Letter | Published:

# Toxic algal bloom induced by ocean acidification disrupts the pelagic food web

## Abstract

Ocean acidification, the change in seawater carbonate chemistry due to the uptake of anthropogenic CO2, affects the physiology of marine organisms in multiple ways1. Diverse competitive and trophic interactions transform the metabolic responses to changes in community composition, seasonal succession and potentially geographical distribution of species. The health of ocean ecosystems depends on whether basic biotic functions are maintained, ecosystem engineers and keystone species are retained, and the spread of nuisance species is avoided2. Here, we show in a field experiment that the toxic microalga Vicicitus globosus has a selective advantage under ocean acidification, increasing its abundance in natural plankton communities at CO2 levels higher than 600 µatm and developing blooms above 800 µatm CO2. The mass development of V. globosus has had a dramatic impact on the plankton community, preventing the development of the micro- and mesozooplankton communities, thereby disrupting trophic transfer of primary produced organic matter. This has prolonged the residence of particulate matter in the water column and caused a strong decline in export flux. Considering its wide geographical distribution and confirmed role in fish kills3, the proliferation of V. globosus under the IPCC4 CO2 emission representative concentration pathway (RCP4.5 to RCP8.5) scenarios may pose an emergent threat to coastal communities, aquaculture and fisheries.

## Main

The ocean plays a key role in the climate system by taking up one quarter of anthropogenic CO2 emissions5. As CO2 reacts with seawater to form carbonic acid, oceanic CO2 uptake increases seawater acidity, a process termed ocean acidification. While elevated CO2 can benefit carbon acquisition in photosynthetic organisms, increased acidity alters the transmembrane potential, which affects a range of cellular processes, including acid–base regulation, nutrient uptake and calcification6,7. The sensitivity to acidification-induced changes in seawater conditions differs between phytoplankton groups and even between species of the same taxonomic group8. By altering the competitive fitness of interacting species or the palatability for predators, this can shift the composition of phytoplankton communities. Basic ecosystem functions are generally maintained if the shifts occur among functionally redundant species. However, if the change in community composition involves the loss of a keystone species or shifts between functional groups, this can alter the functionality of an ecosystem, with potential impacts on food webs and biogeochemical processes9.

The most severe impacts may arise where community change favours the proliferation of harmful algal species. Although expansion of nuisance species is often considered a potential risk of ocean change2, information on the effects of ocean acidification on the physiological performance and cellular toxicity of harmful algal species is scarce and contradictory10. Whether changes in physiological performance can lead to the expansion of harmful algal blooms (HABs) under ocean acidification depends on how they affect the competitive fitness of HAB species over that of other co-existing species. This can best be tested in community experiments, which include genetic variability and cover the complex competitive and trophic interactions of pelagic food webs.

Recent investigations in temperate regions indicate that the responses of plankton communities to ocean acidification are most pronounced under conditions of nutrient limitation11,12. These findings draw attention to the oligotrophic (permanently nutrient-limited) subtropical waters that cover 40% of the Earth’s surface but have not been the focus of ocean acidification research to the same extent as eutrophic temperate regions13. Therefore, in 2014, we conducted a 9-week in situ mesocosm experiment off Gran Canaria (Spain) in the subtropical North Atlantic to assess the effects of ocean acidification on an oligotrophic plankton community.

Details of the experimental design, sampling and measurements are described in Taucher et al.14. Briefly, on 23 September 2014, nine 15 m long mesocosms, each enclosing about 35 m3 of seawater, were deployed off the east coast of Gran Canaria at 27° 55′ 41′′ N, 15° 21′ 55′′ W. The mesocosm bags were left open to allow free exchange with the outside water until their closure on 27 September, which also marks the start of the sampling programme. After 4 days of initial measurements, the carbonate chemistry of the enclosed seawater was manipulated by adding CO2-saturated seawater in four steps over 7 days. Two further CO2 additions were conducted on days 21 and 38 to compensate for the loss of CO2 through air–sea gas exchange. The partial pressure of CO2 ($${p_{{{\rm{CO}}_2}}}$$) averaged over the duration of the experiment covered a range from 352 to 1,025 μatm (Table 1). The experiment lasted for 62 days, starting with the closing of the mesocosms four days before day 0 (1 October, the day of the first CO2 manipulation) and finishing with the last sampling of the sediment trap on day 57.

The pronounced thermal stratification in the Canary Islands region is occasionally disrupted by mesoscale variability, in particular by island eddies that transport nutrient-rich waters from the mesopelagic zone to the surface15 and by upwelling filaments reaching out from the West-African coast16. To simulate such a temporary shift from oligotrophic to eutrophic conditions in the mesocosms, we replaced 20% of the enclosed mesocosm water with nutrient-rich deep water collected nearby at 650 m with a newly developed 85 m3 volume deep water collector during the night between day 24 and 25 (see Taucher et al.14). The addition of deep water increased nutrient concentrations from average values until day 23 of 0.06 ± 0.01, 0.026 ± 0.004 and 0.26 ± 0.04 μmol l−1 for nitrate + nitrite, phosphate and orthosilicic acid, respectively, to 3.15, 0.17 and 1.60 μmol l−1, respectively, in the early morning of day 25.

During the 4-week period of oligotrophic conditions, the phytoplankton species composition changed from a cyanobacteria-dominated community towards a more even distribution of different phytoplankton groups, including diatoms, prymnesiophytes, cryptophytes, cyanobacteria and, to a lesser degree, prasinophytes and chlorophytes14. The addition of deep water on day 24 initiated a phytoplankton bloom (Fig. 1a), with an exponential increase in biomass of nearly all groups, but most pronounced in diatoms, which dominated at the peak of the bloom and contributed 60–80% of the total biomass. The phytoplankton taxa responded differently to the CO2 treatment. While prymnesiophyte and cryptophyte biomass correlated positively with $${p_{{{\rm{CO}}_2}}}$$, autotrophic dinoflagellates showed a strong negative correlation with $${p_{{{\rm{CO}}_2}}}$$(Supplementary Fig. 1). Diatom biomass was unaffected by the CO2 treatment during the bloom phase, but remained at higher levels in the high CO2 mesocosms during the post-bloom phase.

Elevated CO2 triggered a further pivotal shift in phytoplankton composition. Halfway through the oligotrophic phase, V. globosus (Dictyochophyceae, Y. Hara and M. Chihara17; basionym: Chattonella globosa, Y. Hara and M. Chihara) suddenly appeared. This toxic microalga produces haemolytic cytotoxins, which impair membrane permeability and lead to osmotic cell lysis3. V. globosus abundance increased exponentially in CO2 treatments above 600 µatm, while it remained below the detection limit in the low CO2 treatments (Fig. 1b,c). The exponential growth of V. globosus continued after deep water was added on day 24 in the three highest CO2 treatments, reaching maximum abundances 4–6 days later, with cell densities of 600–800 cells ml−1 (Fig. 1b).

Exponential growth of V. globosus was detectable in all mesocosms with $${p_{{{\rm{CO}}_2}}}$$ values above 600 µatm from day 15 onwards (Fig. 1c), suggesting a direct positive effect of elevated CO2 on its cell division rate. In the intermediate CO2 treatments, exponential growth stopped abruptly just before deep water was added, followed by a rapid decline in cell numbers. The abundances of micro- and mesozooplankton were similar across all treatments at the time when V. globosus abundances started to diverge between moderate and high CO2 mesocosms18. Whatever caused the rapid decline in the intermediate CO2 treatments, it was apparently absent or ineffective under high CO2, allowing V. globosus to develop to HAB levels. Possible explanations for the continued net growth in the high CO2 treatments are: (1) higher toxicity under elevated CO2 as observed in other HAB species19,20, reducing predatory loss; (2) increased resistance to viral infection21; or (3) switching off of the programmed cell death directive22 under high CO2/low pH. Dedicated culture studies are necessary to assess the physiological performance of V. globosus under elevated CO2/reduced pH and elucidate the mechanism that favours bloom formation under ocean acidification.

The blooming of V. globosus did not affect other dominant phytoplankton groups. Diatom and prymnesiophyte biomass increased exponentially after deep water was added in mesocosms with and without V. globosus proliferation (Supplementary Fig. 1). In fact, prymnesiophytes reached their highest biomass in the high CO2 treatments concurrently with the blooming of V. globosus, and diatom biomass remained at higher levels in mesocosms with V. globosus proliferation during the post-bloom period. In contrast, autotrophic dinoflagellates never took off in the two mesocosms with high abundances of V. globosus. Since a direct negative effect of ocean acidification on dinoflagellates of this magnitude has not been observed in previous studies23,24,25, we suspect that repressed growth of autotrophic dinoflagellates may have been caused by the HAB species.

V. globosus proliferation under elevated CO2 conditions strongly impacted the zooplankton community. While micro- and mesozooplankton biomass rapidly increased in response to the deep water-induced phytoplankton bloom in low and moderate CO2 treatments, it dropped below pre-bloom levels in the high CO2 mesocosms and remained low until V. globosus abundances started to decline about 20 days after the start of the bloom (Fig. 2). This applied equally to all mesozooplankton species, dominated by the calanoid copepods Paracalanus indicus, Clausocalanus furcatus and Clausocalanus arcuicornis, the appendicularian Oikopleura dioica, and the dominant microzooplankton groups, aloricate ciliates and heterotrophic dinoflagellates18. The strong negative impact across the dominant zooplankton taxa, not all of which necessarily grazed actively on V. globosus, is consistent with the cytotoxic effect of V. globosus cell extracts demonstrated by Chang3. The suppression of zooplankton development largely inhibited the trophic transfer up the food web, including the consumption of the abundant food provided by other palatable phytoplankton groups, which prolonged the residence time of diatoms and prymnesiophytes during the post-bloom period in the presence of V. globosus (Supplementary Fig. 1).

The strong negative effect of V. globosus on the zooplankton community impacted key biogeochemical processes (Fig. 3). Degradation of sinking particles was reduced during the bloom of V. globosus, as indicated by significant differences in the quality of exported organic material26. A treatment effect was also observed for dissolved organic carbon, which accumulated to higher concentrations during the post-bloom period in the high CO2 treatments27. Whether this is a direct effect, for example, due to carbon overconsumption under elevated CO2, or whether an indirect effect mediated by V. globusus induced changes in bacterial activity, cannot be determined based on the available data. While particulate organic matter (POM) build-up in the water column was similar across the CO2 treatments, the post-bloom POM sedimentation peak was delayed in the high CO2 treatments by about a week (Fig. 3b). This may be explained by a reduced zooplankton-mediated downward flux due to low zooplankton standing stocks in the high CO2 treatments (Fig. 3a). In combination, these effects greatly reduced the downward flux of POM in treatments with high V. globosus abundances, where cumulative sedimentation of particulate organic carbon, nitrogen and phosphorus was about half of that obtained from mesocosms with no bloom of V. globosus (Fig. 3c).

The success of V. globosus in the high CO2 treatments probably originated from a combination of two effects: (1) stimulated net growth of V. globosus under elevated CO2, strengthening its competitive fitness over other co-existing phytoplankton species and increasing its relative contribution to the phytoplankton standing stock; and (2) decreased loss, possibly due to grazer repellence through increased toxicity. Both mechanisms increase the likelihood of this species to form harmful blooms in an acidifying ocean. Whether ocean acidification also strengthens the competitive fitness of other toxic microalgae is uncertain at present. Whereas growth rate in some HAB species was found to increase in response to elevated CO2, it remained stable in others, in some cases despite downregulated CO2 concentrating mechanisms and increased photosynthetic O2 evolution at elevated CO2 (see Wells et al.10 for a comprehensive review). There is also limited and conflicting information on the potential impact of ocean acidification on the cellular toxicity of HAB species. Cell toxicity was observed to increase or decrease in response to elevated CO2 depending on culturing methods, growth conditions and species, and even for different strains of the same species10.

Considering that HAB species are taxonomically diverse28, which also implies varying efficiencies of their CO2 concentrating mechanisms, the physiological benefits of rising $${p_{{{\rm{CO}}_2}}}$$ are expected to vary among different taxa. Moreover, given the high diversity of their toxins and of the metabolic pathways that lead to their production28, the effect of ocean acidification on HAB toxicity is bound to be non-uniform. The results of this study are unique and disconcerting because they provide the first evidence that ocean acidification improves the competitive fitness of a toxic microalga over that of other co-existing species in representative concentration pathways (RCPs)4 well below ‘business-as-usual’ CO2 emission scenarios (between RCP4.5 and RCP8.5). Based on the available information, it cannot be excluded that this also applies to other HAB species and that the stimulating effects of ocean acidification on growth and toxicity could lead to an expansion and increased intensity of harmful algal blooms29. Given the wide geographical distribution of V. globosus3, with reports from the coastal waters of Japan, Southern China, New Zealand, South East Asia, Australia, Canada, Greece, Russia and Brazil, and its potential to form harmful blooms disrupting the trophic transfer (this study) and causing mortality of farmed and wild fish30,31, the results of this study should be regarded as a warning call. An emerging threat for human society thereby lies in being unprepared for range expansions of toxic microalgae in currently poorly monitored areas32. This calls for broadening of seafood biotoxin and HAB monitoring programmes and emphasizes the need for further dedicated research on the responses of toxic microalgae to ocean change in an ecosystem context.

## Methods

### Study site

The mesocosm experiment was conducted in Gando Bay on the east coast of Gran Canaria at 27° 55′ 41′′ N, 15° 21′ 55′′ W (for map, see Taucher et al.14). Influenced primarily by the subtropical North Atlantic gyre, the Canary Islands region is characterized by a pronounced thermal stratification and predominantly oligotrophic conditions with low nutrient concentrations and plankton biomass throughout the year37. These stable conditions are occasionally disrupted by island eddies and upwelling filaments from the Canary current upwelling system off the West-African coast, which transport nutrient-rich water from the mesopelagic zone into the surface layer16. Such an event was simulated about half way through the experiment by replacing 20% of the enclosed oligotrophic surface water with nutrient-rich deep water collected nearby at 650 m depth as described in Taucher et al.14.

### Mesocosm operation

Nine 13 m long, 2 m in diameter mesocosms38 were deployed by the research vessel ESPS RV Hesperides in the northern part of Gando Bay (27° 55′ 41′′ N, 15° 21′ 55′′ W) at a depth of 20–25 m and moored in clusters of three on 23 September 2014. The mesocosm bags were left open to the surrounding water for 4 days, with top and bottom ends covered by nets with a 3 mm mesh size, to allow for rinsing of the bags’ interiors and free exchange of plankton < 3 mm. On 27 September, the bottom ends were closed by conical sediment traps and the top ends pulled above the sea surface, separating the enclosed contents from the surrounding water. CO2 manipulation was done by adding CO2-saturated seawater in four steps on 1, 3, 5 and 7 October. The first day of CO2 manipulation represents experimental day 0; sampling days before this are represented by negative numbers. To minimize wall growth, the mesocosm bags were cleaned at regular intervals from the inside with a specifically designed cleaning ring and from the outside by scrubbers and divers14.

### Sampling and sample processing

A comprehensive sampling programme for a wide range of physical, ecological and biogeochemical variables was conducted in the mesocosms and the surrounding water every second day between 9 a.m. and noon (see Taucher et al.14 for a list of parameters, sampling approaches and analytical methods). Water samples were usually taken with depth-integrating water samplers (Hydro-Bios); occasional large-volume samples were taken with a depth-integrating pumping system. Sinking particulate matter was collected every second day in the sediment traps at the bottom of the mesocosms and analysed according to Boxhammer et al.39. Mesozooplankton was sampled with an Apstein net (55 μm mesh size, 0.17 m diameter opening) at 8-day intervals and analysed as described in Algueró-Muñiz et al.18. Phytoplankton (>10 µm) and microzooplankton were enumerated in depth-integrated water samples fixed with acidic Lugol’s solution (1% final concentration) at ×200 and ×400 magnification using a ZEISS Axiovert A1 inverted microscope. Pico- and nanophytoplankton (<10 µm) were quantified using a FACSCalibur flow cytometer (BD Biosciences). Samples for particulate carbon and nitrogen were filtered (<200 mbar) onto pre-combusted (450 °C for 6 h) GF/F filters, dried (60 °C) overnight and measured on an elemental CN analyzer (EuroEA) following the protocol set out by Sharp40. Pigment samples were filtered (<200 mbar) on GF/F filters and stored at −80 °C until analysis. Pigments were extracted in 90% acetone and their concentrations quantified by means of reverse-phase high-performance liquid chromatography41. The contribution of distinct phytoplankton taxa to total chlorophyll a was calculated with the Chemtax software (version 1.95), which classifies phytoplankton based on taxon-specific pigment ratios using the pigment ratios provided by Mackey et al.42.

### Identification of V. globosus

Seawater samples were filtered on Millipore membrane filters (pore size 2 µm), stored at −80 °C until nucleic acid purification43 and used for PCR, quantitative PCR, TOPO TA cloning and Sanger sequencing. Selected samples were used for high-throughput sequencing using Illumina HiSeq technology; 18S recombinant DNA (rDNA) gene sequences were reconstructed from the raw reads with phyloFlash (https://github.com/HRGV/phyloFlash). Assemblies were performed using SPAdes version 3.9044 with standard parameters. Genome binning was performed in Bandage45 by collecting contigs linked to V. globosus. The evolutionary history was inferred using the neighbour joining method46 on a 540-base pair fragment of the 18S DNA. Phylogenetic analysis was conducted in MEGA747. Sequences were submitted to the National Center for Biotechnology Information (submission in progress). The identity of V. globosus was verified using Sanger and Illumina sequencing, with 18S rDNA-based hits revealing a 99% identity to V. globosus strain NIWA1008 (Supplementary Fig. 2). The temporal development of the V. globosus cluster, as determined by a clade-specific quantitative PCR of the 18S rDNA gene, is consistent with the direct cell counts; an increase in the intermediate and high CO2 treatments is followed by a sudden decline in the intermediate treatments and a continued rise and delayed breakdown in the high CO2 mesocosms (Supplementary Fig. 3).

### Statistical information

We applied simple linear regression analysis to examine the effects of simulated ocean acidification on the abundance and biomass of the different plankton taxa. Therefore, ecological data sets were temporally averaged for distinct periods of the experiment, using the corresponding $${p_{{{\rm{CO}}_2}}}$$ levels as the explanatory variable.

### Reporting Summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

## Data availability

The data of this mesocosm study are archived in the World Data Centre MARE/PANGAEA and can be downloaded using the following link: https://www.pangaea.de/?q=campaign%3A%22KOSMOS_2014%22.

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

## References

1. 1.

Kroeker, K. J. et al. Impacts of ocean acidification on marine organisms: quantifying sensitivities and interaction with warming. Glob. Change Biol. 19, 1884–1896 (2013).

2. 2.

Hall-Spencer, J. M. & Allen, R. The impact of CO2 emissions on ‘nuisance’ marine species. Res. Rep. Biodivers. Stud. 4, 33–46 (2015).

3. 3.

Chang, F. H. Cytotoxic effects of Vicicitus globosus (Class Dictyochophyceae) and Chattonella marina (Class Raphidophyceae) on rotifers and other microalgae. J. Mar. Sci. Eng. 3, 401–411 (2015).

4. 4.

IPCC Climate Change 2013: The Physical Science Basis (eds Stocker, T. F. et al.) (Cambridge Univ. Press, 2013).

5. 5.

Le Quéré, C. et al. Global Carbon Budget 2016. Earth Syst. Sci. Data 8, 605–649 (2016).

6. 6.

Beardall, J. & Raven, J. A. The potential effects of global climate change on microbial photosynthesis, growth and ecology. Phycologia 43, 26–40 (2004).

7. 7.

Giordano, M., Beardall, J. & Raven, J. A . CO2 concentrating mechanisms in algae: mechanisms, environmental modulation, and evolution. Annu. Rev. Plant Biol. 56, 99–131 (2005).

8. 8.

Riebesellsbe, U. & Tortell, P. D. in Ocean Acidification (eds Gattuso, J. P. & Hansson, L.) 99–121 (Oxford Univ. Press, New York, 2013).

9. 9.

Riebesell, U. et al. Competitive fitness of a predominant pelagic calcifier impaired by ocean acidification. Nat. Geosci. 10, 19–23 (2017).

10. 10.

Wells, M. L. et al. Harmful algal blooms and climate change: learning from the past and present to forecast the future. Harmful Algae 49, 68–93 (2015).

11. 11.

Sala, M. M. et al. Contrasting effects of ocean acidification on the microbial food web under different trophic conditions. ICES J. Mar. Sci. 73, 670–679 (2016).

12. 12.

Bach, L. T. et al. Influence of ocean acidification on a natural winter-to-summer plankton succession: first insights from a long-term mesocosm study draw attention to periods of low nutrient concentrations. PLoS ONE 11, e0159068 (2016).

13. 13.

Schulz, K. G. et al. Phytoplankton blooms at increasing levels of atmospheric carbon dioxide: experimental evidence for negative effects on prymnesiophytes and positive on small picoeukaryotes. Front. Mar. Sci. 4, 64 (2017).

14. 14.

Taucher, J. et al. Influence of ocean acidification and deep water upwelling on oligotrophic plankton communities in the subtropical North Atlantic: insights from an in situ mesocosm study. Front. Mar. Sci. 4, 85 (2017).

15. 15.

Arístegui, J. et al. The influence of island-generated eddies on chlorophyll distribution: a study of mesoscale variation around Gran Canaria. Deep Sea Res. Pt I 44, 71–96 (1997).

16. 16.

Pelegrí, J. L. et al. Coupling between the open ocean and the coastal upwelling region off northwest Africa: water recirculation and offshore pumping of organic matter. J. Mar. Syst. 54, 3–37 (2005).

17. 17.

Chang, F. H., McVeagh, M., Gall, M. & Smith, P. Chattonella globosa is a member of Dictyochophyceae: reassignment to Vicicitus gen. nov., based on molecular phylogeny, pigment composition, morphology and life history. Phycologia 51, 403–420 (2012).

18. 18.

Algueró-Muñiz, M. et al. Impacts of ocean acidification on the development of a subtropical zooplankton community during oligotrophic conditions and simulated upwelling. Front. Mar. Sci. (in the press).

19. 19.

Tatters, A. O., Fu, F. X. & Hutchins, D. A. High CO2 and silicate limitation synergistically increase the toxicity of Pseudo-nitzschia fraudulenta. PLoS ONE 7, e32116 (2012).

20. 20.

Tatters, A. O., Flewelling, L. J., Fu, F., Granholm, A. A. & Hutchins, D. A. High CO2 promotes the production of paralytic shellfish poisoning toxins by Alexandrium catenella from Southern California waters. Harmful Algae 30, 37–43 (2013).

21. 21.

Danovaro, R. et al. Marine viruses and global climate change. FEMS Microbiol. Rev. 35, 993–1034 (2011).

22. 22.

Bidle, K. D. Programmed cell death in unicellular phytoplankton. Curr. Biol. 26, R594–R607 (2016).

23. 23.

Eberlein, T., Van den Waal, D. B. & Rost, B. Differential effects of ocean acidification on carbon acquisition in two bloom-forming dinoflagellate species. Physiol. Plant. 151, 468–479 (2014).

24. 24.

Rost, B., Richter, K. U., Riebesell, U. & Hansen, P. J. Inorganic carbon acquisition in red tide dinoflagellates. Plant Cell Environ. 29, 810–822 (2006).

25. 25.

Bach, L. T., Alvarez-FernandezS., HornickT., StuhrA. & Riebesell, U. Simulated ocean acidification reveals winners and losers in coastal phytoplankton. PLoS ONE 12, e0188198 (2017).

26. 26.

Stange, P. et al. Ocean acidification-induced restructuring of the plankton food web can influence the degradation of sinking particles. Front. Mar. Sci. 5, 140 (2018).

27. 27.

Zark, M. et al. Ocean acidification experiments in large-scale mesocosms reveal similar dynamics of dissolved organic matter production and biotransformation. Front. Mar. Sci. 4, 271 (2017).

28. 28.

Zingone, A. & Oksfeldt Enevoldsen, H. The diversity of harmful algal blooms: a challenge for science and management. Ocean Coast. Manag. 43, 725–748 (2000).

29. 29.

Glibert, P. M. et al. Vulnerability of coastal ecosystems to changes in harmful algal bloom distribution in response to climate change: projections based on model analysis. Glob. Change Biol. 20, 3845–3858 (2014).

30. 30.

Hiroshi, S., Okada, H., Imai, I. & Yoshida, T. High toxicity of the novel bloom-forming species Chattonella ovata (Raphidophyceae) to cultured fish. Harmful Algae 4, 783–787 (2005).

31. 31.

Imai, I. & Itoh, K. Annual life cycle of Chattonella spp., causative flagellates of noxious red tides in the Inland Sea of Japan. Mar. Biol. 94, 287–292 (1987).

32. 32.

Hallegraeff, G. M. Ocean climate change, phytoplankton community responses, and harmful algal blooms: a formidable predictive challenge. J. Phycol. 46, 220–235 (2010).

33. 33.

Sathyendranath, S. et al. Carbon-to-chlorophyll ratio and growth rate of phytoplankton in the sea. Mar. Ecol. Prog. Ser. 383, 73–84 (2009).

34. 34.

Putt, M. & Stoecker, D. K. An experimentally determined carbon:volume ratio for marine “oligotrichous” ciliates from estuarine and coastal waters. Limnol. Oceanogr. 34, 1097–1103 (1989).

35. 35.

Lehette, P. & Hernández‐León, S. Zooplankton biomass estimation from digitized images: a comparison between subtropical and Antarctic organisms. Limnol. Oceanogr. Methods 7, 304–308 (2009).

36. 36.

Kiørboe, T. Zooplankton body composition. Limnol. Oceanogr. 58, 1843–1850 (2013).

37. 37.

Arístegui, J., Hernández-León, S., Montero, M. F. & Gómez, M. The seasonal planktonic cycle in coastal waters of the Canary Islands. Sci. Mar. 65, 51–58 (2001).

38. 38.

Riebesell, U. et al. Technical note: a mobile sea-going mesocosm system—new opportunities for ocean change research. Biogeosciences 10, 1835–1847 (2013).

39. 39.

Boxhammer, T., Bach, L. T., Czerny, J. & Riebesell, U. Technical note: sampling and processing of mesocosm sediment trap material for quantitative biogeochemical analysis. Biogeosciences 13, 2849–2858 (2016).

40. 40.

Sharp, J. H. Improved analysis for “particulate” organic carbon and nitrogen from seawater. Limnol. Oceanogr. 19, 984–989 (1974).

41. 41.

Barlow, R. G., Cummings, D. G. & Gibb, S. W. Improved resolution of mono- and divinyl chlorophylls a and b and zeaxanthin and lutein in phytoplankton extracts using reverse phase C-8 HPLC. Mar. Ecol. Prog. Ser. 161, 303–307 (1997).

42. 42.

Mackey, M. D., Mackey, D. J., Higgins, H. W. & Wright, S. W. CHEMTAX: a program for estimating class abundances from chemical markers: application to HPLC measurements of phytoplankton. Mar. Ecol. Prog. Ser. 144, 265–283 (1996).

43. 43.

Schunck, H. et al. Giant hydrogen sulfide plume in the oxygen minimum zone off Peru supports chemolithoautotrophy. PLoS ONE 8, e68661 (2013).

44. 44.

Nurk S. et al. Assembling genomes and mini-metagenomes from highly chimeric reads. In Proc. RECOMB 2013: Research in Computational Molecular Biology (eds Deng, M. et al.) 158–170 (Springer, 2013).

45. 45.

Wick, R. R., Schultz, M. B., Zobel, J. & Holt, K. E. Bandage: interactive visualization of de novo genome assemblies. Bioinformatics 31, 3350–3352 (2015).

46. 46.

Saitou, N. & Nei, M. The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4, 406–425 (1987).

47. 47.

Kumar, S., Stecher, G. & Tamura, K. MEGA7: molecular evolutionary genetics analysis version 7.0 for bigger datasets. Mol. Biol. Evol. 33, 1870–1874 (2016).

## Acknowledgements

We thank the Oceanic Platform of the Canary Islands (Plataforma Oceánica de Canarias) for their hospitality and outstanding support and the Marine Science and Technology Park (Parque Científico Tecnológico Marino) for providing access to their facilities. We are grateful to the captains and crews of ESPS RV Hesperides for deploying and recovering the mesocosms (cruise 29HE20140924), and of RV Poseidon for transporting the mesocosms and for their support in testing the deep water collector during cruise POS463. The manuscript greatly benefited from the comments of two anonymous reviewers. This project was funded by the German Federal Ministry of Education and Research (BMBF) in the framework of the coordinated project BIOACID—Biological Impacts of Ocean Acidification, phase 2 (FKZ 03F06550). U.R. received additional funding from the Leibniz Prize 2012 by the German Research Foundation (DFG).

## Author information

### Affiliations

1. #### GEOMAR Helmholtz Centre for Ocean Research Kiel, Kiel, Germany

• Ulf Riebesell
• , Eric P. Achterberg
• , Lennart T. Bach
• , Tim Boxhammer
• , Mathias Haunost
• , Andrea Ludwig
• , Carsten Spisla
• , Michael Sswat
• , Paul Stange
•  & Jan Taucher
2. #### Trondheim Biological Station, Department of Biology, Norwegian University of Science and Technology, Trondheim, Norway

• Nicole Aberle-Malzahn
3. #### Alfred-Wegener-Institut Helmholtz Centre for Polar and Marine Research, Biologische Anstalt, Helgoland, Germany

• María Algueró-Muñiz
• , Santiago Alvarez-Fernandez
• , Maarten Boersma
•  & Henriette G. Horn
4. #### Instituto de Oceanografía y Cambio Global, Universidad de Las Palmas de Gran Canaria, Las Palmas, Spain

• Javier Arístegui
5. #### Department of Marine Biotechnology, School of Laboratory Medicine and Life Science, Wenzhou Medical University, Wenzhou, Zhejiang, China

• Wanchun Guan
6. #### Danish Institute for Advanced Study, Department of Biology, University of Southern Denmark, Odense, Denmark

• Carolin R. Löscher

### Contributions

Experiment conception and design: U.R., J.T. and L.T.B. Experiment performance: all authors. Data analysis: J.T., L.T.B., T.B., M.A.-M., J.A., W.G., C.R.L., H.G.H. and P.S. Manuscript writing: U.R. with input from all co-authors.

### Competing interests

The authors declare no competing interests.

### Corresponding author

Correspondence to Ulf Riebesell.

## Supplementary information

1. ### Supplementary Information

Supplementary Figures 1–3