Vertical transport and sinks of perfluoroalkyl substances in the global open ocean

Belén González-Gaya *abc, Paulo Casal b, Elena Jurado bd, Jordi Dachs b and Begoña Jiménez a
aDepartment of Instrumental Analysis and Environmental Chemistry, Institute of Organic Chemistry (IQOG-CSIC), Juan de la Cierva 3, 28006 Madrid, Spain. E-mail: b.gonzalez@iqog.csic.es
bDepartment of Environmental Chemistry, Institute of Environmental Assessment and Water Research, Spanish National Research Council (IDAEA-CSIC), Jordi Girona 18-26, 08034 Barcelona, Catalonia, Spain
cDepartment of Analytical Chemistry, Plentzia Marine Station of the University of the Basque Country (EHU/UPV), Areatza Pasealekua 47, 48620 Plentzia, Basque Country, Spain
dEngineering Department, La Salle Campus Barcelona, Ramon Llull University (URL), Sant Joan de la Salle, 42, 08022 Barcelona, Catalonia, Spain

Received 31st May 2019 , Accepted 27th July 2019

First published on 31st July 2019


The ubiquitous occurrence of perfluoroalkyl substances (PFAS) in the open ocean has been previously documented, but their vertical transport and oceanic sinks have not been comprehensively characterized and quantified at the oceanic scale. During the Malaspina 2010 circumnavigation expedition, 21 PFAS were measured at the surface and at the deep chlorophyll maximum (DCM) in the Atlantic, Indian and Pacific oceans. In this work, we report an extended data set of PFAS dissolved phase concentrations at the DCM. ∑PFAS at the DCM varied from 130 to 11[thin space (1/6-em)]000 pg L−1, with a global average value of 500 pg L−1. Perfluorooctanesulfonate (PFOS) abundance contributed 39% of ∑PFAS, followed by perfluorodecanoate (PFDA, 17%), and perfluorohexanoate (PFHxA, 12%). The relative contribution of the remaining compounds was below 10%, with perfluorooctanoate (PFOA) contributing only 5% to PFAS measured at the DCM. Estimates of vertical diffusivity, derived from microstructure turbulence observations in the upper (<300 m) water column, allowed the derivation of PFAS eddy diffusive fluxes from concurrent field measurements of eddy diffusivity and PFAS concentrations. The PFAS concentrations at the DCM predicted from an eddy diffusivity model were lower than field-measured concentrations, suggesting a relevant role of other vertical transport mechanisms. Settling fluxes of organic matter bound PFAS (biological pump), oceanic circulation and potential, yet un-reported, biological transformations are discussed.



Environmental significance

Scientific and social concern has risen globally towards poly- and per-fluoroalkyl substances (PFAS), which are bioaccumulative and recalcitrant, and some are even toxic. For instance, perfluorooctanesulfonate (PFOS) has been phased out of production and included in the Stockholm Convention list. Their physico-chemical properties make them prone to using the open ocean as a final sink. However, studies regarding their occurrence in the global ocean are mainly focused on surface waters (top 5–10 m), with few reports for their occurrence and fate in deeper waters. In this work, we report the concentrations of PFAS at the deep chlorophyll maximum depth at 89 stations in the Atlantic, Indian and Pacific oceans with simultaneous estimates of vertical fluxes by eddy diffusivity. A modelling effort and suggestions for biotic and abiotic processes affecting the PFAS fate in the global ocean are also proposed.

Introduction

Perfluoroalkyl substances (PFAS) are persistent synthetic compounds causing concern because of their adverse effects, ongoing use and global occurrence, reaching even remote oceans.1–5 Although earlier studies reported PFAS occurrence in the marine environment, including assessments at the oceanic scale6–8 and a few inter oceanic comparisons,1,5,9,10 an understanding of their ultimate fate represents an important scientific uncertainty.11 Neutral PFAS can be atmospherically transported and deposited globally,12–14 while for ionisable PFAS, such as perfluoroalkyl carboxylic acids (PFCAs) and perfluoroalkyl sulfonic acids (PFSAs), marine currents are thought to be the main transport vector to remote oceans.15 Atmospheric deposition of ionisable PFAS is also possible following photo-oxidation of neutral PFAS,14,16,17 but the relative importance of this input into the ocean has been poorly evaluated. Once in the oceanic water column PFAS residence times and their global fate are still uncertain due to their unusual chemical properties (e.g. lipophobicity and moderate water solubility) compared to those of other Persistent Organic Pollutants (POPs).11

Most previous assessments of PFAS in the marine environment have reported concentrations in surface waters,6,18,19 while the transport routes (horizontally and vertically) and the biogeochemical mechanisms controlling their occurrence have received little attention to date.20 The magnitude of the oceanic sinks of PFAS remains largely uncharacterized,15,21,22 but is probably relevant as suggested previously.2,20 To date, a limited number of studies have focused on understanding PFAS vertical transport and distribution in the water column of the marine environment.15,20,23–26 A remarkable study by Yamashita and co-workers15 reported the vertical distribution of PFAS concentrations, from the surface down to several thousand meters deep at 9 locations in the North Atlantic and South Pacific and offshore Japan. Their study showed perfluorooctanoate (PFOA) and perfluorooctanesulfonate (PFOS) concentrations to be generally higher at the surface compared to deep waters, consistent with these chemicals being introduced at the surface through either riverine inputs or atmospheric deposition. Subsequently, Lohmann and co-workers used this database to test the relevance of vertical eddy diffusion as a process affecting PFOA depletion from surface marine waters and driving their transport to the deep ocean.23 Vertical eddy diffusion was suggested as an oceanic sink with a magnitude three fold larger than that of subduction of water masses, which had previously been identified as a relevant sink for other organic pollutants27 and PFAS.2 Both vertical transport by eddy diffusion and subduction of water masses are part of the “physical pump”. Lohmann and co-workers23 also noted the scarce information on the efficiency of vertical mixing due to the limited estimates of eddy diffusion coefficients derived from field data, as well the lack of knowledge about the occurrence of PFAS in many oceanic regions, including the Indian and Pacific oceans. A recent study on PFAS in the Arctic Ocean26 has provided additional data on PFAS distribution with depth, including vertical profiles of a few hundred meters deep. It revealed the influence of the atmospheric inputs in remote areas, although mixing and riverine inputs are key parameters for PFAS distribution in this polar area. However, the particular conditions of the Arctic Sea regarding temperature stratification, ice coverage and physical isolation, make it a different ecosystem compared to the tropical and subtropical ocean gyres, which are still understudied.

The “biological pump” is a relevant sequestration and transfer mechanism to the deep ocean for POPs that adsorb onto settling organic particles.28–31 This process has been suggested to play a significant role in removing PFAS from the surface oceans,20 as previously observed for other POPs in remote areas, such as the Arctic28,29 and the Southern Ocean32,33 and other marine environments.34 PFAS adsorb onto particulate organic matter in aquatic systems8,35 to a lower extent than other POPs due to their lower octanol–water partition coefficient (Kow), yet PFAS have been described to be affected by this settling process in the oligotrophic open ocean.20 A comparison of the role that the biological pump plays as an oceanic sink of PFAS with the other plausible fluxes affecting their global fate is however still missing.

Therefore, the objective of this work was to report the occurrence of PFAS in a large set of samples from the deep chlorophyll maximum depth, and explore removal fluxes from the surface ocean for a wide range of PFAS through: (i) the evaluation of the horizontal and vertical variability of their concentrations in the global tropical and subtropical open ocean, (ii) the quantification of vertical fluxes of PFAS by turbulent diffusion based on onsite measurements of concentrations and eddy diffusivities and (iii) the approximation to the relative importance of other fluxes like the biological pump, deposition and vertical advection affecting PFAS inventories at the oceanic scale.

Material and methods

Sampling

Samples for PFAS analysis were collected during the Malaspina 2010 circumnavigation expedition carried out on board RV Hespérides from December 2010 to July 2011, providing the first global sampling of PFAS in the tropical and subtropical Atlantic, Pacific and Indian oceans between 35° N and 40° S.

PFAS concentrations in surface waters have been described in a previous study and these are used here when needed.5 A total of 89 water samples at the DCM depth were taken during the Malaspina 2010 circumnavigation (Fig. 1). The DCM depth ranged between 20 and 160 m during the circumnavigation cruise, depending on the light and nutrient availability in the water column.36,37 The DCM depth was chosen because it is where the concentration of chlorophyll, used as a proxy for phytoplankton biomass, is at its maximum, and it has been reported to be an area of special interest regarding biological processes.37–39 As phytoplankton is the first step of the aquatic food web, the study of the DCM depth is especially important to determine the exposure of organisms to dissolved phase bioavailable PFAS. The DCM depth was determined just before water sampling by measuring the fluorescence in the water column as detailed in the ESI (Text S1). One liter of seawater from the DCM depth was sampled using Niskin bottles attached to an oceanographic rosette. The water was then transferred to 1L polypropylene bottles for its subsequent concentration in the boat laboratory (ESI, Text S1). All the sampling station locations and depths for DCM samples are included in the ESI (Table S1).


image file: c9em00266a-f1.tif
Fig. 1 Occurrence of the three PFAS families analyzed. (A) PFSAs, (B) PFCAs and (C) FOSAs at the DCM depth. For the (A) and (B) panels, the squared map shows data with a larger scale to ease the visualization of much larger concentration values.

Sample treatment & instrumental analysis

The DCM water samples were analyzed following the same protocols used for surface water samples as described elsewhere,5 simultaneously with that sample set and in a random order; therefore the results obtained are fully comparable (ESI Text S1 includes a full description of the sampling and sample treatment). Briefly, immediately after collection, water samples were filtered using glass fiber filters (GF/F, 0.7 μm, Whatman) and spiked with a mixture of 13C labeled C4,6,8–12 PFCAs and 18O C6 and 13C C8 PFSAs as surrogates to calculate the PFAS recovery. The filtrate was concentrated using solid phase extraction Oasis WAX cartridges (6cc, 150 mg, 30 μm, supplied by Waters) and kept at −20 °C during the cruise until their further elution in the laboratory.

The instrumental analysis was performed using a Waters Acquity Ultraperformance Liquid Chromatography system coupled with a Waters XEVO TQS triple-quadrupole mass spectrometer (UPLC-MS/MS) with negative ion electrospray ionization (ESI) operating in the multiple-reaction-monitoring (MRM) mode. Further details are provided by González-Gaya et al.5 and in the ESI (Text S1). The analytical method targeted 21 PFAS in the dissolved phase, including PFCAs, PFSAs and perfluoroalkyl sulfonamides (FOSAs). Three labeled PFSAs and PFCAs (PFOS 13C4, PFOA13C8, and PFUnDA 13C7) and one deuterated FOSA (d3-N-MeFOSA) were used as injection standards for internal standard quantification. Of the 21 PFAS analyzed, 9 ionisable PFAS (C6–C10 PFCAs and C4, C6–C8 PFSAs) and 2 neutral FOSA precursor compounds (perfluorooctane sulfonamide (FOSA) and N-methyl perfluorooctane sulfonamide (N-MeFOSA)) were consistently identified (see the ESI, Text S2). Therefore, the present study focused on these 11 compound concentrations in seawater, though the calculations for the fluxes only include the 9 ionisable species and not the precursors.

PFAS concentrations at the DCM from a subset of 28 samples have been reported previously,20 but here we report the extended analysis of PFAS for 89 samples of seawater from the global ocean, which ensure a much wider coverage of the open tropical and subtropical oceans. In addition, the previously reported PFAS concentrations in plankton20 are recalled here to calculate the biological pump fluxes, including PFOS and PFOA isomer concentrations (which were not measured in the water samples). Details on the sample treatment and analysis of the plankton can be found elsewhere.20

Quality assurance/quality control

Each sample was injected in triplicate and all the detection and quantification limits (DL and QL) were calculated as described in detail elsewhere.5 Field and laboratory blanks were analyzed simultaneously with the sample batches, randomized within the surface and DCM water samples, and showed no significant contamination with the target compounds.5 Average recoveries for the DCM water samples ranged from 76% for perfluorononanoate (PFNA) 13C5 to 150% for perfluorohexane sulfonate (PFHxS) 18O2 and samples were not recovery corrected. Further details of blanks, limits and recoveries can be found in the ESI (Text S2).

Estimation of vertical diffusive fluxes of PFAS

The oceanic sink due to vertical eddy diffusion was calculated for individual PFAS with a one dimensional diffusion model as described elsewhere.23 Briefly, the estimation of the turbulent flux (FEddy, ng m−2 d−1) is based on Fick's first law,
 
image file: c9em00266a-t1.tif(1)
where Cw (ng m−3) is the PFAS seawater concentration, z is the depth, and Kρ is the eddy diffusivity (m2 d−1). In the model, it is assumed that no PFAS pollution was present at the marine surface in the initial time step (Cw0 = 0, t0 = 1970) and that the net flux is unidirectional from the surface towards the deep ocean, driven solely by eddy diffusion. T0 was selected as the moment when the production of perfluorochemicals started to be relevant, but their oceanic concentrations were still negligible.40 As boundary conditions, the surface concentrations are allowed to maintain a uniform increase from 1970 until reaching the CW values measured for the surface ocean in 2011, as reported in a previous study.5 Such a time trend of concentrations has been suggested previously.27,41 Time steps are fixed at Δt = 0.5 years and the resolution of the water column is 1 m depth. The vertical profile of concentrations was determined by integrating Fick's second law using the Crank-Nicholson integration method. Details of the model are given by Lohmann et al.23

K ρ values were calculated from measurements of dissipation rates of turbulent kinetic energy carried out during the Malaspina 2010 circumnavigation by using a microstructure profiler, as reported previously.42 The microstructure profiler was deployed, immediately after the water sampling for PFAS analysis, at 50 stations down to a maximum depth of 300 m. Between 2 and 6 turbulence profiles were completed at each station, and Kρ values used in this study correspond to averages from the total number of profiles completed at the same station. Kρ was calculated from the turbulent kinetic energy dissipation rate and local stratification following the Osborn 1980 model.43 An adaptation of the K-profile parameterization described by Large et al.44 was used to compute Kρ at those stations where microstructure turbulence observations were not available (details provided by Fernández-Castro et al.42). It must also be noted that the Osborn model used to estimate the diffusivity from turbulent kinetic energy dissipation rates under stratified conditions might be of limited applicability to the surface layer, as stratification is sometimes undetectable in this part of the water-column. A comparison of the diffusivity derived from the Osborn model with the diffusivity derived from Kρ, which is based on the Monin–Obukov theory and thus is not affected by this limitation, revealed consistent ranges of variability between both methods during the Malaspina Expedition,42 suggesting that the use of the Osborn formula is suitable for the purposes of this study.

The estimated FEddy values are based on the modeled decrease of PFAS concentrations with depth. At the surface, the concentrations were those measured during the sampling cruise,5 and the Kρ was averaged per meter over the water column at each sampling point. Fluxes at surface waters (considering an averaged value for the upper 5 to 15 m) and for the DCM depth of each location are given for 9 ionisable species (C6-10 PFCAs and C4, 6-8 PFSAs) as their detection over various depths in the sampling campaign was consistent, allowing their comparison in all the tropical and subtropical oceans.

Statistical analysis

SPSS Statistics version 21.0 (IBM Corp.) was used for non-parametric statistical analysis as the measured concentrations of individual PFAS were not normally distributed (Kolmogorov–Smirnov normality test, p < 0.01). Correlations and comparisons between groups were done using Spearman's Rho regression analysis and Kruskal–Wallis single factor one-way ANOVA.

Results and discussion

Global occurrence of PFAS at the deep chlorophyll maximum depth

The dissolved phase concentrations of ∑PFAS (11 target PFAS above the QLs) at the DCM depth vary from 130 to 11[thin space (1/6-em)]000 pg L−1, with a global average of 500 pg L−1 (median 440 pg L−1). The high abundance of PFOS was remarkable, contributing 39% of the total average PFAS concentration on a global scale, followed by perfluorodecanoate (PFDA, 17%), and perfluorohexanoate (PFHxA, 12%). The relative contribution of the remaining compounds was below 10%, with PFOA contributing only 5% of ∑PFAS measured at the DCM (Fig. 1 and ESI Table S2 and Fig. S1 and S2).

Regarding global coverage, the Atlantic Ocean (n = 44) showed the highest average concentrations of ∑PFAS at the DCM (median, 460 pg L−1), followed by the Pacific (n = 27; median, 440 pg L−1) and the Indian Ocean (n = 18; median, 380 pg L−1). As found previously,5 concentrations in the Southern hemisphere (median, 520 pg L−1) were significantly higher (p < 0.01) than those in the Northern hemisphere oceans (median, 350 pg L−1), mainly due to the great concentrations found close to the Brazilian, South African and West Australian coasts (Fig. 1 and ESI Fig. S2).

The global distribution of the three studied PFAS families at the DCM displays different patterns (ESI Fig. S2). PFSAs showed a clear coastal gradient with the highest concentrations, even 1 or 2 orders of magnitude higher than baseline concentrations, in the South Atlantic Ocean near the Brazilian coast, as observed previously at the surface5 (Fig. 1A). The high increase of PFSA concentrations in the South Atlantic basin compared with that in the Northern is remarkable (ESI Fig. S2). PFCAs followed a similar spatial occurrence pattern, but the highest concentrations were found near West Africa in the North Atlantic gyre (also one order of magnitude higher than the average), and close to the coastal margins of the Indian Ocean transect (Fig. 1B). Concentrations of the two FOSAs analyzed (Fig. 1C) were 2 or 3 orders of magnitude lower than those of the two ionisable families, very close to their DL. Neutral PFAS compounds have been reported here for the first time in deep tropical waters, although they were detected in merely 60% of the samples and mostly in the Northern hemisphere. This can be due to their chemical characteristics such as high volatility and lower persistence which would lead to a lower occurrence in the deeper remote ocean.7

Observations on the vertical distribution of PFAS

The PFAS geographic pattern found at the DCM depth resembled that described previously for the corresponding surface waters5 for most ocean basins with the exception of the North Atlantic Ocean (ESI Table S2, Fig. S1 and S2). In this sub basin, PFCAs were predominant at the surface (almost 80% of ∑PFAS)5 but at the DCM depth PFCAs and PFSAs had similar average amounts (55% and 42%, respectively), specifically due to PFOA and PFDA relative lower concentration with depth. This could be caused by a contemporary discharge of PFOA and PFDA in the North Atlantic, mainly originating from eastern coasts (e.g. Europe and North Africa) according to the surface occurrence pattern,5 but no evidence for these sources can be shown here.

There is little information on PFAS oceanic depth profiles, which have been summarized in Table S3 (ESI). Still, the general decrease in PFAS concentrations with depth, reaching common concentrations of a few pg L−1, was observed previously in profiles in the Mid Atlantic, Southern Pacific, Japan Sea and Arctic Ocean.15,26,45 Similarly, concentrations of PFOS, PFHxS and PFOA were detected in samples over 1000 m deep in the Sulu Sea (Japan) and even below 4000 m in the southeast Pacific Ocean1 but are always lower than those at the surface. In addition, in a single sample taken at 1390 m in the Mediterranean Sea the concentration of total PFAS was two times lower than the equivalent in surface water.24 Likewise, in the coastal water of Tokyo Bay, with only a 20 m deep water column, a slight decrease of some PFAS was observed with depth, mainly due to adsorption to bottom sediments of single compounds according to their physicochemical properties.8 This indicates a different behavior and fate of congeners depending on their length and polar head characteristics, but also on the suspended organic matter present in the water column. Similarly, four profiles at Baffin Bay (between Greenland and Canada) showed a general decrease of PFOA with depth, although peaking around 100 m depth in all sites, and at 600 m depth in one site, probably due to interactions with organic matter.22 Profound anomalies with depth (deeper than the sampling depths provided here) have also been shown, due mainly to deep oceanic currents, as the North Atlantic inflow towards the Arctic results in higher concentrations of PFAS below the halocline than on surface waters.15

Regarding ∑PFAS measured at the surface versus at the DCM depth, both concentrations were statistically correlated for all individual compounds (p < 0.01), and concentrations were significantly lower at the DCM (p < 0.01) (Fig. 2). In addition, the concentration range at the DCM was lower than that at the surface,5 exhibiting a lower variability of PFAS concentrations on a global scale at the bottom of the surface mixed layer or just below it, than at the surface.


image file: c9em00266a-f2.tif
Fig. 2 Relation between concentrations (pg L−1) measured at the surface and at the DCM depth in the same sampling stations. Each regression curve equation and adjusting value (R2) are included in the graph.

In 18% of the sampling locations shown in this global assessment, some congeners' concentrations at the DCM were higher or not statistically different from those measured at the surface5 for some compounds (ESI Table S2 and Fig. S2). Assuming a mainly superficial input of PFAS, it seems that in some areas the factors affecting depletion of concentrations are more intense than in others, revealing 3 different patterns. The more usual one corresponds to a decreasing concentration of all PFAS with depth. The remaining (<20%) cases correspond to locations where PFAS concentrations at the DCM were in the same range as those at the surface, being those (a) in regions with the lowest surface PFAS concentration (i.e. the central Pacific Ocean), where the total levels are close to the baseline concentrations and therefore, low concentrations along the whole column (ESI Fig. S2) or (b) in areas with extremely high concentrations combined with intense removal fluxes (Central American, South African and Australian coasts) where high levels are found along the analyzed depths (ESI Fig. S2). Apart from vertical settling and diffusion, horizontal transport through oceanic currents26 or inner riverine/coastal discharges7,25 cannot be excluded as other drivers affecting PFAS concentrations with depth. Accounting for these inversions or region-specific conditions, it is relevant to assess at a global scale the magnitude and spatial distribution of PFAS sinking fluxes (eddy diffusion and the biological pump among others) in order to understand their vertical and spatial distribution.

Vertical eddy diffusion fluxes of PFAS

Eddy diffusion coefficients, Kρ, obtained from the measured turbulence, ranged from 5.1 m2 s−1 in surface areas of the North Atlantic (where events of significant atmospheric forcing occurred enhancing turbulent fluxes) to 3.5 × 10−7 m2 s−1 in the deepest areas of the measured mixed layer, causing in the latter a very slow diffusion typical for the inner parts of the ocean42 (ESI Fig. S3). In surface waters, Kρ shows a high variability due to the strong influence of wind speed, radiation and currents, with a global average Kρ of 1.1 × 10−1 m2 s−1 with 2 to 3 orders of magnitude of variability. This average is obtained considering in situ measurements, even if at some sampling points there may be differential stratification conditions in the mixed layer, which may alter the local calculated dissipation rates at the surface. However, for the purposes of this analysis, the averaged values are considered good estimations for the global ranges. Around the DCM, Kρ showed a lower variability with a global average of 1.4 10−5 m2 s−1, few orders of magnitude lower than at the surface and to some extent less variable than the surface measurements. It is necessary to note that the diffusivity values reported reflect the regional and temporal conditions at the precise sampling moment particularly at the surface layer, where turbulent mixing is strongly coupled with the atmospheric forcing which varies on several spatial and temporal scales (from weather to climate), and to a lesser extent, at the DCM level due to the more indirect coupling with the atmospheric forcing. Averaged values over regions, as suggested by Fernández-Castro and co-workers42 would help to have a reasonable mean value of turbulence over a latitudinal band or an oceanic sub-basin. However, regional variability on a global scale can result in broad ranges for diffusive fluxes, even in the ocean interior, due to the variability of energy sources and instabilities in the flow. This can be especially relevant in areas with high shear-induced mixing and salt fingers, where diffusivity can increase more than 20% due to the turbulent mixing.42 In order to preserve this natural large variability, here we used the in situ measured values as proxies of the wide ranges turbulence may show in the global ocean. Due to the diversity of sampling regions and the variability in the local meteorological and sub-surface stratification conditions captured during our study, the range of variability of the diffusivity values that we were able to capture can possibly be considered as a reliable estimate of the global ranges. In order to obtain that data, annual extensive monitoring would be needed, which is out of the scope of this study.

The application of the one-dimensional model using the measured Kρ allowed the prediction of the vertical profile of CW in the water column; thus it was possible to estimate the value of CW at the DCM depth if eddy diffusion was the only mechanism for vertical transport. To test the model's goodness-of-fit we compared the measured field concentration of individual PFAS at the DCM with those obtained from the model at the same depth. Generally, the model underestimated the measured concentrations (the mean relative error (in absolute value) ranging from 63% for PFOA to 120% for perfluoroheptanesulfonate (PFHpS)) (ESI Table S4 and Fig. S4). In the South Atlantic Ocean, where maximal concentrations were found at the surface,5 there was a greater difference between modeled and measured concentrations, particularly for PFSAs (ESI Fig. S4). As explained above, physicochemical characteristics of the target compounds and physical and biological settling forces at the different areas of the analyzed oceans largely affect the real concentrations measured, as can be seen as an example for PFOS and PFOA in the ESI (Fig. S5) when compared with the modeled values that only take into account vertical transport due to turbulence.

F Eddy calculated from the modeled concentrations decreased with depth (Fig. 3 and ESI Table S5), consistent with the microstructure turbulence being 1–7 orders of magnitude lower in the mesopelagic ocean than in the photic layer of the open ocean at tropical and subtropical latitudes.42FEddy in the photic mixed surface waters was generally higher, between the same and up to 4 orders of magnitude greater than at the DCM (ESI Table S5), especially in the Atlantic Ocean (Fig. 3). An inversion of turbulence flux intensity occurred consistently for all analyzed compounds only at 5 sampling stations (located at the Brazilian coast, in the Pacific Equatorial Countercurrent and near the Caribbean sea) corresponding to areas near the coast and/or where conditions like sheared currents or salt-fingers are present, causing increased interior mixing areas as reported previously.42


image file: c9em00266a-f3.tif
Fig. 3 Turbulent flux average profile per ocean for (A) PFOS and (B) PFOA (FEddy, ng per m2 per day).

F Eddy at the surface (averaged top 15 m) oscillated between 2.9 × 10−8 and 5.4 × 10−2 ng per m2 per day for PFOS and between 3.2 × 10−8 and 9.1 × 10−4 ng per m2 per day for PFOA. FEddy at the DCM depth (corresponding value at each station) ranged between 1.5 × 10−5 and 2.0 × 10−2 ng per m2 per day for PFOS and between 1.1 × 10−6 and 1.3 × 10−3 ng per m2 per day for PFOA (Fig. 3, ESI Table S5 and Fig. S6). The turbulent fluxes calculated by Lohmann et al. for PFOA were of the same order of magnitude and also exhibited broad ranges in magnitude depending on the location.23FEddy calculated was lower for the less abundant PFAS (e.g. PFHpS and PFNA). The PFOS eddy diffusion flux was one order of magnitude greater than that of PFOA due to the higher concentrations of PFOS. However, even at the maximal measured Kρ for PFOS, eddy diffusion would sequester 7 ng m−2 of PFOS per year from the oceanic surface, thus representing a minor (slow) flux.

The PFOS and PFOA annual removal fluxes for the global tropical and subtropical surface oceans (computing mean turbulent fluxes over the mixed layer per oceanic sub basin over the total area of each region46) were 2.3 and 0.2 tons per year, respectively. Yamashita et al.15 reported a rate of 0.6 tons per year for PFOS and 1.5 tons per year for PFOA due to deep water formation globally, as an annual average, from samples taken on several cruises between 2002 and 2006.

Lohmann and coauthors23 estimated a total amount of PFOA removed from the top 100 m of the global ocean by turbulence of more than 660 tons (from 1970 to 2009), representing a mean of 17 tons per year, 85 times greater than our calculated values for this compound. These authors used a range of plausible diffusivity values in the thermocline (10−5–10−3 m2 s−1) which results in a wide range of the estimated PFOA export from 1.5 tons to 4800 tons since 1970, placing our values at the lower end of their calculations. The lower-end diffusivities used by Lohmann are of the same order of our diffusivity estimates, and close to the typical values in the main pycnocline, as derived from microstructure measurements and tracer experiments.47,48 However, the higher-end estimates were adjusted to match the vertical distribution of PFOA, in line with other budgeting approaches that represent the integrated effect of several mixing processes, and typically produce values of diffusivity an order of magnitude larger than point-wise microstructure measurements.49 They also estimated the mean flux due to deep water formation of 220 tons of PFOA (80–360 tons) for the period 1970–2009, and thus between 2 and 9 tons per year.23 The global diffusive fluxes reported here are of similar magnitude than the mean fluxes due to the deep water formation reported by Lohmann as well as that reported by Yamashita et al.15 (0.6 tons per year for PFOS and 1.5 tons per year for PFOA). Therefore, our estimation of surface annual FEddy is comparable to the lower end estimate of subduction fluxes according to their calculations.

As a final remark, it is noteworthy that eddy diffusion of the different pollutants is not constant over the open ocean at a fixed depth, and this unpredictability is driven not only by the variability in turbulent mixing but also by the concentration of pollutants. For instance, at the DCM depth, a very high FEddy was recorded for PFOS in the east South Atlantic, probably due to elevated concentrations of the pollutant coupled with an intense internal turbulence42 or due to the Amazon plume intrusion.50 High FEddy peaks for PFOA at certain points of the West African coast in the North Atlantic Ocean were found, probably associated with a significant occurrence of PFOA (Fig. 4). Shallower areas (such as the Tasmanian straight), upwelling waters (near the South African west coast) and natural stratification of the mixed layer (for instance due to riverine discharge or cold water intrusions), among other processes, affect the intensity of the turbulence42 as reflected in our measurements. Thus, local to regional as well as temporal conditions should be taken into account for calculating removal processes of surface pollutants in the open ocean, as shown in the variability captured in this database.


image file: c9em00266a-f4.tif
Fig. 4 Flux assessed for PFOS and PFOA; turbulent fluxes (FEddy) on the top and biological pump fluxes on the bottom (FPhyto and FFecal). Bars with the ≈ symbol have been manually diminished by a factor of 10 in order to ease the global comparison of all the measurements. Real values are indicated close to the reduced corresponding bar.

Vertical fluxes of PFAS due to the biological pump

The settling flux (FSettling) associated with the settling of organic matter bound PFAS (biological pump) has been previously estimated20 on an annual basis using
 
image file: c9em00266a-t2.tif(2)

However, the settling fluxes of organic carbon in the oligotrophic open ocean largely differ between the fluxes associated with algae from those caused by zooplankton sinking pellets,51 therefore it can be stated that the biological pump is equal to

 
image file: c9em00266a-t3.tif(3)
where FOM (mg per m2 per day) is the flux of organic matter that settles out of the euphotic layer to deeper waters with an algal matter flux component (FPhyto) and a zooplankton related matter flux component (FFecal), CPlankton (ng gdw−1) is the concentration of PFAS measured in plankton during the Malaspina 2010 circumnavigation,20 and fOC (g C gdw−1) is the organic carbon fraction measured in plankton.31

Fluxes of settling carbon (FOC, mg C per m2 per day) for algae and zooplankton can be found separately in the climatology of Siegel and coauthors.51 As suggested previously,20,52 these values allow estimating FOM, as organic matter is composed of an average of 55% carbon. Therefore, both the phytoplankton and zooplankton contribution to the biological pump fluxes were estimated separately here for the first time, as previous approximations to POP biological pump flux calculations considered one single settling flux of organic matter.20,28,29 Even if there are no available data for PFAS, fecal pellets have been reported to accumulate several organic pollutants.53–55 It is assumed that the plankton samples include this organic material as well and that PFAS concentrations in fecal pellets are equal to those in plankton, as described for other organic pollutants.55 However, these data are at the moment only available for PAHs, and there is an urgent need for measurements for PFAS and pollutants. Due to this lack of detailed information on PFAS in different types of organic matter, and taking into account the uncertainties, we assume that the measured concentrations in plankton reported previously20 are similar to those in settling organic matter.

Siegel et al.51 modeled FOC for algae (FOC Phyto) and zooplankton fecal matter (FOC Fecal) on a monthly basis in a one degree resolution global grid. We used the corresponding monthly average export for the precise sampled positions to obtain the FOC to calculate the biological pump settling fluxes (ESI Table S6). FSettling was calculated independently for the algae and zooplankton related PFAS sinking fluxes where Cplankton was available (Fig. 4 and ESI Table S7).

Algal sinking fluxes ranged from 10−6 ng per m2 per day for PFHxS to 10 ng per m2 per day for PFPeA and PFHpA. Sinking fluxes attributed to zooplankton fecal pellets ranged between 10−4 ng per m2 per day for PFHxS and 101 ng per m2 per day for linear PFOS. Fluxes associated with the zooplankton fecal matter were found to be between one and two orders of magnitude larger than those associated with phytoplankton for the measured concentrations of PFAS (Fig. 4 and ESI Table S7). The total median FSettling (algal + zooplankton) was 1.4 and 0.72 ng per m2 per day, respectively, for PFOS and PFOA. PFOS and PFOA differed not only in their rates of export from the surface, but also in their global patterns (Fig. 4). A higher FSettling of PFOS was recorded in the equatorial Atlantic Ocean, in the area of influence of the Brazilian coast, where the concentration of this compound has been reported to be extremely high,20,22 but also near South African and Indian coasts and in Australian vicinity areas. ’The FSettling of PFOA was also intense in the Atlantic Ocean and Central American west coast (Costa Rica dome) where high concentrations of PFOA coincided with significant phytoplankton biomass,56 even if the maximum export fluxes were found near the Australian continent. The relative importance of the algal sinking fluxes is remarkable in this region.

The reported PFAS data for the plankton samples include the concentrations of different PFOS and PFOA isomers,20 and therefore, the particular settling fluxes for these homologues could be calculated. The biological pump fluxes for the PFOS and PFOA branched isomers were relatively more intense in the Atlantic basin and near Australia (ESI Table S8), where the percentage of these compounds is higher, especially for PFOA branched isomers.20

The flux intensity does not correlate with the hydrophobicity of the compounds, as for instance the long chain PFCAs that have higher KOW57 exhibit very low fluxes (ESI Fig. S7). This is due to the observed lack of correlation between plankton bioaccumulation factors (BAFs) and KOW reported previously.20 Only PFOA showed a particularly large plankton BAF when compared with the other measured PFAS.20 Parameters other than hydrophobicity have been suggested to modulate the PFAS interaction with organic matter, although more research is needed in this field.58 Independent of their bioaccumulation potential, the removal flux due to the biological pump will tend to deplete PFAS from the photic zone dissolved phase. For instance, there was a negative significant correlation between FSettling and the DCM dissolved phase concentration for PFOA (Spearman's Rho, p < 0.05) (ESI Fig. S8), which supports the role of the biological pump in depleting the water column concentrations.

There are some uncertainties associated with the estimations of the magnitude of the biological pump, including those due to annual and regional variations. For example, zooplankton egg production, included in some studies as a contribution to total planktonic export of POPs,30,55 may enhance the FSettling particularly due to the tendency of PFAS to bind to albumin and proteins.58 Moreover, as zooplankton undergo vertical daily migration, PFAS will equilibrate both at the surface and at a depth due to this migration. Since the concentrations are lower in deeper waters, at the DCM or lower, they would be desorbed from the organic matter due to partitioning, which could induce an additional flux as suggested for other organic pollutants.55,59,60

It would also be interesting to ascertain the biological pump effect on the fractionation and preferential accumulation in the zooplankton versus phytoplankton and bacteria, and explore any potential biodegradation, metabolization, biotransformation or bioaccumulation of PFAS as suggested for other POPs.31,34,59,61–64 However, these processes cannot be estimated here, as we have no information on the kinetics of the partitioning behavior, and no metabolization capacity for PFAS has been described for marine plankton or bacteria yet.

Global oceanic sinks and PFAS fate

Vertical turbulent diffusion fluxes are low, with maximum values generally linked to extraordinary events of mixing. Sinking fluxes due to the biological pump, several orders of magnitude higher than the turbulent fluxes, seem to dominate the removal processes in the water column (Fig. 4 and ESI Tables S4 and S6). Nevertheless, both removal processes seem to be slow, which can be corroborated by the estimates of the residence times of PFAS in the surface ocean. A mean PFAS concentration between the surface and DCM can be calculated to obtain an averaged inventory of PFAS in the global marine upper euphotic zone of the water column. The averaged inventory for PFOS and PFOA in the euphotic layer would be in the order of 40 μg m−2 and 4 μg m−2, respectively. The estimated mean FEddy plus FSettling for the individual compounds would range from 14 ng per m2 per year for PFDA to 520 ng m−2 of the annual export for PFOS (ESI Table S9). Therefore, the estimated mean residence time for PFOS and PFOA in the mixed layer is 360 and 32 years, respectively (ESI Table S9). These values, even if they still are quite large, are far more realistic than those calculated by assuming eddy diffusive fluxes alone (of hundreds and thousands of years). Nevertheless, at the maximal sinking flux reported, the residence time calculated is 1 year residence time for PFOA and 6 years for PFOS, significantly lower than the calculated averages. However, it must be taken into account that turbulence used to perform the FEddy calculations can be enhanced in areas not sampled (e.g. upwellings or deep mixing areas), in different periods of the year, or due to undersampled events. For example, rapid deepening of the mixed layer during convective mixing events can potentially represent a significant pathway for vertical transport, which is likely undersampled in our dataset, where conditions favorable for convection were only sampled in a few stations.42 In addition, monthly averaged climatology is being used for the FSettling calculations; thus in some specific events vertical transport could be higher due to nutrient pulses or to highly productive seasons. Besides, the oceanic sink due to the FSettling estimated here is a lower-end figure of the global oceanic sink, since the regions where a higher biological pump export occurs (like the Northern Atlantic and Pacific oceans, the Southern Ocean and upwelling areas, like the western coast of South America)65 have not been assessed. Thus, it should be noted that these calculations should be taken as mid to lower estimates for sinking fluxes due to diffusion and the biological pump.

Additionally, many other processes may affect the surface removal of PFAS on a global scale. Export of organic matter and particles from the continental shelves to the deep marine environment, through dense shelf water cascading, has been reported for PFAS.66 In closed seas and coastal areas, those processes, riverine deep intrusion and downwelling can be of high relative importance.24 Moreover, the formation of deep oceanic water will also deplete PFAS from the global ocean surface. The latter is only relevant in certain regions of the North Atlantic and Southern oceans26,67 not sampled during the Malaspina 2010 expedition, but should be quantified properly for a global assessment. Also extreme rare events cannot be neglected, and even if very few data are available in the literature,45,68 it can be relevant to account for the relative effect of dispersion due to slow diffusive processes versus fast mixing events.

Another plausible process affecting environmental removal of PFAS from the marine surface is photodegradation, recently reported for a few compounds, but only under precise conditions difficult to reach in the oceanic environment69 although it has been called into question.70 Likewise, biodegradation has not been reported yet for these chemicals.71–74 Nevertheless, surface/DCM concentration ratios for PFOA, for PFCAs, (p < 0.01) and for the total PFAS (p < 0.05) correlated with bacterial activity concurrently measured during the sampling, as well as bacterial size and total bacterial carbon content (p < 0.05), used as a proxy for bacterial communities.75,76 The interaction between PFAS and bacteria needs further research. In a recent study,77 it has been shown that PFOS can modulate bacterial growth and enzymatic activities, an issue that needs further research to elucidate the mechanisms behind these observations. Furthermore, the role of bacteria and the microbial loop is far from being understood and should be the objective of urgent research. The overall vertical transport and fate of PFAS in the oceanic environment needs a multidisciplinary research effort on the interactions of PFAS and the different players of the oceanic carbon cycle (bacteria, phytoplankton, zooplankton, etc.), and how they affect the cycling and biogeochemistry of PFAS in the water column.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was funded by the Spanish Ministry of Economy and Competitiveness (Circumnavigation Expedition Malaspina 2010: Global Change and Biodiversity Exploration of the Global Ocean. CSD2008-00077). The BBVA Foundation is acknowledged for its economic support of the PhD fellowship granted to B. G.-G. The CSIC and MAGRAMA are also acknowledged for additional financial support. P. C. acknowledges a FPI fellowship from the Economy and Competitiveness Ministry. The officers, crew and UTM personnel of the R/V Hespérides are acknowledged for their great support during the field work. We thank Professor David Siegel, from the Earth Research Institute, University of California, for the organic carbon global climatology. Ana Gomes, Prof. Josep M. Gasol (ICM-CSIC), and the rest of the people on board measuring bacterial abundances are also acknowledged. Dr Bieito Fernández Castro and Dr Beatriz Mouriño (Vigo University) are specially acknowledged for their comments on eddy diffusive fluxes calculations.

References

  1. N. Yamashita, K. Kannan, S. Taniyasu, Y. Horii, G. Petrick and T. Gamo, A global survey of perfluorinated acids in oceans, Mar. Pollut. Bull., 2005, 51, 658–668 CrossRef CAS PubMed .
  2. L. Ahrens, Polyfluoroalkyl compounds in the aquatic environment: a review of their occurrence and fate, J. Environ. Monit., 2011, 13, 20–31 RSC .
  3. A. B. Lindstrom, M. J. Strynar and E. L. Libelo, Polyfluorinated Compounds: Past, Present, and Future, Environ. Sci. Technol., 2011, 45, 7954–7961 CrossRef CAS PubMed .
  4. L. Ahrens and M. Bundschuh, Fate and effects of poly- and perfluoroalkyl substances in the aquatic environment: a review: fate and effects of polyfluoroalkyl and perfluoroalkyl substances, Environ. Toxicol. Chem., 2014, 33, 1921–1929 CrossRef CAS .
  5. B. González-Gaya, J. Dachs, J. L. Roscales, G. Caballero and B. Jiménez, Perfluoroalkylated Substances in the Global Tropical and Subtropical Surface Oceans, Environ. Sci. Technol., 2014, 48, 13076–13084 CrossRef PubMed .
  6. L. Ahrens, J. L. Barber, Z. Xie and R. Ebinghaus, Longitudinal and Latitudinal Distribution of Perfluoroalkyl Compounds in the Surface Water of the Atlantic Ocean, Environ. Sci. Technol., 2009, 43, 3122–3127 CrossRef CAS PubMed .
  7. L. Ahrens, W. Gerwinski, N. Theobald and R. Ebinghaus, Sources of polyfluoroalkyl compounds in the North Sea, Baltic Sea and Norwegian Sea: evidence from their spatial distribution in surface water, Mar. Pollut. Bull., 2010, 60, 255–260 CrossRef CAS PubMed .
  8. L. Ahrens, S. Taniyasu, L. W. Y. Yeung, N. Yamashita, P. K. S. Lam and R. Ebinghaus, Distribution of polyfluoroalkyl compounds in water, suspended particulate matter and sediment from Tokyo Bay, Japan, Chemosphere, 2010, 79, 266–272 CrossRef CAS PubMed .
  9. S. Wei, L. Q. Chen, S. Taniyasu, M. K. So, M. B. Murphy, N. Yamashita, L. W. Y. Yeung and P. K. S. Lam, Distribution of perfluorinated compounds in surface seawaters between Asia and Antarctica, Mar. Pollut. Bull., 2007, 54, 1813–1818 CrossRef CAS PubMed .
  10. Z. Zhao, Z. Xie, A. Möller, R. Sturm, J. Tang, G. Zhang and R. Ebinghaus, Distribution and long-range transport of polyfluoroalkyl substances in the Arctic, Atlantic Ocean and Antarctic coast, Environ. Pollut., 2012, 170, 71–77 CrossRef CAS PubMed .
  11. R. C. Buck, J. Franklin, U. Berger, J. M. Conder, I. T. Cousins, P. de Voogt, A. A. Jensen, K. Kannan, S. A. Mabury and S. P. van Leeuwen, Perfluoroalkyl and polyfluoroalkyl substances in the environment: terminology, classification, and origins, Integr. Environ. Assess. Manage., 2011, 7, 513–541 CrossRef CAS PubMed .
  12. A. Jahnke, U. Berger, R. Ebinghaus and C. Temme, Latitudinal Gradient of Airborne Polyfluorinated Alkyl Substances in the Marine Atmosphere between Germany and South Africa (53° N–33° S), Environ. Sci. Technol., 2007, 41, 3055–3061 CrossRef CAS PubMed .
  13. L. Ahrens, M. Shoeib, S. D. Vento, G. Codling and C. Halsall, Polyfluoroalkyl compounds in the Canadian Arctic atmosphere, Environ. Chem., 2011, 8, 399–406 CrossRef CAS .
  14. S. D. Vento, C. Halsall, R. Gioia, K. Jones and J. Dachs, Volatile per- and polyfluoroalkyl compounds in the remote atmosphere of the western Antarctic Peninsula: an indirect source of perfluoroalkyl acids to Antarctic waters?, Atmos. Pollut. Res., 2012, 3, 450–455 CrossRef .
  15. N. Yamashita, S. Taniyasu, G. Petrick, S. Wei, T. Gamo, P. K. S. Lam and K. Kannan, Perfluorinated acids as novel chemical tracers of global circulation of ocean waters, Chemosphere, 2008, 70, 1247–1255 CrossRef CAS PubMed .
  16. J. P. Benskin, L. Ahrens, D. C. G. Muir, B. F. Scott, C. Spencer, B. Rosenberg, G. Tomy, H. Kylin, R. Lohmann and J. W. Martin, Manufacturing Origin of Perfluorooctanoate (PFOA) in Atlantic and Canadian Arctic Seawater, Environ. Sci. Technol., 2012, 46, 677–685 CrossRef CAS PubMed .
  17. K. Y. Kwok, S. Taniyasu, L. W. Y. Yeung, M. B. Murphy, P. K. S. Lam, Y. Horii, K. Kannan, G. Petrick, R. K. Sinha and N. Yamashita, Flux of Perfluorinated Chemicals through Wet Deposition in Japan, the United States, And Several Other Countries, Environ. Sci. Technol., 2010, 44, 7043–7049 CrossRef CAS PubMed .
  18. M. Cai, Z. Zhao, Z. Yin, L. Ahrens, P. Huang, M. Cai, H. Yang, J. He, R. Sturm, R. Ebinghaus and Z. Xie, Occurrence of Perfluoroalkyl Compounds in Surface Waters from the North Pacific to the Arctic Ocean, Environ. Sci. Technol., 2012, 46, 661–668 CrossRef CAS PubMed .
  19. L. Li, H. Zheng, T. Wang, M. Cai and P. Wang, Perfluoroalkyl acids in surface seawater from the North Pacific to the Arctic Ocean: Contamination, distribution and transportation, Environ. Pollut., 2018, 238, 168–176 CrossRef CAS PubMed .
  20. P. Casal, B. González-Gaya, Y. Zhang, A. J. F. Reardon, J. W. Martin, B. Jiménez and J. Dachs, Accumulation of Perfluoroalkylated Substances in Oceanic Plankton, Environ. Sci. Technol., 2017, 51, 2766–2775 CrossRef CAS PubMed .
  21. K. Prevedouros, I. T. Cousins, R. C. Buck and S. H. Korzeniowski, Sources, Fate and Transport of Perfluorocarboxylates, Environ. Sci. Technol., 2006, 40, 32–44 CrossRef CAS PubMed .
  22. J. P. Benskin, D. C. G. Muir, B. F. Scott, C. Spencer, A. O. De Silva, H. Kylin, J. W. Martin, A. Morris, R. Lohmann, G. Tomy, B. Rosenberg, S. Taniyasu and N. Yamashita, Perfluoroalkyl Acids in the Atlantic and Canadian Arctic Oceans, Environ. Sci. Technol., 2012, 46, 5815–5823 CrossRef CAS PubMed .
  23. R. Lohmann, E. Jurado, H. A. Dijkstra and J. Dachs, Vertical eddy diffusion as a key mechanism for removing perfluorooctanoic acid (PFOA) from the global surface oceans, Environ. Pollut., 2013, 179, 88–94 CrossRef CAS PubMed .
  24. M. Brumovský, P. Karásková, M. Borghini and L. Nizzetto, Per- and polyfluoroalkyl substances in the Western Mediterranean Sea waters, Chemosphere, 2016, 159, 308–316 CrossRef PubMed .
  25. Y. Zhou, T. Wang, Q. Li, P. Wang, L. Li, S. Chen, Y. Zhang, K. Khan and J. Meng, Spatial and vertical variations of perfluoroalkyl acids (PFAAs) in the Bohai and Yellow Seas: Bridging the gap between riverine sources and marine sinks, Environ. Pollut., 2018, 238, 111–120 CrossRef CAS PubMed .
  26. L. W. Y. Yeung, C. Dassuncao, S. Mabury, E. M. Sunderland, X. Zhang and R. Lohmann, Vertical Profiles, Sources, and Transport of PFASs in the Arctic Ocean, Environ. Sci. Technol., 2017, 51, 6735–6744 CrossRef CAS PubMed .
  27. R. Lohmann, E. Jurado, M. E. Pilson and J. Dachs, Oceanic deep water formation as a sink of persistent organic pollutants, Geophys. Res. Lett., 2006, 33(12) DOI:10.1029/2006gl025953 .
  28. C. J. Galbán-Malagón, N. Berrojalbiz, R. Gioia and J. Dachs, The “Degradative” and “Biological” Pumps Controls on the Atmospheric Deposition and Sequestration of Hexachlorocyclohexanes and Hexachlorobenzene in the North Atlantic and Arctic Oceans, Environ. Sci. Technol., 2013, 47, 7195–7203 CrossRef PubMed .
  29. C. Galbán-Malagón, N. Berrojalbiz, M.-J. Ojeda and J. Dachs, The oceanic biological pump modulates the atmospheric transport of persistent organic pollutants to the Arctic, Nat. Commun., 2012, 3, 862 CrossRef PubMed .
  30. L. Nizzetto, R. Gioia, J. Li, K. Borgå, F. Pomati, R. Bettinetti, J. Dachs and K. C. Jones, Biological Pump Control of the Fate and Distribution of Hydrophobic Organic Pollutants in Water and Plankton, Environ. Sci. Technol., 2012, 46, 3204–3211 CrossRef CAS PubMed .
  31. L. Morales, J. Dachs, M.-C. Fernández-Pinos, N. Berrojalbiz, C. Mompean, B. González-Gaya, B. Jiménez, A. Bode, M. Ábalos and E. Abad, Oceanic Sink and Biogeochemical Controls on the Accumulation of Polychlorinated Dibenzo-p-dioxins, Dibenzofurans, and Biphenyls in Plankton, Environ. Sci. Technol., 2015, 49, 13853–13861 CrossRef CAS PubMed .
  32. C. Galbán-Malagón, A. Cabrerizo, G. Caballero and J. Dachs, Atmospheric occurrence and deposition of hexachlorobenzene and hexachlorocyclohexanes in the Southern Ocean and Antarctic Peninsula, Atmos. Environ., 2013, 80, 41–49 CrossRef .
  33. C. J. Galbán-Malagón, S. Del Vento, A. Cabrerizo and J. Dachs, Factors affecting the atmospheric occurrence and deposition of polychlorinated biphenyls in the Southern Ocean, Atmos. Chem. Phys., 2013, 13, 12029–12041 CrossRef .
  34. N. Berrojalbiz, J. Dachs, M. J. Ojeda, M. C. Valle, J. Castro-Jiménez, J. Wollgast, M. Ghiani, G. Hanke and J. M. Zaldivar, Biogeochemical and physical controls on concentrations of polycyclic aromatic hydrocarbons in water and plankton of the Mediterranean and Black Seas, Global Biogeochem. Cycles, 2011, 25(4) DOI:10.1029/2010gb003775 .
  35. B. Boulanger, A. M. Peck, J. L. Schnoor and K. C. Hornbuckle, Mass Budget of Perfluorooctane Surfactants in Lake Ontario, Environ. Sci. Technol., 2005, 39, 74–79 CrossRef CAS PubMed .
  36. J. Huisman, N. N. P. Thi, D. M. Karl and B. Sommeijer, Reduced mixing generates oscillations and chaos in the oceanic deep chlorophyll maximum, Nature, 2006, 439, 322 CrossRef CAS PubMed .
  37. J. J. Cullen, The Deep Chlorophyll Maximum: Comparing Vertical Profiles of Chlorophyll a, Can. J. Fish. Aquat. Sci., 1982, 39, 791–803 CrossRef CAS .
  38. A. R. Longhurst and W. Glen Harrison, The biological pump: profiles of plankton production and consumption in the upper ocean, Prog. Oceanogr., 1989, 22, 47–123 CrossRef .
  39. M. G. Weinbauer, Ecology of prokaryotic viruses, FEMS Microbiol. Rev., 2004, 28, 127–181 CrossRef CAS PubMed .
  40. A. G. Paul, K. C. Jones and A. J. Sweetman, A First Global Production, Emission, and Environmental Inventory for Perfluorooctane Sulfonate, Environ. Sci. Technol., 2009, 43, 386–392 CrossRef CAS PubMed .
  41. R. van Zelm, M. A. J. Huijbregts, M. H. Russell, T. Jager and D. van de Meent, Modeling the environmental fate of perfluorooctanoate and its precursors from global fluorotelomer acrylate polymer use, Environ. Toxicol. Chem., 2008, 27, 2216 CrossRef CAS PubMed .
  42. B. Fernández-Castro, B. Mouriño-Carballido, V. M. Benítez-Barrios, P. Chouciño, E. Fraile-Nuez, R. Graña, M. Piedeleu and A. Rodríguez-Santana, Microstructure turbulence and diffusivity parameterization in the tropical and subtropical Atlantic, Pacific and Indian Oceans during the Malaspina 2010 expedition, Deep Sea Res., Part I, 2014, 94, 15–30 CrossRef .
  43. T. R. Osborn, Estimates of the Local Rate of Vertical Diffusion from Dissipation Measurements, J. Phys. Oceanogr., 1980, 10, 83–89 CrossRef .
  44. W. G. Large, J. C. McWilliams and S. C. Doney, Oceanic vertical mixing: A review and a model with a nonlocal boundary layer parameterization, Rev. Geophys., 1994, 32, 363–403 CrossRef .
  45. Y. Miyazawa, N. Yamashita, S. Taniyasu, E. Yamazaki, X. Guo, S. M. Varlamov and T. Miyama, Oceanic dispersion simulation of perfluoroalkyl substances in the Western North Pacific associated with the Great East Japan Earthquake of 2011, J. Oceanogr., 2014, 70, 535–547 CrossRef CAS .
  46. M. J. Costello, A. Cheung and N. De Hauwere, Surface area and the seabed area, volume, depth, slope, and topographic variation for the world's seas, oceans, and countries, Environ. Sci. Technol., 2010, 44, 8821–8828 CrossRef CAS PubMed .
  47. J. R. Ledwell, A. J. Watson and C. S. Law, Mixing of a tracer in the pycnocline, J. Geophys. Res., 1998, 103, 21499–21529 CrossRef .
  48. A. F. Waterhouse, J. A. MacKinnon, J. D. Nash, M. H. Alford, E. Kunze, H. L. Simmons, K. L. Polzin, L. C. St. Laurent, O. M. Sun, R. Pinkel, L. D. Talley, C. B. Whalen, T. N. Huussen, G. S. Carter, I. Fer, S. Waterman, A. C. Naveira Garabato, T. B. Sanford and C. M. Lee, Global Patterns of Diapycnal Mixing from Measurements of the Turbulent Dissipation Rate, J. Phys. Oceanogr., 2014, 44, 1854–1872 CrossRef .
  49. W. H. Munk, Abyssal recipes, Deep-Sea Res. Oceanogr. Abstr., 1966, 13, 707–730 CrossRef .
  50. N. Schmidt, V. Fauvelle, A. Ody, J. Castro-Jiménez, J. Jouanno, T. Changeux, T. Thibaut and R. Sempéré, The Amazon River – a major source of organic plastic additives to the tropical North Atlantic?, Environ. Sci. Technol., 2019, 53(13), 7513–7521 CrossRef CAS PubMed .
  51. D. A. Siegel, K. O. Buesseler, S. C. Doney, S. F. Sailley, M. J. Behrenfeld and P. W. Boyd, Global assessment of ocean carbon export by combining satellite observations and food-web models, Global Biogeochem. Cycles, 2014, 28, 181–196 CrossRef CAS .
  52. J. Dachs, R. Lohmann, W. A. Ockenden, L. Méjanelle, S. J. Eisenreich and K. C. Jones, Oceanic Biogeochemical Controls on Global Dynamics of Persistent Organic Pollutants, Environ. Sci. Technol., 2002, 36, 4229–4237 CrossRef CAS PubMed .
  53. E. Lipiatou, J.-C. Marty and A. Saliot, Sediment trap fluxes of polycyclic aromatic hydrocarbons in the Mediterranean Sea, Mar. Chem., 1993, 44, 43–54 CrossRef CAS .
  54. J. Dachs, J. M. Bayona, S. W. Fowler, J.-C. Miquel and J. Albaigés, Vertical fluxes of polycyclic aromatic hydrocarbons and organochlorine compounds in the western Alboran Sea (southwestern Mediterranean), Mar. Chem., 1996, 52, 75–86 CrossRef CAS .
  55. N. Berrojalbiz, S. Lacorte, A. Calbet, E. Saiz, C. Barata and J. Dachs, Accumulation and Cycling of Polycyclic Aromatic Hydrocarbons in Zooplankton, Environ. Sci. Technol., 2009, 43, 2295–2301 CrossRef CAS PubMed .
  56. S.-J. Royer, A. S. Mahajan, M. Galí, E. Saltzman and R. Simó, Small-scale variability patterns of DMS and phytoplankton in surface waters of the tropical and subtropical Atlantic, Indian, and Pacific Oceans, Geophys. Res. Lett., 2015, 42, 475–483 CrossRef .
  57. Z. Wang, M. MacLeod, I. T. Cousins, M. Scheringer and K. Hungerbühler, Using COSMOtherm to predict physicochemical properties of poly- and perfluorinated alkyl substances (PFASs), Environ. Chem., 2011, 8, 389 CrossRef CAS .
  58. C. A. Ng and K. Hungerbühler, Bioaccumulation of Perfluorinated Alkyl Acids: Observations and Models, Environ. Sci. Technol., 2014, 48, 4637–4648 CrossRef CAS PubMed .
  59. H. Frouin, N. Dangerfield, R. W. Macdonald, M. Galbraith, N. Crewe, P. Shaw, D. Mackas and P. S. Ross, Partitioning and bioaccumulation of PCBs and PBDEs in marine plankton from the Strait of Georgia, British Columbia, Canada, Prog. Oceanogr., 2013, 115, 65–75 CrossRef .
  60. M. Pućko, W. Walkusz, R. W. Macdonald, D. G. Barber, C. Fuchs and G. A. Stern, Importance of Arctic Zooplankton Seasonal Migrations for α-Hexachlorocyclohexane Bioaccumulation Dynamics, Environ. Sci. Technol., 2013, 47, 4155–4163 CrossRef PubMed .
  61. B. González-Gaya, A. Martínez-Varela, M. Vila-Costa, P. Casal, E. Cerro-Gálvez, N. Berrojalbiz, D. Lundin, M. Vidal, C. Mompeán, A. Bode, B. Jiménez and J. Dachs, Biodegradation as an important sink of aromatic hydrocarbons in the oceans, Nat. Geosci., 2019, 12, 119 CrossRef .
  62. W. A. Gebbink, A. Bignert and U. Berger, Perfluoroalkyl Acids (PFAAs) and Selected Precursors in the Baltic Sea Environment: Do Precursors Play a Role in Food Web Accumulation of PFAAs?, Environ. Sci. Technol., 2016, 50, 6354–6362 CrossRef CAS PubMed .
  63. T. M. Boudreau, C. J. Wilson, W. J. Cheong, P. K. Sibley, S. A. Mabury, D. C. G. Muir and K. R. Solomon, Response of the zooplankton community and environmental fate of perfluorooctane sulfonic acid in aquatic microcosms, Environ. Toxicol. Chem., 2003, 22, 2739 CrossRef CAS PubMed .
  64. M. L. Bates, S. M. Bengtson Nash, D. W. Hawker, E. C. Shaw and R. A. Cropp, The distribution of persistent organic pollutants in a trophically complex Antarctic ecosystem model, J. Mar. Syst., 2017, 170, 103–114 CrossRef .
  65. S. Honjo, S. J. Manganini, R. A. Krishfield and R. Francois, Particulate organic carbon fluxes to the ocean interior and factors controlling the biological pump: a synthesis of global sediment trap programs since 1983, Prog. Oceanogr., 2008, 76, 217–285 CrossRef .
  66. A. Sanchez-Vidal, M. Llorca, M. Farré, M. Canals, D. Barceló, P. Puig and A. Calafat, Delivery of unprecedented amounts of perfluoroalkyl substances towards the deep-sea, Sci. Total Environ., 2015, 526, 41–48 CrossRef CAS PubMed .
  67. S. Rahmstorf, Ocean circulation and climate during the past 120,000 years, Nature, 2002, 419, 207 CrossRef CAS PubMed .
  68. E. Yamazaki, N. Yamashita, S. Taniyasu, Y. Miyazawa, T. Gamo, H. Ge and K. Kannan, Emission, Dynamics and Transport of Perfluoroalkyl Substances from Land to Ocean by the Great East Japan Earthquake in 2011, Environ. Sci. Technol., 2015, 49, 11421–11428 CrossRef CAS PubMed .
  69. S. Taniyasu, N. Yamashita, E. Yamazaki, G. Petrick and K. Kannan, The environmental photolysis of perfluorooctanesulfonate, perfluorooctanoate, and related fluorochemicals, Chemosphere, 2013, 90, 1686–1692 CrossRef CAS PubMed .
  70. Z. Wang, I. T. Cousins and M. Scheringer, Comment on “The environmental photolysis of perfluorooctanesulfonate, perfluorooctanoate, and related fluorochemicals”, Chemosphere, 2015, 122, 301–303 CrossRef CAS PubMed .
  71. T. Frömel and T. P. Knepper, in Reviews of Environmental Contamination and Toxicology Volume 208: Perfluorinated alkylated substances, ed. P. De Voogt, Springer, New York, New York, NY, 2010, pp. 161–177 Search PubMed .
  72. X.-J. Zhang, T.-B. Lai and R. Y.-C. Kong, in Fluorous Chemistry, ed. I. T. Horváth, Springer Berlin Heidelberg, Berlin, Heidelberg, 2012, pp. 365–404 Search PubMed .
  73. J. S.-C. Liou, B. Szostek, C. M. DeRito and E. L. Madsen, Investigating the biodegradability of perfluorooctanoic acid, Chemosphere, 2010, 80, 176–183 CrossRef CAS PubMed .
  74. J. Liu and S. Mejia Avendaño, Microbial degradation of polyfluoroalkyl chemicals in the environment: a review, Environ. Int., 2013, 61, 98–114 CrossRef CAS PubMed .
  75. X. A. G. Morán, J. M. Gasol, M. C. Pernice, J.-F. Mangot, R. Massana, E. Lara, D. Vaqué and C. M. Duarte, Temperature regulation of marine heterotrophic prokaryotes increases latitudinally as a breach between bottom-up and top-down controls, Glob. Change Biol., 2017, 23, 3956–3964 CrossRef PubMed .
  76. E. Teira, V. Hernando-Morales, F. M. Cornejo-Castillo, L. Alonso-Sáez, H. Sarmento, J. Valencia-Vila, T. S. Catalá, M. Hernández-Ruiz, M. M. Varela, I. Ferrera, X. A. G. Morán and J. M. Gasol, Sample Dilution and Bacterial Community Composition Influence Empirical Leucine-to-Carbon Conversion Factors in Surface Waters of the World's Oceans, Appl. Environ. Microbiol., 2015, 81, 8224–8232 CrossRef CAS PubMed .
  77. E. Cerro-Gálvez, P. Casal, D. Lundin, B. Piña, J. Pinhassi, J. Dachs and M. Vila-Costa, Microbial responses to anthropogenic dissolved organic carbon in the Arctic and Antarctic coastal seawaters, Environ. Microbiol., 2019, 21, 1466–1481 CrossRef PubMed .

Footnote

Electronic supplementary information (ESI) available. See DOI: 10.1039/c9em00266a

This journal is © The Royal Society of Chemistry 2019