Download (7M) - NERC Open Research Archive

By Martha Rodriguez,2014-02-16 06:56
21 views 0
Download (7M) - NERC Open Research Archive


    230Geomorphological and geochemical evidence (sediment Th anomalies)

    for cross-equatorial currents in the central Pacific

    12Neil C. Mitchell and John M. Huthnance

    1School of Earth, Atmospheric and Planetary Sciences, University of Manchester, Manchester M13 9PL, UK

    2National Oceanography Centre, Joseph Proudman building, 6 Brownlow Street, Liverpool L3 5DA, UK

For correspondence:

    Keywords: sediment focusing, bottom Ekman layer, bottom boundary layer, sediment furrows


     Shallow broad elongated sediment depressions and ridges are revealed in multibeam echo-sounder data collected over the carbonate ooze in the central equatorial Pacific. These features, otherwise called "furrows", have orientations that appear locally distorted by seabed topography as expected of contour-trending currents but at regional scale typically cross contours at high angles. In places, complex patterns suggest that formative currents have a strong time-varying component. From direction indicators, the movement of bottom waters is north to south on average, though with some movement locally south to north. There is a modest 18? average change in orientation crossing from north to south of the equator, with features to the south oriented clockwise of those to the north. This is as expected for a partly developed bottom Ekman layer, with currents in the layer deflected by the Coriolis effect with opposing senses either side of the equator. The features are less prominent on and immediately south of the

    230equator. We evaluated these observations along with reported Th

    accumulation rates in sediment cores., which are curiously enhanced along the


    equator, an observation that has been previously interpreted as suggesting

    230Th bound to particles to the equator. transport of

     Limited current meter and other data and physical oceanographic models help to explain these observations. Data from current meters 1? north of the

    equator show a highly asymmetric mesoscale eddy motion here, aligned with the furrows. Phase relationships between near-bed and upper ocean currents suggest an indirect coupling of upper-ocean eddies with the lower ocean. The bottom Ekman layer is predicted theoretically to thicken towards the equator.

    230The resulting reduced bed shear stress may explain the Th deposition and

    more weakly developed furrows at the equator. Given evidence that equatorial

    2303accumulation rates of Th and extraterrestrial He both fluctuated over the

    Late Pleistocene, we explore how the ideas presented here could help to explain how the geochemical anomalies relate to physical oceanographic processes.


     Multibeam echo-sounder data reveal abundant elongate sedimentary features, which are evidence for a widespread benthic current in the equatorial Pacific. The current is at a high angle to the expected geostrophic flow and, along with Ekman layer theory, may help to explain a previously reported enhanced

    2303deposition of Th and He along the equator.

     In the central Pacific, bottom water temperatures increase eastwards implying a general eastwards movement of ocean waters (Mantyla, 1975), which

    14is predicted by geostrophic calculations (Reid, 1997). On the other hand, C

    ages of water generally increase northwards in the Pacific Ocean (Key et al., 2004; Matsumoto, 2007) though they can be modelled with a bottom water movement from a southern ocean source flowing as a western boundary current northwards and then eastwards into the central Pacific (Roussenov et al., 2004). These and geochemical tracers are scalar properties and, as the numerical modelling of Roussenov et al. (2004) illustrates, they can be explained by complex flow paths and their values are complicated by eddy diffusion, which is poorly characterized in the deep ocean. Vector measurements (with current meters and floats) recording the full time-varying movements of the ocean are preferred but are limited in the abyssal central Pacific. For understanding


    particle transport, as entrainment and deposition involve threshold shear stresses, the full temporal history of water movement and in particular peak currents are important rather than merely the mean current.

     The sediments of the central and eastern equatorial Pacific have been studied extensively over the past few decades from samples recovered by scientific drilling and sediment coring because of potential links between their properties and climate. For example, equatorial upwelling is linked to wind stress driving surface Ekman currents and the upwelling in turn may affect pelagic productivity, implying a link, albeit complex, between wind strength and the production rate of particles (e.g., Mayer et al., 1992a). But such work is potentially compromised if particles move significantly with abyssal currents before depositing. Accumulation rates of particles are related to the shear stress near the seabed amongst other factors (e.g., McCave and Swift (1976)). Therefore, where temporal variations in depositional fluxes in

    paleoceanographic studies have been used to investigate past chemical properties of the ocean or upper-ocean properties (Beaufort et al., 2001; Chugh and Bhattacharji; Griffith et al., 2010; Lyle, 2003; Mitchell et al., 2003; Pälike et al., 2012), those fluxes may also have been affected by changing strength of abyssal currents. Spatial variations in the properties of deposited particles (Boltovskoy, 1991; Murray et al., 2000) may also not necessarily relate solely to variations in their production or input to the ocean if there is significant movement.

     Geomorphologic features on the ocean floor can provide clues to water and sediment movements, which may in turn help to guide the design of future oceanographic fieldwork as well as indicate where uncertainties in paleoceanographic studies may lie. The sedimentary features described here are elongated ridges, which are essentially sediment drifts, and depressions, which are similar to features termed elsewhere as furrows (Dyer, 1970). Although not an ideal descriptor, we use the term "furrow" here for consistency. Furrows in abyssal environments are typically 1-20 m deep sedimentary troughs extending parallel to the flow direction for several kilometres (Hollister et al., 1974), produced by helical secondary flow near the bed (Flood, 1983; Lonsdale and Spiess, 1977; Viekman et al., 1992). Examples have been reported in calcareous pelagic sediment, such near the Samoan Passage, where they were found to be 1


    m deep and spaced at 30 m (Lonsdale, 1981; Lonsdale and Spiess, 1977). Lonsdale and co-workers (Lonsdale, 1976; Lonsdale, 1977; Lonsdale, 1981; Lonsdale and Malfait, 1974) have characterized near-bottom flow patterns in the southwest and east Pacific using these features. Both steady and oscillating currents can produce furrows (Flood, 1983), which indicate the general orientation of flow.

     The earlier studies of these low-relief features relied on sidescan sonar data collected from instruments close to the bed to generate acoustic shadows and strong acoustic backscattering contrasts. The early 16-beam SeaBeam multibeam sonars were noisy (de Moustier and Kleinrock, 1986) so only the largest erosional features could be interpreted from their data (Shipley et al., 1985). Over the last decade or two, however, the multibeam sonar technology has improved significantly. The data acquired with them have better noise characteristics and effective spatial resolution. The context provided by their wider swaths helps the interpreter discriminate these current-generated features from other elongated features, such as abyssal hills.

     The furrows described here are generally more widely spaced than those mentioned in the earlier publications. This difference may merely be an artifact of differing sonar resolution effectively filtering a continuum of bedform scale (the multibeam data shown here have a spatial resolution of ~112 m whereas the Deep Tow data used by Lonsdale and others had much higher resolution). Indeed high-resolution sonar images of such features elsewhere suggest a range of spacings can occur in practice (Cochonat et al., 1989). Alternatively, this difference might have a fluid dynamical origin. Current measurements in Lake Superior showed the secondary circulation extending to a height similar to the furrow spacing (Viekman et al., 1992). The difference may therefore relate to different boundary layer thicknesses, though we have insufficient data to confirm this here. These features are also morphologically similar to shallow gullies observed in upper continental slope settings produced by density currents (Izumi, 2004; Puig et al., 2008). Those off southern California, for example, have spacings of 100-500 m (Field et al., 1999; Mitchell and Huthnance, 2007), overlapping with spacings of the furrows shown here.


     Sediment geochemical data have also been interpreted as evidence for transport of particles by bottom currents to and their deposition on the Pacific equator (Marcantonio et al., 1996), and attempts have been made to use such data to correct depositional fluxes for particle redistribution (Francois et al.,

    3He and 2004). We re-examine those arguments in the light of the new data. 230Th accumulation rates in equatorial Pacific sediments in places deviate from

    3their introduction to the water column (He from interplanetary particle input to

    230234the ocean (Marcantonio et al., 1996) and Th from U decay (Higgins et al.,

    1999; Marcantonio et al., 2001)). Thorium is strongly particle-reactive in water, 230so Th produced by radioactive decay is readily absorbed onto falling particles. Consequently, it has an average residence time in the oceans of only 20 years (Henderson and Anderson, 2003). Uranium, in contrast, is well mixed and

    230production of Th from radioactive decay is predictable from water depth and

    230the uranium concentration. The ratio of accumulation rates of Th to the

    230expected production of Th in the immediately overlying water column then

    suggests whether sediment has been advected into or away from a site. Further measurements of this ratio (“Y”) termed the "focusing factor" compiled by

    Kienast et al. (2007), which are shown in Figure 1a, confirm this apparent anomalous accumulation of particles along the equator.

    230 This simple interpretation of the Th data has been controversial,

    however. Some authors (Broecker, 2008; Lyle et al., 2005; Thomas et al., 2000) have argued that focusing factors are enhanced at the equator because of enhanced scavenging by falling particles in this high pelagic productivity region. Removal of thorium from the equatorial ocean is then compensated by spreading of thorium by eddy diffusion along isopycnals from outside the equatorial region. Siddall et al. (2008), who described this as a particle flux effect (PFE), extended a global circulation model to study it and argued that it was insufficient to explain the magnitude of observed equatorial focusing factors. Thorium is preferentially absorbed onto finer lithogenic particles (in particular, clays), which have larger effective surface area for a given mass (Geibert and Usbeck, 2004; Kretschmer et al., 2010; McGee et al., 2010; Roy-Barman et al., 2005). Their smaller terminal velocities allow them to be more easily advected laterally than larger particles, so varied focusing factors may partly reflect varied transport of fines. For a


    discussion of other potential issues with this method, see Lyle et al. (2007) and Francois et al. (2007). We return to the particle flux and particle size effects in the discussion.

     In the present study, evidence for equatorial bottom currents is examined in multibeam echo-sounder and chirp sediment profiler records collected during cruise AMAT03RR of the RV Revelle (Dubois and Mitchell, 2012; Tominaga et al.,

    2011). Further information on the flow is provided by limited current meter datasets. The discussion examines the origins of the geomorphologic features,

    230Th the nature of time-varying currents crossing the equator here, the elevated anomalies reported previously (Kienast et al., 2007) and theoretical reasons why sediment accumulation may be enhanced along the equator, as well as general consequences for paleoceanographic studies.


    Geophysical data

     Bathymetry data were collected on Revelle using a Kongsberg-Simrad

    EM120 multibeam echo-sounder during site survey cruise AMAT03RR for Integrated Ocean Drilling Program (IODP) expedition 319/320 (Pälike et al., 2010). Data were acquired in small grid surveys over each IODP site as well as in transit, while simultaneously running seismic reflection, chirp sediment profiler and other scientific instruments.

     The EM120 multibeam sonar on Revelle has 191 beams spread over a

    150? sector. Resolution fore-aft is dictated by the 1? transmit acoustic beam

    widths (~70 m in 4000 m of water at nadir, widening with slant range away from nadir). The system uses the split-beam method of bottom echo detection, whereby the time of the echo arriving at the beam centre is determined from a series of differences of acoustic phase detected between two sections of the receive transducers (Hammerstad et al., 1991). Resolution athwartships therefore depends on the time window and method used for that determination, but is probably similar to the fore-aft resolution by inspection of the resulting bathymetry data. Because noise is reduced by filtering the data, resolution also in practice depends on the noise characteristics of the integrated system including motion sensors and acoustic noise arising from sea conditions (Schmitt


    et al., 2008). Data collected during trials with the ship stationary were found to have a standard deviation of less than 0.2% of water depth (<8 m in 4000 m depth) across the swath to 60? off-nadir (de Moustier, 2001).

     The sounding data were processed using standard procedures within the MB-System software (Caress and Chayes, 1996), including manually removing acoustic outliers. The data were binned in 0.001? cells (a step that reduces noise

    somewhat by averaging ~2-3 values in each cell) and gridded at 0.0005?

    resolution using a surface-fitting program (Smith and Wessel, 1990). The effective spatial resolution of the following images is therefore ~112 m. The shaded relief images in Figures 2-4 illustrate the quality of the data. Artifacts include enhanced variability in the outer swath produced by noise probably from the water column multiple of the transmitted pulse, some across-track striping caused by noise from the vessel passing through waves and minor remaining individual sounding outliers. An along-track striping can also be observed at the swath nadir (such as in the centre of Figure 3b), where the system uses a different seabed echo detection method (Mitchell, 1996). These artifacts are all readily identifiable and can be avoided in interpretation. If not acoustically isolated, interference from other sonars transmitting with different cycle periods might lead to artifacts trending obliquely to the swath, but in areas of flat seabed no such artifacts are evident in these data.

     The data were analyzed with the GeoMapApp software

    ( along with other wide-swath multibeam sonar data that

    have been incorporated within the Global Multi-Resolution Topography Synthesis (Ryan et al., 2009). Different artificial sun illumination directions were used to assess trends, using low inclination illumination to enhance subtle features not so easily interpreted in contour maps (Edwards et al., 1984). Such shading calculations typically involve a directional derivative of the topography, which enhances small features relative to large features if appropriately oriented. The shaded-relief imagery in Figures 2-4 was produced with the GMT software system (Wessel and Smith, 1991) with the illumination directions given in the figure captions. Those directions were chosen to show best the seabed features we interpret here. Readers can further study these data by viewing them when available from the National Geophysical Data Center (


    Contour maps of Figures 2-4 are also provided in an electronic supplement to this paper (link). Although contours are noisy from across- and along-track artifacts mentioned above, deviations of contours in the maps can be used to assess relief of the larger furrows.

    {Note to editors/typesetters: text "when available" can be removed if the data have been uploaded to NGDC by the time of publication}

Current meter data and tidal models

     No current meter records are available from the centre of the region studied here, but data from adjacent sites provide some sense of the character of the currents. Individual near-bed current meter measurements are shown as point clouds in Figure 5 for the sites marked MS, MC and MCD in Figure 1a (Manganese Nodule Project (MANOP) sites S, C and CD, respectively). The data in (a), (b), (d) and (e) of Figure 5 were provided by the Buoys Group of the Oregon State University (,

    whereas the record in (c) was reproduced from Mayer (1981). (The magnetic compass data from the latter site may have been biased (Gardner et al., 1984) so only the dispersion of directions and speed can be interpreted from this record.) Figure 5f shows a time-series of current speed recorded at site MANOP Site C1 (20 mab) to illustrate the influence of tides in a relatively quiescent part of the record. Measurement durations, given in Table 1, vary over 7-14 months. Progressive vector diagrams (PVDs) computed from the MANOP C1, C2 and C3 data are shown in Figure 6c (solid circles are shown at weekly intervals). A local bathymetry map and the temperature records from C1 are shown in Figures 6a and 6b, respectively. The PVDs can be contrasted with diagrams for the more northerly MANOP-S sites shown in Figure 7a.

     Further current measurements have been made north of the Clipperton Fracture Zone (Morgan et al. (1999) reported flow data compiled by Demidova et al. (1996)). However, the plots in Figure 6c illustrate that measurements made over much less than a year are unlikely to report the flow properties accurately given the variations of various time-scales including greater-than-annual periods affecting bottom waters here, so we have shown only one of those recordings in Figure 1a (site B, 197 days, 30 mab). The mean current of each set of current


    meter data is shown by vectors in Figure 1a (red and purple for >100 mab and <100 mab, respectively) and is reproduced in Table 1. The vector “MM” represents the mean current at MANOP Site M (200 mab, 13 months) from data in Lyle et al. (1985).

     Theoretical estimates of the barotropic tidal currents were made at the locations shown by cross symbols in Figure 1a and at the MANOP C1 and CD current meter sites. Calculations were carried out using the Oregon State University TPX06.2 global tide model (Egbert and Erofeeva, 2002) using

    TMMatlab Tide Model Driver software. Figure 8 shows scatter plots of the results, each calculated for one lunar month from 1/1/1992 with all main tidal constituents (M2, S2, N2, K2, K1, O1, P1, Q1, MV and MM).

Geochemical data

     230 TheTh accumulation anomaly or focusing factor Y is commonly defined as

    (Suman and Bacon, 1989):

    r22300éùThrdròexbëûr1Y= (1), bH(t-t)21

    2302300where [Th] is the decay-corrected concentration of Th in sediments ex,0

    corrected also for in-situ production, is the sediment dry bulk density and is b

    230the production rate of Th in the overlying water column of depth H. The

    calculation is carried out over the sediment depth range r to r corresponding to 12

    the sediment age range t to t. 12

     The colour-filled circles in Figures 1a and 1b (with scales to right) show the Y-

    values that have been reported in the literature for the Holocene and Last Glacial Maximum (LGM), respectively. These values are from a compilation of Kienast et al. (2007, their Table 4), supplemented with five values for the LGM calculated from data of McGee et al. (2007). Figure 9a shows the Y-values for the region

    109?W to 145?W plotted versus latitude (solid and open circles represent Holocene and LGM values, respectively). Also shown in Figure 9a, large grey-filled and open circles connected by solid lines are average Y-values (Holocene

    and LGM, respectively) each computed over 1? of latitude centred on 0? and 1?N.

     As explored by Kienast et al. (2007), uncertainties in individual Y-values are

    commonly dominated by dating uncertainties and can be difficult to quantify.


    Skinner and Shackleton (2005) documented a 2 ky lag of the Marine Isotope Stage 1/2 boundary in a Pacific core compared with the glacioeustatic (global) curve. This error would bias (increase) Y-values by 15% and may largely explain

    why the LGM values, which have not been corrected for this effect (Kienast, pers. comm., 2013), are larger than the Holocene values in Figure 9a. Kienast et al. (2007) estimated a further 1.5 ky uncertainty could arise from difficulty in identifying the Stage 2/3 boundary. McGee et al. (2007) also estimated 1

    uncertainties in Holocene Y-values by combining uncertainties in sediment age

    and density. For each of these studies (Kienast et al., 2007; McGee et al., 2007), we computed the mean ratio of their uncertainty to value of Y in order to

    construct the "typical" uncertainty bars shown in the left of Figure 9a for Y=1 and


     A further uncertainty (of interpretation rather than values) could arise from local transport of sediments, for example, because of a tendency for sediment to infill basins between abyssal hills (Dubois and Mitchell, 2012; Laguros and Shipley, 1989; Mitchell, 1995; Mitchell et al., 1998; Tominaga et al., 2011) or be removed from basement highs or areas of scour around outcrops (Mitchell, 1995; Mitchell, 1998b). To assess this effect, the core locations (Kienast et al., 2007; McGee et al., 2007) were read into the Geomapapp software so that the physiography of each site could be evaluated using either multibeam bathymetry data if available (Ryan et al., 2009) or otherwise satellite gravity-derived bathymetry capable of resolving larger features (Smith and Sandwell, 1997). The following core site names are from Kienast et al (2007) and McGee et al. (2007). Two sites were found to be close to seamounts of >1000 m elevation (TT013/MC34 and TT013/MC19) and one adjacent to a 600 m hill

    (TT013/MC69). Three sites lie on or near the edge of an abyssal hill where winnowing might be more pronounced (MANOP/B18, ODP852A and ODP853B). The Y-values of these sites are marked in Figure 9a with cross-symbols. These potentially biased values appear to have no particular effect on the overall trend, however; the smaller circles connected by dashed lines are 1?-averages after

    removing these values. Because of potential enhanced scavenging associated with hydrothermal activity (Singh et al., 2013) as well as these effects of topography, the sites closer to the East Pacific Rise were also excluded from the

Report this document

For any questions or suggestions please email