Trapping efficiencies of sediment traps from the deep Eastern North Atlantic:: the 230Th calibration

https://doi.org/10.1016/S0967-0645(00)00176-4Get rights and content

Abstract

Bottom-tethered sediment traps deployed in the deep eastern North Atlantic between 54°N 20°W and 33°N 20°W (L1, L2, L3), at the European continental margin at 49°N (OMEX) and off the Canary Islands (ESTOC) were investigated for the determination of 230Th trapping efficiencies. The ratios of 230Th flux measured in the traps (Fa) to the expected 230Th flux from the production rate of 230Th in the overlying water column (Fp) ranged between 0.09 and 1.26. For the traps with deployment periods >300 days the interannual variation of Fa/Fp ratios (different years but same location and water depth) were up to 10%, suggesting that the average 230Th flux to the sediment traps did not vary significantly. The influence of lateral advection on the 230Th flux was taken into account either by applying a mass balance of 230Th and 231Pa or by assuming a constant removal rate of 230Th from the water column, an assumption based on similar 230Th concentration-depth profiles observed at most locations investigated. 230Th trapping efficiencies were between 9 and 143%, showing a trend of increasing efficiencies with increasing water depth. No relation was found between current velocities and 230Th trapping efficiencies. Our investigations suggest that the observation of constant or even increasing particle flux rates with increasing water depths in several sediment trap arrays investigated may be a result of sediment trap biases. The correction for the trapping biases is important for the understanding of the regional differences in the particle flux in the eastern North Atlantic.

Introduction

One of the major scientific goals of the Joint Global Ocean Flux Studies (JGOFS) is “to determine and understand on a global scale the processes controlling the time-varying fluxes of carbon and associated elements in the ocean” (SCOR, 1990). In order to meet these goals measurements of vertical particle fluxes are a central issue in JGOFS. These flux measurements rely mainly on sediment traps that are deployed in the ocean waters (Knauer and Asper, 1989). They allow time-dependent quantification of particle fluxes and collection of settling material for further compositional analyses.

Several sediment trap studies have shown a clear relation between bioproductivity in surface waters and deep-water particle flux (Bacon et al., 1985; Wefer, 1989, Berger and Wefer, 1990; Haake et al., 1993). High-latitude areas are characterised by a distinct seasonality in the particle flux, with high fluxes during spring blooms, whereas at low latitudes seasonal flux patterns are less pronounced. In upwelling areas variations in particle flux are linked to the dynamics of the upwelling systems (Wefer and Fischer, 1993). A comparison of interannual variability of particle flux and particle composition in the North Atlantic showed that the magnitude of the deep particle flux for two sites about 100 km apart was in the same range, although not all flux events were recorded at both sites at the same time (Newton et al., 1994). Gradients of deep-water mass flux were observed in the east Atlantic from 48° to 19°N (Jickells et al., 1996). On this transect the total biogenic flux increased by about a factor of two from south to north, the particulate organic carbon flux (POC) by about factor of three.

There is no absolute standard for flux measurements because the vertical particle flux in the ocean is a result of complex processes. Therefore, all the attempts to understand the temporal and spatial variability of the particle flux are based on the assumption that the sediment traps estimate the vertical particle flux reasonably well. Field and laboratory investigations showed, however, that various types of errors — e.g., hydrodynamic effects (water flow, advective flow, tilt of trap), solution of trapped material, swimmers — may have potential to influence the accuracy of flux measurements (Baker et al., 1988; Lee et al., 1988; Gust et al (1992), Gust et al (1994); Gardner, 1996; Gardner and Zhang, 1997). Indirect evidence of possible biases in particle flux measurements may be deduced from current measurements. Many JGOFS-related particle flux studies (e.g., Jickells et al., 1996; Neuer et al., 1997; Newton et al., 1994) assume that currents ⩽10–12 cm s−1 do not affect the measured particle flux, an assumption mainly based on the investigations of Baker et al. (1988) who suggested no significant trapping biases for current speeds ⩽12 cm s−1. Gardner (1980) showed, however, that undertrapping can occur when current speeds are about 4–5 cm s−1.

More direct evidence for trapping efficiencies of sediment traps used in particle flux investigations comes from natural radionuclides (Bacon, 1996). The natural radionuclides 234Th (T1/2=24 d), 230Th (T1/2=75.2 kyr), 231Pa (T1/2=32.5 kyr) and 210Pb (T1/2=22.3 yr) are produced at a known rate by radioactive decay in the ocean, and because they are particle reactive their removal from the water column is closely linked to the particle flux. By comparing the water column distribution of the radionuclides with their fluxes into sediment traps, information on trapping efficiencies can be obtained. For instance, measurements of 234Th in the upper water column and in sediment traps from various oceanic regions showed that sediment traps deployed in shallow waters (⩽200 m water depth) may under-and/or over-estimate the true vertical particle flux up to a factor 10 (Buesseler, 1991). It is this high uncertainty of shallow-deployed sediment traps that made 234Th measurements a core parameter in many recent JGOFS-related process studies (Buesseler et al., 1995; Bacon et al., 1996; Murray et al., 1996).

While much attention has been given to the trapping efficiency of shallow traps, little work has been done on investigating possible trapping biases of traps deployed in the deep water column. Most attempts to calibrate these traps were based on a comparison of the 230Th flux measured in the sediment traps with the flux of 230Th expected from the water column. This model assumes a steady state between 230Th production and removal of 230Th by sinking particles. Applying this model Brewer et al. (1980) estimated apparent trapping efficiencies of traps moored in the deep Sargasso Sea and off Bermuda to be between 20 and 70%.

Most open-ocean 230Th water column profiles show a linear increase of total (dissolved plus particulate) 230Th concentrations with increasing water depth. This distribution can be described by a reversible scavenging process (Bacon and Anderson, 1982): dissolved 230Th is adsorbed on suspended particles and aggregation of suspended particles to fast sinking particles removes 230Th from the water column. Adsorption and aggregation are linked with their reverse processes, i.e. desorption of 230Th from suspended particles and disaggregation of fast sinking particles. The resulting residence time of 230Th in the water column ranges between 5 and 40 yr (Anderson et al., 1983).

Lateral advected water masses and their mixing can change the water column 230Th distribution when these water masses have significantly different 230Th concentrations. Such differences may be the result of up and/or down-welling of water masses. For instance in the South Atlantic south of the Polar Front, upwelling of Circumpolar Deep Water significantly increased the 230Th concentrations in the Weddell Sea (Rutgers van der Loeff and Berger, 1993) and an export of 230Th and 231Pa to the Antarctic Circumpolar Current was suggested (Walter et al., 2000). In the Eurasian basins of the Arctic Ocean, in contrast, the intermediate water masses were renewed on time scales of about 50 yr with water masses derived from the Norwegian-Greenland Sea. Since both water masses had comparable 230Th concentrations, the effects on the 230Th distribution in the Arctic basins are negligible (Scholten et al., 1995).

Lateral transport of 230Th and 231Pa in the water column also can occur from the open ocean to the ocean margins due to intensified scavenging at continental margins (Anderson et al., 1983). This boundary scavenging effect depends on the relation between scavenging residence time of the radionuclides and the ventilation time of water masses (Bacon, 1988; Walter et al., 1999). In the Pacific Ocean the ventilation time is long (about 500 yr) relative to the scavenging residence time of 230Th (about 20 yr) and 231Pa (about 150 yr), resulting in a preferential removal of 231Pa at ocean margins. Advection and scavenging of 230Th in the world oceans have been investigated recently using a General Circulation Model (Henderson et al., 1999). For the eastern North Atlantic the model results suggested no import or export of 230Th from/to remote areas.

Bacon et al. (1985) quantified the boundary scavenging effect of 230Th and 231Pa in the western North Atlantic and, correcting for it, they estimated the trapping efficiency of a Paraflux sediment trap in the Sargasso Sea (3200 m water depth) to be 105±17%. Further investigations of trapping biases of bottom-tethered sediment traps deployed in various ocean regions indicate reliable trapping efficiencies of sediment traps deployed in water depths >1200 m, whereas in shallower water depth efficiencies tend to be lower (Yu et al., 2001).

As part of the German JGOFS project time-series traps were deployed in the deep eastern North Atlantic Ocean at three locations (L1, L2, L3) between 54°N 20°W and 33°N 20°W and 100 km north of the Canary Islands within the ESTOC program (European Station for time series in the Ocean, Canary Islands) (Fig. 1). Additional sediment traps were deployed at Goban Spur in the Celtic Sea near the shelf break (OMEX; Ocean Margin Exchange Project). According to Longhurst (1993), the above-mentioned deployment areas include subtropical environments with a permanently stratified region (L1, ESTOC) and transition areas with deep winter mixed layers and a marked seasonality in the particle flux (L2, L3, OMEX). At these locations bottom-tethered sediment traps were deployed in water depths of between 500 and 4000 m. Based on measurements of 230Th and 231Pa we discuss the trapping efficiency of the deployed traps, and it will be shown that trapping biases are primarily to be expected in sediment traps deployed in shallow water depths.

Section snippets

Samples

The sediment traps investigated consisted of a fibre glass cone with an aperture of 0.5 m2 and a sample changer that can collect 21 subsamples. Further details on the trap design (“Kiel” type) are given in Kremling et al. (1996). Details of the locations and of the water depths of moored traps together with collection dates are given in Table 1. Prior to analysis zooplankton swimmers were removed from samples under the binocular by a Teflon tweezer (Kuss and Kremling, 1999). The swimmers were

Results

Results of alpha measurements are only given in this study in cases where no TIMS measurements are available (details of the results obtained by alpha measurements will be presented elsewhere). The 232Th flux varied from 6.51×10−4 to 4.16×10−2 dpm m−2 d−1, and the flux of 230Thex from 3.98×10−3 to 0.291 dpm m−2 d−1 (Table 2). The 231Paex/230Thex ratios varied between 0.029 and 0.068. The ratio of these isotopes in the water column was between 0.149 and 0.293 (Table 3), which is in the range

Discussion

Estimations of trapping efficiency of moored sediment traps by means of 230Th are based on a mass balance between the production rate of 230Th in the water column and the 230Th flux measured in the sediment traps.

The 230Th trapping efficiency (E) is defined asE=Fa/V=Fa/(Fp−H),where Fa is the average 230Th flux measured in a sediment trap, V the vertical flux of 230Th, Fp the expected vertical flux of 230Th from the production rate of 230Th in the water column, and H the horizontal 230Th flux in

Conclusions

Trapping efficiencies of bottom-tethered sediment traps deployed in the eastern North Atlantic were estimated on the basis of the distribution of 230Th and 231Pa in the water column and their average fluxes into sediment traps. 230Th trapping efficiencies were low (<40%) for traps deployed in shallow waters (⩽1000 m), whereas the deep traps seem to sample the 230Th and the particle flux reasonably well. No tight relation between trapping efficiency and current velocities was found. The

Acknowledgements

We express our gratitude to the German JGOFS community for their support and discussions from which our work benefitted. Especially we like to thank R. Botz, J. C. Duinker, K. Kremling, S. Neuer, D. E. Schulz-Bull, and B. Zeitzschel. We are grateful to K. Schmikale and M. Arp for assistance in the laboratory. We also express our gratitude to the captains and crew of F.S. Poseidon and F.S. Meteor for assistance at sea. We are grateful to J.M. Murray and an anonymous reviewer for their

References (59)

  • G. Gust et al.

    Particle fluxes and moving fluidsexperience from synchronus trap collections in the Sargasso Sea

    Deep-Sea Research I

    (1992)
  • G. Gust et al.

    Mooring line motions and sediment trap hydromechanicsin situ intercomparison of three common deployment designs

    Deep-Sea Research I

    (1994)
  • B. Haake et al.

    Seasonality and interannual variability of particle fluxes to the deep Arabian Sea

    Deep-Sea Research I

    (1993)
  • K. Lochte et al.

    Plankton succession and carbon cycling at 47°N 20°W during the JGOFS North Atlantic Bloom Experiment

    Deep-Sea Research II

    (1993)
  • A. Longhurst

    Seasonal cooling and blooming in tropical oceans

    Deep-Sea Research I

    (1993)
  • J.M. Murray et al.

    Export flux of particulate organic carbon from the central equatorial Pacific determined using a combined drifting trap- 234Th approach

    Deep-Sea Research II

    (1996)
  • S. Neuer et al.

    Deep water particle flux in the Canary Island regionseasonal trends in relation to long-term satellite derived pigment data and lateral sources

    Deep-Sea Research I

    (1997)
  • P.P. Newton et al.

    Temporal and spatial variability of biogenic particle fluxes during the JGOFS northeast Atlantic process studies at 47°N 20°W

    Deep-Sea Research I

    (1994)
  • M.M. Rutgers van der Loeff et al.

    Scavenging of 230Th and 231Pa near the Antarctic Polar Front in the South Atlantic

    Deep-Sea Reserach I

    (1993)
  • J.C. Scholten et al.

    230Thex flux into Norwegian-Greenland Sea sedimentsevidence for lateral sediment transport during the past 300,000 years

    Earth and Planetary Science Letters

    (1994)
  • J.C. Scholten et al.

    Distribution of 230Th and 231Pa in the water column in relation to the ventilation of the deep Arctic basins

    Deep-Sea Research II

    (1995)
  • T.J. Shaw et al.

    Scavenging of ex234Th, exTh and exPb by particulate matter in the water column of the California continental margin

    Deep-Sea Research II

    (1998)
  • R.M. Sherrell et al.

    Temporal variability of suspended mass and composition in the Northeast Pacific water columnrelationship to sinking flux and lateral advection

    Deep-Sea Research II

    (1998)
  • L. Van Camp et al.

    Upwelling and boundary circulation off Northwest Africa as depicted by infrared and visible satellite observation

    Progress in Oceanography

    (1991)
  • S. Vogler et al.

    230Th in the eastern North Atlanticthe importance of water mass ventilation in the balance of 230Th

    Earth and Planetary Science Letters

    (1998)
  • H.J. Walter et al.

    Enhanved scavenging of 231Th in the South Atlantic south of the Polar FrontImplications for the use of the 231P/230Th ratio as a paleoproductivity proxy

    Earth Planetary Science Letters

    (1997)
  • H.J. Walter et al.

    Reduced scavenging of 230Th in the Weddell Sea

    Implications for paleoceanographic reconstruction in the South Atlantic. Deep-See Research I

    (2000)
  • G. Wefer et al.

    Seasonal patterns of vertical particle flux in the equatorial and coastal upwelling areas of the eastern Atlantic

    Deep-Sea Research I

    (1993)
  • R.F. Anderson et al.

    Determination of natural actinides and plutonium in marine particulate matter

    Analytical Chemistry

    (1982)
  • Cited by (138)

    • Diatom and coccolithophore species fluxes in the Subtropical Frontal Zone, east of New Zealand

      2021, Deep-Sea Research Part I: Oceanographic Research Papers
      Citation Excerpt :

      Issues with hydrodynamic biases, zooplankton ‘swimmer’ contamination and flux component dissolution have been recognised as potential sources of error associated with the sediment trap method in general (Buesseler et al., 2007; Boyd and Trull 2007). Sediment traps on bottom-tethered, fixed moorings in the mesopelagic zone above ~1000–1200 m depth have been demonstrated to under-estimate downward fluxes by 60–90% based on radio-nuclide inventories, even in low current (<10 m s−1) environments (Scholten et al., 2001; Yu et al., 2001). Even small variations in trap tilt angle have also been shown to significantly reduce trap collection efficiency of bulk flux constituents, such as POC, by 30–50% (Siegel and Deuser 1997; Chiswell and Nodder 2015).

    • Marine radioactivity analysis

      2020, Handbook of Radioactivity Analysis: Volume 2: Radioanalytical Applications
    View all citing articles on Scopus
    View full text