Combined effects of global climate change and nutrient enrichment on the physiology of three temperate maerl species

Abstract Made up of calcareous coralline algae, maerl beds play a major role as ecosystem engineers in coastal areas throughout the world. They undergo strong anthropogenic pressures, which may threaten their survival. The aim of this study was to gain insight into the future of maerl beds in the context of global and local changes. We examined the effects of rising temperatures (+3°C) and ocean acidification (−0.3 pH units) according to temperature and pH projections (i.e., the RCP 8.5 scenario), and nutrient (N and P) availability on three temperate maerl species (Lithothamnion corallioides, Phymatolithon calcareum, and Lithophyllum incrustans) in the laboratory in winter and summer conditions. Physiological rates of primary production, respiration, and calcification were measured on all three species in each treatment and season. The physiological response of maerl to global climate change was species‐specific and influenced by seawater nutrient concentrations. Future temperature–pH scenario enhanced maximal gross primary production rates in P. calcareum in winter and in L. corallioides in both seasons. Nevertheless, both species suffered an impairment of light harvesting and photoprotective mechanisms in winter. Calcification rates at ambient light intensity were negatively affected by the future temperature–pH scenario in winter, with net dissolution observed in the dark in L. corallioides and P. calcareum under low nutrient concentrations. Nutrient enrichment avoided dissolution under future scenarios in winter and had a positive effect on L. incrustans calcification rate in the dark in summer. In winter conditions, maximal calcification rates were enhanced by the future temperature–pH scenario on the three species, but P. calcareum suffered inhibition at high irradiances. In summer conditions, the maximal calcification rate dropped in L. corallioides under the future global climate change scenario, with a potential negative impact on CaCO3 budget for maerl beds in the Bay of Brest where this species is dominant. Our results highlight how local changes in nutrient availability or irradiance levels impact the response of maerl species to global climate change and thus point out how it is important to consider other abiotic parameters in order to develop management policies capable to increase the resilience of maerl beds under the future global climate change scenario.

The physiological response of maerl to global climate change was species-specific and influenced by seawater nutrient concentrations. Future temperature-pH scenario enhanced maximal gross primary production rates in P. calcareum in winter and in L. corallioides in both seasons. Nevertheless, both species suffered an impairment of light harvesting and photoprotective mechanisms in winter. Calcification rates at ambient light intensity were negatively affected by the future temperature-pH scenario in winter, with net dissolution observed in the dark in L. corallioides and P. calcareum under low nutrient concentrations. Nutrient enrichment avoided dissolution under future scenarios in winter and had a positive effect on L. incrustans calcification rate in the dark in summer. In winter conditions, maximal calcification rates were enhanced by the future temperature-pH scenario on the three species, but P. calcareum suffered inhibition at high irradiances. In summer conditions, the maximal calcification rate dropped in L. corallioides under the future global climate change scenario, with a potential negative impact on CaCO 3 budget for maerl beds in the Bay of Brest where this species is dominant. Our results highlight how local changes in nutrient availability or irradiance levels impact the response of maerl species to global climate change and thus point out how it is important to consider other abiotic parameters in order to develop management policies capable to increase the resilience of maerl beds under the future global climate change scenario.

| INTRODUC TI ON
The coastal system is under severe threat and is considered as one of the most vulnerable environments due to the strong influence of global and local anthropogenic pressures. Global climate change is expected to affect abundance, diversity, and productivity of marine populations (Barange & Harris, 2003), thus becoming a major driver of the future state of marine ecosystems (Duarte et al., 2013).
Atmospheric carbon dioxide (CO 2 ) concentrations have risen from 278 parts per million (ppm) at the start of the Industrial Revolution to the current level of 410 ppm. Approximately, one-third of CO 2 emissions are absorbed by the ocean and affect its chemistry and physics (Gattuso et al., 2015), causing a global average increase in temperature of 0.8°C and a decline in pH of 0.1 units of the surface oceans since the industrial revolution, phenomena known as ocean warming and ocean acidification (OA), respectively (Gattuso et al., 2015). Marine coastal ecosystems are also influenced by other, more local anthropogenic pressures, which are growing due to various human activities in coastal areas. Considering this, coastal eutrophication is attributed to the enrichment of nutrients such as phosphates and nitrates resulting from agricultural runoff or wastewater discharged into the sea via rivers (Diaz-Pulido & McCook, 2008;Duarte et al., 2013;Russel & Connell, 2009;Salisbury, Green, Hunt, & Campbell, 2008;Strong, Kroeker, Teneva, Mease, & Kelly, 2014).
Global climate change is expected to alter dominance relationships among primary producers, in particular between calcifying and noncalcifying macroalgal species Celis-Plá et al., 2015;Falkenberg, Connell, & Russel, 2013). However, these changes will depend on local factors such as nutrient availability (Celis-Plá et al., 2015;Fabry, Seibel, Feely, & Orr, 2008). Red calcareous coralline macroalgae are among the most sensitive organisms to OA (Martin & Hall-Spencer, 2017) because they precipitate high-magnesium calcite, which is the most soluble form of biogenic calcium carbonate (CaCO 3 ) (Morse, Andersson, & Mackenzie, 2006). Among them, free-living nongeniculate red coralline algae, called rhodoliths or maerl, are distributed in coastal ecosystems throughout the world (Foster, 2001) and experience different levels of pH and temperature changes depending on their location (Qui-Minet et al., 2018). Previous studies on maerl and other red coralline algae have given insights into their individual responses to OA and warming, which are useful for a better understanding of how species physiology responds to these factors alone or in combination. Although the responses of coralline algae are species-specific (Martin, Charnoz, & Gattuso, 2013;Noisette, Egilsdottir, Davoult, & Martin, 2013;Vazquez-Elizondo & Enriquez, 2016), most studies have revealed adverse effects of the combination of ocean acidification and warming on coralline algal physiology (Martin & Hall-Spencer, 2017).
Owing that productivity is limited by nutrient availability, moderate nutrient enrichment can benefit algae by reducing nutrient limitation and allowing them to cope with the metabolic cost of adapting to global climate change (Celis-Plá et al., 2015;Hofmann, Heiden, Bischof, & Teichberg, 2014). However, nutrient enrichment can threaten maerl survival and favor the development of fleshy epiphytic macroalgae Steller, Riosmena-Rodrίguez, Foster, & Roberts, 2003). Research focusing on the effect of nutrient enrichment under current conditions of pH and temperature in coralline algae has reported that increased nitrate concentrations do not affect photosynthesis and calcification (Belliveau & Paul, 2002;Bjork, Mohammed, Bjorklund, & Semesi, 1995). However, phosphates have been described as "crystal poison" since they have shown to act as an inhibitor of calcification and growth of coralline algae (Bjork et al., 1995;Kinsey & Davies, 1979;Simkiss, 1964). Studies specifically addressing the interaction of global climate change with nutrient enrichment on the physiology of coralline algae are scarce, and none have explored the response of maerl species.
The purpose of this study was to test (a) the individual and combined effects of global climate change (ocean warming and acidification) and local nutrient loading on various species of maerl and (b) if different species coming from the same location respond similarly due to convergent adaptation. We considered present temperature-pH conditions and future temperature-pH conditions projected for 2,100 according to the representative concentration pathway RCP 8.5 global change scenario (Gattuso et al., 2015) as well as enrichment in nitrate (NO 3 − ) and phosphate (PO 4 3− ), the main nutrients originating from human activities affecting the Bay of Brest (Le Pape & Menesguen, 1997).
To consider the potential seasonal variability of the response of maerl to OA and warming (Martin, Cohu, Vignot, Zimmerman, & Gattuso, 2013;Martin & Gattuso, 2009), we carried out our study under winter and summer conditions. We hypothesized that (a) the future scenario of ocean warming and acidification will negatively affect calcification in the three maerl species; (b) nutrient enrichment will exacerbate the impact of global climate change on the three species; and (c) although the three maerl species coexist in the same location and hence are adapted to the same environmental conditions, species-specific responses occur due to divergent ecological traits such as morphology and pigment concentrations.

| Collection site
The three maerl species, L. corallioides, P. calcareum, and L. incrustans ( Figure 1), were collected in the Roz maerl bed (48°19′58″N, 04°19′57″W), located in the southern basin of the Bay of Brest (Brittany, France), where it covers a surface of 1.4 km 2 (Qui-Minet et al., 2018). The three here studied species are present in different abundances, being L. corallioides the most abundant and the only one distributed along the Bay of Brest.
Temperature and pH in the Bay of Brest average 10°C and 8.04 in the winter, and 18°C and 8.00 in the summer, respectively. Local nutrient conditions are strongly variable according to the season as reported by Le Pape andMenesguen (1997), andQui-Minet et al. (2018); nitrate concentrations are significantly higher in winter (±30 μmol/L) relative to summer (<1 μmol/L) while phosphate concentrations remain below 1 μmol/L during both summer and winter seasons. Incident irradiance in the Bay of Brest varies with tides. In the collection site, it can vary from <1 to 320 μmol photons m −2 s −1 in winter and from 1 to 645 μmol photons m −2 s −1 in the summer (Qui-Minet et al., 2018). The Roz bed is located at 0.7 m depth (chart datum), with maximal water depth amplitudes of 8 m (Daniel, 1995).
Physicochemical parameters also vary on a daily basis, in the summer in the Roz bed, temperature and pH can observe a mean variation of 0.7°C and 0.08 pH units, respectively (Qui-Minet et al., 2018).

| Biological material
Living specimens of the three species were collected with a 0.1 m 2 Van Veen grab in January 2016 (winter experiment) and in June 2016 (summer experiment). The algae were brought to the laboratory in a box filled with seawater (seawater temperature did not differ much from air temperature). After cleaning, thalli were randomly distributed into the aquaria of the experimental system and kept under in situ conditions of temperature and pH. In each aquarium, there were two sets of 5-7 g dry weight (DW) of L. corallioides, 5-7 g (DW) of P. calcareum, and 7-9 g (DW) of L. incrustans attached with a nylon wire and labeled with small plastic numbers and additional unlabeled thalli (approximately 10 g per species).
F I G U R E 1 Specimens of (a) Lithothamnion corallioides, (b) Phymatolithon calcareum, and (c) Lithophyllum incrustans collected in the Roz maerl bed in the Bay of Brest (Brittany, France) (photos Coralie Delaunay). Scale bars = 1 cm

| Experimental conditions and setup
Two experiments were conducted for both winter (January-April For the tested present temperature-pH conditions, temperature and pH were determined from mean values measured in situ in winter and summer (temperature: 10 and 18°C, respectively; pH on the total scale (pH T ): 8.04 and 8.00, respectively) (Qui-Minet et al., 2018). Acclimation to the experimental system under controlled conditions lasted 3 weeks. For the tested projected future temperature-pH conditions, values were selected according to the future RCP 8.5 scenario (+2.7°C and −0.33 pH units; Gattuso et al., 2015).
Temperature and pH were progressively increased and decreased by ~0.4°C and ~0.05 pH units, respectively, every 2 days during 2 weeks. An experimental thermostated system was set up for both temperature-pH scenarios (present and future). Each experimental system consisted of a thermostated bath containing six aquaria (10 L), three with nutrient enrichment and three without nutrient enrichment (12 aquaria in total; Figure 2). Aquaria were supplied at a rate of 9 L/hr from a 100 L header tank continuously supplied with filtered seawater pumped in front of the Roscoff Marine Biological Station (Station Biologique de Roscoff). Due to constraining space conditions, only one header tank was used per scenario. However, this was an open-flow system with filtered seawater (5 µm) and water renewal rate (~54 L/hr) with high cleaning frequency (weekly) in the header tank to prevent any contamination event and tank effect.

| Monitoring of the seawater parameters
Temperature and pH were measured on a daily basis; total alkalinity (T A ) was measured every week, and nutrients were measured every 2 weeks. The pH values of the system were adjusted based on daily measurements of pH T and temperature in the 12 aquaria using a pH meter (HQ40D, Hach Lange, Ltd portable LDO™) calibrated with Tris/HCl and 2-aminopyridine/HCl buffers (Dickson, Sabine, & Christian, 2007 , and F I G U R E 2 Schematic diagram of the experimental laboratory setup with two scenarios: present pH and temperature and future (RPC 8.5) projected pH and temperature by 2,100. All aquaria were maintained at constant temperature in a thermostated bath. Each aquarium contained 2 sets of maerl thalli for each species Si(OH) 4 concentrations were done by spectrophotometry using a Technicon autoanalyzer according to Aminot and Kérouel (2007).
NH 4 + concentrations were determined using the Solorzano (1969) method. T A was measured by 0.01 N HCl potentiometric titration on an automatic titrator (Titroline alpha, Schott Si Analytics) calibrated on the National Bureau of Standards scale and by using the Gran method of nonlinear least-squares fit applied to pH variations from 3.5 to 3.0 mEq/L (Dickson et al., 2007).

| Metabolic rate measurements
Metabolic measurements were performed after 3 months of culture in each treatment. Net primary production (NPP) and net calcifi- . Reactive oxygen spots of the chambers were calibrated before the incubations with 0% and 100% oxygen buffers. Calcification rates were estimated using the alkalinity anomaly technique (Smith & Key, 1975), which is a good estimator for short-term incubations. It is based on a decrease of two moles total alkalinity (T A ) by two equivalents per molecule of CaCO 3 precipitated (Wolf-Gladrow, Zeebe, Klaas, Kortzinger, & Dickson, 2007). Seawater was sampled at the beginning and at the end of incubation, and analyses were performed as previously described.
Net primary production (NPP), dark respiration (R), gross primary production (GPP), and light and dark calcification (G) rates were corrected from controls and calculated as Equations (1), (2), and (3), respectively: The ammonium flux was calculated according to the following equation: where ΔO 2 and ΔT A are respectively the differences between initial and final O 2 concentrations (μmol O 2 L −1 ) and T A (μEq/L), V is the volume of the chamber (L), Δt is the incubation time (hr), and DW is the dry weight of the algae (g).
In the absence of inhibition of the net primary production and calcification by high irradiances, NPP and G versus irradiance (E, μmol photons m −2 s −1 ) curve parameters were obtained according to Platt, Gallegos, and Harrison (1980): In the presence of inhibition of the net primary production and calcification by high irradiances, NPP-E and G-E curves were established by fitting Eilers and Peeters (1988) equation: in Equations (5) and (7), R is the dark respiration rate; in Equations (6) and (8), G D is the calcification rate in the dark. GPP max and G max are the maximum rates of gross primary production and calcification (μmol O 2 or CaCO 3 g DW hr −1 ), respectively, and E k (μmol photons m −2 s −1 ) is the saturating irradiance.
In the presence of inhibition of the net primary production and calcification by high irradiances GPP max , G max , and E k were calculated from the parameters a, b, and c in Equations (7) and (8) as:

| Dry weight and CaCO 3 content in algae
After incubations, DW was measured on lyophilized maerl samples. The relative CaCO 3 content (%) was calculated from the ash weight after burning at 550°C (5 hr) and the DW of samples (g CaCO 3 g −1 DW maerl).

| Chlorophyll a content
Maerl samples for chlorophyll a analyses were stored at −80°C prior to lyophilization. Afterward, they were ground in plastic tubes with 0.5 cm stainless steel beads (Brammer) using a Tissue Lyser II (QIAGEN). Extraction was done according to Arar and Collins (1997), that is, 5 ml of 90% acetone were added to 0.05 g of maerl powder. Samples were kept in glass tubes for 12 hr in dark and 4°C. Afterward, incubation tubes were centrifuged at 3,000 g for 10 min, and the supernatant was then transferred to a new tube; fluorescence was then measured using a calibrated Turner 10-AU fluorometer.

| Statistics
Statistical analyses were performed using the open source software R version 3.5.1 (R Core Team, 2017). The chlorophyll a and CaCO 3 contents, and the metabolic rates were averaged for the two sets of thalli of each maerl species per aquarium, and the individual aquaria were thus considered as replicates (n = 3). Because of the small number of replicates per treatment, the normality of data was not verified for any of the analyses. This prevented us from using parametric analyses such as ANOVA or PERMANOVA and led us to use nonparametric analyses. Moreover, the use of rank-based tests has been recommended for small samples (Legendre & Legendre, 1998) and is often as powerful as parametric tests (Scherrer, 2007).
However, we are aware that such an approach increases the number of comparisons, and as such, there is a risk of a type I error, that is, incorrect rejection of the null hypothesis.
A two-way nonparametric ANOVA (the

| Impact of the future temperature-pH scenario and nutrient enrichment in winter conditions under ambient irradiance
Chl a content was significantly and positively affected by nutrient enrichment in the three maerl species (Appendix S1 & Figure 3a-c).
R rates in L. corallioides and L. incrustans were not significantly affected by the temperature-pH scenario or nutrient conditions (Table 2 and Figure 4a,c). R rates in P. calcareum were negatively affected by nutrient enrichment (Table 2 and Figure 4b). GPP rates did not vary significantly across treatments or species (Table 2 and   Figure 5c). G L rates did not differ significantly between treatments in L. corallioides, although they were more than two-fold lower under Fu-LN relative to the other treatments (Table 2 and Figure 6a). P. calcareum showed net dissolution under Fu-LN, but the test did not reveal any significant differences (Table 2 and Figure 6b). In L. incrustans, G L rates were significantly affected by the temperature-pH scenario with net dissolution observed under Fu-LN.
Lithothamnion corallioides and P. calcareum winter G D rates were negatively affected by the future temperature-pH scenario (Table 2), and net dissolution was observed under Fu-LN (Figure 7a,b).
Lithophyllum incrustans G D rates did not differ significantly between treatments (Table 2), although dissolution was observed under Fu-LN (Figure 7c).
In L. incrustans, it ranged from −0.03 ± 0.10 (Fu-LN) to 0.33 ± 0.00 (Pr-HN) and was negatively affected by the future temperature-pH scenario ( Table 2).  Note: Comparisons among seasons were done using a one-way nonparametric (Kruskal-Wallis) test. Analyses significant at the α = .025 level are indicated by asterisks.
Values in bold refer to parameters that are significantly different.

| Impact of the future temperature-pH scenario and nutrient enrichment in summer conditions under ambient irradiance
Lithothamnion corallioides and P. calcareum Chl a content was affected neither by the temperature-pH scenario nor by nutrient conditions (Figure 3d,e) but, in L. incrustans, Chl a was positively affected by the future temperature-pH scenario (Appendix S1 & Figure 3f).  Figure 6f). Calcification was observed in the dark in L. corallioides under all treatments, and no effect was observed (Table 2 & Figure 7d). As in winter conditions, P. calcareum G D rates were negatively affected by the temperature-pH scenario (Table 2) with net dissolution observed under Fu-HN (Figure 7e).
Lithophyllum incrustans G D was significantly affected by nutrient availability (Table 2)

| Impact of the future temperature-pH scenario on the relationship between net primary production/ calcification and irradiance in winter conditions
The curves of the relationship between winter net primary production (NPP) or calcification (G) and irradiance (E) are shown on  Table 3). In the three maerl species, G max was enhanced under the future scenario (Figure 9, Tables 3 and   4). In P. calcareum, inhibition was observed at irradiances above 200 μmol photons/m 2 hr −1 (Figure 10). Calcification E k was only affected in L. corallioides, where it was higher under the future scenario ( Figure 9; Tables 3 and 4).

| Impact of the future temperature-pH scenario on the relationship between net primary production/ calcification and irradiance in summer conditions
The curves of the relationship between NPP or G and E are shown in Figures 10 and 11 (R 2 > 0.90). Only L. corallioides GPP max increased significantly under the future scenario (Tables 3 and 4).
Photosynthetic irradiance of saturation (E k ) did not vary significantly among treatments in any of the species (Figure 10 & Table 4).
Lithothamnion corallioides G max dropped significantly (of 37%) under the future scenario, but did not vary significantly between temperature-pH scenarios in the other species (Figure 11; Tables 3 and 4).
Calcification irradiance of saturation (E k ) did not vary among treatments in any of the species (Figure 11; Tables 3 and 4).

| NH 4 + consumption
NH 4 + consumption did not vary significantly between species and treatments whatever the season (ANOVA, p > .05), but winter consumption was significantly affected by irradiance (ANOVA, F = 96.06, p < .001). The highest value was observed at 400 μmol photons m −2 s −1 (0.06 ± 0.02 μmol NH 4 + g −1 DW hr −1 ) and the lowest at F I G U R E 5 Box plots presenting gross primary production rates of Lithothamnion corallioides, Phymatolithon calcareum, and Lithophyllum incrustans in the winter (a, b, c) and in the summer (d, e, f) (n = 3) in the present (Pr) and future (Fu) scenarios with low (LN) or high nutrients (HN). Boxes extend from the 25th to the 75th percentiles of all the data for each treatment; the central horizontal line represents the median, and bars extend to the 95% confidence limits 30 μmol photons m −2 s −1 (0.006 ± 0.01 μmol NH 4 + g −1 DW s −1 ). The NH 4 + summer consumption did not vary with irradiance (ANOVA, p > .05), but was higher (0.07 ± 0.02 μmol g −1 DW s −1 ) than in winter (ANOVA, p < .001).

| Effect of projected global climate change on physiological parameters
The response of the three studied maerl species to future pH and temperature scenarios varied with the season. This is in agreement with previous studies on coralline algae reporting that the effects of ocean warming and acidification are modified by seasonal variation in species responses and environmental conditions (Martin, Cohu, et al., 2013;Martin & Hall-Spencer, 2017). In temperate ecosystems, maerl beds face strong seasonal variability of physicochemical parameters (Qui-Minet et al., 2018) including variations in temperature and irradiance (length of photoperiod and intensity). Rising sea temperature can be beneficial in winter for coralline algae, allowing an increase in their photosynthetic and calcification rates (Martin, Cohu, et al., 2013). However, increased temperature in summer above the range of temperatures experienced in natural habitats could be detrimental for coralline algae; moreover, as it could be aggravated by decreased pH (Martin & Gattuso, 2009). In addition, illumination factors such as length of photoperiod and intensity affect the physiology of coralline algae (Martin, Castets, & Clavier, 2006;Martin, Charnoz, et al., 2013;Martin et al., 2007;Martin, Cohu, et al., 2013).
The length of the photoperiod directly influences diel patterns of respiration and net photosynthesis and their coupling to calcification.
In the light, the photosynthetic fixation of CO 2 causes an increase in the pH within the intracellular spaces and at the boundary layer of diffusion that induces CaCO 3 precipitation (Cornwall, Hepburn, Pilditch, & Hurd, 2013;De Beer & Larkum, 2001). Conversely, in the dark, the respiration process generates CO 2 , resulting in a decrease of pH within the intracellular spaces and the boundary layer of diffusion (Cornwall et al., 2013;Hurd, Harrison, Bischof, & Lobban, 2014) and thus hindering calcification.
In future climate change scenarios, the three maerl species may thus be particularly sensitive during the winter period when pho- Knowledge of the environmental conditions at the location where the species were collected is essential to understand how they may be impacted by global climate change. Likewise, the maerl bed (Roz bed located in the Bay of Brest), where the three maerl species were collected has a shallow depth (chart datum 0.7 m) and faces a high variability of light intensity resulting from cloud cover, turbidity, and tide variation. According to our results, the three maerl species will be more vulnerable to dissolution and/or their calcification rates will decrease in the near future during the winter season due to low light intensities (30 μmol photons m −2 s −1 ). Although, in the summer, the photoperiod and light intensity may increase; summer epiphyte algal species show high coverage of the Roz maerl bed, which severely decreases the light intensity that maerl receives. However, in this study, photosynthesis and calcification were not impacted negatively by none of the scenarios at low irradiances in the summer.
Species-specific consequences of ocean warming and acidification are also noticed for the respiration of the maerl species.
The respiration rates of L. corallioides and L. incrustans were not affected by the future pH and temperature scenario either in winter or in summer, while those of P. calcareum decreased significantly under the future summer scenario. Compared to the two other maerl species, P. calcareum has a higher affinity for lower temperatures (Adey & McKibbin, 1970) and may be more sensitive to  Note: Comparison of the season effect on the parameters of the NPP-E and G-E curves was done using a nonparametric test (Kruskal-Wallis) (n = 3). GPP max , the maximal gross primary production (μmol O 2 g −1 DW hr −1 ), E k , the irradiance of saturation (μmol photons m −2 s −1 ), G max , the maximal gross calcification (μmol CaCO 3 g −1 DW hr −1 ). Analyses significant at the α = .025 level are indicated by asterisks.
Values in bold refer to parameters that are significantly different.

| Effect of nutrient enrichment on physiological parameters
Coastal ecosystems are impacted by local changes, nutrient enrichment being one of the most relevant (Aufdenkampe et al., 2011).
Species-specific responses were observed. The absence of negative effects of nutrient enrichment on calcification contradicts previous studies that have described phosphate as a crystal poison due to its negative impact on the calcification rates of coralline algae (Bjork et al., 1995;Kinsey & Davies, 1979;Simkiss, 1964), even at the concentrations reported in the present study (1 μmol/L; Bjork et al., 1995).

| Interaction between global climate change and local change on physiological parameters
Environmental variations of natural and anthropogenic origin, including nutrient availability, may increase coralline algae resilience or exacerbate their response to global climate change .
According to some authors, the response of calcified organisms to the future conditions of temperature and pH will directly depend on nutrient availability (Celis-Plá et al., 2015;Langdon & Atkinson, 2005).
Nevertheless, its impact has not been previously studied in maerl/rhodolith species. Several authors reported a negative effect of phosphate on the calcification rates of coralline algae (Bjork et al., 1995;Kinsey & Davies, 1979;Simkiss, 1964). However, higher phosphate concentrations (1 μmol/L) did not decrease the calcification rates in any of the species under projected future temperature-pH conditions (winter and summer) at ambient irradiance and at dark. These results are consistent with previous observations on calcifying organisms suggesting that energy supply helps them to cope with higher pCO 2 (Schoepf et al., 2013;Tanaka et al., 2013;Ramajo et al., 2016). Nevertheless, the ratio between the rates of G L and gross primary production for P. calcareum decreased significantly under the Note: GPP max ,the maximal gross primary production rates (μmol O 2 g −1 DW hr −1 ), E k , the irradiance of saturation (μmol photon m −2 s −1 ), G max , the maximal gross calcification rates (μmol CaCO 3 g −1 DW hr −1 ). Nutrient availability can modify the enzymatic activity of algae and therefore the energetic balance and the response of photosynthesis and calcification processes to elevated CO 2 and increased temperature (Hofmann et al., 2013;Williamson, Perkins, Voller, Yallop, & Brodie, 2017). However, to our knowledge, there is no information regarding the energy allocation of maerl in an ocean acidification and warming context.

F I G U R E 1 0 Relationship between net primary production rates and irradiance in
Since pigments are N-containing components (Hurd et al., 2014), maerl specimens observing inhibition of winter gross primary production and net calcification under high irradiances were probably under deficient conditions of nitrogen.
The relationship between primary production/calcification and irradiance under different nutrient concentrations was studied only in L. corallioides in summer, and we did not observe any interactive effect between scenario and nutrients on the maximal rates of gross primary production and calcification (GPP max and G max ). The nonresponsiveness to these conditions may suggest the nonlimitation of nutrients for this species, likely due to the capacity of macroalgae to stock N during previous seasons and to their low energetic requirements (Hurd et al., 2014). Further studies should consider the response and potential interaction between global climate change and nutrient loading on the light harvesting and calcification mechanisms of coralline algae, in order to give a deeper insight into their capacity to adapt to future scenarios. These results suggest that despite any possible artifact introduced by the experimental system, nutrient enrichment can modify the response of the algae under a given scenario.

| CON CLUS ION
The magnitude of the ongoing global climate change effects on temperate maerl species located in coastal systems will depend on how their fitness responds to it in interaction with seasonal variations of temperature, photoperiod, light intensity, and nutrient availability, among other physicochemical parameters. It will also depend on the interaction with other stressors or sources of local changes, such as freshwater inputs, hydrodynamics, pollution, and among others (Horta et al., 2016;Qui-Minet et al., 2018).
Although care must be taken in the interpretation of the results due to potential pitfalls related to experimental design, our results suggest: • An antagonistic interaction between global climate change and nutrient enrichment with nutrient loading ameliorating some of the negative effects of global climate change.
• The ability of ocean acidification and warming to impair light harvesting and photoprotective mechanisms of maerl algae, while this may depend on the interaction with other abiotic factors.
• Species-specific responses to global climate change suggest that dominance and species distribution may deeply change in the future . For instance, although species can adapt to the same environment, they are sensitive to specific thresholds of environmental parameters. Our study suggests that ocean acidification and warming would render some species more vulnerable to stressful conditions of irradiance than others.
In this context, P. calcareum when living at shallow depths (thus, subject to higher light intensities) may be severely affected by global climate change relative to shallow adapted species such L. incrustans. P. calcareum is also the species with the lowest calcification rates and the only one to endure dissolution in the dark under the future summer scenario, which may threaten its growth rates and survival in shallow depths and lower latitudes.
Given these results, we expect that temperate maerl species will benefit from moderate nutrient enrichment in the winter season, avoiding dissolution under ambient low irradiance levels. Being L. corallioides the most abundant species in the Bay of Brest, the significant decrease of its calcification rates in the summer will have a negative impact on the CaCO 3 budget in this location. Furthermore, previous studies have shown a positive impact of ocean acidification on the development of fleshy epiphytic macroalgae (Johnson, Price, & Smith, 2014). Many of the species developing on North Atlantic maerl beds are consider as opportunists (Bunker, Brodie, Maggs, & Bunker, 2017) and may also benefit from the increase of temperature and nutrient availability with negative consequences for maerl survival.
In this context, fleshy epiphytic macroalgae compete with maerl for light, carbon, and nutrients (Steneck, 1986). They can also modify the seawater chemistry within the diffusive boundary layer and increase the diel pH fluctuations under future global change scenarios (Short, Pedersen, & Kendrick, 2015). However, epiphytic abundance is also affected by other abiotic and biotic parameters such as hydrodynamics (Hily, Potin, & Flocíh, 1992), the antifouling ability of coralline algae (Figueiredo, Norton, & Kain, 1997) and grazing by the benthic fauna (Guillou, Grall, & Connan, 2002). Furthermore, coastal ecosystems are highly heterogeneous, and maerl beds located in these ecosystems are subject to different sources of local change which affect their resilience and capacity to adapt to future scenarios of global change (Horta et al., 2016). Due to their slow growth rates (Potin, Floc'h, Augris, & Cabioch, 1990), activities such as dredging may reduce their abundance at a higher rate than global climate change Hall-spencer, Grall, Moore, & Atkinson, 2003).
We are aware that the small number of replicates resulting from a constraining experimental design prevented us from carrying out a more powerful single multifactorial analysis. This weakness is a consequence of the process of testing and answering a complex but realistic problem, the combined action of climate change and local eutrophication in a multifactorial experimental design. Therefore, the scope of our study was limited, and our results should be considered carefully. Despite interdependence of replicates per scenario, this paper contains useful information to increase our ability to predict the future effects of ocean acidification.
Such experiment could have been improved by increasing the number of header tanks to avoid pseudo-replication as recommended by Cornwall and Hurd (2016). However, the number of tank and aquarium units would multiply rapidly in such an experiment, manipulating both climate change scenarios and nutrient concentrations at multiple levels. This would have been prohibitive.
Further studies considering other abiotic parameters and biotic interactions need to be carried in order to develop management policies capable to increase the resilience of maerl beds under the future global climate change scenario.

ACK N OWLED G M ENTS
We thank the sailors of the Albert Lucas Oceanographic vessel, Franck, and Daniel for their invaluable help during the sampling. This work benefited from a grant of the Consejo Nacional de Ciencia y Tecnología (CONACyT) (391572) and was supported by the French National Research Agency via the "Investment for the Future" program IDEALG (no. ANR-10-BTBR-04). We would like to thank the reviewers and the editor for their thoughtful comments and suggestions toward improving our manuscript.

CO N FLI C T O F I NTE R E S T
None declared.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data supporting this study are provided as a Dryad file: https :// doi.org/10.5061/dryad.zkh18 935j