Isotopic Evidence for the Evolution of Subsurface Nitrate in the Western Equatorial Pacific
Abstract
Subsurface waters from both hemispheres converge in the Western Equatorial Pacific (WEP), some of which form the Equatorial Undercurrent (EUC) that influences equatorial Pacific productivity across the basin. Measurements of nitrogen (N) and oxygen (O) isotope ratios in nitrate (δ15NNO3 and δ18ONO3), the isotope ratios of dissolved inorganic carbon (δ13CDIC), and complementary biogeochemical tracers reveal that northern and southern WEP waters have distinct biogeochemical histories. Organic matter remineralization plays an important role in setting the nutrient characteristics on both sides of the WEP. However, remineralization in the northern WEP contributes a larger concentration of the nutrients, consistent with the older “age” of northern thermocline-depth and intermediate-depth waters. Remineralization introduces a relatively low δ15NNO3 to northern waters, suggesting the production of sinking organic matter by N2 fixation at the surface—consistent with the notion that N2 fixation is quantitatively important in the North Pacific. In contrast, remineralization contributes elevated δ15NNO3 to the southern WEP thermocline, which we hypothesize to derive from the vertical flux of high-δ15N material at the southern edge of the equatorial upwelling. This signal potentially masks any imprint of N2 fixation from South Pacific waters. The observations further suggest that the intrusion of high δ15NNO3 and δ18ONO3 waters from the eastern margins is more prominent in the northern than southern WEP. Together, these north-south differences enable the examination of the hemispheric inputs to the EUC, which appear to derive predominantly from southern hemisphere waters.
Key Points
- Remineralization adds 15N-enriched nitrate to the southern hemisphere thermocline at the southern edge of the equatorial Pacific upwelling
- Isotope tracers indicate the remineralization of 15N-depleted nitrate due to N2 fixation in the northern Pacific thermocline waters
- Mixing estimates suggest a predominantly southern hemispheric source to the Equatorial Undercurrent that supplies the equatorial Pacific upwelling
1 Introduction
The Western Equatorial Pacific (WEP) is a “water mass crossroad” (Fine et al., 1994), where northern and southern hemisphere waters delivered through western boundary currents are exported to the Indian Ocean via the Indonesian Throughflow (Godfrey, 1996; Sprintall et al., 2014) and to the wider equatorial Pacific through the Equatorial Undercurrent (EUC; Butt & Lindstrom, 1994; Johnson et al., 2002; Lindstrom et al., 1987; Figure 1). As such, the WEP and its boundary currents exert an important influence on lower latitude biogeochemistry, particularly because the EUC is the source of nutrients to the equatorial Pacific surface waters (Rafter & Charles, 2012; Rafter & Sigman, 2016; Sloyan et al., 2010; Toggweiler et al., 1991; Wyrtki, 1981), fueling nearly 10% of global primary production (Pennington et al., 2006).
(a) Map showing simplified circulation of surface (solid line) and subsurface (dashed line) waters across the Pacific after Talley (1993), Fine et al. (1994), Sokolov and Rintoul (2000), and Kawabe and Fujio (2010). Colors show nitrate concentrations at 150 m from the WOA13 climatology. Black squares indicate the two study areas north and south of the equator. (b) Detailed view of sampling sites (GeoB174xx) off Mindanao and Papua New Guinea (PNG), with blue colors indicating the northern stations and red colors representing the southern stations. The same color code is used throughout this study to highlight the respective hemispherical location of individual sites. (c) Illustration of major surface (solid lines) and subsurface (dashed lines) currents in the Western Equatorial Pacific (WEP) after Lindstrom et al. (1987), Bingham and Lukas (1994), Fine et al. (1994), and Cravatte et al. (2011). The low-latitude western boundary currents are the Mindanao Current (MC), the New Guinea Coastal Current (NGCC) and the East Australia Current (EAC) at the surface, and the Mindanao Undercurrent (MUC), the New Guinea Coastal Undercurrent (NGCUC) and the Great Barrier Reef Undercurrent (GBRUC) at the subsurface. Additional surface currents in the WEP include the North and South Equatorial Current (NEC, SEC), and the North and South Equatorial Countercurrent (NECC and SECC). Subsurface currents include the Equatorial Undercurrent (EUC), the New Britain Coastal Undercurrent (NBCU), the North Vanuatu Jet (NVJ), the North Caledonian Jet (NCJ), the St. Georges Undercurrent (SGU), the Solomon Island Coastal Undercurrent (SICU), the New Ireland Coastal Undercurrent (NICU), and the Northern and Southern Subsurface Countercurrent (NSCC, SSCC). HE and ME indicate the Halmahera and Mindanao Eddies. Additional abbreviations indicate New Britain (NB), New Ireland (NI), New Hanover (NH), and Solomon Islands (SI).
The water masses and currents that converge at the equator at the western boundary have been investigated in a number of studies (e.g., Bingham & Lukas, 1995; Fine et al., 1994; Kashino et al., 2005; Lindstrom et al., 1987; Nie et al., 2016; Toole et al., 1988; Tsuchiya, 1968). Nevertheless, the relative transports to the EUC from north and south of the equator remain subject to debate. Some studies suggest that both southern and northern hemisphere nutrients contribute substantively (Fine et al., 1987; Izumo et al., 2002; Liu & Huang, 1998), or partially (<20%; Grenier et al., 2011) to the EUC, while others argue that transports from the South Pacific dominate the EUC (Toggweiler et al., 1991; Tsuchiya et al., 1989). In this respect, Southern Ocean nutrients are hypothesized to dominate the resupply to the lower latitude Pacific via Subantarctic Mode Water (SAMW) and WEP boundary currents (Palter et al., 2010; Sarmiento et al., 2004; Toggweiler et al., 1991). The evolution of nutrients in SAMW from its origin in the Southern Ocean en route to the lower latitudes of the South Pacific has been investigated recently from stable isotopes of nitrate, revealing that nitrate in SAMW is substantially altered in transit from the Southern Ocean (Rafter et al., 2013). The nitrate isotope ratios in SAMW bear evidence of its partial consumption at the Southern Ocean surface prior to subduction at the Subantarctic Front (DiFiore et al., 2010; Rafter et al., 2013; Sigman et al., 1999, 2000) and lateral exchange with oxygen deficient zones at the eastern margins (Peters et al., 2017; Rafter et al., 2012, 2013; Yoshikawa et al., 2015). Importantly, nitrate isotope ratios indicate that a substantial amount of nitrate is added to SAMW en route from the Southern Ocean to the tropics from the remineralization of organic material delivered by biological pumping (Peters et al., 2017; Rafter et al., 2012, 2013; Sigman et al., 2009b). Thus, the nutrient content of SAMW-density waters in the tropics is influenced by surface processes in the Southern Ocean as well as by modifications along its flow path in the South Pacific. While SAMW-density waters are too deep (>400 m) to directly upwell to the surface, diapycnal mixing between those nutrient-rich waters and overlaying nitrate-free surface waters of the South Pacific gyre result in the resupply of lower latitude thermocline-depth nutrients (Palter et al., 2010; Rafter et al., 2012, 2013; Toggweiler et al., 1991).
The isotopic evolution of nitrate in intermediate water masses of the North Pacific, which may influence the composition of the EUC, is less well characterized. Measurements of nitrate isotope ratios at the eastern tropical North Pacific margin have provided insights into the N biogeochemistry of oxygen deficient zones (ODZs; Altabet et al., 1999; Brandes et al., 1998; Buchwald et al., 2015; Sigman et al., 2005; Voss et al., 2001). Meridional sections in the eastern tropical Pacific reveal that ODZ waters permeate westward at intermediate depths along the North Equatorial Current (NEC; Reid, 1965; Reid & Mantyla, 1978). Additionally, nitrate isotope profiles at station ALOHA in the North Pacific gyre bear evidence of the remineralization of N deriving from N2 fixation in the surface gyre (Casciotti et al., 2008; Sigman et al., 2009a).
The WEP also appears to be a hot spot for biological N2 fixation at the sea surface, as evidenced by measurements of total organic nitrogen (Hansell & Feely, 2000) and from nitrogen isotope ratios of dissolved nitrate (Kienast et al., 2008) and settling particles (Yoshikawa et al., 2005). Accordingly, 15N2 tracer incubations off the coast of Papua New Guinea (PNG) and in the Solomon Sea indicate exceptionally high N2 fixation rates in both areas, exceeding most estimates reported for oceanic waters (Bonnet et al., 2009, 2015). In contrast to the WEP, 15N2 tracer incubations in the eastern margin of the South Pacific have consistently evidenced very low N2 fixation rates (Knapp et al., 2016; Raimbault & Garcia, 2008), in spite of inverse model diagnoses suggesting that elevated rates of N2 fixation may occur at the eastern gyre boundary (Deutsch et al., 2007). The N isotope composition of nitrate correspondingly bears no clear indication of the remineralization of newly fixed N in the thermocline of the South Pacific subtropical gyre (Peters et al., 2017; Rafter et al., 2013), in contrast to its North Pacific equivalent (Casciotti et al., 2008; Sigman et al., 2009a).
(1a)
(1b)
(1c)Measurements of coupled δ15N and δ18O of nitrate (δ15NNO3 and δ18ONO3) are sensitive to important biogeochemical fluxes and provide information on the origin and history of water masses (Sigman et al., 2000), identifying processes that are not immediately evident from standard hydrographic measurements. The assimilation, denitrification, and production of nitrate (i.e., remineralization) each alter the isotopic composition in distinctive ways, thus revealing the relative contribution of different processes to the nitrate pool. The partial assimilation of nitrate at the sea surface results in a coincident enrichment of N versus O isotope ratios of the unconsumed pool with a characteristic ratio of 1 (Casciotti et al., 2002; Granger et al., 2004). Similarly, in oxygen deficient waters, dissimilatory nitrate consumption (denitrification) also imparts equivalent N versus O enrichment on residual nitrate (Granger et al., 2008; Sigman et al., 2005). In turn, the remineralization of organic material at the subsurface, namely, the ammonification of organic nitrogen followed by its nitrification, produces nitrate with a δ15NNO3 akin to that of the remineralized material, thus reflecting the δ15N of material exported from the sea surface. In this respect, the fixation of atmospheric N2 into reactive nitrogen and its subsequent remineralization generate nitrate with a relatively low δ15N (0‰ to −1‰), deriving from the δ15N of atmospheric N2. Concurrently, the δ18O of remineralized nitrate adopts a value near that of ambient water (Buchwald & Casciotti, 2010; Buchwald et al., 2012; Casciotti et al., 2002; Sigman et al., 2009a). Interpreted in the context of hydrography and nutrient tracers, the coupled δ15NNO3 and δ18ONO3 thus provide a means of tracking the biogeochemical evolution of the nitrate pool along its flow path.
Seawater δ13C in dissolved inorganic carbon (DIC; δ13CDIC) provides additional insight into biogeochemical processing. It is not only a tracer for past water mass ventilation and air-sea gas exchange, but also an indicator of biological activity (Kroopnick, 1985). In the euphotic zone, the preferential removal of 12C during photosynthesis leaves the residual DIC enriched in 13C, while producing organic matter with a low δ13C. Organic matter remineralization “returns” 12C to the ambient water, consequently decreasing the δ13CDIC.
We thus investigate isotopic features that offer insights into distinct biogeochemical influences on nitrate in association with the water masses that converge in the WEP. The isotopic composition of nitrate and DIC, along with standard hydrographic measurements provide a means to identify the fraction of preformed nitrate relative to that remineralized along the flow path, and to nitrate advected laterally from oxygen deficient zones in the eastern Pacific (Rafter et al., 2013). We also examine the degree to which the remineralization of newly fixed N imprints on the subsurface isotopic signal of nitrate in the two regions and consider the origins of distinct features among regional profiles. Finally, the isotopic composition of nitrate previously measured in the EUC (Rafter & Sigman, 2016) provides a basis to assess the relative contributions of northern and southern western boundary currents to the EUC. The observations reveal that stations north and south of the equator have divergent isotopic properties, suggesting limited communication and different hydrographic histories, and providing evidence of differential contributions to the EUC. These observations help elucidate the sources, transformation, and communication of subsurface nutrients in the region and are essential for predicting long-term local and regional biogeochemical variability (Kienast et al., 2008).
2 Materials and Methods
Seawater samples were collected during the EISPAC expedition (SO-228) aboard the RV SONNE in May and June 2013 using a rosette water sampler equipped with 24 10 L Niskin bottles (Mohtadi et al., 2013). Hydrographic data were obtained using a Seabird SBE911 CTD. During the cruise, a total of 18 CTD profiles were taken from 8°N and 126°E off Mindanao to 7°S and 148°E south of the Bismarck Sea (Figure 1). In addition to temperature, salinity, and pressure, the CTD recorded oxygen concentrations using a Clark-type oxygen sensor, and turbidity and fluorescence using a fluorescence sensor. The strong surface and subsurface currents in the study area, especially off Mindanao, resulted in some difficulties in reproducing down-cast and up-cast O2 profiles. Compared to measurements at corresponding WOCE stations, oxygen profiles showed analogous patterns to the WOCE data, but differed in terms of absolute concentrations. The oxygen profiles measured during SO-228 are thus presented as uncorrected instrumental units, and will only be referred to in terms of overall pattern rather than absolute values.
Water samples for stable isotope measurements of DIC were collected at 16 stations from the surface to a depth of around 4,000 m. A portion of the water collected in the Niskin bottles was siphoned into 100 mL glass bottles with water enough to overflow twice to avoid the formation of bubbles. The samples were poisoned with 50 μL of mercury (II) chloride (HgCl2) to prevent alterations of the actual δ13CDIC by biological activity. The glass bottles were sealed with wax and stored at 4°C until further analysis at the MARUM isotope laboratory in Bremen. The analyses were carried out on a gas bench coupled to a Finnigan MAT 252 mass spectrometer using 1 mL of seawater. The routinely performed measurements of the internal standard SHK, which is calibrated against NBS 19 and seawater from the deep Atlantic Ocean, indicated a long-term standard deviation better than 0.1‰. The δ13CDIC was measured in March 2014, nearly 9 months after collection. With a few exceptions, no bubbles were present in the samples, and measured δ13CDIC values agree with historic WOCE values for corresponding locations and water masses.
Water samples for stable nitrogen and oxygen isotope analyses and nutrient measurements (nitrate, nitrite, phosphate, silicic acid, and ammonium) were collected at 12 stations. Seawater from the Niskin bottles was collected into a syringe, then filtered through a 45 μm surfactant-free cellulose acetate (SFCA) membrane into 15 mL high-clarity polypropylene conical (HCPC) tubes for nutrient measurements and into acid-washed and prerinsed 60 mL polyethylene bottles for isotope analyses. The samples were stored at −20°C until analyses.
Nutrient analyses were conducted postcruise at the Bedford Institute of Oceanography (BIO). Nitrate and nitrite in seawater were measured according to the Industrial Method 158-71W adapted from Armstrong et al. (1967) and Grasshoff (1969). The concentration of phosphate was quantified colorimetrically using the Industrial Method 155-71W after Murphy and Riley (1962). For silicic acid measurements, the Industrial Method 186-72W was used according to Strickland and Parsons (1972). Ammonium concentrations were determined fluorometrically after Kérouel and Aminot (1997). Due to long-term storage, however, phosphate and silicic acid concentrations at the same stations could not be reconciled with WOCE measurements at corresponding hydrographic stations.
To discern any excess or deficit of nitrate relative to coincident phosphate, we exploit the semiconservative N* tracer, defined here as N* (µM) = [
] − 16*[
] (Gruber & Sarmiento, 1997). N* quantifies the concentration of reactive N added (from N2 fixation) or lost (to denitrification) assuming that organic material is remineralized in a molar ratio of 16:1 (Redfield, 1934). We also refer to Si*, which corresponds to the difference between silicic acid and nitrate (Si* = [Si(OH)4] − [
]) as defined after Sarmiento et al. (2004). To ensure relative accuracy in calculating N* and Si*, we relied on nutrient concentrations from WOCE at corresponding stations for the end-member mixing estimates. The agreement between our measured nitrate concentrations and those from WOCE data provides assurance that the use of historical nutrient tracers in the mixing calculations is adequate.
To better interpret subsurface nutrient distributions, we calculated the mixed layer depth at each station using a vertical density gradient criterion defined by Wijffels et al. (1994), where the mixed layer depth represents the depth at which the density differs by 0.01 kg m−3 compared to the surface density.
Nitrate 15N/14N and 18O/16O measurements were performed postcruise at the University of Connecticut using the “denitrifier method” (Casciotti et al., 2002; Sigman et al., 2001). This method uses denitrifying bacteria—in this case Pseudomonas chlororaphis f. sp. aureofaciens (ATCC# 13985)—that lack an active nitrous oxide (N2O) reductase, to convert nitrate into a nitrous oxide gas analyte. N and O isotopes in nitrous oxide were measured using a Delta V Advantage continuous flow isotope ratio mass spectrometer interfaced with a Gas Bench II and sample preparation device, dual cold traps, and GC Pal auto-sampler. The 15N/14N reference is N2 in air and the 18O/16O reference is Vienna Standard Mean Ocean Water (VSMOW).
Nitrate samples were standardized to seawater-based reference material USGS-32 (δ15N of 180‰ versus air; δ18O of 25.7‰ versus VSMOW), USGS-34 (δ15N of −1.8‰ versus air; δ18O of −27.9‰ versus VSMOW), and IAEA-N3 (δ15N of 4.7‰ versus air; δ18O of 25.6‰ versus VSMOW) (Böhlke et al., 2003; Gonfiantini, 1984). Standards were prepared in nitrate-free seawater collected in the surface waters of the Sargasso Sea. Standard concentrations were adjusted to that of the corresponding samples in order to account for matrix effects on nitrate δ18O measurements (Weigand et al., 2016). Where nitrite was present, it was removed from samples with sulfamic acid prior to the addition of the denitrifiers according to Granger and Sigman (2009). Based on a minimum of triplicate measurements of individual samples, the analytical precision was 0.2‰ for δ15NNO3 and 0.3‰ for δ18ONO3.
(2)
sources (water masses in this case) and
tracers (potential temperature, salinity, oxygen, nitrate, etc.). Let
be a vector of measured tracer values for a field sample of unknown composition and
the set of unknown mixing fractions that must sum to one. The matrix
is a rectangular matrix of the measured tracer values for each of the sources (supporting information Table S1). The unknown mixing fractions
can be obtained by solving the following linear least squares problem:
(3)
(4)
(5)Equation 4 expresses the fact that the mixing fractions must sum to 1;
is a row vector of
ones and
a column vector of 0s. The third inequality (equation 5) states that mixing fractions must be strictly positive. In order to obtain statistically rigorous solutions to the mixing problem, we use a Monte Carlo approach to generate an ensemble of solutions (Smith, 1984; Van den Meersche et al., 2009). The Monte Carlo method generates samples in the feasible set for the problem defined above: the feasible set consists of values for the mixing fractions that (a) add up to unity and (b) are all strictly positive. Rather than sample the feasible set uniformly, the samples are weighted by the error between the calculated tracer values from the model and the measured values
, as is standard in Markov Chain Monte Carlo calculations. We used the mirroring method described by Van den Meersche et al. (2009), but obtained identical results with the older, simpler sampling methods (Smith, 1984). For all the mixing models, we generated 20,000 samples and adjusted the step size in the sampling algorithm to ensure at least 30% of trial moves were accepted.
3 Study Area
A total of 18 hydrographic stations were visited during the EISPAC expedition (SO-228) aboard the RV SONNE in May and June 2013, off Mindanao at 8°N and 126°E and within and surrounding the Bismarck Sea at 7°S and 148°E (Figure 1). The complex bathymetry and hydrography of the WEP have been investigated, particularly during the Western Equatorial Pacific Observation Circulation Study (WEPOCS; e.g., Bingham & Lukas, 1994; Lindstrom et al., 1987; Toole et al., 1988; Tsuchiya et al., 1989) with the overarching goal of elucidating the convoluted circulation pattern of the low-latitude western boundary currents (LLWBCs). An exhaustive list of the abbreviated names of the numerous currents and water masses of the WEP is provided in Table 1. These LLWBCs consist of the New Guinea Coastal Current (NGCC) and the New Guinea Coastal Undercurrent (NGCUC) in the southern hemisphere, and the Mindanao Current (MC) and Mindanao Undercurrent (MUC) in the northern hemisphere (Lukas et al., 1996; Figure 1c).
South of the equator, the main water transport into the Bismarck Sea occurs through the Vitiaz Strait between the islands of New Guinea and New Britain (Figure 1b). This 1,100 m deep and ∼50 km wide sill separates the Bismarck Sea from the Solomon Sea and restricts deep and bottom water of southern origin from entering the coastal area of PNG. An additional, much smaller transport from the Solomon Sea into the Bismarck Sea has been recorded through St. Georges Channel (SGU; Butt & Lindstrom, 1994). The western boundary currents off PNG are fed by the broad South Equatorial Current (SEC), which crosses the Pacific at ∼3°N–20°S and brings water from the subtropical Pacific to the western boundary (Figure 1a). Upon reaching the western boundary, the SEC splits near 15°S into the North Vanuatu Jet (NVJ), and North and South Caledonian Jet (NCJ and SCJ, respectively). The NVJ flows directly into the Solomon Sea where it joins the NGCUC (Cravatte et al., 2011), while the NCJ crosses the Coral Sea before splitting into two branches, one branch turning south into the Eastern Australian Current (EAC) and a second branch turning north into the Great Barrier Undercurrent (GBRUC) and NGCUC (Figure 1c; Germineaud et al., 2016; Qu & Lindstrom, 2002). One part of the NGCUC crosses the Vitiaz Strait into the Bismarck Sea, while another part turns east south of the Bismarck Sea forming the New Britain Coastal Undercurrent (NBCU), which subsequently feeds the New Ireland Coastal Undercurrent (NICU) that flanks the eastern edge of New Ireland (NI). The NICU, in turn, bifurcates at the northern tip of NI into a western branch flowing into the Bismarck Sea and an eastward branch joining the EUC (Butt & Lindstrom, 1994). Subsurface waters east of NI are also influenced by a northern-more branch of the SEC at 3°S, which splits into a northern and southern branch as it approaches NI and the Solomon Islands (SI). The NICU east of NI is thus fed by waters from the NBCU, the Solomon Island Coastal Undercurrent (SICU) and the low-latitude SEC, with their relative contribution depending on the season (Melet et al., 2010). Along the northeastern PNG coast, the surface NGCC turns to the east feeding the North Equatorial Countercurrent (NECC), while the underlying NGCUC partly turns eastward feeding the EUC (Bingham & Lukas, 1994) and partly crosses the equator and diverges into the Indonesian Throughflow or northward along the coast of the Philippines into the Mindanao Undercurrent (Figure 1c; Qu & Lindstrom, 2004; Tsuchiya, 1968). The eastward deflection of the NGCC into the NECC further results in a quasi-stationary eddy structure called the Halmahera Eddy (HE) northwest of PNG (Fine et al., 1994; Kashino et al., 2013).
North of the equator, the boundary currents are fed by the NEC, which crosses the Pacific at ∼10°N–20°N (Figure 1a). The NEC bifurcates at 14°N into two branches, the northward Kuroshio Current and the southward MC. At the southern tip of the Philippines, one branch of the MC flows southwestward into the Celebes Sea and feeds the Indonesian Throughflow (Gordon & Fine, 1996), while the other branch of the MC flows eastward and feeds the NECC (Figure 1c). The latter branch creates the persistent quasi-stationary Mindanao Eddy (ME) off the coast of the Philippines (Fine et al., 1994; Kashino et al., 2013; Lukas et al., 1991). Both the NECC, as well as its southern counterpart the South Equatorial Countercurrent (SECC), are broad zonal surface currents in between ∼5°N–10°N and ∼3°S–10°S, respectively. Below the surface, the Northern and Southern Subsurface Countercurrents (NSCC and SSCC) flow eastward at ∼3° off the equator (Figure 1c).
Several water masses are associated with the LLWBCs (Figure 2). To the south, off PNG, subsurface water masses are composed of South Pacific Tropical Water (SPTW), apparent as a salinity maximum (∼35.6) in the upper thermocline (∼24.8σθ), and Western South Pacific Central Water (WSPCW) in the lower thermocline (∼26.4σθ). SPTW originates in the oligotrophic subtropical South Pacific gyre due to high evaporation and consequent subduction of surface water (Grenier et al., 2013; Tomczak & Godfrey, 1994). In the WEP, SPTW can further be divided into two branches, one branch reaching the Solomon Sea via the NVJ (Germineaud et al., 2016; Grenier et al., 2013), and a second branch entering the Solomon Sea via the GBRUC (Gasparin et al., 2014). WSPCW forms seasonally through winter-convection in the subtropical convergence zone between Tasmania and New Zealand (Grenier et al., 2014; Roemmich & Cornuelle, 1992; Qu et al., 2009) and reaches the southern study site via the NCJ and GBRUC (Gasparin et al., 2014). Below the thermocline, SAMW-density water (26.8–27.1σθ) is discernible as a pycnostad with relatively high oxygen concentrations and a characteristically low silicate-to-nitrate ratio (Si* minimum; Sarmiento et al., 2004). The underlying Antarctic Intermediate Water (AAIW) is discernible as a salinity minimum (∼34.4) centered around 800 m (27.2σθ) and extending down to ∼1,000 m. Off New Hanover (NH; GeoB1726-1), intermediate water masses are composed of Equatorial Pacific Intermediate Waters (EqPIW; ∼27.3σθ), permeating along the outer edge of the Bismarck Sea. These correspond to the South Pacific Tropical Intermediate Water described by Bingham and Lukas (1995), originating in the eastern South Pacific and reaching the study area below the EUC via the Equatorial Intermediate Current. Upper Circumpolar Deep Water (UCDW) underlies AAIW and EqPIW and forms a distinctive nearly isohaline (34.53 ± 0.03) layer between 1,200 and 2,000 m (27.3–27.7σθ).
(a) Potential temperature (θ) and (b) nitrate concentration versus practical salinity, including summary of prevalent water masses. Gray lines (a) indicate isopycnal lines, and colors show (a) Si* ([Si(OH)4] − [
]) and (b) latitude of stations. Water masses include the North and South Pacific Tropical Water (NPTW and SPTW), Subantarctic Mode Water (SAMW), North Pacific Intermediate Water (NPIW), and Antarctic Intermediate Water (AAIW).
North of the equator off Mindanao, subsurface waters correspond to North Pacific Tropical water (NPTW) in the upper thermocline, apparent as a salinity maximum (∼34.9) centered around ∼140 m (∼24.0σθ), and North Pacific Intermediate Water (NPIW) in the lower thermocline, with a characteristic salinity minimum (≤34.4) centered at ∼350 m (∼26.6σθ; Figure 2). NPTW forms at ∼25°N where high evaporation leads to the subduction of surface water (Katsura et al., 2013; Tsuchiya, 1968). The salinity minimum of the NPIW is formed in the northwestern part of the subtropical gyre between the Kuroshio Extension and the Oyashio front (Talley, 1993). Both NPTW and NPIW are carried westward by the NEC and reach Mindanao via the MC. Pacific Deep Water (PDW), underlaying NPIW (Kawabe & Fujio, 2010; Johnson & Toole, 1993; Wijffels et al., 1996), is a mixture of Antarctic Bottom Water, North Atlantic Deep Water and AAIW, and is characterized by reduced oxygen concentrations and high nutrient loads (Tomczak & Godfrey, 1994).
4 Results
4.1 General Hydrography and Water Mass Distribution
Hydrographic profiles reveal that stations south of the equator near PNG differ markedly from stations north of the equator off the coast of Mindanao. In general, surface and intermediate-depth waters are more saline and warmer south of equator than at corresponding depths off Mindanao (Figures 2 and 3, and Table 2). Both regions are characterized by relatively warm (≥28.6°C), fresh (≤34.0) surface waters and a shallow mixed layer ranging between ≤5 m near the coast to a maximum of 40 m further offshore (GeoB17426-1; Figures 2 and 3). Waters below the surface mixed layer correspond to the “Barrier Layer” (BL) of the Western Pacific Warm Pool and equatorial Pacific (Lukas & Lindstrom, 1991), an intermediate layer isothermal with the surface (28–30°C) that prevents heat flux through the bottom of the mixed layer into the thermocline (Tomczak & Godfrey, 1994).
Water column profiles of (a) salinity, (b) temperature, (c) oxygen, (d) nitrate concentration, and (e) N* versus depth. Colors illustrate different stations with shades of blue/green indicating northern sites and shades of red showing southern sites. (e) N* values are derived from WOA13. Note variations in depth intervals on y axis.
| General | Currents | ||
|---|---|---|---|
| AOU | Apparent Oxygen Utilization | EAC | Eastern Australian Current |
| BL | Barrier Layer | EUC | Equatorial Undercurrent |
| DIC | Dissolved Inorganic Carbon | GBRUC | Great Barrier Reef Undercurrent |
| NB | New Britain | HE | Halmahera Eddy |
| NH | New Hanover | LLWBC | Low-Latitude Western Boundary Current |
| NI | New Ireland | MC | Mindanao Current |
| ODZ | Oxygen Deficient Zone | ME | Mindanao Eddy |
| PNG | Papua New Guinea | MUC | Mindanao Undercurrent |
| SI | Solomon Islands | NBCU | New Britain Coastal Undercurrent |
| WEP | Western Equatorial Pacific | NCJ | North Caledonian Jet |
| NEC | North Equatorial Current | ||
| Water Masses | NECC | North Equatorial Countercurrent | |
| AAIW | Antarctic Intermediate Water | NGCC | New Guinea Coastal Current |
| EqPIW | Equatorial Pacific Intermediate Water | NGCUC | New Guinea Coastal Undercurrent |
| LCDW | Lower Circumpolar Deep Water | NICU | New Ireland Coastal Undercurrent |
| NPIW | North Pacific Intermediate Water | NSCC | Northern Subsurface Countercurrent |
| NPTW | North Pacific Tropical Water | NVJ | North Vanuatu Jet |
| PDW | Pacific Deep Water | SCJ | South Caledonian Jet |
| SAMW | Subantarctic Mode Water | SEC | South Equatorial Current |
| SPTW | South Pacific Tropical Water | SECC | South Equatorial Countercurrent |
| UCDW | Upper Circumpolar Deep Water | SGU | St. Georges Undercurrent |
| WSPCW | Western South Pacific Central Water | SICU | Solomon Island Undercurrent |
| SSCC | Southern Subsurface Countercurrent | ||
| Lon | Lat | Water masses | SigmaT (kg m−3) | Potential temp (°C) | Salinity | Nitrate (μM) | δ13CDIC (‰) | δ15NNO3 (‰) | δ18ONO3 (‰) | Δ(15–18) (‰) | N*a (μM) | AOUb (μmol kg−1) | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Mindanao | 127°E | 8°N | NPTW | 23.0–25.0 | 23.5 ± 1.4 | 34.98 ± 0.08 | 1.9 ± 0.7 | 0.6 ± 0.3 | 5.7 ± 0.2 | 3.5 ± 0.3 | 2.3 ± 0.6 | −3.7 ± 0.3 | 37.6 ± 15.5 |
| NPIW | 26.5–26.8 | 9.2 ± 0.8 | 34.37 ± 0.04 | 29.1 ± 2.5 | 0.1 ± 0.0 | 7.1 ± 0.1 | 3.7 ± 0.2 | 3.5 ± 0.2 | −5.2 ± 0.4 | 177.9 ± 12.9 | |||
| PDW | 27.5–27.7 | 2.7 ± 0.4 | 34.59 ± 0.02 | 39.6 ± 0.5 | 0.1 ± 0.0 | 6.0 ± 0.2 | 2.3 ± 0.1 | 3.7 ± 0.1 | −6.0 ± 0.3 | 218.3 ± 3.7 | |||
| Bottom Water | >27.7 | 1.6 ± 0.1 | 34.64 ± 0.01 | 37.0 | 0.2 | 5.4 | 2.0 | 3.4 | −5.2 ± 0.5 | 193.4 ± 11.5 | |||
| New Hanover | 151°E | 2°S | SPTW | 24.3–25.3 | 21.6 ± 1.6 | 35.63 ± 0.06 | 8.6 ± 0.1 | 0.8 ± 0.1 | 8.3 | 3.0 | 5.3 | −4.0 ± 0.7 | 92.8 ± 8.8 |
| SAMW | 26.8–27.1 | 8.2 ± 0.8 | 34.57 ± 0.04 | 30.5 ± 1.9 | 0.8 ± 0.1 | 6.9 ± 0.1 | 2.7 ± 0.1 | 4.2 ± 0.1 | −3.4 ± 0.5 | 183.4 ± 20.4 | |||
| EqPIW | 27.2–27.4 | 4.7 ± 0.4 | 34.50 ± 0.01 | 35.7 ± 0.8 | 0.6 ± 0.2 | 6.3 ± 0.1 | 2.3 ± 0.0 | 4.0 ± 0.1 | −4.7 ± 0.5 | 188.6 ± 11.4 | |||
| UCDW | 27.4–27.5 | 3.8 ± 0.2 | 34.53 ± 0.01 | 37.3 ± 0.0 | 0.7 ± 0.4 | 6.1 ± 0.0 | 2.2 ± 0.0 | 3.9 ± 0.1 | −5.2 ± 0.1 | na | |||
| Papua New Guinea | 147°E | 6°S | SPTW | 24.3–25.3 | 19.7 ± 0.3 | 35.51 ± 0.02 | 8.8 ± 0.1 | 0.8 ± 0.0 | 7.5 ± 0.1 | 3.0 ± 0.0 | 4.5 ± 0.1 | −2.9 ± 0.2 | 57.9 ± 7.8 |
| WSPCW | 25.8–26.5 | 14.4 ± 1.4 | 35.10 ± 0.12 | 15.5 | na | 6.9 | 3.0 | 3.9 | −3.4 ± 0.2 | 94.2 ± 7.0 | |||
| SAMW | 26.8–27.1 | 7.8 ± 0.9 | 34.53 ± 0.06 | 25.1 ± 1.5 | 1.1 ± 0.1 | 6.5 ± 0.1 | 2.7 ± 0.1 | 3.9 ± 0.1 | −3.3 ± 0.4 | 134.9 ± 17.8 | |||
| AAIW | 27.1–27.3 | 5.3 ± 0.5 | 34.46 ± 0.01 | 32.4 ± 1.2 | 0.7 ± 0.0 | 6.3 ± 0.0 | 2.2 ± 0.1 | 4.0 ± 0.0 | −4.1 ± 0.2 | 160.9 ± 7.9 | |||
| UCDW | 27.3–27.7 | 3.8 ± 0.5 | 34.53 ± 0.03 | 36.8 ± 1.3 | 0.4 ± 0.2 | 6.0 ± 0.1 | 2.1 ± 0.2 | 3.9 ± 0.1 | −5.4 | 207.4 |
- a Nutrient for N* from WOCE (Mindanao and PNG) and WOD (NH).
- b Oxygen for AOU from WOCE (Mindanao and PNG) and PANDORA (NH).
At the southern stations off PNG, the thermocline extends from 60 to 400 m depth. Thermocline waters are comprised of SPTW (24.8σθ, 19.7 ± 0.3°C), characterized by a salinity maximum of 35.51 ± 0.02 centered at ∼175 m in the upper thermocline (Figures 2 and 3, and Table 2), and WSPCW (∼26.2σθ, 14.4 ± 1.4°C) in the lower thermocline. Subthermocline waters correspond to SAMW-density water (26.8σθ, 7.8 ± 0.9°C), with lower salinity than overlaying tropical waters (34.53 ± 0.06) and a pronounced Si* minimum centered around ∼500 m (Figure 2a and supporting information Figure S3). AAIW is discernible as a salinity minimum of 34.46 ± 0.01 centered around 800 m (27.2σθ, 5.3 ± 0.5°C) and extending down to ∼1,000 m. Below intermediate waters, salinity increases to 34.53 ± 0.03 in UCDW (∼27.5σθ, 3.8 ± 0.5°C).
At the northern stations off Mindanao, the warm BL extends to the top of the thermocline at 90 m depth. The upper thermocline consists of NPTW (24.0σθ, 23.5 ± 1.4°C) centered at ∼140 m, with a less pronounced salinity maximum of 34.98 ± 0.08 than the corresponding SPTW. The lower thermocline corresponds to NPIW (26.6σθ, 9.2 ± 0.8°C), characterized by a salinity minimum (≤34.4) centered at ∼350 m. Though not readily apparent, a slight salinity minimum (≥27.2σθ) and lower temperature (0.5°C) at ∼800 m in between NPIW and the underlying PDW (27.6σθ; 2.7 ± 0.4°C) suggest some intrusion of AAIW reaching Mindanao from south of the equator. Below PDW (>3,000 m), salinity increases from 34.59 ± 0.02 in PDW to 34.65 ± 0.01 at ∼3,800 m.
4.2 Nitrate and N* Distributions
The concentration of nitrate is below detection in the surface mixed layer at stations north and south of the equator (Figures 2b and 3d, and Table 2). In both regions, nitrate is detectable in the BL and increases monotonically into tropical waters below: nitrate concentrations in SPTW (8.6 ± 0.1 µM) are higher than in NPTW (1.9 ± 0.7 µM; Figure 2b). At the southern stations, the concentration of nitrate increases in SAMW-density water to ∼25 µM off PNG and to ∼30 µM off NH (GeoB17426-1), respectively. In the underlying intermediate layer, nitrate concentrations average 32.4 ± 1.2 µM in AAIW off PNG and 35.7 ± 0.8 µM in EqPIW off NH (Table 2). At corresponding depths north of the equator, the concentration of nitrate in NPIW is ∼30 µM, increasing to nearly 40 µM in PDW. Bottom waters north and south of the equator have indistinguishable nitrate concentrations of ∼37 µM.
The N* signal, indicative of any stoichiometric excess or deficit in nitrate relative to phosphate, is generally less pronounced at stations south of the equator relative to corresponding depths at northern stations (Figure 3e). Off PNG, N* in SPTW is −2.9 ± 0.2 µM and declines throughout the water column to ∼−3.3 µM in WSPCW and SAMW-density water, −4.1 ± 0.2 µM in AAIW, and −5.4 µM in UCDW, respectively. A similar decline from the thermocline toward deep waters is apparent north of the equator off Mindanao, with N* decreasing from −3.7 ± 0.3 µM in NPTW to −5.2 ± 0.4 µM in NPIW, reaching a minimum of −6.0 ± 0.3 µM in PDW, followed by a slight increase to −5.2 ± 0.5 µM near the bottom.
4.3 Nitrate and DIC Isotope Distributions
Profiles of δ15NNO3 differ between stations north and south of the equator, particularly in the thermocline and toward the surface (Figure 4a). Due to nitrate levels below the lower limit of quantification for nitrate isotope ratios (<0.5 µM), no N and O isotope data are available for surface waters. In the underlying BL, δ15NNO3 values in a number of profiles are relatively elevated in both regions, particularly at southern stations off NH where values reach upward of 17.6‰ at 50 m depth. The δ15NNO3 decreases progressively in the thermocline, with tropical waters being substantially more 15N-enriched at southern stations (7.4–9.4‰) compared to stations off Mindanao (5.5–7.6‰). Values in SAMW-density water and NPIW below are comparable (6.5–7.2‰), albeit with slightly higher values in the core of NPIW. Among southern profiles, the δ15NNO3 in SAMW-density water off NH (6.9 ± 0.1‰) is notably more elevated than at corresponding depths off PNG (6.5 ± 0.1‰). The δ15NNO3 in intermediate waters are similar in both regions (∼6‰ in AAIW and PDW) and remain comparable in deep waters (5.5‰, Figure 4a).
Water column profiles of (a) δ15NNO3, (b) δ18ONO3, (c) nitrate Δ(15–18), and (d) δ13CDIC versus depth. Note differences in intervals on x (a and b) and y axes. Colors illustrate different stations with shades of blue/green indicating northern sites and shades of red showing southern sites.
The δ18ONO3 profiles reveal lower values at most depth intervals for stations south of the equator (Figure 4b). Similar to δ15NNO3, δ18ONO3 values are elevated in the BL at a number of stations, particularly off NH (≤17.6‰). Below the BL, values decrease throughout the upper thermocline to an average of 3.5 ± 0.3‰ in NPTW and 3.0 ± 0.0‰ in SPTW. Values of δ18ONO3 decrease further toward mode and intermediate waters at the southern site (2.7 ± 0.1‰ in SAMW-density water, 2.2 ± 0.1‰ in AAIW) while slightly increasing in intermediate waters off Mindanao (3.7 ± 0.2‰ in NPIW) before decreasing in PDW (2.3 ± 0.1‰). In deep waters, δ18ONO3 values are comparable in UCDW and North Pacific bottom water averaging ∼2.1‰.
The contrasting depth distributions of δ15NNO3 and δ18ONO3 north and south of the equator give way to strikingly divergent profiles of Δ(15–18) between the two regions (Figure 4c). South of the equator, values reach maxima in SPTW ranging from 4.4‰ to 5.3‰, followed by a decrease in mode and intermediate waters (∼4.0‰). Conversely, Δ(15–18) values show distinct minima in NPTW (≥1.6‰) and a subsequent increase toward intermediate and deep waters (3.5 ± 0.2‰ and 3.7 ± 0.1‰ in NPIW and PDW, respectively).
Finally, δ13CDIC values are generally more elevated at the southern sites relative to corresponding depths north of the equator (Figure 4d and Table 2). Surface values are high in both regions (1.2–1.6‰), albeit perceptibly more so at southern stations. Values of δ13CDIC decrease with depth reaching values of 0.8 ± 0.1‰ versus 0.6 ± 0.3‰ in SPTW and NPTW, respectively. Values decrease further to an average of 0.1 ± 0.0‰ in NPIW and in underlying PDW, whereas they increase slightly in SAMW-density water to a maximum of ∼1.1 ± 0.1‰. As with other tracers, δ13CDIC in SAMW-density water off NH differs from that off PNG, posting lower values of 0.8 ± 0.1‰. Below the mode water, δ13CDIC values are 0.7 ± 0.0‰ in AAIW, decreasing further in UCDW. A slight maximum in δ13CDIC at 27.1σθ at stations north of the equator is evident, consistent with the intrusion of AAIW off Mindanao.
5 Discussion
The nutrient characteristics of WEP thermocline waters (≈100–400 m) are important because they fuel primary production locally and across the Pacific basin via the EUC (Rafter & Sigman, 2016; Toggweiler et al., 1991). Our δ15NNO3, δ18ONO3, and δ13CDIC measurements indicate a different biogeochemical history of WEP nutrients north (Mindanao stations) versus south (PNG and NH) of the equator (Figure 4 and Table 2). Specifically, our data indicate interbasin differences in (a) the contribution of remineralization to bulk nutrients in the thermocline, (b) the δ15N of organic material remineralized in the thermocline, and (c) the lateral contribution of nutrients from the eastern margins. We also identify subtle nutrient differences between PNG and NH regions south of the equator that suggest different southern pathways to the “water mass crossroads” of the Pacific (Fine et al., 1994).
To understand the origins of these contrasting thermocline-depth nutrient characteristics, we examine our data beginning with WEP intermediate-depth (≈400–1,200 m) waters, which resupply thermocline nutrients to the respective basins via SAMW (Palter et al., 2010) and NPIW. We then examine tracer distributions in the WEP thermocline and derive the δ15N of sinking organic matter that is necessary to explain nitrate biogeochemical differences in the northern and southern WEP, which has implications for tracing the contribution of newly fixed nitrogen to the sinking organic matter flux. Finally, because nutrient characteristics of the WEP thermocline fuel equatorial Pacific primary production (Rafter & Sigman, 2016), we investigate the contribution of northern and southern WEP waters to the EUC.
5.1 The Biogeochemical History of WEP Intermediate-Depth Waters
To a first approximation, lower oxygen concentrations, lower δ13CDIC values, and higher nutrient concentrations in northern WEP intermediate waters (off Mindanao; Figures 3c, 3d, and 4) point to a greater accumulation of remineralized material relative to the southern sites. The greater contribution of remineralized nitrate to the total pool at the northern sites should be reflected in the δ18ONO3, which is sensitive to nitrification: newly nitrified nitrate adopts a δ18O signature similar to that of ambient seawater (Buchwald et al., 2012; Casciotti et al., 2008; Sigman et al., 2009b), thus tending to lower the δ18ONO3 proportionally. As such, a greater contribution of remineralized nitrate should manifest as a lower δ18ONO3 of subsurface nitrate at the northern stations. Contrary to expectations, however, the δ18ONO3 is more elevated at the northern sites (Figure 4b). This difference suggests that other processes overprint the tendency toward lower δ18ONO3 due to nitrification. Below, we examine the biogeochemical history of intermediate and mode waters in the context of their origin and circulation to explain the apparent discrepancy between δ13CDIC and δ18ONO3.
5.1.1 The Origin of Southern WEP δ18ONO3
With respect to the South Pacific, both SAMW and AAIW form due to winter cooling and deep convection of the surface layer at the Subantarctic Front of the Southern Ocean (McCartney, 1977; Talley, 1996). High surface productivity in the Southern Ocean aided by strong air-sea exchange at cold temperatures (Bostock et al., 2010) imprints relatively elevated initial δ13CDIC values on SAMW and AAIW. In turn, the δ15NNO3 and δ18ONO3 in SAMW and AAIW are set by those of nitrate in the source waters and by biological transformations in the surface layer. In the Southern Ocean, UCDW, which upwells to the surface as part of the Southern Overturning Circulation (Tomczak & Godfrey, 1994), is considered as the source water of SAMW. This mechanism delivers nutrients to the surface mixed layer of the Open Antarctic Zone (Orsi et al., 1995) with characteristic isotopic signatures (δ15NNO3 of ∼5.0‰, δ18ONO3 of ∼2.0 ‰, Table 3; DiFiore et al., 2009; Rafter et al., 2013; Sigman et al., 1999, 2000). In transit from the Open Antarctic Zone and across the Subantarctic Front towards the Subantarctic Zone, isotopic discrimination associated with the partial assimilation of nitrate results in the export of relatively low-δ15N organic matter from the surface (Altabet & François, 1994; Karsh et al., 2003; Lourey et al., 2003), and a parallel increase in both the δ15NNO3 and δ18ONO3 remaining in surface waters that form SAMW at the Subantarctic Zone (DiFiore et al., 2006; Rafter et al., 2013; Sigman et al., 1999; Smart et al., 2015). The resulting δ15NNO3 and δ18ONO3 values of high latitude SAMW (51°S–41°S) and AAIW (56°S–51°S) are 6.2 ± 0.4‰ and 3.5 ± 0.7‰ and 5.5 ± 0.2‰ and 2.8 ± 0.3‰, respectively (DiFiore et al., 2006; Rafter et al., 2013; Sigman et al., 1999; Smart et al., 2015). Prior work illustrated that intermediate-depth δ18ONO3 south of the equator is lower than its high latitude source water (SAMW) because of nitrate added by organic matter remineralization (Rafter et al., 2013). The δ18ONO3 values observed in SAMW-density waters of the southern WEP are similarly lower than the Southern Ocean end-member (Figure 5), suggesting nitrate was added in transit by remineralization.
| Lon | Lat | Water masses | SigmaT (kg m−3) | Potential temp (°C) | Salinity | Nitrate (μM) | δ13CDICa (‰) | δ15NNO3 (‰) | δ18ONO3 (‰) | Δ(15–18) (‰) | N* (μM) | AOU (μmol kg−1) | Reference | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ALOHA | 158°W | 23°N | TW | 23.0–25.0 | 20.6 ± 0.5 | 35.19 ± 0.04 | 1.9 ± 0.6 | na | 4.6 ± 0.1 | 3.3 ± 0.1 | 1.3 ± 0.0 | −1.7 ± 0.4 | na | Sigman et al. ( 2009a, 2009b) |
| NPIW | 26.5–26.8 | 7.4 | 34.08 | 28.2 | −0.5–0 | 6.7 | 3.3 | 3.4 | −4.9 | na | Sigman et al. ( 2009a, 2009b) | |||
| PDW | 27.5–27.7 | 2.5 ± 0.5 | 34.60 ± 0.03 | 40.9 ± 0.5 | na | 5.7 ± 0.3 | 2.3 ± 0.2 | 3.5 ± 0.2 | −5.5 ± 0.4 | na | Sigman et al. ( 2009a, 2009b) | |||
| Abyssal | >27.7 | 1.5 ± 0.1 | 34.68 ± 0.02 | 37.5 ± 1.1 | na | 5.1 ± 0.1 | 1.9 ± 0.1 | 3.2 ± 0.1 | −4.0 ± 0.4 | na | Sigman et al. ( 2009a, 2009b) | |||
| Equatorial Pacific | 165°E–95°W | 0°N | EUC | ∼26.0 | 16.3 ± 2.6 | 34.99 ± 0.12 | 14.0 ± 6.5 | na | 7.1 ± 0.3 | 3.0 ± 0.3 | 4.1 ± 0.3 | −5.1 ± 0.9 | na | Rafter et al. ( 2012, 2016) |
| 110°W | 5°N | NSCC | ∼26.1 | 13.10 ± 3.14 | 34.67 ± 0.03 | 25.9 ± 4.6 | na | 7.2 ± 0.2 | 3.9 ± 0.5 | 3.3 ± 0.5 | na | na | Rafter et al. ( 2012, 2016) | |
| 110°W | 5°S | SSCC | ∼26.4 | 12..97 ± 0.46 | 34.95 ± 0.03 | 28.7 ± 2.2 | na | 5.8 ± 0.6 | 3.5 ± 0.3 | 2.3 ± 0.7 | na | na | Rafter et al. ( 2012, 2016) | |
| South Pacific | 150°W | 20°S | SAMW | 26.8–27.1 | 6.1 ± 0.6 | 34.37 ± 0.01 | 29.0 ± 1.8 | na | 7.0 ± 0.6 | 2.9 ± 0.7 | 4.1 ± 0.1 | −3.6 ± 0.0 | 149.1 ± 7.4 | Rafter et al. ( 2013) |
| 150°W | 20°S | AAIW | 27.1–27.3 | 4.9 ± 0.5 | 34.41 ± 0.03 | 33.0 ± 1.4 | 0.75–1.75 | 6.4 ± 0.1 | 2.2 ± 0.1 | 4.3 ± 0.1 | −4.2 ± 0.3 | 321.3 ± 3.5 | Rafter et al. ( 2013) | |
| Southern Ocean | 150°W | 51°S–41°S | SAMW | 26.8–27.1 | 7.2 ± 0.9 | 34.41 ± 0.07 | 20.6 ± 3.7 | na | 6.2 ± 0.4 | 3.5 ± 0.7 | 2.7 ± 0.3 | −2.4 ± 0.2 | 58.9 ± 18.7 | Rafter et al. ( 2013) |
| 150°W | 56°S–51°S | AAIW | 27.1–27.3 | 3.3 ± 1.0 | 34.19 ± 0.12 | 29.8 ± 1.5 | 0.85–1.6 | 5.5 ± 0.2 | 2.8 ± 0.3 | 2.7 ± 0.2 | −2.9 ± 0.4 | 334.9 ± 8.6 | Rafter et al. ( 2013) | |
| 150°W | 56°S | UCDW | na | 2.5 ± 0.2 | 34.46 ± 0.14 | 33.5 ± 0.7 | na | 5.0 ± 0.1 | 2.0 ± 0.3 | 3.0 ± 0.2 | −3.5 ± 0.3 | 153.3 ± 19.0 | Rafter et al. ( 2013) |
Property-property plots of δ18ONO3 versus δ15NNO3 including (a) all depths, and density ranges of (b) 24–26 (σθ) and (c) 26.8–27.1 (σθ). Colors represent different stations and (a) gray arrows indicate shoaling. The gray line runs through the mean δ18ONO3 and δ15NNO3 of the Southern Ocean SAMW end-member (3.5‰ and 6.2‰, respectively; Table 3), with its slope of 1 reflecting the isotopic fractionation expected from the consumption of nitrate through denitrification and assimilation, and any deflection from the 1:1 line indicating the remineralization of nitrate. Measurements from (a) station ALOHA are from Sigman et al. (2009a), (a, b) EUC and (a, c) SAMW data from the western (165°E to 170°W), central (140° and 155°W) and eastern (110°W) equatorial Pacific are from Rafter et al. (2012, 2016). Note different scales on x and y axes.
(6)Based on the water-mass characteristic values in Tables 2 and 3, the change in SAMW-density nitrate concentrations are 9.9 μM between the SAMW 50°S end-member and SAMW-density waters at the NH station. A δ18Onitr of 1.0‰ is necessary to explain the change in nitrate concentrations and δ18ONO3 between SAMW and NH stations. This δ18Onitr value is in agreement with similar derivations conducted by Rafter et al. (2013), as well as with the empirical value for global ocean nitrification of +1.1‰ derived by Sigman et al. (2009b). Alternatively, we can estimate the fraction of remineralized (relative to preformed) nitrate using the Apparent Oxygen Utilization (AOU) and the respiration stoichiometry of Anderson (1995). The remineralized nitrate expected based on AOU is 13.3 µM, corresponding to a δ18Onitr of 0.8‰ (equation 6), thus coherent with the nitrate-based mass balance.
Another potential source of nitrate isotopic variability is denitrification in the eastern tropical Pacific (Brandes et al., 1998; Sigman et al., 2005; Voss et al., 2001), which has a strong influence on N* (Gruber & Sarmiento, 1997). N* decreases from the 50°S SAWM end-member from −2.4 ± 0.2 to −3.4 ± 0.5 µM off NH (Table 3). This suggests some entrainment of ODZ water from the eastern margins, reaching the eastern coast of NI and NH via the northern branch of the SEC (Rafter et al., 2012). Thus, mass balance calculations suggest that the net decrease in δ18ONO3 at NH can largely be explained by the addition of remineralized nitrate in transit, countered by modest lateral entrainment of 18O-enriched nitrate from the margins.
At stations off PNG, the SAMW-density δ18ONO3 is identical to that off NH (2.7‰), however, the nitrate concentration is markedly lower (25.1 ± 1.5 µM versus 30.5 ± 1.9 µM, respectively). Repeating the δ18O mass balance exercise for SAMW-density nitrate off PNG results in lower δ18Onitr values (−1.0‰ or −0.5‰ using ΔAOU) that appear inconsistent with expectations. Admittedly, these calculations are subject to large uncertainty and fail to account for a variety of processes, such as differential mixing between PNG and NH stations and end-member water mass composition. Indeed, the complex bathymetry at the western boundary off PNG reportedly fosters diapycnal mixing of intermediate water masses (Germineaud et al., 2016; Grenier et al., 2011; Melet et al., 2011), such that the discrepancies in regenerated nitrate estimates between the NH and PNG stations could arise from the mixing of SAMW-density water with overlaying nitrate-deplete gyre water, lowering the nutrient content of SAMW-density water without affecting its isotopic composition. However, hydrographic differences in other parameters such as δ15NNO3 (discussed further below), salinity, δ13CDIC and oxygen concentrations argue for differences in the end-member water masses ventilating SAMW-density waters off PNG versus NH. In this respect, SAMW reaches PNG waters via the southern branch of the SEC, which may have a lesser contribution of eastern ODZ mixing relative to NH (Rafter et al., 2012). Moreover, the lower thermocline of the Coral Sea is partly fed by the NCJ, which carries Central Water (WSPCW; 25.8–26.5σθ) originating off New Zealand (Germineaud et al., 2016; Grenier et al., 2013, 2014; Roemmich & Cornuelle, 1992).
In summary, tracer distributions and mass balance calculations suggest that the net decrease in δ18ONO3 between the Southern Ocean source of SAMW and our southern WEP SAMW-density waters can largely be explained by the addition of remineralized nitrate in transit. Differential influences from the entrainment of eastern ODZ waters, diapycnal mixing, and ventilation of the PNG thermocline by WSPCW may explain tracer differences between NH and PNG intermediate water masses (further discussed below).
5.1.2 The Origin of Northern WEP δ18ONO3
The northern stations off Mindanao show lower δ13CDIC and higher nutrient concentrations in NPIW-depth and PDW-depth waters, suggesting a larger accumulation of remineralized products relative to corresponding depths south of the equator. This is consistent with the fact that—unlike SAMW and AAIW—NPIW does not outcrop at the surface in the area of formation but reflects remineralization and mixing processes along the flow path. This partly explains the higher nutrient concentrations, lower oxygen concentrations, and lower δ13CDIC values at intermediate depths off Mindanao. As noted above, the higher degree of remineralization would be expected to impart a lower δ18ONO3 relative to southern stations. Yet, intermediate-depth δ18ONO3 values are consistently higher north of the equator. The relatively elevated δ18ONO3 may result from a larger contribution of nitrate advected from the eastern tropical Pacific ODZ with elevated δ18ONO3 from denitrification (Sigman et al., 2005). The larger direct contribution of waters from the eastern margin in the northern WEP is consistent with a larger N-deficit in intermediate waters north of the equator (−5.2 ± 0.4 and −4.1 ± 0.2 µM in NPIW and AAIW, respectively). The δ18ONO3 at the WEP is also elevated (3.7 ± 0.2‰) relative to the central gyre (3.3‰, Sigman et al., 2009a), further suggesting a larger lateral contribution from the eastern margin to the northern WEP. The westward movement of the denitrification signal from the eastern margin is likely propagated by the NEC, as previously inferred (Kienast et al., 2008; Rafter et al., 2012).
Using a mass balance to identify the δ18Onitr for intermediate-depth water in the northern WEP (as in section 5.1.1) is challenging because of the few dual N and O isotope measurements from the higher latitude North Pacific location of NPIW source waters. Moreover, the elevated δ18ONO3 of WEP intermediate waters, in light of the high proportion of remineralized nitrate, suggests an important lateral nitrate input from the eastern margins—thus thwarting a single end-member mass balance exercise. The analysis of the northern WEP will be improved by upcoming high-resolution nitrate isotope measurements during the North Pacific GEOTRACES expedition (www.geotraces.org), which will provide fundamental constraints to elucidate the evolution of NPIW as well as other North Pacific water masses converging at the WEP.
5.1.3 Estimating the δ15N of Remineralized Organic Matter
(7)Off Mindanao, NPIW-depth δ15NNO3 (Figure 4a and Table 2) is also higher than its putative higher latitude source off Japan (7.1 ± 0.1‰ versus 6.1 ± 0.2‰; Yoshikawa et al., 2006). The δ15NNO3 at the western boundary is also higher relative to the same density interval at station ALOHA in the central gyre (6.7 ± 0.1‰; Sigman et al., 2009a, Figure 5a). This increase in δ15NNO3 as intermediate waters transit the North Pacific is somewhat surprising considering the widespread occurrence of N2 fixation in the North Pacific subtropical gyre (Karl et al., 1997, 2002), introducing isotopically light (lower δ15N) organic matter to the subsurface (Casciotti et al., 2008; Karl et al., 1997; Sigman et al., 2009a). In this respect, the salient decrease of Δ(15–18) from intermediate waters to the subsurface (Figure 4c) signals the remineralization of low-δ15N material north of the equator. The low-δ15NNO3 introduced by the remineralization of newly fixed nitrogen may in part be overprinted by the lateral advection and mixing of high-δ15NNO3 from denitrification in the ODZ of the eastern equatorial Pacific, which is consistent with the lower N*, lower Δ(15–18) and higher δ18ONO3 values observed north compared to south of the equator (Figures 3 and 4). These patterns are further consistent with the bulk of organic matter remineralization occurring in the subeuphotic zone (Martin et al., 1987)—an assertion that we test in the following section by analyzing the shallower, thermocline-depth waters.
5.2 The Biogeochemical History of WEP Thermocline-Depth Waters
The isotope composition of nitrate in thermocline and near-surface waters (depths shallower than the intermediate water masses) are drastically different between the northern and southern hemisphere WEP. The most obvious difference is at ≈160–200 m, where the southern δ15NNO3 is 7.5‰ (PNG) and 8.3‰ (NH), but northern δ15NNO3 is 5.7‰ (Table 2). Moreover, southern WEP thermocline δ15NNO3 is higher than the underlying intermediate water nitrate, whereas northern WEP nitrate at these depths is lower than that in waters immediately below (Table 2). In the next section, we examine these patterns in the southern then northern WEP thermocline.
5.2.1 Southern WEP Nitrate: From the Thermocline to the Surface
At southern stations, thermocline depths are contiguous with SPTW, originating at ∼20°S, 125°W in the subtropical gyre (Tomczak & Godfrey, 1994; Tsuchiya et al., 1989). In this area of water mass formation, Peters et al. (2017) report low surface nitrate (<5 µM) and a concurrent enrichment in both δ15NNO3 and δ18ONO3 (up to 28‰ and 25‰, respectively). They attribute these high values to the incomplete consumption of surface nitrate due to Fe-limitation in waters entrained from the eastern part of the upwelling system, namely, the equatorial upwelling and/or possibly upwelling at the margins. As originally proposed by Rafter et al. (2013), the Rayleigh distillation of nitrate isotopes as waters move poleward from the equatorial upwelling (Altabet, 2001; Altabet & Francois, 1994; Rafter & Sigman, 2016) gives way to the sinking and remineralization of high-δ15N particles in SPTW, transmitting the 15N-enrichment from the surface into the subsurface dissolved nitrate pool. This signal is evident as high-δ15NNO3 in the tropical water along the southern WEP.
We note that SPTW entering the Bismarck Sea through the Vitiaz Strait has comparatively lower δ15NNO3 (by 0.8‰), which can be explained by the influence of the southern branch of SPTW. The lower δ15NNO3 of the southern SPTW branch likely reflects a regional contribution of N2 fixation to thermocline nitrate (Rafter et al., 2012; Yoshikawa et al., 2015) and/or a diminishing influence of the northern SPTW branch.
Other ocean regions recognized to host significant in situ N2 fixation, such as the central North Pacific gyre and the North Atlantic, are characterized by low-δ15NNO3 in the upper thermocline, reflecting the input of isotopically “light” nitrogen from diazotrophy, with thermocline δ15NNO3 as low as 1.5‰ and ∼2.6‰ at station ALOHA and in the Sargasso Sea, respectively (Casciotti et al., 2008; Knapp et al., 2008). The comparative absence of an unambiguous N2 fixation imprint in the southern WEP thermocline nitrate pool may owe to (i) limited remineralization of newly fixed N, and/or (ii) an overprinting by isotopically “heavy” nitrate reaching the WEP via the SEC. Against previous model predictions of high N2 fixation rates in the eastern subtropical and tropical South Pacific (Deutsch et al., 2007), Knapp et al. (2016) reported low (≤24 µmol N m−2 d−1) to undetectable rates for that area. Limited N2 fixation in the eastern and central South Pacific gyre may indicate a high iron requirement and the coherent iron-dependent occurrence and spatial distribution of N-fixers (Mills et al., 2004; Moore & Doney, 2007), restricting high N2 fixation rates to areas of elevated iron input such as the western Pacific (Mackey et al., 2002; Slemons et al., 2010). Conforming to these notions, elevated rates of N2 fixation are reported for both the Solomon Sea and Bismarck Sea (Berthelot et al., 2017; Bonnet et al., 2009, 2015). We thus hypothesize that the relatively elevated δ15NNO3 in the southern WEP thermocline owes primarily to the vertical flux of high-δ15N material at the southern edge of the equatorial upwelling overprinting the regionally restricted influence of newly fixed nitrogen.
(8)
(a) δ15NNO3 and (b) δ18ONO3 versus the natural logarithm of [
]. Colors show the latitude of the station, with blue colors indicating the northern site off Mindanao and red colors illustrating southern sites of PNG and NH, respectively.
5.2.2 Northern WEP Nitrate: From the Thermocline to the Surface
North of the equator, NPTW, which feeds the upper thermocline, originates in the North Pacific subtropical gyre along ∼25°N between 150°E and 130°W and enters the WEP and our study site off Mindanao via the broad westward flowing NEC and MC (Fine et al., 1994; Katsura et al., 2013; Nie et al., 2016; Tsuchiya, 1968). Off Mindanao, upper thermocline waters exhibit a clear minimum in δ15NNO3, with values as low as 4.4‰. The decrease in δ15NNO3 from intermediate waters to the lower thermocline occurs despite a coincident decrease in nitrate concentration (Figure 6a). Above the salient δ15NNO3 minimum, a concurrent increase in both δ15NNO3 and δ18ONO3, along with decreasing nitrate concentration, suggest some degree of subsurface nitrate assimilation. The isotopically light thermocline nitrate in the NPTW and corresponding low Δ(15–18) within the same depth range suggest remineralization of organic matter with a low δ15N due to N2 fixation. This conclusion is consistent with previous studies in the area (Kienast et al., 2008). Moreover, the observed δ15NNO3 is comparable to that reported in the corresponding density range elsewhere in the subtropical North Pacific (∼4‰ at station ALOHA, Casciotti et al., 2008; Sigman et al., 2009a; Figure 5a).
While we see a clear signal of N2 fixation in the central North Pacific and at the western boundary, explicit evidence in the South Pacific gyre seems to be missing. As pointed out above, the high δ15N associated with nitrate advected to the surface layer of the subtropical gyres from the eastern equatorial upwelling system (Peters et al., 2017; Rafter et al., 2013) imprints onto the nitrate pool of SPTW. In contrast, this dynamic is not seen in the North Pacific gyre (Rafter et al., 2013), allowing the remineralization of newly fixed N (with a low δ15N) to have a larger influence on the δ15N of the nitrate pool.
5.2.3 Implications for δ15N of Particle Flux at the Equatorial Upwelling
An interesting implication of the apparent transport of elevated δ15N nitrate from the low latitudes across the South Equatorial Current (SEC) into the South Pacific gyre is that this process essentially fractionates the isotopes of reactive N upwelled at the equator among regions. Incomplete consumption of nitrate at the equator leads to the lateral divergence of nutrients away from the equator (Kessler, 2006). The poleward decrease in nitrate and concurrent fractionation toward the lighter isotope during assimilation by phytoplankton creates an inverse correlation between nitrate versus the δ15N of near-surface organic matter (Altabet, 2001; Altabet & Francois, 1994). This process produces an elevated δ15N particle flux in the northern gyre, balanced by a lower δ15N particle flux in the proximity to the upwelling system. The subsurface remineralization of the lower δ15N particle flux is consistent with a zonal band of lighter δ15N nitrate (∼5‰) focused along 3–5°S (Rafter et al., 2012, 2013) within the west-to-east flowing Tsuchiya Jets (called SSCC), which flow underneath the nitrate-rich surface waters on and south of the equator (Figure 1). This water mass is thought to originate in the Coral Sea, such that the associated δ15N signal was initially hypothesized to reflect the transport of low-δ15N nitrate from the remineralization of newly fixed N in the western South Pacific by the jets (Rafter et al., 2012). However, as pointed out by Yoshikawa et al. (2015), the δ15NNO3 in the Coral Sea thermocline is not sufficiently low to explain the signal observed in the Tsuchiya Jets off the equator. We submit that the low δ15NNO3 transported in the jets originates from the vertical flux of low δ15N generated from the partial consumption of nitrate in waters upwelling at and to the south of the equator (see Rafter & Sigman, 2016). By comparison, the Northern Subsurface Counter Current (NSCC) may have a higher δ15NNO3 of ≥6‰ (Rafter et al., 2012) because it does not transit the Pacific basin underneath nitrate-rich waters. Sinking organic matter δ15N is also higher north of the equator (Altabet et al., 1999), suggesting that nitrification of this material may also elevate the δ15NNO3 of the NSCC.
5.3 Quantifying the Northern and Southern Hemisphere Sources of the EUC
The eastward-flowing, thermocline-trapped EUC is the source water of the equatorial upwelling system (Dugdale et al., 2002; Rafter & Sigman, 2016; Wyrtki, 1981). The EUC source region is composed of northern and southern-sourced waters (Butt & Lindstrom, 1994; Fine et al., 1994; Lindstrom et al., 1987; Melet et al., 2010; Tsuchiya et al., 1989; Ueki et al., 2003). Estimated contributions of each hemisphere to the EUC vary among studies and investigated latitude, with some studies arguing for a dominance of the northern-sourced MC (Fine et al., 1994) or southern attributions via the NGCUC and NICU (Toggweiler et al., 1991; Tsuchiya et al., 1989). Attributions are also divergent among modeling studies, ranging from a roughly balanced input of the southern and northern hemispheres (Izumo et al., 2002) to a net dominance of southern LLWBCs (Blanke & Raynaud, 1997; Fukumori et al., 2004; Grenier et al., 2011; Rodgers et al., 2003).
To a first approximation, nitrate isotope profiles measured at stations along the EUC (Rafter & Sigman, 2016) suggest a disproportionate contribution of southern hemisphere water masses to EUC nitrate (Figure 5). The N and O isotopic compositions of nitrate at three longitudinally distinct sections of the EUC are confined to a restricted range of 6.8–7.5‰ for δ15NNO3 and 2.5–3.2‰ for δ18ONO3, respectively. Compared to our stations in the southern and northern WEP, nitrate δ15N and δ18O values in the EUC appear more closely aligned with values observed at southern stations, particularly in lower density layers in the upper EUC (24–26σθ; Figure 5b): southern profiles generally show a similar range in δ18ONO3 and somewhat higher δ15NNO3 values relative to the upper EUC, whereas profiles at the northern sites off Mindanao indicate greater δ18O values and markedly lower δ15N values than at comparable densities in the EUC (Figure 5b). At the SAMW-density interval (26.5–27.1σθ), southern and northern stations are less distinct with respect to nitrate isotope ratios, with δ15NNO3 and δ18ONO3 in the EUC being slightly higher than PNG and NH, and similar to lower than values off Mindanao (Figure 5c). Considering that the EUC may issue from diapycnal mixing of thermocline and intermediate waters, however, the nearly uniform δ15NNO3 and δ18ONO3 in the EUC could arguably be explained as a mixture of upper and lower density levels of southern stations (Figures 5b and 5c).
To further interrogate the provenance of tracers in the EUC, profiles of salinity, temperature, nitrate, silicic acid, oxygen, δ15NNO3, δ18ONO3, and δ13CDIC measured at 0°N/165°E (Rafter et al., 2012; Rafter & Sigman, 2016) were used to calculate the mixture of LLWBC end-members and respective depth intervals that best account for the properties observed in the shallow and midlayer of the EUC. Individual values for each end-member are summarized in supporting information Table S1 and the results of the mixing model are illustrated in Figure 7, where the contribution of source waters to the upper and lower EUC is given in relation to source area and density layer, with relative proportions adding up to 1. Considering only salinity and temperature in the mixing model, the optimal solution diagnoses significant contributions to the lower and upper EUC from both hemispheres (Figure 7a). Putative contributions from Mindanao decrease with the addition of the isotopic tracers and nitrate to the mixing model (Figure 7b). The upper EUC then derives predominantly from the NGCUC, while the lower EUC shows significant contributions from both the NICU and NGCUC. Finally, considering silicic acid and oxygen concentrations in addition to the other tracers renders a solution wherein the NICU contributes dominantly to both the upper and lower EUC (Figure 7c). The mixing model exercise thus illustrates that the distribution of combined physical and biogeochemical tracers in the EUC is best explained by a dominance of southern WEP waters, albeit, with uncertainties regarding the differential contributions of NGCUC and NICU.
Relative contributions of different LLWBC end-members and respective density intervals to the upper EUC (24–25.5σθ, red) and lower EUC (25.5–26σθ, blue). Abbreviations on the x axis indicate the three source regions (Vitiaz Strait (V), New Hanover (N), Mindanao (M)) further divided into four density layers: upper EUC layer (U; 24–25.5σθ), lower EUC layer (L; 25.5–26σθ), deep EUC layer (D; 26–26.5σθ), and SAMW layer (S; 26.5–27.1σθ) according to the end-members listed in supporting information Table S1. Mixing calculation include (a) salinity and temperature only, (b) salinity, temperature, nitrate, and isotope tracers, and (c) salinity, temperature, isotope tracers, oxygen, and nutrients (nitrate, silicic acid). Solution box plots were computed from respective ensembles of solutions (see text). All boxes have a notch at the median solution value, the box covers the interquartile range, and the whiskers indicate 95% coverage intervals (e.g., 2.5–97.5 percentile).
Regarding the specific provenance of nitrate in the EUC, the mixing model similarly indicates a disproportionate contribution from southern boundary currents (≥70%), with the NICU accounting for ≥50% of total EUC nitrate (supporting information Table S2). Among density intervals, the model diagnoses that nitrate at intermediate depths (26.5–27.1) contributes most to lower EUC nitrate, whereas all density intervals considered contribute comparably to upper EUC nitrate (supporting information Table S3). In all, this analysis suggests that EUC nutrients originate predominantly from southern hemisphere boundary currents. Southern hemispheric processes that begin with Southern Ocean overturning, and include lower latitude organic matter remineralization, mixing with the eastern ODZ, and N2 fixation, thus exert an important influence on the biogeochemical properties of waters upwelled at the equatorial Pacific.
6 Summary
We examined the N and O isotopic composition of nitrate and the distribution of complementary biogeochemical tracers to elucidate the hydrography of the WEP and the biogeochemistry and evolution of water masses that feed the EUC, which has implications for understanding controls on the productivity of the tropical Pacific. Tracer distributions reveal remarkably distinct biogeochemical features at stations north and south of the equator.
Partial assimilation along the equatorial upwelling system results in the export of high-δ15N organic matter at the northern edge of the South Pacific gyre, the remineralization of which is manifested as elevated-δ15NNO3 in SPTW advected to the western boundary—a high δ15NNO3 signal that may overprint the influence of newly fixed nitrogen near the western boundary. The remineralization of high-δ15N organic matter is further evident in underlying SAMW-density waters, which have a higher δ15NNO3 and lower δ18ONO3 (and thus higher Δ(15–18)) than the SAMW end-member at 50°S. Additionally, the relatively modest N* at intermediate depths further points to restricted direct lateral advection of denitrified waters from the Eastern Tropical South Pacific, and to the remineralization of high-δ15N organic matter as the dominant contributor of elevated δ15NNO3 at intermediate and thermocline depths. Differences in tracer distributions in the southern WEP thermocline off PNG compared to NH reflect the influence of waters ventilated further south, and are also consistent with diapycnal mixing of intermediate water masses. In contrast, lower δ15NNO3 and lower Δ(15–18) in the thermocline-depth waters off Mindanao show the contribution of nitrate from the remineralization of newly fixed nitrogen in the North Pacific. The relatively elevated δ18ONO3 in intermediate waters below suggests greater direct lateral contribution of nitrate from the eastern ODZ compared to the southern WEP. These strong hemispheric differences in WEP biogeochemical tracers allow us to place constraints on the source of EUC waters, suggesting that most EUC waters—and nutrients therein—derive from southern hemisphere sources.
Acknowledgments
We gratefully acknowledge the captain and crew of the RV Sonne SO-228 for contributing to a successful expedition, funded by the German Ministry of Education and Research (BMBF; grant 03G0228A, EISPAC). The Natural Science and Engineering Research Council (NSERC) Canada provided funding for this project, in particular to NL through the NSERC CREATE Transatlantic Ocean System Science and Technology (TOSST) grant. Further support for this work came from the United States National Science Foundation grant OCE-1435002 to J.G. Data are available on https://doi.org/10.1594/PANGAEA.885589.
utilization in the equatorial Pacific
along the US GEOTRACES Eastern Pacific Zonal Transect (GP16) yield insights into nitrate supply, remineralization, and water mass transport. Marine Chemistry, in press. 
