Decoupled growth and grazing rates of diatoms and green algae drive increased phytoplankton productivity on HNLC sub‐Antarctic plateaux

The combination of iron limitation and microzooplankton grazing controls phytoplankton productivity and taxonomic composition in high‐nutrient low‐chlorophyll (HNLC) regions. While increased productivity and diatom contribution triggered by iron enrichment support this view, direct measurements of underpinning group‐specific growth and grazing rates are scarce for the Southern Ocean. To assess these rates, we conducted dilution experiments coupled to high‐performance liquid chromatography and flow‐cytometry in sub‐Antarctic waters on and off Campbell Plateau, southeast of Aotearoa‐New Zealand. Off the plateau, growth and grazing were closely balanced for all groups despite a two‐fold difference between slow‐ and fast‐growing groups. On Campbell Plateau, where HNLC conditions were alleviated, the balance was disrupted, mainly by the preferential growth of diatoms and green algae, which was stimulated beyond grazing. Our results expand the recognized ability of diatoms to escape grazing control to picoplanktonic green algae that also avoid grazing and contribute significantly to phytoplankton productivity and biomass accumulation.

Vast areas of the Southern Ocean and the subarctic and equatorial Pacific Ocean have low chlorophyll concentrations despite abundant macronutrients due to iron limitation (Martin 1990).The "Ecumenical hypothesis" proposed that the combination of iron limitation and microzooplankton grazing regulates phytoplankton biomass and composition in these so-called high-nutrient low-chlorophyll (HNLC) waters (Morel et al. 1991).Mesoscale studies demonstrated that increase iron supply stimulates primary productivity, with grazing dynamics further regulating phytoplankton biomass, community composition and fate (Boyd et al. 2007).Current understanding states that low iron supply limits the growth of large (diatom) cells, facilitating the dominance of small phytoplankton, less affected by iron limitation and more strongly controlled by protistan grazing (Landry et al. 1997;Boyd et al. 2000;McNair et al. 2021), although small phytoplankton can also accumulate in response to increase iron supply (Barber and Hiscock 2006).
Taxon-specific rate measurements conducted in the equatorial (Latasa et al. 1997;Selph et al. 2011) and subarctic Pacific (Strom and Welschmeyer 1991) support this view showing that: (1) microzooplankton grazing pressure was stronger on picocyanobacterial and smaller eukaryotic phytoplankton, than on diatoms, and (2) the fast growth capacity and lower grazing susceptibility of diatoms underpinned their dominance following HNLC condition alleviation.
The sub-Antarctic zone is one of the largest HNLC regions globally (Dugdale et al. 1995;Longhurst 2007), with primary productivity typically low due to iron and/or silicate limitation (Boyd et al. 1999;Sedwick et al. 1999;Dugdale et al. 2007) and strong microzooplankton grazing (Banse 1996;Evans and Brussaard 2012;Hall et al. 2004).Despite small phytoplankton dominance, the contribution of larger forms (e.g., diatoms) typically increase in bottle (Sedwick et al. 1999;Boyd et al. 2001) and mesoscale iron addition experiments (Coale et al. 2004), as well as in naturally fertilized sub-Antarctic island systems and plateaux (Korb et al. 2008;Queguiner 2013;Irion et al. 2020).These observations support the combined role of iron and grazing in regulating phytoplankton dynamics in sub-Antarctic HNLC waters (Banse 2013;Christaki et al. 2021).However, phytoplankton taxon-specific growth and grazing rate measurements that contributed critically to evolve the "Ecumenical hypothesis" into the current paradigm (Strom and Welschmeyer 1991;Latasa et al. 1997;Landry et al. 2011) are lacking for the sub-Antarctic zone.
Studies have reported picomolar iron concentration in sub-Antarctic waters southeast of New Zealand (Boyd et al. 2005;Ellwood et al. 2008), with low photosynthetic competence, growth and productivity measurements (Boyd et al. 1999;McKay et al. 2005;Peloquin et al. 2011) supporting ironlimited, HNLC conditions.Yet, patches of elevated chlorophyll are observed around submarine rises and islands of the Campbell Plateau (Heath and Bradford 1980;Banse and English 1997); a submerged continental platform extending 1000 km southeast of New Zealand (Neil et al. 2004).This topographic feature acts as a barrier to the circumglobal sub-Antarctic and Subtropical fronts, which are diverted around the plateau (Smith et al. 2013;Shao et al. 2015) creating a shallow body of low velocity water (Stanton and Morris 2004) isolated from the deeper Southwest Pacific Basin with implications for sub-Antarctic mode water formation (Morris et al. 2001;Forcén-V azquez et al. 2021) and phytoplankton dynamics (Heath and Bradford 1980;Banse and English 1997;Murphy et al. 2001;Boyd et al. 2004).The presence of economically important fisheries (Bradford-Grieve et al. 2003) reinforces distinctive productivity dynamics on the plateau.Recent in situ measurements displayed enhanced net primary production (NPP) and photochemical efficiency (Gutiérrez-Rodríguez et al. 2020) consistent with natural iron fertilization processes observed in similar sub-Antarctic systems (Blain et al. 2007;Pollard et al. 2007;Korb et al. 2008).However, the Gutiérrez-Rodríguez et al. ( 2020) study lacked taxon-specific rates needed to demonstrate differences in bottom-up and top-down controls on large and small phytoplankton central to the "Ecumenical hypothesis." The primary goal of this study is to determine if phytoplankton community and taxon-specific dynamics in sub-Antarctic waters align with current understanding of iron-limited HNLC systems.Specifically, we aim to test that: (1) sub-Antarctic waters on (< 500 m isobath, Onplateau) and off (> 1000 m, Off-plateau) the Campbell Plateau differ in HNLC conditions, (2) this difference influence growth, grazing and net growth rates of all phytoplankton groups, with some groups (i.e., diatoms) responding more strongly to improved conditions, and (3) the resulting imbalance in taxon-specific rates and carbon fluxes leads to differences in community structure On-and Off-Plateau.We assessed phytoplankton community composition, taxon-specific growth and grazing rates, carbon production and microzooplankton consumption from high-performance liquid chromatography (HPLC) and flow-cytometry (FCM) analysis of serial dilution experiments conducted in sub-Antarctic waters On-and Off-Plateau during the austral late-summer autumn period when elevated chlorophyll has been reported (Heath and Bradford 1980;Banse and English 1997;Morris et al. 2001).

Study area and sample collection
This study was conducted on RV Tangaroa during the TAN1702 voyage that surveyed the Campbell Plateau and the Southwest Pacific Basin.Details of the voyage sampling strategy can be found in Gutiérrez-Rodríguez et al. (2020).In the present study we distinguish between a subset of "Biomass" stations (n = 7), where depth-resolved samples (5-100 m) for nutrients, total and size-fractionated chlorophyll a, HPLC, and FCM were collected, and additional "Experimental" stations (n = 8), where in addition to these stock variables, dilution experiments coupled to HPLC and FCM analysis, were carried out with surface (10 m) water to assess taxon-specific growth and grazing dynamics (Fig. 1).

FCM, HPLC, and CHEMTAX analysis
For FCM, duplicated 1.5-mL seawater samples were preserved with 0.25% glutaraldehyde and 0.01% pluronic acid final concentration (Marie et al. 2014), flash-frozen in liquid nitrogen and stored at À80 C. Samples were analyzed using a BD-FACSCalibur instrument as described in Hall and Safi (2001).For HPLC, 1.7-2-L seawater samples were vacuumfiltered onto GFF filters, kept at À80 C and analyzed using an Agilent HPLC 1200 system with response factor of detectors calibrated against pigment standards (DHI Group, Denmark) and following details in Latasa et al. (2022).For CHEMTAX analysis, initial samples at "Experimental" stations and depth-resolved samples from "Biomass" stations were clustered using the logarithm of the pigment to total Chl a (TChl a) ratios of the following pigments: Chlc3, Chlc2, peridinin, 19 0 -butanoyloxyfucoxanthin, fucoxanthin, 19 0hexanoyloxyfucoxanthin, alloxanthin, zeaxanthin, Chl b, prasinoxanthin, neoxanthin, and violaxanthin as described in Latasa et al. (2022).CHEMTAX analysis was applied separately to each of the three clusters obtained (Fig. S1; Latasa 2007) to estimate the relative contributions of cryptophytes, diatoms, dinoflagellates, green algae (Chlorophyta), haptophytes, pelagophytes, and Synechococcus to TChl a.

Serial dilution experiments and microzooplankton biomass
Dilution experiments were set up as described in Gutiérrez-Rodríguez et al. ( 2020).Samples were incubated on deck under in situ simulated temperature and irradiance levels, and initial and final (24-h) samples were collected for HPLC and FCM analysis.Phytoplankton apparent growth rate k (day À1 ) in each incubation bottle were estimated as k ¼ 1=t ð ÞÂ ln N t =N 0 ð Þx D, where t is the incubation time and D the dilution factor, from changes between the initial (N 0 ) and final (N t ) pigment concentration or cell abundance for Synechococcus and picoeukaryotes (PEUK), and can be found in Gutiérrez-Rodríguez et al. (2023).Nutrient-amended growth rate (μ nut ) was estimated as the y-intercept of the linear regression between k and D. Grazing rate (m) was estimated as the difference between μ nut and the mean k (μ nutnet ) in nondiluted, nutrient-amended bottles (m = μ nut À μ nutnet ).Intrinsic growth rate (μ) was calculated as the sum between the mean apparent growth rate in nondiluted, unamended bottles (μ net ) and m.
Fucoxanthin was used as pigment marker for diatoms, 19 0 -hexanoyloxyfucoxanthin for haptophytes, 19 0butanoylxyfucoxanthin for pelagophytes, peridinin for dinoflagellates, alloxanthin for cryptophytes, and Chl b for green algae.Single pigment markers instead of CHEMTAXbased Chl a were used to assess group-specific growth and Blue circles indicate stations where macronutrients, phytoplankton community biomass and taxonomic composition (Size fractionated Chl a, HPLC, FCM) were quantified ("Biomass" stations).Orange circles indicate the stations where in addition to these stock variables, serial dilution grazing experiments coupled to HPLC and FCM analysis were conducted with surface water ("Experimental" stations).
grazing rates because amplification of relatively small error (Chl a ≈ 0.5 ngL À1 ) in CHEMTAX-based estimates caused large variability in estimated μ and m, particularly in oligotrophic waters Off-plateau.Pigment-based estimates of k were corrected using PEUK cell red fluorescence (FL3) and side scatter (SSC) from FCM initial and final measurements to account for photoacclimation-related changes (Gutiérrez-Rodríguez et al. 2020) and corrected k then used to calculate instantaneous growth and grazing coefficients from linear regressions.

Primary production and consumption rates
Daily Chl a synthesis and consumption rates of specific phytoplankton groups were estimated according to Frost (1972) as Chl where N 0 is the initial Chl a biomass of each group and μ and m the taxon-specific instantaneous growth and grazing rates.Phytoplankton community Chl a : C ratios were calculated at each station from daily Chl a synthesis and carbon uptake rates ( 14 C-NPP; Gutiérrez-Rodríguez et al. 2020).The inverse C : Chl a ratio was used to convert Chl a synthesis and consumption rates to carbon equivalents.

Statistical analysis
Data analysis and statistics were performed with Prism 9.5.1 software and R Studio, using alpha = 0.05 for statistical significance.Principal component and non-metric multidimensional scaling analysis, followed by permutational multivariate ANOVA (PERMANOVA) and similarity percentage (SIMPER) analysis were used to assess regional differences in environmental and CHEMTAX-based Chl a data, transformed by the fourth root.Two-way multivariate ANOVA (MANOVA) with region and phytoplankton group as factors was conducted to investigate region and taxonomic main effects on growth and grazing dynamics and univariate two-way ANOVA followed by Holm-Slidak's method (Holm 1979) to assess differences between each taxonomic group.For statistical analysis m and m : μ ratios were arc tangent transformed (Calbet and Landry 2004).
The relationship between μ and m provides further insights into the taxon-specific control exerted by microzooplankton.Offplateau, most phytoplankton groups fell close to the 1 : 1 line (Fig. 4a), suggesting that μ and m were closely balanced and biomass accumulation was largely controlled by microzooplankton grazing.This was the case even for green algae, diatoms, and cryptophythes showing relatively high growth rates μ ≥ 0.5 day À1 ).On-plateau, all groups, except slow-growing haptophytes and pelagophytes, fell below the 1 : 1 line reflecting the ability of diatoms and other small phytoplankton groups to avoid microzooplankton grazing control (Fig. 4b).

Carbon production and consumption rates
Carbon (C) synthesis and consumption rates were higher On-than Off-plateau for every group (Table 1), although their relative contribution to these flows differed between regions.Off-plateau, haptophytes dominated C-synthesis (44 AE 6%) followed by green algae (39 AE 4%) and dinoflagellates (11 AE 6%; Table 1).Net C-synthesis rates were on average negligible for most groups (<0.1 mgC m À3 day À1 ) and negative for dominant haptophytes (Table 1).On-plateau, green algae (39 AE 7%) and diatoms (21 AE 6%) channeled the majority of C-synthesis (Table 1) despite the greater contribution of haptophytes to the community biomass (Fig. 2).Phytoplankton net C-synthesis was positive and driven mainly by green algae and diatoms (Table 1).

Discussion
The "Ecumenical hypothesis" ascribes the low phytoplankton biomass and small cell dominance in HNLC systems to the combination of iron limitation and microzooplankton grazing control (Morel et al. 1991).Our results obtained in sub-Antarctic waters of contrasting productivity on and off Campbell plateau, are consistent with this view and provide new insights on phytoplankton taxon-specific dynamics.Table 1.Phytoplankton community and group-specific mean daily C-synthesis, C-consumption, and net C-synthesis rates in the surface mixed-layer On-and Off-plateau.The relative contribution (%) of each phytoplankton group to C-synthesis is included as a reference.
Off-plateau, growth and grazing were tightly coupled for all groups in agreement with balanced conditions expected in sub-Antarctic HNLC waters (Banse 2013).However, taxon-specific rates differed with diatoms, cryptophytes, and green algae exhibiting faster growth than dinoflagellates, haptophytes, and pelagophytes, a pattern reported in other HNLC regions (Strom and Welschmeyer 1991;Selph et al. 2011;Latasa et al. 2014).In our study, pelagophytes and haptophytes showed consistently low growth rates Off-plateau that increased only slightly under improved nutrient conditions On-plateau.These results suggest that the systematic slow-growth patterns observed across these regions might reflect the adaptation of these groups to the prevalent low iron availability in HNLC waters rather than a suboptimal physiological condition (Sofen et al. 2021).
This balance was disrupted On-plateau, where growth of most phytoplankton groups exceeded grazing and lead to positive net growth.Although excluding zooplankton in dilution experiments can cause trophic cascades, this typically leads to enhanced microzooplankton grazing (Zeldis et al. 2002;Zhao et al. 2020;Landry et al. 2022), suggesting such exclusion would have worked against the increased μ net observed.This rate imbalance was translated into significant net C-synthesis, particularly for green algae, which due to their consistently higher biomass compared to diatoms, emerged as the main contributor to C-flows On-plateau.The relative contribution of these groups should be viewed with caution since the same C : Chl a ratio was used to transform Chl a fluxes into carbon equivalents, but differences can emerge between phytoplankton of different cell size or taxonomic affiliation co-existing in the same environment (Chan 1980;Finkel 2001;Tozzi et al. 2004).Culture laboratory experiments have reported higher Chl a : C in diatoms compared to prasinophyte species under nutrient replete conditions, and the reverse pattern under nutrient starvation (Liefer et al. 2018).Off-plateau, diatoms and green algae μ ≈ 0.4-0.5 day À1 was substantially higher than μ ≈ 0.2 day À1 reported in early starvation phase cultures (≈ 0.2 day À1 ; Liefer et al. 2018), suggesting nutrient limitation Off-plateau was mild.C : Chl a reported for prasinophytes (≈ 33 g : g) and diatom (25-50 g:g) at this early starvation phase (Liefer et al. 2018) did not seem systematically different, and were in close agreement with the communitywise C : Chl a estimated Off-plateau (42 AE 6 g : g).Based on these trends, one would expect picoplanktonic green algae to have higher C : Chl a than larger diatoms under nutrient richer conditions (Finkel 2001;Tozzi et al. 2004;Liefer et al. 2018).Hence, community-wise C : Chl a we applied On-plateau may underestimate carbon biomass and fluxes of green algae relative to those of diatoms in our study.
While On-plateau diatom dynamics are consistent with their bloom-forming capacity following the alleviation of HNLC conditions (Boyd et al. 2007;Queguiner 2013), the prominent response of green algae is at odds with the capacity of protistan grazers to keep control on small phytoplankton populations central to the "Ecumenical hypothesis" (Morel et al. 1991;Landry et al. 1997;McNair et al. 2021).Experiments conducted in the equatorial and subarctic Pacific showed growth rates of small phytoplankton matched by grazing (Strom and Welschmeyer 1991;Rivkin et al. 1999), even when improved physico-chemical conditions associated with seasonal or climatic variability favored the growth of phytoplankton (Landry et al. 1993;Latasa et al. 1997).In our study, however, the decoupling between growth and grazing rates and high net C-synthesis of green algae observed On-plateau, argues against this view and supports the capacity of small phytoplankton to avoid microzooplankton grazing control and contribute significantly to phytoplankton accumulation in a way consistent with the Rising tide hypothesis (Barber and Hiscock 2006).
In the Northern Hemisphere, prasinophytes have been reported as important contributors to community biomass during winter deep mixing and the onset of the spring bloom (Bustillos-Guzman et al. 1995;Latasa et al. 2010;Gutiérrez-Rodríguez et al. 2011;Bolaños et al. 2020).In the present study, prasinoxanthin concentration and prasinoxanthin : Chl b ratios (0.12 AE 0.02 g : g) On-plateau (Fig. S4) were higher and similar to those observed in prasinoxanthin-containing prasinophytes (Latasa et al. 2004), suggesting that green algae biomass increase On-plateau was likely driven by these prasinophytes.Across the entire region, biomass and relative contribution of green algae were positively correlated with the depth of the mixed-layer and negatively correlated with the mean irradiance exposure (Fig. S5), supporting the ecological preference of prasinophytes for deep-mixing and low-light conditions.These results together with observations in other systems emphasize the bloom-forming capacity of this picoplanktonic group, and downplays the perceived dominant role of cell size in the formation and demise of phytoplankton blooms.The mechanism allowing prasinophytes to escape microzooplankton grazing control is unclear, but their tendency to thrive under deep-mixing, low-light conditions suggests their ecological advantage in such environments.This view fits with Beherenfeld's hypothesis of growth and grazing decoupling caused by deepening mixed-layers (Behrenfeld 2010), and suggest that prasinophytes can maintain relatively high growth rates under deep-mixing low-light conditions while they benefit from a more diluted predator environment (Behrenfeld 2014;Sallée et al. 2015;Gutiérrez-Rodríguez et al. 2020).The large decrease in grazing mortality and lack of growth stimulation observed for green algae between On-and Off-plateau suggests that the alteration of predatorprey encounter rate caused by physical mixing may be more important than the increased nutrient supply for them to thrive.Conversely, the stronger growth stimulation observed for diatoms On-plateau, highlights the greater importance of iron supply for this group.These observations suggest the mechanisms underpinning the decoupling leading to positive net growth and biomass accumulation in HNLC regions may differ between these groups.
The present study provides evidence supporting the hypothesis that iron limitation and microzooplankton grazing both control of phytoplankton biomass and productivity (observed for the HNLC equatorial and subarctic Pacific), also applies to the sub-Antarctic zone of the Southern Ocean.Our results show that in addition to diatoms, several other small phytoplankton groups, represented mainly by picoplanktonic green algae, can also escape grazing control and contribute significantly to productivity and biomass accumulation following the alleviation of HNLC conditions.Understanding environmental conditions and ecological mechanisms that can unleash the potential of small phytoplankton to bloom is paramount to predicting the phytoplankton response under a future warming Southern Ocean.

Fig. 1 .
Fig. 1.Map of the Campbell Plateau region surveyed during the TAN1702 voyage (15 March-1 April 2017) and stations sampled for the present study.

Fig. 2 .
Fig. 2. Phytoplankton community composition.(A) Relative contribution of each group to total Chl a (TChl a) estimated by CHEMTAX in the surface mixed-layer of stations On-and Off-plateau.Experimental stations where dilution experiments were conducted are labeled in bold.(B) Mean Chl a concentrations of main phytoplankton groups in surface mixedlayer waters On-and Off-plateau (error bars correspond to standard deviation of the mean, SEM), and (C) their average contribution to TChl a at stations On-and Off-plateau.Adjusted p-value of unpaired t-test corrected for multiple comparisons by Holm-Slidak method are indicated (* p < 0.05; ** p < 0.01;) and value shown if adjusted p-value between 0.05 and 0.1.

Fig. 3 .
Fig. 3. Phytoplankton community and group-specific mean (A) growth rates (μ), (B) grazing rates (m), (C) grazing to growth ratios (m : μ) and (D) net growth rate μ net estimated On-plateau (n = 5) and Off-plateau (n = 3).Error bars represent the standard error of the mean (SEM).Statistical post-hoc analyses of m and m : μ were conducted on arc tangent transformed individual values and the inverse tangent function used to transform back to the regional average values plotted.Adjusted p-value of Holm-Slidak post-hoc multiple comparison test are indicated if statistically significant (* p < 0.05; ** p < 0.01; *** p < 0.0001) and value shown if adjusted p-value between 0.05 and 0.1.

Fig. 4 .
Fig. 4. Mean regional growth and grazing rates of phytoplankton community and specific groups calculated for stations On-(n = 5) and Off-(n = 3) the Campbell Plateau.Error bars represent the standard error of the mean.The 1 : 1 line represents balanced conditions between growth and grazing rates.