Upward nitrate transport by phytoplankton in oceanic waters: balancing nutrient budgets in oligotrophic seas

Abundance and depth distribution of Rhizosolenia mats

Rhizosolenia mats (Fig. 1) were observed by divers at every station sampled over the 19-year period spanning the cruises (Fig. 2). At low abundance (<0.5 mat m^-3), patterns in the vertical distribution were difficult to detect. At high abundance (>0.5 mats m^-3), a surface maximum was often evident (up to ∼12.5 mats m^-3) with decreasing abundance at depth (Fig. 3). Mats were visible below depths the divers could access and were visible to the limit of vertical visibility (40–60 m). While mats were present at all stations in Fig. 2, abundance was quite variable and occasionally (4 of 96 stations) below detection limit of the sampling frame at each of the 4–6 depths measured (∼1 mat in 30 m³). For the 1989–2003 cruises (the 2008 cruise was snorkel only with no abundance data collected), average integrated abundance determined by divers was 4.1 ± 5.7 mats m^-3 with a range of 0.03–27.5 mats m^-2 excluding the 4 stations where mats were below enumeration limits (Fig. 4). These values were combined with literature reports from this area in Fig. 4, generating a 26-year summary of Rhizosolenia mat distribution and diver-estimated abundance.

The 2003 VPR imagery (RoMP 2003) revealed an abundance of Rhizosolenia mats that were ≤1 cm in size. These small-sized mats have previously been noted and are under-counted in diver surveys (Pilskaln et al., 2005; Villareal et al., 1999b). Our observations in RoMP2003 along a transect line from 168 to 177°W found mats were distributed to at least 150 m (Fig. 5). The vertical distribution had no consistent pattern with some stations (Sta. 7) showing a surface maximum, while other stations (Sta. 5) had a maximum at depth. In all cases, abundance did not decline to zero at the deepest strata (150 m). Integrated values (Table 1) ranged from 188–17,062 mats m^-2. The station with two sets of tows approximately 10 h apart (Sta. 12) showed good agreement between profiles with the two samples within 2% of the mean. When VPR and diver counts were compared, divers consistently under-estimated mat abundance. The 0–150 m integrated VPR counts were up 6–2,843 times higher than the diver-based 0–20 m integrated counts (Table 1) and a comparison of Figs. 3 and 5 reveals the much higher mat densities observed in general by the VPR. VPR-based integrated abundance varied nearly 200-fold from 80–17,062 mats m^-2 with 90% below diver accessible depths and had no relationship (r² = 0.08) to diver-based abundance in the 0–20 m range (Table 1). The 2003 counts were up to 100-fold higher than VPR-based abundance data collected 2,000 km to the east in 1996 (Villareal et al., 1999b).

Nitrogen isotope values

Rhizosolenia mats in all years were collected only in the upper mixed layer (<10 m depth), a zone that routinely has only nanomolar nitrate concentrations (Brzezinski, Villareal & Lipschultz, 1998; Casciotti et al., 2008; Dore & Karl, 1996). The buoyancy status of the mats is closely tied to their N status. Villareal et al. (1996) noted that when compared to negatively buoyant mats (sinkers), positively buoyant mats (floaters) have significantly higher internal nitrate pools (∼1–2 vs 8+ mM), lower C:N ratios (7–8 vs 9–11) and higher protein:carbohydrate ratios (0.4–0.6 to 1.1–1.6). These patterns are consistent with increasing nitrogen stress concurrent with the mats becoming negatively buoyant. During 2002–2003, C:N ratios in sinking mats were significantly higher than in ascending mats across the entire longitudinal gradient (Table 2), a marker resulting from unbalanced uptake of N and C and consistently tied to a vertical migration strategy (Villareal et al., 1996).

In the 2002 and 2003 samples, mat δ¹⁵N was uniformly elevated across the basin (Figs. 6 and 7) and averaged 2.91 ± 0.28 (95% C.I., n = 181) when combined with historical data (Villareal, Altabet & Culver-Rymsza, 1993; Villareal et al., 1999b; Villareal et al., 1996). Ascending mats were significantly (2.17 ± 0.32 per mil versus 3.53 ± 0.30 per mil, p = 0.05) depleted in ¹⁵N relative to sinking mats (Fig. 6, Table 2). Mats were also enriched in ¹⁵N relative to the suspended particulate material at the surface (Fig. 7). The δ¹⁵N of the ambient NO₃⁻ pool in 2002 (RoMP2002) at 200–400 m ranged from 5.29 to 6.73 per mil and was consistent with historical observations (Fig. 7). Inclusion of additional data from Station ALOHA (Casciotti et al., 2008) highlighted the lighter isotopic values of NO₃⁻ in the nutricline expected as the result of the remineralization of diazotrophically derived N. The light δ¹⁵N of mixed layer particulates is a pattern also seen at Station ALOHA and is considered typical of the oligotrophic gyres (Dore et al., 2002; Montoya, Carpenter & Capone, 2002; Montoya, Wiebe & McCarthy, 1992). The similarity of the Rhizosolenia mat δ¹⁵N to the deep NO₃⁻δ¹⁵N is evidence that mats are generally migrating to the 150–200 m depth range. In the mats, the isotopic signature results from luxury nitrate uptake at high (µM) NO₃⁻ concentrations at depth in the dark (Joseph & Villareal, 1998; Richardson et al., 1996) and assimilation from mM INP via nitrate/nitrite reductases in the low concentration (nM) NO₃⁻ of the upper euphotic zone. Nitrate uptake does not fractionate (Comstock, 2001; Robinson, Handley & Scrimgeour, 1998) and the internal pools initially represent the source nitrate δ¹⁵N.

Migrators cannot acquire sufficient NO₃⁻ in the mixed layer to create large INP. While rapid surge uptake at nanomolar NO₃⁻ concentrations has been demonstrated in the diatom Pheodactylum tricornutum (Raimbault, Slawyk & Gentilhomme, 1990), INP can only accumulate when assimilation via nitrate reductases is slower than uptake. Direct measurements of the migrating giant diatom Ethmodiscus indicate NO₃⁻ uptake at ambient concentrations were several orders of magnitude too low to support the observed nitrate reductase activity (Villareal et al., 1999a). Dark periods in elevated NO₃⁻ concentrations are required for INP to build up in Rhizosolenia spp. found in mats (Joseph & Villareal, 1998). Moreover, if these cells were building INP as a result of assimilating NO₃⁻ at the ambient concentrations, there should be only minimal differences between sinking and floating mats. Such uptake is inconsistent with the multiple proxies of N-limitation that are found in sinking mats and not found in floating mats (Joseph, Villareal & Lipschultz, 1997; Villareal et al., 1996). Due to the strictly internal N assimilation in the mixed layer, there is no net change of mat δ¹⁵N at the surface due to reductase fractionations and/or NO₃⁻uptake. However, mats in both years were actively excreting NO₃⁻ (Singler & Villareal, 2005). While we could not measure the isotopic composition of this excreted N, relatively light NO₃⁻ would preferentially leak out and result in the observed pattern of lower mat δ¹⁵N in floating (recently ascended) mats compared to sinking (depleted INP).

These data provide a picture of Rhizosolenia mat abundance across the Pacific Ocean as well as within their vertical migration range. The latitudinal distribution extends from ∼24° to ∼35°N with additional observations near Oahu, Hawaiʻi (Cowen & Holloway, 1996), the coastal California current, and equatorial Pacific (Alldredge & Silver, 1982). Mats were observed over 50° of longitude (∼1/2 the width of the Pacific Ocean) and were abundant at the western terminus of the cruise set. We found no further records in the Pacific Ocean west of this point, but the broad distribution in the Indian Ocean (Carpenter et al., 1977; Wallich, 1858; Wallich, 1860), North and South Atlantic Ocean (Caron et al., 1982; Carpenter et al., 1977; Darwin, 1860), equatorial Atlantic Ocean (Bauerfeind, 1987) and north and south Central Pacific Ocean (Alldredge & Silver, 1982) supports a reasonable expectation that their distribution extends across the entire Pacific Ocean (Villareal & Carpenter, 1989). Abundance is considerably lower in the Sargasso Sea (Carpenter et al., 1977), although they are still present. Our 2003 VPR observations confirm the earlier report that small Rhizosolenia mats dominate both numerically (Villareal et al., 1999b) and in particulate Si contribution (Shipe et al., 1999) in the N. Pacific. These small mats are virtually invisible to divers due to the low contrast of small mats, and the depth limitations imposed on blue-water SCUBA techniques (∼20 m) preclude diver enumerations at depths (Villareal et al., 1999b). We conclude that the pattern of numerically dominant small mats extending to depth is the prevailing distribution of Rhizosolenia mats and that the mats are both widespread and abundant in the Pacific Ocean.

Rhizosolenia mat δ¹⁵N values show a pattern dominated by values typical of sub-euphotic zone nitrate. Prior to this study, only a handful of values were published raising the possibility that these were not representative of larger scales. However, our current data set spans nearly 1/2 the Pacific Ocean and clearly shows high δ¹⁵N NO₃⁻ pools as an N source. Vertical migration is a consistent feature of their biology and occurs across the entire distributional range. A re-assessment of the quantitative importance of mat N transport is required and is particularly timely given the need to identify mechanisms capable of closing euphotic zone nitrate budgets (Ascani et al., 2013; Johnson, Riser & Karl, 2010). In the next section, we will consider the implications for nutrient cycling and the role of ascending motion in general in phytoplankton.

Significance to oceanic nutrient cycles

The upward nitrate flux problem derives from budgeting analysis that concludes that nitrate introduced at the base of the euphotic zone must be transported upward many 10s of meters to zones of net community production and export, and that this transport occurs along a negligible concentration gradient (Ascani et al., 2013; Johnson, Riser & Karl, 2010). In order to assess the potential role of Rhizosolenia mats in the Pacific to this process, we calculated flux rates using our new data and previously published models. Nitrogen transport rates are calculated from abundance data coupled to turnover models. Negative buoyancy increases as the mats undergo progressive N limitation and sink to depth (Villareal et al., 1996). Nitrate uptake occurs at depth and in the absence of light, leading to buoyancy reversals and ascent to the surface (Richardson et al., 1996). At the surface, the pattern repeats with some fraction of the nitrate being lost via excretion (Singler & Villareal, 2005). Protist parasitism has been noted and probably results in nitrate release as well (Caron et al., 1982; Villareal & Carpenter, 1989). The overall migration cycle is shown conceptually in Fig. 8. Note that the cell growth rates of the component diatoms can be as high as 0.6 d^-1 (Moore & Villareal, 1996a). Migration-based growth rates are slower due to time spent at sub-optimal light conditions while ascending/descending and during dark uptake of NO₃⁻.

Two estimates of migration-based growth rates have been published. Villareal et al. (1996) considered a simple box model that used measured mat ascent rates to calculate transit time to and from the nutricline, nitrate uptake rates based on a large diatom, and carbon-based doubling times based on oxygen evolution measurements of Rhizosolenia mats. Richardson et al. (1998) used a 13 layer model over the upper 120 m derived from Kromkamp & Walsby (1990) where mat biomass is represented in carbon and nitrogen units. Photosynthesis was modeled by a standard photosynthetic model (Platt, Gallegos & Harrison, 1980) modified to incorporate diel changes in irradiance, depth-dependent attenuation, and temperature Q₁₀ effects. Nitrate uptake was based on Richardson et al. (1996). Ascent and descent rates were based on changes in the carbon:nitrogen ratio derived from Villareal et al. (1996). Loss rates are unknown, and were adjusted in the model to maintain a stable migration pattern for 5 cycles. The Villareal et al. model yielded total migration growth rates of 0.19–0.28 d^-1, while the Richardson et al. model produced 0.11–0.15 d^-1. While both models converge on values in 0.1–0.2 d^-1 range, we have used a lower, more conservative estimate (0.14 d^-1) for our calculations of mat turnover.

The N transport rates calculated from the VPR abundance ranged from 6–444 µmol N m^-2 d^-1 (Table 1) with an average daily rate of 172 µmol m^-2 d^-1. However, our abundance data are not uniformly distributed across the year. Rhizosolenia mat observations are biased towards June–October due to weather constraints on diving operations. We have only limited reports from April/May (Alldredge & Silver, 1982; Villareal & Carpenter, 1989) and no quantitative estimates for the balance of the year. Therefore, we restrict our calculation to a conservative 6 month distribution window to calculate the impact of only a six month period on the annual budget (i.e., 6 month rates = the annual input via migration) based on abundance at each of our stations (range = 1.1–79.9 mmol N m^-2 y^-1). The upper value is directly comparable to the eddy injection N (88 mmol N m^-2 y^-1) calculated to balance the N budget in the upper 250 m (Johnson, Riser & Karl, 2010). These calculations suggest that N transport via Rhizosolenia mats scales on an event basis that is comparable to eddy injection of nitrate to the euphotic zone, while recognizing that upward transport is not sustained at that level. This calculation is a conservative underestimate since anecdotal observations indicate mats are present year round in the eastern Pacific (Alldredge & Silver, 1982).

Finding the proper spatial and temporal scales for comparison is a challenge. Eddy injection (a physical process) and Rhizosolenia mat dynamics (a biological process) likely operate, and are certainly recorded, on different time scales. For example, the nutrient budgets were assembled for the Hawaiʻi Ocean Time-Series region at Station ALOHA (22°45′ N), a latitude that has high trade winds much of the year that inhibit diving operations. Rhizosolenia mat data were collected several hundred kilometers to the north (∼28–30°) where wind conditions permit divers to routinely enter the water. Eddy turbulent kinetic energy and numbers of eddies in the mat collection areas are low (Chelton, Schlax & Samelson, 2011) relative to Station ALHOA. We have no site where both long-term N budgets and Rhizosolenia mat abundance are available. In addition, Rhizosolenia mats are not unique in their migration strategy, and comprehensive consideration of phytoplankton upward nitrate transport requires inclusion of other migrating phytoplankton taxa. A brief review is presented here to provide the required perspective and background to justify inclusion of these taxa in the subsequent calculations.

Other vertically migrating phytoplankton taxa: life history and abundance

The literature on other migrating, non-flagellated phytoplankton in the open sea is dispersed and the natural history of this group poorly represented in the literature of the past several decades. There are several taxa that must be represented and spanning a broad taxonomic range: Pyrocystis, Halosphaera, Ethmodiscus, and free living Rhizosolenia.

Pyrocystis species are positively buoyant warm water, non-motile dinoflagellates with a dominant cyst-like non-motile stage typically 10⁷ µm³ (Rivkin et al., 1984). They undergo a migration to the nutricline (Rivkin et al., 1984) and have been considered members of the shade flora (Sournia, 1982). Reproduction occurs by release of a brief reproductive stage from a cyst-like vegetative form (Swift & Durbin, 1971). Bilobate reproductive stages release immature vegetative stages that swell up to near full size in ∼10 min (Swift & Durbin, 1971), become positively buoyant within 13 h and indistinguishable from the cyst-like form after 15 h (Swift, Stuart & Meunier, 1976b). Thecate, dinoflagellate stages appear as swarmers in some species (Meunier & Swift, 1977; Swift & Durbin, 1971). Buoyancy reversals in the cyst form occur when negatively-buoyant nutrient-depleted stages descending to the nutricline are resupplied with NO₃⁻ and become positively buoyant, consistent with acquiring NO₃⁻ at depth (Ballek & Swift, 1986). Non-motile stage cells take up NO₃⁻ and NH₄⁺ at almost equal rates in the light and dark (Bhovichitra & Swift, 1977) and field-collected cells at the surface contain up to 8 mM INP (Villareal & Lipschultz, 1995). Growth rates in culture range up to 0.2 div day^-1 (Bhovichitra & Swift, 1977), with doubling times of 4–14 (P. fusiformis) and 10–22 days (P. noctiluca) in field populations (Swift, Stuart & Meunier, 1976a). Abundance is reported up to 200 cells m^-3 in the Atlantic Ocean (Rivkin et al., 1984; Swift, Stuart & Meunier, 1976a) and 40–50 cells m^-3 in the Pacific Ocean (Sukhanova, 1973; Sukhanova & Rudyakov, 1973). Photosynthetic and light acclimation curves from field populations showed a time-averaging of the light field such that C fixation at the surface supported a near-constant doubling rate throughout the euphotic zone (Rivkin et al., 1984).

Halosphaera is a genus of positively buoyant non-motile phycomate prasinophytes noted throughout the oceans (poles to tropics) from the earliest days of oceanography (Agassiz, 1906; Schmitz, 1878; Sverdrup, Johnson & Fleming, 1942). It is listed as a member of the shade flora (Sournia, 1982). Reproduction occurs by swarmer formation with up to 50,000 flagellated swarmers released from a phycoma (Parke & den Hartog-Adams, 1965). Individual swarmers can vegetatively reproduce, and then after 14–21 days start to increase in size at 5–10 µm d^-1 to reach a species-specific diameter of several hundred microns. At this time, the cytoplasm undergoes numerous divisions to form flagellated swarmers (Parke & den Hartog-Adams, 1965). Size and photosynthetic rates (3–6 ng C cell^-1 h^-1) are similar to Pyrocystis (Rivkin & Lessard, 1986). Growth rates are poorly known; reproduction is linked to lunar rhythms in the North Sea and adjacent waters. INP up to 100 mM have been documented (Villareal & Lipschultz, 1995), and deep populations with seasonal descent and ascent are noted in the Mediterranean Sea (Wiebe, Remsen & Vaccaro, 1974). Abundance ranges from ∼10^-3 cells m^-3 (Wiebe, Remsen & Vaccaro, 1974) to 340 cells L^-1 (Gran, 1933). Halosphaera is representative of a number of species that reproduce by phycoma and swarmer formation, including members of the genus Pterosperma. Typical concentrations reported for the Mediterranean are 1–3 L^-1; Pterosperma is reported at ∼40 cells L^-1 in HNLC areas of the Pacific Ocean (Gomez et al., 2007). In the text calculation on N-transport, we have assumed an abundance of 200 cells m^-3 (0.2 cells L^-1) as a conservative mid-range value of the 9 order of magnitude abundance range for this group.

Ethmodiscus spp. are solitary centric diatoms and are the largest diatoms known with a diameter of >2,000 µm in the Pacific Ocean; cells are somewhat smaller in the Atlantic Ocean (Swift, 1973; Villareal & Carpenter, 1994; Villareal et al., 1999a). Internal nitrate concentrations from surface-collected samples reached 27 mM in the Sargasso Sea (Villareal & Carpenter, 1994). Cells become increasing negatively buoyant as INP are depleted with positively buoyant cells having significantly higher internal nitrate concentrations than sinking cells (Villareal & Lipschultz, 1995). Nitrate reductase activity, C doubling and Si uptake rates can support doubling times of 45–75 h in large Pacific cells (Villareal et al., 1999a); cell cycle analysis suggests division rates of 0.24–0.42 div d^-1 in smaller Atlantic cells (Lin & Carpenter, 1995). Pooled cells had δ¹⁵N values of 2.56–5.09 per mil (Villareal et al., 1999a). Maximum reported abundance is 27.3 cells m^-3 in equatorial waters off Chile (Belyayeva, 1972), but abundance generally ranges from 0.03–4.7 cells m^-3 in the Atlantic and 0.02–6 cells m^-3 in the central Pacific gyre (Belyayeva, 1968; Belyayeva, 1970; Villareal et al., 2007). In the Pacific, abundance increases westward into the open Pacific Ocean with the highest values near the equator (Belyayeva, 1970). Ascent rates reach 4.9 m h^-1 (Moore & Villareal, 1996b) and like Pyrocystis and Rhizosolenia, appears to result from active ionic regulation of inorganic (Woods & Villareal, 2008) and organic compounds required for osmoregulation (Boyd & Gradmann, 2002). Living cells have been collected in downward facing sediment traps at 5400 m (Villareal, 1993) indicating living cells with positive buoyancy at great depth.

Several of the Rhizosolenia species that are found in mats also exist as free-living diatom chains. These species exhibit similar characteristics to mat-forming spp. INP are present at up to 26 mM (Moore & Villareal, 1996a). Individual species (non-aggregated) ascend at up to 6.9 m h^-1, depending on species and are also listed as members of the shade flora (Sournia, 1982). Growth rates for buoyant species range from 0.37 to 0.78 div d^-1 in the laboratory and up to 1.0 div d^-1 in the field (Moore & Villareal, 1996a; Yoder, Ackleson & Balch, 1994). Other characteristics are similar to Rhizosolenia mats (Moore & Villareal, 1996a). Little abundance information is available. R. castracanei is reported at up to 10³ cells L^-1 from the Bay of Naples (Marino & Modigh, 1981) and 50 cells m^-3 in Sargasso Sea warm core rings (TA Villareal and TJ Smayda, unpublished data). R. debyana reached 10⁶ cells L^-1 in the Gulf of California in bloom conditions (Garate-Lizarraga, Siqueiros-Beltrones & Maldonado-Lopez, 2003); similar abundance was likely in the equatorial Pacific “Line in the Sea” front accumulation (Yoder, Ackleson & Balch, 1994).

Significance of migrating phytoplankton to the North Pacific N budget

In this final step of the calculation, we incorporated these additional migrating taxa into the model. In order to compare the spatially limited input of a mesoscale eddy with the broader distribution patterns of phytoplankton, we combined conservative abundance data and growth rate estimates for Halosphaera, Ethmodiscus, Pyrocystis and solitary Rhizosolenia spp. (Table 1) and calculated their combined contributions to NO₃⁻flux to be 62.5 µmol N m^-2 d^-1. These estimates are very generalized and can only be used to scale the fluxes. Using profiler-derived estimates of eddy NO₃⁻ injection from the Pacific Ocean (Johnson, Riser & Karl, 2010), we considered the nitrate input via eddy injection over the entire time frame of measurement (145 mmol NO₃⁻ m^-2 over 600 days), and computed an average daily eddy injection rate of 242 µmol NO₃⁻ m^-2 d^-1. Nitrate transport of Rhizosolenia mats (2002/2003 data; 172 µmol NO₃⁻ m^-2 d^-1) combined with other taxa (values from Table 1) equals 235 µmol N m^-2 d^-1. This nearly equals the average daily eddy injection of nitrate (242 µmol NO₃⁻ m^-2 d^-1). Our previous VPR estimates of mats (Villareal et al., 1999b) is lower, and reduces the upward transport to 179 µmol N m^-2 d^-1 if we include those abundance estimates. However, within the uncertainties of both calculations, this is remarkably good agreement. On a timescale of weeks to months, migrating phytoplankton can transport sufficient N from deep euphotic zone pools to the upper euphotic zone to significantly impact nutrient budgets. Upward biological transport of nitrate is quantitatively important to the biogeochemistry of surface waters in the N. Pacific gyre. Other mechanisms may exist, but migration alone appears to be sufficient to dominate the required upward transport.

The abundance range found in the vertically migrating flora is not trivial; Halosphaera abundance records span 9 orders of magnitude and the abundance used profoundly affects the calculations. While Halosphaera may be extreme, it highlights the difficulties in enumerating a frequently rare and largely ignored component of the marine phytoplankton. Moreover, there are significant gaps in our knowledge of the biology of these taxa, their life cycles and migration timing that create uncertainties in how to apply this information.

Acquisition of imported N by other phytoplankton requires release of internal nitrate pools or remineralization by grazers. Rhizosolenia mats directly release NO₃⁻. Using excretion rates (Singler & Villareal, 2005) for NO₃⁻ (2 cruise range: 22.8–23.7 nmol N µg chl^-1 h^-1) and published N:Chl ratios (1.7 µmol N: µg chl a) (Villareal et al., 1996), we calculate N-specific release of ∼1.3% h^-1 or up to 31% d^-1. Release rates vary with Fe status, buoyancy status and location along the E-W gradient (Singler & Villareal, 2005); however, it is clear that over time scales of days to weeks, Rhizosolenia mats (and by inference, other high nitrate cells) will release NO₃⁻. Grazers on this size class are poorly known. Hyperiid amphipods are associated with mats, as well as parasitic dinoflagellates and ciliates (Caron et al., 1982; Villareal & Carpenter, 1989; Villareal et al., 1996). Nitrate is probably released during feeding by myctophids as well (Robison, 1984). Such release by both Rhizosolenia and other ascending, high NO₃⁻ cells provides the needed mechanism for transporting NO₃⁻ to the required depths for net community production (Johnson, Riser & Karl, 2010), balancing isotope budgets (Altabet, 1989), and providing sources for the observed difference in the δ¹⁵N of NO₃⁻ in small pro- and eukaryotes (Fawcett et al., 2011). Energy dissipation via reduction of oxidized N and subsequent release also provides additional pathways to the environment from nitrate-using cells (Lomas, Rumbley & Glibert, 2000) although the isotope systematics of the various products would need to be carefully considered for their contribution to the resultant particulate signal. However, this N would contribute to meeting N budget requirements.

Using flow-cytometer sorted populations coupled with mass spectrometry, Fawcett et al. (2011) noted uniformly low δ¹⁵N values in prokaryotic phytoplankton from the Sargasso Sea while small (<30 µm) eukaryotes showed a higher mean value with significant variability suggestive of nitrate utilization from beneath the euphotic zone. Our results suggest that some of this nitrate may be made available by migrating phytoplankton. However, this requires differential availability of nitrate to the small pro- and eukaryote populations. In general, picoplankton can respond to low (<100 nM) nitrate additions (Glover, Garside & Trees, 2007). However, the response of specific taxa in this size class may not be uniform. Synechococcus, in most cases, appears to be able to utilize both NO₃⁻ and NH₄⁺ while Prochlorococcus initially appeared unable to utilize nitrate (Moore et al., 2002), although recent work has noted evidence growth on NO₃⁻ (Casey et al., 2007). The general patterns suggest NH₄⁺ utilization supports Procholorococcus while Synechococcus appears to have retained the capacity to utilize both NO₃⁻and NO₂⁻ in the presence of NH₄⁺ (Bird & Wyman, 2003; Wyman & Bird, 2007). Exceptions exist in both groups (Fuller et al., 2003) and considerable genomic diversity in N uptake exists in the two groups (Scanlan et al., 2009). Temporal changes in populations can be expected as they adapt to seasonal patterns of nutrient availability in the water column (Bragg et al., 2010). While this example focuses on prokaryotes due to the greater body of work available on oceanic forms, the relevant point is that NO₃⁻ injections above the nutricline can reasonably be expected to be differentially assimilated by subpopulations within the phytoplankton, and that, at times, significant components of the prokaryotic phytoplankton may not have access to NO₃⁻.