- Split View
-
Views
-
Cite
Cite
Ana Fernández, Emilio Marañón, Antonio Bode, Large-scale meridional and zonal variability in the nitrogen isotopic composition of plankton in the Atlantic Ocean, Journal of Plankton Research, Volume 36, Issue 4, July/August 2014, Pages 1060–1073, https://doi.org/10.1093/plankt/fbu041
- Share Icon Share
The zonal (∼15°W–40°W along 26°N–29°N) and meridional (∼30°N–30°S along 28°W–29°W) variability of δ15N of suspended particles and zooplankton (>40 µm) was studied to assess the influence of nitrogen fixation in the isotopic budget of the tropical and subtropical Atlantic ocean. Two cruises were conducted in October–November 2007 and April–May 2008 comprising a zonal and meridional transect each. In the region between 30°N and 15°N, the concurrently measured nitrogen fixation was insufficient to explain the consistent patch of suspended particles with δ15N < 2‰ and points to a significant contribution of atmospheric deposition of light nitrogen to the isotopic budget. The equatorial region (15°N–10°S) is subject to intense nitrogen fixation, which, according to a two-end-member mixing model, may explain 40–60% of the observed δ15N in suspended particles and 3–30% in zooplankton. In the South region between 10°S and 30°S, low values (<4‰) were measured in suspended particles and zooplankton during 2008. The values of δ15N of suspended particles suggest that nitrogen fixation, which is usually low (<10 µmol N m−2 day−1), may represent 50–60% of phytoplankton nitrogen in this region. Hence, diazotrophy in the South Atlantic may be more important than previously thought.
INTRODUCTION
In many marine ecosystems, primary production is limited by the availability of nitrogen (Vitousek and Howarth, 1991; Karl et al., 2002; Moore et al., 2013). Reactive nitrogen is supplied to the euphotic zone by different physical, chemical and biological processes such as advective diffusion, atmospheric deposition and biological nitrogen fixation. The latter is mediated by organisms and, in the oligotrophic regions of the oceans, is a relevant source of new nitrogen (Paerl and Zehr, 2000). The ratio of stable isotopes in phytoplankton (15N:14N expressed as δ15N in ‰) is variable, due to the contrasting preferences of the organisms for each isotope. The metabolic pathways usually discriminate against the heavy isotope (15N), this discrimination is measured by the isotopic fractionation factor (Montoya, 2008). Additionally, the different forms of inorganic nitrogen have distinct signatures of δ15N. Deep-nitrate typically ranges between 3 and 6‰ (Montoya, 2008), atmospheric dinitrogen is, by definition, 0‰, and deep-ammonium lies between 6 and 8‰ (Miyake and Wada, 1967). Hence, a very different δ15N of organic matter is expected, according to the source of nitrogen, if this is completely consumed. The isotopic signature of phytoplankton will depend then on the signature of the source of nitrogen and the degree of fractionation during uptake. Yet, the interpretation of δ15N is not so straightforward. In the case of animals (i.e. upper trophic levels), a trophic effect is also observed whereby the tissues of the consumer are usually 2–4‰ heavier than the food, whereas the animal's excreta, mainly in the form of ammonium, can be 2–4‰ lighter than the food (Montoya, 2008; and references herein). In addition, cultured cyanobacteria growing on excess nitrate showed a strong fractionation factor, yielding δ15N values similar to those produced by growth on dinitrogen (Bauersachs et al., 2009).
In the Atlantic Ocean, experimental data retrieved during large-scale surveys show that Trichodesmium, the most well-studied diazotroph, is distributed preferentially between 0°N and 20°N (Tyrrell et al., 2003; Moore et al, 2009; Fernández et al., 2010; Luo et al., 2012). In addition, nitrogen fixation, mostly measured with the method of Montoya et al. (Montoya et al., 1996), is more significant between 0°N and 15°N (Moore et al., 2009; Fernández et al., 2010; Luo et al., 2012). The δ15N of diazotrophs usually ranges between −1 and −2‰ (Montoya et al., 2002). However, the measured isotopic signature of nitrogen in suspended particles and the biogeochemical estimates of excess nitrogen available in the literature suggest that nitrogen fixation is more relevant in a region further north, between 15°N and 30°N (Gruber and Sarmiento, 1997; Mahaffey et al., 2003, 2004; Hansell et al., 2004; Reynolds et al., 2007). The time scales reflected by these measurements are different: in situ nitrogen fixation rates generally represent instantaneous rates over a few hours to 1 day while δ15N and excess nitrogen are indicators of the diazotrophic activity over longer periods of days to months. However, the determinants of this disagreement remain undefined. Duce et al. (Duce et al., 2008) argued that the atmospheric deposition of reactive nitrogen in the oceans has increased due to human activities and is fast approaching the marine N2 fixation budget. Other studies have also shown an increase of the atmospheric deposition of 15N-depleted nitrogen in high and temperate latitudes (Hastings et al., 2009; Mara et al., 2009; Morin et al., 2009; Holtgrieve et al., 2011), as a result of the increasing anthropogenic production of reactive nitrogen and/or natural speciation processes. In addition, Baker et al. (Baker et al., 2007) and Knapp et al. (Knapp et al., 2010) reported depositional fluxes of low δ15N similar to measured N2 fixation rates in the Atlantic Ocean.
As part of a wider project, we have previously described the latitudinal and longitudinal distribution of measured community nitrogen fixation in the tropical and subtropical Atlantic Ocean (Fernández et al., 2010, 2013) and the relative contribution of nitrogen fixation and nitrate eddy diffusion in supplying new nitrogen to the euphotic layer (Mouriño-Carballido et al., 2011). Here, we report on the distribution of δ15N in suspended particles and two size-fractions of plankton, with the aim of describing the large-scale latitudinal variability of nitrogen isotopic signatures in the Atlantic Ocean and comparing these inferred patterns of diazotrophy with concurrent, direct measurements of in situ N2 fixation rates.
METHOD
Sampling, hydrography and chlorophyll a
The vertical distribution of temperature, salinity, dissolved oxygen and fluorescence was measured by an SBE 911plus CTD attached to a rosette equipped with 12-L Niskin bottles that were fired to 300 m depth, always before dawn. The vertical profiles of fluorescence and oxygen at each station were used to choose the sampling depths for the determination of inorganic nutrients concentration, chlorophyll a (chl-a) concentration, community 15N2 fixation and natural abundance of nitrogen isotopes in suspended particles.
The concentration of chl-a was measured at six to seven depths distributed through the euphotic layer. At each depth, a 250-mL sample was filtered, using low vacuum pressure, through 0.2-μm pore-size polycarbonate filters. The pigments were extracted overnight in 90% of acetone at −4°C. Fluorescence was subsequently measured on board with a Turner Designs 700 fluorometer, which was calibrated with pure chl-a (Fluka).
Rates of N2 fixation by the whole planktonic community in a 24-h incubation period were determined in each station at the surface (5 m), an intermediate depth (30–80 m) and the depth of the deep chlorophyll maximum (DCM) and are already described in Fernández et al. (Fernández et al., 2010, 2013). Briefly, we incubated triplicate samples following the Montoya et al. (Montoya et al., 1996) protocol for the 15N2-uptake technique with the modifications of Rees et al. (Rees et al., 2009). The equations of Weiss (Weiss, 1970) and Montoya et al. (Montoya et al., 1996) were used to calculate the initial N2 concentration (assuming equilibrium with atmosphere) and N2 fixation rates, respectively. The limit of detection, estimated following Montoya et al. (Montoya et al., 1996), was 0.001 µmol N m−3 day−1.
Natural abundance of nitrogen isotopes in suspended particles
For the determination of δ15N signature in suspended particles (δ15Nsp), 2-L samples were taken at six depths through the euphotic layer at each pre-dawn station and filtered through a 25-mm diameter GF/F filter (Whatman). All filters were dried at 40°C for 24 h and then stored until pelletization in tin capsules. The measurement of particulate organic nitrogen (PON) and 15N atom% was carried out with an elemental analyser combined with a continuous flow stable isotope mass spectrometer (FlashEA112 + Deltaplus, ThermoFinnigan) and using an acetanilide standard as reference. The limit of detection of the equipment was 0.20 µg N.
The isotopic signature observed in the suspended particles may be affected by the presence of other types of material in addition to phytoplankton (i.e. bacteria, detritus, zooplankton). The existence of a relationship between the PON to chl-a ratio and the δ15N of suspended particles is an indicator of such a trophic effect (Waser et al., 2000). The Pearson product-moment correlation coefficient of PON : chl-a and δ15Nsp was calculated to test this possibility.
where δ15Ndiazotroph is the nitrogen isotopic composition of diazotrophs (15N : 14N, ‰) and is the nitrogen isotopic composition of deep-nitrate (15N : 14N, ‰). As indicated by these authors, this two-end-member mixing model is sensitive to the values of the end-members chosen (δ15Ndiazotroph and ). In order to represent only the nitrate in the upper thermocline, and avoid the effect of recently fixed nitrogen recycled between the upper water column and the thermocline in the calculations, the used was 4.5‰, which is the global average of deep-nitrate (Liu and Kaplan, 1989; Sigman et al., 1997). Due to the fact that most of our stations were oligotrophic, no additional fractionation factor during nitrate uptake was added. As a conservative choice representing the least contribution of nitrogen fixers, and considering the fact that little fractionation occurs during N2 fixation (Montoya, 2007), the δ15Ndiazotroph used was −2‰ (Montoya et al., 2002).
Natural abundance of nitrogen isotopes in plankton
At each pre-dawn station, zooplankton were collected by vertical tows of a 40-µm net, 30 cm in diameter, through the upper 200 m of the water column at a constant towing speed of 60 m min−1. The content of the collector was suspended in 500 mL of 20-µm filtered seawater. Two 60 mL sub-samples were preserved, one in Lugol's solution and the other in formaldehyde, for the determination of the abundance of Trichodesmium and other plankton by microscopic examination. Trichodesmium trichomes were more abundant in the fraction of 40–200 µm while colonies were present in the >200 µm fraction. The rest of the sample was separated into two size-fractions by passage through nylon sieves of 40 and 200 µm. Each fraction was then re-suspended in 200 mL of 20-µm filtered seawater and subsequently filtered on pre-weighted 45-mm diameter GF/F filters by low vacuum pressure. All filters were dried for 24 h at 40°C and stored until measurement of PON and 15N atom% as previously described.
where δ15Nplankton stands for the nitrogen isotopic composition of the plankton size-fraction (15N : 14N, ‰), and δ15Nreference pl is the δ15N of reference zooplankton. Again, a conservative value of −2‰ was used for δ15Ndiazotroph. The δ15N of the reference plankton was calculated as the mean of the δ15N40 or δ15N200 measured in the stations where the lowest abundance of Trichodesmium and nitrogen fixation was found, i.e. the stations between 0°S and 20°S on the latitudinal leg of 2007 cruise, where no influence of nitrogen fixation in the samples is expected. The values used were: 4.6‰ for the δ15Nreference 40 and 5.9‰ for δ15Nreference 200. This model is based on the use of reference plankton to account for the trophic effect, i.e. the reference plankton serves as a proxy in both terms of the calculation; therefore, no additional fractionation term for the trophic effect was needed. The assumptions that are implied are: (i) the size distribution of grazers in the sample and the reference plankton are similar, (ii) the trophic fractionation in the sample and the reference are similar and (iii) in both locations the isotopic composition of the nitrate supporting the food web is the same (J. P. Montoya, personal communication).
RESULTS
Hydrography and fluorescence
The hydrographic settings found in these regions were similar on both legs. Surface waters in the equatorial region were always warmer (>24°C) and less saline (<35 psu) than in the gyres in both seasons. In turn, the stability of the water column in the gyres was weaker than that found in the equatorial region, where the average Brunt–Väisäla frequency in the upper 125 m was higher (Fernández et al., 2010). The fluorescence profiles showed a well-defined DCM associated with the thermocline on both transects (Fig. 2e and f). This DCM was shallower and better defined in the equatorial region than in the gyres. In contrast, on the longitudinal sections, waters were warmer and slightly more saline in autumn 2007 than in spring 2008 (Fig. 2e), leading to a stronger stability of the water column, as indicated by the higher Brunt–Väisäla frequency measured on this cruise (Fernández et al., 2013). The DCM was located at ∼100 m on both zonal legs, and no apparent trend in depth was observed (Fig. 2e and f).
Stable nitrogen isotopes in suspended particles (δ15Nsp) and PON
The correlations of δ15N of suspended particles with ammonium concentration and with nitrate concentration are shown in Table I. Considering all the stations on each cruise, δ15Nsp correlated with ammonium concentration in 2007 (P < 0.05, n = 128) and with nitrate in 2008 (P < 0.05, n = 119).
. | 2007 cruise . | 2008 cruise . | ||||||
---|---|---|---|---|---|---|---|---|
All stations . | North gyrea . | Equatorial region . | South gyre . | All stations . | North gyrea . | Equatorial region . | South gyre . | |
NH4 | 0.22* | 0.55* | 0.43** | n.s. | n.s. | n.s. | 0.47** | −0.38* |
[128] | [54] | [38] | [42] | [36] | ||||
NO3 | n.s. | 0.30* | n.s. | n.s. | 0.21* | n.s. | n.s. | 0.51** |
[54] | [120] | [36] |
. | 2007 cruise . | 2008 cruise . | ||||||
---|---|---|---|---|---|---|---|---|
All stations . | North gyrea . | Equatorial region . | South gyre . | All stations . | North gyrea . | Equatorial region . | South gyre . | |
NH4 | 0.22* | 0.55* | 0.43** | n.s. | n.s. | n.s. | 0.47** | −0.38* |
[128] | [54] | [38] | [42] | [36] | ||||
NO3 | n.s. | 0.30* | n.s. | n.s. | 0.21* | n.s. | n.s. | 0.51** |
[54] | [120] | [36] |
Numbers in brackets represent the total number of samples used for the analysis.
n.s., no significance.
aIncludes the zonal and meridional legs in the North gyre.
*P < 0.05, **P < 0.01.
. | 2007 cruise . | 2008 cruise . | ||||||
---|---|---|---|---|---|---|---|---|
All stations . | North gyrea . | Equatorial region . | South gyre . | All stations . | North gyrea . | Equatorial region . | South gyre . | |
NH4 | 0.22* | 0.55* | 0.43** | n.s. | n.s. | n.s. | 0.47** | −0.38* |
[128] | [54] | [38] | [42] | [36] | ||||
NO3 | n.s. | 0.30* | n.s. | n.s. | 0.21* | n.s. | n.s. | 0.51** |
[54] | [120] | [36] |
. | 2007 cruise . | 2008 cruise . | ||||||
---|---|---|---|---|---|---|---|---|
All stations . | North gyrea . | Equatorial region . | South gyre . | All stations . | North gyrea . | Equatorial region . | South gyre . | |
NH4 | 0.22* | 0.55* | 0.43** | n.s. | n.s. | n.s. | 0.47** | −0.38* |
[128] | [54] | [38] | [42] | [36] | ||||
NO3 | n.s. | 0.30* | n.s. | n.s. | 0.21* | n.s. | n.s. | 0.51** |
[54] | [120] | [36] |
Numbers in brackets represent the total number of samples used for the analysis.
n.s., no significance.
aIncludes the zonal and meridional legs in the North gyre.
*P < 0.05, **P < 0.01.
Nitrogen isotopic signature in the euphotic layer
In the autumn 2007 meridional transect, the isotopic signature of suspended particles showed two minima (<−2‰) in the North gyre and South gyre regions. In the equatorial region, δ15Nsp oscillated ∼0‰ (Fig. 6). The δ15N40 and δ15N200 roughly followed these patterns. In contrast, the distributions were dome shaped in spring 2008, reaching peak values in the equatorial region. On both cruises, the gyres showed low δ15N values at most of the stations. Besides, a positive statistical correlation between δ15N in the three fractions suggests a regular impact of light nitrogen across trophic levels (Table II).
. | δ15Nsp . | δ15N40 . | ||
---|---|---|---|---|
2007 . | 2008 . | 2007 . | 2008 . | |
δ15N40 | 0.66** | 0.82** | − | − |
δ15N200 | 0.66** | 0.74** | 0.90** | 0.78** |
. | δ15Nsp . | δ15N40 . | ||
---|---|---|---|---|
2007 . | 2008 . | 2007 . | 2008 . | |
δ15N40 | 0.66** | 0.82** | − | − |
δ15N200 | 0.66** | 0.74** | 0.90** | 0.78** |
**P < 0.01, n = 17.
. | δ15Nsp . | δ15N40 . | ||
---|---|---|---|---|
2007 . | 2008 . | 2007 . | 2008 . | |
δ15N40 | 0.66** | 0.82** | − | − |
δ15N200 | 0.66** | 0.74** | 0.90** | 0.78** |
. | δ15Nsp . | δ15N40 . | ||
---|---|---|---|---|
2007 . | 2008 . | 2007 . | 2008 . | |
δ15N40 | 0.66** | 0.82** | − | − |
δ15N200 | 0.66** | 0.74** | 0.90** | 0.78** |
**P < 0.01, n = 17.
A two-way factorial ANOVA indicated significant differences between regions and cruises, and for the comparison of δ15N40 and δ15N200, the interaction region-cruise was also significant, and enhanced the observed difference (Table III). The differences between regions appeared to be significant only for the North gyre-equatorial region (post hoc Tukey's HSD test) as can be also seen in Fig. 6.
. | DF . | δ15Nsp . | δ15N40 . | δ15N200 . | |||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
SS . | MS . | F . | P . | SS . | MS . | F . | P . | SS . | MS . | F . | P . | ||
Region | 2 | 23.6 | 11.8 | 6.3 | 0.004 | 26.6 | 13.3 | 11.5 | 0.000 | 25.7 | 12.8 | 15.4 | 0.000 |
Cruise | 1 | 16.0 | 16.0 | 8.6 | 0.006 | 4.8 | 4.8 | 4.2 | 0.048 | 15.9 | 15.9 | 19.2 | 0.000 |
Region vs. cruise | 2 | 6.0 | 3.0 | 1.6 | 0.213 | 20.9 | 10.5 | 9.0 | 0.000 | 21.7 | 10.9 | 13.1 | 0.000 |
. | DF . | δ15Nsp . | δ15N40 . | δ15N200 . | |||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
SS . | MS . | F . | P . | SS . | MS . | F . | P . | SS . | MS . | F . | P . | ||
Region | 2 | 23.6 | 11.8 | 6.3 | 0.004 | 26.6 | 13.3 | 11.5 | 0.000 | 25.7 | 12.8 | 15.4 | 0.000 |
Cruise | 1 | 16.0 | 16.0 | 8.6 | 0.006 | 4.8 | 4.8 | 4.2 | 0.048 | 15.9 | 15.9 | 19.2 | 0.000 |
Region vs. cruise | 2 | 6.0 | 3.0 | 1.6 | 0.213 | 20.9 | 10.5 | 9.0 | 0.000 | 21.7 | 10.9 | 13.1 | 0.000 |
DF, degrees of freedom; SS, sums of squares; MS, mean of squares; F, F statistic; P, probability.
. | DF . | δ15Nsp . | δ15N40 . | δ15N200 . | |||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
SS . | MS . | F . | P . | SS . | MS . | F . | P . | SS . | MS . | F . | P . | ||
Region | 2 | 23.6 | 11.8 | 6.3 | 0.004 | 26.6 | 13.3 | 11.5 | 0.000 | 25.7 | 12.8 | 15.4 | 0.000 |
Cruise | 1 | 16.0 | 16.0 | 8.6 | 0.006 | 4.8 | 4.8 | 4.2 | 0.048 | 15.9 | 15.9 | 19.2 | 0.000 |
Region vs. cruise | 2 | 6.0 | 3.0 | 1.6 | 0.213 | 20.9 | 10.5 | 9.0 | 0.000 | 21.7 | 10.9 | 13.1 | 0.000 |
. | DF . | δ15Nsp . | δ15N40 . | δ15N200 . | |||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
SS . | MS . | F . | P . | SS . | MS . | F . | P . | SS . | MS . | F . | P . | ||
Region | 2 | 23.6 | 11.8 | 6.3 | 0.004 | 26.6 | 13.3 | 11.5 | 0.000 | 25.7 | 12.8 | 15.4 | 0.000 |
Cruise | 1 | 16.0 | 16.0 | 8.6 | 0.006 | 4.8 | 4.8 | 4.2 | 0.048 | 15.9 | 15.9 | 19.2 | 0.000 |
Region vs. cruise | 2 | 6.0 | 3.0 | 1.6 | 0.213 | 20.9 | 10.5 | 9.0 | 0.000 | 21.7 | 10.9 | 13.1 | 0.000 |
DF, degrees of freedom; SS, sums of squares; MS, mean of squares; F, F statistic; P, probability.
We tried to estimate if Trichodesmium could be the major influence on the patterns observed but no significant correlation (Pearson's r) was found between the measured filament abundance (Fernández et al., 2010, 2013) and the δ15N of suspended particles (P = n.s., n = 42), the 40–200 µm (P = n.s., n = 41) or the >200 µm plankton size-fractions (P = n.s., n = 42).
Nitrogen fixation rates (Fig. 6) were previously reported in Fernández et al. (Fernández et al., 2010, 2013). Briefly, on the longitudinal transects, no apparent trend was observed in 2007 while a clear increasing pattern to the East appeared in 2008 (Fig. 6b and d). On the spring 2008 zonal leg, the average vertically integrated N2 fixation was 7-fold higher than that in autumn 2007 (8.3 ± 3.3 vs 1.2 ± 0.5 µmol N m−2 day−1). On both meridional transects, the highest integrated rates (∼250 and 150 µmol N m−2 day−1 in 2007 and 2008, respectively) were measured at stations located within the equatorial region (Fig. 6a and c). Besides, the North gyre showed higher diazotrophic activities than the South gyre. However, while N2 fixation south of the Equator was almost undetectable during the 2007 cruise, substantial rates were measured in the Southern Hemisphere in 2008 (Fig. 6a and c).
Diazotroph nitrogen contribution to δ15N in the euphotic layer
The contribution of diazotrophs to the observed δ15N of suspended particles, 40–200 µm and >200 µm plankton size-fractions, estimated by the two-end-member mixing models, decreased to the South on the 2007 cruise (Table IV). In 2008, the minimum was observed in the equatorial region (Table IV). The importance of this contribution is higher in 2007, except in the South gyre, where the contribution of diazotroph nitrogen was higher in all size-fractions. On both cruises, diazotrophy explains, on average, 61% of the observed δ15Nsp, 27% of δ15N40 and 30% of δ15N200.
Region . | Suspended particles . | 40–200 µm zooplankton . | >200 µm zooplankton . | |||
---|---|---|---|---|---|---|
2007 . | 2008 . | 2007 . | 2008 . | 2007 . | 2008 . | |
Longitudinal transect | 81 (29) | 59 (3) | 52 (14) | 16 (9) | 43 (4) | 31 (6) |
North gyre region | 85 (14) | 61 (10) | 48 (22) | 21 (11) | 41 (12) | 36 (8) |
Equatorial region | 62 (27) | 39 (8) | 25 (18) | 3 (2) | 11 (3) | 29 (12) |
South gyre region | 49 (12) | 58 (18) | 15 (20) | 36 (16) | 13 (15) | 40 (17) |
Region . | Suspended particles . | 40–200 µm zooplankton . | >200 µm zooplankton . | |||
---|---|---|---|---|---|---|
2007 . | 2008 . | 2007 . | 2008 . | 2007 . | 2008 . | |
Longitudinal transect | 81 (29) | 59 (3) | 52 (14) | 16 (9) | 43 (4) | 31 (6) |
North gyre region | 85 (14) | 61 (10) | 48 (22) | 21 (11) | 41 (12) | 36 (8) |
Equatorial region | 62 (27) | 39 (8) | 25 (18) | 3 (2) | 11 (3) | 29 (12) |
South gyre region | 49 (12) | 58 (18) | 15 (20) | 36 (16) | 13 (15) | 40 (17) |
The reference zooplankton used in each fraction corresponded to the average of the stations sampled in the South gyre during 2007 where Trichodesmium abundance was <1 trichome L−1, δ15N40 = 4.6‰ and δ15N200 = 5.9‰. The values of % of diazotroph N >100 and <0 were discarded in the calculation of the regions average. Number of samples is indicated in parentheses.
Region . | Suspended particles . | 40–200 µm zooplankton . | >200 µm zooplankton . | |||
---|---|---|---|---|---|---|
2007 . | 2008 . | 2007 . | 2008 . | 2007 . | 2008 . | |
Longitudinal transect | 81 (29) | 59 (3) | 52 (14) | 16 (9) | 43 (4) | 31 (6) |
North gyre region | 85 (14) | 61 (10) | 48 (22) | 21 (11) | 41 (12) | 36 (8) |
Equatorial region | 62 (27) | 39 (8) | 25 (18) | 3 (2) | 11 (3) | 29 (12) |
South gyre region | 49 (12) | 58 (18) | 15 (20) | 36 (16) | 13 (15) | 40 (17) |
Region . | Suspended particles . | 40–200 µm zooplankton . | >200 µm zooplankton . | |||
---|---|---|---|---|---|---|
2007 . | 2008 . | 2007 . | 2008 . | 2007 . | 2008 . | |
Longitudinal transect | 81 (29) | 59 (3) | 52 (14) | 16 (9) | 43 (4) | 31 (6) |
North gyre region | 85 (14) | 61 (10) | 48 (22) | 21 (11) | 41 (12) | 36 (8) |
Equatorial region | 62 (27) | 39 (8) | 25 (18) | 3 (2) | 11 (3) | 29 (12) |
South gyre region | 49 (12) | 58 (18) | 15 (20) | 36 (16) | 13 (15) | 40 (17) |
The reference zooplankton used in each fraction corresponded to the average of the stations sampled in the South gyre during 2007 where Trichodesmium abundance was <1 trichome L−1, δ15N40 = 4.6‰ and δ15N200 = 5.9‰. The values of % of diazotroph N >100 and <0 were discarded in the calculation of the regions average. Number of samples is indicated in parentheses.
DISCUSSION
Our data contribute to the existing studies in the Atlantic Ocean (Waser et al., 2000; Mino et al., 2002; Montoya et al., 2002; Mahaffey et al., 2003, 2004; Reynolds et al., 2007; Landrum et al., 2011; Mompeán et al., 2013) providing basin-scale distribution of δ15N in suspended particles and two plankton size-fractions during two contrasting seasons. We found a consistent 15N-depleted signal (<4‰) in suspended particles (δ15Nsp) in the euphotic layer at most of the stations (Figs 5 and 6). This implies that nitrogen fixation and/or atmospheric deposition were supplying an important fraction of new nitrogen in most of the tropical and subtropical Atlantic Ocean during our cruises. The trends in the δ15N of the two plankton size-fractions closely matched that of suspended particles, indicating an impact of light nitrogen even in upper trophic levels, at least for some regions (Fig. 6, Table II).
On the meridional transect in 2008, the δ15N signal in the >200 µm size-fraction (δ15N200) was lower than that in the 40–200 µm size-fraction (δ15N40) at most of the stations (Fig. 6), contrary to the usually observed enrichment in 15N of upper trophic levels (Montoya, 2008). One possible reason is that the longer turnover times of mesozooplankton relative to phytoplankton and microplankton could result in the uncoupling of different size-fractions, producing this inversion of the expected increasing pattern (Landrum et al., 2011; Mompeán et al., 2013). Another possible explanation could be the presence of Trichodesmium colonies, which were large enough to be retained in this size-fraction, and would have lowered the isotopic signature of zooplankton.
Zonal and meridional variations in δ15N in the North gyre
In the North gyre, the difference in δ15Nsp between stations was higher in autumn 2007 than in spring 2008 (6 and 2‰, respectively). However, the vertical change in δ15Nsp at each station was small on both cruises (Fig. 5) and the measured values of δ15Nsp were always <2‰. Our data agree with previous reports in the subtropical North Atlantic, which show a range of variation between −2 and 4‰ in the signature of suspended particles (Montoya et al., 2002; Mahaffey et al., 2003; Reynolds et al., 2007; Landrum et al., 2011). These authors described a consistently depleted signal between 7°N and 32°N, which is also confirmed by geochemical tracers (Gruber and Sarmiento, 1997; Hansell et al., 2004) that point to a persistent excess nitrate relative to phosphate in this area, indicative of intense nitrogen fixation. Those light nitrogen signatures have been associated with a large impact of diazotrophic nitrogen in the isotopic budget of this area, discarding the influence of other sources of new light nitrogen because of their weak strength or unlikely occurrence (Reynolds et al., 2007; Landrum et al., 2011). However, later studies provided new insights that suggest a more important influence of alternative sources such as the atmospheric deposition of 15N-depleted nitrogen (Baker et al., 2007; Hastings et al., 2009; Mara et al., 2009; Morin et al., 2009; Knapp et al., 2010; Holtgrieve et al., 2011; Mouriño-Carballido et al., 2011).
The δ15N of deep-water nitrate typically ranges between 3 and 6‰ with a global average of 4.8‰ (Montoya, 2008). In the presence of excess nitrate, the isotopic fractionation, due to the incomplete exhaustion of the nitrate pool by phytoplankton, could result in values of δ15Nsp <3‰ (Montoya, 2008). A recent study showed that cyanobacteria, especially Trichodesmium, growing on nitrate could express a nitrogen isotopic signal similar to that of nitrogen fixation depending on the isotopic composition of the nitrogen source, the degree of fractionation and the species of cyanobacterium (Bauersachs et al., 2009). However, no excess dissolved inorganic nitrogen was found in surface waters in our zonal or meridional legs, where the concentration of nitrate in the euphotic layer was always <130 nM (Mouriño-Carballido et al., 2011; Fernández et al., 2010, 2013). We also recorded measurable but low abundances of Trichodesmium (<60 trichomes L−1) in the euphotic layer (Fernández et al., 2010, 2013), which is also an indication of potential diazotrophy in the area. But, no significant correlation appeared between Trichodesmium abundances and the δ15Nsp (Pearson's r). Hence, we would not expect that a strong isotopic fractionation associated with cyanobacteria or other phytoplankters was responsible for the observed δ15Nsp during our cruises.
The lack of data on atmospheric deposition of nitrogen during our study limits any direct comparison with the measured nitrogen fixation and the distribution of the δ15N signature, but we can attempt to use an indirect analysis instead. The two-end-member model proposed by Montoya et al. (Montoya et al., 2002) yields a contribution of N2 fixation to δ15Nsp in the range of 81–85% in 2007 and 59–61% in 2008 (Table IV), which is close to the previous estimation of 74% by Reynolds et al. (Reynolds et al., 2007). However, experimental measurements of community nitrogen fixation in this region indicate modest rates of diazotrophy (<60 µmol N m−2 day−1) throughout the year (Fig. 6; Moore et al., 2009; Benavides et al., 2011). During the spring 2008 cruise, Mouriño-Carballido et al. (Mouriño-Carballido et al., 2011) calculated the relative importance of nitrate eddy diffusion and measured rates of nitrogen fixation as sources of new nitrogen to the euphotic layer in the North gyre. They estimated that the average contribution of nitrogen fixation during this cruise was only 2% over daily timescales. We acknowledge that the comparison of these two fluxes is difficult as they represent different timescales, i.e. the δ15Nsp represents timescales of days to weeks, while the measured nitrogen fixation timescale is 1day. But the difference between the fluxes was 30-fold in 2008. This suggests that other sources than diazotrophy may be contributing to our observed δ15Nsp, and that the low values are not only a consequence of intense nitrogen fixation. The importance of the atmospheric deposition of low δ15N nitrogen, natural or anthropogenic, is increasing in high and temperate latitudes (Hastings et al., 2009; Mara et al., 2009; Morin et al., 2009; Holtgrieve et al., 2011). According to the model of Duce et al. (Duce et al., 2008), the atmospheric supply of anthropogenic reactive nitrogen in the central North Atlantic is usually higher in the latitudinal range between 5°N and 25°N. The δ15N of this anthropogenic N depends on its origin and is extremely variable (Fang et al., 2011). For instance the δ15N of fuel NOx produced by power plants ranges between 5 and 13‰ (Heaton, 1990; Kiga et al., 2000) but that of thermal NOx produced by vehicle exhausts ranges between −13 and −2‰ (Heaton, 1990). Additionally, in the Atlantic Ocean between 45°N and 45°S, Morin et al. (Morin et al., 2009) found a δ15N of atmospheric nitrate that ranged between −7 and −1.6‰ and mainly representative of natural sources. In the Mediterranean Sea (Crete), Mara et al. (Mara et al., 2009) described a consistent source of low δ15N nitrate throughout the year with a potential impact on the isotopic budget of intermediate and deep-waters, which could lead to an overestimation of N2 fixation if atmospheric nitrate is neglected. In spite of that, previous studies discarded the effect of this process in the analysis of the nitrogen isotopic budget, based on the assumption that the flux is small compared with nitrogen fixation (Landrum et al., 2011) or to the export flux of nitrogen out of the euphotic layer (Reynolds et al., 2007). In contrast, Baker et al. (Baker et al., 2007) and Knapp et al. (Knapp et al., 2010) measured atmospheric depositional fluxes of 15N-depleted nitrogen similar to those of N2 fixation in the North Atlantic Ocean. Considering this information, we suggest that the observed δ15Nsp in the North gyre region during our cruises could be the result of the supply of light nitrogen through both nitrogen fixation and atmospheric deposition. Therefore, discarding the effect of this atmospheric supply in the analysis of δ15Nsp would result in the overestimation of nitrogen fixation.
The signature of light nitrogen found in suspended matter spreads over the food web. First, the δ15N distributions of the two zooplankton size-fractions and the suspended particles are significantly correlated (Fig. 6, Table II). Secondly, the observed difference between plankton and particles at each station (2–4‰) agrees with previous studies reporting that zooplankton nitrogen is typically 3‰ heavier than phytoplankton (Minagawa and Wada, 1984) due to the enrichment of zooplankton tissues by the isotopic fractionation associated with metabolic and excretory processes (Montoya, 2008). Thirdly, both plankton size-fractions showed relatively low values of δ15N throughout the transects (<4‰). Again, the data provided by the two-end-member model proposed by Montoya et al. (Montoya et al., 2002) point out that the contribution of nitrogen fixation to upper trophic levels in 2007 represents 48–52% in the 40–200 µm size-fraction and 41–43% in the >200 µm size-fraction (Table IV). In 2008, it represented roughly 16–21% of the 40–200 µm fraction signal and the 31–36% of the >200 µm size-fraction (Table IV). However, we would expect an overestimation of this contribution due to the combined effect of atmospheric deposition of 15N-depleted nitrogen and nitrogen fixation in the isotopic budget of this region.
The δ15N measured in suspended particles and zooplankton suggests a consistent supply of light nitrogen in this region of the Atlantic Ocean throughout the year, which coincides with previous studies (Montoya et al., 2002; Mahaffey et al., 2003; Reynolds et al., 2007; Landrum et al., 2011). However, we did not measure the isotopic composition of the depositional fluxes, which could be compared with measured community nitrogen fixation on our cruises to determine the actual contribution of each flux (Baker et al., 2007; Knapp et al., 2010). Further studies, characterizing the strength, frequency and δ15N of the atmospheric sources of nitrogen relative to in situ measured nitrogen fixation, will help to unequivocally ascertain the relative importance of each process in determining the δ15N signatures in the North Atlantic.
Meridional variations in δ15N in the equatorial region
In the equatorial region (15°N–10°S), the meridional trends of the δ15N of suspended particles coincided with those previously described by Mahaffey et al. (Mahaffey et al., 2004). However, our absolute values are lower than theirs, in the range of −2 to 4‰, and closer to those measured by Reynolds et al. (Reynolds et al., 2007) in the water column and by Mino et al. (Mino et al., 2002) in surface waters. The upwelling in this region allows a persistent diffusion of deep-nitrate to surface waters, which was reflected in the increase of nitrate concentration during our cruises (Fernández et al., 2010, 2013; Mouriño-Carballido et al., 2011), and is likely to support a substantial fraction of primary production. Therefore, heavy deep-nitrate is probably determining part of the δ15N of suspended particles in the equatorial region.
In autumn 2007, the difference between the δ15N of 40–200 and >200 µm plankton size-fractions (2‰) suggests either a different time scale in the integration of the signal or a low efficiency in the transfer of nitrogen to upper trophic levels. The latter could be attributed to the loss of isotopically light ammonium through excretory processes, which has been proposed as a major source of light nitrogen in oligotrophic regions (Checkley and Miller, 1989; Montoya, 2008). However, the positive correlation between δ15N of suspended particles and ammonium concentration on our cruises (Table I) suggests that the increase in ammonium is increasing the δ15Nsp and may not be related to the excretion of plankton.
The cyanobacterium Trichodesmium exudes up to 50% of recent fixed N2 as dissolved organic nitrogen, which can be easily assimilated by other phytoplankters and/or bacteria (Glibert and Bronk, 1994). Furthermore, both nitrogen fixation (Fig. 6) and Trichodesmium abundances typically reach high values in this region (Tyrrell et al., 2003; Moore et al., 2009; Fernández et al., 2010). Thus, the supply of light ammonium linked to diazotrophs is probably determining an important fraction of the nitrogen isotopic budget in the equatorial region. The two-end-member mixing model (Montoya et al., 2002) yields an average contribution of this diazotroph nitrogen to δ15Nsp of 62 ± 27% in autumn 2007 and 39 ± 8% in spring 2008 (Table IV). On the other hand, Mouriño-Carballido et al. (Mouriño-Carballido et al., 2011) estimated that the daily contribution of N2 fixation to total (N2 fixation + vertical diffusion of nitrate) input of new nitrogen was 22% on the 2008 cruise. Again, these fluxes represent different time scales, but both suggest that nitrogen fixation accounts for a relevant fraction of the supply of nitrogen to the euphotic layer in this region, and is consistent with previous experimental measurements.
The diazotroph nitrogen was inefficiently transferred to upper trophic levels as it accounted for 25 ± 18% in the 40–200 µm size-fraction and 11 ± 3% in the >200 µm size-fraction in 2007 cruise and for 3 ± 2% in the 40–200 µm size-fraction and 29 ± 12% in the >200 µm size-fraction in 2008 cruise (Table IV). Trichodesmium, the dominant diazotroph in this region, is toxic to many species of zooplankton (Hawser et al., 1992) and only a few groups of copepods are known to graze it (O’Neil and Roman, 1994). Besides, these groups seem to excrete a major fraction of the ingested nitrogen (O’Neil et al., 1996, Wannicke et al., 2010). Thus, diazotroph nitrogen is preferentially transferred through dissolved pools when Trichodesmium dominates the community (Mulholland, 2007).
Meridional variations in δ15N in the South gyre region
The distribution of δ15N of suspended particles (δ15Nsp) and zooplankton in the South gyre region showed contrasting trends in 2007 and 2008 (Fig. 6). Even though the vertical distribution of δ15Nsp varied within a range of 6‰, data were <4‰ in all stations (Fig. 5) suggesting that a 15N-depleted source of nitrogen is contributing significantly to the signals. In autumn 2007, the general meridional pattern largely coincided with that described by Mino et al. (Mino et al., 2002) in surface waters, but not with those given by Mahaffey et al. (Mahaffey et al., 2004) and Reynolds et al. (Reynolds et al., 2007), who found a general increasing trend to the South with values of >2‰. The light patch of δ15Nsp in the range of −2 to 0‰, which was found between 20°S and 30°S in the 2007 cruise (Fig. 6), seems to be a persistent feature also described by Mino et al. (Mino et al., 2002) with values close to −1‰, and Reynolds et al. (Reynolds et al., 2007) with values close to 0‰. In contrast, this is the first time that a decreasing trend in δ15N such as the one depicted in spring 2008 has been described in this region.
The flux of atmospheric deposition of nutrients in the South Atlantic Ocean is extremely weak (Gao et al., 2001; Duce et al., 2008); thus, we may discard the effect of light atmospheric nitrogen in the isotopic budget. The small difference in the δ15N of both zooplankton size-fractions suggests a strong coupling between trophic levels, with low isotopic fractionation in the loss of nitrogen by excretion (Checkley and Miller, 1989). The uptake of dissolved organic nitrogen and their inorganic degradation products, originated by the nitrogen fixers and processed by microbes, may explain such coupling, as isotopic fractionation in microbial food webs is generally low (Rau et al., 1990). Mahaffey et al. (Mahaffey et al., 2004) suggested that the relatively important dissolved organic nitrogen pool of the South Atlantic could account for the high δ15N measured in their study. On the contrary, Knapp et al. (Knapp et al., 2011) found that a long lived and poorly reactive DON pool in other regions of the Atlantic and Pacific Oceans could be a source of light ammonium through deamination. However, we propose that the supply of diazotroph nitrogen is significantly determining the observed δ15Nsp. The few experimental measurements performed to date in the South Atlantic show that nitrogen fixation is persistent in this region with rates in the range of 2–50 µmol N m−2 day−1 (Moore et al., 2009; Fernández et al., 2010; Grosskopf et al., 2012), which are similar to those reported in the equatorial and North gyre region (Fig. 6). This suggests that nitrogen fixation could be responsible of the persistent light patch of δ15Nsp in the South gyre between 10°S and 30°S. The contribution of nitrogen fixation to the δ15N of suspended particles was 49 ± 12% in 2007 and 58 ± 18% in 2008, according to a two-end-member model based on nitrate and diazotrophy (Montoya et al., 2002). Besides, the daily contribution of nitrogen fixation to the total (N2 fixation + nitrate diffusive flux) input of nitrogen to the euphotic layer was 44% during strong stratification conditions in April 2008 (Mouriño-Carballido et al., 2011).
This diazotroph nitrogen was transferred to upper trophic levels with relatively high efficiency in 2008, and it represented 36 ± 16% of the 40–200 µm size-fraction and 40 ± 17% of the >200 µm size-fraction (Table IV). Hence, both experimental measurements and estimations seem to agree in that nitrogen fixation could be supporting an important fraction of primary production in the South gyre, despite the fact that the absolute rates of both processes are low.
CONCLUSIONS
A persistent and consistent signature of low δ15N of suspended particles (δ15Nsp) is found in the North gyre region (30°N–15°N) in both zonal and meridional transects, which is usually associated with a significant input of nitrogen fixed by diazotrophs (Mahaffey et al., 2003; Reynolds et al., 2007; Landrum et al., 2011). However, the experimental measurements of nitrogen fixation show modest rates in comparison with other regions of the Atlantic Ocean and do not seem to support this argument. The atmospheric deposition of light nitrogen, which has increased in recent years, is likely to complete the required supply that produces this depleted δ15N signal. However, few studies have addressed the depositional and diazotrophic fluxes together (Baker et al., 2007; Knapp et al., 2011) and further studies are needed to accurately define the strength, frequency and isotopic composition of the atmospheric depositional flux against the flux of nitrogen fixation in the North Atlantic. The equatorial region (15°N–10°S) is subject to relatively intense nitrogen fixation throughout the year (Moore et al., 2009; Fernández et al., 2010; Grosskopf et al., 2012) which may explain 40–60% of the observed δ15Nsp signal. However, this nitrogen of diazotrophic origin seems to be inefficiently transferred to upper trophic levels. In the South gyre, the low δ15Nsp and the daily estimated contribution of nitrogen fixation to the supply of new nitrogen (Mouriño-Carballido et al., 2011) suggest that diazotrophs can contribute up to half of the nitrogen in phytoplankton at different time scales (Fig. 4). Even though the measured nitrogen fixation rates are low (Moore et al., 2009; Fernández et al., 2010; Grosskopf et al., 2012), their impact in the nitrogen isotopic budget of this region may be large. Hence, a re-evaluation of the importance of diazotrophy in the South Atlantic Ocean is needed through new studies that should address the annual variability in nitrogen fixation rates as well as the distribution and relative importance of the different groups of diazotrophs.
FUNDING
A.F. was supported by grant PGIDIT05PXIC31201PN of the Xunta de Galicia. This is a contribution of the project TRYNITROP (Trichodesmium and N2 fixation in the tropical Atlantic Ocean) funded by the Spanish Ministry of Science and Technology through grants CTM2004-05174-C01 and CTM2004-05174-C02 to A.B. and E.M., respectively.
ACKNOWLEDGEMENTS
The authors appreciate the careful revision and thorough comments of two anonymous reviewers who helped to improve the quality of the manuscript. We thank N. Lluch and P. Chouciño for technical assistance. Stable isotopes were analysed at SXAI-Universidade da Coruña. We also thank the officers and crew of the BIO Hespérides and the staff of the Marine Technology Unit (UTM), for their support during the work at sea.
REFERENCES
Author notes
Corresponding editor: John Dolan