Localized nutrient colimitation of phytoplankton growth rates across the subtropical South Pacific Ocean

Significance

Oceanic phytoplankton support marine food webs and influence global climate. Their growth typically depends on the availability of nutrients; principally nitrogen, phosphorus, and iron. An ongoing conundrum is whether multiple nutrient depletion leads to nutrient “colimitation” of phytoplankton growth rates. This has been difficult to address due to a lack of appropriate methods. Here, we use a different type of experiment to show that even when all three nutrients were simultaneously depleted across large extents of the South Pacific, phytoplankton growth was limited by nitrogen alone. However, we also found a zone where phytoplankton growth rates were robustly colimited by nitrogen and iron. Our initial findings suggest that alternative phytoplankton models might be needed to better understand ocean productivity.

Abstract

The simultaneous depletion of multiple nutrients in seawater potentially leads to colimitation of phytoplankton growth across large oceanic extents. Single limitation versus colimitation carries implications for mathematically predicting growth, its response to environmental forcing, and evaluating biogeochemical feedbacks. However, identifying colimited growth has proved challenging due to a lack of appropriate methods. Here, we present the results of 12 experiments conducted across the South Pacific that used a matrix of nutrient additions to strongly diluted surface seawater. Dilution restricted both grazing rates and nutrient drawdown due to phytoplankton accumulation. We find that despite simultaneous depletion of nitrate, phosphate, and iron concentrations throughout the oligotrophic gyre, community-level phytoplankton growth rates were only constrained by nitrogen. In contrast, zones of colimitation and serial limitation by nitrogen and iron were found along the eastern gyre margin. At the nitrogen-iron co-/serially limited sites, growth response surfaces to nutrient additions varied, suggesting the need for dynamic models to accurately represent colimited phytoplankton growth in the ocean.

Multiple nutrients are often simultaneously depleted in the surface ocean, suggesting potential for colimitation of phytoplankton growth (1, 2) (SI Appendix, Text 1). Nutrient colimitation, rather than limitation by a single nutrient, would change the way phytoplankton growth is calculated in the ocean (for example, in ocean biogeochemical models) with important ramifications for understanding ocean productivity, ecology, and biogeochemistry in the past, present, and future ocean (2, 3) (SI Appendix,Fig. S1 and Text 2).

Various forms of nutrient colimitation have been identified that operate on a range of scales (2, 4–6) (SI Appendix, Text 1). Identification of simultaneous nutrient colimitation in the ocean has relied on factorial nutrient addition experiments (7, 8). In these experiments, two or more nutrients are supplied at assumed-saturating levels both individually and in combination and changes in phytoplankton biomass are determined after several days. Nutrient colimitation is then defined as cases where biomass does not increase in response to addition of individual nutrients alone but does in response to combined supply of two or more nutrients. A form of nutrient limitation distinct from colimitation is serial limitation, whereby the addition of one nutrient alone increases biomass, but addition of a second nutrient in combination with the first increases biomass further. In this case, the experiment potentially provides an indication of how close the system is to nutrient colimitation. Experiments conducted in this way have found widespread colimitation and serial limitation (7).

However, there are challenges to interpreting the responses of such factorial nutrient addition experiments. First, observed biomass changes are the net result of gain (growth) and loss (grazing, viral lysis) processes. Therefore, independent assessment of only the gain (growth) process in response to nutrient supply is not possible. This could be important, for example, where increases in growth following nutrient supply are rapidly compensated by significant increases in loss processes (e.g., grazing; ref. 9). Such effects are likely greater in systems with smaller phytoplankton that are consumed by small, rapidly responding grazers, introducing an ecological dependence on the result. Second, experimental additions of nutrients to high concentrations potentially allow for elevated biomass accumulation and thereby rapid drawdown of nonadded nutrients that are initially nonlimiting, increasing the chance of observing serial limitation where multiple nutrients are added simultaneously (10). Third, observations of serial limitation in these types of experiments do not rule out colimitation of phytoplankton growth rates by multiple nutrients. Specifically, growth rates in situ might be depressed by the restricted availability of two (or more) nutrients, but when supplied individually, phytoplankton still grow in response to the supply of one of the nutrients, for example, due to a much higher flexibility in requirements for one nutrient than another.

Aside from experimental approaches, multiple indices of nutrient stress (defined as physiological responses implying nutrient shortage, including identification of nutrient stress genes, transcripts or proteins, enzyme activities, or fluorescence signals; ref. 8) have been utilized to infer single or costress (e.g., refs. 11–16). While these approaches have the major advantage that they reflect the initial, unperturbed microbial community as well as identifying costress at finer taxonomic resolution and potential subtle colimitations (2), it is not well constrained in which cases measurements can be inferred as colimitation of growth rates. Experimental approaches are therefore still required.

Here, we conducted a set of 12 factorial nutrient addition experiments under strong dilution with a matrix of nutrient additions across the breadth of the South Pacific Ocean through gradients in nitrate, dissolved iron (dFe), and phosphate concentrations (Fig. 1 and SI Appendix, Fig. S2). Experimental seawater was diluted by 73% with filtered trace-metal-clean seawater to i) restrict grazer consumption of phytoplankton biomass and produce net growth rates closer approximating specific growth rates; and ii) to prevent large absolute phytoplankton biomass increases and nutrient drawdown over the experiment duration (of either the amended or nonamended nutrients).

Fig. 1.

Experimental sites. Map of the South Pacific Ocean showing the cruise track (black line) and experimental sites (numbered circles) on a background of satellite-derived chlorophyll-a in March 2022.

Results and Discussion

Gradients in Nutrient Availability across the South Pacific Ocean.

A total of 12 experiments were conducted across the South Pacific Ocean at ca. 30°S along major gradients in nutrients as well as phytoplankton biomass and community structure (Table 1 and SI Appendix, Figs. S2 and S3). At the experimental sites, nitrate concentrations were always depleted (<53 nM) but showed moderately elevated values in the east (11 to 53 nM) relative to the central and western South Pacific (4 to 8 nM). Phosphate concentrations declined strongly from east to west (414 to 1 nM) while dFe concentrations showed moderately elevated values in the east and west and lowest values in the subtropical gyre (down to 0.13 nM). Based on these background nutrient concentrations, the first 10 experiments on transiting east-to-west focused on nitrogen and iron as potentially colimiting nutrients, whereas the final two experiments focused on nitrogen and phosphorus (Table 1 and SI Appendix, Fig. S2). The concentrations of supplemented nutrients across experiments were varied to i) take into account expected variability in absolute background concentrations through the ocean basin, and ii) to explore variability in responses to different applied concentration regimes. Phytoplankton concentrations (as indicated by chlorophyll-a concentrations) were strongly depleted in the central subtropical gyre (<0.05 µg L^–1) and were moderately elevated to the east and west (Table 1). Phytoplankton diagnostic pigments indicated communities were dominated by Prochlorococcus, Synechococcus, and haptophytes, with enhanced contributions of haptophytes in the east (SI Appendix, Fig. S3A).

Table 1.

Initial conditions for bioassay experiments

Experiment	1	2	3	4	5	6	7	8	9	10	11	12
Longitude (°E)	–72.84	–78.96	–86.40	–93.44	–102.32	–111.99	–128.00	–141.00	–152.00	–167.50	–177.18	177.97
Latitude (°N)	–32.50	–32.50	–32.50	–32.50	–31.67	–31.43	–31.43	–31.43	–26.17	–26.17	–25.87	–25.72
Duration (h)	38	38	34	37	33	46	46	49	50	48	42	42
SST (°C)	19.54	19.94	20.77	21.87	22.48	24.60	24.54	22.90	26.73	26.69	26.37	25.81
Salinity (PSU)	34.41	34.16	34.73	34.95	35.13	35.54	35.38	35.18	35.18	35.00	35.04	35.29
Nitrate (nM)	53	11	35	19	8	8	7	6	4	5	6	6
Phosphate (nM)	414	334	208	152	144	118	74	67	14	6	1	1
DOP (nM)	60	220	197	205	99	157	92	112	190	164	164	138
dFe (nM)	0.51	0.23	0.32	0.25	0.13	0.15	0.21	0.17	0.14	0.17	0.16	0.20
Nitrate/dFe	104	48	109	76	62	53	33	35	29	29	31	30
Nitrate/Phosphate	0.13	0.03	0.17	0.13	0.06	0.07	0.09	0.09	0.29	0.83	5.0	6.0
Chlorophyll-a (µg L–1)	0.720	0.210	0.130	0.090	0.062	0.037	0.046	0.046	0.064	0.102	0.118	0.129
Pro (cells mL–1)	213,791	92,771	36,632	ND	ND	ND	41702	ND	ND	152,281	ND	161,709
Syn (cells mL–1)	204,888	9,473	13,226	ND	ND	ND	784	ND	ND	1,580	ND	1,457
Pico (cells mL–1)	3,668	2,216	10,197	ND	ND	ND	325	ND	ND	309	ND	425

SST, sea surface temperature; DOP, dissolved organic phosphorus; dFe, dissolved iron; Pro, Prochlorococcus; Syn, Synechococcus; Pico, picoeukaryote; ND, not determined.

Response of Phytoplankton Growth Rate to Nutrient Amendment.

Concentrations of chlorophyll-a (SI Appendix, Fig. S4) and derived growth rates (Fig. 2) showed evidence for nutrient limitation at all sites. In the far east near the Chilean continental shelf (Experiments 1 and 2), in the central subtropical gyre (Experiments 4 to 10), as well as in the far western edge of the basin (Experiments 11 and 12), growth was only enhanced following the supply of nitrogen. This was evidenced by the diluted phytoplankton community growing in response to the lowest nitrogen amendment (25 to 100 nM) alone, without any supplemental iron or phosphate supply. Serial growth responses (i.e., supplemental responses) were observed following the combined supply of nitrogen and iron in Experiment 2. At this site, provided the lowest concentration of nitrogen was added (100 nM), amendment with either progressively higher nitrogen or iron enhanced growth further and amendment with the highest nitrogen and iron concentrations generally enhanced growth the most. Finally, Experiment 3, located slightly further west than Experiment 2, demonstrated colimitation by nitrogen and iron; specifically, no enhancement in growth rate was observed following the lowest nitrogen (100 nM) or iron (0.1 nM) amendments, but growth rate was progressively enhanced following combined amendment with both nutrients.

Fig. 2.

Response surfaces of chlorophyll-a based growth rates following two-dimensional nutrient additions. (A–J) Five-by-five matrix of nitrogen (N) and iron (Fe) additions in Experiments 1 to 10. (K and L) Five-by-five matrix of N and phosphorus (P) additions in Experiments 11 to 12. For all plots the x and y axes show the concentrations of the added nutrients and the z axis shows the growth rate, which is also visualized in the color shading. All plots show the mean values of triplicate, biologically independent replicates.

These findings contrast with standard experiments employing undiluted seawater with single high-concentration nutrient amendments (SI Appendix, Fig. S5; ref. 17). Standard experiments conducted in parallel on the same research cruise showed serial nitrogen-iron limitation extending throughout the subtropical gyre (as far west as 164°W, approximately the position of Experiment 10 of the current study; SI Appendix, Fig. S5; ref. 17). This difference was caused by the standard experiments having a combination of i) single, high amended nitrogen concentrations (2,000 nM), and ii) undiluted phytoplankton. The former (point “i”) will likely favor the growth of initially relatively rare, fast-growing, opportunistic phytoplankton (e.g., diatoms and other larger eukaryotes) that are effective at rapid nutrient consumption when nutrient limitation is fully relieved (8). This would lead to more rapid depletion of the second most deficient nutrient (i.e., iron) over the experimental timescale and make an observation of a serial response to this nutrient more likely. Similarly, the latter (point “ii”) results in larger absolute phytoplankton biomass changes following addition of the primary limiting nutrient, with an equivalently larger drawdown of the second most deficient nutrient in the seawater, again increasing the chance of observing a serial response to its supply.

That we found no serial iron limitation of diluted phytoplankton following nitrogen amendments of up to 800 nM in the central subtropical gyre strongly implies that the low dFe concentrations at these sites were not regulating in situ community-level growth rates (requiring greater phytoplankton biomass accumulation and nutrient drawdown before iron becomes limiting). Although direct comparisons between methods, specific locations, and seasons are challenging, our experimental results would then argue against previously observed iron stress signals (that is, physiological response to nutrient shortage) within the South Pacific Gyre (16, 18) as reflecting either iron limitation or nitrogen-iron colimitation of community-level phytoplankton growth rates.

The discrepancy for nitrogen and iron was also found for nitrogen and phosphate. In Experiments 11 and 12 in the far western portion of the transect, surface phosphate concentrations were strongly depleted (<10 nM; Table 1). In these waters, the standard, undiluted, high-level nutrient addition experiments demonstrated serial phosphate limitation (SI Appendix, Fig. S5; ref. 17), matching a number of experimental results found in other low-phosphate, low-nitrate regions of the ocean (7, 10, 19). In contrast, for the diluted phytoplankton community, we found that phytoplankton growth rate responses to nitrogen additions up to 200 nM were the same whether or not phosphate was supplied (Fig. 2 K and L). Similar to the nitrogen-iron experiments further to the east, this result strongly implies phosphate was not regulating in situ community-level growth rates.

The initial phosphate concentrations at Experiment sites 11 and 12 were highly depleted (measured as 1 nM; Table 1), therefore, even a very small amount of phytoplankton growth following nitrogen addition to the diluted seawater over the experimental timescale might be expected to deplete this further and lead to a serial limitation response. Specifically, calculations using a range of plausible ratios of phytoplankton nitrogen: phosphorus and the measured amount of phytoplankton accumulation over the experimental timescales suggested that phosphate should become serially limiting following even a low nitrogen amendment (SI Appendix, Fig. S6 and Text 3). Either one or a combination of i) even lower and/or highly flexible phosphorus requirements of the extant phytoplankton community (11), ii) sufficient cellular phosphate reserves (20, 21), and/or iii) active utilization of the much larger, partially bioavailable dissolved organic phosphorus pool (13) (Table 1) could explain this observation. Our experimental findings furthermore imply that measurements of phosphorus stress (physiological response to nutrient shortage), which are typically very strong in low phosphate waters such as these (11, 13, 16, 22), neither reflect phosphate limitation, nor nitrogen-phosphate colimitation, of community-level phytoplankton growth rates.

In contrast to the single nitrogen limitation found elsewhere, strong evidence for colimitation of phytoplankton growth rates by nitrogen and iron was found in Experiment 3 (Fig. 2C). Although measured dFe concentrations at this site were not particularly low (0.32 nM), nitrate concentrations were relatively enhanced (35 nM) and the dissolved seawater nitrate: iron concentration ratio was the highest of all encountered sites (Table 1). This concentration ratio potentially reflects enhanced supply rates of nitrate relative to iron to surface seawaters, which would favor nitrogen–iron colimitation relative to nitrogen-only limitation found at other sites (23, 24). Shoaling isolines of temperature, salinity, and nitrate around this site suggested that this was driven by northward intrusion of a different water mass (with lower temperature, lower salinity, and higher macronutrients; SI Appendix, Fig. S7).

Summary and Implications.

To our knowledge, Experiments 2 and 3 are the first observations of growth responses of nitrogen-iron serial/colimited phytoplankton to a matrix of dissolved nitrogen-iron concentrations (Fig. 2 B and C). For the range of amended nitrogen and iron concentrations, we observe a distinct “L”-shaped zero net growth isoclines (ZNGI) in Experiment 3, characteristic of a Liebig limitation model where nutrient limitation switches stepwise between nitrogen and iron limitation and a relatively narrow transition zone of colimitation in between (SI Appendix, Fig. S8). In comparison, in Experiment 2, following relief of the proximal nitrogen limitation via the lowest nitrogen amendment, growth rates in the remaining nitrogen-iron concentration space showed more curved ZNGIs characteristic of other models (e.g., multiplicative model; ref. 3). Although our five-by-five matrix of amended nutrient concentrations is restrictive in distinguishing subtly different response surfaces associated with different mathematical nutrient limitation models, the results of Experiments 2 and 3 suggest oceanic biogeochemical systems might not be uniquely characterized into one (if any) model. For example, despite being geographically close, a Bray–Curtis analysis of phytoplankton community composition strongly distinguished the sites of Experiments 2 and 3 (SI Appendix, Fig. S3), which possibly contributed to this difference.

The occurrence of colimitation in the ocean and growth rate responses to nutrient amendment are likely strongly regulated by nutrient supply rates and ratios to the surface ocean as well as microbial demands (8, 24). In line with previous suggestions (14, 24), our results suggest that at broad ocean scales, simultaneous nutrient colimitation of growth rates appears restricted to the boundaries between biogeochemical provinces that are limited by different nutrients, and more specifically, in between nitrogen- and iron-limited systems. While our results showed no widespread nitrogen–iron or nitrogen–phosphate colimitation of growth rates throughout the subtropical gyre, the extent of nitrogen–iron colimitation found at the gyre boundary was also restricted to one site. This restricted geographic extent was likely strongly influenced by our research cruise track, which had limited coverage of the high–low nitrate transition that generally characterizes subtropical gyre boundaries globally. Future experiments such as those conducted here through and along such boundaries could be used to investigate this further. At the same time, this would enable a more informed classification of nitrogen–iron colimitation within the available mathematical models (3) as well as potential environmental and ecological drivers of variability in response surfaces (Fig. 1 and SI Appendix, Figs. S2 and S3). Ultimately, this would inform choices on how phytoplankton growth is calculated in biogeochemical ocean models used to predict the impacts of climate change (25).

Materials and Methods

Research Cruise and Sample Collection.

Sample collection and experiments were conducted on the RV Sonne SO289 cruise from 18th February to 8th April 2022, representing section GP21 of the GEOTRACES programme. Surface (~2 to 3 m depth) seawater was sampled from a custom-built towed-fish fitted with acid-washed tygon tubing with suction provided by an acid-washed Teflon bellows pump (Dellmeco DM15/55TTT-E). Seawater was pumped directly into a purpose-built clean-air plastic tent with positive air pressure maintained via inward airflow passed through a high-efficiency particulate air filter. All sample collection and nutrient spiking were carried out within the clean-air plastic tent. Depth profiles (ca. 0 to 250 m) of nutrients were obtained using a trace-metal-clean titanium-frame CTD system fitted with Niskin-X bottles (Ocean Test Equipment), collecting samples at six discrete depths. A separate stainless steel CTD, equipped with an in vivo fluorescence sensor and photosynthetically active radiation sensor, guided target depth selection.

Nutrient Addition Matrix Experiments.

A total of 12 five-by-five matrix experiments with low-level nitrogen (N) and iron (Fe), or N and phosphorus (P) additions were performed with diluted (27% unfiltered) seawater to estimate phytoplankton growth rates. Here, dilution reduces grazing rates (i.e., by reducing grazer–phytoplankton interaction) as well as reducing overall phytoplankton biomass increases and dissolved nutrient drawdown over the experimental duration. We assumed net growth rates at the 27% dilution level to be approximately equivalent to specific growth rates (26–29), with further dilution potentially causing issues with accurately measuring chlorophyll-a concentrations in highly oligotrophic waters. To achieve the 27% dilution level, an acid-washed 60-L carboy (Nalgene) was used to collect seawater, by filling with 16 L of natural unfiltered seawater and then 44 L of filtered seawater (through a 0.8/0.2 μm pore-size filter capsule, AcroPak 500, Pall). The filling time was ~30 min. The diluted seawater was immediately dispensed (800 mL) into 75 acid-washed 1-L polycarbonate bottles (Nalgene). Before filling the carboy, additional unfiltered seawater was collected for initial measurements of dissolved macronutrient and trace element concentrations, chlorophyll-a concentrations, cell counts, and diagnostic phytoplankton pigments. Incubation bottles were spiked in triplicate with a matrix of nutrients. Nitrogen was added as a combined nitrate and ammonium solution, with equal contributions of both N species for each N amendment. Amendments were i) N (0, 100, 200, 400, 800 nM NaNO₃ + NH₄Cl) and Fe (0, 0.1, 0.2, 0.4, 0.8 nM FeCl₃) for Experiments 1 to 5, or ii) N (0, 50, 100, 200, 400 nM NaNO₃ + NH₄Cl) and Fe (0, 0.05, 0.1, 0.2, 0.4 nM FeCl₃) for Experiments 6 to 8, or iii) N (0, 25, 50, 100, 200 nM NaNO₃ + NH₄Cl) and Fe (0, 0.025, 0.05, 0.1, 0.2 nM FeCl₃) for Experiments 9 to 10, or iv) N (0, 25, 50, 100, 200 nM NaNO₃ + NH₄Cl) and P (0, 5, 10, 20, 40 nM NaH₂PO₄) for Experiments 11 to 12. Macronutrient spikes were previously passed through a column of prepared cation exchange resin (Chelex-100, BioRad) to remove contaminating trace elements. The iron spike was prepared from 99+% purity solid standards dissolved in 0.01 M hydrochloric acid (Fisher Optima grade diluted in Milli-Q deionized water). Incubation bottles were tightly capped and placed in on-deck incubators flushed with continuously replenished surface waters from the ship’s underway flow-through system to maintain temperatures. The incubators were screened with Blue Lagoon screening (Lee Filters), which maintained irradiance at ~30% of that of the surface ocean. Following 33 to 50 h incubation, experiments were taken down, and each replicate was subsampled for chlorophyll-a.

Macronutrient Concentrations.

Samples for nitrate plus nitrite (hereafter simply referred to as nitrate) and phosphate were collected into 15-mL acid-washed polypropylene tubes, and then analyzed on board using a segmented flow injection nutrient autoanalyzer (QuAAtro, Seal Analytical), with detection limits of 5 nM and 1 nM, respectively (30). Samples for total dissolved phosphorus (TDP) were collected into 50-mL acid-washed polypropylene tubes. TDP samples were digested under elevated pressure (1.5 bar) and temperature (121 °C) for 30 min, following the addition of potassium peroxide, sodium hydroxide, and boric acid. The oxidized TDP samples were then analyzed using the nutrient autoanalyzer (QuAAtro, Seal Analytical). Samples were blank-corrected for P content of digestion chemicals. Dissolved organic phosphorus concentration was subsequently calculated as the difference between TDP and phosphate.

Trace Element Concentrations.

Samples for dFe were filtered (0.8/0.2 μm pore-size filter capsule, AcroPak 500, Pall) into 125-mL acid-washed low-density polyethylene bottles (Nalgene) and acidified to pH 1.9 with ultrapurehydrochloric acid (UpA, Romil) onboard. After 6 mo, samples were measured following preconcentration using SeaFAST (Elemental Scientific Inc.), with the subsequent analyses completed by an inductively coupled plasma mass spectrometry (ICP-MS; Element XR, Thermo Fisher Scientific), with quantification by standard addition (31). Certified reference materials (SAFe D2, CASS-6 and NASS-7, National Research Council Canada) were analyzed alongside the samples and produced good agreement with their consensus values (weighted mean absolute percentage error < 5.5%).

Chlorophyll-a Concentrations.

Samples for chlorophyll-a were filtered (~200 mL for natural seawater; ~800 mL for diluted seawater) onto 25-mm diameter glass microfiber filters (GFF, nominal pore size of 0.7 μm, Fisher MF300), extracted in the dark for 16 to 24 h in 10 mL 90% (by volume) acetone at –20 °C, and then analyzed on a precalibrated Turner Designs Trilogy fluorometer (32).

Flow Cytometry.

Flow cytometry samples (2 mL) were fixed with paraformaldehyde at a 1% final concentration (methanol-free 16% 10-mL glass ampules, Alfa Aesar/Thermo Fisher), mixed with a vortex, left for 10 min at room temperature in the dark, and then directly frozen at –80 °C. Samples were analyzed upon return to shore using a FACSCalibur flow cytometer (Becton Dickenson, Oxford, United Kingdom). Cell counts were carried out in CellQuest software (Becton Dickenson). Plots of orange fluorescence versus red fluorescence were used to identify and enumerate Synechococcus from other picophytoplankton, and plots of side scatter versus red fluorescence (with Synechococcus gated out) were used to enumerate Prochlorococcus.

Diagnostic Phytoplankton Pigment Concentrations.

Samples for phytoplankton pigment analyses were filtered (2 to 4 L for natural seawater) onto 25-mm diameter Fisher MF300 GFF filters and frozen immediately at –80 °C. Upon return to land, pigments were extracted in 90% acetone in plastic vials by homogenization of the filters using glass beads in a cell mill, centrifuged (10 min, 5,200 rpm, 4 °C), then the supernatant was filtered through 0.2-μm polytetrafluoroethylene filters (VMR International) and subsequently quantified by reverse-phase high-performance liquid chromatography (HPLC, Dionex UltiMate 3000 LC system, Thermo Fisher Scientific) (33). Pigment standards were from Sigma-Aldrich (USA) and the International Agency for ¹⁴C Determination (Denmark). Measured pigment assemblages were used to estimate the contributions of 11 phytoplankton groups to total chlorophyll-a (the sum of chlorophyll-a and divinyl chlorophyll-a) via CHEMTAX (34), employing published field-derived starting pigment ratios (35).

Phytoplankton Growth Rate Calculation and Simulation.

Calculation of chlorophyll-a-based growth rates (${\mathrm{\mu }}_{\mathrm{measured}}$) used the equation:

$${\mathrm{\mu }}_{\mathrm{measured}}=\raisebox{1ex}{$\ln \left(\frac{{\text{Chl}}_{\mathrm{t}}}{{\text{Chl}}_{\mathrm{I}}}\right)$}\!\left/ \!\raisebox{-1ex}{$\mathrm{t}$}\right.,$$

where ${\mathrm{Chl}}_{\mathrm{t}}$ is the final chlorophyll-a concentration in the treatment bottles after incubation, ${\mathrm{Chl}}_{\mathrm{I}}$ is the initial chlorophyll-a concentration, and t is the duration of the experiment.

Statistical Analyses.

Statistical analyses and visualizations were performed using R version 4.3.2 (36), with specific functions from the “agricolae,” “ggplot2,” and “vegan” packages. To explore patterns in environmental variation among sampling sites, a Principal Coordinates Analysis (PCoA) was performed based on a dissimilarity matrix calculated from community composition. Vector fitting was applied to assess the relationship between environmental variables and the ordination space using the envfit function from the vegan package. The significance of each environmental variable was determined through 999 permutations, and variables with P < 0.05 were considered to contribute significantly to the ordination. Arrows representing environmental variables were overlaid on the PCoA plot, with asterisks indicating significant contributors.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

Experimental data are available in Yuan and Browning (2025) (37). Satellite chlorophyll-a and temperature at 4 km resolution were extracted from the Aqua Moderate Resolution Imaging Spectroradiometer (Aqua/MODIS) Level-3 product from the NASA ocean color website (https://oceancolor.gsfc.nasa.gov) (38).