Short-term effects of CO 2 , nutrients and temperature on three marine macroalgae under solar radiation

Three macroalgal species belonging to Chlorophyta (Ulva rigida), Rhodophyta (Ellisolandia elongata) and Phaeophyceae (Heterokontophyta; Cystoseira tamariscifolia), naturally growing at the same shore level and representing 3 morpho-functional groups, were exposed to short-term changes in temperature under different carbon and nitrogen regimes. Experiments were conducted in outdoor tanks at 4 combinations of carbon and nitrogen levels under reduced solar radiation. In vivo chl a fluorescence parameters and pigment contents were monitored to assess diurnal physiological responses and potential for recovery. Strong fluctuations in chl a fluorescence parameters, but not in chl a content, were observed in response to diurnal variation in solar radiation and light climate within the tanks; sensitivity varied between algal species and, in some cases, depended on the carbon and nitrogen regime. Nitrogen uptake was similarly high in U. rigida and E. elongata and lowest in C. tamariscifolia. In U. rigida and E. elongata, chl a concentrations decreased after high-carbon treatments. Effective photosystem II quantum efficiency was reduced in all species at noon, and lowest in C. tamariscifolia. The results highlight the complexity of physiological short-term acclimations which were most likely linked to biochemical changes at the cellular level. Long-term experiments are required in future for more comprehensive investigation of the observed interactive effects of the different environmental parameters.

To date most studies have focused on the isolated impact of CO 2 levels (and/or reduced pH) on algal physiology such as carbon exchange (e.g.Cornwall et al. 2012, Ní Longphuirt et al. 2013), calcification (e.g.Martin & Gattuso 2009), growth (e.g.Riebesell et al. 2007), reproduction (e.g.Roleda et al. 2012), or several parameters in selected species (e.g.Bach et al. 2013).However, in natural systems, predicted and already detectable changes in seawater pH will interact with other ambient physico-chemical conditions and potential stressors such as temperature (Roleda & Hurd 2012) and solar light climate, including UV radiation (Gao & Zheng 2010, Russell et al. 2011, Yildiz et al. 2013), at species and community levels (Porzio et al. 2011, Hofmann et al. 2012).Whilst interactive impacts of coastal eutrophication and OA on water chemistry have been demonstrated (e.g.Cai et al. 2011), little is known about their combined effects on algal physiology.
Physiological characteristics of macroalgae, such as light absorption and conversion efficiency, carbon exchange and nutrient uptake rates, are impacted by thallus morphology and have been reported to differ between thin species (such as Ulva spp.), slowgrowing calcareous species (e.g.red coralline algae) and more complex brown algae (such as Cystoseira spp.).Additionally, species-specific features, such as growth in dense clumps within algal turfs (Hay 1981) and physiological adaptations to facilitate protection against stressful environments (e.g. the capacity to exude phenolics during high light stress; Shibata et al. 2006), may influence measurable responses to climate change parameters of interest.Several studies have previously assessed interactive effects of nitrate and UV radiation on photosynthesis (pulse-amplitude modulated [PAM]-chlorophyll fluorescence), pigments, mycosporine-like amino acids (MAAs) and nutrient uptake characteristics of red and green macroalgae from southern Spain (e.g.Figueroa et al. 2003).Additionally, a study on nutrient uptake characteristics of algae with different life strategies suggested that ni trogen uptake characteristics were aligned to growth strategies and the adaptation of individual species to local nitrogen supply regimes (short-term pulses vs. long-term exposure) (Martínez et al. 2012).Figueroa et al. (2009) demonstrated the effect of nutrient regime on the capability of Ulva lactuca to respond to temperature and light stress; how such short-term responses are influenced by additional im pacts of high-CO 2 levels is not certain.On the shores of southern Spain, Ulva rigida C. Agardh, Ellisolandia elongata (J.Ellis & Solander) K. R. Hind & G. W. Saunders (formerly Corallina elongata) and Cystoseira tamariscifolia (Hudson) Papenfuss inhabit exposed rocky mid-to low-intertidal shores in close vicinity (Figueroa et al. 2014a).Thus, a comparative analysis of their respective short-term responses to such stressors, most likely linked to morphological features and life strategies of these species and with potential ecological consequences, is of interest.
This paper represents the first in a series of outputs from the 2012 International Workshop of the Group on Aquatic Primary Productivity (GAP) hosted by the University of Málaga, southern Spain, and includes a description of the experimental set-up employed by the working group on macroalgae, and some primary results.Objectives of the Working Group on Macroalgae were to assess impacts of pulse-supply of nitrogen on primary productivity of selected macroalgae.Here, we describe physico-chemical parameters within tanks with different carbon and nitrogen regimes to which the 3 marine macroalgal species described above were exposed.Nitrate concentration pulses used in this experiment were in the range of those observed in the coastal water in the Bay of Málaga (Ramírez et al. 2005).Impacts of diurnal variation in light climate and am bient temperature (including an experimental 4°C temperature increase) on short-term physio logical responses (PAM fluorometry: effective Photosystem II [PSII] quantum yield) and pigments are reported.The capability of different algal species to recover from mid-day exposure to high irradiances was in vestigated under different carbon and nitrogen re gimes; this capability was further assessed after imposing additional short-term temperature stress (4°C increase for 3 days).
Results of further detailed physiological measurements (P/E curves measured via PAM chl a fluorescence under controlled conditions) and biochemical analyses, to assess impacts of the broader photosynthetic processes and overall metabolism, will be presented and discussed elsewhere (Figueroa et al. 2014b,c, this Theme Section).

Experimental design
The experiment was designed to examine interactive effects of the current pCO 2 (380 ppm, pH target above 8.0) and predicted future concentration for the year 2100 (700 ppm, pH target 7.92; Orr et al. 2005, Meehl et al. 2007) in a crossed combination with 2 pulsed nitrate concentrations (5 and 50 µM) according to nitrate data reported for the Alborán Sea, southern Spain (Ramírez et al. 2005).The 4 treatments were de signated as low-carbon/ low-nitrogen (LCLN), low-carbon/ high-nitrogen (LCHN), highcarbon/ low-nitrogen (HCLN) and high-carbon/ highnitrogen (HCHN) (see next section).In addition, the effect of a short-term temperature increase was tested in the second part of the experiment.
The experiments were conducted in a multi-tank system placed under semi-natural solar conditions, i.e. photosynthetically active radiation (E PAR ; 400-700 nm) reduced by 35%, and UVA (320-400 nm) and UVB (280-320 nm) reduced by 39%, using a green mesh, in the Unit for Microbiology, Ecophysiology and Aquatic Organisms of Málaga University (UMEGOA), located in the Grice-Hutchinson Experimental Centre.The experimental system was composed of 3 open vessels (0.094 m 2 surface area, 14 l volume) per treatment and species (n = 3) connected in parallel to a separate tank of 60 l capacity (1 tank per treatment and species, 12 tanks in total, Fig. 1a).The water flow between each vessel and its header tank was 0.84 ± 0.05 l min -1 , representing a turnover rate of 26 ± 1% h -1 .In addition to the series connection circuit causing permanent seawater supply and movement and mixing within vessels, a second closed circuit re-circulated the water along a 3 m pipe within each header tank.This secondary recirculated system facilitated water mixing in the header tank.Aeration with ambient air caused further mixing within vessels.The seawater used originated from the Bay of Málaga and was stored in two 12 000 l tanks on site.The entire system was placed within 8 tanks of 1000 l with circulating freshwater which were permanently cooled using 2 cooling units (Titan; Aqua Medic).
The collected algal material (per replicate: 200 g fresh weight [FW] of U. rigida; 390 g FW for C. tamariscifolia and 400 g FW for E. elongata) were randomly assigned to the different replicates of the treatments and maintained under the same conditions (LCLN and ambient temperature) for 4 d.Subsequently, the experimental conditions of the different treatments (i.e.LCLN, HCLN, LCHN, HCHN) were established and algae acclimated at ambient temperature for 6 d before the temperature was raised by 4°C (on average) for 3 further experimental days.Data presented here are based on physiological measurements conducted immediately before the temperature increase ('ambient') and 3 d after the temperature increase ('+ 4°C').

Environmental parameters: control and measurement
Carbon treatments.In order to achieve an HC treatment, a computer-operated pH control system (AT control System, Aqua Medic) was used, with pH sensors (Aqua Medic T2001HC, Aqua Medic) located inside each of the twelve 60 l header tanks (Fig. 1b).
The system automatically recorded one measurement every 15 min and was programmed to initiate the supply of pure CO 2 via a solenoid valve as soon as the pH exceeded a threshold of 7.92 in the header tanks (corresponding to 700 ppm of CO 2 , HC treatment).When the pH returned to this value, CO 2 injection stopped.CO 2 was supplied to the secondary recirculating system of each header tank.This water circuit performed the role of a CO 2 reactor, facilitating rapid CO 2 dissolution and avoiding a sharp drop in pH.No upper pH limit was set for the LC treatment.
Nitrogen treatments.Nitrate was added every morning at 06:00 h GMT to each header tank using a 1 M KNO 3 pulse to achieve either a final nitrate concentration of 5 µM corresponding to LN conditions, or a concentration of 50 µM corresponding to HN conditions.The volumes of the 1 M KNO 3 solution required to reach each experimental concentration were estimated from the last sampling result of the previous day.In addition, 1 µM K 3 PO 4 was added in the experimental tanks to avoid phosphate limitation.For nutrient analysis, water samples (50 ml) were collected every day just before nutrient addition, after 1 h, at noon, and again in the evening.Water samples were kept in a freezer at −20 ± 1°C until applying segmented flow analysis using a Bran-Luebbe AA3 Autoanalyzer following the methods described by Grasshoff et al. (1999).The detection limit was 0.05 µM.
Temperature control.Water temperature was monitored using the same control system as for pH, with   temperature sensors also located in the header tanks.
In the second part of the experiment, temperatures were raised to simulate a daily temperature curve of 4°C above the ambient temperatures experienced during the previous days.Temperature was controlled in the freshwater circulating system in the eight 1000 l tanks.In addition, a diurnal temperature curve was established in each header tank using an AT control System connected to 150 W submersible heaters (Fig. 1b).
Monitoring conditions in vessels.Temperature, pH and conductivity were measured in each of the 48 vessels and header tanks 3 times per day using a multi-parameter probe (HI 982X; Hanna).No significant differences (ANOVA, p > 0.05) were found between the values measured in the vessels and their corresponding header tank (data not shown).Water loss via evaporation was monitored, and distilled water was added every day after sunset to the header tanks if salinity was higher than 38.
Solar radiation.Ambient photosynthetically active radiation (E PAR ) at the UMEGOA experimental site beside the tanks (underneath the mesh) was measured in air at 5 min intervals from sunrise to sunset using a spherical quantum sensor (Li 193 SA,Licor) connected to a Data Logger (Li-1400).These data are presented together with effective PSII quantum yield (ΔF/F m ').In addition, incident solar radiation was measured continuously in air using a UV-PAR Multifilter radiometer NILU-6 (Geminali).Levels of UVA (320−400 nm) and UVB (280−320 nm) radiation were calculated from the data of the different UV filters according to Høiskar et al. (2003).The integrated daily light fluence ex pressed as kJ m -2 was calculated for the duration of the experiment.The NILU-6 was located on the roof of the building of the Central Services Research of Málaga University (3.7 km from the UMEGOA experimental site).
In order to characterize the underwater light field within vessels, solar radiation (E PAR and UVR) was measured in the morning of the first day of the experiment just below (1 cm) the water surface and at the bottom of each vessel using Zippo-Hobo ®- U12-UV sensors (PAR: SQ-212, UVA: LPUVA01; Onset Computer).

Biological parameters
Chl a fluorescence of algae in tanks.The effective quantum yield of PSII (ΔF/F m ') of algae subjected to the conditions above was determined in 15 samples from every species and treatment every 2 h from sunrise to sunset using Diving-PAM Fluoro meters (Walz).
Pigment contents.For pigment analysis (chlorophyll and phycobiliproteins), 1 algal sample per vessel was frozen in liquid nitrogen until extraction.For chlorophyll determination, samples were then ground with a mortar and pestle for 20 min in 1 ml of 90 % (v/v) acetone neutralized with magnesium carbonate hydroxide and sea sand.After centrifugation at 1500 g for 10 min, each supernatant was used to determine chlorophyll concentrations spe ctrophotometrically (UV mini-1240 Spectrophotometer; Shimadzu) following Ritchie's equations (Ritchie 2008); chlorophyll contents are presented in mg g −1 dry weight (DW).
Phycobiliproteins (i.e.phycoerythrin and phycocyanin) were extracted from 0.1 to 0.2 g FW of E. elongata ground in phosphate buffer at pH 7.0.Pigments were quantified spectrophotometrically (UV mini-1240 spectrophotometer; Shimadzu) according to Beer & Eshel (1985) and are expressed in mg g −1 DW of seaweed material.
To acquire the ratio of FW to DW, 15 to 20 samples of known FW for every species were dried in an oven at 60°C for 24 h, followed by determination of final DW.

Statistical analyses
For each macroalgal species, experiments investigating impacts of OA and eutrophication were carried out in triplicate (n = 3), with 3 independent vessels used at each of the defined conditions of LCLN, LCHN, HCLN and HCHN.Chlorophyll a fluorescence (i.e.ΔF/F m ') was determined for 15 different specimens per vessel, with the average of 15 values reflecting the overall photosynthetic performance of algae within 1 tank, i.e. the average was considered as 1 replicate measurement and used for further statistical analyses (n = 3).Pigment concentrations were measured in 1 specimen per species per tank.All data are presented as mean ± SD.
Main effects of the fixed factors 'time of day' (TD, i.e. morning, noon and evening), 'carbon regime' (CR, i.e. low and high carbon) and 'nitrogen regime' (NR, i.e. low and high nitrogen) as well as their respective interactions on the dependent variables 'pigment content') of each species were analysed using 3-way model I ANOVAs.All data were normally distributed (Kolmogorov-Smirnov test), variances were homogeneous (Levene's test) and mean values were generally considered as significantly dif-ferent at p < 0.05.Where the Levene's test indicated heterogeneity of variances, even after transformation (this was the case for some pigment data), a more conservative α of 0.01 was applied to minimise the probability of a Type I error (Krassoi et al. 2008).The least significant difference (LSD) test was applied to differentiate between means when significance was found.

Light climate
The solar radiation measured at the University of Málaga campus (3.7 km from the experimental site) was stable throughout the acclimation period but declined on September 18 due to clouds and associated rainfall; the daily integrated light fluence dropped from nearly 10 000 to 4000 kJ m -2 and UVA+UVB from 1600 to 600 kJ m -2 (Fig. 2a).For the remainder of the experimental period, solar conditions were stable again and similar to the initial period.
Despite similar light conditions measured on tank surfaces, the 3 algal species experienced different light climates within their respective vessels (Fig. 2b−d).Even measurements of irradiance within the surface layer of Ulva rigida vessels showed some variation compared to the 2 other species due to U. rigida thalli intermittently floating above the Hobo ® sensor.Both PAR and UV radiation at the bottom of the U. rigida tanks were considerably reduced due to heavy shading by moving thalli (Fig. 2b); by contrast, there was only a small light reduction in Ellisolandia elongata tanks, as algal thalli were resting at the bottom of the tanks (Fig. 2d).Regarding Cystoseira tamariscifolia, both PAR and UVR were considerably lower at the bottom compared to surface solar radiation (Fig. 2c), and nearly no UV radiation was available at depth due to absorption by exuded phenolic compounds (Figueroa et al. 2014c).
Temperature within the header tanks showed considerable diurnal variation, but, in contrast to pH, the pattern was consistent for all species and treatments (Fig. 4).During the ambient temperature experi mental period, nighttime temperatures decreased from about 23 to 19°C until sunrise, and increased to maxima of about 24.5°C during midafternoon (17:00 h GMT) in all species and treatments.Lowest temperatures were observed in all tanks around 08:00 h GMT.Following the 4°C temperature increase, nighttime temperatures ranged from 24 to 25.5°C, and in creased sharply after sunrise at 08:00 h GMT to reach maxima of 29°C at around 17:00 h GMT.Although all tanks were cooled permanently, heating by ambient daytime air temperatures resulted in a diurnal increase in water temperature in all tanks; temperatures in '+ 4°C' treatments were higher both during night-and daytime (p < 0.00001; 4-way ANOVA) but despite variability did not vary significantly between species (p = 0.6945).
Temperatures were similar in all tanks containing the 3 different species and for all carbon and nitrogen treatments (Fig. 4).

Nitrate uptake
The nitrate provided in the morning was more rapidly taken up by U. rigida and E. elongata than by C. tamariscifolia (Fig. 5).One hour after the addition of nitrate, nearly 50% was taken up by U. rigida and 20% by the other 2 species; by evening, seawater in almost all vessels contained nitrate levels close to zero.When nitrate was added, uptake by C. tama ris cifolia was faster under LC than HC conditions, and HCHN was the only treatment where some nitrate remained in the tanks by evening (Fig. 5b).

Temporal variation in the effective PSII quantum yield
Results of 2 d of fluorescence measurements are presented in Fig. 6: 1 d at ambient temperature (after 6 d under the different treatments) and 1 d at ambient temperature '+ 4°C' (after 3 d at the in creased .ΔF/F m ' exhibited a diurnal pattern regardless of treatment and temperature regime in all species, but such variations were generally more pronounced in U. rigida than in C. tama riscifolia or E. elongata (Fig. 6).During the early part of the day, ΔF/F m ' decreased with increasing solar radiation (simultaneously measured beside the experi mental tanks underneath the mesh) and represented by E PAR in air (Fig. 6); the reduction in ΔF/F m ' was maximal at noon when the radiation was highest, and this was followed by an increase in ΔF/F m ' towards the evening.
In U. rigida, the reduction in ΔF/F m ' at noon appeared to be greater for LCLN than for the other car- bon-nitrogen regimes applied, i.e.U. rigida probably benefitted from elevated carbon and nitrogen levels at ambient temperatures (Fig. 6a).At increased temperature a more pronounced reduction in ΔF/F m ' at noon was again observed in both LCLN and HCLN, indicating that the low nitrogen may have caused the observed decrease in ΔF/F m ' (Fig. 6d).
In contrast to U. rigida, different carbon-nitrogen regimes had only marginal impacts on ΔF/F m ' and its diurnal pattern in C. tamariscifolia and E. elongata at either temperature (Fig. 6b,c,e,f).ΔF/F m ' of E. elongata appeared to be slightly lower for LCLN than for the other treatments in the '+ 4°C' temperature regime (Fig. 6f).

Chl a concentrations
Chl a levels differed considerably between the 3 algal species, with concentrations ranging from 1 to 3 mg g -1 DW for C. tamariscifolia, 0.5 to 1.5 mg g -1 DW for U. rigida, and 0.1 to 0.3 mg g -1 DW for E. elongata (Fig. 7).After the increase in temperature, there was no consistent diurnal pattern, and lower chl a content was measured in the morning in C. tama ris cifolia but higher levels occurred in E. elongata (Table 1).Nitrogen and/or carbon significantly im pacted the different algal species: higher chl a content was observed under higher nitrogen regimes in U. rigida at ambient temperature and in C. tamariscifolia at both temperatures.The HC treatment resulted in lower chl a levels in U. rigida and E. elongata at both temperatures, and the interaction time of day × carbon was significant for U. rigida at both temperatures (Table 1).

Accessory pigment concentrations
Chl b content of U. rigida ranged from 0.4 to 1.0 mg g -1 DW (Fig. 8a).At ambient temperature, carbon and nitrogen had significant impacts on chl b content with higher levels observed in LC than in HC, and in HN than in LN treatments (Table 2).U. rigida grown in the LCHN treatment had the highest chl b level; also, the interactions carbon × nitrogen and time of day × nitrogen were significant.At increased temperature, only effects of carbon and the inter action time of day × carbon were significant.
In C. tamariscifolia, 0.2 to 0.6 mg g -1 DW of chl c were found, with no variation amongst carbon and nitrogen treatments at ambient or increased temperature (Fig. 8b; Table 2).At ambient temperature, a slight, but not significant, increase in chl c occurred; following the temperature increase, chl c values at noon were significantly higher than those measured in the morning or evening.
E. elongata contained between 1.5 and 3.0 mg phycoerythrin g -1 DW and between 0.30 and 0.65 mg treatments were adjusted every morning at 06:00 GMT phycocyanin g -1 DW (Fig. 8c,d).Neither carbon nor nitrogen treatments, nor the increase in temperature had an impact on phycobiliprotein contents (Table 2).Phycoerythrin contents were slightly, but not significantly, higher in the evening at ambient temperature.On the other hand, time of day affected phycocyanin levels of E. elongata at ambient temperature, with the highest levels observed in the evening, and similar levels in the morning and noon samples.

Experimental setup
This paper describes the conditions in the experimental tanks used during the GAP workshop at the University of Málaga in September 2012, and the short-term photosynthetic responses of Ulva rigida, Cystoseira tamariscifolia and Ellisolandia elongata The experimental set-up exposed 3 macroalgal species to 4 combinations of nitrogen and carbon treatments in replicated outdoor tanks; natural solar irradiance was reduced by about 35 and 39% (E PAR and UVR, respectively).Temperature and pH (by CO 2 -enrichments) in tanks were automatically adjusted and resulted in 4 distinct nutrient and carbon conditions for each species.Additionally, pH exhibited considerable diurnal variability in the uncontrolled LC treatments, prob ably caused by a high metabolic activity, especially in the tanks containing U. rigida.This is consistent with previous studies on Ulva (formerly Enteromorpha) intestinalis.from rockpools where photosynthetic activities significantly increase pH in small volumes of water (pH > 9.4, Larsson et al. 1997), though no negative impacts on photosynthesis were observed here.Previous studies have reported an insen sitivity of Ulva spp. to high pH due to its capacity to utilize bicarbonate (Axelsson et al. 1995, Frost-Christensen & Sand-Jensen 1990).Averaged pH in LC and HC in U. rigida tanks differed considerably throughout the duration of the experiment, whereas for C. tamariscifolia and E. elongata the differences were less pronounced.The ob served species-specific differences in pH profiles for the different treatments suggest that interpretation of physiological responses should be restricted to a comparison of responses to HC/LC treatments within species, and be interpreted only with caution to explain differences amongst species.On the other hand, daily fluctuations in tem peratures in tanks were more consistent for all species (and treatments); the observed diurnal increases were consistently impacting all species and treatments, and the required increase in temperature by 4°C was successfully achieved, with only small, negligible variability between species and treatments.Any changes in physiological parameters observed in response to increases in temperature should thus be comparable amongst species.

Light climate
Spectral irradiances measured at the University of Málaga campus (3.7 km from the experimental site) were representative of natural conditions before the 35% reduction in solar irradiance caused by the mesh; some small shading effects were likely to occur due to different sun angles, but it can probably be assumed that these were, on average, similar for all tanks throughout a daily cycle and thus did not have any impacts on particular treatments, species or  LN: low nitrogen, HN: high nitrogen) as well as interactions of each of these fixed factors were computed for each species by a 3-way ANOVA; bold = significant p-values (α = 0.05 was used if variances were homogenous; if they were heterogeneous, a more conservative α = 0.01 was employed, indicated by italic font).Where a significant difference was found by ANOVA, means were compared using the least significant difference (LSD) procedur replicate.In addition to the expected diurnal natural variation in light climate, daily radiation differed between experimental days due to partial cloud cover, but again differences were similar for all species and treatments.Changes in natural irradiance may potentially have impacted on observed differences be tween measurements taken on different days which also coincided with the increase in temperature by 4°C; on the other hand, most significant changes in light climate, expectedly, occurred over the daily cycle and caused the diurnal variation in e.g.photosynthetic parameters such as ΔF/F m '.

Light climate within vessels
Whilst every attempt was made to provide different species and vessels with the same light climate, species-specific effects were observed which may have resulted in different micro-climates within vessels, and these were likely to be similar for all treatments within species.Thus differences in physiological parameters measured were most likely due to dif ferent treatments, but a comparison between such effects in different species needs to be interpreted in view of the different light climates experienced by different species within vessels.The resultant light micro-climates were caused by the different morphologies of the different species, e.g.severe shading was observed within U. rigida vessels due to movement of thalli caused by water motion, which produced efficient mixing of algal fronds which were then used for the assessment of physiological para meters.In U. rigida vessels, the light climate fluctuated greatly and may also have resulted in light flecks (Kübler & Raven 1996).Sam-  2. Statistical summary of the effects of time and carbon-nitrogen treatment on accessory pigments (chl b in Ulva rigida, chl c in Cystoseira.tamariscifolia, and phycobiliproteins in Ellisolandia elongata) kept at ambient or + 4°C temperature.Main effects of time of day (TD; M: morning, N: noon, E: evening), carbon regime (CR; LC: low carbon, HC: high carbon), and nitrogen regime (NR; LN: low nitrogen, HN: high nitrogen) as well as interactions of each of these fixed factors were computed for each species by a 3-way ANOVA; bold = significant p-values (α = 0.05 was used if variances were homogenous; if they were heterogeneous, a more conservative α = 0.01 was employed, indicated by italic font).Where a significant difference was found by ANOVA, means were compared using the least significant difference (LSD) procedure ples of U. rigida would thus have experienced a mixed and rapidly changing light climate, but it is also possible that some pieces were based at the bottom of the tank for extended periods and thus became shade-acclimated.For physiological analyses, samples were consistently taken from the surface, but it is possible that individual algae had only just reached the surface from lower layers following water movement within the vessels that allowed mixing within tanks.By contrast, due to their weight, all C. tamariscifolia and E. elongata specimens were permanently placed at the bottom of the vessels, even though the same vigorous water movement was applied, and it is likely that these 2 species experienced a more constant light climate.In the case of C. tama riscifolia, water colour changed over time, which explained the differences in light levels between surface and bottom measurements; most likely this change in water colour and light penetration was related to the exudation of phenolics released by C. tamariscifolia (Figueroa et al. 2014c); such ex u da tion of screening compounds has already been observed under high light for the brown macroalga Macro cystis integrifolia (Swanson & Druehl 2002).In the confined environment of the experimental vessels, this exudation resulted in reduced light, particularly UV, availability.It is also likely that samples removed from C. tamariscifolia vessels on different days had experienced slightly different light climates due to the summative release (and subsequent re duction/change in light) over time.By contrast, the removal of biomass over time (for different measurements and analyses) may have reduced the degree of shading and thus increased the availability of light within U. rigida vessels.
The most consistent light climate was probably experienced by E. elongata, with only small differences measured between surface and bottom light within tanks, and again the use of samples from the surface was always attempted, although some movement of samples may have occurred that caused some variability between replicate samples.

Nitrogen uptake
Despite spiking all HN tanks equally every morning, resultant N levels in tanks varied between species, although in each case nitrogen-enrichment (HN) resulted in concentrations about 4 times those of unenriched (LN) treatments.This suggests that significantly different N-regimes were achieved.As might be expected due to its high surface-to-volume ratio, U. rigida rapidly took up nitrogen provided to HN treatments within about 5 h, with concentrations measured during mid-and late afternoon reaching those of unenriched vessels.An assessment of possible effects attributed to N-enrichment (or the lack thereof) should consider that elevated N conditions were still present in C. tamariscifolia vessels, whereas in the case of U. rigida and E. elongata, such effects may be caused by nutrients previously absorbed and thus by associated internal N-storage; it is likely that N-depletion occurred from noon onwards.With the exception of N levels in C. tamariscifolia exhibiting some variability, there were no differences between HCHN and LCHN treatments.
The conditions provided in the tanks therefore represented realistic nutrient regimes experienced by algae in their natural habitats on the southern coast of Spain, where nutrient pulsing is a well-documented phenomenon (e.g.Ramírez et al. 2005).

Pigments
Nitrogen enrichment is likely to affect pigment contents including chlorophyll (Figueroa et al. 2009), protein (Ribeiro et al. 2013) andMAA (Figueroa et al. 2010) concentrations.Previous studies have shown a direct relationship between N-uptake and chl a and b levels in Ulva sp.(Altamirano et al. 2000) and chl a in Gracilaria conferta (Figueroa et al. 2010).Indeed, in our study, all specimens of U. rigida grown under HN had higher chl a and chl b concentrations.The relationship between seawater N concentrations and pigment contents in brown algae is less well documented, although under controlled conditions, pigment levels increased e.g. in Ascophyllum nodosum (Stengel & Dring 1998) and several kelp species (Korb & Gerard 2000); also several observations from field populations report on pigment changes in response to in situ nutrient levels (Henley & Dunton 1995).In line with its coarsely branched morphology and thus lower surface-to-volume ratio compared to the other species, N uptake by C. tamariscifolia (as indicated by disappearance from the media) was less pronounced than that by the other species, although both LCHN and HCHN treatments resulted in elevated chl a levels.
Chl a levels generally exhibited some variability.The most striking effects were observed in response to N-enrichment and these were most pronounced in U. rigida under LC conditions before the temperature rise, but relatively consistently higher levels also oc -curred in HN treatments of C. tamariscifolia.Differences observed in E. elongata were more likely due to HC treatments with no consistent trends observed for HN impacts, even though N was taken up.At elevated temperatures, observed changes in chl a levels were caused by interactions between light, nutrients and carbon treatments.In U. rigida and E. elongata in particular, HC appeared to reduce chl a concentrations.
In this study, phycobiliprotein (i.e.phycoerythrin and phycocyanin) contents of E. elongata were not influenced by carbon and nitrogen treatment, contrary to previous studies: in Gracilaria spp.phycoerythrin increased with seawater N concentration and internal N content (Rosenberg & Ramus 1982, Figueroa et al. 2010).On the other hand, Gao & Zheng (2010) observed a decrease in the phycoerythrin content of Corallina sessilis when exposed to high CO 2 (1000 ppmv) for 30 d.The short experimental duration of the present study may explain the lack of change in phycobiliprotein levels.However, some diurnal variation was observed, with the highest phycocyanin levels measured in E. elongata in the evening.
To assist with the interpretation of the impacts of the different treatments described here, further biochemical analyses were conducted (e.g.soluble proteins, fatty acids, MAAs, phenolics; Figueroa et al. 2014c).

Effective PSII quantum yield
In all species, considerable diurnal variation was observed on the 2 experimental days, with lower yields during periods of high solar radiation in the afternoon; greatest impacts on yield under different carbon and nitrogen treatments were observed for U. rigida where LN caused a slow recovery at ambient temperature and a fast decline under the 'ambient + 4°C' treatment.Recovery of yields occurred in all species and under both temperature regimes, but, typically, yields of U. rigida were higher than those of C. tamariscifolia and E. elongata, corresponding to values previously reported for these species and other Chlorophyta, Phaeophyceae and Rhodophyta (e.g.Büchel & Wilhelm 1993).
In U. rigida, yields decreased more rapidly at elevated temperatures but on this day, natural solar radiation was also higher.Under these more extreme conditions, recovery under LN conditions was slower and suggests a N requirement for photoprotective mechanisms in U. rigida (Henley et al. 1991).

Conclusions
The established multi-tank system allowed the successful conditioning of 3 macroalgal species under semi-natural light conditions and modified carbon, nitrogen and temperature regimes.Different micro-scale light environments were observed within tanks which potentially influenced physiological responses, but most of the significant physiological changes occurred in response to diurnal variation in light climate.However, responses were species-specific and may have further been affected by in vitro experimental conditions (self-shading, release of phenolics).Although water temperatures were highly variable diurnally, the experimental approach succeeded in increasing the average temperature by 4°C; however this was accompanied by a larger temperature variation within tanks.This short-term temperature in crease did not appear to impact the ability of any species to recover to morning levels in any of the treatments.The most significant impacts of the carbon and nitrogen treatments were observed in U. rigida, probably due to its sheet-like morphological structure, high N dependency and generally high metabolic rates; the observed physiological responses were impacted by interactions of the environmental factors carbon and nitrogen, whereas temperature effects were less pronounced for all species.Physiological responses of U. rigida appeared to be most sensitive to rapid changes in irradiances and both carbon and nutrient treatment, whilst C. tamariscifolia was impacted the least by any environmental factor.
The 3 species grow in close proximity on exposed rocky outcrops in the low intertidal in southern Spain; however, the differences in their physiological responses were likely due to a combination of morphological structures (e.g.shrubby nature of C. tamariscifolia), low productivity of the calcified E. elongata and photoprotective mechanisms in C. tamariscifolia and E. elongata.In their natural habitat all species occur predominantly in dense clumps, with ex treme self-shading providing at least partial protection from certain stressors, including solar radiation and desiccation, at least for E. elongata.
Analysis of further biochemical parameters and more detailed PAM fluorescence measurements conducted under controlled conditions are presented in other papers (Figueroa et al. 2014b,c) and will help further elucidate the protective mechanisms employed by the 3 species.Further, longer-term studies should be conducted which incorporate different temporal scales and assess the impacts at multiple, physiological, biochemical and genetic levels.
Acknowledgements.The authors are grateful for the following financial contributions supporting the GAP 9 Workshop 'Influence of the pulsed-supply of nitrogen on primary productivity in phytoplankton and marine macrophytes: an experimental approach' from: Walz GmbH, Effeltrich, Germany (also for making available several PAM fluorometers), Redox, the University of Málaga, Ministerio de Economía y Competitividad del Gobierno de España (Acción Complementaria CTM2011-15659-E), the Spanish Institute of Oceanography, the (Irish) Marine Institute and the Marine Research Sub-programme of the National Development c

Fig. 2 .
Fig. 2. Light climate: (a) Daily fluence values of photosynthetically active radiation (PAR; 400−700 nm), UVA (320−400 nm) and UVB (280−320 nm) in the air during the experiment; the arrows show the sampling dates.Light climate (relative to surface) at different depths (1 cm below surface and at bottom) within vessels at the beginning of the experiment (19 Sep 2012) with UV radiation (dotted lines) and PAR (solid line) for (b) Ulva rigida, (c) Cystoseira tamariscifolia and (d) Ellisolandia elongata.Note the variability in light climate at both surface and bottom of the U. rigida tanks due to floating thalli causing shading.There was little movement of other species within the other tanks

Fig. 3 .
Fig. 3. Daily variation of pH in the different carbon and nutrient treatments (C: carbon; N: nitrogen; L: low; H: high) during day and night periods on (a) Ulva rigida, (b) Cystoseira tamariscifolia and (c) Ellisolandia elongata.Data are means ± SD of the values measured during the 4 days of the experiment at different times of day

166Fig. 4 .
Fig. 4. Daily variation in temperature in tanks during the different carbon and nutrient treatments on (a) Ulva rigida, (b) Cystoseira tamariscifolia and (c) Ellisolandia elongata before increasing temperature (left) and after increasing temperature by 4°C (right); see Fig. 3 for treatment abbreviations.Data are means ± SD of values averaged over 4 and 3 d for the first and second part of the experiment, respectively

Fig. 5 .
Fig. 5. Daily variation of water nitrate concentration in tanks during the different carbon and nutrient treatments on (a) Ulva rigida, (b) Cystoseira tamariscifolia and (c) Ellisolandia elongata; see Fig. 3 for treatment abbreviations.Data are means ± SD of the values measured during the 7 d of the experiment.The water nitrate concentrations of the different treatments were adjusted every morning at 06:00 GMT

Fig. 6 .
Fig. 6.Daily evolution of the effective PSII quantum yield (ΔF/F m ') in 3 macroalgae exposed to different carbon and nitrogen regimes and reduced solar radiation at an ambient (a−c) and 4°C-increased ambient temperature (d−f).Ulva rigida (a,d), Cystoseira tamariscifolia (b,e) and Ellisolandia elongata (c,f) were subjected to either low or high carbon (LC, HC respectively) at low and high nitrogen (LN, HN respectively) levels.ΔF/F m ' was determined on algae in tanks; data are means ± SD (−SD for LN, + SD for HN; n = 3).E PAR was measured in air immediately beside the experimental tanks using a spherical quantum sensor connected to a data logger

Fig. 7 .
Fig. 7. Morning (M), noon (N) and evening (E) chl a contents in 3 macroalgae exposed to different carbon and nitrogen regimes and natural solar radiation at ambient (left-hand side) and 4°C-increased ambient temperature (right-hand side).Ulva rigida (a), Cystoseira tamariscifolia (b) and Ellisolandia elongata (c) were subjected to either low or high carbon (LC, HC respectively) at low and high nitrogen levels (LN, HN respectively).Data are means + SD

Table 1 .
).Data are means + SD Statistical summary of the effects of time and carbon-nitrogen treatment on chl a content in 3 different macroalgae (Ulva rigida, Cystoseira tamariscifolia and Ellisolandia elongata) kept at ambient or + 4°C temperature.Main effects of time of day (TD; M: morning, N: noon, E: evening), carbon regime (CR; LC: low carbon, HC: high carbon), and nitrogen regime (NR;