4.1.1 Surface and thermocline waters (< 26.9 kg⁻¹ m⁻³; < 800 m)

A high salinity surface layer (> 35.5 PSU) centred at ∼25.5 kg⁻¹ m⁻³ (Fig. 5a) is described in several studies. It has been termed “southern subtropical surface water” by Muromtsev (1959), “subtropical surface water” by Wyrtki (1973) and “subtropical subsurface water” (SSW) by Schott and McCreary (2001). For further descriptions, we adopt the definition of Wyrtki (1973) and use the abbreviation SSW. The SSW is formed in the subtropical gyre of the Southern Hemisphere due to an excess of evaporation over precipitation (Schott and McCreary, 2001) at latitudes between 25 and 35^∘ S (Baumgartner and Reichel, 1975). It is subducted into the thermocline of the subtropical gyre (Schott and McCreary, 2001), is detectable as far north as 15.08^∘ S at CTD 05 (Fig. 3b) and is not discernable further north in the southern equatorial Indian Ocean (Figs. 5c, 6).

Less saline surface water (< 35.1 PSU) occurs above the density level of the salinity maximum (> 23 ^∘C; < 24.0 kg⁻¹ m⁻³; Fig. 5b) and is described by Wyrtki (1971) and Warren (1981). These low salinity values reflect an excess of precipitation over evaporation at latitudes between 0 and 10^∘ S (Baumgartner and Reichel, 1975) accompanied by the influx of low salinity water (34.0–34.5 PSU) from the Pacific Ocean through the Indonesian Archipelago, called “Indonesian Throughflow” (ITF). The ITF carries less saline water westwards via the SEC within the entire thermocline (Wyrtki, 1971; You and Tomczak, 1993). Emery (2001) named this less saline surface water (34.4–35.0 PSU) “Indonesian Upper Water” (IUW; Fig. 6).

Table 1Average nitrate and phosphate concentrations and average δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ and δ¹⁸O–${\mathrm{NO}}_{\mathrm{3}}^{-}$ values within water masses of the southern Indian Ocean, defined along their potential density surfaces (sigma-theta) and separated into four latitudinal sections (27.78–26.05^∘ S; 23.91–20.36^∘ S; 15.08^∘ S; and 8.81–2.98^∘ S). Water mass abbreviations are as follows: Indonesian Upper Water (IUW), Subtropical Surface Water (SSW), Subantarctic Mode Water (SAMW), Indian Equatorial Water (IEW), Antarctic Intermediate Water (AAIW), Indonesian Intermediate Water (IIW), Red Sea–Persian Gulf Intermediate Water (RSPGIW), Indian Deep Water (IDW), and Circumpolar Deep Water (CDW).

^a δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ in per mille (‰) vs. air and ^b δ¹⁸O–${\mathrm{NO}}_{\mathrm{3}}^{-}$ in per mille (‰) vs. VSMOW; D.A. refers to “deviation from the average value”; N/D represents “not detectable due to insufficient nitrate concentrations”.

Download Print Version | Download XLSX

The oxygen maximum south of 20^∘ S in a density range of 26.4–26.9 kg⁻¹ m⁻³(250–750 m; Fig. 5d) corresponds to the “Subantarctic Mode Water” (SAMW; Figs  5, 6), described by McCartney (1977). It is formed at latitudes between 40 and 50^∘ S and injects oxygen saturated waters at a temperature range of 6–14 ^∘C into the subtropical gyre. The SAMW in the southern Indian Ocean can be separated into three modes by slightly different density distributions (Herraiz-Borreguero and Rintoul, 2011) which originate in different ocean regions. For example, a lighter mode of the SAMW is formed in the western Indian basins and is limited to the southwest portion of the subtropical gyre, whereas the denser mode is found south off Australia and is carried further north by the outer portion of the subtropical gyre, ventilating a larger fraction of the gyre interior (Herraiz-Borreguero and Rintoul, 2011). However, for our purposes we assume the SAMW to be one homogenous water mass flowing above the density surface of 26.9 kg⁻¹ m⁻³. On its transition to the north, the oxygen concentrations rapidly decrease from > 4.6 mL L⁻¹ (CTD 05; Fig. 5e) to < 1.9 mL L⁻¹ (CTD 01, 03; Fig. 5f) due to respiration and the absence of effective ventilation in the northern Indian Ocean. The reduced vertical changes in salinity north of ∼15^∘ S mark the “Indian Equatorial Water” (IEW; Fig. 6). This is described by Sharma (1976), Warren (1981), Quadfasel and Schott (1982), You and Tomczak (1993) and Schott and McCreary (2001) as a mixture of thermocline water masses from the northern and southern Indian Ocean.

4.1.2 Intermediate water masses (26.9–27.4 kg⁻¹ m⁻³; 800–1000 m)

The salinity minimum (< 34.6 PSU) south of 15^∘ S, in a density range of 26.9–27.4 kg⁻¹ m⁻³(core density σ= 27.2 kg⁻¹ m⁻³; Fig. 5a) is associated with the “Antarctic Intermediate Water” (AAIW; Fig.4; Bindoff and McDougall, 2000; Deacon, 1933; Fine, 1993; Schott and McCreary, 2001; Toole and Warren, 1993; Warren, 1981; Wyrtki, 1973; You, 1998). The AAIW is transported eastwards by the “Antarctic Circumpolar Current” (ACC), penetrates into all three oceans and extends towards the Equator to feed the intermediate waters (Fine, 1993; McCartney, 1977; Piola and Gordon, 1989; Reid, 1986, 1989; Sverdrup et al., 1942; Talley, 1996; Wüst, 1935).

Figure 5Water mass properties represented as salinity vs. sigma-theta diagrams (a, b, c) and as oxygen vs. sigma-theta diagrams (d, e, f) for CTD stations at latitudes between 20.36 and 27.78^∘ S, 15.08^∘ S, and 2.98 and 8.81^∘ S. For water mass abbreviations see Table 1.

Download

The salinity minimum (< 34.6 PSU) observed at station CTD 05 (15.08^∘ S; Fig. 5b) has a slightly divergent core density (27.0 kg⁻¹ m⁻³) compared with the AAIW (Fig. 5a). This implies a further source for the salinity minimum of the AAIW. A low-salinity water mass (∼34.8 PSU) flows along 10–15^∘ S (Schott and McCreary, 2001; Wyrtki, 1971; You and Tomczak, 1993) and originates from the ITF. At intermediate depths it has been called “Indonesian Intermediate Water” (IIW; Fig. 6) by Emery and Meincke (1986) and Emery (2001).

The increase in salinity (> 34.9 PSU; Fig. 5c) further north, at the same density level as the AAIW, is caused by the inflow of saline water from the Arabian Sea, mainly from the Red Sea outflow (Warren, 1981) and is additionally fed by the outflow of the Persian Gulf (Emery and Meincke, 1986). Therefore, this water mass is called “Red Sea–Persian Gulf Intermediate Water” (RSPGIW; Fig. 6). The RSPGIW is transported towards the Equator and beyond to as far south as 10^∘ S (You, 1998), recirculates in the tropical gyre, and creates the absolute oxygen minimum (< 1.1 mL L⁻¹) caused by biogeochemical processes in the Arabian Sea (see Sect. 4.2.1).

4.1.3 Deep water masses (> 27.4 kg⁻¹ m⁻³; > 1000 m)

Overall, low oxygen concentrations in the northern study area underneath the AAIW (> 27.4 kg m⁻³; Fig. 5f) are caused by in situ consumption (Wyrtki, 1962) and reduced ventilation in the northern Indian Ocean. The deep oxygen minimum extends towards the south (∼3.0 mL L⁻¹) and is associated with the water mass of the “Indian Deep Water” (IDW). The IDW has higher salinities than the overlying AAIW (Bindoff and McDougall, 2000; Mantyla and Reid, 1995; Schott and McCreary, 2001; Talley, 2013) with values of > 34.6 PSU below the density range of the AAIW (Fig. 5a). The IDW ($\mathit{\sigma }=\sim \mathrm{27.5}$ kg⁻¹ m⁻³) flows in the density range just above the “Circumpolar Deep Water” (CDW; Fig. 6) and a further increase in salinity (34.62–34.73 PSU) and in the oxygen concentration at the 2 ^∘C temperature level (Emery, 2001) marks the transition between the IDW and the underlying CDW.

Figure 6Water mass distribution model from 30^∘ S to the Equator. The CTD stations taken into account for this transect ranged between 67.07 and 77.16^∘ E. Dotted lines represent N–S current directions and circled crosses indicate latitudinal directions from E to W. For water mass abbreviations see Table 1.

Download

4.2.1 Nutrient supply in the oligotrophic subtropical gyre and lateral transfer across the gyre boundaries

Intense downwelling in the centre of the IOSG is induced by the convergence of horizontal Ekman volume flux (Williams and Follows, 2003) and creates the thick layer of nutrient-depleted surface waters within the IUW and SSW (Fig. 7a, b, Table 1), and also within the underlying SAMW. The northward increase in nutrients at ∼15^∘ S (CTD 05, 2017) marks the northern boundary of the subtropical gyre and the maximum extension of the IUW, SSW and SAMW (Table 1). Further increase in nutrient concentrations within the IEW indicates the transition from the subtropical gyre towards the southern equatorial Indian Ocean identified by the characteristic shoaling of the nutricline at the northern fringe of the gyre (Table 1, Fig. 7a, b). The IEW is not a well-defined water mass, but rather a mixture of thermocline waters from the southern Indian Ocean and from the nutrient-enriched northern Indian Ocean. Therefore, just below the upper 100 m nutrient concentrations increase up to ∼23 µmol kg⁻¹ nitrate and ∼1.6 µmol kg⁻¹ phosphate (Table 1) at the northernmost stations (CTD 01, 03; 2017) and indicate the increasing influence of the nutrient-enriched northern Indian Ocean (Gaye et al., 2013). This increased northern influence is also reflected by the ${\mathrm{NO}}_{\mathrm{3}}^{-}/{\mathrm{PO}}_{\mathrm{4}}^{\mathrm{3}-}$ ratios, which exhibit values of less than 8 in the upper 200 m of the subtropical gyre, but increase towards the southern equatorial Indian Ocean, tracking the outcropping nutriclines (Fig. 7c). Low ${\mathrm{NO}}_{\mathrm{3}}^{-}/{\mathrm{PO}}_{\mathrm{4}}^{\mathrm{3}-}$ ratios are typical in surface waters of oligotrophic regions because nitrate commonly becomes depleted prior to phosphate (Sarmiento and Gruber, 2006; Deutsch et al., 2007). Due to the intense downwelling in the centre of the IOSG, the supply of nutrients by vertical mixing is reduced or absent in the gyre (Williams and Follows, 1998). Thus, lateral transfer across the gyre boundaries and biological N₂ fixation are major processes supplying nutrients to the euphotic zone of the subtropical gyre.

Figure 7Latitudinal profiles from south to north with an overlay of the water mass distribution model (white contour lines) in the southern Indian Ocean for nitrate (a) and phosphate (b) concentrations, the ${\mathrm{NO}}_{\mathrm{3}}^{-}/{\mathrm{PO}}_{\mathrm{4}}^{\mathrm{3}-}$ ratio (c), N^* (d), δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ (e) and δ¹⁸O–${\mathrm{NO}}_{\mathrm{3}}^{-}$ of nitrate (f) and nitrate Δ(15–18) as the difference between δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ and δ¹⁸O–${\mathrm{NO}}_{\mathrm{3}}^{-}$. For water mass abbreviations see Fig. 6.

Download

The water masses entering the study area from the Southern Ocean and from the northern Indian Ocean have characteristic nutrient concentrations and isotope fingerprints of reactive nitrogen, so that some of the water masses are clearly discernible by the distribution of nutrients and the isotopic composition of nitrate within the IOSG. Our samples show ${\mathrm{NO}}_{\mathrm{3}}^{-}/{\mathrm{PO}}_{\mathrm{4}}^{\mathrm{3}-}$ ratios of 14.56 on average (Fig. 8). These ${\mathrm{NO}}_{\mathrm{3}}^{-}/{\mathrm{PO}}_{\mathrm{4}}^{\mathrm{3}-}$ ratios are lower than the global ocean mean of 16:1 (Redfield, 1934, 1963). Furthermore, measurements in the Arabian Sea reveal typical ${\mathrm{NO}}_{\mathrm{3}}^{-}/{\mathrm{PO}}_{\mathrm{4}}^{\mathrm{3}-}$ ratios of 12.81 (Codispoti et al., 2001), even lower than our detected ${\mathrm{NO}}_{\mathrm{3}}^{-}/{\mathrm{PO}}_{\mathrm{4}}^{\mathrm{3}-}$ ratios. Consequently, the average ${\mathrm{NO}}_{\mathrm{3}}^{-}/{\mathrm{PO}}_{\mathrm{4}}^{\mathrm{3}-}$ ratio of 14.56 falls between the global ocean mean of 16:1 (Redfield, 1934, 1963) and the typical ratio in the Arabian Sea of 12.81 (Codispoti et al., 2001). This alone indicates the mixing of water masses of southern and northern Indian Ocean origin.

The deviation from the Redfield stoichiometry (Redfield, 1934, 1963) is quantified by the tracer N*. The analytical error on N^* estimate based on the relative error for nitrate and phosphate analysis was below 1.5 % for duplicate sample measurements. The Arabian Sea is characterized by an extensive oxygen deficit zone (ODZ) that induces denitrification in mid-water depths (150–400 m) (Gaye et al., 2013) and leads to an N deficit and, in turn, to negative N^* values (e.g. Bange et al., 2005; Rixen et al., 2005; Gaye et al., 2013). Our data set reveals values of about −1 µmol kg⁻¹ within the RSPGIW and values lower than −4 µmol kg⁻¹ within the IDW (Fig. 7d), which coincide with the oxygen minimum (see Sect. 4.1). Consequently, negative N^* values are a result of the influx of water masses from the Arabian Sea, which are affected by denitrification. To strengthen this assumption and to compensate for the limitations of the previously mentioned N^* approach, we use stable isotope measurements.

Figure 8Correlation of nitrate vs. phosphate concentrations. Regression line of the sample pool (solid, grey line) indicates a ratio of ∼14.56 (r²=0.99), intermediate between the Redfield ratio of 16:1 (black, dashes line) and the mean ratio in the Arabian Sea with a slope of ∼12.81 (grey, dotted-dashed line) following Codispoti et al. (2001). Colour coding of dots indicates the potential density sigma-theta in kilograms per cubic metre (kg m⁻³).

Download

We observe elevated δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ and δ¹⁸O–${\mathrm{NO}}_{\mathrm{3}}^{-}$ values of > 7 and > 3 ‰ within the RSPGIW at CTD 01 and 03 (Table 1, Fig. 7e, f) accompanied by nitrate concentrations of > 30 µmol kg⁻¹ (Fig. 7a). Denitrification discriminates against the heavier isotope of nitrate (¹⁵N, ¹⁸O) and raises δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ as well as δ¹⁸O–${\mathrm{NO}}_{\mathrm{3}}^{-}$. In the Arabian Sea average δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ and δ¹⁸O–${\mathrm{NO}}_{\mathrm{3}}^{-}$ values of > 20 and > 15 ‰, respectively, are observed at mid-water depths (150–400 m), within the oxygen minimum zone (Gaye et al., 2013; Martin and Casciotti, 2017). The significant progressive reduction of δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ and δ¹⁸O–${\mathrm{NO}}_{\mathrm{3}}^{-}$ towards the southern Indian Ocean is a result of mixing with subtropical thermocline water masses and remineralization/N-assimilation processes along the flow path.

Within the RSPGIW and the lower IEW, we observe a deviation from the O-to-N isotope effect of ${}^{\mathrm{18}}\mathit{\epsilon }{:}^{\mathrm{15}}\mathit{\epsilon }\sim \mathrm{1}$, which is typical for consumption processes like denitrification (Granger et al., 2004; Rafter et al., 2013; Sigman et al., 2003, 2005). The difference between N and O isotopes (δ¹⁵N, δ¹⁸O) can be quantified by the tracer Δ(15–18) that indicates values of > 4 ‰ within the RSPGIW and the lower IEW (Fig. 7g). The RSPGIW injects nitrate that is remineralized from ¹⁵N-enriched organic matter, and originates in a region of strong denitrification. During remineralization of organic matter, the N isotope effect associated with ammonium production and nitrification does not affect the δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ but depends on the biomass being remineralized (Rafter et al., 2013). In contrast, the δ¹⁸O of newly nitrified nitrate is independent of the isotopic composition of the organic matter. However, δ¹⁸O–${\mathrm{NO}}_{\mathrm{3}}^{-}$ depends on the isotope effect during ${\mathrm{NH}}_{\mathrm{4}}^{+}$ and ${\mathrm{NO}}_{\mathrm{2}}^{-}$ oxidation, water incorporation (δ¹⁸O–H₂O of ∼0 ‰), and the exchange of oxygen atoms with water that should generate a δ¹⁸O of newly produced ${\mathrm{NO}}_{\mathrm{3}}^{-}$ between −8 ‰ and −1 ‰ (Buchwald and Casciotti, 2010; Casciotti et al., 2010). Therefore, the RSPGIW adds nitrate that is enhanced in δ¹⁵N compared with the ambient water and has a relative lower δ¹⁸O, which consequently drives the decoupling of N and O isotopes. Furthermore, the source nitrate for N assimilation in the lower IEW is this regenerated nitrate and also results in the decoupling of N and O isotopes in this depth range. Consequently, the elevated Δ(15–18) can be explained by a remineralisation/N-assimilation cycle and by the lateral influx of ¹⁵N-enriched nitrate induced by strong denitrification in the oxygen minimum zone of the Arabian Sea. Furthermore, alteration processes within the oxygen minimum zone have the possibility to modify the Δ(15–18). However, it is still unclear how this signal is preserved along the flow path, and we suggest that the dominant mechanism that elevates Δ(15–18) values is due to the influx of ¹⁵N-enriched nitrate due to denitrification.

In the IOSG, we observe elevated δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ and δ¹⁸O–${\mathrm{NO}}_{\mathrm{3}}^{-}$ values of > 7 ‰ and > 4 ‰ (Table 1, Fig. 7e, f), respectively, within the SAMW (400–500 m), which originates in the subantarctic thermocline of the Southern Ocean. In general, N assimilation has an isotopic effect of about 5 ‰–10 ‰ (Montoya and McCarthy, 1995; Sigman et al., 2005; Waser et al., 1998) and produces biomass that is relatively depleted in ¹⁵N and ¹⁸O in comparison with the nitrate source. Consequently, this drives the elevation in δ¹⁵N and δ¹⁸O of the remaining nitrate as uptake proceeds. However, in oligotrophic waters, such as in the IOSG, this isotopic effect is not observable (Montoya et al., 2002), as nitrate is typically drawn down to the limit of detection by complete N assimilation. Nitrate in surface waters of the Southern Ocean is only partially assimilated due to light limitation and lower iron availability (Boyd et al., 2000; DiFiore et al., 2006, 2010; Hutchins et al., 2001; Sigman et al., 1999) and leads to δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ values of up to ∼13 ‰ (DiFiore et al., 2006; Sigman et al., 1999, 2000). Seasonal mixing and remineralization processes result in δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ values of 5 ‰–9 ‰ within subantarctic thermocline waters (McCartney, 1977; Sigman et al., 1999). On its flow path towards the north, this isotope trace of incomplete assimilation causes the elevated isotope values within the SAMW, which enters the subtropical Indian Ocean thermocline with δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ values of > 7 ‰ (Table 1, Figs. 4c, 7e, 9a).

Because subantarctic thermocline waters are the source water of the SAMW and the underlying AAIW in the IOSG, we compare the nitrate isotope properties of subantarctic thermocline waters with our results. Sigman et al. (1999, 2000) and DiFiore et al. (2006) use the correlation of δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ and the fraction of nitrate remaining – ln(${\mathrm{NO}}_{\mathrm{3}}^{-}$) – to quantify the isotope fractionation effect during N assimilation in the Antarctic and subantarctic region. If N assimilation occurs with a constant effect and no new nitrate is added to the surface ocean, then the uptake process can be described in terms of Rayleigh fractionation kinetics (Mariotti et al., 1981). To fulfil the conditions of Rayleigh fractionation, the nitrate samples plot along a straight line in δ¹⁵N∕ln(NO₃) space, where the slope of the line represents the isotope effect of N uptake or mixing of different nitrate pools. Sigman et al. (1999, 2000) and DiFiore et al. (2006) compare the theoretical Rayleigh utilization trend of δ¹⁵N∕ln(NO₃) =5 ‰ with their measured nitrate utilization trend within the subantarctic thermocline, where Sigman et al. (1999, 2000) determined a slope of δ¹⁵N∕ln(NO₃) ∼1.3 ‰ (Fig. 9b). Our results reveal a similar, but even shallower slope of δ¹⁵N∕ln(NO₃)∼0.93 ‰ (Fig. 9b). It is clearly a mixing signal that causes the moderate slopes of δ¹⁵N∕ln(NO₃) in both the gyre region and in the subantarctic, as biological utilization of nitrate is unlikely at this depth range. The explanation for the slightly shallower slope in our data set compared with the results in the subantarctic is vertical mixing with the overlying SSW that has lower δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ values (< 6 ‰) with nitrate concentrations of < 5 µmol kg⁻¹. This process does not fulfil the conditions of Rayleigh fractionation due to fundamentally different formation background (Table 1, Fig. 9a; see Sect. 4.2.2). Deep-water nitrate concentrations vary little and δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ increases towards shallower water depths; in addition, the nitrate signal also slightly differs from that of the Southern Ocean due to the influence of the IDW, which originates in the northern Indian Ocean.

Figure 9Nitrate concentrations vs. δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ (a) and δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ vs. ln(${\mathrm{NO}}_{\mathrm{3}}^{-}$) (b) for CTD stations within the IOSG (20.36–27.78^∘ S). The colour code of dots in Fig. 9a indicates the potential density sigma-theta (kg m⁻³). In Fig. 9b, data are grouped for the Subtropical Surface Water (blue), SAMW (green), AAIW (yellow) and deep water masses (red).

Download

Within the isotope maxima of the SAMW (∼500 m), the uniform evolution of N and O isotopes of nitrate breaks down and reveals an offset of about 100 m between the N and O isotopic maxima at 23.91–20.36^∘ S (Fig. 4c, d; see Supplement Tables S2 and S3). Both, δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ and δ¹⁸O–${\mathrm{NO}}_{\mathrm{3}}^{-}$, are elevated within the SAMW, but δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ is less elevated than δ¹⁸O–${\mathrm{NO}}_{\mathrm{3}}^{-}$. The decoupling of N and O isotopes leads to low Δ(15–18) values of < 3 ‰ (Fig. 7g) within the SAMW, which originates in the partially assimilated subantarctic region. Isotope fractionation during the initial phase of partial N assimilation leads to sinking organic matter that is more depleted in ¹⁵N than the source nitrate (Sigman et al., 1999; Rafter et al., 2013). The influx of the SAMW into the subtropical gyre injects ¹⁵N-depleted organic matter and its remineralization adds nitrate that lowers the δ¹⁵N of the ambient nitrate, leading to low Δ(15–18) in subtropical thermocline waters (Fig. 7g). To conclude, remineralization of organic matter produced by partial N assimilation in the Southern Ocean is reflected in lower Δ(15–18) and leads to the deviation of the isotope maxima with an offset of 100 m within the depth range of the SAMW. However, the remineralization of ¹⁵N-depleted organic matter formed out of newly fixed nitrogen from N₂ fixation in surface waters may also influence the decrease of Δ(15–18) (see the following Sect. 4.2.2).

4.2.2 Evidence for N₂ fixation in the IOSG

The mixing of source water signals from the lateral influx of the neighbouring northern Indian Ocean and Southern Ocean significantly affects the nutrient distribution and isotopic composition of nitrate in the gyre region. However, N^* and δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ suggest that N₂ fixation introduces new nitrate into the surface waters of the IOSG. The increase in N^* up to 2 µmol kg⁻¹ at 60–200 m indicates a positive deviation of the ${\mathrm{NO}}_{\mathrm{3}}^{-}/{\mathrm{PO}}_{\mathrm{4}}^{\mathrm{3}-}$ ratio from the Redfield stoichiometry (Gruber and Sarmiento, 1997; Redfield, 1934, 1963) and is evidence for the input of newly fixed N into the surface water. Studies in the southwestern Pacific Ocean also indicate positive N^* anomalies of ∼2 µmol kg⁻¹ in the productive layer, reflecting diazotrophic N₂ fixation (Fumenia et al., 2018). Although N₂ fixation is the first-order driver of positive N^* (Bourbonnais et al., 2009; Monteiro and Follows, 2006), other processes, such as atmospheric deposition or the preferential remineralization of N over P, may also be responsible for excess N and result in an overestimation of N*-based N₂ fixation (Bourbonnais et al., 2009; Monteiro and Follows, 2006). However, the southern Indian Ocean is less affected by the influx of nutrient-enriched mineral aerosols from atmospheric deposition (Duce at al., 2008; Duce and Tindale, 1991) and we neglect this factor in our further discussion. Furthermore, we use δ¹⁵N and δ¹⁸O of nitrate, as well as Δ(15–18), as additional indicators of N sources to overcome the weakness associated with the N^* approach.

Diazotrophic N₂ fixation produces organic matter that has a low δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ relative to average oceanic combined nitrogen (Carpenter et al., 1997; Minagawa and Wada, 1986; Montoya et al., 2002; Wada and Hattori, 1976). Within the upper 200 m of the IOSG (20.36–23.91^∘ S), the δ¹⁵N of nitrate varies between 4.3 ‰ and 5.9 ‰ with a mean of ∼5.0 ‰ (Figs. 7e, 9a, b). These values are higher compared with other regions of intense N₂ fixation, such as in the subtropical northeastern Atlantic, where values of 2 ‰–5 ‰ occur in surface waters (Bourbonnais et al., 2009). At first sight, the high values of surface waters in the IOSG do not speak for significant N₂ fixation in surface waters. However, considering that SAMW is the source of nitrate with elevated δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ values of 7.4 ‰ on average (20.36–23.91^∘ S) and peak values of 7.9 ‰ (Fig. 9a, b, see also Table S2), the δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ in surface waters (4.3 ‰–5.9 ‰) is lowered by ∼2.4 ‰. This is similar to the decrease of ∼3 ‰ in surface waters of the north Atlantic, where N₂ fixation is verified (Bourbonnais et al., 2009; Knapp et al., 2008). Therefore, N₂ fixation is most likely the main driver of the upward decrease of δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ in surface waters and adds isotopically light nitrogen from the atmosphere into the gyre region. To prove this, we take the nitrate δ¹⁸O into account, which exhibits values of < 3 ‰ and shows a minor decrease compared to δ¹⁵N. The resulting decoupling of N and O isotopes of nitrate leads to smaller differences between δ¹⁵N and δ¹⁸O and reveals a Δ(15–18) of < 3 ‰ (Fig. 7g). To conclude, positive N^* values, the distinct upward decrease of δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$, reduced Δ(15–18) and the distance from any external nitrate sources are unambiguous signs of diazotrophic activity within the surface layer.

Figure 10${\mathrm{NO}}_{\mathrm{3}}^{-}/{\mathrm{PO}}_{\mathrm{4}}^{\mathrm{3}-}$ ratio vs. nitrate concentrations of seawater samples at 20.36–23.91^∘ S and 26.05–27.78^∘ S (a). The solid grey line indicates the calculated N assimilation (${\mathrm{NO}}_{\mathrm{3}}^{-}$_cal vs. ${\mathrm{NO}}_{\mathrm{3}}^{-}$ ∕ ${\mathrm{PO}}_{\mathrm{4}}^{\mathrm{3}-}$_cal) with a preformed ${\mathrm{NO}}_{\mathrm{3}}^{-}$ ∕ ${\mathrm{PO}}_{\mathrm{4}}^{\mathrm{3}-}$_in ratio of 13.25 for the region of the IOSG and 14.25 for the southern equatorial Indian Ocean in addition to progressive nutrient assimilation with a Redfield ratio of 16 (${\mathrm{NO}}_{\mathrm{3}}^{-}$ ∕ ${\mathrm{PO}}_{\mathrm{4}}^{\mathrm{3}-}$_ass). In (b) we present the portion of nitrate formed out of newly fixed N (${\mathrm{NO}}_{\mathrm{3}}^{-}$_new) vs. depth at latitudes between 20.36 and 23.91^∘ S.

Download

To estimate the supply of newly fixed N to the nitrate pool within the IOSG, we can calculate the fraction of nitrate coming from atmospheric N₂ fixation and the fraction that is added from the underlying source water using the observed δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ within the upper 200 m and the following equation modified after Bourbonnais et al. (2009, 2013):

where “δ¹⁵N_surface” is the detected δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$, which shows an average of ∼5.0 ‰ in the upper 200 m (range 4.3 ‰–5.9 ‰) at latitudes between 20.36 and 23.91^∘ S; the “δ¹⁵N_fix” is the isotope value of atmospheric N₂, assumed to be about 0 ‰; the factor “a” is the percentage of nitrate coming from atmospheric N₂ fixation; the δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ of the source water, which is the SAMW, with values about 7.4 ‰, is represented by “δ¹⁵N_source”; and “b” is the corresponding percentage.

In this equation we neglect the point of isotope fractionation via N assimilation due to the fact that in oligotrophic environments a complete N assimilation takes place (Montoya et al., 2002). Thus, no net expression of the isotope fractionation occurs. Consequently, the organic matter produced, which is again remineralized, also has a similar isotope signal to the assimilated nitrate. This assimilated nitrate is the mixing product of newly fixed nitrogen from the atmosphere and the input from the underlying SAMW, which is expressed by Eq. (1). We resolve the equation to a ($=\mathrm{1}-b$) and b ($={{\mathit{\delta }}^{\mathrm{15}}\mathrm{N}}_{\mathrm{surface}}/{{\mathit{\delta }}^{\mathrm{15}}\mathrm{N}}_{\mathrm{source}}$); hence, we calculate that about 32 % of the assimilated nitrate is provided from newly fixed nitrogen by N₂ fixation.

To prove this first approach, indicating that about 32 % of the assimilated nitrate results from atmospheric N input by N₂ fixation, we can calculate the ${\mathrm{NO}}_{\mathrm{3}}^{-}/{\mathrm{PO}}_{\mathrm{4}}^{\mathrm{3}-}$ ratio resulting from nutrient assimilation without any external N input by N₂ fixation and compare this with our measured ${\mathrm{NO}}_{\mathrm{3}}^{-}/{\mathrm{PO}}_{\mathrm{4}}^{\mathrm{3}-}$ ratio within surface waters. To provide an estimate of the excess nitrogen in surface waters supplied by the remineralization of cyanobacterial biomass, we use the following equations:

where “${\mathrm{NO}}_{\mathrm{3}}^{-}$_in” is the initial nitrate concentration of the source water (SAMW within the IOSG), “${\mathrm{NO}}_{\mathrm{3}}^{-}$_ass” denotes the assimilated nitrate (${\mathrm{NO}}_{\mathrm{3}}^{-}$_in-${\mathrm{NO}}_{\mathrm{3}}^{-}$_sample) and “${\mathrm{NO}}_{\mathrm{3}}^{-}$_sample” represents the sample concentrations. The initial nitrate to phosphate pool ${\mathrm{NO}}_{\mathrm{3}}^{-}/{\mathrm{PO}}_{\mathrm{4}}^{\mathrm{3}-}$_in is defined as the mean ratio of the source water. For the region of the IOSG, we presume that the mean ratio within the SAMW is 13.25. For the general N assimilation in the euphotic zone we assume Redfield conditions of ${\mathrm{NO}}_{\mathrm{3}}^{-}/{\mathrm{PO}}_{\mathrm{4}}^{\mathrm{3}-}$_ass=16. To calculate the residual nitrate, we multiply the calculated nitrate to phosphate ratio (${\mathrm{NO}}_{\mathrm{3}}^{-}/{\mathrm{PO}}_{\mathrm{4}}^{\mathrm{3}-}$_cal) by the measured phosphate concentrations:

The difference of ${\mathrm{NO}}_{\mathrm{3}}^{-}$_sample and ${\mathrm{NO}}_{\mathrm{3}}^{-}$_cal represents the portion of the nitrate supplied by nitrification out of newly fixed N. At latitudes between 20.36 and 23.91^∘ S, our samples indicate elevated ${\mathrm{NO}}_{\mathrm{3}}^{-}/{\mathrm{PO}}_{\mathrm{4}}^{\mathrm{3}-}$ ratios and a resulting positive deviation from the calculated line of N assimilation at nitrate concentrations of < 10 µmol kg⁻¹ (Fig. 10a), which indicates an external input of N into the surface waters of the IOSG. We presume that N₂ fixation leads to the local elevation in the ${\mathrm{NO}}_{\mathrm{3}}^{-}/{\mathrm{PO}}_{\mathrm{4}}^{\mathrm{3}-}$ ratios due to the input of new N and coincides with the decrease of δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ and the decoupling of N and O isotopes, leading to low Δ(15–18). The quantity of newly fixed nitrate (${\mathrm{NO}}_{\mathrm{3}}^{-}$_new) is given by the formula:

which is presented in Fig. 10b. This indicates a distinct upward increase in the upper 200 m at 20.36–23.91^∘ S with an average portion of fixed nitrate of about 34 %. Consequently, our first approach that suggests that 32 % of the nitrate measured in the upper 200 m is derived from newly fixed nitrogen agrees quite well with the 34 % calculated using ${\mathrm{NO}}_{\mathrm{3}}^{-}/{\mathrm{PO}}_{\mathrm{4}}^{\mathrm{3}-}$ ratios. Bourbonnais et al. (2009) stated that N₂ fixation accounts for ∼40 % of newly supplied nitrate in the subtropical North Atlantic. This is slightly higher than our assumption for the subtropical southern Indian Ocean. However, in the subtropical North Atlantic higher N^* values (3.5 µmol kg⁻¹), a slightly stronger upward decrease of δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ and a stronger decoupling of N and O isotopes are observed in surface waters, suggesting higher fixation rates.

Further south at 26.05–27.78^∘ S, samples plot close to the line of N assimilation and no significant input of fixed nitrate is indicated (Fig. 10a). This agrees with the δ¹⁵N–${\mathrm{NO}}_{\mathrm{3}}^{-}$ values in surface waters, which demonstrate an abrupt increase at about 26^∘ S to similar values as in the underlying SAMW (> 7 ‰), whereas δ¹⁸O–${\mathrm{NO}}_{\mathrm{3}}^{-}$ still shows low values of < 3 ‰. Consequently, N and O isotopes reveal a counteracting behaviour that differs from the region at 20.36–23.91^∘ S, resulting in a high Δ(15–18) of > 4.5 ‰ (Fig. 7g). This is a strong indication of the absence of N₂ fixation in this region but also leads to the assumption that δ¹⁸O remains low due to ongoing nitrate production by nitrification. This sudden termination of N₂ fixation may be due to a temperature-limiting factor, mentioned by Capone et al. (1997) and Breitbarth et al. (2007). They argue that N₂ fixation by Trichodesmium, which is the dominant nitrogen-fixing cyanobacteria in subtropical oligotrophic waters (Berman-Frank et al., 2001; Pearl et al., 1994), decreases dramatically at seawater temperatures below 22 ^∘C (Berman-Frank et al., 2001). Thus, modelling N₂ fixation with this assumption resulted in very low fixation rates south of about 25^∘ S in the Indian Ocean (Paulsen et al., 2017). However, N₂ fixation by other diazotrophs (e.g. unicellular diazotrophic cyanobacteria) has been shown to occur at higher latitudes than Trichodesmium (Moisander et al., 2010). Another reason for the decline of N₂ fixation south of 26^∘ S may be the limited availability of iron and other micronutrients. Atmospheric iron deposition is low in the Southern Hemisphere oceans, and iron availability gradually decreases towards high southern latitudes (Boyd et al., 2000; Duce and Tindale, 1991; Duce et al., 2008). Reduced iron availability is suggested to limit the growth of nitrogen-fixing organisms in regions where iron availability is already restricted (Sanudo-Wilhelmy et al., 2001). Berman-Frank et al. (2001) calculated the potential of nitrogen fixation by Trichodesmium and suggested that iron availability limits nitrogen fixation in 75 % of the global ocean. However, until now no concrete studies on iron or other micronutrient availability and N₂ fixation have been conducted within the southern Indian Ocean.