Shallow particulate organic carbon regeneration in the South Pacific Ocean
Edited by David M. Karl, University of Hawaii, Honolulu, HI, and approved April 8, 2019 (received for review January 31, 2019)
Significance
Plankton in the sunlit surface ocean photosynthesize, fixing dissolved CO2 into particulate organic carbon (POC). This POC sinks and is respired, releasing CO2 into subsurface waters that are sequestered from the atmosphere. The depth scale over which this regeneration happens strongly affects atmospheric CO2, but estimates to date have been sparse and challenging to interpret. We use a new geochemical method to determine POC regeneration depth scales at unprecedented resolution in the South Pacific Ocean, finding shallow regeneration in both oxygen-deficient zone and oligotrophic gyre settings. Our results imply decreased future ocean carbon storage due to gyre expansion and two opposing feedbacks to expanding oxygen-deficient zones, the net effects of which on ocean carbon storage require future research.
Abstract
Particulate organic carbon (POC) produced in the surface ocean sinks through the water column and is respired at depth, acting as a primary vector sequestering carbon in the abyssal ocean. Atmospheric carbon dioxide levels are sensitive to the length (depth) scale over which respiration converts POC back to inorganic carbon, because shallower waters exchange with the atmosphere more rapidly than deeper ones. However, estimates of this carbon regeneration length scale and its spatiotemporal variability are limited, hindering the ability to characterize its sensitivity to environmental conditions. Here, we present a zonal section of POC fluxes at high vertical and spatial resolution from the GEOTRACES GP16 transect in the eastern tropical South Pacific, based on normalization to the radiogenic thorium isotope 230Th. We find shallower carbon regeneration length scales than previous estimates for the oligotrophic South Pacific gyre, indicating less efficient carbon transfer to the deep ocean. Carbon regeneration is strongly inhibited within suboxic waters near the Peru coast. Canonical Martin curve power laws inadequately capture POC flux profiles at suboxic stations. We instead fit these profiles using an exponential function with flux preserved at depth, finding shallow regeneration but high POC sequestration below 1,000 m. Both regeneration length scales and POC flux at depth closely track the depths at which oxygen concentrations approach zero. Our findings imply that climate warming will result in reduced ocean carbon storage due to expanding oligotrophic gyres, but opposing effects on ocean carbon storage from expanding suboxic waters will require modeling and future work to disentangle.
Sign up for PNAS alerts.
Get alerts for new articles, or get an alert when an article is cited.
The oceanic biological pump encompasses a series of processes by which phytoplankton at the sea surface photosynthetically fix carbon dioxide (CO2) to form particulate organic carbon (POC), a portion of which is exported from the upper ocean and sinks to depth, where it is regenerated by microbial respiration (1, 2). The first two components of the biological pump, primary production and export of POC from the upper ocean, have been sufficiently characterized to enable their parametrization in terms of variables that can be measured by satellites, allowing for comprehensive estimates of their global rates and spatiotemporal variability (3–6). However, the fate of exported POC upon sinking into the ocean interior has proved to be an elusive oceanographic target. Because the time scale that waters are sequestered from the atmosphere increases with depth, the length scale over which POC regeneration occurs exerts a strong control on oceanic carbon storage and atmospheric CO2 levels (7). Consequently, assessing how environmental conditions influence POC regeneration length scales provides critical insights that can be incorporated into ocean carbon cycle models to improve projections of future oceanic CO2 uptake, including the response to global warming.
Historical estimates of carbon regeneration in the ocean interior have come from POC flux profiles generated either by compilations of sediment traps (8); by individual free-floating sediment trap profiles, typically with three to six depths in the upper 500 m (9, 10); or by combining 234Th-based euphotic zone POC fluxes with those from bottom-moored sediment traps below 1,500 m (11, 12). POC regeneration length scales are then determined by fitting either power laws (8) or exponential functions (13) to the vertical profiles of POC flux. However, these methods are respectively limited by their spatial resolution, vertical resolution, and integration across different temporal and spatial domains. The methods also provide conflicting results on the spatial patterns of regeneration depths, precluding the development of a comprehensive mechanistic understanding of the processes that control POC regeneration (9, 11).
We determine POC regeneration length scales in the eastern tropical South Pacific by adapting the paleoceanographic 230Th-normalization method (14) to the water column. Our study is the first application of this approach to generate internally consistent, high-resolution POC flux profiles that resolve differences in POC flux characteristics across biogeochemical gradients on annual to multiannual time scales. By analyzing particulate 230Th (230Thp) and POC collected by in situ filtration, we calculate POC fluxes, integrated across ∼1- to 3-y time scales, at each measurement depth (15, 16) (see Materials and Methods). A recent intercomparison of sediment trap and radiochemical methods at the Bermuda Atlantic Time-Series Station found that 230Thp-normalized POC fluxes agreed (within 2-σ uncertainty) with other radiochemical methods for estimating POC flux in the upper water column (17). In further support of this approach, we find that 230Thp-derived POC fluxes on the GEOTRACES GP16 transect are within uncertainty of nearby annually averaged sediment trap POC fluxes (SI Appendix, Fig. S1).
Samples were collected on the GP16 transect (Research Vessel Thomas G. Thompson, cruise TN303) spanning from Peru to Tahiti (Fig. 1). The GP16 section traversed a strong zonal gradient in upper water column conditions, particularly in productivity and subsurface O2 (Fig. 1). The Peru oxygen-deficient zone (ODZ) in the eastern portion of the section hosts nanomolar to subnanomolar O2 levels, making it functionally anoxic (18). Oxygen concentration minima from GP16 were below the detection limit of 1 μmol/kg at stations 1 to 13 (Fig. 1B). Pigment and fluorescence data indicate that there is a transition in microbial community structure moving offshore within the ODZ, from autotrophic at station 9 to heterotrophic at station 11 (19). Our 230Thp-normalized POC flux profiles have sufficient vertical resolution to provide statistically significant constraints on the spatial variability and mechanisms controlling POC regeneration length scales and carbon transfer to the deep ocean across the sharp biogeochemical gradients spanning from the Peru ODZ to the highly oligotrophic South Pacific subtropical gyre (SPSG).
Fig. 1.
Results and Discussion
GP16 230Thp-normalized POC flux profiles have highest values in the subsurface near the deep chlorophyll maximum and base of the mixed layer (SI Appendix, Fig. S2), and decrease subsequently with depth. Maximum POC fluxes of 5 mmol⋅m−2⋅d−1 are found nearest to the continental shelf in the Peru upwelling region, decreasing to 2 mmol⋅m−2⋅d−1 in the oligotrophic South Pacific gyre (Fig. 2). The greatest flux attenuation occurs in the upper 300 m of the water column, indicative of shallow POC regeneration. At stations 1 to 9, corresponding to the Peru ODZ, POC flux decreases rapidly through the upper oxycline, stays constant through the depths of lowest oxygen, then decreases again through the lower oxycline (SI Appendix, Fig. S3). The lack of POC flux attenuation within the ODZ must reflect negligible regeneration of vertically sourced POC supply from above (see SI Appendix, Supplementary Information Text).
Fig. 2.
Regeneration length scales are traditionally expressed using a power law relationship (8). We fit power laws of form to POC fluxes in the upper 1,000 m at each station to predict the POC flux (Fz) at depth z, relative to a reference depth z0, with the exponent b describing the rate of flux attenuation with depth. Stations 15 to 36, west of the Peru upwelling region, have average b values of 1.29 ± 0.12 (Fig. 3), much higher (i.e., shallower, faster regeneration) than previous estimates for the SPSG derived from the combination of bottom-moored sediment traps and 234Th (b = 0.52) (11, 12) and a follow-up approach that included constraints from particle imaging (b = 0.84) (12). Our results are much more consistent with neutrally buoyant sediment trap deployments in the North Pacific subtropical gyre (b = 1.33 ± 0.15), which had greater resolution through the upper water column depths at which the bulk of POC regeneration occurs (10). Transfer efficiencies derived from b values fitted to bottom-moored sediment trap observations may not be an ideal benchmark for evaluating biogeochemical model representations of POC flux and regeneration in the mesopelagic, as previously suggested (9, 20).
Fig. 3.
POC fluxes are constant with depth within the suboxic waters of the Peru ODZ (Fig. 4A). Correlations of POC flux with depth from 60 to 600 m show that the stations within the ODZ have no statistically significant decrease in POC flux with depth (Fig. 4A), while stations with no suboxic waters have highly statistically significant (P < 10−10) decreasing POC flux in the same depth range (Fig. 4B). Stations 1 and 7 in the Peru ODZ have lower b values of 0.74 ± 0.15 and 0.66 ± 0.18, respectively, compared with a range of b values from 1.11 to 1.52 at oxic stations 15 to 36 (Fig. 3). Previous studies have also found low b values for POC flux profiles from sediment trap deployments in ODZs, attributed to greater POC preservation under low oxygen conditions (21–23). However, the goodness of fit for power laws at stations 1 to 7, where the POC flux at depth is greatest and the top of the ODZ is shallowest, was much lower than at the offshore stations (SI Appendix, Fig. S4). The residuals of the power-law fits to stations 1 to 7 are also correlated with depth (rho = −0.66, P = 0.002), indicating that a power law fails to adequately capture the functional form of POC flux profiles at ODZ stations.
Fig. 4.
We instead fit POC flux profiles using an exponential function (13, 24), , including an asymptotic flux F∞ preserved as depth approaches infinity to quantify the effect of the ODZ on POC regeneration length scale (L) and transfer to the deep ocean. Unlike the power law, the residuals of exponential fits at ODZ stations are not significantly correlated with depth (rho = −0.18, P = 0.40). We used the exponential fits to generate bootstrapped probability distribution functions for L and F∞ for suboxic and oxic station groupings. The distributions are nearly disjoint, with suboxic stations having both shallower regeneration (Fig. 4C) and nearly 4 times more carbon flux preserved into the deep ocean (Fig. 4D) than oxic stations. This is not simply a consequence of larger export fluxes at suboxic stations. The transfer efficiency, computed as the best-fit F∞/Fmax, where Fmax is the maximum POC flux at each station, is 2 to 5 times higher at suboxic stations 5, 1, and 7 than at oxic stations (SI Appendix, Fig. S5). Thus, the flux profiles at suboxic stations are qualitatively and significantly different from those at oxic stations. We argue that the low b values previously inferred in ODZs overlook the importance of shallow POC regeneration in the upper oxycline. In addition to its implications for ocean carbon storage, accurately representing the vertical pattern of POC flux and regeneration in ODZs is critical for determining the depth distribution and magnitude of nitrogen loss processes (anammox and denitrification) in ODZs, which are directly linked to the supply, regeneration, and stoichiometry of organic matter flux (25).
Our results have important implications for feedbacks in the global carbon cycle under future climate change. The oligotrophic subtropical gyres are projected to expand due to increased vertical stratification (26, 27). The efficiency of POC transport to the deep ocean in the subtropics has been debated, with bottom-moored sediment trap observations suggesting efficient subtropical C storage (11), but inverse modeling (20) and neutrally buoyant sediment trap results (9, 10) suggesting the opposite. Shallow subtropical POC regeneration inferred from 230Thp normalization in both the North Atlantic gyre (17) and the South Pacific (this study) are consistent with inefficient carbon storage in the oligotrophic ocean. Thus, gyre expansion from CO2 warming is predicted to drive a positive feedback involving shallower carbon regeneration and less efficient carbon sequestration in the deep ocean.
ODZs are expected to expand in area and shoal under climate warming (28, 29); however, the relative importance of increased respiration and decreased ventilation is unknown (30, 31). Both L and F∞ are well correlated with the depth of the upper oxycline (19) (Fig. 5), indicating that changes in ODZ extent and POC regeneration will be intimately coupled. Our results show that regeneration dynamics in ODZ regions have two potentially very large, offsetting effects on ocean carbon storage. Shallower regeneration length scales will return respired CO2 to the atmosphere more quickly, but greater POC preservation to depths below 1,000 m will result in greater abyssal carbon storage. The expansion and shoaling of ODZs, therefore, will not necessarily result in enhanced overall ocean carbon storage, as previously proposed (22). The feedbacks between ODZ expansion and ocean carbon storage will require the implementation of more flexible and spatially variable regeneration length scales in global carbon cycle models and should be a high priority target for future study.
Fig. 5.
Materials and Methods
Particulate Sample Collection.
Particulate samples on the GP16 section were collected via in situ filtration using McLane pumps (WTS-LV) with two flow paths. Each flow path was equipped with a 142-mm-diameter filter holder containing baffles to ensure homogenous particle distributions on the filters (32). The holders both had a 51-μm Sefar Polyester mesh prefilter, followed by either paired acid-leached, precombusted quartz-fiber Whatman QM-A filters with a 1-μm pore size, or acid-leached paired Pall Supor800 0.8-μm polyethersulfone filters (33).
Blank filters were simultaneously deployed with the pumps on each cast, either on specially adapted filter holders disconnected from pumped water flow or in polypropylene containers zip-tied to the frame of a pump. The blank filters were in contact with ambient seawater at pump depth for the entire cast. These dipped blanks were used for background corrections of POC and Th isotopes. Previous publications (33, 34) have provided more detailed documentation of the collection of in situ pumped particles on the TN303 cruise.
Sample Analysis.
Measurement techniques for dissolved oxygen (35), POC (33), and Th isotopes (36) on the GP16 section have been previously documented. We provide here a brief overview containing the salient details of the measurements techniques, but refer readers to the publications containing the original data for complete methods.
Dissolved oxygen was determined via modified Winkler titration, according to standard procedures established in the WOCE, CLIVAR, and GO-SHIP Repeat Hydrography programs (https://www.go-ship.org/HydroMan.html). The detection limit for discrete oxygen samples was 1 μmol/L, with ∼0.1% precision (35). The finalized oxygen dataset is archived online at the Biological and Chemical Oceanography Data Management Office (BCO-DMO, https://www.bco-dmo.org/dataset/503145), as well as the GEOTRACES Intermediate Data Product (37).
POC in the 0.8- to 51-μm small-size fraction (SSF) was measured on two 12-mm-diameter punches taken from the top QM-A filter, representing ∼20 L of pumped seawater. The filters were dried at sea, fumed with concentrated hydrochloric acid (HCl) to remove inorganic carbon, and then dried again before the punches were taken. SSF POC was measured using a FlashEA 1112 Carbon/Nitrogen Analyzer using a Dynamic Flash Combustion technique. Dipped blank QM-A filters (n = 47) were used for blank subtraction, and the SD of the dipped blank measurements was assigned as the uncertainty for SSF POC measurements. POC data are available online at BCO-DMO (https://www.bco-dmo.org/dataset/668083) and the GEOTRACES Intermediate Data Product (37).
Particulate 230Th (230Thp) was measured in two laboratories: Lamont-Doherty Earth Observatory (LDEO), and University of Minnesota (UMN). Intercalibration showed no detectable differences between the methods of the two laboratories. At LDEO, one-fourth–filter aliquots were placed in 60-mL Savillex jars, a 229Th-233Pa spike and 25 mg of purified iron carrier were added, and the filters sat overnight in concentrated HNO3 at room temperature. The filters were then completely digested in concentrated perchloric acid (HClO4) to dissolve the polyethersulfone material. Particles were subsequently digested in concentrated HNO3 and HF, followed by iron coprecipitation. Thorium fractions were isolated using anion exchange chromatography (Bio-Rad AG1-X8, 100 to 200 μm). Measurements of 230Th and 232Th were made on a Thermo Element XR inductively coupled plasma mass spectrometer (ICP-MS) instrument, using an Aridus desolvating nebulizer for sample introduction to improve sensitivity (38). Full details of the LDEO method have been published previously (36, 39).
At UMN, one-eighth–filter aliquots were folded into 30-mL Teflon beakers, a 229Th-233Pa spike was added, and filters were submerged in 7N HNO3 and 10 drops of concentrated HF. The beakers were capped and heated under pressure for 10 h at 200 °F to leach/digest the samples. After heating, the leach solution was quantitatively transferred to a separate 30-mL Teflon beaker, and five drops of concentrated HClO4 were added. The leach solution was dried down and taken up in 2N HCl, followed by iron hydroxide coprecipitation. The precipitate was dissolved, dried down, and taken up again in 7N HNO3, which was then loaded onto Bio-Rad AG1-X8 100- to 200-μm mesh resin for separation of Th fractions via anion exchange chromatography. Thorium isotope measurements were made on a Thermo Neptune multicollector ICP-MS instrument (40, 41). In both laboratories, measured 230Th and 232Th were blank corrected using average dipped blank values. Errors in measured 230Th include uncertainties from ICP-MS counting statistics, spike concentrations, and blank corrections. Particulate 230Th and 232Th data are archived at BCO-DMO (https://www.bco-dmo.org/dataset/676231) and in the GEOTRACES Intermediate Data Product (37). More details on the measurements techniques in this study can be found in SI Appendix, Supplementary Information Text.
Application of 230Th normalization to POC fluxes.
230Th normalization is a widely used method in paleoceanography for correcting sediment mass accumulation rates for syndepositional redistribution (42, 43). Most 230Th in seawater is produced in the water column by the decay of 234U. Uranium is highly soluble in seawater, stabilized as carbonate complexes (44, 45) with a residence time of hundreds of thousands of years (46). As such, uranium is conservative in seawater, with only minor (parts per thousand) spatial variations in concentration as a function of salinity, allowing for the prediction of oceanic uranium concentrations from salinity (47, 48). These uranium–salinity relationships are used to predict the activity of the major uranium isotope, 238U, which is multiplied by the seawater 234U/238U activity ratio of 1.1468 (49) to estimate 234U. Thus, the production rate of 230Th integrated to a depth horizon z can be predicted anywhere in the water column:
Unlike its parent 234U, 230Th is highly insoluble in seawater. Upon production by 234U decay, 230Th rapidly adsorbs to particles, with a scavenging residence time of 20 to 40 y (50), much shorter than both its half-life [75,584 y (51)] and the time scale of whole-ocean mixing. The removal of 230Th from a given location is potentially driven by two processes: scavenging removal by particles, and lateral redistribution by advective-diffusive fluxes. Where the latter can be either ignored or corrected, the concentrations of both dissolved and particulate 230Th are expected to increase linearly with depth in a process known as reversible scavenging (52). In this formulation, the integrated production of 230Th to a depth z is balanced in one dimension by its downward export on particles sinking through that depth.
The equation for calculating 230Thp-normalized POC fluxes is nearly identical to that used in paleoceanography to determine vertical constituent fluxes:
where the integrated production rate is in μBq⋅m−2⋅d−1, [230Th]p is the activity of particulate 230Th in μBq⋅m−3, and [POC] is the concentration of POC in mmol⋅m−3. The resulting POC fluxes we report (Dataset S1) are in units of mmol⋅m−2⋅d−1.
We calculate POC fluxes on particles in the 0.8- to 51-μm SSF. Due to low 230Th activity on particles >51 μm, larger filter aliquots were required for analysis than could be routinely measured across the entire section. The actual size of sinking particles carrying 230Th downward to balance its water column production is unknown. However, scavenging removal of 230Th is a two-step process involving adsorption of 230Th onto small particles, which subsequently undergo repeated cycles of aggregation into larger “sinking” particles and disaggregation into smaller “suspended” particles (53, 54). Thus, provided that the aggregation-sinking process is in equilibrium on the time scales of 230Th removal, the POC fluxes recorded by 230Thp normalization on 0.8- to 51-μm particles will be valid.
Statistical procedures.
Power laws of form were fit to POC flux profiles at each station (Fig. 3) using data only at or above 1,000 m. Because the depths of the mixed layer, the deep chlorophyll maximum, and the oxycline varied between stations (SI Appendix, Fig. S2), we used the depth of maximum POC flux at each station as the reference depth z0 rather than interpolating onto a common reference depth (e.g., the base of the euphotic zone or 100 m) across all stations. We show in SI Appendix, Supplementary Information Text that our findings are not sensitive to the choice of reference depth.
For Fig. 4 A and B, data from the suboxic stations at the depths where oxygen concentrations were near zero (60 to 600 m) were grouped, as were data from the oxic stations over the same depth range. Three correlation tests were performed—Pearson’s correlation, Spearman’s rank correlation, and Kendall’s rank correlation (also known as Kendall’s tau)—and associated P values were computed for both groups of data (55). By any usual significance threshold, fluxes from suboxic depths of the suboxic stations are not significantly correlated with depth, whereas fluxes from the same depths of the oxic stations significantly decrease with depth.
To quantify the differences between the flux–depth relationships in the oxic versus suboxic stations, we estimated uncertainty in the parameters of the exponential fits using a bootstrap analysis (56). For each group of stations, we generated 10,000 replicate datasets via resampling with replacement and then fit the functional form via nonlinear least-squares regression to each replicate. These 10,000 estimates for each parameter are shown in Fig. 4 C and D for F∞ and L, respectively. Based on the intersection of these parameters’ estimated probability distributions, we can state with 98% and 90% confidence, respectively, that F∞ is larger and that L is smaller for the suboxic data than for the oxic data. The median L for the suboxic data is 56 m and the median L for the oxic data is 102 m. The median F∞ for the suboxic data is 0.36 mmol⋅m−2⋅d−1 and the median F∞ for the oxic data is 0.093 mmol⋅m−2⋅d−1.
Acknowledgments
We thank the captain and crew of the Research Vessel Thomas G. Thompson during the TN303 cruise. The pump team led by Dan Ohnemus was responsible for the collection of the particulate samples used in this study. We thank Kassandra Costa for comments on an early draft of the paper, as well as two anonymous reviewers and the handling editor for constructive feedback. This work was funded by US National Science Foundation Awards OCE-1233688 (to R.F.A.), OCE-1233903 (to R.L.E.), and OCE-1518110 (to P.J.L.), and by the NSF Graduate Research Fellowship DGE-16-44869 (to F.J.P.).
Supporting Information
Appendix (PDF)
- Download
- 720.17 KB
Dataset_S01 (CSV)
- Download
- 5.44 KB
References
1
MP Hain, DM Sigman, GH Haug, The biological pump in the past. Treatise on Geochemistry, eds Holland HD, Turekian KK (Elsevier, Oxford), 2nd Ed, Vo. 8, pp 485–517. (2013).
2
CL De La Rocha, U Passow, The biological pump. Treatise on Geochemistry, eds Holland HD, Turekian KK (Elsevier, Oxford), 2nd Ed, Vol 8, pp 93–122. (2014).
3
MJ Behrenfeld, PG Falkowski, Photosynthetic rates derived from satellite-based chlorophyll concentration. Limnol Oceanogr 42, 1–20 (1997).
4
MJ Behrenfeld, et al., Biospheric primary production during an ENSO transition. Science 291, 2594–2597 (2001).
5
DA Siegel, et al., Global assessment of ocean carbon export by combining satellite observations and food‐web models. Global Biogeochem Cycles 28, 181–196 (2014).
6
SA Henson, et al., A reduced estimate of the strength of the ocean’s biological carbon pump. Geophys Res Lett 38, L04606 (2011).
7
EY Kwon, F Primeau, JL Sarmiento, The impact of remineralization depth on the air–sea carbon balance. Nat Geosci 2, 630–635 (2009).
8
JH Martin, GA Knauer, DM Karl, WW Broenkow, VERTEX: Carbon cycling in the northeast Pacific. Deep Sea Res A Oceanogr Res Pap 34, 267–285 (1987).
9
CM Marsay, et al., Attenuation of sinking particulate organic carbon flux through the mesopelagic ocean. Proc Natl Acad Sci USA 112, 1089–1094 (2015).
10
KO Buesseler, et al., Revisiting carbon flux through the ocean’s twilight zone. Science 316, 567–570 (2007).
11
SA Henson, R Sanders, E Madsen, Global patterns in efficiency of particulate organic carbon export and transfer to the deep ocean. Global Biogeochem Cycles 26, GB1028 (2012).
12
L Guidi, et al., A new look at ocean carbon remineralization for estimating deepwater sequestration. Global Biogeochem Cycles 29, 1044–1059 (2015).
13
RA Armstrong, C Lee, JI Hedges, S Honjo, SG Wakeham, A new, mechanistic model for organic carbon fluxes in the ocean based on the quantitative association of POC with ballast minerals. Deep Sea Res Part II Top Stud Oceanogr 49, 219–236 (2002).
14
R Francois, M Frank, MM Rutgers van der Loeff, MP Bacon, 230Th normalization: An essential tool for interpreting sedimentary fluxes during the late Quaternary. Paleoceanography 19, PA1018 (2004).
15
K Hirose, A new method to determine depth‐dependent carbon export fluxes using vertical 230Th profiles. Geophys Res Lett 33, L05609 (2006).
16
RF Anderson, et al., How well can we quantify dust deposition to the ocean? Philos Trans A Math Phys Eng Sci 374, 20150285 (2016).
17
CT Hayes, et al., Flux of particulate elements in the North Atlantic Ocean constrained by multiple radionuclides. Global Biogeochem Cycles 32, 1738–1758 (2018).
18
B Thamdrup, T Dalsgaard, NP Revsbech, Widespread functional anoxia in the oxygen minimum zone of the Eastern South Pacific. Deep Sea Res Part I Oceanogr Res Pap 65, 36–45 (2012).
19
DC Ohnemus, et al., Elevated trace metal content of prokaryotic communities associated with marine oxygen deficient zones. Limnol Oceanogr 62, 3–25 (2017).
20
T Weber, JA Cram, SW Leung, T DeVries, C Deutsch, Deep ocean nutrients imply large latitudinal variation in particle transfer efficiency. Proc Natl Acad Sci USA 113, 8606–8611 (2016).
21
BAS Van Mooy, RG Keil, AH Devol, Impact of suboxia on sinking particulate organic carbon: Enhanced carbon flux and preferential degradation of amino acids via denitrification. Geochim Cosmochim Acta 66, 457–465 (2002).
22
EL Cavan, M Trimmer, F Shelley, R Sanders, Remineralization of particulate organic carbon in an ocean oxygen minimum zone. Nat Commun 8, 14847 (2017).
23
RG Keil, JA Neibauer, C Biladeau, K van der Elst, AH Devol, A multiproxy approach to understanding the “enhanced” flux of organic matter through the oxygen-deficient waters of the Arabian Sea. Biogeosciences 13, 2077–2092 (2016).
24
KO Buesseler, PW Boyd, Shedding light on processes that control particle export and flux attenuation in the twilight zone of the open ocean. Limnol Oceanogr 54, 1210–1232 (2009).
25
AR Babbin, RG Keil, AH Devol, BB Ward, Organic matter stoichiometry, flux, and oxygen control nitrogen loss in the ocean. Science 344, 406–408 (2014).
26
JL Sarmiento, et al., Response of ocean ecosystems to climate warming. Global Biogeochem Cycles 18, GB3003 (2004).
27
JJ Polovina, EA Howell, M Abecassis, Ocean’s least productive waters are expanding. Geophys Res Lett 35, 270 (2008).
28
L Stramma, GC Johnson, J Sprintall, V Mohrholz, Expanding oxygen-minimum zones in the tropical oceans. Science 320, 655–658 (2008).
29
WF Gilly, JM Beman, SY Litvin, BH Robison, Oceanographic and biological effects of shoaling of the oxygen minimum zone. Annu Rev Mar Sci 5, 393–420 (2013).
30
A Oschlies, P Brandt, L Stramma, S Schmidtko, Drivers and mechanisms of ocean deoxygenation. Nat Geosci 11, 467–473 (2018).
31
D Breitburg, et al., Declining oxygen in the global ocean and coastal waters. Science 359, eaam7240 (2018).
32
JKB Bishop, PJ Lam, TJ Wood, Getting good particles: Accurate sampling of particles by large volume in-situ filtration. Prog Oceanogr 10, 681–710 (2012).
33
PJ Lam, et al., Size-fractionated distributions of suspended particle concentration and major phase composition from the U.S. GEOTRACES Eastern Pacific Zonal Transect (GP16). Mar Chem 201, 90–107 (2018).
34
JM Lee, MI Heller, PJ Lam, Size distribution of particulate trace elements in the U.S. GEOTRACES Eastern Pacific Zonal Transect (GP16). Mar Chem 201, 108–123 (2018).
35
BD Peters, et al., Water mass analysis of the 2013 US GEOTRACES eastern Pacific zonal transect (GP16). Mar Chem 201, 6–19 (2018).
36
F Pavia, et al., Intense hydrothermal scavenging of 230Th and 231Pa in the deep Southeast Pacific. Mar Chem 201, 212–228 (2018).
37
R Schlitzer, et al., The GEOTRACES Intermediate Data Product 2017. Chem Geol 493, 210–223 (2018).
38
MS Choi, et al., Rapid determination of 230Th and 231Pa in seawater by desolvated micro-nebulization inductively coupled plasma magnetic sector mass spectrometry. Mar Chem 76, 99–112 (2001).
39
RF Anderson, et al., GEOTRACES intercalibration of 230Th, 232Th, 231Pa, and prospects for 10Be. Limnol Oceanogr Methods 10, 179–213 (2012).
40
C-C Shen, et al., High-precision and high-resolution carbonate 230Th dating by MC-ICP-MS with SEM protocols. Geochim Cosmochim Acta 99, 71–86 (2012).
41
C-C Shen, et al., Uranium and thorium isotopic and concentration measurements by magnetic sector inductively coupled plasma mass spectrometry. Chem Geol 185, 165–178 (2002).
42
GM Henderson, RF Anderson, The U-series toolbox for paleoceanography. Rev Mineral Geochem 52, 493–531 (2003).
43
MP Bacon, Glacial to interglacial changes in carbonate and clay sedimentation in the Atlantic Ocean estimated from 230Th measurements. Chem Geol 46, 97–111 (1984).
44
R Djogić, L Sipos, M Branica, Characterization of uranium(VI) in seawater. Limnol Oceanogr 31, 1122–1131 (1986).
45
D Langmuir, Uranium solution-mineral equilibria at low temperatures with applications to sedimentary ore deposits. Geochim Cosmochim Acta 42, 547–569 (1978).
46
RM Dunk, RA Mills, WJ Jenkins, A reevaluation of the oceanic uranium budget for the Holocene. Chem Geol 190, 45–67 (2002).
47
JH Chen, RL Edwards, GJ Wasserburg, 238U, 234U and 232Th in seawater. Earth Planet Sci Lett 80, 241–251 (1986).
48
SA Owens, KO Buesseler, KWW Sims, Re-evaluating the 238U-salinity relationship in seawater: Implications for the 238U–234Th disequilibrium method. Mar Chem 127, 31–39 (2011).
49
MB Andersen, CH Stirling, B Zimmermann, AN Halliday, Precise determination of the open ocean 234U/238U composition. Geochem Geophys Geosyst 11, Q12003 (2010).
50
CT Hayes, et al., A new perspective on boundary scavenging in the North Pacific Ocean. Earth Planet Sci Lett 369–370, 86–97 (2013).
51
H Cheng, et al., Improvements in 230Th dating, 230Th and 234U half-life values, and U–Th isotopic measurements by multi-collector inductively coupled plasma mass spectrometry. Earth Planet Sci Lett 371–372, 82–91 (2013).
52
MP Bacon, RF Anderson, Distribution of thorium isotopes between dissolved and particulate forms in the deep sea. J Geophys Res 87, 2045–2056 (1982).
53
MP Bacon, C-A Huh, AP Fleer, WG Deuser, Seasonality in the flux of natural radionuclides and plutonium in the deep Sargasso Sea. Deep Sea Res A Oceanogr Res Pap 32, 273–286 (1985).
54
W-L Wang, RA Armstrong, JK Cochran, C Heilbrun, 230Th and 234Th as coupled tracers of particle cycling in the ocean: A maximum likelihood approach. Deep Sea Res Part I Oceanogr Res Pap 111, 61–70 (2016).
55
WH Press, SA Teukolsky, WT Vetterling, BP Flannery Numerical Recipes in C: The Art of Scientific Computing (Cambridge Univ Press, 2nd Ed, Cambridge, UK, 1992).
56
B Efron, Bootstrap methods: Another look at the jackknife. Ann Stat 7, 1–26 (1979).
Information & Authors
Information
Published in
Classifications
Copyright
© 2019. Published under the PNAS license.
Submission history
Published online: April 29, 2019
Published in issue: May 14, 2019
Keywords
Acknowledgments
We thank the captain and crew of the Research Vessel Thomas G. Thompson during the TN303 cruise. The pump team led by Dan Ohnemus was responsible for the collection of the particulate samples used in this study. We thank Kassandra Costa for comments on an early draft of the paper, as well as two anonymous reviewers and the handling editor for constructive feedback. This work was funded by US National Science Foundation Awards OCE-1233688 (to R.F.A.), OCE-1233903 (to R.L.E.), and OCE-1518110 (to P.J.L.), and by the NSF Graduate Research Fellowship DGE-16-44869 (to F.J.P.).
Notes
This article is a PNAS Direct Submission.
Authors
Competing Interests
The authors declare no conflict of interest.
Metrics & Citations
Metrics
Citation statements
Altmetrics
Citations
Cite this article
116 (20) 9753-9758,
Export the article citation data by selecting a format from the list below and clicking Export.
Cited by
Loading...
View Options
View options
PDF format
Download this article as a PDF file
DOWNLOAD PDFLogin options
Check if you have access through your login credentials or your institution to get full access on this article.
Personal login Institutional LoginRecommend to a librarian
Recommend PNAS to a LibrarianPurchase options
Purchase this article to access the full text.