We report measurements of the nitrogen isotopic composition of nitrate (the δ15N of NO3−) across the equatorial Pacific, for zonal transects from 165°E to 95°W and meridional transects across 95° and 110°W. The δ15N of NO3− is similar in the equatorial thermocline (≈100 m) and intermediate depth waters (≈150 to 600 m), averaging (7.1 ± 0.3)‰ and (7.1 ± 0.1)‰, respectively. These values are more than 2‰ higher than subthermocline waters of the Southern and Atlantic Oceans and are ≈1‰ higher than putative source waters in the high latitude South Pacific (Subantarctic Mode Water, SAMW). The combined constraints of nitrate concentration and δ15N of NO3−in the equatorial Pacific require (1) lateral exchange between the high-latitude source waters and the zones of denitrification in the eastern tropical Pacific and (2) the accumulation of remineralized nutrients at depth. The zonal uniformity of the subsurface equatorial Pacificδ15N of NO3− indicates rapid transport within the equatorial zone, which works to homogenize the δ15N of NO3− across the Pacific basin. Against this backdrop of high δ15N of NO3−in the tropical Pacific, we find a discrete off-equatorial core of lowerδ15N of NO3− (5.5 ± 0.3)‰ concentrated at 5°S and 150 to 200 m along the 110° and 95°W transects and in apparent association with the Southern Subsurface Counter Current (SSCC). We propose that the remineralized products of nitrogen fixation, at the source of the SSCC in the western south Pacific, are the origin of the low δ15N of NO3− in these waters.
 The 15N/14N of nitrate is sensitive to important biogeochemical fluxes [e.g., Knapp et al., 2005] and therefore provides diagnostic constraints on nutrient budgets in the global ocean. It also provides information on the origin and history of water masses [e.g., Sigman et al., 2000], identifying processes that are not immediately evident from standard hydrographic measurements such as temperature, salinity, and nutrient concentrations. These multiple applications for the 15N/14N of nitrate (hereafter referred to as “δ15N of NO3−” where δ15N is expressed relative to atmospheric N2 and equals R (sample)/R (N2 of the atmosphere) – 1, with the isotope ratio R = [15N]/[14N]) are being pursued in both low- and high-latitude regions of the world's oceans [Sigman et al., 2000; Voss et al., 2001; Sigman et al., 2005; DiFiore et al., 2006, 2009]. Our intention here is to highlight these various applications with a new set of δ15N of NO3− measurements spanning the tropical Pacific.
 Within the tropical Pacific, a number of biogeochemically important processes are imprinted on the δ15N of NO3−. For example, upwelling along the equator and on the Pacific's eastern margin delivers abundant nutrients to the surface ocean, but the entire NO3− pool is not utilized (Figure 1a). This allows for 14N of NO3− to be preferentially assimilated, leading to an increase in the δ15N of the residual NO3− in the surface ocean and producing sinking material with a lower δ15N [Altabet and Francois, 1994]. As NO3− consumption proceeds to completion, which occurs relatively close to the equatorial and coastal zones of upwelling, the integrated δ15N of the N exported to depth converges on the δ15N of the NO3− upwelled from below.
 N fixation, the fixation of N2 into organic matter by a discrete set of ocean phytoplankton, is an additional influence on NO3−isotopic composition, opposing the denitrification-driven loss of NO3−. N fixation introduces low δ15N of organic matter to the surface ocean and, upon sinking to depth, produces NO3− with a low δ15N, as is apparent in the North Atlantic [Knapp et al., 2005, 2008] and also the Indo-Pacific [Brandes et al., 1998; Casciotti et al., 2008]. Because the δ15N of NO3− is sensitive to these transformations (discussed further below), NO3− influenced by these biological processes (and the organic matter produced from this NO3−) is isotopically “tagged,” providing a view into the history of water masses.
 Here we use measurements of tropical Pacific δ15N of NO3−along with standard oceanographic measurements to characterize the water masses of this region based on near-surface circulatory features and their intermediate (>200 m) and deep-sea (>1000 m) high-latitude source waters (Figure 3). These δ15N of NO3−measurements are then interpreted within the context of known and concurrently measured circulation to provide additional information on the patterns of Pacific Ocean subsurface flow and their interaction with Pacific Ocean biogeochemistry. Of particular interest are the intermediate-depth South Pacific waters, which are on the same density level as the eastern tropical Pacific ODZs. It has been suggested that these intermediate waters (specifically, the Subantarctic Mode Water, SAMW) are the major source of nutrients to the equatorial Pacific thermocline [Toggweiler et al., 1991; Sarmiento et al., 2004; Palter et al., 2010], even though there is a considerable change in high to low latitude NO3− characteristics (Figures 1b and 1c) at the density of this water mass. The δ15N of NO3− data reported here along with published analyses from the eastern South Pacific and Subantarctic Pacific impose significant new constraints on the subsurface pathways and apparent modification of SAMW NO3− between the high and low latitude South Pacific.
 While understanding the distribution of nutrients is important for studies of surface ocean productivity, the complete characterization of subsurface Pacific circulation is also necessary for understanding aspects of global climate, such as the global ocean heat balance [Gordon, 1986], the upper ocean conditions that give rise to the El Niño Southern Oscillation [Gu and Philander, 1997], and even ice age cycling [Spero and Lea, 2002; P. A. Rafter and C. D. Charles, Pleistocene equatorial Pacific dynamics described by zonal gradients in nitrogen isotopes, submitted to Nature Geoscience, 2011]. In a more general sense, these measurements provide an essential measure of environmental conditions for regional ecological studies [e.g., Graham et al., 2007; Newsome et al., 2006] as well as sediment nitrogen isotope studies of past biogeochemical change [Farrell et al., 1995; Ganeshram et al., 1995; Robinson et al., 2009].
 We organize this work by first using our measurements to describe the NO3− characteristics of the tropical Pacific in section 3. In this section, we also discuss the NO3− properties of the Eastern Tropical South Pacific and Subantarctic Pacific observed in previous studies. In section 4, we describe some of the physical and biogeochemical processes that can explain the surface to the seafloor characteristics of equatorial Pacific NO3−, and the implications of these findings are discussed in section 5. Among other results discussed here, the measurements of tropical pacific δ15N of NO3− demonstrate the capacity of tropical Pacific subsurface circulation to transport the geochemical signals of denitrification and nitrogen fixation westward and eastward, respectively.
2. Materials and Methods
 Tropical Pacific water samples were collected during the National Oceanic and Atmospheric Administration's (NOAA) routine maintenance of the Tropical-Atmosphere-Ocean (TAO) array. Collection of water by the ship's rosette was accompanied by Conductivity-Temperature-Depth (CTD) measurements and a dissolved oxygen sensor package on some occasions. Each ship has a hull-mounted Acoustic Doppler Current Profiler (ADCP), to identify direction and speed of circulatory features from 30 m to a maximum depth of 700 m.
 Water sampling locations and depths (diamonds in Figure 1) were as established by the TAO program. In October and November of 2003, water samples were collected on board the NOAA ship RV Ronald H. Brown for meridional transects at 95°W and 110°W. The zonal transects are constructed from samples retrieved on the RVs Ka'imimoana and Ronald H. Brown at 0°N (between 165°E and 125°W) in fall 2004 and at 1°N (between 165°E and 95°W) in late summer/fall 2005, respectively. Differences in nutrient concentrations are small between 0°N and 1°N, especially below the mixed layer, but the EUC may extend over a slightly smaller depth range at 1°N [Firing et al., 1983; Johnson et al., 2002]. The “gyre” stations at 7°N, 155°W and 8°S, 155°W were sampled in October and December 2005, respectively.
 All sample bottles (60 ml Nalgene HDPE) were acid-washed and rinsed with DI water before shipping. At sea, sample bottles were rinsed three times with sample water, filled to a volume of ≈50 ml, frozen on site, and shipped overnight to laboratories at the Scripps Institution of Oceanography or Princeton University. The samples from 2003 and 2004 were analyzed for nitrate concentration [NO3−] at Princeton University by reducing NO3− (and nitrite, NO2−, if present) to NO using a vanadium(III) reagent, followed by chemiluminescence measurement of NO [Braman and Hendrix, 1989]. The full suite of nutrient analyses (NO3−, nitrite, phosphate and silicate) were performed on the summer/fall 2005 water samples at the UCSB Marine Science Institute Analytical Lab and showed very low (<0.1 μmol/L) to undetectable nitrite (NO2−) below the mixed layer (below the depth where temperature is ≈ 0.5°C different than sea surface temperature). All samples were analyzed for the 15N/14N ratio of NO3− via the “Denitrifier Method” of Sigman et al.  and Casciotti et al. , with measurement error ± 0.3‰. Because the presence of nitrite could result in an underestimation of the δ15N of NO3− when using the Denitrifier Method [Casciotti and McIlvin, 2007], we removed NO2− in potentially impacted samples using the methods of Granger et al.  and Granger and Sigman .
3.1. General Distribution of Tropical Pacific δ15N of NO3−
 There is a broad depth range along the equatorial Pacific (from ≈150 to at least 600 m) where weighted average δ15N of NO3− is (7.1 ± 0.1)‰ (Figure 4a), more than 2‰ higher than comparable depths in the North Atlantic and the Antarctic [Sigman et al., 2000; DiFiore et al., 2006; Knapp et al., 2008]. While the δ15N of NO3− decreases below ≈700 m in the equatorial Pacific, the relative δ15N elevation persists, with a δ15N of > 5.5‰ at 3000 m. Our stations at 7°N and 8°S and 155°W (Figure 4d) are essentially “upstream” of the equatorial waters [e.g., Fine et al., 1987; McCreary and Lu, 1994], and show that the subsurface δ15N of NO3− is > 6‰ down to at least 1000 m on both sides of the equator.
 The elevated δ15N of NO3− measured throughout the tropical Pacific at ≈400 m is consistent with the deficit of NO3− suggested by the N* parameter. Assuming that NO3− uptake and regeneration follows the 16:1 ratio derived by Redfield , measured NO3− to phosphate can be normalized as “N*” (from Gruber and Sarmiento  and simplified here as N* = [NO3−] – 16[PO43−], where [PO43−] is the concentration of phosphate in units of μM). The negative N* generated from denitrification in the eastern tropical Pacific ODZs (as well as continental margin sediments) communicates with the broader Pacific through the generalized intermediate-depth circulation (Figure 1c), resulting in negative N* across much of the Pacific basin. The widespread distribution of negative N* and the prevalence of elevated δ15N of NO3− in the equatorial Pacific at the same depths illustrates the broad geographic scale of the influence of the denitrification process on the Pacific NO3− pool.
 Within this large and apparently basin-wide distribution of elevatedδ15N of NO3−, a few lower values (6.5–6.8‰) are seen in the eastern equatorial Pacific between 150 and 200 m (Figures 4 and 5). This is a surprising result, in that this represents a lateral decrease in δ15N of NO3− toward the longitudes of the ODZs. However, the meridional transects crossing the equator at 95°W and 110°W reveal that this slightly lower eastern equatorial δ15N of NO3− is an equatorward extension of even lower δ15N of NO3− (5.5 ± 0.3)‰ at 5°S and within the 100–200 m depth range (Figures 4 and 5). ADCP data (shading in Figure 5) shows that this low δ15N of NO3− water at 5°S appears to be associated with the Southern Subsurface Counter Current (SSCC) at both meridional transects (Figures 4 and 5). While not measured concurrently, N* values from the World Ocean Circulation Experiment (WOCE) do not show a distinct anomaly for these waters, with respect to surrounding water masses (Figure 5c).
3.2. Characterizing the Tropical Pacific Water Masses
 The δ15N of NO3− measurements can be considered in the context of the broader Pacific Ocean circulation. First, we relate the δ15N of NO3−(and other parameters) to the near-surface currents (upper 200 to 300 m) using concurrent and historical ADCP data based on the work ofJohnson et al.  (Figures 3 and 5 and Table 1).
Table 1. Average Values of Select Parameters for Water Masses in the Equatorial Pacific
The Equatorial Under Current (EUC) and Northern and Southern Subsurface Counter Currents (NSCC and SSCC) were identified using climatological locations and concurrent ADCP data. The intermediate and deeper water masses (potential density > 26.5) were identified based on potential density and salinity characteristics of their putative South Pacific source waters seen in Figure 3 (Subantarctic Mode Water and Antarctic Intermediate Water: SAMW and AAIW).
Average values found in the “Av.” row and standard deviation in the row labeled “SD.”
Dissolved oxygen was not measured at all longitudes.
 We define the Equatorial Under Current (EUC) as all eastward-flowing water at thermocline depth (dashed line on equator inFigure 1a) at a depth of ≈75 m in the east and 150 m in the west and within 1° of the equator. The weighted average δ15N of NO3− for the EUC is (7.1 ± 0.3)‰ (averaged below the depths where algal NO3− assimilation increases the δ15N of NO3−). The other eastward subsurface jets—the Northern and Southern Subsurface Counter Currents (NSCC and SSCC)—are identified as eastward flow at their climatological positions (dotted lines in Figure 1a; from ≈4° to 6°N and 4° to 6°S and 100–300 m depth east to west), in which the δ15N of NO3− averages (6.8 ± 0.4)‰ and (5.5 ± 0.3)‰, respectively. The waters upwelled at the equatorial Pacific surface—the EUC—are ultimately ventilated by Subtropical Mode Water (STMW), which is carried within the subtropical gyres of the North and South Pacific [e.g., Tsuchiya et al., 1989; Fine et al., 1987] within the potential density range of 25.0 to ≈26.5 (roughly 50 to 200 m depth along the equator). Our stations at 7°N and 8°S, 155°W are within the currents that essentially feed the EUC (the NEC and SEC, respectively). The 7°N station (NEC) δ15N of NO3− averages (7.5 ± 0.1)‰ (n = 2), but there are no samples within this depth range for the 8°S (SEC) station.
 Deeper waters along the equator are characterized on the basis of potential density and salinity in order to identify their source (labeled to the right of Figure 3b). The intermediate-depth equatorial Pacific waters are primarily ventilated in the South Pacific [Fine et al., 1994] and are labeled for their respective ventilation sources, such as “EqPac(SAMW),” which is ventilated by Subantarctic Mode Water along the 26.5 to 27.1 isopycnals (approximately 200 to 600 m at the equator). Likewise, “EqPac(AAIW)” is a salinity minimum water below EqPac(SAMW) that is ventilated by Antarctic Intermediate Water [Hanawa and Talley, 2001; Sloyan et al., 2003]. We do not include the North Pacific Intermediate Water (NPIW) as a source for intermediate equatorial waters because of its limited input to equatorial waters at densities deeper than 26.5 [e.g., Fine et al., 1994]. More recent work also affirms the dominance of the south Pacific on water mass tracers at intermediate depths [Bostock et al., 2010]. However, NPIW most likely contributes to intermediate waters at the North Pacific subtropical gyre station (7°N, 155°W), which averages (7.8 ± 0.3)‰ between 26.5 to 27.1 potential density (Figure 1a). At these depths, the SEC station (8°S, 155°W) is ventilated by SAMW, and δ15N of NO3− averages (7.3 ± 0.5)‰ (Table 1). From 1500 to 3000 m, Pacific Deep Water (PDW) is a mixture of North Atlantic Deep Water, Upper and Lower Circumpolar Deep Waters.
3.3. Characterizing the High Latitude and ODZ Waters of the South Pacific
 To allow for a direct comparison between EqPac(SAMW) (defined above) and SAMW, we compile average values for SAMW and AAIW from the Southern Ocean (white circles in Figure 1) in Table 2 [Sigman et al., 2000; DiFiore et al., 2006]. Recent work using ARGO autonomous profiling floats illustrates that only SAMW formed in the southeast Pacific reaches the lower latitudes, with SAMW from the central Pacific remaining within the gyre circulation [Herraiz-Borreguero and Rintoul, 2011]. Therefore, we use measurements from the southeast Pacific for our estimates [Sigman et al., 2000]. To capture the characteristics of the newly formed SAMW, we average measurements made within 5° latitude north of the Subantarctic Front and within the 26.5–27.1 kg/m3 potential density range (as defined by Hanawa and Talley ) and excluded mixed layer data (where δ15N of NO3− is seasonally elevated by summertime NO3− assimilation). The mean δ15N of NO3− of SAMW (SE Pacific) is (6.3 ± 0.2)‰ (Table 2), an elevation of ≈1‰ above subsurface Antarctic zone water, which results from the incorporation of NO3− that has been partially consumed by phytoplankton assimilation in Southern Ocean surface waters [Sigman et al., 1999; 2000] and presumably also some admixture of lower latitude thermocline waters influenced by denitrification. AAIW is defined as salinity minimum water (S < 34.5) below SAMW and has a potential density between 27.1 and 27.3 [Hanawa and Talley, 2001].
Table 2. Average Values of Select Parameters for Subantarctic Mode Water (SAMW), Antarctic Intermediate Water (AAIW), and the Oxygen Deficient Zone of the Eastern Tropical South Pacific (N* < −15)
 In addition to these high southern latitude water masses, we estimate representative values from ETSP ODZ (pink squares in Figure 1). The ETSP values (from the Peruvian and Chilean Margins [Liu, 1979; De Pol Holz et al., 2009]) are averages of low oxygen waters with N* less than −15 μM (by the definition used here) and within the density range of SAMW.
3.4. South Pacific–Equatorial Pacific Connections at Intermediate Depths
 We compare the nitrate-related characteristics of the equatorial Pacific (both averaged along the equator and at 8°S, 155°W), Subantarctic, and Eastern Tropical South Pacific waters inFigure 6 (δ15N of NO3−, [NO3−], and N*; all within the density range of 26.5 to 27.1). As is expected from Figures 1b and 1c, the nitrate concentration and N* of SAMW are considerably different than the equatorial Pacific in this density range, and the δ15N of NO3− adds an additional constraint: EqPac(SAMW) is ≈1‰ higher than SAMW (Figure 6). Furthermore, the south Pacific subtropical gyre δ15N of NO3−(SEC) is essentially the same as the equatorial Pacific, indicating that the modification of SAMW-derived waters takes place before arrival at the equator. These few SEC measurements are important here because the same depth and density north of the equator (at our NEC station) are similarly elevated inδ15N of NO3− (Figure 4d) and negative in N*. Without the SEC measurements, it could be argued that a major component of the equatorial Pacific NO3− pool at these depths (≈200 to 600 m) is derived from the NEC, but our data are consistent with a primarily South Pacific origin [Fine et al., 1994]. The differences between high and low latitude waters of this density range also extend to temperature and salinity, with the lower latitude waters being warmer and more saline (Figure S2).
 In order to understand how equatorial Pacific NO3− acquires these characteristics, one approach is to consider the processes by which SAMW may be modified into the NO3− characteristics of the equatorial Pacific subsurface. Simple calculations can account for the three processes by which SAMW can be modified to generate the properties of equatorial Pacific waters of similar density: lateral exchange (“mixing”), the addition of NO3− through remineralization of organic matter, and denitrification.
 Our first approach focuses on the [NO3−], δ15N of NO3−, and N* for mixing between SAMW and the region of denitrification in the ETSP ODZ (gray lines in Figure 6). The simplest formulation of this process is given by the equation
where subscript “mix” indicates the values for the mixed product, the subscripts “a” and “b” indicate the values for the mixing end-members, “Xa” is the proportion of water mass mixing of end-member “a” in the mixture, “C” is the nitrate concentration. This end-member mixing calculation is narrowly focused on explaining the equatorial Pacific characteristics, given these two end-members; the origin of the characteristics of these end-members is not addressed here.
 Lateral exchange between these water masses (broad gray lines) produces a convex curve in Figure 6 because the mixture δ15N of NO3−is biased toward the mixing end-member with higher [NO3−] (the ETSP ODZ). As can be seen in Figures 6a and 6b, this simple mixing scheme between SAMW and the ETSP ODZ can produce the same δ15N of NO3− and N* of the equatorial Pacific intermediate waters, but it cannot explain the higher equatorial Pacific [NO3−] (Figure 6a).
 The second approach is to simulate the addition of remineralized nutrients while assuming a regenerated nitrate: phosphate = 16 and no isotopic fractionation during remineralization. This is demonstrated in Figure 6 for a δ15N of NO3− addition of 6.3‰ (dashed blue arrow) and 7.1‰ (dashed green arrow) based on SAMW and EqPac(SAMW) values. The starting point along the mixing line was chosen to produce a final δ15N of NO3− equal to the EqPac(SAMW). In this scheme, it is apparent that two processes working in tandem—mixing with the ETSP ODZ and the addition of remineralized nitrate—can modify SAMW to match the equatorial Pacific waters (both along the equator and at 8°S, 155°W) of similar density. The arrow representing 7.1‰ in Figures 6b and 6d is not seen because, with the assumption of regenerated nitrate to phosphate ratio equaling 16:1, there is no change in N* or δ15N of NO3−. The arrow representing the addition of 6.3‰ nitrate is slightly to the left of the EqPac(SAMW) marker in Figures 6c and 6d. Variations in the nutrient uptake and regenerated nitrate to phosphate ratio (as have been suggested by Mills and Arrigo ) would alter the N* of the “regenerated nitrate” addition, angling the arrows toward the left in Figures 6c and 6d in the case of regenerated N: P ratio lower than 16. If so, this might require a change in the assumption of the mean δ15N of the remineralized NO3−and / or the proportions of the end-members in the water mass mixing calculation.
 We also provide the approximate trajectories in the nitrate characteristics of SAMW-derived water were it to undergo denitrification (red lines). Three different sets of assumptions are shown: a Rayleigh model (closed-system) assuming a denitrification isotope effect (ε) of −25‰ (equation (1)) [Brandes et al., 1998; Sigman et al., 2003]; a steady state model (open system) with a much smaller isotope effect of −12.5‰ (equation (2)); and a final calculation using a 0‰ isotope effect (equation (3)) meant to represent sedimentary denitrification. The equations are the Rayleigh fractionation,
and the Steady State fractionation,
where “f” is the fraction of NO3− remaining and “ε” is the isotope effect, defined here as 15k/14k - 1 (15k and 14k are the rate coefficients for the reactions of NO3− with 15N and NO3− with 14N, respectively). For calculating changes to NO3− and phosphate concentrations during denitrification, we assume that for every 104 moles of NO3− respired, 1 mol of phosphate is regenerated [Anderson, 1995]. In the sedimentary denitrification model—denit. (3)—there is no change to δ15N of NO3− in this water mass (because all NO3− is consumed during denitrification), but [NO3−] and N* decrease [Brandes and Devol, 2002].
 It is readily apparent that none of the models of “denitrified” SAMW match the equatorial Pacific waters at similar density range, especially with respect to [NO3−] because these denitrification models reduce the already low SAMW values (Figure 6).
 The NO3− characteristics of the tropical Pacific (elevated δ15N of NO3−and negative N*) demand some degree of mixing between this well-oxygenated region and the Oxygen Deficient Zones of the eastern tropical Pacific. Balancing this constraint with the additional requirement of nutrient sources associated with Subantarctic Mode Water [Toggweiler et al., 1991; Sarmiento et al., 2004; Palter et al., 2010] requires the admixture of water with higher δ15N of NO3− and more negative N* (Figures 1 and 6). Below, we propose where and how SAMW-derived waters are modified and how these waters influence the equatorial Pacific thermocline. This proposal is followed by a more general consideration of the nitrate properties of the entire equatorial Pacific waters (from the EUC to the abyss).
4.1. Modifying SAMW Nitrate
 Our simple calculations (Figure 6) illustrate that mixing between SAMW and the ETSP ODZ along with the addition of remineralized nutrients can explain the apparent change in SAMW nitrate characteristics between the high and low latitude waters (see Figures 1b and 1c). While additional observations would help delineate the regional expression of these processes, the station within the South Pacific subtropical gyre (8° S, 155° W) indicates that the modification of SAMW occurs before these waters reach the equator, confining this alteration to the extra-tropical South Pacific. In fact, recent observations confirm the Chilean Margin as a region where SAMW mixes with the warm, salty, oxygen-poor waters of the eastern tropical Pacific [Johnson and McTaggart, 2010]. This eddy-diffusive mixing takes place in the southeast Pacific along the boundary between the lower nitrate / higher N* SAMW and the higher nitrate / more negative N* waters of the lower latitudes (stars inFigures 1b and 1c). Measurements of δ15N of NO3− from the Chilean Margin (black triangles in Figure 1 [De Pol-Holz et al., 2009]) also appear to fall on the two end-member mixing trend between SAMW and the equatorial Pacific in a plot ofδ15N of NO3− versus N* (Figure 6d).
 When [NO3−] is plotted versus δ15N of NO3− (Figure 6c), some Chilean Margin nitrate measurements fall to the left of the mixing line (resulting from nitrate-loss during active denitrification), but most are to the right and downward of the mixing line. This trend is the expected effect of the accumulation of regenerated nitrate at depth—a process that could substantially increase [NO3−] while having a weak effect on δ15N of NO3− and N*. Assuming SAMW is the ultimate source of low latitude surface ocean nitrate, the addition of regenerated nutrients increases the original (SAMW) nitrate concentration by ≈ 30%, with an unknown effect on the δ15N of NO3−. With the above assumptions on remineralization, the near-complete consumption of nitrate in the surface layer leads to a correspondence between theδ15N of NO3− brought into the tropical surface and that of the N sinking out of the surface ocean [Sigman et al., 2000].
 Upwelling along the eastern margins and the equator brings nutrient-bearing water to the surface. These nutrients are eventually consumed from the upwelled water, and are transferred back to the shallow interior via remineralization. This process is frequently described as “nutrient trapping” [Najjar et al., 1992]. While this equatorial/tropical nutrient trapping is certainly important, there are other mechanisms for concentrating nutrients in SAMW. In particular, in the eastern South Pacific, nutrient bearing surface water from the Southern Ocean impinges equatorward, transported by the South Pacific subtropical gyre (Figure 1c). These surface nutrients are consumed along the South American coast, combining with the coastal upwelling to drive high productivity. The conversion of these eastern South Pacific surface waters from nutrient-rich to nutrient-poor represents a second source of regenerated nutrient to the underlying SAMW-derived waters. In any case, while it is apparent that nutrients have already accumulated to substantial levels in the SEC, it is not yet possible to attribute relative importance to these distinct mechanisms of accumulating nutrients in Pacific mid-depth waters: it could result either through mixing with tropical mid-depth waters that have abundant trapped nutrients or through the addition of regenerated nutrients in the temperate South Pacific, or both.
 This question aside, the pathway these modified high latitude intermediate waters take to the equatorial Pacific is likely within the SEC/subtropical gyre to the western boundary [Tsuchiya and Talley, 1996], where there are two main routes to the equator: north of the New Ireland/Solomon Islands [Butt and Lindstrom, 1994] and slightly toward the west, off the Papua New Guinea coast [Tsuchiya et al., 1989; Cravatte et al., 2011]. While the SEC feeds both of these equatorward flows (see arrows in Figure 1b) and they are similarly saline, water on the New Ireland pathway is from the lower latitude SEC and has higher nutrients and lower oxygen [Tsuchiya et al., 1989]. Using the nutrient data of Tsuchiya et al. , we calculate a slight difference in N* between these water masses as well (more negative off New Ireland from the lower latitude SEC flow), which suggests the lower latitude SEC water taking the New Ireland pathway is more influenced by mixing with the ETSP ODZ. The nutrient and dissolved oxygen differences between these two equatorward pathways of intermediate-depth SEC waters is intriguing and suggests that not all SAMW-derived waters approaching the equator are equally modified by the process we have highlighted above. Alternations in the strength of these separate equatorward flows may play a role in equatorial Pacific nitrate variability. Regardless of the pathway to the equator, we find that much of the nitrate characteristics of the SAMW-derived waters can be established in the Southern Hemisphere by interacting with the ETSP ODZ.
4.2. Source of Equatorial Pacific Thermocline Waters
 The δ15N of NO3− in the equatorial Pacific thermocline (≈100 to 150 m depth, Table 1) is not statistically different from that in the deeper “EqPac(SAMW)” (Table 1). This similarity exists despite different histories for each water mass; the equatorial thermocline is ventilated by both the South and North Pacific STMW [Tsuchiya et al., 1989; Fine et al., 1987] and is warmer and more saline, with higher [O2] and lower NO3−concentrations than the deeper, SAMW-derived waters (Table 1). For SAMW to deliver NO3−to the lower latitude thermocline, there must be diapycnal mixing between the essentially nutrient-deplete STMW and the deeper SAMW-derived waters (as has also been suggested byToggweiler et al. ). This diapycnal mixing would increase STMW NO3− concentrations and give these waters a δ15N of NO3− equal to the deeper SAMW. Recent work suggests that substantial STMW modification occurs via diapycnal mixing in the Solomon Sea [Melet et al., 2011]. Mixing with STMW also explains the observed warming and increased salinity of EqPac(SAMW) relative to preformed SAMW (Figure S2). Further, a similar process likely applies in the North Pacific, where subtropical gyre waters also show elevated δ15N of NO3− in the shallow subsurface (Figure 4d).
 Equatorial Pacific thermocline waters take a similar western boundary pathway as the SAMW-depth waters described above. However, there is also a North Pacific source to the equatorial thermocline near the western boundary [Fine et al., 1994] along with contributions from both hemispheres in the central and eastern equatorial Pacific via the lower branch of the “Sub Tropical Cell” (STC: white dashed arrows toward the equator in Figure 1) [Fine et al., 1994; Johnson and McPhaden, 1999] (the upper branch of the cell is the upwelled equatorial water). That the equatorial thermocline waters are relatively homogenous despite this variety of sources suggests that the lower branch of the STC consists of waters with similar nitrate characteristics from north and south of the equator.
4.3. Downward Propagation of the Isotopic Signal by Sinking Organic Matter
 Equatorial upwelling delivers NO3− with elevated δ15N of NO3− from the EUC to the surface, where organic matter with similarly elevated δ15N is produced. The NO3− upwelled along the equator is transported meridionally as a part of the upper branch of the STC [Johnson, 2001] and this surface NO3− is completely consumed within a few degrees of the equator. While the δ15N of organic matter produced in these NO3−-replete waters will be lower than the upwelledδ15N of NO3− [Altabet and Francois, 1994], the meridionally integrated δ15N of this organic matter will carry the elevated δ15N of the EUC to deeper waters. The remineralization of this organic matter at depth introduces elevated δ15N of NO3− throughout the water column, including deeper waters (>600 m). The properties of deeper equatorial waters supports this interpretation, considering that AAIW in the equatorial zone has higher δ15N of NO3−, higher [NO3−], and lower [O2] than the source waters defined in the Subantarctic Zone (Tables 1 and 2). This sinking flux of high-δ15N PN also partially explains the elevation of δ15N of NO3− in ≈3000 m depth water in the equatorial Pacific relative to deep waters upstream in the circulation of Pacific Deep Water (see Sigman et al.  for a similar explanation for the deep North Pacific). In summary, through diapycnal mixing and the downward flux of high δ15N of organic matter, the elevation of δ15N of NO3− in the intermediate depths of the South Pacific (the depths impacted by processes in the Subantarctic and ETSP ODZ) can also result in an increase of δ15N of NO3− throughout the equatorial Pacific water column.
4.4. Zonal Exchange by the Deep Tropical Gyres
 Another important feature of equatorial Pacific δ15N of NO3−distribution is the lateral homogeneity for the depth interval between ≈100 m (the EUC) and 600 m. The standard deviation of the measurements is ±0.3‰ in the shallowest depths and ±0.1‰ in the depths of SAMW-derived waters (seeTable 1). This homogeneity is surprising because a zonal gradient in δ15N of NO3− might be expected related to the distance from the eastern tropical Pacific ODZs (where much higher δ15N of NO3− is found; Figures 2 and 4). That there is no such basin-wide gradient suggests that equatorial waters are an ODZ source [Johnson and McPhaden, 1999; Kessler, 2006].
 This homogeneity in the δ15N of NO3− is likely maintained by the intensive zonal mixing occurring just below the EUC and between the NSCC and SSCC known as the “Deep Tropical Gyres” [Rowe et al., 2000]. The NSCC and SSCC are subsurface eastward jets that retroflect equatorward in the east and add to the westward flowing Equatorial Intermediate Current and the North and South Equatorial Currents [Gouriou and Toole, 1993; Johnson and Moore, 1997; Rowe et al., 2000; Sloyan et al., 2003]. This circulatory pattern results in significant mixing and homogenization of water located between the NSCC and SSCC, which is the most likely explanation for the subsurface homogeneity in δ15N of NO3− and also [NO3−] (not shown), temperature, and salinity (a.k.a. the “Thermostad” [Lukas, 1986]). The homogeneity at even deeper intermediate water depths strongly suggests the presence of an even “deeper” tropical gyre consisting of the Equatorial Intermediate Current and the off-equatorial Intermediate Counter Currents [Firing et al., 1998; Stramma et al., 2010].
4.5. Evidence for Eastward Transport of Newly Fixed N
 So far, we have described processes that work to produce elevated δ15N of NO3− in most of the subtropics and tropics for both the North and South Pacific. However, the relatively low δ15N of NO3−, averaging (5.5 ± 0.3)‰, associated with the high-velocity off-equatorial Southern Subsurface Counter Current (SSCC) (Figures 2 and 4) must have a different origin. Deep Southern Ocean waters have similar δ15N of NO3−, but these colder, denser, deeper, and higher [NO3−] waters are an unlikely source of the tropical subsurface (Tables 1 and 2). The lower-than-expectedδ15N suggests the influence of N2 fixation, which produces organic matter with δ15N ≈ −1‰. While there have been suggestions of N2 fixation in the eastern tropical Pacific [Deutsch et al., 2007; Westberry and Siegel, 2006], there is no reason to expect remineralization of newly fixed N so clearly organized by latitude and depth within the eastern SSCC. Furthermore, strong vorticity barriers poleward of the SSCC make it unlikely that lower δ15N of NO3− is introduced via the subtropical gyre [Johnson and Moore, 1997; Tsuchiya, 1981]. Instead, we propose that the lower δ15N of NO3− in the SSCC derives from N fixation prevalent in the western South Pacific where the SSCC has its origins [Dupouy et al., 1988; Campbell et al., 2005; Masotti et al., 2007; Garcia et al., 2007; Biegala and Raimbault, 2008; Raimbault and Garcia, 2008; Bonnet et al., 2009; Law et al., 2011].
 Prior work has assumed the EUC and SSCC source waters were the same [Tsuchiya, 1981], but our measurements propose distinct sources for each current. This was also suggested by Butt and Lindstrom  who found an eastward subsurface current beginning around 2°S, 153°E. Future measurements in the eastern and western tropical Pacific should be able to determine whether this signal is omnipresent in the SSCC and if the lower δ15N of NO3− signal has temporal variability. Such variability may also explain why N* (not measured concurrently with δ15N of NO3−) appears to show no clear elevation in the SSCC (Figure 5), as might be expected in regions of N2 fixation [Gruber and Sarmiento, 1997].
 It is worth noting that the eastward moving SSCC jet identified by ADCP contains nearly all the low δ15N of NO3− along 95°W, but only part of the lowest δ15N of NO3− along 110°W (Figure 5). There are two possible explanations for this incongruence. First, in the east, some portion of SSCC water moves equatorward and is entrained in the westward moving currents as part of the “deep tropical gyre” [Rowe et al., 2000]. Second, the distribution of tropical Pacific nutrients depends on the integrated transport of the preceding month [Strutton et al., 2008], which may not be perfectly represented in ADCP measurements at the time of sample collection. For example, tropical instability waves may shift water masses meridionally on the scale of days.
 In contrast to the SSCC, the NSCC has a North Pacific source [Johnson et al., 2002] and nearly the same δ15N of NO3− as other tropical Pacific subsurface waters (Table 1). Assuming this subsurface jet is not altered en route from the western equatorial Pacific (as with the SSCC), the elevated δ15N of NO3− points to the influence of eastern tropical Pacific denitrification in the eastern tropical North Pacific, perhaps carried to that region by the northern SEC or NEC. Regardless of their different sources, the SSCC and NSCC play a part in zonally mixing tropical Pacific waters through the actions of the deep tropical gyre.
4.6. Explanation for Lower δ15N of NO3− Above the Pacific ODZs
 The vigor of lateral transport and exchange in the tropical Pacific may also explain the relatively lower δ15N of NO3− above the very high δ15N of NO3− in the ODZ core (Figure 2b). It has been argued that this vertical gradient in δ15N of NO3− results from the remineralization of organic matter produced through N fixation in surface waters above the ODZ [Brandes et al., 1998]. This explanation assumes that NO3− in waters above the ODZ is supplied from the ODZ and therefore starts with an elevated δ15N of NO3− (>15‰). In this hypothesis, the elevated δ15N of NO3− is then reduced to relatively “low” values (≈7‰) by the remineralized products of N fixation. However, with a more expansive data set, it is apparent that the typical δ15N of NO3− for tropical Pacific water (≈7‰) is very similar to what was previously assumed to be anomalously low values. Therefore, the vertical gradient in δ15N of NO3− above the ODZ could be explained simply by lateral mixing of this typical tropical water. In fact, the waters that are mixing above the ODZs may come directly from the poleward retroflection of the subsurface counter currents [Tsuchiya, 1981], such as the NSCC, which has “a persistent northern component” in the east [Rowe et al., 2000]. It remains to be seen whether this explanation for lower δ15N of NO3− and [NO3−] is also applicable to similar observations from the Arabian Sea ODZ [Brandes et al., 1998]. The equatorward edges of the ETNP and ETSP ODZs also have large gradients in δ15N of NO3−, [NO3−], and [O2], especially in the 100 to 300 m depth range (Figures 4 and 5). Like the waters above the ODZs, these strong lateral gradients may be maintained by the vigorous subsurface flow of the SSCC/NSCC and the retroflection of the Deep Tropical Gyres, which circulates the outward mixing characteristics of ODZ water across the entire Pacific and thus dilutes them locally.
 From the lower thermocline (≈100 m) down to 600 m in the equatorial Pacific, the δ15N of NO3− is 7.1‰, an elevation of more than 2‰ relative to similar depths in North Atlantic and the Antarctic [Sigman et al., 2000; Knapp et al., 2005, 2008; DiFiore et al., 2006]. Coupled with the negative N* found throughout the Pacific basin, these measurements strongly suggest that denitrification in the eastern tropical Pacific influences nutrient characteristics of the equatorial Pacific waters. In the face of strong meridional gradients in δ15N of NO3− in the eastern Pacific, the zonal homogeneity of the δ15N of NO3− argues that the Deep Tropical Gyres and other deeper zonal mixing processes rapidly exchange NO3− between east and west Pacific. The depth homogeneity of δ15N of NO3− implicates (i) diapycnal mixing in the upward transport of NO3−from mid-depths to the equatorial undercurrent (EUC) and (ii) the sinking of highδ15N organic N into the deep ocean.
 These conclusions have implications for the nutrient pathway from the deep Southern Ocean to the equatorial Pacific surface waters via Subantarctic Mode Waters [Toggweiler et al., 1991; Sarmiento et al., 2004; Palter et al., 2010]. In particular, while deep-sea nutrients are required to replenish the low latitude thermocline, if the pathway is primarily within SAMW, its nutrient characteristics are strongly modified en route to the equatorial Pacific. We identify two processes that, working together, can shape SAMW into subsurface equatorial Pacific water: (1) circulation- and mixing-driven exchange with water influenced by eastern tropical Pacific denitrification and (2) the addition of NO3− through remineralization of organic matter (i.e., “nutrient trapping”). We propose that much of the mixing occurs on the Chilean margin, a suggestion that is supported by measurements in the region (Figure 6). The regenerated NO3−added to the equatorial Pacific intermediate-depth water has aδ15N similar to SAMW and the equatorial Pacific; this is consistent with the complete consumption of upwelled NO3− in tropical waters, leading to sinking N that is similar in δ15N to the upwelled NO3−. Subtropical South Pacific NO3− (water that is “upstream” of the equatorial Pacific) shows that much of the biogeochemical modification of SAMW has occurred before this water reaches the equator.
 In the eastern tropical Pacific, we find the expected elevation of δ15N of NO3− in the Oxygen Deficient Zones (>18‰ in very low oxygen waters), but we also find surprisingly low δ15N of NO3− (5.5 ± 0.3) ‰ being advected eastward within the Southern Subsurface Counter Current. We propose that the low δ15N of NO3− in the SSCC is derived from the remineralized products of N fixation in the North Tasman Sea, Coral Sea, and/or the south Pacific gyre. Because of these δ15N of NO3− differences in the eastern tropical Pacific, the meridional exchange between these water masses caused by the Deep Tropical Gyre retroflection can be clearly observed (Figures 4 and 5). Along these lines, we suggest that the lower δ15N of NO3− found at the shallowest depths of the ODZs (≈7‰ compared to >14‰ in the ETNP ODZ core) is due at least in part to the lateral exchange with tropical Pacific waters outside of the ODZs.
 While we have described a pathway for Southern Ocean waters to the equatorial Pacific thermocline, there are several predictions we can also make regarding the ultimate fate of this water. For example, the EUC slows east of 95°W [Lukas, 1986], where it appears to split around the Galapagos [Karnauskas et al., 2010] and approach the South American coastline, perhaps entering the Peru-Chile Undercurrent where it can influence the ODZ upwelling waters off Peru. The off-equatorial Northern and Southern Subsurface Counter Currents likely feed the Costa Rica Dome and Peru Margin upwelling respectively [Kessler, 2006]. These subsurface counter currents may also be the principle source of lateral ventilation for the ETNP and ETSP oxygen deficient zones [Kessler, 2006; Stramma et al., 2010], making these water masses important with regard to the strength of denitrification.
 While the application of δ15N of NO3− measurements has primarily been targeted toward understanding biogeochemical cycling, our study employs δ15N of NO3− as a tool in the study of ocean circulation. For example, we have identified key modifications of Subantarctic waters en route to the equatorial Pacific. Through our dense sampling of eastern equatorial waters, we have also shown that SSCC source waters are different than the other subsurface jets (the EUC and NSCC; see Figure 4). These results highlight the need for a more extensive application of this geochemical tool not only as a tracer for the cycling of nutrients, but also to investigate ocean circulation.
 We thank the National Oceanic and Atmospheric Administration (NOAA), the crews of the Ka'imimoana and Ronald H. Brown, Mike McPhaden, Brian Lake, Chris Beaverson, Shawn Gendron, and Tonya Watson. Invaluable laboratory help was provided by Greg Cane, Dennis Graham, Yi Wang, and Julien Foriel. Several figures were prepared using Ocean Data View (R. Schlitzer, Ocean Data View, 2011, http://odv.awi.de). We also thank two anonymous reviewers and the Editor for their comments. The measurements at Princeton were supported by the US NSF through grant OCE-0447570.