UBC Faculty Research and Publications

Refined Estimates of Net Community Production in the Subarctic Northeast Pacific Derived From ΔO₂/Ar… Izett, Robert W.; Manning, Cara C.; Hamme, Roberta C.; Tortell, Philippe Daniel, 1972- Mar 7, 2018

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-Izett_R_et_al_Refined_estimates_net_community.pdf [ 1.65MB ]
JSON: 52383-1.0364314.json
JSON-LD: 52383-1.0364314-ld.json
RDF/XML (Pretty): 52383-1.0364314-rdf.xml
RDF/JSON: 52383-1.0364314-rdf.json
Turtle: 52383-1.0364314-turtle.txt
N-Triples: 52383-1.0364314-rdf-ntriples.txt
Original Record: 52383-1.0364314-source.json
Full Text

Full Text

Refined Estimates of Net Community Productionin the Subarctic Northeast Pacific DerivedFrom ΔO2/Ar Measurements With N2O-Based Corrections for Vertical MixingRobert W. Izett1 , Cara C. Manning1 , Roberta C. Hamme2 , and Philippe D. Tortell1,3,41Department of Earth, Ocean and Atmospheric Sciences, University of British Columbia, Vancouver, British Columbia,Canada, 2School of Earth and Ocean Sciences, University of Victoria, Victoria, British Columbia, Canada, 3BotanyDepartment, University of British Columbia, Vancouver, British Columbia, Canada, 4Peter Wall Institute for AdvancedStudies, University of British Columbia, Vancouver, British Columbia, CanadaAbstract We present the first field application of a N2O-based approach to correct for vertical mixing inthe estimation of net community production (NCP) from mixed layer O2 measurements. Using newship-based observations of N2O and biological oxygen saturation anomalies (ΔO2/Ar) from the SubarcticNortheast Pacific, we provide refined mixed layer NCP estimates across contrasting hydrographic regimesand a comprehensive assessment of the methodological considerations and limitations of the approach.Increased vertical mixing coefficients at the base of the mixed layer, derived using N2O measurements,corresponded with periods of heightened wind speed and coastal upwelling. Corrections were mostsignificant in coastal regions where the vertical supply of low-O2 water can otherwise falsely imply netheterotrophy from negative ΔO2/Ar measurements. After correcting for the mixing flux, all coastal stationsshowed autotrophic signatures, with maximum NCP exceeding 100 mmol O2 m2 d1 in the spring andsummer. Vertical fluxes were lower in off-shelf waters but often contributed more than 50% to correctedNCP. At some oceanic stations, however, the cooccurrence of N2O minima and O2 maxima resulted inbiased (overestimated) N2O corrections. Evaluating vertical fluxes in these regions remains a challenge forship-based studies. Nonetheless, our refined NCP estimates show better coherence with surfacechlorophyll, temperature, and mixed layer depth than uncorrected values. Potential mixed layer N2Oproduction introduces some uncertainty in the approach, but errors are likely to be small. Ultimately, thiswork provides rationale for the adoption of the N2O correction to refine NCP estimates, particularly incoastal waters.1. IntroductionMarine organic carbon production via photosynthesis regulates higher trophic level biomass (Ware &Thomson, 2005) and influences the strength of the ocean’s biological pump (Volk & Hoffert, 1985).Accurate quantification of primary production is thus critical for understanding the functioning of marineecosystems and their response to environmental change. As a currency of life, oxygen (O2) can be used asa tracer of biological production in the ocean. The net community production (NCP) of O2 represents thebalance between primary production and community respiration (both autotrophic and heterotrophic)and is equivalent to carbon export at steady state and on interannual timescales (Laws, 1991).A number of field studies have previously quantified NCP by examining themixed layer O2mass balance (e.g.,Emerson et al., 1997; Giesbrecht et al., 2012; Luz & Barkan, 2009; Reuer et al., 2007). One common approach isto normalize O2 concentrations to argon (Ar), a biologically inert gas, with solubility and diffusivity propertiesthat are nearly identical to O2 (Craig & Hayward, 1987). This normalization removes the influence of physicalprocesses, such as bubble injection and temperature change, that affect O2 saturation state. Following Kaiseret al. (2005) and Reuer et al. (2007), steady state NCP (mmol O2 m2 d1) is equated to the air-sea flux ofbiologically produced O2:NCP ¼ ΔO2=Ar· O2½ eq·kO2 ; (1)where ΔO2/Ar is the biological O2 saturation anomaly (unitless; equation (2)), [O2]eq is the O2 equilibriumconcentration (mmol m3), and kO2 is the gas transfer velocity of O2 (m d1).IZETT ET AL. 1PUBLICATIONSGlobal Biogeochemical CyclesRESEARCH ARTICLE10.1002/2017GB005792Key Points:• Nitrous oxide measurements are usedto correct for mixing fluxes inship-based estimates of netcommunity production• The magnitude of the mixingcorrection is highest in coastal watersduring spring and summer upwelling• Corrected net community productionshows better coherence withhydrographic variables thanuncorrected valuesSupporting Information:• Supporting Information S1Correspondence to:R. W. Izett,rizett@eoas.ubc.caCitation:Izett, R. W., Manning, C. C., Hamme, R. C.,& Tortell, P. D. (2018). Refined estimatesof net community production in theSubarctic Northeast Pacific derived fromΔO2/Ar measurements with N2O-basedcorrections for vertical mixing. GlobalBiogeochemical Cycles, 32. https://doi.org/10.1002/2017GB005792Received 29 AUG 2017Accepted 7 FEB 2018Accepted article online 12 FEB 2018©2018. American Geophysical Union.All Rights Reserved.Developments in sea-going mass spectrometry over the past decade (Cassar et al., 2009; Kaiser et al., 2005;Tortell, 2005) have made it possible to obtain high-resolution estimates of NCP using continuousship-board measurements of O2/Ar. High spatial sampling resolution is achieved, and biases associatedwith discrete methods for measuring primary production are avoided, making this approach ideal foropen-ocean environments. A number of groups have now applied ship-based ΔO2/Ar measurements ina variety of regions, greatly expanding the spatial coverage of NCP data (Hamme et al., 2012; Kaiseret al., 2005; Lockwood et al., 2012; Manning et al., 2017; Stanley et al., 2010; Tortell et al., 2015, 2012;Ulfsbo et al., 2014).Despite significant advances in NCP measurements, current estimates remain limited by uncertainty in thecontribution of vertical mixing to the surface O2 budget (Hamme & Emerson, 2006). When O2 is undersatu-rated (supersaturated) beneath the mixed layer, failure to account for the mixing supply of this low-O2(high-O2) water (through mixed layer entrainment or vertical diffusivity and advection) can lead to an under-estimation (overestimation) of derived NCP. This problem has been widely recognized, particularly in coastalregions, and limits the applicability of ΔO2/Ar as a productivity tracer (Giesbrecht et al., 2012; Hamme &Emerson, 2006; Jonsson et al., 2013). In some cases, such regions may falsely appear to be net heterotrophicunless a mixing correction is made.Recent modeling work by Cassar et al. (2014) has suggested that measurements of nitrous oxide (N2O), withinand below the mixed layer, may be used to trace vertical mixing and correct ΔO2/Ar-based NCP estimates.Nitrous oxide is largely produced in the oceans during organic matter remineralization, as a by-product ofammonium oxidation (AO) to nitrate (i.e., nitrification) (Ward, 2000). As a result, a consistent relationshipbetween the N2O surplus and O2 deficit is observed in much of the world’s subsurface oceans (Cohen &Gordon, 1979; Nevison et al., 2003; Yoshinari, 1976). Vertical mixing, including advection, diffusion, andentrainment, can therefore provide a source of water that is both undersaturated in O2 and supersaturatedin N2O. Such a signature has been observed in surface waters of upwelling margins globally (Bange et al.,1996; Capelle & Tortell, 2016). It has generally been believed that nitrification is photoinhibited and primarilyconstrained to subsurface waters (Horrigan et al., 1981). Mixed layer N2O supersaturation is thus assumed tolargely reflect vertical mixing, such that surface N2O measurements can, in theory, be used to track theupward transport of low-O2 water. To our knowledge, there have not yet been any direct tests of this methodand its assumptions with field data.In this article, we present new NCP estimates from the Subarctic NE Pacific Ocean, using N2O-corrected ship-based measurements of ΔO2/Ar across a range of coastal (on-shelf) and open-ocean (off-shelf) waters.Surface productivity in the off-shelf waters of the Subarctic Pacific is limited by iron (Fe) availability duringthe summer (Boyd et al., 1996; Maldonado et al., 1999), while the persistence of a steep halocline and strongdensity stratification (Tabata, 1975) limits exchange between the surface and subsurface. In contrast, coastalwaters of this region are characterized by complex physical dynamics, including seasonal upwelling andmixing over heterogeneous bottom topography (Crawford & Thomson, 1991; Whitney et al., 2005). Thesephysical processes drive high productivity but also introduce uncertainty in ΔO2/Ar-based NCP estimates.Our study thus presents an opportunity to evaluate O2-based estimates of NCP across contrasting oceano-graphic regimes. Our new observations provide refined estimates of NCP in the Subarctic Pacific and demon-strate that neglecting the physical contributions to the mixed layer O2 budget can lead to significantunderestimates in some cases. We also evaluate the seasonal and spatial trends in NCP and present anin-depth evaluation of the limitations, uncertainties, and methodological considerations of the N2O correc-tion approach. Our work builds on that of Cassar et al. (2014) and provides guidance for future studies.2. Methods2.1. Study AreaWe collected data on five oceanographic cruises in the Subarctic NE Pacific on the CCGS John P. Tully, betweenFebruary and September 2016 (Figure 1). Line P cruises to Ocean Station Papa (OSP/Station P26; 50°N, 145°W)were conducted in winter (February), spring (June), and summer (August), covering a transect from thecoastal Juan de Fuca (JF) Strait and southern West Coast Vancouver Island (WCVI) to the open-ocean watersof the Gulf of Alaska. Coastal waters around Vancouver Island, including the JF and WCVI regions, wereGlobal Biogeochemical Cycles 10.1002/2017GB005792IZETT ET AL. 2surveyed extensively during cruises in May and September, as part of the La Perouse monitoring program,run by Fisheries and Oceans Canada.The WCVI region is generally influenced by the seasonal timing and intensity of summer upwelling andwinter downwelling (Bylhouwer et al., 2013). During the summer and fall, local currents near the mouth ofthe JF region enhance upward transport of low-O2, nutrient-rich deep water onto the shelf (Crawford &Peña, 2013). In the open-ocean waters of our survey region, strong winter winds drive upper ocean turbu-lence and mixed layer deepening, whereas increased solar radiation and weaker winds in the summer causeshoaling (Whitney & Freeland, 1999). In the spring, water column stratification enhances phytoplanktongrowth and nutrient drawdown prior to the onset of off-shelf summer macronutrient and Fe limitation andsubsequent resupply via turbulent winter mixing (Peña & Varela, 2007; Whitney & Freeland, 1999).Year-round, deep mixing in the off-shelf waters is restricted by the presence of a steep halocline.2.2. Underway O2/Ar MeasurementsContinuous measurements of mixed layer O2/Ar were conducted using membrane inlet mass spectrometry(Tortell et al., 2011), with a Hiden Analytical HAL20 quadrupole mass spectrometer. Seawater from the ship’sunderway seawater loop (intake at 4.5 m) was circulated at constant flow rate through rigid polypropylenetubing, past a 0.007″ thick water-impermeable silicone membrane. To minimize changes in the membranegas permeability, which can affect measurements of O2 and Ar, we maintained constant temperature(10°C) of inflowing seawater by circulating the water through a 6 m heat exchange coil immersed in a waterbath. Measurements were obtained every ~20 s as the ratio of ion currents at the mass-to-charge ratios of 32and 40 and averaged into 1 min bins.Seawater standards, used to correct for instrument drift and to derive biological O2 saturation anomalies (seesection 2.4), consisted of air-equilibrated, 0.2 μm filtered seawater in 4 L polycarbonate bottles (Tortell et al.,2011). The bottles were incubated at ambient sea surface temperature (SST) and bubbled using an air pumpto obtain gas equilibration with the atmosphere. While bubbling can introduce supersaturation in O2, itnegligibly affects the O2/Ar ratio. At 2 h intervals, we measured the standards for 2 min, using the same flowrate and sampling temperature as during the underway data acquisition.2.3. N2O Sample CollectionDiscrete samples for N2O analysis were collected on all cruises from Niskin bottles. Depth profiles wereobtained at various stations along the respective cruise tracks (markers in Figure 1), while surface measure-ments (5 m Niskin bottles) of N2O supersaturation were collected at higher spatial resolution. During theAugust Line P trip, additional samples for surface N2O analysis were obtained in the main laboratoryFigure 1. Map of representative Line P (gray) and La Perouse (black) cruise tracks, with markers showing the locations of N2O profile sampling. The contour lines onthe inset represent the 100, 500, 1,000, and 2,000 m isobaths. JF, WCVI, and QCS denote the Juan de Fuca canyon, West Coast Vancouver Island, and Queen CharlotteSound regions, respectively. P4, P8, P12, P16, P20, and OSP (P26) represent the major Line P stations, and LB, LC, LD, and LG refer to transects along the La Perousesampling grid. Note that the ship tracks varied slightly between cruises and that discrete surface samples for N2O measurements were collected at higher spatialresolution than the plot markers (not shown). The inset shows an expanded view of the JF and southern WCVI regions. The open squares indicate the locations atwhich Bakun Upwelling Index values were obtained (Figure S5).Global Biogeochemical Cycles 10.1002/2017GB005792IZETT ET AL. 3directly from the ship’s underway seawater system. These samples and corresponding Niskin samples(obtained at the same station from 5 m Niskin bottles) showed good agreement (average root-mean-squareerror of 0.5 μmol m3; Figure S1 in the supporting information), and no statistically significant difference wasobserved between paired samples (n = 12). Underway samples obtained between stations were thereforeincluded in our data.For all N2O measurements, duplicate subsamples were collected into 60 mL glass vials using flexible siliconetubing. The vials were overfilled three volumes avoiding bubble contamination, immediately poisoned with50 μL saturated mercuric chloride (HgCl2), and stored at 4°C in the dark until analysis (within 3 months).Nitrous oxide was measured in the laboratory using an automated purge-and-trap gas chromatography,mass spectrometry system (Capelle et al., 2015). Measurement precision was 3.5%, based on the averagepercent standard deviation of 125 triplicate air-equilibrated standards processed at the same time as ourseawater samples.2.4. Calculations of NCPUncorrected mixed layer NCP was calculated according to equation (1), where the biological O2 saturationanomaly is defined, as per Craig and Hayward (1987), asΔO2=Ar ¼ O2=Ar½ measO2=Ar½ eq–1: (2)The meas and eq subscripts refer to the seawater and atmospheric equilibrium ratios measured from theunderway seawater system and the bubbled air standards, respectively.We calculated corrected NCP (mmol O2 m2 d1) following the approach of Cassar et al. (2014), whichcombines the mixed layer O2 and N2O budgets.NCP ¼ kO2 · ΔO2=Ar· O2½ eq–kN2OkO2·∂ O2½ B∂ N2O½ B· N2O½ B !(3)Here kN2O is the gas transfer velocity for N2O (m d1), ∂ O2½ B∂ N2O½ B is the vertical gradient of biological-O2 tobiological-N2O concentrations (mmol O2 (mmol N2O)1), and [N2O]B is the surface biological-N2O saturationconcentration (mmol N2O m3). Cassar et al. (2014) assumed a constant kN2OkO2 ratio of 0.92, but we indepen-dently calculated the ratio (range 0.87–0.94) for each observation based on the Schmidt number parameter-izations reported in Wanninkhof (2014). The approach assumes that N2O production (through nitrification) isphotoinhibited in the mixed layer, but we discuss this assumption below (section 4.1.2). The NCP correctionterms (i.e., the second part of equation (3)), obtained at each station with surface N2O measurements (meanspatial resolution of approximately 102, 37, and 47 km in winter, spring, and summer respectively), werelinearly interpolated to the resolution of the underway ΔO2/Ar data (mean spatial resolution of 0.2 km).The interpolated correction terms were thus applied to equation (1) to obtain high-resolution estimates ofNCP along our entire cruise tracks. We assume steady state in biological O2 concentrations, and that horizon-tal fluxes of O2 can be neglected in calculating NCP (see sections and S2 of the supporting informationfor an analysis of these assumptions).As measurements of O2/Ar reflect processes occurring over the residence time of O2 within the mixed layer(between approximately 1 and 4 weeks), we applied the approach of Reuer et al. (2007), with corrections byTeeter (2014), to calculate weighted gas transfer velocities for O2 and N2O over a 30 day period prior to ourobservations. We used the gas transfer velocity parameterization of Ho et al. (2006) based on 6-hourly CCMP(Cross-Calibrated Multi-Platform) wind vector analysis product data (retrieved from http://www.remss.com/measurements/ccmp/; Atlas et al., 2011).The [O2]B and [N2O]B concentrations (mmol m3) represent the biological saturation concentrations of therespective gases. Within the mixed layer, [O2]B is defined asO2½ B ¼ ΔO2=Ar· O2½ eq; (4)where ΔO2/Ar (equation (2)) accounts for bubble injection and physically induced changes in solubility.Global Biogeochemical Cycles 10.1002/2017GB005792IZETT ET AL. 4Below the mixed layer, where we do not have Ar data, we assume that Ar concentrations are at saturation forthe calculation of [O2]B (this assumption is evaluated in section 4.1.4). Since there is no similar inert analog forN2O, we follow the recommendation of Cassar et al. (2014) and apply a thermal correction to surface mea-surements to account for solubility changes resulting from recent heat flux. Using this correction, mixed layer[N2O]B is calculated asN2O½ B ¼ N2O½ meas– N2O½ eq– N2O½ thermal: (5)Here [N2O]thermal is derived following the equation of Keeling and Shertz (1992), with kinetic corrections anddelay as per Jin et al. (2007).N2O½ thermal ¼11:3· –∂ N2O½ eq∂T·Qcp·ρ·86; 400 s1 day ·1kN2O(6)The term in parentheses represents the thermally induced sea-to-air flux of N2O (mmol N2O m2 d1). Rapidtemperature changes cause a disequilibrium between observed surface N2O and the saturation concentra-tion. The thermally induced flux of N2O should be equal to the disequilibrium concentration gradient,multiplied by the gas transfer velocity of N2O. We thus divide the thermal flux by kN2O to obtain theconcentration offset, [N2O]thermal, produced by solubility effects averaged over the period one half monthbefore the measurement time. The negative operator in equation (6) converts the air-to-sea heat flux (Q,W m2, the sum of latent heat, sensible heat, longwave, and shortwave radiative fluxes) and positive heatcapacity (cp, J kg1 C1; multiplied by sea surface density, ρ, kg m3) into a sea-to-air N2O flux. Here Q andcp were estimated as localized mean values over the mixed layer residence time of N2O from NCPE/NCARReanalysis 1 surface flux (6 h resolution data retrieved from https://www.esrl.noaa.gov/psd/data/gridded/data.ncep.reanalysis.html; Kalnay et al., 1996) and Level 3 MODIS Aqua SST (1 day, 9 km resolution meanof 11 μ daytime and nighttime data retrieved from https://oceancolor.gsfc.nasa.gov/cgi/l3; NASAGoddard Space Flight Center, Ocean Ecology Laboratory, Ocean Biology Processing Group, 2014a,2014b). Below the mixed layer, no thermal correction for N2O is made in the calculation of [N2O]B (i.e.,[N2O]thermal = 0 in equation (5)).The solubility equations of Garcia and Gordon (1992, 1993) and Weiss and Price (1980) were used to derivethe saturation concentrations of O2 and N2O, respectively, from SST and sea surface salinity measurements.In these calculations, we accounted for changes in atmospheric N2O concentrations using observations fromBarrow, AK (data provided by the NOAA ESRL Global Monitoring Division HATS Network and retrieved fromhttps://www.esrl.noaa.gov/gmd/dv/iadv/; Hall et al., 2007).The vertical [O2]B/[N2O]B gradient (hereafter referred to as the “supply ratio”) represents the stoichiome-try of vertical mixing, with respect to biological O2 and N2O. Cassar et al. (2014) used a discrete two-point slope ( ∂ O2½ B∂ N2O½ B ¼O2½ Bdeep– O2½ BMLN2O½ Bdeep– N2O½ BML), based on end-members within (ML) and below (deep) the mixedlayer. We found that this approach produced significant variability in our NCP correction (discussed insection 4.1.1). We thus utilized depth profile measurements to estimate the supply ratio from the slopeof the [O2]B versus [N2O]B relationship in the water column from the base of the mixed layer to 150 mbelow (mld to mld + 150 m). The supply ratio was therefore characterized independently for each cruiseas the slope through the respective pooled data set of all profile measurements in the 150 m below themixed layer. A similar approach was previously used by Castro-Morales et al. (2013) and Manning et al.(2017) in estimating vertical fluxes from O2 gradients. As discussed below (section 4.1.1), the magnitudeof the NCP correction is sensitive to the approach used to estimate the supply ratio term but is largelyinsensitive to the depth of integration or the threshold criterion of the mixed layer depth (mld)estimate.2.5. Calculations of Apparent Mixing Coefficients, ΚmixTo validate the NCP corrections, we calculated an apparent mixing coefficient (Κmix, m2 s1) based on themixed layer N2O budget. The coefficient represents all vertical advection, diffusion, and entrainmentprocesses over the mixed layer N2O residence time. Assuming steady state and that N2O is produced onlyGlobal Biogeochemical Cycles 10.1002/2017GB005792IZETT ET AL. 5below the euphotic zone (see section 4.1.2), the mixing coefficient is calculated from the surface N2O massbalance as follows:Kmix ¼ kN2O· N2O½ B∂ N2O½ B=∂Z·1 day86; 400 s; (7)where Κmix · (∂[N2O]B/∂Z) is the vertical mixing flux of [N2O]B (mmol m2 s1) and kN2O· N2O½ B is the air-seaflux (mmol m2 d1). The vertical gradient of [N2O]B, ∂[N2O]B/∂Z (mmol m4), was estimated at each stationwith profile N2O measurements as the depth-dependent regression slope of all N2O measurements in the150 m below the mixed layer.2.6. Ancillary Hydrographic Data and CalculationsAncillary hydrographic data were collected and provided by the Institute of Ocean Sciences laboratory of theDepartment of Fisheries and Oceans Canada. Continuous underwaymeasurements of SST and sea surface sali-nity were obtained from a thermosalinograph (Sea-Bird CTD SBE-21) connected to the ship’s surface seawatersupply. Depth profiles of temperature and salinity were obtained fromCTD casts using a Sea-Bird CTD (SBE-911-plus) mounted on the Rosette frame. We used a density-difference criterion of 0.125 kg m3 to define the mld.Water column measurements of O2 (at depths corresponding to our N2O samples) were made by discreteWinkler titrations, using an automated titration system (Metrohm Dosimat 876) (Carpenter, 1965).Chlorophyll (chl) a samples, collected onto 25 mm GF/F filters prior to extraction in 90% acetone, were mea-sured on a Turner 10 AU fluorometer before and after acidification to correct for phaeopigments(Welschmeyer, 1994). Nitrate and nitrite (NO3 + NO2) samples were measured spectrophotometricallyusing an Astoria autoanalyzer (Barwell-Clarke & Whitney, 1996).Higher-resolution (1 m depth intervals) O2 measurements were made using a SBE-43 O2 sensor mounted onthe CTD platform. These data were used to evaluate the gradient of biological O2 (∂[O2]B/∂Z, mmol m4)below the mixed layer at each station along the Line P and La Perouse transects. The gradient was taken asthe slope of the regression line through each station’s data set in the 25 m interval below the mixed layer(i.e., ∂[O2]B/∂Z|25). We selected an integration depth of 25m as it represents the depth range immediately belowthe mixed layer and the direct water source for vertical mixing to the surface. As we explain in section,we use these O2 gradients to diagnose situations where the N2O correction could not be applied.3. ResultsIn the presentation of our results, we divide our sampling area into two distinct hydrographic regions, basedon the spatial patterns we observed in NCP, chl a, and nutrient concentrations. The off-shelf region consists ofthose portions of the Line P transect west of station P4, while the on-shelf (coastal) region includes all of theLa Perouse track, and the Line P transect east of and including P4. This separation corresponds approximatelywith the location of the continental slope (500–1,000 m isobaths; Figure 1) and is similar to previous divisionsof the Line P transect (Whitney et al., 1998) and other physical representations of the continental shelf region(Crawford & Thomson, 1991). We first present our observations of hydrography and plankton biomass acrossthese two regions (section 3.1). Then we examine the relationship between O2 and N2O and assess how N2Odata can be used to derive mixing coefficients to correct NCP estimates (sections 3.2 and 3.3). We finish byexamining the spatial patterns of NCP across our survey region in relation to other oceanographic variables(section 3.4). Limitations of the N2O correction approach are addressed in the discussion (section 4.1).3.1. Hydrographic Conditions and Phytoplankton BiomassDuring our cruises, the spatial and temporal variability in mixed layer chl a generally followed typical patternsfor our study area, with the highest chl a in the on-shelf region (up to 38 mg m3 during the August Line Pcruise) and lower values (~0.2–0.8 mg m3) in much of the off-shelf region (Figures 2a–2c). We observed theexpected seasonal cycle in surface chl a in the coastal waters, with values increasing from <1.2 mg m3 inFebruary to >10 mg m3 in May and June, and >30 mg m3 in August. In the off-shelf region, winter andspring concentrations (~0.2–0.8 mg m3) were usually greater than in August (<0.4 mg m3 west of stationP8). Peaks in chl a were observed near OSP during the spring and summer. The maximum chl concentrationobserved at OSP on the June cruise (1.2 mg m3) was unusually high (relative to the long-term summer aver-age of ~0.4 mg m3).Global Biogeochemical Cycles 10.1002/2017GB005792IZETT ET AL. 6Dissolved NO3 + NO2 showed maximum concentrations (>10 μmol L1) closest to shore in all threeseasons (Figures 2d–2f). In the oceanic waters, concentrations were highest in February, ranging from~5 μmol L1 at the boundary between the on-shelf and off-shelf zone (approximately 80 km from shore)to nearly 14 μmol L1 at OSP (Figure 2d). June and August had a similar spatial pattern, but concentrationsdecreased through the spring and summer. In general, spring and summer nutrient concentrations in thesouthern shelf region were low to about 40 km off shelf (<6 μmol L1) but increased sharply toward the coastto values above 10 μmol L1 (Figures 2e and 2f). Between stations P4 and P14 (~400 km from shore), weobserved near depletion (<0.5 μmol L1) of mixed layer NO3 + NO2 in June. In August, the off-shelf deple-tion extended to over 600 km. Beyond the regions of macronutrient depletion, concentrations increased butnever reached maximum values observed in the winter. The elevated summer macronutrient concentrationsin off-shelf waters are taken as evidence of Fe limitation (Varela & Harrison, 1999).3.2. O2 and N2ODepth profiles of O2 and N2O in the upper water column exhibited a linear relationship at all stations. Ourmeasurements, made in 2016, are consistent with those from previous Line P and La Perouse cruises datingback to 2012 (Capelle & Tortell, 2016) (Figure 3a). Both O2 and N2O typically exhibited homogeneous distri-butions and supersaturation in the mixed layer ([O2]B and [N2O]B > 0; surface [N2O]B data are presented inFigure S2). With the exception of some off-shelf stations during spring and summer, N2O supersaturationincreased monotonically below the mixed layer, while O2 became significantly undersaturated (Figures 3aand S3; deviations are discussed below in section From all available profile data (2012 to present),we thus derived an inverse linear relationship between [O2]B and [N2O]B, which yielded a narrow distributionin the subsurface-to-surface (mld to mld + 150 m) supply ratio terms (Figure 3b). Based on 2016 profile data,our calculated supply ratios ranged from3.2 × 104 to1.0 × 104 (mean1.5 × 104) mmol O2 (mmol N2O)1.Figure 2. Mixed layer chl a (a–c) and NO3 + NO2 (d–f) concentrations along the 2016 Line P and La Perouse transects. Data are separated into the off-shelf(a, d) and on-shelf (b, e) regions of the Line P transect and the La Perouse cruise (c, f). Note that the winter on-shelf symbols in Figure 2b are obscured behind thespring symbols and that the x and y scales differ among some of the panels.Global Biogeochemical Cycles 10.1002/2017GB005792IZETT ET AL. 7These values fall well within the range we observed along Line P since 2012 (1.2 × 105 to6.6 × 103 mmol O2 (mmol N2O)1), with 99% of 2016 gradients being within one standard deviation ofthe 2012–2016 mean (1.8 × 104 ± 1.3 × 104 mmol O2 (mmol N2O)1).Estimates of the [O2]B gradient in the 25 m below the mixed layer ranged from 8.4 to 1.1 mmol O2 m4(Figure 3c). Thus, at some stations, we observed increasing or near-constant [O2]B concentrations belowthemixed layer (∂[O2]B/∂Z|25 ≥0.1). All of the positive values (i.e., higher O2 concentrations below themixedlayer) occurred during the spring or summer in the off-shelf region. In contrast, only negative gradients wereobserved in the coastal regions during all seasons (Figure S4). Where N2O profiles were obtained, we alsoobserved N2O deficits (i.e., ∂[N2O]B/∂Z|25 < 0) coincident with the O2 maxima and therefore linear negative[O2]B/[N2O]B gradients with depth (Figure S3). The supply ratio for these profiles was not significantly differ-ent than the other stations or seasons.3.3. Apparent Mixing CoefficientsBased on our N2O measurements, we derived apparent mixing coefficients (representing all vertical advec-tion, diffusion, and entrainment processes) between 1.1 × 104 and 1.7 × 103 (mean 6.7 × 104) m2 s1(Figure 4). Values are presented only where stations showed the expected negative O2 gradient below theFigure 3. Relationship between N2O and O2 concentrations in the upper water column (a), showing [N2O]B and [O2]B from all 2016 measurements to 150 m (white)and 500 m (black) below the mixed layer, and previous observations to 500 m below the mixed layer (2012–2015, gray). The dashed lines indicate 100% saturation.(b) The distribution of supply ratio estimates derived from 2016 data only, and 2012–2015 data, calculated for each station in the 150 m below the mixed layer(see text for further explanation). The solid horizontal line denotes a slope of 1 × 104 mmol O2 (mmol N2O)1, representing the global approximate averagestoichiometry of O2/N2O (e.g., Nevison et al., 2003). (c) Estimates of the subsurface [O2]B gradient by season (2016 cruises only), defined in the 25 m below the mixedlayer. The absolute sensitivity of NCP estimates to the supply ratio term, the method of estimating it, and the correction approach is shown in (d). Corrected NCPwasNCPestimated using a supply ratio derived in the 150 m below the mixed layer, 500 m below, using the global stoichiometric average value, and a two-pointslope. We also evaluated NCP by making corrections using a literature-derived eddy diffusivity value and our estimates of ∂[O2]B/∂Z|25 (see section for moredetails). In (b)–(d), the boxes show the 25th, 50th, and 75th percentile values (bottom, middle, and top lines, respectively). Whiskers extend to the 1st and 99thpercentile ranges, and crosses represent outlying data beyond this range. The lower limit in (b) was restricted to4 × 104 mmol O2 (mmol N2O)1, and the scale in(d) was limited to ±300 mmol O2 m2 d1.Global Biogeochemical Cycles 10.1002/2017GB005792IZETT ET AL. 8mixed layer (see sections 3.4 and Similarly, we did not calculate a Kmix value at several stations wherewe observed undersaturated N2O concentrations in the mixed layer (Figure S2).The largest Kmix values were observed in the on-shelf regions following periods of intense spring and summerupwelling (mean 8.2 × 104 m2 s1). By comparison, mixing coefficients were lower in most of the off-shelfregion during the same seasons (mean 3.2 × 104 m2 s1), except following periods of increased wind speed(e.g., P8 in August). In the winter, we observed mixed layer N2O undersaturation at P20 and P26, resulting inderived Kmix < 0 that cannot be used to correct NCP estimates. Elsewhere off-shelf, Kmix had intermediatevalues in February (mean 2.7 × 104 m2 s1) and the lowest value at the coastal P4 station (2.6 × 104 m2 s1),relative to the other Line P cruises.3.4. Net Community ProductionFigure 5 shows the spatial distribution of NCP during the winter (a and b), spring (c–e), and summer (f–h)cruises, with and without the N2O correction. We present both uncorrected and corrected estimates,noting that some of the corrected values are biased by subsaturated surface [N2O]B, or subsurface[N2O]B deficits (and [O2]B maxima), which yield (nonsensical) negative mixing terms (discussed below insection Our final NCP estimates are thus compiled from a combination of corrected anduncorrected values, with uncorrected NCP selected for regions where the N2O correction could not beapplied. In total, the N2O correction could not be applied at 19 out of 137 locations (<15% of data points).A summary of mean and standard deviation values for final NCP and the N2O correction is presentedin Table S1.Figure 4. Vertical mixing coefficients derived from profile measurements of N2O for the off- and on-shelf regions ((a) and (b), respectively). The asterisks in (a) repre-sent the Ocean Station Papa (P26) summer and autumn mean eddy diffusivity (1 × 104 and 3 × 104 m2 s1, respectively) estimated by Cronin et al. (2015).Note that our Kmix accounts for vertical diffusivity, advection, and entrainment fluxes, while these latter values do not include advection and entrainment. Thehorizontal dashed lines represent the upper and lower Kmix values after accounting for mixed layer nitrification (see section 4.1.2). The y axis in (a) and (b) is on alogarithmic scale. (c and d) The corresponding subsurface [N2O]B gradients (derived in the 150 m below the mixed layer), which are used to estimate Kmix. Values inall panels are only presented for stations where we obtained profile measurements of N2O; Kmix is shown only where the NCP correction was not biased by asubsurface maximum in O2 concentrations (see section Biogeochemical Cycles 10.1002/2017GB005792IZETT ET AL. 9The N2O correction always increased our estimated NCP over uncorrected values. The magnitude of thecorrection factor was generally highest in the on-shelf region during the summer, with an average of75 mmol O2 m2 d1 difference (3–190 mmol O2 m2 d1 range) between corrected and uncorrected data.These corrections are equivalent to 23–327% of corrected NCP values. Prior to correction, several stations inthe southern on-shelf region had negative NCP values in spring and summer, suggesting net heterotrophicconditions. After correction, however, all estimates became positive, implying net autotrophy. The absolutemagnitude of the correction was lower in the off-shelf waters, with mean values of 23, 25, and31 mmol O2 m2 d1, during the winter, spring, and summer respectively, but the relative contributions tocorrected NCP were still high, (81, 52, and 74%, on average, respectively).We observed negative NCP near OSP during the winter. However, the observation of undersaturated N2O inthe mixed layer at these stations meant that a correction could not be applied. True NCP here is likely higherthan our uncorrected value, because our uncorrected value may be underestimated due to mixing of low-O2water from below (we discuss other correction approaches in section In the spring, NCP was slightlyhigher at OSP relative to other oceanic stations, consistent with the coincident positive anomaly in chl a(Figure 2a).Our final NCP shows a spatial gradient characterized by generally low values in the off-shelf region and high-est values on shelf (Figures 5 and 6). For each spring and summer cruise, high NCP was observed in the south-ern WCVI area, with maximum values of 162, 182, 396, and 189 mmol O2 m2 d1 in May, June, August, andFigure 5. Uncorrected and corrected NCP estimates in winter (February Line P; a and b), spring (June Line P, c and d; May La Perouse; e), and summer (August Line P, fand g; September La Perouse, h). The dashed black lines trace our final NCP estimates, taken as the unbiased N2O-corrected values (black markers) oruncorrected values (open-face markers) when corrected values were biased (asterisks). See text, section, for details of biased corrections. The Line P data arepresented as the average of the outbound and inbound transects, and the La Perouse data show all stations arranged by their proximity to shore. The dashedvertical gray lines show the locations of major stations along Line P, while the solid horizontal line indicates 0 NCP. Note the difference in scale between the off-shelf(a, c, and f) and on-shelf regions (b, d, e, g, and h).Global Biogeochemical Cycles 10.1002/2017GB005792IZETT ET AL. 10September, respectively. Wintertime measurements in the same region revealed significantly lower NCP, atjust ~24 mmol O2 m2 d1 (maximum) and a weaker spatial gradient between coastal and oceanicstations. Our repeat sampling of this region suggests that NCP is persistently high over the on-shelfsection of the Line P transect and between the LB and LC lines of the La Perouse survey during the springand summer (Figure 6). The off-shelf extent of this high productivity region varies seasonally, with elevatedNCP extending to ~50 km (126°W) from the coast in June and ~100 km of the coast (127°W) in August.Further north on the La Perouse track, we observed increased NCP on the transects adjacent to the coast(May; Figure 6b) and in Queen Charlotte Sound (September; Figure 6c). Off-shelf, during the spring andsummer, NCP was mostly low, except for the high anomalies observed at OSP and between stations P16and P20 in the spring. Relatively high NCP was also observed between P8 and P16 during the winter.4. DiscussionTo our knowledge, our work represents the first field application of the Cassar et al. (2014) method ofN2O-based corrections for ΔO2/Ar-derived NCP estimates. This approach is significant in enabling morerobust productivity estimates in physically dynamic regions, where vertical mixing is a nonnegligible termin the mixed layer O2 mass balance. Below, we discuss our results in terms of both methodological andpractical considerations (section 4.1) and the interpretation of spatial and temporal patterns in NCP(section 4.3). In section 4.1.5, we outline some key limitations of, and alternatives to, the current approachand provide a discussion of uncertainties. The supporting information contains a detailed explanation ofthe uncertainties.4.1. Methodological Considerations, Assumptions, and LimitationsWe begin by addressing methodological considerations of the NCP correction, expanding on the work ofCassar et al. (2014), with emphasis on the implications for field studies. We focus on the use of differentapproaches to derive the O2/N2O supply ratio term, the possible impact of mixed layer nitrification, thermalsolubility effects on mixed layer N2O, and potential deviations from subsurface Ar saturation. We summarizeby discussing the limitations of the current approach, outlining remaining uncertainties, and providingrecommendations for future applications.Figure 6. Spatial distribution of NCP in winter (a), spring (b), and summer (c), showing all outbound and inbound Line P data (overlaid). The N2O correction was lin-early interpolated to the resolution of our underway ΔO2/Ar data.Global Biogeochemical Cycles 10.1002/2017GB005792IZETT ET AL. 114.1.1. Considerations for the [O2]B/[N2O]B Supply Ratio TermCassar et al. (2014) estimated the subsurface-to-surface supply ratio of O2/N2O based on a two-point slope,with end-members within and below the mixed layer. Using this approach, we obtained some supply ratioterms that resulted in unrealistic NCP values (<300 or >400 mmol O2 m2 d1; Figure 3d) based on therange of previous estimates in our study area (see section 4.3). The two-point slope approach is highly sensi-tive to the selected depth of the end-members (i.e., the depth of discrete Niskin bottle sampling) and is thussusceptible to even minor irregularities in the profile data. For these reasons, we used an alternativeapproach, estimating the supply ratio as the multipoint slope through the pooled data sets of [O2]B and[N2O]B in the 150 m below the mld for each respective cruise.The depth-dependent increase in [N2O]B and corresponding decrease in [O2]B that we observed at moststations yielded a narrow range of supply ratio estimates and is consistent with nitrification being the domi-nant source of N2O in our study area (Capelle & Tortell, 2016; Grundle et al., 2012). Our supply ratio estimatesare also consistent in space and time along the Line P and La Perouse transects, and the difference betweenvertical [O2]B/[N2O]B gradients for individual profiles during a single cruise was not statistically significant.This result justifies the use of a pooled data set for each cruise. We selected an integration depth of 150 mas it extends fully beyond the maximum depths of the mixed layer and euphotic zone observed duringthe 2016 cruises, thereby capturing the subsurface regions of increased N2O production. A similar depth ofintegration for vertical mixing (120 m) was used by Ianson and Allen (2002) in a physical model of theWCVI region.The supply ratios derived by our multipoint slope approach did not produce any unrealistic negative NCPvalues (Figure 3d), and our estimates from the 2012–2016 data agree well with O2/N2O gradients reportedin the region and with the global average of ~1 × 104 mmol O2 (mmol N2O)1 (Cohen & Gordon, 1979;Nevison et al., 2003; Oudot et al., 1990; Rees et al., 1997). In the absence of depth profile data, a gradientof 1 × 104 mmol O2 (mmol N2O)1 may therefore constitute a suitable alternative in most regions.Indeed, we observed only small differences between NCP estimated using our data-derived supply ratioterms and the global average stoichiometric value (mean difference of 11.3 mmol O2 m2 d1; Figure 3d),with the largest deviations occurring in coastal waters with high surface [N2O]B concentrations. Moreover,the NCP correction is generally insensitive to the depth of integration of the supply ratio term, provided thatsufficient points are selected to accurately estimate a regression slope and that a region of subsurface nitri-fication is included (i.e., using integration depths that extend beyond the euphotic zone). When the gradientwas estimated as the slope integrated over the 500 m below the mld, for example, we observed a meandeviation of 6.6 mmol O2 m2 d from our corrected NCP estimates (Figure 3d). Similarly, we found thatvarying our mld estimates by ±5 m resulted in less than 3.8% variation in the supply ratio terms(<1 mmol O2 m2 d1 NCP). This demonstrates that our results are not overly sensitive to our choice of a0.125 kg m3 density difference criterion to define the base of the mixed layer. In all cases, the spatialpatterns of NCP remain similar using different slope-derived supply ratio terms (data not shown), but theamplitude of the correction changes slightly. As we describe below in section 4.1.5, caution should nonethe-less be exercised before applying the N2O based approach in some scenarios.4.1.2. Mixed Layer NitrificationNitrous oxide is producedmainly through nitrification, as a by-product of the oxidation of ammonium (AO) toNO2 and NO3 or through incomplete denitrification under low-O2 conditions (Codispoti et al., 2005; Ward,2000). It is believed that nitrification is photoinhibited in surface waters (Horrigan et al., 1981), while highmixed layer O2 concentrations are not favorable for denitrification. The net result is that N2O has typicallybeen assumed to be produced mainly in subsurface waters (e.g., Dore & Karl, 1996), and this assumption isapplied in the N2O correction (equation (3)). Yet recent work suggests that N2O production may occur withinthe euphotic zone (Grundle et al., 2013; Smith et al., 2014; Zamora & Oschlies, 2014). Cassar et al. (2014) sug-gested that mixed layer N2O production would lead to an overestimate of the vertical mixing term and over-correction of NCP. Below, we attempt to quantify the potential significance of this limitation.During nitrification, [N2O]B increases, while ΔO2/Ar decreases as O2 is consumed through AO. A similar netchange occurs through mixing, such that the contribution of mixing, and subsequently the NCP correction,is overestimated when nitrification is neglected. We define NCP as net O2 production resulting from grossphotosynthesis, minus community respiration of organic carbon, such that only respiratory processesGlobal Biogeochemical Cycles 10.1002/2017GB005792IZETT ET AL. 12affecting the organic carbon pool are considered. As such, nitrification is not an inherent component of ourdefinition of NCP, and the NCP equation (equation (3)) can be adjusted, accounting for the effects of this pro-cess on both O2 and N2O (refer to the supporting information for the derivation).NCP ¼ kO2 · ΔO2=Ar· O2½ eq–kN2OkO2·∂ O2½ B∂ N2O½ B· N2O½ B !–N· 1–ΔN2OΔO2·∂ O2½ B∂ N2O½ B !(8)Here N is the nitrification rate in terms of O2 consumption (≤0 mmol O2 m2 d1), and ΔN2OΔO2 is the mixed layerN2O stoichiometric yield of nitrification (≤0 mmol N2O (mmol O2)1).Applying equation (8), we calculate the potential impact of mixed layer nitrification on our 2016 NCP esti-mates, considering only those locations where the NCP correction was not biased (see section et al. (2013) reported mean mixed layer nitrification rates of ~20 nM NH4+ d1 along the off-shelfLine P transect, which is equivalent to O2 consumption of 0.04 mmol O2 m3 d1, based on the stoichiometryof the reaction: NH4+ + 2O2→NO3 + H2O + 2H+. Since rates of nitrification vary geographically, we use thisregional value in our calculations (Table 1), recognizing that it falls within the range of globally reported rateestimates for near-surface waters (0–~130 nM NH4+ d1; Dore & Karl, 1996; Santoro et al., 2010; Ward, 2005).Based on 2016 mlds from our data, we derived mixed layer-integrated nitrification rates (N in equation (8))between 0.4 and 3.6 mmol O2 m2 d1.The impact of surface water nitrification on the N2O correction approach depends on the N2O yield of thisprocess. The stoichiometric N2O yield of nitrification (ΔN2OΔO2) varies widely under different substrate concentra-tions (Frame & Casciotti, 2010; Goreau et al., 1980) and will often differ from that of vertical mixing (∂ N2O½ B∂ O2½ B ; thereciprocal of the supply ratio). In the case where these two terms are the same, nitrification will have no neteffect on the NCP correction, since equation (8) simplifies to equation (3). For the other (more likely) scenario,we can derive lower and upper bounds on our calculations, using low (1.4 × 104 mol N2O (mol O2)1;Yoshida et al., 1989) and high (3.8 × 104 mol N2O (mol O2)1; Santoro et al., 2011) N2O yields reportedunder O2 and NO2 concentrations similar to those observed in our study area.The results of our calculations show that nitrification could produce between 0.04 and 0.26 μmol m3 N2O inthe mixed layer, which is equivalent to 0.4–2.2% of mean 2016 surface concentrations (Table 1). These valuestranslate into net reductions of NCP between 0.3 and 11.2 mmol O2 m2 d1, through the combined influ-ence of altered O2 and N2O concentrations. On one hand, nitrification acts to decrease apparent NCP byconsuming O2 (equation (3)). Conversely, nitrification increases the surface N2O saturation anomaly, therebyleading to an overestimate of the mixing correction term (i.e., leading to overcorrection of NCP). In our calcu-lations, we found that nitrification would increase NCP by between 0.4 to 3.6 mmol O2 m2 d1 through itseffects on surface O2 and decrease NCP by between 0.8 and 14.9 mmol O2 m2 d1 through increased N2O.These independent effects sum to produce the net influence on NCP (presented in Table 1). Similarly, Kmix(equation (S1)) could be between 0.7 and 58% lower than values calculated without consideration of nitrifi-cation (Figure 4).We thus conclude that the effect of nitrification on NCP estimates could be substantial in our study regionunder conditions where background NCP and mixing rates are low (e.g., in open-ocean waters). Indeed, atthe upper limit, the bias in NCP derived from mixed layer nitrification could be as much as 42% of the appar-ent NCP. By comparison, in productive coastal waters, the effect of surface nitrification tends to introduceonly a small error (<4% in spring and summer). Direct observations of nitrification rates are required to betterconstrain the influence of this process on surface NCP estimates. In the absence of such data, we continue tosupport the assumption made by Cassar et al. (2014), since the uncertainty estimated here (mean5.5 mmol O2 m2 d1, or 16% of estimated NCP based on high N2O yields) is of similar magnitude and typi-cally smaller than errors in the gas transfer parameterizations (mean 4.4 mmol O2 m2 d1; see the support-ing information). We note, however, that NCP measurements made by this approach may still constitute anupper bound on true values.4.1.3. Thermal Correction for [N2O]BFollowing Cassar et al. (2014), we applied a thermal adjustment to mixed layer measurements of [N2O]B(equations (5) and (6)) to account for solubility effects. Recent warming (cooling) prior to samplingGlobal Biogeochemical Cycles 10.1002/2017GB005792IZETT ET AL. 13Table1ThePotentialEffectsofSurfaceNitrificationonMixedLayerN2OConcentrations,NCPEstimates(ΔNCP),andMixingCoefficients(ΔK mix)Region2016dataCalculationsmldτMeanN2O∂O2½B∂N2O½B∂N2O½B∂ZMixedlayernitrificationrateN2OproducedΔNCPΔK mix(m)(days)(μmolm3)(mmolmmol1)(mmolm4)(mmolO2m2d1)(μmolm3)(%)(mmolO2m2d1)(%)(m2s1)(%)WinterOff90.516.511.951.08×1041.07×1043.×10520.×10457.8On69.016.311.472.×10515.×10443.9SpringOff28.110.810.781.40×1049.55×1051.×1057.×10520.5On14.18.812.790.×1061.×1053.7SummerOff25.58.39.401.29×1041.05×1041.×1054.×10512.1On10.916.816.×1060.×1051.9Note.WeestimatedtheproportionalcontributionofinsitunitrificationtosurfaceN2Oconcentrations(N2Oproduced)basedonliterature-derivedvaluesofmeanammoniumoxidationratesandN2Oyieldsofnitrification(seetextfordetails).TheeffectsonNCPandK mixwereestimatedfollowingequations(8)and(S1).MixedlayerdepthsandN2Oresidencetimes(τ)representtheaveragevaluesfortheoff-shelfandon-shelfregions,respectively,ineachseason.Thetworowsofcalculationsforeachregionrepresentvaluesbasedonthelow(shadedgray)andhighN2Oyieldesti-mates,respectively.Percentvalues(%)representtheproportionalchangeinavariablerelativetovaluesestimatedwhennitrificationwasneglected.Global Biogeochemical Cycles 10.1002/2017GB005792IZETT ET AL. 14decreases (increases) N2O solubility and increases (decreases) theapparent saturation anomaly, causing a N2O efflux (influx) at thesurface. Since the residence time of gases in the mixed layer is typi-cally 2–3 weeks, atmospheric gas exchange lags behind changes ingas solubility. As a result of warming (cooling), mixed layer [N2O]Bwould appear oversaturated (undersaturated), and the NCP correctionwould be too large (small). Figure S2 shows [N2O]B with and withoutthe thermal adjustment, and the resulting effect onNCP is proportionalto the difference in observed concentrations. Using our unbiasedcorrected NCP, we find that the effect of the N2O thermal correctionis generally small in absolute magnitude (<11.4 mmol O2 m2 d1)but proportionately large (up to 25%), in low productivity, off-shelfregions (Table 2). For air-sea heat fluxes exceeding 100 W m2 inmagnitude (spring and summer), estimated NCP would be more than7.8 mmol O2 m2 d higher in the absence of thermal adjustment.When NCP is relatively low, as in much of the off-shelf regions, suchlarge heat fluxes can contribute >20% bias in final NCP estimates.We thus recommend the continued application of this correction forNCP estimates in off-shelf waters, despite the relatively large errors associated with estimates of surface heatfluxes (see the supporting information Table S2). By comparison, the thermal correction makes a relativelysmall (mean ~9%) contribution to the overall uncertainty in coastal NCP, due to the large magnitude of NCPestimates here but should still be included since thermal fluxes are likely to vary widely in space and time.4.1.4. Ar Saturation Below the Mixed LayerIn the calculation of the vertical [O2]B term, the NCP correction approach assumes that subsurface Arconcentrations are at saturation. Departures from this impact the derived supply-ratio term. Recent workin the North Pacific by Hamme and Emerson (2002) and Emerson et al. (2012) suggests that subsurfaceAr concentrations rarely deviate from saturation by more than ±4%. In the absence of correspondingprofile measurements of Ar, we recalculated all supply ratio terms, assuming ΔAr = ±0.04, to estimatethe upper limit on the potential uncertainty in ∂ O2½ B∂ N2O½ B . We found that gradients were up to 3.2% steeper(shallower) when Ar was supersaturated (undersaturated). This source of error is small compared withthe larger uncertainties in the gas transfer velocity parameterization and supply ratio terms (see thesupporting information), and we thus suggest that deviations in Ar from equilibrium can be neglectedwithout a significant influence on the overall results.4.1.5. Discussion of the Approach Limitations and Uncertainties4.1.5.1. Scenarios Where the N2O Correction Approach Is InvalidThere are several scenarios for which NCP corrections based on the O2/N2O supply ratio may not be suitable.In much of the Arctic Ocean, for example, a complex N-cycle (Brown et al., 2015) and lack of a consistent rela-tionship between O2 and N2O (Fenwick et al., 2017) invalidate this approach and likely preclude the use ofN2O corrections. Furthermore, regions where the subsurface [O2]B/[N2O]B supply ratio does not representthe relevant gradient for vertical O2 mixing pose a limitation to the correction approach of Cassar et al.(2014). Principally, this occurs under situations where subsurface O2 maxima cooccur with localized minimain N2O and a negative relationship between [O2]B and [N2O]B persists. Here the intuitive NCP correctionshould be to lower uncorrected values, since the upward supply of relatively high-O2 water from beneaththe mixed layer would introduce a positive ΔΟ2/Αr anomaly. However, since the supply ratio stoichiometryof [O2]B/[N2O]B remains negative, a false positive correction term is derived (i.e., correctedNCP > uncorrected NCP), leading to an overestimation of NCP, if equation (3) is applied. We observed this“biased” scenario at a number of off-shelf stations in June and August and chose to present uncorrected datafor these regions in the absence of a suitable correction factor (see section below). The same limitationwould also occur in regions where [N2O]B and [O2]B both decline below the mixed layer (∂[O2]B/∂[N2O]B> 0),as might be seen where near-surface oxygen minimum zones result in N2O loss through denitrification (e.g.,Farías et al., 2009). In these cases, the positive supply ratio would yield a negative correction term, eventhough vertical mixing would have introduced a low-ΔO2/Ar anomaly.Table 2Influence of Air-Sea Heat Flux (Q, Equation (6) on Mixed Layer N2O Budgets DuringOne Residence Time of N2O Prior to SamplingRegionAir-seaheat fluxSea-air thermalN2O flux ΔNCP(W m2) (μmol m2 d1) (mmol O2 m2 d1) (%)Winter Off 18.8 0.13 1.4 5.3On 17.0 0.11 1.2 5.6Spring Off 126.7 0.75 11.4 25.2On 110.3 0.61 8.2 11.6Summer Off 120.4 0.57 7.8 18.3On 123.6 0.65 8.0 7.5Note. The effect of thermally derived N2O changes on NCP was estimated bycomparing calculations with and without the thermal N2O correction (ΔNCPrepresents the change in NCP if the thermal adjustment is not made).Negative heat fluxes represent ocean cooling, which induce negative sea-airN2O fluxes according to equation (6).Global Biogeochemical Cycles 10.1002/2017GB005792IZETT ET AL. 15It is thus evident that the NCP approach applied here is invalidated if N2O decreases below themixed layer. Ingeneral, the approach works well when subsurface [N2O]B increases, and [O2]B decreases, as was the case atall coastal stations (where we found the largest correction terms) and much of the off-shelf region. Yet ourextensive O2 and N2O profile measurements show that the approach, as described by Cassar et al. (2014),is not likely to be valid everywhere. At several stations, we observed deficits in subsurface N2O that corre-sponded with relative maxima in O2 (e.g., Figure S3). Although we lack profile measurements of N2O at everysampling station, we can utilize our CTD-derived O2 measurements to diagnose this limitation. By indepen-dently evaluating the depth-dependent O2 slope in the 25 m below the mixed layer (∂[O2]B/∂Z|25) in everyprofile, we observed a number of stations with positive gradients in the off-shelf region during the springand summer (Figures 3 and S4). Where profiles of N2O existed, corresponding deficits in N2O, resulted in anegative O2/N2O gradient that was not significantly different than all other stations (Figure S3e). The obser-vation of a consistent negative supply ratio term in all profile data resulted in positive correction terms, andoverestimation (i.e., positive bias) of true NCP when the N2O-based approach was used.It is important to note that the O2 minima/N2O maxima that we describe here are shallow features, localizedto the waters just below the mixed layer (Figure S3). Beyond these depths, we typically observe declining O2and increasing N2O concentrations. For this reason, the ∂[N2O]B/∂Z values we use to derive Kmix terms (inte-grated down to 150 m below the mixed layer) are not explicitly negative for the regions where the NCPcorrection is biased (Figures 4c and 4d). We simply lack the depth resolution of N2O measurements to betterconstrain the relevant slope over a shallower depth range. We find the presence of these N2O minima to besomewhat counterintuitive and difficult to explain, since O2 concentrations are clearly high enough toprevent N2O loss through denitrification.Finally, we also note that the current approach breaks down at other stations where mixed layer N2O isundersaturated ([N2O]B< 0; e.g., P20–P26 in February). This scenario results in a nonsensical negative mixingcorrection term.We therefore suggest that stations that do not exhibit significantly negative subsurface [O2]B gradients, orwhere surface N2O is undersaturated, should be precluded from the [O2]B/[N2O]B supply ratio-based NCPcorrection. We use an upper ∂[O2]B/∂Z|25 cutoff threshold of 0.1 mmol O2 m4 (Figure S4) as mixing fluxesusing this value are small (<4 mmol O2 m2 d1) for the mixing rates we observe off shelf. If O2 profiles arenot available to diagnose the suitability of the N2O correction, this limitation may apply in oceanic regionswith strong density stratification and deep euphotic zones, where subsurface primary productivity maygenerate an O2 maximum or when water masses containing signatures of high productivity are subductedbeneath surface layers. This scenario is prevalent in much of the North (tropical to midlatitude) Pacific duringthe summer time (Shulenberger & Reid, 1981). Yet strong density stratification in these regions will likely actto minimize the impact of vertical mixing fluxes, thus limiting the potential importance of NCP corrections(Giesbrecht et al., 2012). Nonetheless, future studies should focus on O2 and N2O dynamics in these waters,to understand the cause of subsurface N2O deficits and to evaluate whether N2O-based corrections couldbe feasible. Alternatives to the N2O Correction ApproachAs an alternative to the N2O approach, a correction can be made using estimates of a mixing rate andobserved subsurface O2 gradients (i.e., Kmix · ∂[O2]B/∂Z). When we apply this correction to our data usingaverage off-shelf Kmix values for the respective seasons, we obtained correction terms exceeding75 mmol O2 m2 d1 at stations P20–P26 in February and final NCP of ~50mmol O2 m2 d1. These elevatedNCP values are more than twofold higher than previous estimates (e.g., Giesbrecht et al., 2012; Palevsky et al.,2016; Yang et al., 2017). We therefore believe these values to be the unrealistic result of a combination of verystrong subsurface O2 gradients and overestimation of Kmix using the off-shelf (station P8–P16) average value.Indeed, the subsurface [N2O]B gradients at stations P20 and P26 were steeper than at the other off-shelfstations (Figure 4c), such that the Kmix terms are likely to be smaller than predicted by the average value.Similarly, off-shelf ∂[N2O]B/∂Z differs between stations in the summer, and average Kmix values may be a poorrepresentation of the regional mixing rates.Our study clearly demonstrates the spatial and temporal variability of oceanic mixing rates (Figure 4). We thuscaution against the use of non-site-specific and literature-derived estimates of Kmix or eddy diffusivity whenGlobal Biogeochemical Cycles 10.1002/2017GB005792IZETT ET AL. 16deriving NCP corrections. To examine the sensitivity of estimates to the choice of mixing rates used forcorrection, we recalculated corrected NCP using the mean summertime eddy diffusivity value of Croninet al. (2015) (1 × 104 m2 s1) and our independent subsurface O2 gradient estimates (Figure 3d). On average,these estimates differed from our N2O-corrected values by 32.4 mmol O2 m2 d1 (~50% mean). This resultunderscores the importance of independent estimates of mixing rates for each respective station. In theabsence of such direct estimates, we suggest the use of uncorrected NCP as the most viable option atpresent. This seems reasonable given the low apparent contribution of mixing fluxes to surface ΔO2/Ar inthe summertime (Giesbrecht et al., 2012; Palevsky et al., 2016; Yang et al., 2017). In coastal regions, wheremix-ing corrections are largest, the O2 and N2O profiles permit the use of the N2O-based correction approach. Insummary, our analyses highlight the importance of examining the shape of the O2 profiles before applyingthe N2O-based NCP correction. Additional Limitations of In Situ NCP EstimatesAdditional limitations and uncertainties (see section S2 for a detailed error analysis) of the present approachare inherent to the majority of ship-based estimates of NCP from surface ΔΟ2/Ar measurements. Specifically,NCP derived from O2 mass balance equations (equations (1) and (3)) assumes steady state of surface O2concentrations and negligible horizontal fluxes. The present approach provides corrections for verticalmixing (diffusion, entrainment, and advection) but makes no attempt to rectify potential fluxes from theseadditional sources.In a Southern Ocean Lagrangian study, Hamme et al. (2012) observed absolute changes in ΔO2/Ar up to 0.6%,partially attributable to diurnal cycles in productivity. We found that ΔO2/Ar varied by 3% (from ~3.5 to 6.5%)during a 1 day occupation of OSP in June, with lower variability (mean 0.1% d1 off-shelf and 0.5% d1 on-shelf; overall maximum 2.5% d1) observed between outbound and inbound sampling of the same locationsalong the Line P cruises. Thus, while the timing of our sampling (i.e., during the diel cycles) has an effect onthe magnitude of our ΔO2/Ar measurements, the uncertainty in NCP attributed to temporal changes in ΔO2/Ar (~0.6 and 1.1 mmol O2 m2 d1 off-shelf and on-shelf, respectively) is smaller than errors associated withgas flux parametrizations (~4.4 mmol O2 m2 d1).Off-shore, lateral advection is likely to be negligible, but near the coast, a combination of high current speedsand patchiness in phytoplankton biomass may make horizontal fluxes significant. Using net current velocitydata (data provided by the NASA Physical Oceanography Distributed Active Archive Center and retrievedfrom https://doi.org/10.5067/OSCAR-03D01; Bonjean & Lagerloef, 2002; Earth Space Research, 2009) andderived estimates of the surface lateral [O2]B gradient (from our underway ΔO2/Ar and thermosalinographdata), we estimated mean horizontal fluxes of ~1.6 and 22.7 mmol O2 m2 d1 off-shelf and on-shelf,respectively. This level of uncertainty applies to measurements of NCP in Eulerian studies. However, in aLagrangian sense, lateral advection influences the location assigned to a given NCP measurement butnot the rate itself, so these flux estimates are in many cases an exaggeration of the influence of lateraladvection on NCP rates.Finally, the relatively low spatial resolution of our surface N2O measurements (mean of 102, 37, and 47 km inour winter, spring, and summer cruises respectively) also limits the application of this approach on highlyresolved space scales. We obtained estimates of NCP on the sub-kilometer scale (as in Figure 6) by linearlyinterpolating the N2O correction term to the resolution of our underway ΔΟ2/Αr measurements. However,we estimate a mean uncertainty in these values of 5.5 mmol O2 m2 d1 by comparing estimates derivedthrough linear and nearest-neighbor interpolation schemes (see the supporting information for moredetails). This additional uncertainty motivates the need for improved resolution of N2O measurements.Our comparison between samples obtained from Niskin and the ship’s seawater supply (Figure S1) showsthat high quality N2O measurements can be derived without the need for discrete, on-station bottlesampling, suggesting good possibilities for continuous underway analysis. Nitrous oxide measurementsare not common on most oceanographic field programs, and existing data are largely based on discretesampling. However, recent developments in instrumentation (Arévalo-Martínez et al., 2013; Grefe &Kaiser, 2014) should provide additional impetus for the development of ship-based underway N2Omeasurements to complement mass spectrometry-based ΔO2/Ar data. This would constitute an importantnext step in the application of the correction approach, enabling sub-kilometer spatial resolution of N2O-corrected NCP estimates.Global Biogeochemical Cycles 10.1002/2017GB005792IZETT ET AL. 17Overall, in spite of the above mentioned limitations and uncertainties, the N2O approach we evaluate hererepresents an improvement over existing techniques for evaluating NCP from ship-based measurementsof ΔΟ2/Αr. Specifically, the total mean analytical and parameterization uncertainty in our data is9.2 mmol O2 m2 d1 (~34 and 21% of off- and on-shelf estimates), with the correction factor contributingroughly 4.3 and 13.7 mmol O2 m2 d1 in each region (Table S2). Although the methodological uncertaintiesare large in some cases (combined maximum of ~16 and 26 mmol O2 m2 d1 off-shelf and on-shelf), theserepresent conservative estimates. Moving forward, constraining the rate of upper ocean N2O production,improving reanalysis products for the estimation of oceanic heat flux, and reconciling surface measurementsof ΔO2/Ar for lateral advection would significantly improve confidence in N2O-derived vertical mixing correc-tion terms, making this an important tool for resolving spatial and temporal patterns in NCP.4.2. Correction Factor and KmixPrevious studies that have corrected for mixing processes on the mixed layer O2 mass balance have usedliterature values for eddy diffusivity coefficients (e.g., Bushinsky & Emerson, 2015; Weeding & Trull, 2014).These indirectly obtained values may obscure potentially significant variability in both space and time (seesection Moreover, these corrections only represent diffusive fluxes, such that additional means ofquantifying advection or entrainment are required. This introduces additional uncertainty to the estimatedmixing terms. Some approaches have relied on model-derived mixing fluxes to evaluate all three terms(e.g., Plant et al., 2016). Our approach does so empirically.In general, the magnitude of Kmix (Figure 4) reflects regional and temporal differences in upwelling, watercolumn stratification, and wind-induced mixed layer turbulence. We observed the largest apparent mixingcoefficients (and NCP correction factors) in the on-shelf region following periods of upwelling (based onBakun Upwelling Index values at 48°N, 125°W and 50°Ν, 131°W provided by the Pacific FisheriesEnvironmental Laboratory and retrieved from http://www.pfeg.noaa.gov/products/PFEL/modeled/indices/upwelling/NA/data_download.html; Bakun, 1973). Indeed, our mean on-shelf Kmix values strongly correlatewith the regional upwelling index (R2 = 63%; Figure S5). This result, and the occurrence of elevated on-shelfsurface [N2O]B concentrations during the spring and summer upwelling periods (Figure S2), is supported byprevious observations showing that upwelling is the dominant N2O transport process over the continentalshelf along the WCVI (Capelle & Tortell, 2016). Strong stratification in off-shelf regions, resulting from highsolar radiation and mixed layer shoaling, is likely responsible for the smaller mixing fluxes and coefficientsin this region, during the summer. Higher winter and spring Kmix at the off-shelf stations, relative to summer,is likely due to increased wind speed (Figures S6a–S6c), which is directly relatable to the strength of eddyturbulence and Ekman pumping (Kundu et al., 2012). The relative decline in Kmix from P12 to P26 in the springis mostly attributed to reduced wind intensities along the same gradient. Although wind speeds are highestduring the winter, strong temperature stratification, caused by the persistent thermocline located just belowthe mixed layer at this time of year (data not shown), resists particularly large mixing rates. The low winter-time Kmix values at the coastal Line P station, P4, reflects the net downwelling that occurs during this time(Figure S5a). The observation of undersaturated N2O in the winter waters at OSP and profiles with negativeN2O gradients (positive O2 gradients) in the spring and summer makes it difficult to estimate mixing atsome locations.Overall, our estimates of Kmix are consistent with the multiyear OSP time series of Cronin et al. (2015), whoused heat and salt budgets to estimate vertical diffusivities. The results of their work suggest an annual rangeof mixing rates from ~1 × 105 to 1 × 102 m2 s1. Our derived estimates, based on N2O observations (withand without considerations for potential mixed layer nitrification), are fully encompassed in the rangereported by Cronin et al., as well as the direct depth-resolved measurements of eddy diffusivity at OSP(~1 × 105 –1 × 103 m2 s1 below the mixed layer) of Rousseau et al. (2010). Our Kmix values also showreasonable seasonal and spatial patterns. The increase in mixing coefficients closer to shore (in spring andsummer) fits the global data set of Whalen et al. (2012), who identified elevated mixing rates over roughbottom topography, and near the coast. Indeed, the on-shelf WCVI region is characterized by heterogeneousbottom topography, with numerous canyons, that can result in intensified mixing (Allen & Durrieu deMadron, 2009). This result indicates that the N2O correction approach can provide significant insight intophysical processes occurring in near-surface ocean waters.Global Biogeochemical Cycles 10.1002/2017GB005792IZETT ET AL. 18It is important to note that our Kmix represent the combined fluxes due to diffusion, advection, and entrain-ment, whereas other estimates typically express mixing rates in terms of diffusivity coefficients. While we donot evaluate the relative contributions of each process to the total vertical flux, others (e.g., Yang et al., 2017)have shown that vertical diffusivity dominates over entrainment or advection in the off-shelf waters. Incoastal waters, during upwelling events, however, vertical advection is likely to play a considerable role. Itis therefore understandable that our Kmix values are comparable to, but on the upper range of, values derivedfrom other data sets, particularly in regions impacted by upwelling, or subject to significant entrainment.4.3. Spatial Patterns of NCPOur data contribute to our knowledge of NCP in the Subarctic Pacific by providing refined estimates acrossbroad spatial scales and contrasting oceanographic regimes. Much of our understanding of NCP in theSubarctic NE Pacific comes from previous studies in the vicinity of OSP (Bushinsky & Emerson, 2015;Fassbender et al., 2016; Howard et al., 2010; Lockwood et al., 2012; Palevsky et al., 2016; Plant et al., 2016) orat discrete locations along the Line P transect (Giesbrecht et al., 2012; Hamme et al., 2010). By comparison,fewer studies have quantified NCP in the coastal regions of the Subarctic Pacific, owing to the difficulties inconstraining vertical mixing fluxes (e.g., Jonsson et al., 2013), limiting our understanding of this ecologically sig-nificant regime. Here we compare our corrected NCP estimates to values reported elsewhere along the Line Ptransect and discuss the apparent spatial patterns in the context of observed biological and physical forcing.Our NCP estimates for the oceanic regions are consistent with those previously derived from amixed layer O2mass balance approach in the Subarctic Pacific, but the upper range of our values exceeds any previous esti-mates (~50, 81, and 180 mmol O2 m2 d1 in winter, spring, and summer, respectively). At OSP, for example,previous spring and summer O2-based NCP estimates have ranged from ~10 to 60 mmol O2 m2 d1(Bushinsky & Emerson, 2015; Fassbender et al., 2016; Giesbrecht et al., 2012; Plant et al., 2016). Our correctedresults from OSP measurements (58 mmol O2 m2 d1 in spring) are at the upper range of estimates of thesestudies and are similar to the anomalously high NCP (~65 mmol O2 m2 d1) observed following a largeatmospheric dust deposition event in 2008 (see below; Hamme et al., 2010). Only one previous study(Lockwood et al., 2012) reported summertime rates (~84 mmol O2 m2 d1) higher than ours. Likewise, alongthe rest of the off-shelf Line P transect, many of our springtime NCP estimates (>40 mmol O2 m2 d1)exceed or represent the upper range of any previously measured values (~10–48 mmol O2 m2 d1;Giesbrecht et al., 2012; Hamme et al., 2010; Palevsky et al., 2016).While interannual variability in NCP partially explains differences between our observations and those of pre-vious studies, it is clear that the contribution of mixing fluxes is also significant. The difference between ourresults and previous studies is particularly evident in the coastal waters over the continental shelf. At stationP4, our average corrected NCP value in August (180 mmol O2 m2 d1) is more than fourfold higher than theprevious summertime estimates of Hamme et al. (2010) and Giesbrecht et al. (2012). Notably, our uncorrectedNCP estimate for this station agrees well with these prior measurements, suggesting that the discrepancyresults from an underestimation of NCP in previous studies that have neglected the influence of vertical mixing.This result emphasizes the importance of accounting for negative O2 fluxes due to subsurface-to-surface mix-ing, especially in coastal regimes. Net community production estimates derived fromprevious studies without acorrection for mixing should thus be considered as a lower bound on true values, particularly near to the coast.To our knowledge, only Bushinsky and Emerson (2015), Palevsky et al. (2016), Plant et al. (2016), and Yanget al. (2017) have included physical vertical transport terms when evaluating NCP from O2 measurementsin the Subarctic Pacific. Only Palevsky et al. evaluated spatial patterns in NCP from ship-based data, usingmodel-derived corrections for mixing. Giesbrecht et al. (2012) considered mixing, but found vertical fluxesto be negligible in the spring and summer and therefore excluded any explicit corrections. This resultsupports our choice to not apply corrections in off-shelf waters where the observed N2O gradients resultin biased mixing coefficients (see section the consideration of absolute values, our NCP estimates show spatial and temporal patterns that areconsistent with expectations for the Subarctic NE Pacific. Specifically, our data broadly reflect trends in phy-toplankton biomass (using chl a as a proxy) and demonstrate variability that suggests environmental controlfrom surface nutrient concentrations, seawater temperature, and light availability. The strong NCP gradientfrom high on-shelf, to relatively low off-shelf production in spring and summer, is an example of this. InGlobal Biogeochemical Cycles 10.1002/2017GB005792IZETT ET AL. 19the off-shelf waters of the Line P transect, our measurements likely reflect a progression fromwintertime lightlimitation, to a N- and Fe-limited regime in the late summer, based on insights derived from previous studies(Boyd et al., 1996; Maldonado et al., 1999; Peña & Varela, 2007). The on-shelf waters show typical spring andsummer blooms that exceed wintertime production, but these trends are not observed in the uncorrectedNCP data, due to the effects of summer upwelling and deep tidal mixing (e.g., in the JF region) thatpartially mask in situ O2 production. In some cases, these mixing processes result in apparent netheterotrophy over parts of the continental shelf.Across broad spatial and temporal scales, our corrected NCP shows strong correlations with chl a, SST, andmld (proxy for mixed layer light levels and stratification) (Figures 7a–7c), except under conditions ofpresumed macronutrient or micronutrient limitation. We infer N limitation in the spring and summer fromnear-depleted macronutrient concentrations (~50–400 km off-shelf in June and ~50–700 km in August;Figures 2d and 2e), and summertime Fe limitation west of these regions where the significant increase inNO3 + NO2 concentrations is indicative of previously observed high-nutrient, low-chlorophyll (HNLC) con-ditions (Harrison et al., 1999; Maldonado et al., 1999). Significantly, we found that the correlations betweenNCP and the ancillary data were weaker when we used the uncorrected data (based on the R2 values ofempirical best fit equations; Figure 7). This suggests that the corrected values provide a better estimate ofthe broad spatial patterns and environmental controls on NCP across our survey region. The relationshipswe observed between NCP and chl, SST, and mld are consistent with results from previous studies in otheroceanic systems (Eppley, 1972; Li & Cassar, 2016; Tortell et al., 2015).Our results also show some localized NCP patterns, such as the wintertime net heterotrophy at OSP (P26), andincreased NCP during the spring and summer in the off-shelf waters. While the mechanism driving heterotro-phy in subarctic gyres is currently debated, this phenomenon has been observed at OSP previously(Bushinsky & Emerson, 2015; Fassbender et al., 2016; Plant et al., 2016). It should be noted, however, thatFigure 7. Relationship between corrected NCP and chl a (a), SST (b), and mld (c). (d–f) The same relationships for uncorrected NCP. The numbers in each panel repre-sent the correlation coefficients (R2) for the respective empirical fits. The gray asterisks represent data from regions of presumed N or Fe limitation; these data wereexcluded from the correlations.Global Biogeochemical Cycles 10.1002/2017GB005792IZETT ET AL. 20we could not apply a mixing correction at station P26 during the winter, and therefore, our NCP represents alower bound. In contrast, OSP appeared to experience unusually high phytoplankton biomass and NCP in thespring of 2016, coincident with an anomalous increase in chl a (Figure 2a). Intermittent Fe fertilization eventshave been proposed as a mechanism for elevating phytoplankton production at OSP (Boyd et al., 1996). Forexample, enhanced aerosol deposition (associated with the eruption of Aleutian volcanoes) during summer2008 was implicated as a source of Fe to the Subarctic Pacific that stimulated enhanced phytoplankton pro-ductivity (Hamme et al., 2010). To examine the potential impact of aerosol deposition events at OSP, we eval-uated the time series of aerosol optical thickness data, derived from MODIS Aqua satellite products, prior toour occupation in June 2016 (data provided by the NASA Goddard Earth Sciences Data and InformationServices Center Giovanni database and retrieved from https://giovanni.gsfc.nasa.gov/giovanni/; Acker &Leptoukh, 2007). This analysis suggested increased aerosol concentrations near OSP during the precedingmonth (Figure S6h), providing a potential explanation for the elevated chl a and NCP. Other regions of highoff-shelf productivity (e.g., ~200–500 km off-shelf in February and ~250–600 km in June; Figures 5 and 6) maybe attributed to periodic nutrient inputs via coastal eddies (Johnson et al., 2005), periods of mixed layer dee-pening and entrainment of high subsurface nutrient concentrations, elevated SST, or enhanced stratification.In support of this, satellite-based altimetry showed sea surface height anomalies (data provided by the NASAPhysical Oceanography Distributed Active Archive Center and retrieved from https://doi.org/10.5067/SLREF-CDRV1; Zlotnicki et al., 2016) indicative of anticyclonic eddies near the Line P transect in June and August(Figures S6e and S6f).5. ConclusionsCurrent estimates of ship-based NCP, particularly in coastal waters, will often be biased if investigatorsneglect vertical mixing fluxes when evaluating surface O2 budgets. We present refined NCP estimates fromfive coastal and oceanic cruises in the Subarctic Northeast Pacific, derived using the first field applicationof a N2O-based approach to correct for vertical mixing. Our NCP values exceed most past measurementsin the study region (due to a combination of natural variability and our consideration of a mixing correction)and show strong correlations with chl a concentrations, SST, and mld, providing insight into the environmen-tal controls on biological productivity. We also provide a comprehensive assessment of the N2O approach infield settings, making recommendations for future applications in our study area and other oceanographicregions. We find that potential surface N2O production (from nitrification) likely introduces only a small over-bias in NCP derived by this approach, although simultaneous measurements of surface N2O production ratesand nitrification yields are required to fully assess the impact of this process. We also discuss scenarios wherethe N2O approach should be applied cautiously, or omitted altogether. In particular, N2O-based correctionsare problematic in locations with subsurface minima in [N2O]B and corresponding [O2]B maxima.Quantifying the magnitude of vertical fluxes from ship-based observations remains a challenge in theseregions, since independent and site-specific estimates of vertical mixing rates are necessary. Nonetheless,we conclude that the N2O correction approach is robust across a range of oceanographic settings, especiallyin coastal regions where mixing fluxes constitute a substantial part of mixed layer O2 budgets and wherestudies of NCP are generally sparse. These coastal waters contribute disproportionately to ocean biogeo-chemical cycles and are often characterized by significant fine-scale variability, highlighting the utility ofunderway ship-board ΔO2/Ar measurements to estimate surface water NCP.In light of recent anomalous conditions in the Subarctic Pacific such as large volcanic eruptions (Hammeet al., 2010), the warm-water anomaly (the “blob”; Bond et al., 2015), and observed longer-term trends insea surface warming, freshening, and deoxygenation (Crawford & Peña, 2013; Whitney & Freeland, 1999), itis important to quantify marine productivity on ecologically relevant spatial and temporal scales. Themethod employed here has the potential to provide unrivaled surface coverage of NCP estimatescorrected for physical mixing. Developing technologies for underway, high-resolution measurements ofsurface N2O concentrations and reducing the uncertainties around key assumptions of the method wouldconstitute significant progress in applying this technique. Subsequently, the accumulation of a large NCPdatabase would help to produce regionally tuned satellite algorithms providing basin-scale coverage andenabling better assessment of the future evolution of the ocean’s capacity for biomass production andcarbon sequestration.Global Biogeochemical Cycles 10.1002/2017GB005792IZETT ET AL. 21ReferencesAcker, J. G., & Leptoukh, G. (2007). Online analysis enhances use of NASA Earth science data. Eos, Transactions American Geophysical Union,88(2), 14–17.Allen, S. E., & Durrieu de Madron, X. (2009). A review of the role of submarine canyons in deep-ocean exchange with the shelf. Ocean Science,5(4), 607–620. https://doi.org/10.5194/os-5-607-2009Arévalo-Martínez, D. L., Beyer, M., Krumbholz, M., Piller, I., Kock, A., Steinhoff, T., et al. (2013). A new method for continuous measurements ofoceanic and atmospheric N2O, CO and CO2: Performance of off-axis integrated cavity output spectroscopy (OA-ICOS) coupled to non-dispersive infrared detection (NDIR). Ocean Science, 9(6), 1071–1087. https://doi.org/10.5194/os-9-1071-2013Atlas, R., Hoffman, R. N., Ardizzone, J., Leidner, S. M., Jusem, J. C., Smith, D. K., & Gombos, D. (2011). A cross-calibrated, multiplatform oceansurface wind velocity product for meteorological and oceanographic applications. Bulletin of the American Meteorological Society, 92(2),157–174. https://doi.org/10.1175/2010BAMS2946.1Bakun, A. (1973). Coastal upwelling indices, west coast of North America, 1946-71. U.S. Dept. of Commerce, NOAA Tech. Rep., NMFS SSRF-671, 1–103.Bange, H., Rapsomanikis, S., & Andreae, M. (1996). Nitrous oxide in coastal waters. Global Biogeochemical Cycles, 10(1), 197–207. https://doi.org/10.1029/95GB03834Barwell-Clarke, J., & Whitney, F. (1996). Institute of Ocean Sciences nutrient methods and analysis. Canadian Technical Report of Hydrographyand Ocean Sciences, 182, 1–42.Bond, N. A., Cronin, M. F., Freeland, H., & Mantua, N. (2015). Causes and impacts of the 2014 warm anomaly in the NE Pacific. GeophysicalResearch Letters, 42, 3414–3420. https://doi.org/10.1002/2015GL063306Bonjean, F., & Lagerloef, G. S. E. (2002). Diagnostic model and analysis of the surface currents in the tropical Pacific Ocean. Journal of PhysicalOceanography, 32(10), 2938–2954. https://doi.org/10.1175/1520-0485(2002)032%3C2938:DMAAOT%3E2.0.CO;2Boyd, P. W., Muggli, D. L., Varela, D. E., Goldblatt, R. H., Chretien, R., Orians, K. J., & Harrison, P. J. (1996). In vitro iron enrichment experiments inNE subarctic Pacific. Marine Ecology Progress Series, 136, 179–193. https://doi.org/10.3354/meps136179Brown, Z. W., Casciotti, K. L., Pickart, R. S., Swift, J. H., & Arrigo, K. R. (2015). Aspects of the marine nitrogen cycle of the Chukchi Sea shelf andCanada Basin. Deep-Sea Research Part II: Topical Studies in Oceanography, 118, 73–87. https://doi.org/10.1016/j.dsr2.2015.02.009Bushinsky, S. M., & Emerson, S. (2015). Marine biological production from in situ oxygen measurements on a profiling float in the subarcticPacific Ocean. Global Biogeochemical Cycles, 29, 2050–2060. https://doi.org/10.1002/2015GB005251Bylhouwer, B., Ianson, D., & Kohfeld, K. (2013). Changes in the onset and intensity of wind-driven upwelling and downwelling along the NorthAmerican Pacific coast. Journal of Geophysical Research: Oceans, 118, 2565–2580. https://doi.org/10.1002/jgrc.20194Capelle, D. W., Dacey, J. W., & Tortell, P. D. (2015). An automated, high through-put method for accurate and precise measurements of dis-solved nitrous-oxide and methane concentrations in natural waters. Limnology and Oceanography: Methods, 13(7), 345–355. https://doi.org/10.1002/lom3.10029Capelle, D. W., & Tortell, P. D. (2016). Factors controlling methane and nitrous-oxide variability in the southern British Columbia coastalupwelling system. Marine Chemistry, 179(2016), 56–67. https://doi.org/10.1016/j.marchem.2016.01.011Carpenter, J. H. (1965). The Chesapeake Bay institute technique for the Winkler dissolved oxygen method. Limnology and Oceanography,10(1), 141–143. https://doi.org/10.4319/lo.1965.10.1.0141Cassar, N., Barnett, B. A., Bender, M. L., Kaiser, J., Hamme, R. C., & Tilbrook, B. (2009). Continuous high-frequency dissolved O2/Ar measure-ments by equilibrator inlet mass spectrometry. Analytical Chemistry, 81(5), 1855–1864. https://doi.org/10.1021/ac802300uCassar, N., Nevison, C. D., & Manizza, M. (2014). Correcting oceanic O2/Ar-net community production estimates for vertical mixing using N2Oobservations. Geophysical Research Letters, 41, 8961–8970. https://doi.org/10.1002/2014GL062040Castro-Morales, K., Cassar, N., Shoosmith, D. R., & Kaiser, J. (2013). Biological production in the Bellingshausen Sea from oxygen-to-argonratios and oxygen triple isotopes. Biogeosciences, 10(4), 2273–2291. https://doi.org/10.5194/bg-10-2273-2013Codispoti, L. A., Yoshinari, T., & Devol, A. H. (2005). Suboxic respiration in the oceanic water column. In Respiration in aquatic ecosystems(pp. 225–247). Oxford, UK: Oxford University Press. https://doi.org/10.1093/acprof:oso/9780198527084.003.0012Cohen, Y., & Gordon, L. I. (1979). Nitrous oxide production in the ocean. Journal of Geophysical Research, 84(C1), 347–353. https://doi.org/10.1029/JC084iC01p00347Craig, H., & Hayward, T. (1987). Oxygen supersaturation in the ocean: Biological versus physical contributions. Science, 235(4785), 199–202.https://doi.org/10.1126/science.235.4785.199Crawford, W. R., & Peña, M. A. (2013). Declining oxygen on the British Columbia continental shelf. Atmosphere-Ocean, 51(1), 88–103. https://doi.org/10.1080/07055900.2012.753028Crawford, W. R., & Thomson, R. E. (1991). Physical oceanography of the western Canadian continental shelf. Continental Shelf Research,11(8-10), 669–683. https://doi.org/10.1016/0278-4343(91)90073-FCronin, M. F., Pellan, N. A., Emerson, S. R., & Crawford, W. R. (2015). Estimating diffusivity from the mixed layer heat and salt balances in theNorth Pacific. Journal of Geophysical Research: Oceans, 120, 7346–7362. https://doi.org/10.1002/2015JC011010Dore, J. E., & Karl, D. M. (1996). Nitrification in the euphotic zone as a source for nitrite, nitrate, and nitrous oxide at station ALOHA. Limnologyand Oceanography, 41(8), 1619–1628. https://doi.org/10.4319/lo.1996.41.8.1619Emerson, S., Ito, T., & Hamme, R. C. (2012). Argon supersaturation indicates low decadal-scale vertical mixing in the ocean thermocline.Geophysical Research Letters, 39, L18610. https://doi.org/10.1029/2012GL053054Emerson, S., Quay, P., Karl, D., Winn, C., Tupas, L., & Landry, M. (1997). Experimental determination of the organic carbon flux from open-ocean surface waters. Nature, 389(6654), 951–954. https://doi.org/10.1038/40111Eppley, R. W. (1972). Temperature and phytoplankton growth in the sea. Fishery Bulletin, 70, 1063–1085.Earth Space Research (2009). OSCAR third degree resolution ocean surface currents. Ver. 1. PO.DAAC, CA, USA. https://doi.org/10.5067/OSCAR-03D01Farías, L., Castro-González, M., Cornejo, M., Charpentier, J., Faúndez, J., Boontanon, N., & Yoshida, N. (2009). Denitrification and nitrous oxidecycling within the upper oxycline of the eastern tropical South Pacific oxygen minimum zone. Limnology and Oceanography, 54(1),132–144. https://doi.org/10.4319/lo.2009.54.1.0132Fassbender, A. J., Sabine, C. L., & Cronin, M. F. (2016). Net community production and calcification from 7 years of NOAA Station Papa Morringmeasurements. Global Biogeochemical Cycles, 30, 250–267. https://doi.org/10.1002/2015GB005205Fenwick, L., Capelle, D., Damm, E., Zimmermann, S., Williams, W. J., Vagle, S., & Tortell, P. D. (2017). Methane and nitrous oxide distributionsacross the North American Arctic Ocean during summer, 2015. Journal of Geophysical Research: Oceans, 122, 390–412. https://doi.org/10.1002/2016JC012493Global Biogeochemical Cycles 10.1002/2017GB005792IZETT ET AL. 22AcknowledgmentsWe would like to thank all of the indivi-duals involved in each of our 2016cruises, including chief scientists M.Robert and D. Yelland, participatingscientists from the Institute of OceanSciences, the University of Victoria, theUniversity of British Columbia, and theUniversity of Washington, and theCaptains and crews of the CCGS J. P.Tully. We also thank W. Burt for hiscomments and input on the manuscriptand L. Fenwick for her assistance run-ning N2O samples. Funding for thisresearch was provided by the NaturalSciences and Engineering ResearchCouncil of Canada (NSERC). All datapresented in this manuscript are avail-able on Pangaea.de (doi:10.1594/PANGAEA.885867). We declare thatthere are no conflicts of interest asso-ciated with the production of this work.Frame, C. H., & Casciotti, K. L. (2010). Biogeochemical controls and isotopic signatures of nitrous oxide production by a marine ammonia-oxidizing bacterium. Biogeosciences Discussions, 7(9), 2695–2709. https://doi.org/10.5194/bg-7-2695-2010Garcia, H. E., & Gordon, L. I. (1992). Oxygen solubility in seawater: Better fitting equations. Limnology and Oceanography, 37(6), 1307–1312.https://doi.org/10.2307/2837876Garcia, H. E., & Gordon, L. I. (1993). Erratum: Oxygen solubility in seawater: Better fitting equations. Limnology and Oceanography, 38, 656.Giesbrecht, K. E., Hamme, R. C., & Emerson, S. R. (2012). Biological productivity along Line P in the subarctic northeast Pacific: In situ versusincubation-based methods. Global Biogeochemical Cycles, 26, GB3028. https://doi.org/10.1029/2012GB004349Goreau, T. J., Kaplan, W. A., Wofsy, S. C., McElroy, M. B., Valois, F. W., & Watson, S. W. (1980). Production of NO2and N2O by nitrifying bacteriaat reduced concentrations of oxygen. Applied and Environmental Microbiology, 40(3), 526–532.Grefe, I., & Kaiser, J. (2014). Equilibrator-based measurements of dissolved nitrous oxide in the surface ocean using an integrated cavityoutput laser absorption spectrometer. Ocean Science, 10(3), 501–512. https://doi.org/10.5194/os-10-501-2014Grundle, D. S., Juniper, S. K., & Giesbrecht, K. E. (2013). Euphotic zone nitrification in the NE subarctic Pacific: Implications for measurements ofnew production. Marine Chemistry, 155(2013), 113–123. https://doi.org/10.1016/j.marchem.2013.06.004Grundle, D. S., Maranger, R., & Juniper, S. K. (2012). Upper water column nitrous oxide distributions in the northeast subarctic Pacific Ocean.Atmosphere-Ocean, 50(4), 475–486. https://doi.org/10.1080/07055900.2012.727779Hall, B. D., Dutton, G. S., & Elkins, J. W. (2007). The NOAA nitrous oxide standard scale for atmospheric observations. Journal of GeophysicalResearch, 112, D09305. https://doi.org/10.1029/2006JD007954Hamme, R. C., Cassar, N., Lance, V. P., Vaillancourt, R. D., Bender, M. L., Strutton, P. G., et al. (2012). Dissolved O2/Ar and other methods revealrapid changes in productivity during a Lagrangian experiment in the Southern Ocean. Journal of Geophysical Research, 117, C00F12.https://doi.org/10.1029/2011JC007046Hamme, R. C., & Emerson, S. R. (2002). Mechanisms controlling the global oceanic distribution of the inert gases argon, nitrogen and neon.Geophysical Research Letters, 29(23), 2120. https://doi.org/10.1029/2002GL015273Hamme, R. C., & Emerson, S. R. (2006). Constraining bubble dynamics and mixing with dissolved gases: Implications for productivity mea-surements by oxygen mass balance. Journal of Marine Research, 64(1), 73–95. https://doi.org/10.1357/002224006776412322Hamme, R. C., Webley, P. W., Crawford, W. R., Whitney, F. A., Degrandpre, M. D., Emerson, S. R., et al. (2010). Volcanic ash fuels anomalousplankton bloom in subarctic northeast Pacific. Geophysical Research Letters, 37, L19604. https://doi.org/10.1029/2010GL044629Harrison, P. J., Boyd, P. W., Varela, D. E., Takeda, S., Shiomoto, A., & Odate, T. (1999). Comparison of factors controlling phytoplankton pro-ductivity in the NE and NW subarctic Pacific gyres. Progress in Oceanography, 43(2-4), 205–234. https://doi.org/10.1016/S0079-6611(99)00015-4Ho, D. T., Law, C. S., Smith, M. J., Schlosser, P., Harvey, M., & Hill, P. (2006). Measurements of air-sea gas exchange at high wind speeds in theSouthern Ocean: Implications for global parameterizations. Geophysical Research Letters, 33, L16611. https://doi.org/10.1029/2006GL026817Horrigan, S. G., Carlucci, A. F., & Williams, P. M. (1981). Light inhibition of nitrification in sea-surface films. Journal of Marine Research, 39(3),557–565.Howard, E., Emerson, S., Bushinsky, S., & Stump, C. (2010). The role of net community production in air-sea carbon fluxes at the NorthPacific subarctic-subtropical boundary region. Limnology and Oceanography, 55(6), 2585–2596. https://doi.org/10.4319/lo.2010.55.6.2585Ianson, D., & Allen, S. E. (2002). A two-dimensional nitrogen and carbon flux model in a coastal upwelling region. Global BiogeochemicalCycles, 16(1), 1011. https://doi.org/10.1029/2001GB001451Jin, X., Najjar, R. G., Louanchi, F., & Doney, S. C. (2007). A modeling study of the seasonal oxygen budget of the global ocean. Journal ofGeophysical Research, 112, C05017. https://doi.org/10.1029/2006JC003731Johnson, W. K., Miller, L. A., Sutherland, N. E., & Wong, C. S. (2005). Iron transport by mesoscale Haida eddies in the Gulf of Alaska. Deep-SeaResearch Part II: Topical Studies in Oceanography, 52(7-8), 933–953. https://doi.org/10.1016/j.dsr2.2004.08.017Jonsson, B. F., Doney, S. C., Dunne, J., & Bender, M. (2013). Evaluation of the Southern Ocean O2/Ar-based NCP estimates in a model fra-mework. Journal of Geophysical Research: Biogeosciences, 118, 385–399. https://doi.org/10.1002/jgrg.20032Kaiser, J., Reuer, M. K., Barnett, B., & Bender, M. L. (2005). Marine productivity estimates from continuous O2/Ar ratio measurements bymembrane inlet mass spectrometry. Geophysical Research Letters, 32, L19605. https://doi.org/10.1029/2005GL023459Kalnay, E., Kanamitsu, M., Kistler, R., Collins, W., Deaven, D., Gandin, L., et al. (1996). The NCEP/NCAR 40-year reanalysis project. Bulletin of theAmerican Meteorological Society, 77(3), 437–471. https://doi.org/10.1175/1520-0477(1996)077%3C0437:TNYRP%3E2.0.CO;2Keeling, R. F., & Shertz, S. R. (1992). Seasonal and interannual variations in atmospheric oxygen and implications for the global carbon cycle.Nature, 358(6389), 723–727. https://doi.org/10.1038/358723a0Kundu, P. K., Cohen, I. M., & Dowling, D. R. (2012). Fluid mechanics (5th ed.). Amsterdam, Netherlands: Elsevier.Laws, E. A. (1991). Photosynthetic quotients, new production and net community production in the open ocean. Deep Sea Research, 38(1),143–167. https://doi.org/10.1016/0198-0149(91)90059-OLi, Z., & Cassar, N. (2016). Satellite estimates of net community production based on O2/Ar observations and comparison to other estimates.Global Biogeochemical Cycles, 30, 735–752. https://doi.org/10.1002/2015GB005314Lockwood, D., Quay, P. D., Kavanaugh, M. T., Juranek, L. W., & Feely, R. A. (2012). High-resolution estimates of net community production andair-sea CO2 flux in the northeast Pacific. Global Biogeochemical Cycles, 26, GB4010. https://doi.org/10.1029/2012GB004380Luz, B., & Barkan, E. (2009). Net and gross oxygen production from O2/Ar,17O/16O and18O/16O ratios. Aquatic Microbial Ecology, 56, 133–145.https://doi.org/10.3354/ame01296Maldonado, M. T., Boyd, P. W., Harrison, P. J., & Price, N. M. (1999). Co-limitation of phytoplankton growth by light and Fe during winter in theNE subarctic Pacific Ocean. Deep-Sea Research Part II: Topical Studies in Oceanography, 46(11-12), 2475–2485. https://doi.org/10.1016/S0967-0645(99)00072-7Manning, C. C., Stanley, R. H. R., Nicholson, D. P., Smith, J. M., Pennington, J. T., Fewings, M. R., et al. (2017). Impact of recently upwelled wateron productivity investigated using in situ and incubation-based methods in Monterey Bay. Journal of Geophysical Research: Oceans, 122,1901–1926. https://doi.org/10.1002/2016JC012306NASA Goddard Space Flight Center, Ocean Ecology Laboratory, Ocean Biology Processing Group (2014a). Moderate-Resolution ImagingSpectroradiometer (MODIS) Aqua daytime 11 μ SST data. NASA, OB.DAAC. https://doi.org/10.5067/MODSA-1D9D4NASA Goddard Space Flight Center, Ocean Ecology Laboratory, Ocean Biology Processing Group (2014b). Moderate-Resolution ImagingSpectroradiometer (MODIS) Aqua nighttime 11 μ SST data. NSAS OB.DAAC. https://doi.org/10.5067/MODSA-1D9N4Nevison, C., Butler, J. H., & Elkins, J. W. (2003). Global distribution of N2O and the ΔN2O-AOU yield in the subsurface ocean. GlobalBiogeochemical Cycles, 17(4), 1119. https://doi.org/10.1029/2003GB002068Global Biogeochemical Cycles 10.1002/2017GB005792IZETT ET AL. 23Oudot, C., Andrie, C., & Montel, Y. (1990). Nitrous oxide production in the tropical Atlantic Ocean. Deep Sea Research, 37(2), 183–202. https://doi.org/10.1016/0198-0149(90)90123-DPalevsky, H. I., Quay, P. D., Lockwood, D. E., & Nicholson, D. P. (2016). The annual cycle of gross primary production, net community pro-duction, and export efficiency across the North Pacific Ocean. Global Biogeochemical Cycles, 30, 361–380. https://doi.org/10.1002/2015GB005318Peña, M. A., & Varela, D. E. (2007). Seasonal and interannual variability in phytoplankton and nutrient dynamics along Line P in the NEsubarctic Pacific. Progress in Oceanography, 75(2), 200–222. https://doi.org/10.1016/j.pocean.2007.08.009Plant, J. N., Johnson, K. S., Sakamoto, C. M., Jannasch, H. W., Coletti, L. J., Riser, S. C., & Swift, D. D. (2016). Net community production at OceanStation Papa observed with nitrate and oxygen sensors on profiling floats. Global Biogeochemical Cycles, 30, 859–879. https://doi.org/10.1002/2015GB005349.ReceivedRees, A. P., Owens, N. J. P., & Upstill-Goddard, R. C. (1997). Nitrous oxide in the Bellingshausen Sea and Drake Passage. Journal of GeophysicalResearch, 102(C2), 3383–3391. https://doi.org/10.1029/96JC03350Reuer, M. K., Barnett, B. A., Bender, M. L., Falkowski, P. G., & Hendricks, M. B. (2007). New estimates of Southern Ocean biological productionrates from O2/Ar ratios and the triple isotope composition of O2. Deep-Sea Research Part I: Oceanographic Research Papers, 54(6), 951–974.https://doi.org/10.1016/j.dsr.2007.02.007Rousseau, S., Kunze, E., Dewey, R., Bartlett, K., & Dower, J. (2010). On turbulence production by swimming marine organisms in the openocean and coastal waters. Journal of Physical Oceanography, 40(9), 2107–2121. https://doi.org/10.1175/2010JPO4415.1Santoro, A. E., Buchwald, C., McIlvin, M. R., & Casciotti, K. L. (2011). Isotopic signature of N2O produced by marine ammonia-oxidizing archaea.Science, 333(6047), 1282–1285. https://doi.org/10.1126/science.1208239Santoro, A. E., Casciotti, K. L., & Francis, C. A. (2010). Activity, abundance and diversity of nitrifying archaea and bacteria in the centralCalifornia Current. Environmental Microbiology, 12(7), 1989–2006. https://doi.org/10.1111/j.1462-2920.2010.02205.xShulenberger, E., & Reid, J. L. (1981). The Pacific shallow oxygen maximum, deep chlorophyll maximum, and primary productivity, recon-sidered. Deep Sea Research, 28(9), 901–919. https://doi.org/10.1016/0198-0149(81)90009-1Smith, J. M., Casciotti, K. L., Chavez, F. P., & Francis, C. A. (2014). Differential contributions of archaeal ammonia oxidizer ecotypes to nitrifi-cation in coastal surface waters. The ISME Journal, 8(8), 1704–1714. https://doi.org/10.1038/ismej.2014.11Stanley, R. H. R., Kirkpatrick, J. B., Cassar, N., Barnett, B. A., & Bender, M. L. (2010). Net community production and gross primary productionrates in the western equatorial Pacific. Global Biogeochemical Cycles, 24, GB4001. https://doi.org/10.1029/2009GB003651Tabata, S. (1975). The general circulation of the Pacific Ocean and a brief account of the oceanographic structure of the North Pacific Ocean.Atmosphere, 13, 1034–1054. https://doi.org/10.1080/00046973.1976.9648398Teeter, L. (2014). Modelling oxygen and argon to improve estimation of net community productivity in a coastal upwelling zone using ΔO2/Ar. (MSc. Thesis). University of Victoria, Victoria, BC, Canada. Retrieved from https://dspace.library.uvic.ca//handle/1828/5819Tortell, P. D. (2005). Dissolved gas measurements in oceanic waters made by membrane inlet mass spectrometry. Limnology andOceanography: Methods, 3(1), 24–37. https://doi.org/10.4319/lom.2005.3.24Tortell, P. D., Bittig, H. C., Körtzinger, A., Jones, E. M., & Hoppema, M. (2015). Biological and physical controls on N2, O2, and CO2 dis-tributions in contrasting Southern Ocean surface waters. Global Biogeochemical Cycles, 29, 994–1013. https://doi.org/10.1002/2014GB004975Tortell, P. D., Guéguen, C., Long, M. C., Payne, C. D., Lee, P., & DiTullio, G. R. (2011). Spatial variability and temporal dynamics of surface waterpCO2, ΔO2/Ar and dimethylsulfide in the Ross Sea, Antarctica. Deep-Sea Research Part I: Oceanographic Research Papers, 58(3), 241–259.https://doi.org/10.1016/j.dsr.2010.12.006Tortell, P. D., Merzouk, A., Ianson, D., Pawlowicz, R., & Yelland, D. R. (2012). Influence of regional climate forcing on surface water pCO2, O2/Arand dimethylsulfide (DMS) along the southern British Columbia coast. Continental Shelf Research, 47(2012), 119–132. https://doi.org/10.1016/j.csr.2012.07.007Ulfsbo, A., Cassar, N., Korhonen, M., van Heuven, S., Hoppema, M., Kattner, G., & Anderson, L. G. (2014). Late summer net community pro-duction in the central Arctic Ocean using multiple approaches. Global Biogeochemical Cycles, 28, 1129–1148. https://doi.org/10.1002/2014GB004833Varela, D. E., & Harrison, P. J. (1999). Seasonal variability in nitrogenous nutrition of phytoplankton assemblages in the northeastern subarcticPacific Ocean. Deep-Sea Research Part II: Topical Studies in Oceanography, 46(11-12), 2505–2538. https://doi.org/10.1016/S0967-0645(99)00074-0Volk, T., & Hoffert, M. I. (1985). Ocean carbon pumps: Analysis of relative strengths and efficiencies in ocean-driven atmospheric CO2 changes.In E. T. Sundquist & W. S. Broecker (Eds.), The carbon cycle and atmospheric CO2: Natural variations Archean to present (pp. 99–110).Washington, DC: American Geophysical Union. https://doi.org/10.1029/GM032p0099Wanninkhof, R. (2014). Relationship between wind speed and gas exchange over the ocean revisited. Limnology and Oceanography:Methods, 97(C5), 7373–7362. https://doi.org/10.1029/92JC00188Ward, B. B. (2000). Nitrification and the marine nitrogen cycle. In D. Kirchman (Ed.), Microbial ecology (pp. 427–454). New York: Wiley-Liss.Ward, B. B. (2005). Temporal variability in nitrification rates and related biogeochemical factors in Monterey Bay, California, USA. MarineEcology Progress Series, 292, 97–109. https://doi.org/10.3354/meps292097Ware, D. M., & Thomson, R. E. (2005). Bottom-up ecosystem trophic dynamics determine fish production in the northeast Pacific. Science,308(5726), 1280–1284. https://doi.org/10.1126/science.1109049Weeding, B., & Trull, T. W. (2014). Hourly oxygen and total gas tension measurements at the Southern Ocean time series site reveal winterventilation and spring net community production. Journal of Geophysical Research: Oceans, 119, 348–358. https://doi.org/10.1002/2013JC009302Weiss, R. F., & Price, B. A. (1980). Nitrous oxide solubility in water and seawater.Marine Chemistry, 8(4), 347–359. https://doi.org/10.1016/0304-4203(80)90024-9Welschmeyer, N. A. (1994). Fluorometric analysis of chlorophyll a in the presence of chlorophyll b and pheopigments. Limnology andOceanography, 39(8), 1985–1992. https://doi.org/10.4319/lo.1994.39.8.1985Whalen, C. B., Talley, L. D., & MacKinnon, J. A. (2012). Spatial and temporal variability of global ocean mixing inferred from Argo profiles.Geophysical Research Letters, 39, L18612. https://doi.org/10.1029/2012GL053196Whitney, F. A., Crawford, W. R., & Harrison, P. J. (2005). Physical processes that enhance nutrient transport and primary productivity in thecoastal and open ocean of the subarctic NE Pacific. Deep-Sea Research Part II: Topical Studies in Oceanography, 52(5-6), 681–706. https://doi.org/10.1016/j.dsr2.2004.12.023Whitney, F. A., & Freeland, H. J. (1999). Variability in upper-ocean water properties in the NE Pacific Ocean. Deep-Sea Research Part II: TopicalStudies in Oceanography, 46(11-12), 2351–2370. https://doi.org/10.1016/S0967-0645(99)00067-3Global Biogeochemical Cycles 10.1002/2017GB005792IZETT ET AL. 24Whitney, F. A., Wong, C. S., & Boyd, P. W. (1998). Interannual variability in nitrate supply to surface waters of the northeast Pacific Ocean.Marine Ecology Progress Series, 170, 15–23. https://doi.org/10.3354/meps170015Yang, B., Emerson, S. R., & Bushinsky, S. M. (2017). Annual net community production in the subtropical Pacific Ocean from in-situ oxygenmeasurements on profiling floats. Global Biogeochemical Cycles, 31, 728–744. https://doi.org/10.1002/2016GB005545Yoshida, N., Morimoto, H., Hirano, M., Koike, I., Matsuo, S., Wada, E., et al. (1989). Nitrification rates and15N abundances of N2O and NO3inthe western North Pacific. Nature, 342(6252), 895–897. https://doi.org/10.1038/342895a0Yoshinari, T. (1976). Nitrous oxide in the sea. Marine Chemistry, 4(2), 189–202. https://doi.org/10.1016/0304-4203(76)90007-4Zamora, L. M., & Oschlies, A. (2014). Surface nitrification: A major uncertainty in marine N2O emissions. Geophysical Research Letters, 41,4247–4253. https://doi.org/10.1002/2014GL060556Zlotnicki, V., Qu, Z., & Willis, J. (2016). JPL MEaSUREs gridded sea surface height anomalies version 1609. Ver. 1609. PO.DAAC, CA, USA. https://doi.org/10.5067/SLREF-CDRV1Global Biogeochemical Cycles 10.1002/2017GB005792IZETT ET AL. 25


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items