Characterization of a Southern Ocean deep chlorophyll maximum: Response of phytoplankton to light, iron, and manganese enrichment

Southern Ocean phytoplankton growth is limited by low iron (Fe) supply and irradiance, impacting the strength of the biological carbon pump. Unfavorable upper ocean conditions, such as low nutrient concentrations, can lead to the formation of deep chlorophyll or biomass maxima (DCM/DBM). While common in the Southern Ocean, these features remain understudied due to their subsurface location. To increase our understanding of their occurrence, we studied the responses of phytoplankton communities from a Southern Ocean DCM to increasing light, Fe, and manganese (Mn) levels. The DCM communities were light‐ and Fe‐limited, but light limitation did not increase phytoplankton Fe requirements. The greatest physiological responses were observed under combined Fe/light additions, which stimulated macronutrient drawdown, biomass production and the growth of large diatoms. Combined Mn/light additions induced subtle changes in Fe uptake rates and community composition, suggesting species‐specific Mn requirements. These results provide valuable information on Southern Ocean DCM phytoplankton physiology.

DCM to increasing light, Fe, and manganese (Mn) levels.The DCM communities were light-and Fe-limited, but light limitation did not increase phytoplankton Fe requirements.The greatest physiological responses were observed under combined Fe/light additions, which stimulated macronutrient drawdown, biomass production and the growth of large diatoms.Combined Mn/light additions induced subtle changes in Fe uptake rates and community composition, suggesting species-specific Mn requirements.These results provide valuable information on Southern Ocean DCM phytoplankton physiology.
Southern Ocean (SO) phytoplankton plays a key role in transferring carbon from the atmosphere into the ocean through the biological carbon pump (Boyd et al. 2019).Yet, low iron (Fe) concentrations combined with low irradiance and temperatures limit SO phytoplankton growth (Boyd et al. 2007), thereby limiting the strength of this carbon pump.The accumulation of the photosynthetic pigment chlorophyll below the ocean surface is known as deep chlorophyll maxima (DCM; Holm-Hansen et al. 2005;Baldry et al. 2020).Generally, DCM forms at the optimum depth between light attenuation from above and nutrient supply from below.They can be co-located with deep biomass maxima (DBM) or result from photoacclimatory processes, leading to an increase in cellular chlorophyll a (Chl a) concentrations (Cullen 2015).Southern ocean DCM biota are composed of large diatoms (Gomi et al. 2010) and, other than light/nutrient balance, it is hypothesized that water column stratification, grazing from higher trophic levels, and diatom buoyancy regulation are potential drivers of these features (Holm-Hansen et al. 2005;Baldry et al. 2020).
Several open questions related to the formation, ecology, and persistence of SO DCM remain.Here, we examine the most likely environmental drivers influencing the physiology of SO DCM phytoplankton communities.Low irradiance is expected to limit phytoplankton photosynthesis in DCM (Parslow et al. 2001;Baldry et al. 2020).However, the concurrent low light (LL) and dissolved Fe (dFe) concentrations found in the SO at depth where DCM occur (< 10% irradiance; Holm-Hansen et al. 2005 and $ 0.1-0.2nmol L À1 dFe at 100 m; Bowie et al. 2009) raises the additional question: does LL increase phytoplankton Fe requirements, leading to more pronounced Fe limitation of DCM populations compared to surface waters?Some phytoplankton species exhibit Fe-light antagonism, whereby cellular Fe requirements increase under LL, due to the increased number of Fe-rich photosynthetic units as cells photoacclimate (Sunda and Huntsman 1997).However, some SO phytoplankton have adapted to low Fe environment by increasing their light-harvesting antennae size at LL, which does not increase their Fe requirements (Strzepek et al. 2012(Strzepek et al. , 2019;;Ryan-Keogh et al. 2017).Recent field studies of coastal Antarctic surface communities support this strategy, where more severe Fe stress was observed under saturating light compared to lightlimiting conditions (Alderkamp et al. 2019;Vives et al. 2022), but no studies have examined Fe-light interactions in LL acclimated SO DCM communities (Baldry et al. 2020).
The use of Fe for photosynthesis occurs alongside manganese (Mn), an essential component of the oxygen-evolving complex of photosystem II (PSII ;Raven 1990) and other metalloenzymes, including superoxide dismutase (Peers and Price 2004).Dissolved Mn (dMn) concentrations are sufficiently low to limit phytoplankton growth in the SO (Browning et al. 2021).However, Mn limitation has not been investigated in SO DCM, where phytoplankton Mn requirements may be modulated by photoacclimatory strategy (Hawco et al. 2022), and where the supply of dMn may be enhanced by internal mixing with subsurface waters containing higher dMn concentrations (Middag et al. 2011).
This study investigated the effects of increasing light, Fe, and Mn supply on the physiology and composition of a SO DCM/DBM phytoplankton assemblage.This > 20 m-thick DCM is hypothesized to have formed after the decline of the surface phytoplankton community following Fe and silicic acid limitation (Boyd et al. 2023).We performed bioassay experiments under different dFe/dMn concentrations and light levels, and measured growth rates, Fe/carbon uptake, nutrient utilization, photophysiological parameters, and community composition.This enabled us to probe the environmental drivers-and their interactions-influencing the productivity of SO DCM.

Experimental design
The experiment was performed during the Southern Ocean Large Area Carbon Export voyage onboard RV Investigator (IN2020_V08) during the austral summer (December 2020-January 2021).A SO DCM was identified through increased fluorescence and extracted Chl a concentrations at 55.47 S, 138.34 E (Supporting Information Fig. S1).Unfiltered (i.e., grazers present) trace-metal clean seawater was collected at 87 m depth using a trace metal rosette and incubated in deckboard temperaturecontrolled recirculating seawater incubators (3.4 AE 0.2 C, n = 12; in situ DCM temperature 2.7 C), allowing phytoplankton to follow their regular diel light : dark cycles.Control (no additions), +Mn, +Fe, and +FeMn treatments were performed in triplicate within 2 L polycarbonate bottles (Nalgene).Iron and Mn additions were prepared in 0.01 mol L À1 hydrochloric acid (HCl) using ultrapure salts of Fe and Mn chloride, added to reach a final concentration of at least 2 nmol L À1 within incubation bottles.We employed two light treatments: LL (1% of incident irradiance), mimicking light conditions at the DCM and high light (HL; 12.4% of incident irradiance), simulating shoaling up to 40 m within the surface mixed layer observed on station (Supporting Information Fig. S2, Supporting Information Table S1; Boyd et al. 2023).The appropriate light intensities were determined from the proximate CTD photosynthetic active radiation profile (PAR; Supporting Information Fig. S2) and reproduced within incubation bottles using neutral-density mesh screening (Supporting Information Table S1).This resulted in the daily median of PAR ranging from 27 to 69 μmol photons m À2 s À1 for the HL treatment and 2-6 μmol photons m À2 s À1 for the LL treatment (Supporting Information Fig. S3).
Prior to the final sampling, incubated seawater was dispensed into 300 mL acid-washed polycarbonate bottles and spiked with 20 μCi of sodium 14 C-bicarbonate (NaH 14 CO 3 ; specific activity 1.85 GBq mmol À1 ; PerkinElmer) and 0.2 nmol L À1 of an acidified 55 Fe solution ( 55 FeCl 3 in 0.1 mol L À1 Ultrapure HCl; specific activity 30 MBq mmol À1 ; PerkinElmer).Bottles were incubated ($ 4 h post-dawn) in the shipboard incubators for 24 h under the same experimental conditions before sequential filtration through 20, 2, and 0.2 μm polycarbonate filters (Poretics).The filters were washed with titanium(III) ethylenediaminetetraacetic acid-citrate reagent to remove Fe bound to particle surfaces (Hudson and Morel 1989).Finally, filters were placed in 20 mL glass vials (Wheaton), acidified with 200 μL of 1.2 mol L À1 HCl, and stored at room temperature for analysis on shore.

Sample analyses
Photophysiology, macronutrients, and Chl a concentrations were measured onboard.The photochemical efficiency (F v /F m ) and functional absorption cross-section of PSII (σ PSII ) were measured using a Fast Repetition Rate Fluorometry (445 nm excitation; Soliense) on LL-acclimated (≤ 2 μmol photons m À2 s À1 ) communities following Latour et al. (2023b).Macronutrients were measured by segmented flow analysis (Rees et al. 2018).Chlorophyll a concentrations were measured by pigment extraction in 90% acetone for 18-24 h at À20 C before reading the fluorescence before and after the addition of 10% HCl on a Turner Trilogy fluorometer, against a 90% acetone blank (Holm-Hansen et al. 1965).Onshore, BSi concentrations were determined after BSi conversion to silicic acid through leaching with 0.1 mol L À1 sodium hydroxide at 85 C for 2.25 h, before determination using spectrophotometry (Paasche 1973).Filters for POC determination were exposed to 2 mol L À1 fuming HCl for about 12 h to remove carbonates before measurement on a Sercon-Callisto continuous flow isotope ratio mass spectrometer (using alanine OAS [Sercon] as external standard and sucrose, gelatine, USGS 65 and USGS 61 as in-house standards).Carbon and Fe uptake rates were determined by measuring disintegrations per minute on a liquid scintillation counter (PerkinElmer Tri-Carb 2910 TR).Filters were incubated for over 24 h before analysis in 10 mL of Ultima Gold liquid scintillation cocktail (PerkinElmer).Daily carbon incorporation rates were estimated following Hoppe et al. (2017), after correction for ambient dFe and dissolved inorganic carbon concentrations.Dissolved Fe and Mn concentrations were measured through high-resolution inductively coupled plasma mass spectrometry (Element XR; Thermo Fisher Scientific) after preconcentration and matrix removal (Ellwood et al. 2020).We calculated Mn deficiency relative to Fe (Mn*) according to Browning et al. (2021).
Flow cytometry samples were analyzed using an Aurora Cytek flow cytometer (Ferderer et al. 2022).Three populations were distinguished in each sample: (1) picoeukaryotes, (2) nanoeukaryotes, and (3) microeukaryotes based on violet laser excitation, red Chl a fluorescence, and the area-integrated forward scatter (FSC-A) signal as a proxy for phytoplankton size.Gate positions were adjusted per light treatment due to photoacclimation (Cullen 2015).The relative importance of each group in terms of population biovolume (F pop ) was calculated following Bach et al. (2018) using FSC-A converted to biovolume with a power regression (Selfe 2022; see Supporting Information).Bacterial counts were measured on the same instrument after staining the samples Table 1.Initial characteristics of the DCM water incubated and overlying waters: dissolved Fe (dFe), dissolved Mn (dMn), biogenic silica (BSi), particulate organic carbon (POC), chlorophyll a concentrations (Chl a), photochemical efficiency of PSII (F v /F m ) and functional absorption cross-section of PSII (σ PSII ).Blank spaces indicate no data were collected.n = 1 for all measures.with SYBRG I (1000-fold dilution) and following a 15 min dark incubation period at room temperature (Marie et al. 1999).

Statistical analyses
To study the effect of metal additions and light on the tested response variables, linear mixed effect models were performed using R (R "stats" packages; R Core Team 2020).Models were fit with the "lme" function of the "lme4" package, using maximum likelihood.The variability between replicates (or bottle effect) was included as a random effect in all analyses.Using the "drop1" function with a Chi-squared test (null hypothesis of independence), the best model fit was selected by sequentially eliminating variables among the fixed effects: Mn, Fe, light, and in some cases, size class (Fe/carbon uptake) or gated population (flow cytometry).To normalize model residuals, data were log 10 transformed before analyses.Iron and Mn addition treatments were treated as two separate factors, each possessing two levels ("True" for addition or "False" for no addition).When treatment effects were suggested by the model fit (with more than twofactor levels), we performed pairwise comparisons using the "emmeans" package with the Tukey method.All statements of significance (p < 0.05) hereafter refer to the linear mixed effect model results and pairwise comparisons.

Results
Initial characteristics of the DCM/DBM feature, along with oceanographic conditions, are discussed by Boyd et al. (2023).
Trace metal samples revealed little change in dFe and dMn concentrations from the surface to the DCM depth (87 m; Table 1).The DCM phytoplankton community showed relatively elevated photochemical efficiency (F v /F m = 0.48; Table 1) and was dominated by picoeukaryotes in terms of cell counts (Supporting Information Table S2).

Physiological responses to light and micronutrients
The best-fit model identified influences from Fe and light on macronutrient drawdown and the production of Chl a/POC.After 10 d of incubation, increased light conditions (without metal additions) stimulated macronutrient drawdown as well as POC, BSi, and Chl a accumulation (Fig. 1).Iron additions (+Fe and +FeMn) had no detectable effect on macronutrient, POC and BSi concentrations under LL.However, Fe additions stimulated Chl a production under LL (Fig. 1D).With increased light levels, +Fe and +FeMn stimulated significant macronutrient drawdown along with significant POC/Chl a accumulation (Fig. 1A-F).No clear statistical differences were observed from pairwise comparisons of BSi concentrations.However, the linear mixed effect model suggested interacting effects of Fe, Mn, and light on BSi production.In addition, we observed lower BSi : POC ratios under +Fe and +FeMn additions with HL, compared to the control and +Mn treatments (Supporting Information Fig. S4).
The best-fit model identified effects from Fe and light on F v /F m .After 10 d of incubations under LL, phytoplankton communities had high F v /F m (> 0.55; Fig. 1G).Increased light levels (with no metal additions) led to a significant decrease in F v /F m .However, Fe additions resulted in significantly higher F v /F m compared to the control treatment under HL (Fig. 1G).The σ PSII was solely influenced by Fe (Fig. 1H), and significantly decreased when Fe was added.

Phytoplankton community composition
Regarding the evolution of the phytoplankton population biovolume, the best-fit model identified influences from Fe, light, and the population studied.Picoeukaryotes dominated phytoplankton cell counts across all treatments and light conditions (Supporting Information Table S2), while nanoeukaryotes dominated community biovolume across all treatments under LL.Under HL, nanoeukaryotes remained dominant in the control and +Mn treatments, while Fe additions decreased their population biovolume (Fig. 2).Within the +Fe treatment, microeukaryotes dominated population biovolume while the +FeMn treatment led to equal contributions of nano-and microeukaryotes.

Carbon and iron uptake
Regarding Fe and carbon uptake rates, the best-fit model identified influences from Fe, Mn, light, and size classes.Carbon uptake rates were 8-82 times higher under HL compared to LL, while Fe uptake rates were 8-30 times faster under HL (Figs.S5, S6).Under LL, microeukaryotes (> 20 μm) dominated primary productivity (Fig. 3A) with significantly higher carbon uptake rates under +Fe and +FeMn compared to the control.Under HL, microeukaryotes also dominated carbon uptake rates, but all size classes showed significantly elevated rates under +Fe and +FeMn compared to the control and +Mn treatments (Fig. 3B).
In the control treatments under both light conditions, Fe uptake was dominated by picoplankton (picoeukaryotes and bacteria, 0.2-2 μm).However, Fe additions (+Fe and +FeMn) stimulated Fe uptake rates of nano-(2-20 μm) and microeukaryotes (> 20 μm) under LL (Fig. 3C) and all size classes under HL (Fig. 3D), to the point where microeukaryotes dominated Fe uptake rates (+Fe treatment of HL).In addition, Mn was observed to stimulate picoplankton Fe uptake rates compared to the control under both light conditions.
The Fe : C uptake ratios were generally higher in picoplankton and lower within microeukaryotes (Fig. 3E,F).Under LL, no differences in Fe : C ratios were observed within picoplankton, while nano-and microeukaryotes were characterized by significantly higher Fe : C ratios under +Fe and +FeMn compared to the control and +Mn treatments (Fig. 3E).Under HL, no significant changes in Fe : C uptake ratios were observed.

Discussion
Iron and light both limited the growth of phytoplankton sampled from the DCM.However, under our experimental conditions, light had the greater effect in stimulating macronutrients drawdown (up to 1.8-fold) and Chl a/POC accumulation (1.8and 2.5-fold), when comparing HL to LL control treatments.Still, significantly higher Chl a production (1.6-fold, Fig. 1D), decreased σ PSII (Fig. 1H) and higher net primary productivity (NPP) from microeukaryotes (> 20 μm; Fig. 3) with Fe additions under LL indicated Fe stress within the DCM phytoplankton community (Alderkamp et al. 2019).Despite healthy cells being present in the natural LL conditions (F v /F m > 0.55; Fig. 1G), carbon uptake rates remained low (NPP < 0.3 μmol L À1 d À1 ) as has been previously observed in DCM of this region (gross primary productivity < 0.7 μmol L À1 d À1 , Westwood et al. 2011).These results suggest that, while physiologically healthy, the studied DCM population had lower productivity than mixed-layer populations.Yet, this DCM persisted for 3 months, and by extrapolating from the daily carbon uptake rate and integrating over the thickness of the DCM, we calculate that NPP could be 150 mmol L À1 m À2 month À1 , suggesting this feature may still be important in terms of carbon fixation (Boyd et al. 2023).

Iron-light interactions
The transfer of resident cells to HL and Fe led to a major upregulation in their physiology, suggesting these cells were primed for more optimal conditions.Under HL, Fe additions significantly stimulated macronutrient drawdown (up to 8fold) and Chl a/POC production (4-or 10-fold, respectively) compared to the control (Fig. 1).The community shift toward microeukaryotes (> 20 μm; Fig. 2) under +Fe addition, coincident with an increase in BSi production and silicic acid consumption (Fig. 1), indicated Fe stimulated the growth of large diatoms.These results agree with previous DCM field studies that observed strong responses from large diatoms to increasing Fe and light conditions (Hopkinson et al. 2007;Hopkinson and Barbeau 2008).Under HL, lower BSi : POC ratios with Fe additions supported Fe stress relief (Conway et al. 2016), while low Fe : C ratios measured in microeukaryotes suggested these large diatoms are particularly effective at carbon uptake per unit Fe.
Table 2. Three estimates of growth rates for control and +Fe treatments at high light (HL) and low light (LL) using net primary production (NPP), particulate organic carbon (POC), and chlorophyll a (Chl a).NPP was calculated by summing the fractions presented in Fig. 3A,B The upregulation of phytoplankton physiology in response to a combined increase in Fe and light has, in previous bioassay experiments, been taken as evidence for an antagonist Felight relationship: that is, more Fe is needed for photosynthesis at LL (Maldonado et al. 1999;Moore et al. 2007).To test for Fe-light antagonism, we calculated the growth rates of low Fe control (μ ÀFe ) and high Fe treatments (μ +Fe ) under HL and LL using two approaches (Table 2): we calculated the turnover rate of carbon on the final day of incubation (μ (d À1 ) = NPP (μmol L À1 d À1 )/POC (μmol L À1 )), and from the change in both Chl a and POC over the 10 d incubation period (μ (d À1 ) = lnΔChl a or POC/time).If photoacclimation to LL increases phytoplankton Fe demand, we would expect μ ÀFe /μ +Fe to be lower under LL compared to HL conditions.Instead, we observed that μ ÀFe /μ +Fe was higher at LL (0.41-0.88) compared to HL (0.28-0.50).Despite the caveats associated with these μ calculations (POC biased with detrital material and Chl a decoupled from carbon content due to photoacclimation), these calculations produce consistent results.
Thus, we found no evidence for additional Fe demand at LL from DCM phytoplankton, as previously observed in Antarctic field studies (Ryan-Keogh et al. 2017;Alderkamp et al. 2019;Vives et al. 2022), and consistent with the interpretation of Strzepek et al. (2012Strzepek et al. ( , 2019)).Namely, if Fe is increased, but light supply is kept low, light absorption still limits cell growth.Conversely, if the light supply is increased, but Fe concentrations are kept low, cells are unable to fully use the light they absorb due to a lack of Fe.Cells can only use higher light when provided enough Fe to effectively catalyze photosynthetic electron transport.

Responses to manganese
Manganese additions did not significantly stimulate the DCM community.However, initial dMn concentrations were elevated ($ 0.3 nmol L À1 ), resulting in a high Mn* of 0.23, compared to the central Drake Passage where Mn limitation was previously observed (0.08-0.2 nmol L À1 and Mn* < 0.16; Browning et al. 2021;Balaguer et al. 2022).Moreover, the phytoplankton community had low growth rates under DCM conditions (0.02-0.08 d À1 ; Table 2) and exhibited a photoacclimatory strategy consistent with an increase in photosynthetic antenna size rather than increasing Mn-containing PSII reaction centers, both of which would decrease cellular Mn requirements (Hawco et al. 2022;Anugerahanti and Tagliabue 2023).We surmise that this combination of photoacclimation strategy and low (lightlimited) growth rates act in concert to minimize Mn limitation in the DCM.
Manganese additions were observed to increase picoplankton (0.2-2 μm) Fe uptake rates and to stimulate nanoeukaryotes (2-20 μm) biovolume against microeukaryotes (> 20 μm) in the +FeMn treatment compared to +Fe (Figs. 2, 3).This suggests that as growth rates increase (7-to 10-fold compared to the LL control; Table 2) and community composition changes, cellular Mn demands exceed supply for some phytoplankton species within the community, as observed in subantarctic waters (Latour et al. 2023b).In addition to PSII, Mn is required in metalloenzymes, including superoxide dismutase, the expression of which is thought to be influenced by light and Fe limitation but is poorly characterized in SO phytoplankton (McCain et al. 2021;McCain and Bertrand 2022).We hypothesize the Mn addition responses reflect species-specific Mn requirements, but additional field and lab-based studies remain essential to test this hypothesis.

Conclusion
This study aimed to characterize factors driving growth and primary production in a SO DCM/DBM feature.No clear signal of Mn limitation was observed, but subtle changes in Fe uptake rates and community composition suggested speciesspecific responses to Mn additions.We found DCM phytoplankton communities were limited by Fe and light, and increasing both together led to a major upregulation in phytoplankton physiology, especially for large diatoms.This suggested the DCM community was primed for more optimal conditions.We observed little increase in phytoplankton Fe requirements under LL and that an increase in either Fe or light will not suffice to stimulate photosynthetic electron transport.This study helps constrain the role of Fe and light in controlling SO DCM phytoplankton growth, and how ongoing changes occurring in the SO may impact this.

Fig. 1 .
Fig. 1.Concentrations of nitrate (A), phosphate (B), silicic acid (C), chlorophyll a (Chl a, D), particulate organic carbon (POC, E) and biogenic silica (BSi, F) measured in the initial water incubated ("Initial") and after 10 d of incubations in each treatment: Control, +Mn, +Fe, and +FeMn.Panel G shows the photochemical efficiency of photosystem II (F v /F m ) and panel (H) the functional absorption cross-section of PSII (σ PSII ), with both parameters blank-corrected.The two colors represent the light treatments: light gray for high light (HL) and dark gray for low light (LL).Error bars represent the standard deviations (n = 3, except for the Initial treatment where n = 1).Asterisks (*) show significant differences observed in the treatment additions compared to the control, identified after running linear mixed effect models and pairwise test comparison.Only Fe was observed to influence σ PSII , not allowing us to perform pairwise comparisons.

Fig. 2 .
Fig. 2. Relative population contribution to biovolume (F pop ) of the three gated phytoplankton populations: microeukaryotes, nanoeukaryotes and picoeukaryotes for the initial communities incubated ("Initial") and after 10 d of incubation in each treatment: Control, +Mn, +Fe, and +FeMn.The results are separated by light treatments: HL, high light and LL, low light.Error bars represent the standard deviations (n = 3, except for the Initial treatment where n = 1).The asterisk (*) shows a significant difference observed in an addition treatment compared to the control, identified after running linear mixed effect models and pairwise test comparisons.

Fig. 3 .
Fig. 3. Carbon uptake rates in μmol L À1 d À1 (A, B), Fe uptake rates in pmol L À1 d À1 (C, D) and Fe : C uptake ratio in μmol mol À1 (E, F) per size class (microeukaryotes: > 20 μm, nanoeukaryotes: 2-20 μm, and picoeukaryotes and bacteria: 0.2-2 μm), treatments (Control, +Mn, +Fe, +FeMn) and light conditions (low light, LL vs. high light, HL) assessed over a 24 h period on subsamples collected after 10 d of incubation.Note the different y-axis scale between light conditions for panels A-D.Asterisks (*) show significant differences observed in the treatment additions compared to the control, identified after running linear mixed effect models and pairwise test comparisons. .