Sediment focusing in the Panama Basin, Eastern Equatorial Pacific Ocean

Abstract

Age-model derived sediment mass accumulation rates (MARs) are consistently higher than ²³⁰Th-normalized MARs in the Equatorial Pacific Ocean during the past 25 ka. The offset, being highest in the Panama Basin, suggests a significant role for deep-sea sediment redistribution (i.e., sediment focusing) in this region. Here, we test the hypothesis that downslope transport of sediments from topographically high regions that surround the Panama Basin is the cause of higher-than-expected xs²³⁰Th inventories over the past 25 ka in the deeper parts of the basin. We find little difference in xs²³⁰Th inventories between the highest and lowest reaches of the basin. Furthermore, there is no correlation between xs²³⁰Th-derived sediment focusing factors and water depth which suggests that the topographic highs do not serve as a source of xs²³⁰Th. A spatial analysis suggests that there may be an enhanced scavenging effect on xs²³⁰Th concentrations in sediment closest to the equator where productivity is the highest, although further data is necessary to corroborate this. At the equator xs²³⁰Th-derived focusing factors are high and range from about 1 to 5 during the Holocene and about 1 to 11 during the last glacial. In contrast, non-equatorial cores show a smaller range in variability from about 0.7 to 2.8 during the Holocene and from 0.7 to 3.6 during the last glacial. Based on ²³²Th flux measurements, we hypothesize that the location at which eolian detrital fluxes surpass the riverine detrital fluxes is approximately 300 km from the margin. While riverine fluxes from coastal margins were higher during the Holocene, eolian fluxes were higher during the last glacial.

Introduction

Essentially all sediments that reach the pelagic ocean floor are either derived from the continents through weathering processes or formed as a result of biological productivity in surface water. Contemporaneous climatic conditions largely affect the processes and mechanisms that bring these sediments to the ocean floor. For example, sediment intervals that record higher biogenic fluxes are often interpreted as being deposited at a time during which export production to the seafloor was increased. Similarly, intervals during which lithogenic particle fluxes are high can represent an intensified transport of continental material via rivers (if close to a continental margin) and/or wind. Thus, reconstruction of particle fluxes from oceanic sedimentary archives can broaden our understanding of past climate conditions.

Historically and to the present, sedimentary mass accumulation rates (MARs) have been estimated by multiplying the linear sedimentation rate (LSR), estimated using dated horizons (oxygen-isotope- or radiocarbon-derived), with sediment dry bulk density. This method measures the amount of sediment preserved at the sea floor but does not discriminate between vertically falling particles and those redistributed by a variety of horizontal advection processes. The xs²³⁰Th constant-flux proxy (CFP) method of determining mass accumulation rates is thought to “see through” such sediment redistribution processes and is purported to measure the true vertical flux (Bacon, 1984, Francois et al., 2004). The idea behind this constant-flux proxy lies in the different geochemical behavior of thorium and uranium in the oceanic water column. In seawater, ²³⁰Th is produced by the α-decay of ²³⁴U. Unlike uranium, which has a constant seawater concentration, thorium is extremely particle reactive, and the decay product, ²³⁰Th, is rapidly scavenged onto sinking particles so that the flux of ²³⁰Th to the ocean floor is identical to its rate of production in the water column (Bacon, 1984). Hence, MARs within an interval of sediment can be calculated by dividing the known production rate of ²³⁰Th by the concentration of ²³⁰Th within the same interval. Hence, the vertical flux of any component preserved in the sediment, F_i, can be calculated theoretically using the following equation:

$$F_1=\frac{{\left(conc\right)}_i*\beta *Z}{\left[x{s}^{230}T{h}_0\right]}$$

in which (conc)_i is the concentration of component i; β is the production rate of ²³⁰Th in the water column (0.0267 dpm m^− 3 yr^− 1); Z is the water depth in m; and [xs²³⁰Th_o] is the measured sedimentary ²³⁰Th activity corrected for decay, in situ production of ²³⁰Th from authigenic ²³⁴U, and detrital ²³⁰Th. The elegance of Eq. 1 is that the derived sedimentary flux is, by definition, solely the vertical component of the preserved sedimentary flux. Furthermore, one can solve for the ‘normalized’ sedimentary flux by measuring the concentration of xs²³⁰Th alone. The extent that such normalization works depends on how realistic the assumption is that the production of ²³⁰Th in the water column is equal to the flux of the scavenged ²³⁰Th to the underlying sediments (Bacon, 1984, Francois et al., 2004). If the postulated behavior of oceanic ²³⁰Th is correct, an added benefit to the constant-flux proxy methodology is that syndepositional sediment redistribution can be quantified by integrating the xs²³⁰Th inventory within an interval of sediment and comparing it to the integrated production of ²³⁰Th in the overlying water column over the time of accumulation (Suman and Bacon, 1989). Indeed, the ratio of these two parameters has been defined by a physical parameter called the “focusing factor, (Ψ)” (Suman and Bacon, 1989). A Ψ value of one implies that sediment has not been redistributed at the studied site. A Ψ value greater than one implies sediment in excess of what has been delivered vertically has been advected by deep-sea horizontal advection (i.e., focusing) to the studied site, while a Ψ value less than one implies winnowing or removal of sediment from the studied site at the time of sediment deposition (Francois et al., 2004). Model studies have shown that 70% of the ocean floor receives a ²³⁰Th flux within 30% of its production in the water column (Henderson and Anderson, 2003, Siddall et al., 2008), implying a sensitivity of the xs²³⁰Th profiling technique that is typically within +/−30%. This estimate must be considered somewhat tentative given the reliance of this result on the assumption that isopycnal and vertical diffusion is a reasonable approximation of ocean mixing processes inherent in many ocean models (Siddall et al., 2008).

Although age-model-derived and xs²³⁰Th-normalized MARs have been widely used in paleoceanographic research, in some cases the differently calculated MARs are significantly different, and, therefore, yield competing interpretations. Perhaps the best known of these discrepancies exists in the equatorial (west, central and east) Pacific Ocean (Broecker, 2008, Francois et al., 2007, Higgins et al., 1999, Kienast et al., 2007, Koutavas et al., 2002, Koutavas and Sachs, 2008, Kowsmann, 1973, Loubere et al., 2004, Lyle et al., 2005, Lyle et al., 2007, Marcantonio et al., 1996, Marcantonio et al., 2001a, Paytan et al., 1996, Thomas et al., 2000). Here, xs²³⁰Th-derived focusing factors, suggest that horizontal sediment transport almost always is higher (sometimes several times higher) than the vertical flux. The highest focusing factors (as high as 5.5; Kienast et al., 2007) are observed during the last glacial in the eastern equatorial Pacific (EEP) Ocean in the Panama Basin.

In the Panama Basin, using age-model-derived MARs, many investigators have concluded that particle fluxes during the last glacial were as much as 100% higher than those during the Holocene, and are caused by enhanced primary productivity (Lyle, 1988, Lyle et al., 2002, Paytan et al., 1996, Pedersen, 1983). However, xs²³⁰Th normalized MARs for sediments deposited during the last glacial suggest calcite fluxes that are 30–50% lower than those during the Holocene (Loubere et al., 2004). These authors contend that the higher glacial age-model-derived fluxes are due to sediment focusing processes in the Panama Basin. In addition, Kienast et al. (2007) reexamined several sites that were studied by others (Loubere et al., 2004, Lyle et al., 2005) in the Panama Basin and came to a similar conclusion; namely, that xs²³⁰Th-normalized MARs are lower and less variable than age-model-derived MARs, indicating varying degrees of sediment focusing.

Lyle et al. (2005) disagree with the interpretation that sediment focusing is widespread in the Panama Basin, and argue that xs²³⁰Th normalization overestimates the degree to which sediment redistribution processes are occurring in the EEP. They reason that the observed larger-than-expected inventories of sedimentary xs²³⁰Th in the EEP, that are in excess of those expected from a constant water column production rate of ²³⁰Th, can be attributed to increased boundary scavenging at the surface due to increased productivity close to the equator, in agreement with an analysis by Broecker (2008). However, within an efficient (low resolution) ocean circulation model, Siddall et al. (2008) found that particle scavenging effects are not sufficient to explain the additional xs²³⁰Th inventories measured in the Panama Basin. Using the Bern3D ocean model, they considered particle scavenging over a broad range of particle fluxes reaching up to 10 times higher than actual measurements in the equatorial Pacific region. Even at the highest end of this range of particle fluxes, the model by Siddall et al. (2008) suggests only a two-fold increase in the flux of ²³⁰Th over the production of ²³⁰Th in water column due to particle scavenging effects.

Kienast et al. (2007) propose that downslope transport of sediment from the east-west trending Carnegie Ridge, which forms the southern boundary of the Panama Basin, might explain the additional xs²³⁰Th in sediments of the Panama Basin. In this study, we test this downslope transport hypothesis by measuring xs²³⁰Th inventories of sediments deposited on the Cocos and Carnegie Ridges—regional topographic highs that surround the Panama Basin. In general, xs²³⁰Th inventories in sediment from the tops of ridges suggest sediment focusing factors that are greater than 1 for both the Holocene and glacial sediments. More importantly, sediment xs²³⁰Th inventories on the ridge tops are similar to those in the previously studied deeper cores (Kienast et al., 2007). If ridge tops were the source of extra xs²³⁰Th inventory in the basin, one would expect their focusing factors to be less than one and/or lower than those measured in the basin. We explore the potential causes for the larger-than-expected xs²³⁰Th inventories throughout the Panama Basin, including the effects of particle scavenging on ²³⁰Th fluxes to the seafloor.