Response of phytoplankton to enhanced atmospheric and riverine nutrient inputs in a coastal upwelling embayment

1 Response of phytoplankton to enhanced atmospheric and riverine nutrient inputs 1 in a coastal upwelling embayment 2 Teixeira, I.G.1,3*; Arbones, B.1; Froján, M.1; Nieto-Cid, M.1; Álvarez-Salgado, X.A.1; 3 Castro, C.G.1; Fernández, E.2; Sobrino, C.2; Teira, E.2; Figueiras, F.G.1 4 1 CSIC, Instituto de Investigacións Mariñas. Eduardo Cabello 6, 36208 Vigo. Spain. 5 2 Universidade de Vigo, Campus Lagoas-Marcosende, 36310 Vigo, Spain. 6 3 CESAM – Centre for Environmental and Marine Studies & Department of Biology, 7 University of Aveiro, Campus de Santiago, 3810‑193 Aveiro, 8 Portugal. 9 *Corresponding author: [email protected] 10 KEYWORDS: phytoplankton, nutrients, atmospheric deposition, riverine inputs, 11 upwelling system, NW Spain, Ria de Vigo 12 13 ABSTRACT 14 Over the past decades, as a consequence of human activity, there was an increase in 15 nutrient inputs to the ocean and they are expected to enhance even more in the future. 16 Coastal areas, accounting for a significant proportion of marine primary productivity, 17 are the most vulnerable zones to anthropogenic impacts. The response of phytoplankton 18 communities to an increase in organic and inorganic nutrients levels from natural 19 allochthonous sources was assessed in microcosm experiments conducted in a coastal 20 system affected by intermittent upwelling events (Ría de Vigo, NW Iberia). Three 21 nutrient addition experiments were performed in spring, summer and autumn, when 22 2 surface water was supplemented with 5 and 10% of atmospheric and riverine matter. 23 Pico-, nanoand microphytoplankton abundances, chlorophyll a concentration (Chl a) 24 and primary production rates (PP) were measured and compared with those in the 25 control seawater sample (without additions) after 48 h of incubation. Simultaneous 26 experiments with controlled additions of inorganic and organic nutrients were also 27 performed in order to describe the limiting nutrient for phytoplankton growth at each 28 experiment. The composition of the matter inputs and the structure of the phytoplankton 29 communities determined the type of response observed. Phytoplankton responses varied 30 among seasons, being positively correlated with dissolved inorganic nitrogen (DIN) 31 concentrations. As expected, the phytoplankton responses to external nutrient inputs 32 were stronger under low nutrient levels (summer) than when phytoplankton was already 33 growing in nutrient replete conditions (spring). Null and negative responses to the 34 natural inputs were observed in autumn, which suggests that the oceanic phytoplankton 35 advected to this coastal system during downwelling events could be occasionally 36 inhibited by these nutrient inputs. In a future global change scenario, characterized by 37 enhanced nutrient inputs from riverine and atmospheric origin, the response of 38 phytoplankton communities will strongly depend on the concentration and chemical 39 composition of these inputs and on the structure of phytoplankton communities able to 40 respond to them. 41 42 1. Introduction 43 Phytoplankton organisms are the main primary producers in the sea being 44 responsible for about 50% of the world primary production (Field et al. 1998). Coastal 45 areas, despite representing only 7% of ocean’s surface, account for 15-20% of global 46 3 marine primary productivity (Wollast 1993, 1998; Laruelle et al. 2009). Enhanced 47 nutrient fluxes from continental runoff, the atmosphere and the adjacent open ocean, 48 and the efficient nutrient recycling based on a closed coupling between pelagic 49 production and benthic regeneration are the reasons behind the high productivity of the 50 coastal zone (Walsh 1991; Wollast 1998). All these biogeochemical processes are 51 specially intensified in coastal upwelling regions because of the enhanced entry of 52 nutrients from the adjacent ocean in response to intense and persistent equatorward 53 winds (Walsh 1991; Wollast 1998). 54 Due to its close proximity to human populations, coastal areas are also the most 55 vulnerable zones to anthropogenic forcing (Jickells 1998). Inorganic and organic 56 nutrients and pollutants resulting from human activities reach coastal areas through 57 atmospheric deposition, continental runoff and groundwater effluents (Jickells 1998; 58 Doney 2010; Statham 2012). Over the past decades, there was an increase in these 59 nutrient inputs to the ocean and they are expected to expand even more in the future 60 (Anderson et al. 2002; Galloway and Cowling 2002; Galloway et al. 2004; Duce et al. 61 2008). Alterations in the magnitude and composition of matter inputs in coastal areas 62 may induce significant changes in phytoplankton communities inhabiting those systems 63 (Paerl 1997; Peierls and Paerl 1997; Seitzinger and Sanders 1999; Spatharis et al. 2007). 64 Extreme cases of nutrient enrichment may involve the development of massive 65 phytoplankton blooms, including harmful algal species (Paerl 1997; Anderson et al. 66 2002). 67 The Ría de Vigo is a productive embayment in the coastal upwelling system of 68 the Northwest Iberian Peninsula (Fraga 1981). Seasonal patterns in phytoplankton 69 composition and productivity in this system are closely related to the hydrographic 70 variability (Figueiras and Ríos 1993; Figueiras et al. 2002). Upwelling events, occurring 71 4 usually between March and October, introduce nutrient rich subsurface waters into the 72 photic layer and are followed by high concentrations of phytoplankton cells and 73 chlorophyll. On the contrary, the winter period is favorable to downwelling and it is 74 characterized by lower biomass accounted for small-size phytoplankton cells (Figueiras 75 et al. 2002). River discharge and atmospheric deposition may also constitute significant 76 sources of nutrient inputs to this coastal embayment, although to a lesser extent as 77 compared with upwelling episodes (Gago et al. 2005; Rodríguez and Macías 2006; 78 Alonso-Pérez and Castro 2014; Fernández et al. in 2016). The high productivity 79 characteristic of this system sustains an intensive mussel production and significant 80 catches of several fish and shellfish species, which are relevant for the economy of the 81 region (Figueiras et al. 2002; Froján et al. 2014). Thus, understanding the response of 82 the organisms in the basis of the food web to an eventual increase in nutrients inputs in 83 the Ría de Vigo may be determinant to anticipate its consequences to the whole 84 ecosystem. 85 The outcomes of an increase in nutrients on coastal microbial plankton in this 86 system was tested in previous experimental studies, which focused on the effect of 87 controlled inorganic and organic nutrient additions (Martínez-García et al. 2010; Teira 88 el al. 2011) and natural additions of rainwater (Teira et al. 2013; Martínez-García et al. 89 2015). In the present study, we aimed at examining the response of phytoplankton 90 communities in the Ría de Vigo to realistic increasing amounts of dissolved matter 91 inputs from both atmospheric and riverine origin. To calculate the current average 92 riverine and atmospheric inputs to the Ría de Vigo we considered the average river flow 93 of the River Oitabén-Verdugo, 17 m3 s–1 (Gago et al. 2005) and the average 94 precipitation to the Ría de Vigo, 7.7 mm d–1. Considering the surface area of the ría is 95 (174 km2), the mean surface mixing layer (2 m) and the average flushing time of this 96 5 layer (5 days), the surface mixing layer of the ría must contain about 2% of river and 97 2% of rainwater. Therefore, the additions of 5% and 10% tested here would serve to 98 obtain the response of the Ría de Vigo to future global change scenarios in which 99 human activities increase the quantity without altering the quality of riverine and 100 atmospheric inputs. Furthermore, concomitant experiments with controlled additions of 101 inorganic and organic nutrients were also performed in order to describe the limiting 102 nutrient for phytoplankton growth at each season. 103 104 2. Materials and Methods 105 2.1 Sampling 106 Seawater for the experiments was sampled from the middle of the Ría de Vigo 107 (42º14.09’ N, 8º47.18’ W) during the productive season in spring (May) and summer 108 (July), and after the transition to the unproductive period in autumn (October) 2013 109 (Fig. 1). Temperature, salinity and in situ fluorescence down to 25 m depth were 110 obtained with a SBE 9/11 CTD probe and a Seatech fluorometer attached to a rosette 111 sampler. Sub-surface seawater (3-4 m) was collected in 12 L acid-clean Niskin bottles 112 and filtered through a 200 µm pore size mesh to remove larger zooplankton into a large 113 acid-clean carboy to transfer it to the coastal station (Fig. 1) to prepare the addition 114 experiments. 115 116 2.2 Ekman transport and runoff 117 The Ekman transport (Qx; m2 s−1) perpendicular to the coast, a proxy for the 118 occurrence and intensity of coastal upwelling, was calculated from wind direction and 119 velocity, recorded at Cabo Silleiro buoy (Fig. 1) according to Bakun (1973): 120 6 Ría de Muros Ría de Arousa Ría de Pontevedra A Cape Finisterre Ría de Vigo -10.0 -9.5 -9.0 -8.5 -8.0 Longitude (E) 42.0 42.5 43.0 43.5 Latitude (N) Silleiro Buoy -8.9 -8.8 -8.7 Longitude (E) 42.1 42.2 42.3 42.3 42.4 Latitude (N) 75m 20m 40m B Ría de Vigo + IIM - CSIC X City Hall Coastal station 𝑄𝑥= 𝜌𝑎∗𝐶∗ [𝑉]∗𝑉 𝑦 𝑓∗𝜌𝑤 (1) 121 where ρa is the air density (1.22 kg m−3), C is an empirical drag coefficient (1.3 × 10−3, 122 dimensionless), [V] is the wind speed (m s−1) with component Vy, ƒ is the Coriolis 123 parameter at this latitude (9.95 × 10−5 s−1), and ρw is the density of seawater (~1025 kg 124 m−3). The sign of Qx was changed to associate positive values with offshore transport 125 (upwelling) of surface waters. 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 Fig. 1: (A) Map of the west Galician coast with the location of the Silleiro buoy (◆) and (B) 144 map of the Ría de Vigo with the location of the sampling station (●), the river Oitabén-Verdugo (▲) and 145 the meteorological stations in the IIM (+) and in the City Hall (X). The coastal station where addition 146 experiments took place is indicated by an arrow. 147 148 7 Runoff was calculated using data from River Oitabén-Verdugo (Fig. 1), according to the 149 method described in Otero et al. (2010). 150 151 2.3 Preparation of natural waters concentrates for the addition experiments 152 River and rainwater samples, and <10 µm atmospheric particles were collected 153 and processed to obtain concentrates of the riverine and atmospheric matter inputs to the 154 Ría de Vigo. The objective was to reduce the volume of the original water samples 10-155 fold maintaining their chemical composition, i.e. without addition or loss of any 156 component present in the natural samples. In this way, it was possible to test the effect 157 of an increase in natural matter using small water volumes and without significant 158 changes in salinity. 159 The River Oitabén-Verdugo (Fig. 1) was sampled a week before the addition 160 experiments in April, July and October 2013. The water samples were collected 161 upstream of the freshwater-seawater interface, to guarantee that the chemical 162 composition of the river samples represented that of the water that mixes with the 163 seawater of the Ría de Vigo. Five liters of each sample were gravity filtered through a 164 pre-washed (with 10 L of ultrapure water) dual-stage (0.8 and 0.2 µm) filter cartridge 165 (Pall-Acropak supor Membrane). The filtrate was then concentrated 10-fold using 166 rotatory evaporation with a Buchi R215 evaporator under mild conditions (bath 167 temperature: 25 ºC, vacuum: 13 mbar, condenser: acetone/CO2) to avoid breakage of 168 any organic compound present in the original water samples. Analysis of the 169 concentration of inorganic (ammonium, nitrite, nitrate and phosphate) and organic 170 (dissolved organic carbon and nitrogen) substrates confirmed that the samples were 171 concentrated quantitatively but maintaining their original composition. 172 8 A MTX rainwater sampler (model FAS005AB) and a high volume PM10 MCV 173 PM1025 sampler (model CAV-A/MS) installed at the roof of the Instituto de 174 Investigaciones Marinas (IIM-CSIC) (Fig. 1) collected samples of wet and dry 175 deposition to the Ría de Vigo. The MTX sampler was equipped with a humidity sensor 176 opening the system only when it was raining and allowing sampling just the wet 177 fraction of the atmospheric deposition. Rainwater was collected from four weeks to one 178 week before the addition experiments. Samples were taken daily and frozen 179 immediately after collection. A week before the experiment, the daily samples were 180 thawed at ambient temperature, mixed in one volume (6 L), and quantitatively 181 concentrated following the same procedure as for the riverine samples. Rainwater was 182 collected only for the experiments in spring (May) and autumn (October) 2013 because 183 wet deposition was very scarce the weeks before the summer experiment in July (36 184 mm accumulated from 11 June to 10 July; meteorological station of the Vigo city hall – 185 Fig. 1). The high volume sampler was used to collect atmospheric particles (1−10 µm) 186 on precombusted (450 ºC, 4 h) 140 mm GF/F filters (48 h sampling). The particles were 187 collected the week before each of the three addition experiments, operating during 48 h 188 at a rate of 30 m3 h–1. Then, the water-soluble fraction (WSF) of one eighth of the filter 189 was extracted in 400 mL of the corresponding rainwater concentrate by mechanical 190 stirring during 40 min. In the summer experiment, the WSF was extracted in milli-Q 191 water. These proportions (1/8 of the filter in 400 mL of water) were decided to obtain 192 10-fold the expected concentrations based on previous information about the 193 composition of wet and dry deposition to the Ría de Vigo (Teira et al. 2013; Martínez-194 García et al. 2015). Final mixed extracts were filtered through precombusted (450 ºC, 4 195 h) 47 mm diameter Whatman GF/F filters in an acid-cleaned glass filtration system, 196 under low N2 flow pressure, to be chemically characterized and used in the experiments 197 9 NATURAL Riverine Atmospheric 0% 5% 10% Control Inorganic Organic Mixed CONTROLLED as atmospheric concentrate. Similarly to riverine concentrates, quantitative 198 concentration was maintained except for the silicate because the reduction of the 199 rainwater volume was carried out in a glass rotary evaporator and the atmospheric 200 particles were collected onto a glass fiber filter. 201 202 2.4 Natural and controlled addition experiments 203 For both natural and controlled addition experiments, 4 L UV-transparent Whirl-204 pak® bags were gently filled under dim light conditions until 2 L of capacity. For the 205 natural addition experiments, riverine and atmospheric concentrates were added to 206 subsurface seawater, collected as described in Section 2.1, in proportions of 0% (1% 207 ultrapure water: 99% seawater), 5% (0.5% concentrate: 0.5% ultrapure water: 99% 208 seawater) and 10% (1% concentrate: 99% seawater) of the original (previous to 209 concentration) riverine and atmospheric materials (Fig. 2). These proportions also 210 ensured that the final salinity of the samples was kept constant independently of the 211 amount of extract added. 212 213 214 215 216 217 218 219 220 Fig. 2: Schematic representation of the natural and controlled addition treatments performed at each 221 season (see details in section 2.4). 222 223 16 The levels of significance of t-tests, performed between each treatment and the control, are: *p<0.05, 354 **p<0.01, ***p<0.001. 355 In summer, Chl a and PP rates responded positively in all treatments, including 356 natural and controlled additions (Fig. 4c, d). There were only two exceptions: PP rates 357 did not respond to organic inputs and the response of Chl a to the 10% riverine input 358 was marginally significant (P=0.07). In general, the response ratio in this experiment 359 was higher than in spring, with the response ratios to the natural inputs being higher 360 than 1.5 and to the controlled inorganic inputs reaching 4.5 for Chl a and 7 for PP (Fig. 361 4a-d). In autumn, there was a negative response in Chl a to the natural inputs (~0.85) 362 and to the organic controlled additions (~0.40) (Fig. 4e). On the contrary, inorganic and 363 mixed controlled additions led to a slight increase in Chl a concentrations. PP rates 364 increased in response to the 10% riverine inputs and to the mixed treatment, and 365 decreased in response to the organic additions (Fig. 4f). 366 Table 3: Output of multivariate General Linear Model analysis for Chl a, PP and the several 367 phytoplankton groups’ response ratios as dependent variables and Experiment and Treatment as fixed 368 factors. Synecho: Synechococcus-type cyanobacteria, APF: autotrophic picoflagellates, ANF: autotrophic 369 nanoflagellates. 370 Tests of Between-Subjects Effects Source F Sig. Experiment Chl a 23.860 0.001 PP 34.076 0.000 Synecho 9.405 0.010 APF 20.515 0.001 ANF 24.452 0.001 Diatoms 1.333 0.323 Treatment Chl a 15.742 0.001 PP 24.922 0.000 Synecho 3.788 0.052 APF 57.782 0.000 ANF 10.840 0.003 Diatoms 0.946 0.519 Experiment * Treatment Chl a 3.385 0.057 PP 10.331 0.002 Synecho 1.638 0.262 APF 11.154 0.002 ANF 3.579 0.050 Diatoms 1.073 0.484 17 Table 4: Bonferroni post hoc multiple comparisons from General Linear Models analysis with the 371 response ratios of Chl a, PP and the several phytoplankton groups as dependent variables and the 372 experiment and treatment as fixed factors. Synecho: Synechococcus-type cyanobacteria, APF: autotrophic 373 picoflagellates, ANF: autotrophic nanoflagellates. For treatment, only significant results are shown. 374 Experiment Sig. Chl a Spring Summer 0.024 Autumn 0.011 Summer Autumn 0.001 PP Spring Summer 0.001 Autumn 0.052 Summer Autumn 0.000 Synecho Spring Summer 0.012 Autumn 0.136 Summer Autumn 0.510 APF Spring Summer 0.545 Autumn 0.001 Summer Autumn 0.012 ANF Spring Summer 0.182 Autumn 0.001 Summer Autumn 0.013 Diatoms Spring Summer 1.000 Autumn 0.682 Summer Autumn 0.499 Treatment Sig. Chl a Inorganic Natural <0.01 Organic 0.003 PP Inorganic All <0.01 Synecho Atm 10% Mixed 0.037 APF Mixed All <0.01 organic All <0.01 ANF Inorganic Natural <0.05 Organic 0.002 Mixed Organic 0.017 375 376 377 378 18 Statistical analyses (Table 3) demonstrated that both experiment and treatment 379 had a significant effect in the response ratios of Chl a and PP. The interaction between 380 these two factors was also significant for PP response ratios. Thus, Chl a and PP 381 response ratios were significantly different among the three experiments (p < 0.01), 382 except for the PP response ratio between spring and autumn (Table 4). Among 383 treatments, significant differences were found for Chl a and PP response ratios between 384 natural inputs and the inorganic treatment in the controlled additions. The response to 385 the inorganic treatment was also significantly different from the organic treatment for 386 Chl a and from the mixed and organic treatments for PP. No significant differences 387 were found in Chl a and PP response ratios among the samples treated with both types 388 and levels of natural inputs: riverine and atmospheric. 389 The different phytoplankton groups also responded differently to the 390 experiments and treatments (Fig. 5, Table 3 & 4). While Synechococcus showed a 391 significant increase in the organic and mixed treatment in spring (Fig. 5a), a significant 392 decrease was observed in autumn in response to atmospheric and organic additions (Fig. 393 5g). In summer, the low abundances or even absence of Synechococcus in some samples 394 did not allow inferring reliable responses to the treatments. Among APF, positive 395 responses were observed for 10% atmospheric, inorganic and mixed additions in spring 396 and summer, and negative in all experiments for organic additions (Fig. 5b, e, h). ANF 397 showed positive responses in the controlled additions and in the 5% addition of riverine 398 inputs in spring (Fig. 5c), and in the inorganic and mixed treatments in summer (Fig. 399 5f). No response was observed in the autumn experiment for ANF, except a decrease in 400 the response to the organic additions (Fig. 5i). Diatoms showed a high positive response 401 to the inorganic treatment in the summer experiment (Fig. 5d) while a slight positive 402 response was also observed for the mixed addition. In spring and autumn, no responses 403 19 or negative effects were observed on diatoms, especially for organic and mixed 404 additions (not shown). 405 406 407 408 409 410 411 412 413 414 415 Fig. 5: Response ratios (treatment/control) of phytoplankton groups showing significant responses to 416 treatments in spring (a-c), summer (d-f) and autumn (g-i). Synecho: Synechococcus-type cyanobacteria, 417 APF: autotrophic picoflagellates, ANF: autotrophic nanoflagellates. Horizontal line denotes the response 418 ratio = 1. Atm.: Atmospheric inputs; I: controlled inorganic additions; O: controlled organic additions; M: 419 controlled mixed additions. The levels of significance of t-tests, performed between each treatment and 420 the control, are: *p<0.05, **p<0.01, ***p<0.001. In the experiments of July and October, only one bag of 421 each treatment was analyzed for nano and microphytoplankton at 48h, so we don’t have statistical 422 significance levels for ANF and diatoms at these experiments. 423 424 It is noticeable that APF abundances were better stimulated by the combination 425 of inorganic and organic compounds (the mixed treatment) when compared with the 426 addition of only inorganic or only organic nutrients (Fig. 5b, e, h). In fact, the organic 427 20 treatment seems to be in some way detrimental for APF, showing lower abundances 428 than in the control. By contrast, ANF and diatoms typically grew more with the addition 429 of only inorganic nutrients (Fig. 5c, d, f) than in the mixed treatment. Synechococcus 430 showed a similar pattern to APF organisms, with highest growth under the combination 431 of organic and inorganic compounds (Fig. 5a, g). However, in this case, the organic 432 treatment seemed to be harmful for these organisms only in autumn (Fig. 5g). 433 Also for the different phytoplankton groups’ response ratios, the experiment and 434 the treatment were statistically significant factors (Table 3). For Synechococcus, 435 significant differences were found between the spring and summer experiments, while 436 for the small flagellates (APF and ANF) the significant differences occur between the 437 autumn experiment and the other two experiments (Table 4). No differences were found 438 between the responses to the different natural addition treatments. Most significant 439 differences in response ratios between treatments were found between natural and 440 controlled addition treatments. 441 442 Table 5: Pearson correlation coefficients for analyses performed between Chl a concentrations, PP rates 443 and abundances of the several phytoplankton groups at the end of the incubation, and initial inorganic or 444 organic nutrients levels (including in situ and added) at each experiment. Synecho: Synechococcus-type 445 cyanobacteria, APF: autotrophic picoflagellates, ANF: autotrophic nanoflagellates. 446 Chla PP Synecho APF ANF Diatoms SPRING DIN 0.975*** 0.930*** 0.563 0.729* 0.920*** -0.229 DOC 0.422 0.236 0.565 0.341 0.216 -0.497 DON 0.421 0.232 0.573 0.349 0.210 -0.520 SUMMER DIN 0.902*** 0.687* 0.506 0.271 0.922*** 0.710* DOC 0.117 -0.221 0.038 -0.216 0.099 -0.118 DON 0.120 -0.214 0.088 -0.254 0.117 -0.093 AUTUMN DIN 0.657* 0,413 0.642* 0.070 0.104 0.120 DOC -0.208 -0.290 0.281 -0.563 -0.686* -0.775** DON -0.151 -0.289 0.340 -0.621 -0.697* -0.738* 447 21 In order to explain the different phytoplankton responses at each experiment, we 448 analyzed them separately searching for correlations within the several variables. Chl a 449 concentration was positively correlated with DIN in the three experiments (Table 5). 450 Similarly, also several phytoplankton groups showed positive correlations with DIN, 451 but not for all experiments. DIN was positively correlated with APF and ANF in spring, 452 with ANF and diatoms in summer and with Synechococcus in autumn. No significant 453 correlations were found between the several phytoplankton groups and organic 454 compounds (both DOC and DON) in any of the three experiments, except for ANF and 455 diatoms in autumn which showed a negative significant correlation with organic 456 compounds. 457 Some differences in phytoplankton community composition were observed in 458 response to nutrient amendments (Fig. 6). In spring, the proportion of diatoms increased 459 with riverine and atmospheric inputs whereas with controlled additions the relative 460 diatoms abundance was depressed and ANF favored (Fig. 6a). In autumn an increase in 461 the proportion of Synechococcus abundance was observed in response to the organic 462 treatment (Fig. 6c). In this same treatment, it was observed a decrease in the proportion 463 of APF abundance (Fig. 6c). 464 22 465 Fig. 6: Relative abundance of major phytoplankton groups at the end of the experiment in spring (a), 466 summer (b) and autumn (c). Synecho: Synechococcus-type cyanobacteria, APF: autotrophic 467 picoflagellates, ANF: autotrophic nanoflagellates. C: Control (no additions); I: controlled inorganic 468 additions; O: controlled organic additions; M: controlled mixed additions. 469 470 471 % 0 20 40 60 80 100 % 0 20 40 60 80 100 % 0 20 40 60 80 100 Synecho APF ANF Diatoms Pigmented dinoflagellates SpringSummerAutumn a b c Riverine Atmospheric C5%10% C I O MC 5%10% Controlled 23 4. Discussion 472 Performing the experiments in spring (May), summer (July) and autumn 473 (October) allowed investigating the different hydrographic conditions and biological 474 communities typical of this coastal embayment (Figueiras and Ríos 1993; Figueiras et 475 al. 2002; Arbones et al. 2008). Just after the first intense upwelling period of the year in 476 spring, high Chl a and nutrient concentrations were observed in surface waters, with a 477 typical community dominated in biomass by chain forming diatoms (Chaetoceros spp.). 478 On the contrary, in summer, despite the previous upwelling events, we found that the 479 surface layer was poor in nutrients and dominated by small plankton organisms. This 480 fact was due to the sinking of the surface phytoplankton communities (dominated by 481 large diatoms developed in the previous weeks - data not shown), resulting from the 482 relaxation of upwelling favorable winds. In autumn, typical oceanic communities were 483 found at the surface as a result of a previous strong downwelling event which pushed 484 continental shelf waters into the Ría. High abundance of Synechococcus denotes the 485 presence of this offshore water inside the Ría de Vigo (Rodríguez et al. 2006). 486 The statistically different response ratios of phytoplankton among the 487 experiments performed in the three different seasons indicate that the initial conditions 488 at each experiment were important to determine the magnitude and the type of response 489 of phytoplankton communities to nutrients inputs. Differences in initial conditions 490 among the experiments which can influence these phytoplankton responses may include 491 dissimilarities in environmental variables (i.e. nutrients availability), in the composition 492 of the natural matter inputs or in the structure of phytoplankton communities (Martínez-493 García et al. 2015). The high positive correlations found between Chl a (and several 494 phytoplankton groups) and DIN concentrations in the incubation bags suggest that DIN 495 availability is an important factor determining the response of phytoplankton 496 24 communities in this investigation. In addition, the highest phytoplankton responses to 497 the several treatments were observed in summer, when in situ DIN levels were the 498 lowest. In this case, phytoplankton growth seemed to be severely limited by nutrients 499 and both types of inputs (natural and controlled) promoted enhanced phytoplankton 500 growth. A similar situation, although with a lower magnitude, occurred in spring. In 501 autumn, however, when in situ inorganic nutrient levels were the highest, phytoplankton 502 did not respond or even decreased in abundance. In this case, interactions between 503 several factors besides DIN must have occurred as discussed later on. Nitrogen has been 504 widely recognized as the main limiting nutrient in marine systems when light intensity 505 is sufficient (Wollast 1998). In fact, several works have also observed stimulation in 506 phytoplankton communities due to increases in inorganic nitrogen from atmospheric 507 deposition (Paerl et al. 1990; Cui et al. 2016) or in controlled additions (Piehler et al. 508 2004; Xu et al. 2014). 509 In addition to inorganic nutrients, it has been observed that organic compounds 510 may also stimulate phytoplankton growth in coastal areas (Peierls and Paerl 1997; 511 Seitzinger and Sanders 1999). In fact, prior results in the Ría de Vigo showed that 512 organic nitrogen concentration also contributed to the variability in the response of 513 primary production and Chl a to rainwater additions (Martínez-García et al. 2015). 514 However, in our work, no significant correlations were found between phytoplankton 515 abundance and organic compounds, except a negative correlation with ANF and 516 diatoms in autumn. The few significant positive (although low) increases in 517 phytoplankton abundance in response to the organic treatment (Synechococcus and ANF 518 in spring and Chl a in summer) may be related to the occurrence of mixotrophy 519 (including direct consumption of organic compounds) in both seasons (Flynn and Butler 520 1986; Eiler 2006) or to an increase in nutrients remineralization mediated by enhanced 521 25 bacterial activity in the organic treatment in summer (Teira et al. 2016). Nutrient 522 remineralization in spring is unlikely to be important, because there were high in situ 523 nutrient levels and bacteria did not respond to the organic inputs in this experiment 524 (Teira et al 2016). As the chemical composition of the rainwater additions here was very 525 different from that in Martínez-García et al. (2015), it should be expected that different 526 phytoplankton responses may occur depending on the composition of natural matter 527 inputs in the Ría de Vigo. 528 Moreover, some phytoplankton groups appeared to be favored by the 529 combination of inorganic and organic forms. This is the case of APF in spring and 530 summer, which presented a higher response to the mixed treatment than to the inorganic 531 or organic additions alone (Fig. 5b, e). This response can be related to the concomitant 532 use of organic forms, namely amino acids, by these small phytoplankton cells (Flynn 533 and Butler 1986; Hernández-Ruiz et al., 2018). The same trend was observed for 534 Synechococcus in spring (Fig. 5a) likely related to their mixotrophic capabilities (Eiler 535 2006). Other explanations may also include the need of external secondary metabolites 536 for phytoplankton growth (eg. B12 vitamin) produced by heterotrophic bacteria (Croft et 537 al. 2005; Prieto et al. 2016) or the occurrence of bacterivory by the small eukaryotes 538 (Sanders and Gast 2012; Hartmann et al. 2013). In both cases, phytoplankton would be 539 benefiting from the stimulation of bacterial growth by the combination of the organic 540 and inorganic compounds in the mixed treatment (Teira et al. 2016). On the other hand, 541 the lower response of ANF and diatoms to the mixed compared to the inorganic 542 treatment could be explained by competition for inorganic nutrients with the organisms 543 stimulated by the mixed treatment, including APF and bacteria (Joint et al. 2002; 544 Martínez-García et al. 2010). 545 32 Joint, I., Henriksen, P., Fonnes, G.A., Bourne, D., Thingstad, T.F., Riemann, B., 2002. 684 Competition for inorganic nutrients between phytoplankton and bacterioplankton in 685 nutrient manipulated mesocosms. Aquat. Microb. Ecol. 29, 145-159. 686 Kérouel, K., Aminot, A.1997. Fluorimetric determination of ammonia in sea and 687 estuarine waters by direct segmented flow analysis. Marine Chemistry 57, 265–275. 688 Laruelle, G.G., Roubeix, V., Sferratore,A., Brodherr, B., Ciuffa,D., Conley, D. J., Durr, 689 H. H., Garnier, J., Lancelot, C., Le Thi Phuong, Q., Meunier, J.-D., Meybeck, M., 690 Michalopoulos, P., Moriceau, B., Nı´ Longphuirt, S., Loucaides, S., Papush, L., Presti, 691 M., Ragueneau, O., Regnier, P., Saccone, L., Slomp, C. P., Spiteri, C., and Van 692 Cappellen, P., 2009. Anthropogenic perturbations of the silicon cycle at the global 693 scale: Key role of the land-ocean transition. Global Biogeochem. Cycles, 23, GB4031, 694 doi:10.1029/2008GB003267 695 Lessard, E.J., Swift, E., 1986. Dinoflagellates from the North Atlantic classified as 696 phototrophic or heterotrophic by epifluorescence microscopy. J. Plankton Res. 8, 1209-697 1215. 698 Martínez-García, S., Arbones, B., García-Martín, E.E., Teixeira, I.G., Serret, P., 699 Fernández, E., Figueiras, F.G., Teira, E., Álvarez-Salgado, X.A., 2015. Impact of 700 atmospheric deposition on the metabolism of coastal microbial communities. Estuar. 701 Coast. Shelf S. 153, 18-28. 702 Martínez-García, S., Fernández, E., Álvarez-Salgado, X.A., González, J., Lønborg, C., 703 Marañón, E., Morán, X.A.G., Teira, E., 2010. Differential responses of phytoplankton 704 and heterotrophic bacteria to organic and inorganic nutrient additions in coastal waters 705 off the NW Iberian Peninsula. Mar. Ecol. Prog. Ser. 416, 17-33. 706 33 Menden-Deuer, S., Lessard, E.J., 2000. Carbon to volume relationships for 707 dinoflagellates, diatoms, and other protist plankton. Limnol. Oceanogr. 45, 569-579. 708 Otero, P., Ruiz-Villarreal, M., Peliz, A., Cabanas, J.M., 2010. Climatology and 709 reconstruction of runoff time series in northwest Iberia: influence in the shelf buoyancy 710 budget off Ría de Vigo. Sci. Mar. 74, 247-266. 711 Paerl, H.W., 1997. Coastal eutrophication and harmful algal blooms: Importance of 712 atmospheric deposition and groundwater as “new” nitrogen and other nutrient sources. 713 Limnol. Oceanogr. 42, 1154-1165. 714 Paerl, H.W., Rudek, J., Mallin, M.A., 1990. Stimulation of phytoplankton production in 715 coastal waters by natural rainfall inputs: Nutritional and trophic implications. Mar. Biol. 716 107, 247-254. 717 Peierls, B.L., Paerl, H.W., 1997. Bioavailability of atmospheric organic nitrogen 718 deposition to coastal phytoplankton Limnol. Oceanogr. 42, 1819-l1823 719 Pérez, F.F., Padín, X.A., Pazos, Y., Gilcoto, M., Cabanas, M., Pardo, P.C., Doval, M.D., 720 Farina-Busto, L., 2010. Plankton response to weakening of the Iberian coastal 721 upwelling. Glob. Change Biol. 16, 1258-1267. 722 Piehler, M.F., Twomey, L.J., Hall, N.S., Paerl, H.W., 2004. Impacts of inorganic 723 nutrient enrichment on phytoplankton community structure and function in Pamlico 724 Sound, NC, USA. Estuar. Coast. Shelf S. 61, 197-209. 725 Prieto, A., Barber-Lluch, E., Hernández-Ruiz, M., Martínez-García, S., Fernández, E., 726 Teira, E., 2016. Assessing the role of phytoplankton–bacterioplankton coupling in the 727 response of microbial plankton to nutrient additions. J. Plankton Res. 38, 55-63. 728 34 Rodríguez, F., Garrido, J.L., Crespo, B.G., Arbones, B., Figueiras, F.G., 2006. Size-729 fractionated phytoplankton pigment groups in the NW Iberian upwelling system: impact 730 of the Iberian Poleward Current. Mar. Ecol. Prog. Ser. 323, 59-73. 731 Rodríguez, L., Macías, F., 2006. Eutrophication trends in forest soils in Galicia (NW 732 Spain) caused by the atmospheric deposition of nitrogen compounds. Chemosphere 63, 733 1598-1609. 734 Sanders, R.W., Gast, R.J., 2012. Bacterivory by phototrophic picoplankton and 735 nanoplankton in Arctic waters. FEMS Microbiol. Ecol. 82, 242-253. 736 Seitzinger, S.P., Sanders, R.W., 1999. Atmospheric inputs of dissolved organic nitrogen 737 stimulate estuarine bacteria and phytoplankton. Limnol. Oceanogr. 44, 721-730. 738 Spatharis, S., Tsirtsis, G., Danielidis, D.B., Chi, T.D., Mouillot, D., 2007. Effects of 739 pulsed nutrient inputs on phytoplankton assemblage structure and blooms in an enclosed 740 coastal area. Estuar. Coast. Shelf S. 73, 807-815. 741 Statham, P.J., 2012. Nutrients in estuaries — An overview and the potential impacts of 742 climate change. Sci. Total Environ. 434, 213-227. 743 Teira, E., Hernández-Ruiz, M., Barber-Lluch, E., Sobrino, C., Teixeira, I.G., Álvarez-744 Salgado, X.A., Nieto-Cid, M., Martínez-García, S., Figueiras, F.G., Fernández, E., 745 2016. Bacterioplankton responses to riverine and atmospheric inputs in a coastal 746 upwelling system (Ría de Vigo, NW Spain). Mar. Ecol. Prog. Ser. 542, 39-50. 747 Teira, E., Hernando-Morales, V., Martínez-García, S., Figueiras, F.G., Arbones, B., 748 Álvarez-Salgado, X.A., 2013. Response of bacterial community structure and function 749 to experimental rainwater additions in a coastal eutrophic embayment. Estuar. Coast. 750 Shelf S. 119, 44-53. 751 35 Teira, E., Martínez-García, S., Carreira, C., Morán, X.A.G., 2011. Changes in 752 bacterioplankton and phytoplankton community composition in response to nutrient 753 additions in coastal waters off the NW Iberian Peninsula. Mar. Ecol. Prog. Ser. 426, 87-754 104. 755 Walsh, J.J., 1991. Importance of continental margins in the marine biogeochemical 756 cycling of carbon and nitrogen. Nature 350, 53-55. 757 Wollast R. Interactions of carbon and nitrogen cycles in the coastal zone, In: Wollast R., 758 Mackenzie F.T. and Chou L. (Eds.), Interactions of C, N, P and S Biogeochemical 759 Cycles and Global Change, 1993, NATO ASI Series (Series I: Global Environmental 760 Change), vol 4. Springer, Berlin, Heidelberg.) 761 Wollast, R., 1998. Evaluation and comparison of the global carbon cycle in the coastal 762 zone and in the open ocean, in: Brink, K.H., Robinson, A.R. (Eds.), The Sea. John 763 Wiley & Sons, New York, pp. 213-252. 764 Xu, Z.-H., Guo, Z.-R., Xu, X., Huang, D.-J., Sun, X.-X., Jiang, L.-M., Yang, J.-F., 765 2014. The impact of nutrient enrichment on the phytoplankton and bacterioplankton 766 community during a mesocosm experiment in Nan'ao of Daya Bay. Marine Biology 767 Research 10, 374-382. 768     Fig.S1:Temporalevolutionoftheupwellingindexandrunoffvaluesthroughout2013. Experimentdatesaremarkedwithdiscontinuouslines.     Upwelling index (m3 s-1 km-1) -4000 -2000 0 2000 4000 Runoff (m3 s-1) 0 50 100 150 200 250 300 350 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Spring Summer Autumn