2.2.1 Zooplankton

The zooplankton compartment represents copepod-type zooplankton (COPs, organisms larger than 200 µm, Fig. 3) whose biomass depends on prey ingestion, respiration, excretion, egestion (faecal pellets), and predation by higher trophic levels. Copepod prey ingestion is represented using the formulation by Auger et al. (2011). Copepods ingest smaller prey, and grazing rates depend on prey type preference, as well as on temperature and light due to their effect on prey abundance. Copepods feed with decreasing preference for NCMs, nano- and micro-phytoplankton (NMPHYTO), and CM (Verity and Paffenhofer, 1996) and release ammonium (${{\mathrm{NH}}_{\mathrm{4}}}^{+}$), phosphate (${{\mathrm{PO}}_{\mathrm{4}}}^{\mathrm{3}-}$), and dissolved organic carbon (DOC) through excretion, contributing to the POM compartment through egestion and mortality. Mortality due to predation by higher trophic levels represents a closure term (Fig. 2).

Figure 3Repartition of modeled organisms (COPs: copepods; PICO: picophytoplankton; NMPHYTO: nano- and micro-phytoplankton; and BAC: heterotrophic bacteria) in size classes and trophic interactions between them. Preference values are indicated in gray for copepods (Verity and Paffenhofer, 1996) and NCMs (Epstein, 1992; Price and Turner, 1992; Christaki et al., 2009) and CMs (Christaki et al., 2002; Zubkhov and Tarron, 2008; Millette et al., 2017; Livanou et al., 2019). Sizes are expressed in micrometers (µm).

Download

2.2.2 Phytoplankton

We considered two types of phytoplankton based on size (Fig. 3): picophytoplankton (PICO) and nano- and micro-phytoplankton (NMPHYTO). PICO includes autotrophic prokaryotic organisms such as Prochlorococcus spp. and Synechococcus spp., which are ubiquitous in the Mediterranean (Mella-flores et al., 2011). NMPHYTO aims to represent phytoplankton larger than 2 µm and smaller than 200 µm. It mainly includes diatoms and autotrophic nanoflagellates. As diatoms are an important component of Mediterranean spring blooms (Margalef, 1978; Leblanc et al., 2018) and cover a wide size range, we decided to consider them to be representative of the NMPHYTO.

Both the NMPHYTO and PICO biomasses are affected by photosynthesis, respiration, nutrient uptake, exudation, and grazing. Photosynthesis depends on light, nutrients, and temperature (based on Geider et al.'s (1998) formulation). Respiration depends on photosynthesis (a constant fraction of photosynthetically produced carbon) and nutrient uptake. Nutrient uptake is temperature dependent. NMPHYTO and PICO both consume nitrate (${{\mathrm{NO}}_{\mathrm{3}}}^{-}$), ${{\mathrm{NH}}_{\mathrm{4}}}^{+}$, and ${{\mathrm{PO}}_{\mathrm{4}}}^{\mathrm{3}-}$, while PICO also consume dissolved organic nitrogen (DON) and dissolved organic phosphorus (DOP) (Duhamel et al., 2018). The uptake of DON and DOP depends on temperature and the cell's nutritional state. If the cell is replete in N (P), then DON (DOP) uptake is null. Both phytoplankton groups exude DOC, DON, and DOP proportionally to their internal content of carbon (C), nitrogen (N), and phosphorus (P) (Fig. 2).

2.2.3 Heterotrophic bacteria

Heterotrophic bacterial biomass results from balancing growth and losses due to bacterial production, respiration, nutrient uptake, remineralization, predators grazing, and natural mortality (Kirchman and Gasol, 2018; Faure et al., 2010a, b). Bacterial production depends on DOC and particulate organic carbon (POC) and is limited by temperature and substrate availability. Heterotrophic bacteria consume particulate organic nitrogen (PON), particulate organic phosphorus (POP), DON, DOP, ${{\mathrm{NH}}_{\mathrm{4}}}^{+}$, and ${{\mathrm{PO}}_{\mathrm{4}}}^{\mathrm{3}-}$, which they remineralize to ${{\mathrm{NH}}_{\mathrm{4}}}^{+}$ and ${{\mathrm{PO}}_{\mathrm{4}}}^{\mathrm{3}-}$. They contribute to the DOM pool through natural mortality, which depends on temperature (Fig. 2).

2.2.4 Dissolved inorganic matter

The DIM compartment consists of the nutrients ${{\mathrm{NO}}_{\mathrm{3}}}^{-}$, ${{\mathrm{NH}}_{\mathrm{4}}}^{+}$, and ${{\mathrm{PO}}_{\mathrm{4}}}^{\mathrm{3}-}$, as well as dissolved oxygen (O₂) and the carbonate system variables (total alkalinity: TA, dissolved inorganic carbon: DIC, pH_T, partial pressure of CO₂: pCO₂, and calcium carbonate: CaCO₃). Nutrient concentrations are affected by heterotrophic bacterial remineralization, uptake, and excretion of organisms (${{\mathrm{NH}}_{\mathrm{4}}}^{+}$ and ${{\mathrm{PO}}_{\mathrm{4}}}^{\mathrm{3}-}$ only) and nitrification (${{\mathrm{NO}}_{\mathrm{3}}}^{-}$ and ${{\mathrm{NH}}_{\mathrm{4}}}^{+}$ only). Nitrification (i.e., ${{\mathrm{NO}}_{\mathrm{3}}}^{-}$ production from ${{\mathrm{NH}}_{\mathrm{4}}}^{+}$) is temperature and O₂ dependent. O₂ concentration is calculated from photosynthesis, respiration, nitrification, and air–sea exchanges. The other variables included in the DIM compartment are the carbonate system variables (see Barré et al., 2023b, for details).

2.2.5 Particulate and dissolved organic matter

In Eco3M_MIX-CarbOx, we only considered detrital POM and labile DOM. The POM and DOM compartments are affected by zooplankton, mixotrophs, phytoplankton, and heterotrophic bacteria (see above and Fig. 2).

The state equations, process formulations, and associated parameter values for all compartments can be found in Appendices B to E.

2.3.1 Implementation of NCMs

NCMs (type IIIB) are defined as photosynthetic protozoa; i.e., they are primarily phagotrophic but can complement their carbon uptake through photosynthesis (Stoecker, 1998). In Eco3M_MIX-CarbOx, the NCMs are based on ciliates and belong to microplankton (Esteban et al., 2010; Fig. 3). Their dynamics are governed by the following set of balance equations (see Appendix C for a more detailed description of each term):

Being primarily phagotrophic, NCM grazing is implemented in a similar way to zooplankton grazing in that they can ingest preferentially smaller prey items while having certain preferences for different prey types. From most to least preferred prey, NCMs feed on heterotrophic bacteria, picophytoplankton, CMs, and nano- and micro-phytoplankton (Epstein et al., 1992; Price and Turner, 1992; Christaki et al., 1999). By ingesting photosynthetic prey, NCMs acquire the capacity to photosynthesize by temporarily sequestering chloroplasts (Putt, 1990). This process is modeled as a grazing flux between the chlorophyll concentrations of photosynthetic prey and NCMs (Eq. 2). The NCMs' capacity to photosynthesize degrades over time unless fresh chloroplasts are sequestered (Eq. 3, based on Leles et al., 2018).

In the above equations, $\mathrm{PREY}\in [\mathrm{CM},\mathrm{NMPHYTO},\mathrm{PICO}]$, G_MAX, K_NCM, Φ, and k_MORT,Chl represent the maximum grazing rate, the grazing half-saturation constant, the NCM preference for a specific prey type, and the loss rate of captured photosystems, respectively (see Appendix E for details). NCM_X and PREY_X are the NCM and PREY concentrations of element X, respectively. ${\text{Gra}}_{{\mathrm{NCM}}_{\mathrm{Chl}}}^{{\mathrm{PREY}}_{\mathrm{Chl}}}$ and ${\text{Degrad}}_{{\mathrm{NCM}}_{\mathrm{Chl}}}$ are in millimoles per cubic meter per second ($\mathrm{mmol}{\mathrm{m}}^{-\mathrm{3}}{\mathrm{s}}^{-\mathrm{1}}$).

As NCM photosynthesis depends on the sequestered chloroplasts from prey, we created a prey-dependent formulation to represent it (Eq. 4). We based our formulation on Geider et al. (1998), who provide a photosynthesis flux nutrient, temperature, and light dependence. In this formulation, a maximum photosynthetic rate is first calculated (${\mathrm{P}}_{\mathrm{MAX}}^{\mathrm{C}}$) based on the C-specific photosynthetic rate at a reference temperature of the photosynthetic organism (${\mathrm{P}}_{\mathrm{REF}}^{\mathrm{C}}$). This rate is nutrient and temperature dependent and is next multiplied by the light limitation function. We applied parameters of the prey except for the nutrient limitation, which is calculated based on NCM internal content in N and P as the process takes place inside the NCM cells, albeit using the prey's chloroplasts.

In the above equation, $\mathrm{PREY}\in [\mathrm{CM},\mathrm{NMPHYTO},\mathrm{PICO}]$ and ${\mathrm{P}}_{\mathrm{MAX}}^{\mathrm{C}}$ are the maximum photosynthetic rate per second, and ${\text{Photo}}_{{{\mathrm{NCM}}_{\mathrm{C},\mathrm{PREY}}}_{\mathrm{C}}}^{\text{DIC}}$ is the NCM photosynthetic flux associated with the chloroplast from the considered prey in millimoles per cubic meter per second ($\mathrm{mmol}{\mathrm{m}}^{-\mathrm{3}}{\mathrm{s}}^{-\mathrm{1}}$). ${\mathrm{P}}_{\mathrm{REF}}^{\mathrm{C}}$ is the C-specific photosynthetic rate at a reference temperature (see Appendix E for values for each prey). f^T and limI are temperature and light limitation functions, respectively (see Appendix C for detailed formulations). $f_{\mathrm{Q}}^{\mathrm{G}}$ is a nutrient limitation function which express the nutritional state of the cell and is based on the X ($X\in [\mathrm{N},\mathrm{P}]$)-to-C ratio (i.e., NCM_X to NCM_C in this case).

$f_{\mathrm{Q}}^{\mathrm{G}}$ is dimensionless. ${\mathrm{Q}}_{\mathrm{c},\text{min}}^{\mathrm{N}}$, ${\mathrm{Q}}_{\mathrm{c},\text{min}}^{\mathrm{P}}$, ${\mathrm{Q}}_{\mathrm{c},\text{max}}^{\mathrm{N}}$, and ${\mathrm{Q}}_{\mathrm{c},\text{max}}^{\mathrm{P}}$ represent the minima and maxima of the X-to-C ratios (see Appendix E for values used for NCM). When the cellular C content is high relative to other elements, then the $f_{\mathrm{Q}}^{\mathrm{G}}$ value approaches 0 and vice versa.

The photosynthetic fluxes from each prey type were weighted by NCM prey preference and summed according to

where $\mathrm{PREY}\in [\mathrm{CM},\mathrm{NMPHYTO},\mathrm{PICO}]$ and ${\text{Photo}}_{{\mathrm{NCM}}_{\mathrm{C}}}^{\text{DIC}}$ are the NCM photosynthetic flux in millimoles per cubic meter per second ($\mathrm{mmol}{\mathrm{m}}^{-\mathrm{3}}{\mathrm{s}}^{-\mathrm{1}}$), and Φ is the NCM prey type preference (values in Appendix E).

Finally, respiration, exudation, and excretion are based on grazing fluxes and nutrient limitations. Grazed C is consumed through respiration, and excess C is exuded as DOC. The amount of respired or exuded C is determined by the cell's nutritional state. Respiration and exudation fluxes are high when NCM C content is high relative to N or P and vice versa. We used the same reasoning for grazed N (P) which is exuded as DON (DOP) or excreted as ${{\mathrm{NH}}_{\mathrm{4}}}^{+}$ (${{\mathrm{PO}}_{\mathrm{4}}}^{\mathrm{3}-}$) when NCM N (P) content is high (see Appendix C for details).

2.3.2 Implementation of CMs

CMs (type IIA) are defined as phagotrophic algae; i.e., they are primarily phototrophic but can ingest prey to obtain limiting nutrients (Stoecker, 1998). CMs are based on dinoflagellates which belong mainly to nanoplankton but can also be found in microplankton (Stoecker, 1999, Fig. 3). Their dynamics are governed by the following set of balance equations (see Appendix C for details):

CM photosynthesis is temperature, light, and nutrient dependent following Geider et al. (1998):

where ${\mathrm{P}}_{\mathrm{MAX}}^{\mathrm{C}}$ is the maximum photosynthetic rate per second, ${\text{Photo}}_{{\mathrm{CM}}_{\mathrm{C}}}^{\text{DIC}}$ is the CM photosynthetic flux in millimoles per cubic meter per second ($\mathrm{mmol}{\mathrm{m}}^{-\mathrm{3}}{\mathrm{s}}^{-\mathrm{1}}$), and ${\mathrm{P}}_{\mathrm{REF}}^{\mathrm{C}}$ is the C-specific photosynthetic rate at a reference temperature (see Appendix E for CM value). f^T, $f_{\mathrm{Q}}^{\mathrm{G}}$, and limI are temperature, nutrient, and light limitation functions, respectively (see Appendix C for detailed formulations of f^T and limI and Eq. (5) for the formulation of $f_{\mathrm{Q}}^{\mathrm{G}}$).

Like picophytoplankton, CMs assimilate dissolved inorganic nutrients (${{\mathrm{NO}}_{\mathrm{3}}}^{-}$, ${{\mathrm{NH}}_{\mathrm{4}}}^{+}$ and ${{\mathrm{PO}}_{\mathrm{4}}}^{\mathrm{3}-}$) and DOM (DON and DOP). Uptake fluxes are calculated by using a Michaelis–Menten equation and are limited by temperature. DOM uptake also depends on the nutritional state of the cell in that the higher the cell's N (P) content is, the lower the DON (DOP) uptake will be.

When DIN and/or DIP limits the growth, CMs can ingest smaller prey to supplement their N and/or P needs (Stoecker et al., 1997). CMs feed on heterotrophic bacteria (preferred) and picophytoplankton (less preferred; Christaki et al., 2002; Zubkhov and Tarron, 2008; Millette et al., 2017; Livanou et al., 2019), and the same grazing formulation as for zooplankton and NCMs is used, except that CM grazing is limited by DIN (DIP) concentration and light (Stoecker, 1997, 1998; Eq. 9).

In the above equations, $\mathrm{PREY}\in [\text{BAC},\mathrm{PICO}]$ and ${\text{Gra}}_{{\mathrm{CM}}_{\mathrm{C}}}^{{\mathrm{PREY}}_{\mathrm{C}}}$ are in millimoles per cubic meter per second ($\mathrm{mmol}{\mathrm{m}}^{-\mathrm{3}}{\mathrm{s}}^{-\mathrm{1}}$). $f_{\mathrm{I},\text{inhib}}^{\mathrm{CM}}$ and $f_{\text{NUT,inhib}}^{\mathrm{CM}}$ are the (dimensionless) inhibitions of grazing by light and nutrients, respectively. G_MAX, K_CM, Φ, α_Chl, ${\mathrm{P}}_{\mathrm{REF}}^{\mathrm{C}}$, and K_NUT represent the maximum grazing rate, the grazing half-saturation constant, the CM prey preference, the chlorophyll-specific light absorption coefficient, the C-specific photosynthesis rate at a reference temperature, and the half-saturation constant for the considered nutrient (${{\mathrm{NO}}_{\mathrm{3}}}^{-}$, ${{\mathrm{NH}}_{\mathrm{4}}}^{+}$, or ${{\mathrm{PO}}_{\mathrm{4}}}^{\mathrm{3}-}$), respectively (values in Appendix E). ${\mathrm{Q}}_{\mathrm{C}}^{\mathrm{Chl}}$ is the chlorophyll-to-carbon ratio, and E_PAR is the irradiance value. The grazing is also affected by CM internal content in N and P ($f_{\mathrm{Q}}^{\mathrm{G}}$ term, Eq. 5).

CMs ingest prey to supplement their needs in N and P only, exuding grazed C as DOC (Stoecker, 1998; Eq. 10). Hence, DOC is released through two metabolic pathways' exudation of carbon acquired via (i) photosynthesis and (ii) grazing.

In the above equation, $\mathrm{PREY}\in [\text{BAC},\mathrm{PICO}]$ and ${\text{Exu}}_{{\mathrm{CM}}_{\mathrm{C}}}^{\text{DOC}}$ are in millimoles per cubic meter per second ($\mathrm{mmol}{\mathrm{m}}^{-\mathrm{3}}{\mathrm{s}}^{-\mathrm{1}}$). ${\text{Photo}}_{{\mathrm{CM}}_{\mathrm{C}}}^{\text{DIC}}$ is the photosynthetic flux in millimoles per cubic meter per second ($\mathrm{mmol}{\mathrm{m}}^{-\mathrm{3}}{\mathrm{s}}^{-\mathrm{1}}$) (Eq. 8), and ${\text{Gra}}_{{\mathrm{CM}}_{\mathrm{C}}}^{{\mathrm{PREY}}_{\mathrm{C}}}$ is the grazing flux for the considered prey in millimoles per cubic meter per second $\mathrm{mmol}{\mathrm{m}}^{-\mathrm{3}}{\mathrm{s}}^{-\mathrm{1}}$ (Eq. 9). frac_resp represents the fraction of respired carbon from photosynthesis (values and units in Appendix E).

The formulations for DON and DOP exudation are similar. Exudation only occurs on the N and P obtained from nutrient uptake. In other words, neither N nor P obtained from grazing are released through exudation. When DIN (DIP) concentration is limiting, CMs will ingest prey in addition to the uptake of nutrient. As their internal content in N (P) is particularly low, exudation of DON (DOP) is not allowed (equal to 0). When DIN (DIP) concentration is high, CMs only perform nutrient uptake (no grazing as it only supplements N and P needs in limiting conditions). Then, all the N (P) from uptake is exuded as the cell is already loaded in N (P), and as no grazing is performed, no N (P) from grazing is exuded in these conditions. Respiration uses the same formulation as phytoplankton; i.e., a constant fraction of photosynthesis and nutrient uptake is respired (Sect. 2.2.2 and Appendix C).

2.4.1 Assessment of mixotrophs

To be considered as correctly represented by the model, NCMs and CMs must fulfill the properties listed in Table 2. These properties have been stated by Stoecker (1998) to provide conceptual models to represent the different types of mixotrophs.

Table 2Summary of NCM and CM properties based on Stoecker (1998). DIN represents the sum of ${{\mathrm{NO}}_{\mathrm{3}}}^{-}$ and ${{\mathrm{NH}}_{\mathrm{4}}}^{+}$, and DIP represents ${{\mathrm{PO}}_{\mathrm{4}}}^{\mathrm{3}-}$. Food is represented by prey concentration.

Download Print Version | Download XLSX

To verify these properties, we designed several numerical experiments (Tables 3 and 4) in which we modify one of the following features: prey biomass, DIN and DIP concentrations, or irradiance. We first ran a reference simulation (referred to as Replete in Table 3) in which we set all the previous features to a maximum value during the entire simulation. Maximum prey biomass was obtained by multiplying the initial condition by 2 (sum of the initial carbon prey biomass multiply by 2), maximum DIN and DIP concentrations were chosen based on high values observed at SOLEMIO (Pujo-Pay et al., 2011), and maximum irradiance corresponds to the mean value of simulated irradiance for the SOLEMIO station by the meteorological model WRF (Yohia, 2017). Next, we ran the low nutrients (low nutrient values observed at SOLEMIO multiplied by 0.1 – low-nut simulation in Table 3 for NCMs and CMs), low prey concentration (maximum prey concentration multiplied by 0.5 – low-food simulation in Table 3 for NCMs and CMs), and low light (maximum value multiplied by 0.05 – low-light simulation in Table 3 for CMs only). For NCMs, to verify the light-dependent property (NCMP3), it is also necessary to set the NCM concentration to a constant during the entire simulation; we performed another reference simulation and a low-light simulation in which the NCM concentration was constant (initial condition, NCM replete with constant and NCM low light with constant in Table 4, respectively). Finally, we compare the simulations to their associated reference simulation.

Table 3Summary of the simulations performed to check NCM and CM properties (excluding NCMP3). For NCMs, [PREY] stands for the sum of CMs, nano- and micro-phytoplankton, picophytoplankton, and heterotrophic bacterial biomasses. For CMs, [PREY] stands for the sum of picophytoplankton and heterotrophic bacterial biomasses.

Download Print Version | Download XLSX

Table 4Summary of the simulations performed for NCMP3. [PREY] stands for the sum of CMs, nano- and micro-phytoplankton, picophytoplankton, and heterotrophic bacterial biomasses.

Download Print Version | Download XLSX

2.4.2 Typical vs. limited conditions

After verifying mixotroph properties, we simulated three types of light and nutrient regimes for the BoM: typical, nutrient limited, and light limited (Table 5). With these three regimes, we aim to reproduce typical and limited conditions (i.e., nutrient and light limited) in the BoM. Simulations are run for 2017 at SOLEMIO station. The Eco3M_MIX-CarbOx spin-up period is about 3 months. To avoid initial conditions impacting on our results, we ran 3 years of simulations (i.e., repetition of 2017 three times), and we present the results for the second year of simulation.

Table 5Summary of simulation properties. Configurations without mixotrophs are detailed in the Supplement.

Download Print Version | Download XLSX

For the typical scenario, light was modeled using the solar irradiance from the WRF meteorological model for SOLEMIO station (Table 1), and ${{\mathrm{NO}}_{\mathrm{3}}}^{-}$, ${{\mathrm{NH}}_{\mathrm{4}}}^{+}$, and ${{\mathrm{PO}}_{\mathrm{4}}}^{\mathrm{3}-}$ concentrations were based on in situ observations at SOLEMIO during 2017 (values from SOMLIT) using a linear interpolation between fortnightly data points (Fig. 4). In these conditions, we also performed two simulations without mixotrophs as control simulations. These simulations correspond to two configurations of Eco3M_MIX-CarbOx without mixotrophs: a first one in which mixotrophs and their associated biogeochemical processes are simply deleted (D configuration in Table 5) and a second one in which we replaced mixotrophs with organisms with strict diets (R configuration in Table 5). These configurations and their results are presented in the Supplement.

Figure 4Time series of interpolated surface (a) ${{\mathrm{PO}}_{\mathrm{4}}}^{\mathrm{3}-}$ concentration, (b) ${{\mathrm{NO}}_{\mathrm{3}}}^{-}$ concentration, and (c) ${{\mathrm{NH}}_{\mathrm{4}}}^{+}$ concentration (lines) from fortnightly measurements with SOLEMIO data (markers) during 2017. The studied events are shaded in gray.

Download

In the nutrient-limited scenario, the ecosystem is limited by DIN and DIP concentrations only, using values 10 times lower than the minima observed at SOLEMIO, keeping both DIN (sum of ${{\mathrm{NO}}_{\mathrm{3}}}^{-}$ and ${{\mathrm{NH}}_{\mathrm{4}}}^{+}$, $\mathrm{6.75}\times {\mathrm{10}}^{-\mathrm{3}}$ and $\mathrm{7.5}\times {\mathrm{10}}^{-\mathrm{4}}$ mmol m⁻³, respectively) and DIP constant for the duration of the simulation. The Eco3M_MIX-CarbOx model was initially developed to be run with low nutrient concentrations, representative of the Mediterranean Sea (Morel and Andre, 1991). To ensure that organisms were not limited by light, we multiplied the typical irradiance by 2.

In the light-limited scenario, we only applied 5 % of the typical irradiance, while DIN ($\left[{{\mathrm{NO}}_{\mathrm{3}}}^{-}\right]=\mathrm{1.35}$ mmol m⁻³, $\left[{{\mathrm{NH}}_{\mathrm{4}}}^{+}\right]=\mathrm{0.15}$ mmol m⁻³) and DIP concentrations were set to winter values at SOLEMIO.

For the three simulations, we used typical values of the BoM to represent temperature, salinity, wind speed, and atmospheric pCO₂, as described in Table 1.

3.2.1 Annual scale

Through the year of 2017, phytoplankton biomass was dominated by CMs, closely followed by PICO and at some distance by NMPHYTO (Fig. 6a).

CM and PICO chlorophyll concentrations show similar patterns, with values varying between 0.1 (on 18 February) and 0.3 mg Chl m⁻³ (on 24 May). The highest variability occurred between May and October. NMPHYTO chlorophyll concentrations varied between 0.01 (on 25 June) and 0.16 mg Chl m⁻³ (on 20 March), with the lowest values occurring between May and July (Fig. 6e). The in situ values reached a maximum of 1.71 mg Chl m⁻³ on 15 March, linked to the Rhône River intrusion event. Between June and November, in situ values were generally lower compared to the other months, and a minimum of 0.1 mg Chl m⁻³ was reached on 11 October. The modeled chlorophyll concentration shows less variations than the in situ data, especially since the model was unable to reproduce either the maximum related to the Rhône intrusion on 15 March or the minimum on 11 October. Nevertheless, the modeled values, ranging from 0.25 to 0.64 mg Chl m⁻³, are generally of the same order of magnitude as the in situ observation. Both the model results and in situ data yielded the same mean chlorophyll concentrations of 0.4 mg Chl m⁻³.

Total nutrients (Fig. 6f) varied between 0.08 (in summer and autumn) and 5.6 mmol m⁻³ (reached on 15 March). CM and PICO nutrient limitation status remained fairly stable near the mean value of 0.71; however, organisms are more limited in late spring and summer (between May and July). NMPHYTO nutrient limitation status is more variable, showing higher limitations in late spring and summer (between late April and July) and lesser limitation in early spring and late summer.

The light limitation status clearly reflects the diurnal and seasonal variations in incident irradiance (Fig. 6g). Throughout the year, all three phytoplankton groups show nearly identical levels of limitation.

3.2.2 Winter mixing event

During the winter mixing event, a ${{\mathrm{PO}}_{\mathrm{4}}}^{\mathrm{3}-}$ maximum was recorded at SOLEMIO station (0.21 mmol m⁻³, Fig. 4a). In terms of C biomass, CMs were most dominant, followed by NMPHYTO and PICO (Fig. 6b).

CMs and PICO chlorophyll decreased slightly, while NMPHYTO chlorophyll remained constant (Fig. 6e). The decrease in CMs and PICO chlorophyll is also visible in the total chlorophyll, which dropped from 0.41 to 0.28 mg Chl m⁻³ (Fig. 5e).

The nutrient limitation remained fairly stable for all phytoplankton groups (Fig. 6f).

During the event, irradiance was low (<40 W m⁻², Fig. 6g) and decreased at the end of January due to bad weather. CMs, NMPHYTO, and PICO light limitation status remained similar throughout this event and at a relatively low value (0.3).

3.2.3 Rhône River intrusion

The Rhône River intrusion resulted in a ${{\mathrm{NO}}_{\mathrm{3}}}^{-}$ maximum at SOLEMIO station (5.48 mmol m⁻³, Fig. 4b). Model results indicate that, during the event, phytoplankton was dominated by NMPHYTO followed by CMs and PICO (Fig. 6c).

All three chlorophyll concentrations increased, with the most significant increase occurring for NMPHYTO (from 0.11 to 0.15 mg Chl m⁻³), which surpassed PICO at the beginning of the event (Fig. 6e).

The intrusion also led to a significant increase in modeled total nutrients (reaching 5.5 mmol m⁻³). Nutrient limitation status was similar for all groups and remained between 0.67 and 0.75, showing no significant variations during the event (Fig. 5f).

While irradiance levels were moderate (around 60 W m⁻²), all three groups were still light limited (values of about 0.5, Fig. 6g).

3.2.4 Cortiou water intrusion

During the Cortiou water intrusion, the in situ ${{\mathrm{NH}}_{\mathrm{4}}}^{+}$ concentration reached a maximum of 1.06 mmol m⁻³ (Fig. 4c) at SOLEMIO station. In the model, phytoplankton composition was dominated by PICO and CMs, with NMPHYTO being a distant third (Fig. 6d).

Chlorophyll increased in all groups, resulting in an increase in total chlorophyll from 0.36 to 0.52 mg Chl m⁻³ (Fig. 6e).

During the event, the sum of nutrients reached 1.53 mmol m⁻³ with a clear ${{\mathrm{NH}}_{\mathrm{4}}}^{+}$ maximum. Nutrient limitation status was similar across groups and remained stable around 0.7 (Fig. 6f).

Irradiance levels were moderate (between 70 and 112 W m⁻²), leading to only a slight light limitation (values between 0.58 and 0.62, Fig. 6g).

3.3.1 Nutrient-limited conditions

In nutrient-limited conditions, the modeled yearly total C biomass, i.e., the sum of daily C biomass of each organism, was 349.5 mmol C m⁻³, divided between copepods (148.7 mmol C m⁻³), NCM (129.8 mmol C m⁻³) and heterotrophic bacteria (26.2 mmol C m⁻³), followed by the three phytoplankton groups, of which NMPHYTO had the lowest biomass (4.5 mmol C m⁻³, Fig. 7a).

Figure 7Yearly ecosystem C biomass composition and dynamics for copepods (COPs), NCMs, nano- and micro-phytoplankton (NMPHYTO), CMs, picophytoplankton (PICO), and heterotrophic bacteria (BAC). Yearly totals under (a) nutrient- and (b) light-limited conditions. Time series of daily averages under (c) nutrient- and (d) light-limited conditions. Note the different scales in panels (a) and (b), as well as in (c) and (d).

Download

Copepods and NCMs dominated the ecosystem, with copepods being more abundant between October to June, while NCMs dominated during the other months of the year. In early June, NCM biomass started to increase and reached a maximum of 0.56 mmol C m⁻³ on 18 July. Copepod biomass peaked shortly after (0.44 mmol C m⁻³ on 1 September). Heterotrophic bacteria biomass also started to increase in June and reached a maximum of 0.12 mmol C m⁻³ on 29 June. CM and PICO biomasses show similar dynamics, starting to increase in April and reaching a maximum in mid-June before decreasing into September. NMPHYTO biomass remained low and close to its mean value of 0.01 mmol C m⁻³ throughout the year (Fig. 7c).

3.3.2 Light-limited conditions

In light-limited conditions, the modeled yearly total C biomass was about 3 times higher than with nutrient limitation (1192.5 mmol C m⁻³). NCM dominated the ecosystem (462.3 mmol C m⁻³) followed by copepods (417.3 mmol C m⁻³). NMPHYTO biomass was the lowest (59.2 mmol C m⁻³, Fig. 7b).

Between late autumn and late spring, copepods dominate, while NCMs become dominant in terms of biomass between mid-February and September. During this period, NCM biomass appears to be more variable compared to copepods and reaches a maximum of 2.2 mmol C m⁻³ on 14 June. Also, heterotrophic bacteria showed a high variability, particularly in summer, while remaining close to 0.15 mmol C m⁻³ during the rest of the year. CM and PICO showed similar dynamics, with their biomass starting to increase in early March before decreasing from mid-April and increasing again from mid-May till summer. They also showed their highest variability in summer. NMPHYTO biomass oscillated between 0.12 and 0.2 mmol C m⁻³, showing a similar overall behavior to CMs and PICO, except that the NMPHYTO maximum was reached on 23 April and not in summer (Fig. 7d).

3.4.1 Carbon fluxes

In typical and nutrient-limited conditions, NCMs can meet their metabolic needs by ingesting prey and by photosynthesizing using sequestered chloroplasts. In typical conditions (Fig. 8a), NCMs obtained about three-quarters of their C through prey ingestion (74.2 %) and the remaining quarter through photosynthesis (25.8 %). The most significant loss terms are, in descending order, grazing by copepods, exudation of DOC, and respiration. In nutrient-limited conditions (Fig. 8b), C uptake by photosynthesis and predation are more balanced (43.4 % and 56.6 %, respectively), while the losses are similar to the typical scenario.

Figure 8Sankey diagrams showing the carbon (C) fluxes for NCMs (a, b) and CMs (c, d) in typical (a, c) and nutrient-limited (b, d) scenarios. Numbers represent the yearly averaged C fluxes. PS prey: photosynthetic prey; InPrey: ingested prey; SChlo: sequestered chloroplast; Chlo: chloroplast.

Download

In contrast, when CMs find themselves in typical conditions, they meet their metabolic needs almost entirely through photosynthesis, while grazing is almost negligible (Fig. 8c). The most important loss terms are grazing, followed by respiration and DOC exudation. In nutrient-limited conditions, the role of grazing increases, but only slightly, and photosynthesis remains the dominant source of C (Fig. 8d). Interestingly, C loss terms change considerably under nutrient limitation: predation decreased significantly to become the least important loss term, while more than half those losses now occur via DOC exudation, while respiration decreased slightly.

3.4.2 Nitrogen and phosphorus fluxes

CM can complement their normal N and P uptake, i.e., DIM and DOM uptake (referred to as total N or P uptake in Fig. 9), by grazing. In typical conditions, grazing is insignificant to both N and P uptake (Fig. 9a and c), while losses occur predominantly through exudation of DON and DOP, with predation representing only about one-third.

Figure 9Sankey diagrams showing (a, b) nitrogen (N) and (c, d) phosphorus (P) fluxes for CMs in (a, c) typical and (b, d) nutrient-limited conditions. Numbers represent the yearly averaged fluxes. InPrey refers to ingested prey, and Chlo refers to chloroplast. Total N (P) represents the sum of DIN (DIP) and DON (DOP) uptakes.

Download

In nutrient-limited conditions (Fig. 9b and d), the role of grazing has increased substantially and now provides about 40 % of the N and a quarter of the P requirements. Also, the loss terms have changed considerably, with N losses occurring almost exclusively due to grazing (Fig. 9b), while P losses appear to be equally split between DOP exudation and grazing (Fig. 9d).

4.2.1 Light

In our light-limited scenario, nutrient levels were kept artificially elevated throughout the year to prevent nutrients from becoming limiting and affecting the results. Light limitation had a considerable effect on total C biomass, which was almost halved under low light compared to typical conditions (1192.5 vs. 2016.3 mmol C m⁻³).

Ecosystem composition remained almost identical between light-limited and typical conditions (Fig. 10a). In fact, light limitation only directly impacts the three phytoplankton groups, while copepods and NCM are only impacted indirectly through the effect of light on their prey. Heterotrophic bacteria do not become light limited in our model (Appendix C).

Considering the fact that nutrients were kept artificially elevated in the light-limited scenario, it is not surprising that CM nutrition is almost entirely based on photosynthesis (99 %, result not shown); i.e., they behaved like strict autotrophs, and their mixotrophy did not represent a competitive advantage in this case. For instance, Stoecker et al. (1997) showed that, in low-light and high-nutrient conditions, the CM Prorocentrum minimum tend to photosynthesize rather than feed on prey as this latter mechanism only becomes relevant when inorganic nutrients are limiting. Thus, in light-limited conditions, the phytoplankton arrangement only depends on the organism's ability to photosynthesize.

Although CM biomass remains high in low light, its share of the pie decreases in favor of NMPHYTO, which seem to gain a slight edge. While the share of NMPHYTO increases slightly under low light, PICO appears to be unaffected (Fig. 10a, b). In this simulation, nutrient levels were kept artificially high to prevent nutrient limitation. By lifting the nutrient limitation, NMPHYTO, which is particularly sensitive to nutrients concentration, can grow more easily. In addition, NMPHYTO includes mainly diatoms, which are known to be advantaged in low-light environments (Fisher and Halsey, 2016). CMs are more affected by low light and are not able to use mixotrophy in these conditions (nutrient concentration is high). The low effect of light on PICO agrees with observations by Timmermans et al. (2005), who showed that, when nutrients are not co-limiting, picophytoplankton still developed well.

The winter mixing event is a useful example that illustrates the impact of light on phytoplankton. During this event, the weather was particularly cloudy, yielding low levels of ambient light and several decreases. These decreases in light level are reflected in the three phytoplankton groups' limitation status, which also decreased (which indicates an increase in limitation) (Fig. 6g).

4.2.2 Nutrients concentration

When nutrients are limiting, the shares of NCMs, CMs, PICO, and NMPHYTO decrease, while copepods and heterotrophic bacteria show a relative increase (Fig. 10a). We found that, when nutrient concentration was low, the ability of NCMs to photosynthesize was particularly useful as it provided nearly half their C uptake (Fig. 8). Nevertheless, NCM yearly total biomass does not exceed that of the copepods (Figs. 7a and 10a). In fact, despite their ability to photosynthesize, NCMs remained highly dependent on prey abundance. To prove this strong dependence of NCMs on their prey, Mitra et al. (2016) performed several simulations involving different planktonic communities such as heterotrophic bacteria, phytoplankton, and NCMs. They found that NCM biomass quickly increased, but once the available prey was consumed, it dropped just as quickly. Due to this strong prey dependency, NCMs cannot dominate the ecosystem throughout the year. Instead, we found that NCM biomass increased in summer (even exceeding copepods, Fig. 7c), right after CM and PICO biomass had increased, which in turn replenished the prey concentration.

Our modeled phytoplankton showed significant reactions to changes in nutrient concentration. While low nutrients led to an almost complete disappearance of NMPHYTO (Fig. 10a), CMs and PICO appeared to handle low nutrient concentrations more easily. On the one hand, PICO are known to be able to cope with nutrient-limited environments more efficiently than larger cells, mainly due to their small size, which results in higher nutrient affinity (Agawin et al., 2000). On the other hand, nutrient limitation allowed CMs to take full advantage of mixotrophy, which allows them to compensate for a lack in DIN and DIP by grazing. Thus, by using two different competitive strategies, both PICO and CMs can tolerate low nutrient conditions, allowing them to become the dominant phytoplankton groups in this scenario. Leles et al. (2018) also found relative increases in CMs when nutrient concentration decreased.

The Rhône River and Cortiou water intrusions are useful examples that illustrate the impact of nutrient concentrations on the ecosystem and phytoplankton compositions. The Rhône River intrusion led to high ${{\mathrm{NO}}_{\mathrm{3}}}^{-}$ concentrations, which in turn led to increased NMPHYTO growth, illustrating their high sensitivity to nutrient concentrations. NCM also fared well in this scenario and reached a dominant 39 % of the total C biomass (results not shown). In these conditions, NCM nutrition is mainly based on grazing (75.3 %) due to the high prey concentration, but photosynthesis still represents a high percentage of the nutrition. In fact, some mixotrophic ciliates (e.g., Laboea strobila) are known to be highly dependent on photosynthesis (Stoecker et al., 1988; Sanders, 1991; Esteban et al., 2010). Stoecker et al. (1988) calculated that, in this case, photosynthesis via sequestered chloroplasts could contribute up to 37 % of the ciliate's total carbon demand in resource-rich conditions. The Cortiou water intrusion led to high ${{\mathrm{NH}}_{\mathrm{4}}}^{+}$ concentrations, alleviating the nutrient limitation for the three phytoplankton groups, particularly in NMPHYTO (Fig. 6f). In fact, immediately before this intrusion event, the ambient nutrient concentration was very low, which explains the sudden response of phytoplankton. However, NMPHYTO still only represented 15 % of the total phytoplanktonic C biomass at the time (Fig. 6d), indicating that other factors are at play as well. As the Cortiou water intrusion took place during the summer upwelling period, we can hypothesize that temperature also played a role in shaping the phytoplankton composition.