8 to 2419 ± 5 µmol kg −1 SW and 2063 ± 29 to 2124 ± 17 µmol kg −1 SW, respectively. Significant correlations were found between temperature and A T (Pearson: −0.88, R 2 : 0.78, p < 0.01), temperature and DIC (Pearson: −0.78, R 2 : 0.61, p < 0.01), and temperature and bicarbonate ion (HCO 3 -) (Pearson: −0.71, R 2 : 0.50, p < 0.05). Although, we did not find any significant correlations between temperature and the other seawater chemistry parameters, we observed that average Ω ara decreased from 3.82 ± 0.24 to 3.34 ± 0.23 with values dropping as low as 3.07 during upwelling season. Despite these changes, other parameters in the carbonate chemistry showed little seasonal variation (Table 2). Table 2. Environmental variables, temperature in Celsius (Temp • C) and salinity. Total alkalinity before the incubations (A T (i)), total alkalinity after the incubations (A T (f)), net calcification (Gnet) and net dissolution. Measurements denoted by [--.--] were not performed in those dates. Values are given as mean ± s.d.

Discussion
Localized upwelling events caused changes in net calcification of Madracis auretenra due to natural variability in environmental conditions commonly experienced in the study area. Mean temperature changes of ~5 °C (from 29 °C to 24 °C) and Ωara of ~0.5 (from 3.82 to 3.34) characterized the upwelling, which came by pulses from January through March with clear periods of relaxation/intensification mainly driven by the strengthening of the

Discussion
Localized upwelling events caused changes in net calcification of Madracis auretenra due to natural variability in environmental conditions commonly experienced in the study area. Mean temperature changes of~5 • C (from 29 • C to 24 • C) and Ω ara of~0.5 (from 3.82 to 3.34) characterized the upwelling, which came by pulses from January through March with clear periods of relaxation/intensification mainly driven by the strengthening of the trade winds (>10 m s −2 ). From mid-December to mid-February 2020, we observed a period of transitioning to upwelling, then an intensification during February to March and transitioning to non-upwelling conditions again from mid-April to May 2020. As seen in the present study and in accordance with other studies using satellite images of SST (AVHRR) [37], the main cooling event occurs during January-April, with 2 to 4 upwelling pulses and a maximum detected during February-March.
Upwelling and its influence in the carbonate chemistry dynamics: Upwelling conditions in the Colombian Caribbean during the dry season include cold surface waters (21-24 • C), high salinity (>35 ups) and oxygen saturation <90% in the surface layers [41,42,63]. We observed temperatures~28-29 • C and salinity~34 for December when the upwelling had not happened, which indicates the presence of a stratified water column of the Caribbean Surface Waters (CSW) within the area [41]. This water mass is characteristic at depths shallower than 30 m predominant in the non-upwelling months (e.g., December in this study) [41]. For the months that followed the transition and peak upwelling conditions, temperature dropped to 24 • C and salinity increased to~36. These conditions were created by the influence of the Subtropical Underwater Mass (SUW) that upwell from a depth of about 50 to 100 m, and is characteristic of the Guajira upwelling system [41,42]. Andrade and Barton [40] suggest that the subsurface waters that supply the upwelling system in La Guajira may be influenced by a coastal branch from the Panama-Colombia Gyre, whose surface origin and moderately young waters determines its physical-chemical characteristics, as well as its lower nutrient concentration (e.g., [48]) compared to other upwelling centers within the southern Caribbean upwelling systems [37].
We observed an increase in total alkalinity (A T ) and dissolved inorganic carbon (DIC) during upwelling months. Although, these results are baseline data measured for the Tayrona area, we found that are within the reported values for a similar upwelling system (Cariaco Basin, Venezuela) that has been subject of different studies focused on the carbonate chemistry [64][65][66]. During the peak upwelling (February-March) we found values > 2400 µmol kg −2 for A T and >2100 µmol kg −2 for DIC, which explained the chemical nature of the upwelling and the SUW in the area [66,67]. The little variation in pCO 2 (from 453 to 456) and pH T (~8.01) may be related to the relatively shallow-water mass that upwell within the study area, which may not create considerable changes in these carbonate chemistry variables due to its young age [40]. Other factors associated to metabolic pulses (respiration-photosynthesis and calcification-dissolution) and water mass residence time within shallow-water coral reef areas, have the ability to influence the local carbonate chemistry creating a "buffer" effect raising the levels of CO 2 and pH T [68,69]. It is important to note that discrete water samples for carbonate chemistry measurements were taken just above the reef (~1.5 m) and not at the surface or at different depths in the water column; thus, it is necessary to carry out specific studies to be able to answer these questions related to buffer effect in this particular reef system.
Calcification under seasonal upwelling: Seasonal differences in temperature and Ω ara had a main effect on the calcification rate of M. auretenra. Net calcification increased by 42% under upwelling when temperature and Ω ara were lower with average values of 41.8 ± 11.7 µmol CaCO 3 m −2 h −1 for upwelling season compared to 23.8 ± 7.3 µmol CaCO 3 m −2 h −1 for non-upwelling. Both variables, temperature and Ω ara , have been central in the study of climate change, since both have the capacity to alter metabolic functions [21]. Although comparisons with other studies are variable and depends on multiple factors, such as type of coral growth and method employed to measure calcification, as well as community vs. single species approach [11,70], our results indicate that the average net calcification rates for ambient conditions are in agreement with other studies using similar approach and working with the same coral species [16].
When comparing physiological response of marine calcifiers (calcification rates, metabolic functions) with other studies performed in other upwelling areas (e.g., eastern Pacific, northern Arabian Sea), there is the consensus that changes in the carbonate chemistry variables such as decreased temperature, pH T and Ω ara , exert strong effect on functional performance [29][30][31]. For example, seasonal upwelling in the Indian Ocean, which had values of Ω ara ranging from~4 to~3.2 [30], exert a negative response in the extension and calcification rates of Porites sp., similar to what has been found on Pocilloporid corals from the Gulf of Panama [71]. One plausible explanation is that elevated primary productivity in upwelling season is causing a loss in photosynthetic efficiency that is reflected in the decrease in calcification rates and linear extension [30], which might also apply to corals in the Tayrona area, where nutrient concentrations are lower than other more productive upwelling systems [40,48].
We found that calcification rates had a negative function with respect to temperature and Ω ara , despite different studies performed to date that have concluded that calcification is a positive function of temperature and carbonate ion concentration (CO 3 2− ) [6,9,[72][73][74][75][76]. The upwelling-induced counteractive mechanism observed in the present study in the calcification of M. auretenra might be explained in part from the holobiont internal carbonate chemistry control, as has been shown in other studies working with tropical scleractinian corals. As oceans become more acidic, mechanisms of coral calcification to counteract the carbonate depletion involve pH up-regulation, in which corals are able to control the internal chemistry of the calcifying space [13], and switching from CO 3 2to HCO 3 as the primary substrate for CaCO 3 deposition [16,77,78]. Although we found a significant positive correlation between net calcification and bicarbonate ion concentration [HCO 3 -] (Pearson: 0.61, R 2 = 0.37, p < 0.01), the relative contribution of increased bicarbonate at constant CO 3 2was not measured in the present study. One study performed on M. auretenra, indicates that this species responds more to [HCO 3 -] than to [CO 3 2− ], thus giving this species resistance to OA [16]. To date, this is the only evidence, despite other studies having found opposite results [74,76]. Moreover, studies using in situ approach indicated that calcification rates are dependent on the Ω ara , with approximately 52% reduction per unit decrease in saturation state, even in coral species considered tolerant to changes in environmental conditions [79]. Another scenario with counteractive response will favor coral calcification and growth due to increments in DIC that has been shown to enhance photosynthetic efficiency, thus physiological performance [80].
Ecological implications: Global models for ocean acidification indicate a decrease in the saturation state of calcium carbonate, particularly aragonite, important precursor for coral calcification [7], and can be more accentuated in areas naturally experiencing changing environmental conditions such as upwelling systems [33]. However, the success and persistence of M. auretenra in the study area is an example of adaptability of a scleractinian coral to natural variability in ecological and physiological conditions. Over the course of this study, changes in seawater chemistry, particularly Ω ara at Chengue bay, remained adequately to marginally saturated (from~3.8 to~3.3), the pH T remained above 8.00 for >95% of the time except one measurement performed in March that we recorded 7.96 (Table 2), and pCO 2 was within the average for current ambient atmospheric values (420-450 µatm). This means that a moderate upwelling such as the one in PNNT benefits the coral community and could potentially act as refuge for climate change in general. For example, ecological studies performed within the Tayrona area indicated that bleaching susceptibility of corals is lower than in other areas of coral reefs in the South-Western Caribbean due, presumably, to the cooling effect caused by the upwelling processes [48]. Moreover, when bleaching was relatively high within the area, corals recovered quickly during the upwelling, indicating local resilience against climate change stressor [48,81,82]. Ecological restoration studies have been modelling artificially-generated upwelling conditions to see if it can effectively reduce sea surface temperature (SST) and degree heating weeks (DHW), and slow future coral bleaching events [83,84]. Future research should evaluate the calcification response in different climate change scenarios in order to test hypothesis concerning adaptability and resilience not only of M. auretenra, but also comparing it with conspecifics from the nearby reefs.

Conclusions
This study showed that variables of the carbonate chemistry system varied according to seasonal upwelling. These results represent a key baseline input for future studies on carbonate chemistry and ocean acidification for the area. Differences were found in surface temperature, salinity and some variables of the carbonate system such as A T , DIC, HCO 3 -and Ω ara . The other parameters remained relatively homogeneous during the sampling period, such as pCO 2 and pH T , opposite to the patterns expected for an upwelling system. The pattern may be explained by the physical nature of the upwelling, which emerges from a shallow water mass (50-100 m) on the Colombian Caribbean coast. Net calcification (Gnet) of M. auretenra, which increased for the upwelling period, was significantly correlated to temperature and Ω ara . It is worth pointing out that there are still covarying factors affecting these correlations, since about 62% and 79% other than temperature and aragonite, respectively, are affecting calcification rates. Nevertheless, these results still indicate negative feedback in the physiological response where M. auretenra is able to control calcification to counteract changes in temperature and carbonate chemistry. Many questions remain unanswered associated with different physical (geomorphology, residence time and water exchange within the bay) as well as biological factors such as metabolic pulses (respiration/photosynthesis-calcification/dissolution) that can have an external buffer effect on the reef. Moreover, we did not differentiate between the relative effects of day and night calcification vs. respiration and photosynthesis, which will account for the calcium carbonate budget that should be included in future studies.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/oceans4020012/s1, Figure S1. Experimental design in Chengue bay. (A) incubation chambers within the pendulant table-type design, (B) fragment close-up showing the natural appearance of the extend polyps during incubation trials, and (C) deployment of the chamber by divers. Figure S2. Coral reef in Chengue bay where the main growth of M. auretenra were found. Table S1. Raw data for each of the incubation fragments (replicates). Temperature (Temp) in celcius, salinity (Sal) in practical salinity units (PSU), chamber volume (net), fragment volume, depth (meters) and incubation time (minutes). Table S2. Raw data for each of the incubation fragments (replicates). Total alkalinity (A T -µmol CaCO 3 m 2 h −1 ) initial (i) and final (f), dissolved inorganic carbon (DIC-µmol CaCO 3 m 2 h −1 ) initial (i) and final (f), net calcification (G-µmol CaCO 3 m 2 h −1 ) and net dissolution (µmol CaCO 3 m 2 h −1 ).