Potential drivers and consequences of regional phosphate depletion in the western subtropical North Pacific

In regions of the nitrogen limited low latitude ocean, phosphate can also be depleted to levels initiating stress responses in marine microbes. Here, we associate a broad region of phosphate depletion in the subtropical North Pacific with different levels of phosphorus stress. Nutrient and aerosol addition experiments demonstrated primary nitrogen limitation of the bulk phytoplankton community, with supply of aerosols relieving this limitation. At northern sites with depleted phosphate, alkaline phosphatase activities were enhanced, indicating elevated phosphorus stress. Analysis of satellite‐ and model‐derived aerosol loading showed that aerosol deposition was elevated in these regions. Surface rate measurements suggested that the regional enhancement in phosphate depletion was predominantly driven by elevated nitrogen fixation, likely stimulated by the coincident supply of aerosol iron. Such observations are important for predicting future biogeochemical responses in the subtropical North Pacific to changing aerosol supply.

nitrogen fixation, likely stimulated by the coincident supply of aerosol iron. Such observations are important for predicting future biogeochemical responses in the subtropical North Pacific to changing aerosol supply.
Nitrogen (N) limits phytoplankton growth throughout the majority of the low latitude oceans, but in certain regions phosphate (DIP) can be depleted to levels approaching colimitation (Moore et al. 2013). Most evidence for DIP depletion has been provided from the (sub)tropical North Atlantic, where numerous studies have demonstrated that DIP concentrations can be depleted to < 10 nmol L À1 , levels leading to microbial phosphorus stress (Wu et al. 2000;Mather et al. 2008;Moore et al. 2008Moore et al. , 2009Van Mooy et al. 2009;Mahaffey et al. 2014;Browning et al. 2017). However, evidence has also accumulated for low DIP concentrations over a broad extent of the (sub)tropical North Pacific Kitajima et al. 2009;Shiozaki et al. 2010;Martiny et al. 2019;Browning et al. 2022). This has important implications for projecting future changes of the broad oligotrophic North Pacific ecosystems, given the fact that DIP concentrations have been suggested to be declining in recent decades in this region (Kim et al. 2014).
The processes of denitrification, N 2 fixation, and atmospheric N deposition are all believed to be crucial to regulating the distribution of DIP and its excess relative to dissolved inorganic nitrogen (DIN) in surface waters of the low latitude oceans (Gruber and Sarmiento 1997;Wu et al. 2000;Kim et al. 2014). For example, in the (sub)tropical North Atlantic, enhanced N 2 fixation in particular has been proposed as an essential driver of DIP drawdown, which in turn has been related to elevated iron (Fe) supply rates from aerosol deposition (Wu et al. 2000;Moore et al. 2009). In contrast to the North Atlantic, high DIP concentrations in the South Atlantic have been suggested to reflect Fe limitation of N 2 fixation (Moore et al. 2009). A similar control has been suggested to drive a regional DIP gradient in the North Pacific Kitajima et al. 2009;Shiozaki et al. 2010).
Microbial communities growing under scarce phosphorus (P) initiate a range of stress responses, including the substitution of phospholipids for non-phosphorus containing forms (Van Mooy et al. 2009), utilization of reduced P(III) species (phosphite and phosphonates; Van Mooy et al. 2015;Repeta et al. 2016), and upregulation of P scavenging enzymes including alkaline phosphatases (APases) for hydrolysis of dissolved organic phosphorus (DOP) (Karl 2014). In such regions, the concentrations of DOP can exceed DIP by orders of magnitude, with the labile components of this pool potentially supporting microbial P requirements and thereby overall productivity in these systems (Mather et al. 2008;Letscher et al. 2016). Subsequently, the ranges of microbially produced DOP acquisition enzymes have diverse trace metal requirements (Duhamel et al. 2021), presenting the possibility that the restricted availability of these elements could limit DOP acquisition rates in the ocean. For example, evidence has been found for Fe (Browning et al. 2017) and Zn (Mahaffey et al. 2014) limitation of APases activity (APA) in the tropical North Atlantic. However, robust links between DIP availability, N 2 fixation, aerosol deposition, and the controls on microbial access to the DOP pool are lacking in the subtropical northwest Pacific, a vast ecosystem that is subject to rapid changes (e.g., Kavanaugh et al. 2018).
Here, we report results from simultaneous measurements of low-level macronutrient concentrations, trace elements, N 2 fixation and APA across the subtropical northwest Pacific, and supplement these with bioassay experiments testing the short-term microbial response to (micro-)nutrient supply. In parallel, desert dust and anthropogenically impacted aerosols, potentially releasing a range of (micro-)nutrients simultaneously, were further supplied in experiments to simulate the potential short-term biogeochemical impact of their supply.

Methods
Experiments and sample collection were conducted onboard RV Tan Kah Kee from 7 th to 18 th January 2021 (Fig. 1a). Surface seawater for amendment experiments and ambient nutrient and trace metal concentrations measurements (3-h time interval) was pumped from $ 2 m depth into a trace-metalclean laboratory from a towed sampling device (Zhang et al. 2019). Additional discrete depth profile samples were collected throughout the upper 1000 m using 12-L Niskin sampling bottles alongside a conductivity-temperaturedepth (CTD) profiler. The photosynthetically active radiation (PAR) sensor attached to the CTD was used to evaluate 1% of the surface PAR depth (euphotic zone, Lee et al. 2007) and 0.1% of the surface PAR depth.
All laboratory analysis details are provided in Supporting Information Text S1. The monthly average climatology for surface phosphate concentration was obtained from the World Ocean Altas 2018 (https://www.ncei.noaa.gov/ products/world-ocean-atlas). The satellite-derived average UV aerosol index was downloaded from GIOVANNI (https:// giovanni.gsfc.nasa.gov/giovanni). Daily sea level anomaly was obtained from the Copernicus Climate Data Store (https://cds. climate.copernicus.eu/cdsapp#!/dataset/satellite-sea-levelglobal?tab=overview). Aerosol depositions of Fe, fixed N, and DIP were extracted from the published model result of Chien et al. (2016).

Results and discussion
The cruise track transited the North Pacific Subtropical Gyre, from the northern boundary to its southern extent at the north equatorial current (NEC) (Fig. 1a). Surface waters were consistently depleted in DIN (mean DIN = 3.8 nmol L À1 , SD = 3.6 nmol L À1 , n = 34). In contrast to DIN, DIP displayed a pronounced gradient, increasing more than fivefold from $ 20 nmol L À1 in the north to > 100 nmol L À1 in the south of the study area ( Fig. 1b; Table 1). This trend is consistent with the overall pattern in climatological values from the WOA dataset ( Fig. 1a) alongside other previous observations Kitajima et al. 2009;Shiozaki et al. 2010;Martiny et al. 2019). Surface N 2 fixation rates measured at the four stations demonstrated an opposite trend to that of surface DIP concentrations, with rates decreasing abruptly between the two northerly stations (1.46 and 1.15 nmol L À1 d À1 at Stas. M30 and M22, respectively) and two stations in the south (0.26 and 0 nmol L À1 d À1 at Stas. M18 and K8a, respectively). Although our number of N 2 fixation observations is restricted, the inverse gradients in N 2 fixation rates and DIP concentrations between the gyre circulation and the NEC in the western (sub)tropical Pacific have consistently been observed in prior studies Kitajima et al. 2009;Shiozaki et al. 2010).
Our observed gradients in DIP and N 2 fixation were also consistent with that documented in the (sub)tropical Atlantic, where a major aerosol Fe deposition gradient regulates where N 2 fixers, with exceptionally high Fe requirements, can become established (Moore et al. 2009). A similar mechanism has been proposed for the gradients observed in the western (sub)tropical Pacific Kitajima et al. 2009;Shiozaki et al. 2010). In this study, concentrations of dissolved Fe were measured along parts of the transect, which demonstrated an elevated value in the northern-most site (0.60 nmol L À1 ; Sta. M30) that exceeded values further to the south approximately twofold to sixfold ( Fig. 1b; Table 1). Overall, this agreed very well with previous results along exactly the same transect (Nishioka et al. 2020), although differed somewhat to less clear trends found further to the east (Tanita et al. 2021); the latter potentially reflective of the episodic nature of aerosol deposition and the relatively short lifetime of Fe in surface waters. This trend was generally consistent with enhanced Fe availability for N 2 fixation in the northern sites. To further investigate if this trend was related to enhanced aerosol deposition, we analyzed satellite observations of UV aerosol index, a proxy for atmospheric aerosol loading, for the time of our cruise (Torres et al. 2013) and published climatological model estimates of nutrient deposition (Chien et al. 2016). Both the UV aerosol index and model Fe deposition displayed a trend consistent with enhanced aerosol Fe deposition in the north, with more than fourfold greater atmospheric aerosol loading indicated by the UV index and greater than fourfold higher aerosol Fe deposition in the northern (M30 and M22) compared to southern (M18 and K8a) stations. Concentrations of dissolved Zn, Cu, and Ni concentrations showed much less variability than Fe, with the average values of 0.14, 0.45, and 1.92 nmol L À1 , respectively ( Fig. 1b; Table 1), consistent with aerosols probably being a less important source of these elements (Mahowald et al. 2018).
In line with the highly depleted DIN concentration throughout the transect (Fig. 1b), nutrient addition experiments conducted at the four stations demonstrated that N was always the primary limiting nutrient for the bulk phytoplankton community (Fig. 2). No Chl a enhancements were observed following the addition of any nutrient combination that did not contain N, including those amended with P, Fe, Cu, Zn, or Ni. At the northernmost station (M30, Experiment 1), which was host to the highest N 2 fixation rates and the lowest DIP concentrations (Table  1), serial Chl a enhancement (that is, a greater response to the addition of N alone) was observed following addition of N + P and N + Fe combinations. The serial response to P addition can be readily explained, reflecting a condition where depleted initial seawater DIP concentrations (17.5 nmol L À1 ) are further lowered to limiting levels following artificial addition of bioavailable N (Moore et al. 2008;Browning et al. 2017Browning et al. , 2022. In contrast, the serial limitation response to N + Fe is less easy to reconcile with the elevated dissolved Fe concentrations at this site (0.60 nmol L À1 ). Furthermore in Experiment 1, no significant Chl a enhancement was observed following the addition of N + Fe + Cu, suggesting that the Cu addition prevented stimulation of phytoplankton growth (Paytan et al. 2009); however, this was difficult to reconcile with the significant enhancements that were observed in Experiments 2-4. The addition of the variety of aerosol treatments appeared to relieve the primary N limitation at all sites, by enhancing Chl a to levels observed in N addition treatments, suggesting both types of aerosols at both loading concentrations supplied bioavailable N (Fig. S1).
In addition to the depleted concentrations of DIP in the upper water column (Fig. 3c), alongside the serial Chl   a (a, c, e, g). Points denote individual values; bar heights and lines indicate the means and ranges, respectively (n = 3). Horizontal dashed lines represent initial values. Alkaline phosphate activity (b, d, f, h). Bar heights and lines indicate the means and standard errors, respectively (n = 3). The added aerosols comprised of desert-type dust [A1] and anthropogenic perturbed aerosol [A2], each at 0.02 and 0.2 mg L À1 , referred to as A1_0.02, A1_0.2, A2_0.02, and A2_0.2, respectively. The sample of zinc (Zn) addition in Experiment 2 was not determined. Statistically distinguishable means are labeled with different letters (using a one-way ANOVA and a Tukey honest significant difference [HSD] means comparison test, p < 0.05). The red bars and labels indicate that the treatments are significantly increased relative to controls. a enhancement to N + P supply at the northern station (M30; Fig. 2a), evidence for in situ P stress was provided by rates of APA (Fig. 3g). Rates of APA were enhanced approximately twofold in surface waters at northern sites, and more broadly throughout the upper water column (matching previous observations in the region; Suzumura et al. 2012), likely suggesting more rapid utilization of DOP in these waters. Across the whole dataset, APA rates were generally inversely correlated with DIP concentrations (Fig. S3). This matches previous observations (Lomas et al. 2010;Suzumura et al. 2012;Mahaffey et al. 2014), although we observed substantial variations in this trend, including cases where APA was elevated but DIP concentrations were not depleted ( Fig. S3; Sebasti an et al. 2004;Duhamel et al. 2011;Davis and Mahaffey 2017). Responses of APA to nutrient addition were thus less clear than that for Chl a, however some broad trends emerged. First, in contrast to several previous observations (Tanaka et al. 2006;Duhamel et al. 2010;Mahaffey et al. 2014;Browning et al. 2017), but consistent with depth profiles (Fig. 3c,g), P addition did not suppress rates of APA. Nitrogen and Fe addition at the northern, most P-depleted sites led to significant APA increases relative to untreated controls (Fig. 2b,d), presumably due to (1) stimulating phytoplankton growth by the provision of the primary limiting nutrient (N) and thereby further decreasing DIP concentrations, but also (2) provision of Fe, which is a required cofactor for widespread forms of APases enzymes (PhoX and PhoD; Luo et al. 2009;Rodriguez et al. 2014;Yong et al. 2014). Moreover, additions of high concentration of desert-type aerosol (A1_0.2) consistently elevated APA across all experiments (Fig. 2b,d,f,h). This could not be reconciled with the higher amounts of Fe and fixed N being supplied  (Redfield et al. 1963); (f) chlorophyll a concentration; and (g) rate of APA. Error bars represent the standard errors of triplicates. Nutrients measured via the AA3 Auto-Analyzer and nanomolar techniques are distinguished by open circles and filled circles, respectively. Diamonds and asterisks in (e) and (f) indicate 1% surface PAR and 0.1% surface PAR, respectively. from A1_0.2 (Fig. S1), as the parallel N + Fe treatment was already to replete values for both nutrients, but did not initiate APA responses at the southerly sites (Fig. 2f,h). Alternatively, the supply of aerosols might have also supplied significant labile organic carbon, which could have increased concentrations and activity of heterotrophic bacteria and thereby APA (Nicholson et al. 2006;Cao et al. 2010;Luo et al. 2011). Finally, the reason for the significant increase in APA after Ni addition at Experiment 1 was not clear, as Ni neither led to an increase in Chl a concentrations nor has a known role in APases (Duhamel et al. 2021). Ni is a cofactor in urease, which could have increased urea utilization as an N source, leading to P drawdown and thereby the increases in APA; however, such a mechanism would be expected to be associated with an increase in Chl a concentrations following Ni addition, which were not observed.
Ultimately, the relative supply rate of bioavailable N vs. P to the euphotic upper ocean would set how close P becomes to be limiting phytoplankton growth, with denitrification, N 2 fixation, and atmospheric deposition playing major roles (Gruber and Sarmiento 1997;Wu et al. 2000;Kim et al. 2014). Model aerosol nutrient deposition values for our transect suggest that aerosols supply both more Fe and fixed N to the northern part of the region, while the aerosol supply of DIP is predicted to be negligible, consistent with observations ( Fig. 1b; Table S1; Baker et al. 2010;Martino et al. 2014). Iron stimulation of N 2 fixation, in combination with the direct aerosol supply of fixed N, will both partially relieve N limitation of the bulk phytoplankton community in the north, contributing to the drawdown in surface DIP concentrations in conjunction with the increase in Chl a concentrations ( Fig. 3; Table 1). However, estimating the depth-integrated N 2 fixation rate by either integrating the measured surface N 2 fixation rates through the entire euphotic zone, or using published relationships between surface (ρ surface ) and depth-integrated ( Ð ρ) N 2 fixation rates ( Ð ρ ¼ 61:4 ρ surface , r 2 = 0.92, n = 22, p < 0.001; Wen et al. 2022), produces a new N supply from N 2 fixation of either 142 or 80 μmol N m À2 d À1 , respectively (Table S1). Either of these values is an order of magnitude higher than the estimated model aerosol N deposition ($ 9 μmol N m À2 d À1 ). Therefore, acknowledging the caveat of poor data constraint on both N supply terms, this points toward N 2 fixation being the dominant N supply mechanism leading to P drawdown. A fingerprint of this elevated N input is furthermore potentially reflected in the concentrations of DIN and DIP below the euphotic zone (Figs. 3, S4). Ratios of N : P concentrations are elevated in the northern sites relative to the south at the base of the euphotic layer, or 0.1% surface PAR depth, potentially reflecting the accumulation of diazotroph-derived DIN below the euphotic zone (Gruber and Sarmiento 1997;Wu et al. 2000;Kim et al. 2014), although the role of advection of subsurface waters with elevated N : P ratios from elsewhere cannot be ruled out. Upward effective diapycnal fluxes from these relatively N-enriched subsurface waters into the euphotic zone will act to further sustain the latitudinal surface DIP gradient (ratios of N : P fluxes at the 0.1% surface PAR depth, F_ DIN : F_ DIP , decreasing from 22 in the north to 15 in the south; Table 1). As a result, the reduced supply of DIP into the euphotic zone, relative to N, would thus drive the northern microbial community being more reliant on the rapid internal recycling of P ( Fig.3g; Hashihama et al. 2021; including unmeasured forms of reduced P).
We reconcile the broadscale meridional phosphate gradient through the western subtropical North Pacific Kitajima et al. 2009;Shiozaki et al. 2010;Martiny et al. 2019) with (1) primary N limitation of the bulk phytoplankton community, and (2) aerosol Fe regulating the latitudinal distribution of N 2 fixation rates, which introduces new bioavailable nitrogen and leads to phosphate drawdown (Wu et al. 2000;Hashihama et al. 2009). Secondary mechanisms likely contribute to the consistency of the low phosphate throughout the northern region despite more expected variability in aerosol Fe deposition and N 2 fixation (Fig. 1b); including direct aerosol supply of N, sustained upward diffusion of elevated N : P (ultimately derived from enhanced N 2 fixation in waters above) and potentially also Fe (ultimately aerosol-derived, but accumulated in the subsurface; Conway and John 2014; Rigby et al. 2020), alongside lateral advection and mixing of waters throughout the region (Lomas et al. 2010;Martiny et al. 2019). Phosphate depletion in the northern part of the study region in the season of the present study led to enhanced rates of APA and serial limitation of the bulk phytoplankton community by P, suggesting that the system is potentially approaching a state of N-P colimitation. Building on the observations presented here, future work employing lower-level nutrient addition experiments combined with simultaneous assessments of N 2 fixation rates and nutrient concentrations would further resolve just how close this system is to N-P colimitation. Regardless, enhanced N supply without equivalent P, via aerosol N deposition or N 2 fixation, would further draw down surface phosphate, strengthen P stress, and enhance microbial reliance on the DOP pool for phosphate (Lomas et al. 2010). However, while the contributions of N 2 fixation and atmospheric N supply to phytoplankton growth, and subsequent phosphate drawdown, are additive on short timescales, they have the potential to become strongly decoupled if enhanced aerosol N inputs reduce the niche for diazotrophs, via competition with non-diazotrophs for P and/or Fe (Krishnamurthy et al. 2007). Increasing aerosol N inputs to the North Pacific have been documented (Kim et al. 2014), underscoring the need to better understand the biogeochemical impacts of this forcing. Observations such as those presented here are a starting point for making such assessments.