Macroalgal responses to ocean acidification depend on nutrient and light levels

Ocean acidification may benefit algae that are able to capitalise on increased carbon availability for photosynthesis but is expected to have adverse effects on calcified algae through dissolution. Shifts in dominance between primary producers will have knock-on effects on marine ecosystems and will likely vary regionally, depending on factors such as irradiance (light vs shade) and nutrient levels (oligotrophic vs eutrophic). Thus experiments are needed to evaluate interactive effects of combined stressors in the field. In this study, we investigated the physiological responses of macroalgae near a CO2 seep in oligotrophic waters off Vulcano (Italy). The algae were incubated in situ at 0.2 m depth using a combination of three mean CO2 levels (500, 700-800 and 1200 µatm CO2), two light levels (100 and 70% of surface irradiance) and two nutrient levels of N, P, and K (enriched vs non-enriched treatments) in the non-calcified macroalga Cystoseira compressa (Phaeophyceae, Fucales) and calcified Padina pavonica (Phaeophyceae, Dictyotales). A suite of biochemical assays and in vivo chlorophyll a fluorescence parameters showed that elevated CO2 levels benefitted both of these algae, although their responses varied depending on light and nutrient availability. In C. compressa, elevated CO2 treatments resulted in higher carbon content and antioxidant activity in shaded conditions both with and without nutrient enrichment - they had more Chla, phenolic and fucoxanthin with nutrient enrichment and higher quantum yield (Fv/Fm) and photosynthetic efficiency (αETR) without nutrient enrichment. In P. pavonica, elevated CO2 treatments had higher carbon content, Fv/Fm, αETR, and Chla regardless of nutrient levels - they had higher concentrations of phenolic compounds in nutrient enriched, fully-lit conditions and more antioxidants in shaded, nutrient enriched conditions. Nitrogen content increased significantly in fertilised treatments, confirming that these algae were nutrient


INTRODUCTION 72
Ocean acidification due to increased atmospheric CO2 levels is altering the concentrations 73 of dissolved inorganic carbon (DIC) in surface waters; CO3 2levels are falling, which is 74 expected to corrode marine carbonates, whilst CO2 and HCO3levels are rising which can 75 stimulate photosynthesis (Connell et al., 2013;Cornwall et al. 2015). As some primary 76 producers are better able to capitalize on increasing carbon availability than others, this is 77 expected to alter marine communities (Hepburn et al., 2011;Connell et al., 2013;Koch et al., 78 2013;Gaylord et al., 2015). In the Mediterranean, surveys of coastal CO2 seeps have repeatedly 79 shown that coralline algae and sea urchins become less common as pH and CO3 2fall, whereas 80 brown algae, such as Cystoseira spp., Dictyota spp., Sargassum vulgare and Padina pavonica, 81 proliferate as CO2 and HCO3levels rise (Porzio et al., 2011;Baggini et al., 2014). The ways 82 in which ocean acidification affects communities of primary producers is likely to vary 83 regionally, depending on the species present and abiotic factors such as temperature, light and 84 nutrient availability (Giordano et al., 2005;Brodie et al., 2014;Hofmann et al. 2014). 85 To begin to understand the influence of physicochemical factors on the responses of 86 macroalgae to ocean acidification, we grew common types of brown algae (from the Families 87 Fucales and Dictyotales) at CO2 seeps in a multifactorial experiment in which we manipulated 88 light (irradiance) and nutrient levels. At low light levels, macroalgae are thought to be more 89 likely to rely on carbon uptake via diffusion than use energetically expensive carbon 90 concentrating mechanisms (Raven and Beardall, 2014;Raven et al., 2014) which has led to the 91 idea that any benefits of ocean acidification on growth would only be seen at lower light levels 92 for the majority of species (Hepburn et al., 2011). However, ocean acidification also has the 93 potential to damage photoprotective mechanisms which kick-in at high light levels (Pierangelini 94 et al., 2014). Algae minimize damage from high irradiance by down-regulating photosystems, 95 they also produce chemicals, such as phenolic compounds in the brown algae, which screen 96 ultraviolet light and dissipate energy (Figueroa et al., 2014a). In oligotrophic waters, such as 97 those of the Mediterranean, nutrient availability generally limits macroalgal growth (Ferreira et 98 al., 2011), photosynthetic capacity (Pérez-Lloréns et al., 1996) and photoprotective 99 mechanisms (Celis-Plá et al., 2014b). 100 Our study centres upon a highly oligotrophic region (the Tyrrhenian Sea) which is 101 undergoing rapid changes in carbonate chemistry coupled with coastal eutrophication and 102 increased land run-off (Oviedo et al., 2015). In this region, as with elsewhere in the world, 103 canopy-forming brown algae have undergone a decline in abundance due to anthropogenic 104 perturbation (Scherner et al., 2013;Strain et al., 2014;Yesson et al., 2015). Here, we investigate 105 5 N (NH4 + and NO3 -), 17% P (P2O5) and 17% K (Multicote®, Haifa Chemicals, USA) were fixed 140 below nutrient enriched cylinders. For the non-enriched treatments, a bag with 100 g of sand 141 was used as a control. The nutrient treatments were set 20 m apart from each other so that non-142 enriched treatments were unaffected. 143

Environmental conditions 144
The seawater carbonate system was monitored at each site (Table 1). A 556 MPS YSI 145 (Yellow Springs, USA) probe was used to measure salinity, pH and temperature (°C). The pH 146 sensor was calibrated using NBS scale standard buffers. On 20 th March 2013, water samples for 147 total alkalinity (TA) were strained through 0.2 µm filters, poisoned with 0.05 ml of 50% HgCl2, 148 and then stored in the dark at 4° C. Three replicates were analyzed at 25° C using a titrator 149 (Mettler Toledo, Inc.). The pH was measured at 0.02 ml increments of 0.1 N HCl. 150 Total alkalinity was calculated from the Gran function applied to pH variations from 4.2 to 151 3.0, from the slope of the curve HCl volume versus pH. The pCO2 and the saturation state of 152 aragonite were calculated from pHNBS, TA, temperature and salinity using the CO2 SYS package 153 (Pierrot et al., 2006), using the constants of Roy et al., (1993) andDickson, (1990 Several physiological variables were obtained from the algae within each cylinder at the 173 end of the experiment. These variables were also measured in C. compressa and P. pavonica 174 from ambient CO2 site (500 µatm) populations at 0.5 m depth. Carbon and nitrogen contents 175 were determined using an element analyzer CNHS-932 model (LECO Corporation,Michigan,176 USA). 177 In vivo chlorophyll a fluorescence associated with Photosystem II was determined by using 178 a portable pulse amplitude modulated (PAM) fluorometer (Diving-PAM, Walz GmbH, 179 Germany). Macroalgal thalli were collected from natural populations (initial time) and after 180 four days of incubation in the experiment (for each treatment or cylinder), and were put in 10 181 mL incubation chambers to obtain rapid light curves for each treatment. photons m -2 s -1 , A is the thallus absorptance as the fraction of incident irradiance that is absorbed 195 by the algae (see Figueroa et al., 2003) and FII is the fraction of chlorophyll related to PSII (400-196 700 nm) being 0.8 in brown macroalgae (Figueroa et al., 2014a). ETR parameters as maximum 197 electron transport rate (ETRmax) and the initial slope of ETR versus irradiance function (αETR) 198 as estimator of photosynthetic efficiency were obtained from the tangential function reported 199 by Eilers and Peeters, (1988). Finally, the saturation irradiance for ETR (EkETR) was calculated 200 from the intercept between ETRmax and αETR. Non-photochemical quenching (NPQ) was 201 calculated according to Schreiber et al., (1995)  Maximal NPQ (NPQmax) and the initial slope of NPQ versus irradiance function (αNPQ) 206 were obtained from the tangential function of NPQ versus irradiance function according to 207 Eilers and Peeters (1988). 208 Pigments were extracted from 20 mg fresh weight of thalli using 2 mL of 100% acetone 209 and analyzed using an ultra-high-performance liquid chromatographer (Shimadzu Corp.,210 Kyoto, Japan) equipped with a photodiode array detector to measure peaks in the range 350-211 800 nm. After extraction samples were centrifuged at 16200 g for 5 min (Sorvall Legend Micro 212 17, Thermo Scientific, Langenselbold, Germany) and then the extracts were filtered (0.22 μM 213 nylon filters). The separation, was achieved with one column C-18 reversed phase (Shim-pack 214 XR-ODS column; 3.0 × 75mm i. d.; 2.2 um particle size; Shimadzu, Kyoto, Japan) protected 215 by a guard column TR-C-160 K1 (Teknokroma, Barcelona, Spain). The carotenoid composition 216 was determined according to García-Plazaola and Becerril, (1999) with some modifications 217 (García-Plazaola and Esteban, 2012), using commercial standards (DHI LAB Products). The 218 mobile phase consisted of two components: Solvent A, acetonitrile: methanol: Tris buffer (0.1 219 M, pH 8) (84:2:14); and solvent B, methanol: ethyl acetate (68:32). The pigments were eluted 220 using a linear gradient from 100% A to 100% B for the first 7 min, followed by an isocratic 221 elution with 100% B for the next 4 min. This was followed by a 50 seconds linear gradient from 222 100% B to 100% A and an isocratic elution with 100% B for the next 3 min to allow the column 223 to re-equilibrate with solvent A, prior to the next injection. 224 Total phenolic compounds were determined using 0.25 g fresh weight samples pulverized 225 with a mortar and pestle with sea-sand and 2.5 mL of 80% methanol. After keeping the samples 226 overnight at 4ºC, the mixture was centrifuged at 2253 g for 30 min at 4ºC, and then the 227 supernatant was collected. Total phenolic compounds were determined colorimetrically using 228 Folin-Ciocalteu reagent and phloroglucinol (1,3,5-trihydroxybenzene, Sigma P-3502) as 229 standard. Finally the absorbance was determined at 760 nm using a spectrophotometer (UV 230 Mini-1240, Shimadzu) (Celis-Plá et al., 2014a). Total phenolic content was expressed as mg g 231 -1 DW after determining the fresh to dry weight ratio in the tissue (5.2 for C. compressa and 4.5 232 P. pavonica, respectively). The results are expressed as average ± SE from three replicates of 233 each treatment. Antioxidant activity was measured on polyphenol extracts according to Blois 234 (1958); 150 µL of DPPH (2,2-diphenyl-1-picrylhydrazyil) prepared in 90% methanol were 235 added to each extract. The reaction was complete after 30 min in darkness at ambient 236 temperature (~20°), and the absorbance was read at 517 nm in a spectrophotometer (UVmini-237 1240, Shimadzu). The calibration curve made from DPPH was used to calculate the remaining 238 concentration of DPPH in the reaction mixture after incubation. Values of DPPH concentration 239 8 (mM) were plotted against plant extract concentration expressed as the EC50 value (oxidation 240 index, mg DW mL -1 ) required to scavenge 50% of the DPPH in the reaction mixture. Ascorbic 241 acid was used as a control (Celis-Plá et al., 2014a). 242

Statistical analysis 243
The effects of the in situ treatments on the physiological responses of C. compressa and P. 244

Environmental conditions 256
Cystoseira compressa and Padina pavonica were abundant at all three stations; P. pavonica 257 was visibly less calcified at the site with the highest levels of CO2. The seawater temperature 258 was about15 ºC and the salinity was 38 at all stations; at the Ambient site, mean pH was 8.11, 259 at the Medium CO2 site (700-800 µatm), mean pH was 7.97 and at the High CO2 site (1200 260 µatm), it was 7.86 (Table 1). 261 The average daily irradiance for the experimental period was 5360 kJ m -2 for PAR and 666 262 kJ m -2 for UVA. The nutrient enriched treatments had approximately 100 times the nitrate 263 concentration of the ambient seawater; ambient vs enriched ratios were 0.16 ± 0.04 vs 106.17 ± 264 9.37 µM for the ambient site, 0.13 ± 0.01 vs 106.33 ± 9.37 µM at the medium CO2 site and 0.25 265 µM ± 0.01 vs 106.42 ± 9.37µM at the high CO2 site (mean ± SE, n = 3). 266

Physiological and biochemical responses 267
The carbon content of C. compressa increased with increasing CO2, whereas in P. pavonica 268 showed interactive effects between all factors. Carbon, in P. pavonica showed maximal values 269 279.9 ± 6.5 with increased CO2, in non-enrichment enriched treatments and minimal values 270 225.3 ± 2.4 mg g -1 DW with decreased CO2, non-nutrient enriched and 70%PAB conditions 271 (Table 2, S1). The nitrogen content of C. compressa was greatest in the high CO2, nutrient 272 enriched and 70%PAB treatment (Figure 2A, S1); conversely, in P. pavonica the nitrogen content 273 9 was highest at the reference site, ambient CO2 treatment ( Figure 2B, S1). The ratio C:N of C. 274 compressa did not show significant differences between treatments ( Figure 3A, S1), whereas 275 in P. pavonica showed significant effects for CO2 levels and nutrient enrichment ( Figure 3B, 276 S1). The C:N ratio , in P. pavonica showed maximal values (19.5 ± 5.8) with increased CO2, 277 non-nutrient enriched in 100%PAB conditions and minimal values 15.9 ± 0.5 in medium CO2, 278 nutrient enrichment and 70%PAB conditions ( Figure 3B, S1). 279 The maximal quantum yield (Fv/Fm) was significantly different between CO2 treatments, 280 nutrient and irradiance in both macroalgae (Figure 4, S2). In C. compressa, the Fv/Fm was 281 greatest in 70%PAB treatments with high CO2, and non-enriched enrichment ( Figure 4A, S2), 282 but in P. pavonica this was greatest in the nutrient enriched treatments ( Figure 4B, S2). The 283 αETR values also varied significantly between treatments in both species ( compressa was highest in high CO2, 70%PAB and non-nutrient enrichment, also in 100%PAB and 287 nutrient enrichment, and also this was higher with decreased CO2, 100%PAB, in non-nutrient 288 enrichment. In P. pavonica, ETRmax varied significantly depending on nutrient and irradiance, 289 without interactions (Table 2, S2). In contrast, the EkETR in C. compressa had one significant 290 interaction among nutrient x irradiance. P. pavonica had significant interactions between CO2 291 level, nutrient and irradiance. The EkETR, in C. compressa, was greatest in the 100%PAB 292 treatments that had no CO2 or nutrient enrichment, but in P. pavonica EkETR was greatest in 293 70%PAB conditions (Table 2, S2). In both species, the maximal non-photochemical quenching 294 (NPQmax) was affected by the interaction of all factors. In C. compressa, NPQmax increased 295 significantly with increasing CO2 conditions, under nutrient enriched and 100%PAB, also 296 increased in ca 700-800 µatm but in 70%PAB. As well as, NPQmax increased under ambient CO2 297 conditions in 100%PAB in nutrient non-enriched. Finally, in P. pavonica, the NPQmax was 298 significantly higher in 70%PAB at 700 µatm CO2 treatment with nutrient enrichment (Table 2, 299 S2). 300 Nutrient enrichment increased Chla significantly in C. compressa. In contrast, in P. 301 pavonica significant differences were found for the following interactions: CO2 level x nutrient, 302 CO2 level x irradiance and nutrient x irradiance (Table 3, S3). The same occurred for Chlc in 303 P. pavonica; but there was no significant difference in C. compressa (Table 3, S3). The 304 carotenoids, fucoxanthin and violaxanthin in C. compressa did not differ among factors (Table  305 3, S3) but in P. pavonica the fucoxanthin and violaxanthin contents were affected by the 306 interaction of all factors. Fucoxanthin increased in 70%PAB, non-enriched treatments in ambient 307 CO2 whereas violoxanthin levels were highest in 70%PAB, ca 700-800 µatm CO2, nutrient 308 enriched treatment (Table 3, S3). 309 Phenolic content (PC) was affected by the interaction of all factors in both species (Figure  310 5, S4). In C. compressa, PC was highest in CO2 and nutrient enriched conditions ( Figure 5A, 311 S4). In P. pavonica at 1200 µatm CO2, PC was high in 100%PAB and nutrient enriched 312 treatments and in 70%PAB treatments non-nutrient enrichment ( Figure 5B, S4). Antioxidant 313 activity (EC50) showed a significant interaction between CO2 level x nutrient and CO2 level x 314 irradiance in C. compressa; however in P. pavonica the only significant difference found in 315 antioxidant activity was between CO2 level and irradiance. In C. compressa and P. pavonica, 316 EC50 was lowest (i.e. it had higher antioxidant activity) in the high CO2, 70%PAB light conditions 317 and nutrient enriched treatments ( Recent reviews surmise that ocean acidification is likely to increase macroalgal 321 productivity due to beneficial effects of increased dissolved inorganic carbon (DIC) levels 322 which can stimulate the growth of algae and allows them to divert more resources into anti-323 herbivore and photo-protective compounds (Harley et al., 2012;Brodie et al., 2014). Here we 324 show that calcified and non-calcified macroalgae can benefit physiologically from increases in 325 DIC, but that the benefits, and the extent of the algal response, depends upon nutrient and light 326 availability. Figure 6 summarizes our projections that brown macroalgal stands will both 327 proliferate in the shallows (because of up-regulation of anti-herbivore and photo-protective 328 compounds) and extend deeper due to the greater availability of DIC and nutrients due to a 329 combination of ocean acidification and anthropogenic nutrient input, whereas other work on 330 Mediterranean CO2 seeps has established that sea urchins and coralline algae will be adversely 331 affected by acidification (Baggini et al., 2014). In vivo chlorophyll fluorescence parameters 332 (maximal quantum yield or Fv/Fm and maximal electron transport rate or ETRmax) and algal 333 biochemical composition (Chla, total phenolic compounds and antioxidant activity, % C) helps 334 explain the dominance of phaeophytes at a variety of coastal Mediterranean CO2 seeps. 335 Increases in brown macroalgal cover at CO2 seep sites are probably due to a combination of the 336 direct stimulus of increased DIC for photosynthesis for species with inefficient carbon 337 concentrating mechanisms (CCMs), and the decreased grazing since sea urchins for example 338 are excluded by hypercapnia (Calosi et al., 2014). 339 Other seep locations show similar trends to Vulcano, with increases in Cystoseira and 340 Padina species at elevated CO2 locations compared to reference locations (Johnson et al., 2012;341 benefit some macroalgae, such as Gracilaria lemaneiformis in China (Zou and Gao, 2009) and 343 the mat-forming Feldmannia spp. in Australia (Russell et al., 2011), as well as canopy-forming 344 phaeophytes such as Nereocystis luetkeana and Macrocystis pyrifera (Swanson and Fox, 2007;345 Roleda et al., 2012). We found that the benefits of increased DIC were even more pronounced 346 when combined with increased nutrients. This is what we expected, given that macroalgae tend 347 to be nutrient-limited in oligotrophic waters such as those of the Mediterranean Sea (Ferreira et 348 al., 2011). Both our study species increased electron transport rates and the accumulation of 349 photoprotectors when exposed to a Nitrogen Phosphorus Potassium fertilizer, but we should 350 bear in mind that these were short-term experiments with macroalgae grown in isolation. We 351 suspect that chronic eutrophication combined with ocean acidification may benefit more 352 opportunistic algal groups, to the detriment of brown macroalgae based on research by Russell 353 et al. (2009) andFalkenberg et al. (2013). In our study C. compressa and P. Pavonica had 354 increased carbon content when CO2 levels increased, which was augmented by increases in a 355 range of other physiological parameters when nutrient levels were also increased. The Fv/Fm 356 ratios was highest at increased CO2 concentrations with no nutrient enrichment in C. compressa, 357 but highest at increased CO2 with nutrient enrichment for P. pavonica (Figure 4). The maximal 358 photosynthetic activity (ETRmax) in C. compressa was reduced at high nutrient levels in shaded 359 conditions but in fully lit conditions nutrients did not have significant effects under high DIC 360 conditions. In other Cystoseira species, such as C. tamariscifolia, both Fv/Fm and ETRmax also 361 decrease in nutrient enriched treatments in field experiments at various depths (Celis-Plá et al., 362 2014b). In another experiment, Celis-Plá et al. (2014a) found the highest ETRmax in C. 363 tamariscifolia in thalli with the lowest internal nitrogen stores i.e. winter compared to summer 364 grown algae. On the other hand, EkNPQ increased in all cases with the increased CO2 as an 365 acclimation to high light levels. On this basis, it is clear that the responses of coastal macroalgal 366 communities to ocean acidification will depend on nutrient availability, and will be species-367 specific. Given these results we expect that in temperate waters, brown algae will benefit from 368 increases in CO2 if sufficient nutrients are available (Johnson et al., 2012). However, as with all 369 ecology, we can expect that there will be a region-specific balancing act. We show here that in 370 oligotrophic conditions brown macroalgae were unable to take full advantage of increased 371 inorganic carbon availability. There is added complexity when we consider that many regions 372 have experienced a die-back of canopy-forming brown algae due to excess nutrients or 373 sedimentation (Strain et al., 2014); ocean acidification may exacerbate this problem since 374 increased DIC may further benefit those algae that presently compete with fucoids and kelps in 375 eutrophic conditions (Connell et al., 2013). 376 Light quantity and quality drive physiological processes in macroalgae (Hanelt and López-377 Figueroa, 2012), so we were not surprised to find that shading affected their responses to ocean 378 acidification. We expected two outcomes of the effects of light: we expected ETR rates to be 379 higher as the most obvious response to light, but we also expected low-light macroalgae to 380 increase ETR rates and % C when transplanted to higher CO2 concentrations. Our first 381 expectations were met, as maximum quantum yield, photosynthetic efficiency, irradiance of 382 saturation and non-photochemical quenching for chlorophyll fluorescence all increased at 383 higher light levels and were, at times, amplified by increasing CO2 and nutrient levels. The only 384 instance where our second expectation was met was for P. pavonica under ambient nutrients, 385 which had significantly higher % C (and non-significantly higher ETRmax) when transplanted 386 to elevated CO2 sites. Previous studies at the same sites found elevated ETRmax when comparing 387 P. pavonica at an elevated CO2 site compared to an ambient CO2 site (Johnson et al., 2012). If 388 the duration of our experiment had been longer, our transplanted P. pavonica may also have 389 significantly increased their ETRmax. Our results emphasise the likelihood that ocean 390 acidification will act upon primary production differently at different latitudes and depths, not 391 always according to our expectations. This is important since increases in land nutrient run-off, 392 due to changes in land use and/or rainfall, are altering light levels in coastal waters (Scherner et 393 al., 2013). 394 One of the most important photoprotective mechanisms available to algae is an ability to 395 dissipate excess thermal energy (Adams et al., 2006). Thermal dissipation measured as non-396 photochemical PSII fluorescence quenching (NPQ) is triggered by the trans-thylakoidal proton 397 gradient (ΔpH) and zeaxanthin (ZEA) synthesis through the xanthophyll cycle (Gilmore et al., 398 1994) and is recognized as the most important photoprotective mechanisms in higher plants and 399 several algal divisions (Rodrigues et al., 2002). Fucoxanthin and violaxanthin levels were not 400 affected in C. compressa whereas in P. pavonica fucoxanthin and violaxanthin increased under 401 70%PAB conditions, nutrient enrichment and medium CO2 levels.We used NPQmax as an 402 indicator of photoprotective energy dissipation efficiency (Celis-Plá et al., 2014a), and we also 403 measured phenolic content and antioxidant activity (EC50), both of which can be used as 404 photoprotectors (Celis-Plá et al., 2014b). In C. compressa and P. pavonica NPQmax was higher 405 in all shaded treatments with nutrient enrichment, but not in the fully lit treatments, indicating 406 higher photoprotection when nutrients were elevated and light was reduced. Phenols usually 407 accumulated under higher irradiance and (for C. compressa) higher CO2 treatments, as per past 408 13 studies on kelp grown at high CO2 (Swanson andFox, 2007), or measured under higher 409 irradiance (Connan et al., 2004). However, the effects of CO2 on autotroph phenol production 410 are not straight forward, as previous work has shown that both seagrass (Arnold et al., 2012) 411 and the macroalga Cystoseira tamariscifolia (Figueroa et al., 2014b) decrease phenol 412 production when CO2 increased. In C. compressa and P. pavonica, antioxidant activity and 413 EC50 were affected by the interactions between light levels and CO2. EC50 tended to be higher 414 in shaded, high CO2 treatments with and without nutrient addition, suggesting a positive 415 correlation with phenolic compounds and their use as antioxidants to prevent photodamage. 416 Together, NPQmax, phenol production and EC50 indicate that in elevated CO2 conditions some 417 species will have a higher capacity for photoprotection. 418 Macroalgae regulate their biochemical composition to changes in solar radiation (Bischof 419 et al., 2006;Figueroa et al., 2014aFigueroa et al., , 2014b. Whilst light obviously affects photosynthesis, other 420 variables such pH, nutrients and the availability of different DIC species all have the potential 421 to affect photosynthetic rates (Raven and Beardall, 2014). As interactions among such factors 422 will determine the success of algal species and the amount of primary productivity in any time 423 and place, it is crucial to know how the effects of ocean acidification are modified by other key 424 drivers of photosynthesis. Research similar to our study, but with more species, in more 425 locations and for longer durations, is clearly required before solid conclusions can be made with 426 respect to the effects of ocean acidification on macroalgal productivity. 427 In conclusion, our study shows that ongoing ocean acidification can be expected to increase 428 photosynthetic efficiency and algal productivity. The magnitude of these effects, and the species 429 that benefit, will depend on light and nutrient levels. We show that C. compressa and P. 430 pavonica are able to benefit from an increase in CO2 levels, rapidly changing their physiology 431 and biochemical composition over three day alterations in DIC, irradiance and nutrients. These 432 factors had interactive effects on photosynthetic and photoprotective systems in both species 433 and help explain why brown algae proliferate at CO2 seeps. Longer-term growth studies 434 involving algal interactions would be useful: we remain concerned that chronic eutrophication 435 combined with ocean acidification may benefit more opportunistic algal groups to the detriment 436 of canopy-forming brown macroalgae. As ocean acidification is not happening in isolation, but 437 alongside a plethora of other anthropogenic changes, an understanding of the interactive effects 438 of multiple stressors is critical to plan for global ocean change. We have shown that elevated 439 CO2 levels can enhance brown algal productivity, and may boost the kelp and fucoid forests of 440 the planet, but the effects will depends upon interactions with other physicochemical parameters 441 such as light and nutrient availability. Bull. 72(1), 107-118. 469 Bischof, K., Gómez, I., Molis, M., Hanelt, D., Karsten, U., Lüder, U. et al. (2006). Ultraviolet 470 radiation shapes seaweed communities. Rev. Environ. Sci. Biot. 5, 141-166. 471 Blois, M.S. (1958). Antioxidant determinations by the use of a stable free radical. Nature. 181, 472 1199-1200. 473 Table 1. Seawater carbonate chemistry at three sites off Vulcano Island. The three sites were at 637 a decreasing distance from CO2 seeps, with an ambient CO2 site, a Medium CO2 and a High 638 CO2 site. Temperature (ºC), Salinity and pH (NBS scale) were collected on different days in 639