Distinguishing current effects in sediments delivered to the ocean by ice. I. Principles, methods and examples

Abstract

There are climatically important ocean flow systems in high latitudes, for example the East and West Greenland and Labrador Currents and Nordic Sea overflows in the North, and Antarctic Circumpolar Current in the South, for which it would be useful to know history of flow strength. Most of the sediment records under these flows contain evidence of supply from glacial sources, which has led to the supposition that fine sediment records, which in other settings provide evidence of vigour of flow from the sortable silt proxy, are fatally contaminated by unsorted glacial silt. It is suggested here that if the fine fraction (<63 μm) has been transported and sorted, then it does not matter that it may have been released from icebergs, sea ice or meltwater plumes. Here we show that correlation between sortable silt mean and percentage provides a good indicator of whether a fine sediment record has been sufficiently well current-sorted to provide a reliable flow history. The running downcore correlation (rrun) (5 to 9-point depending on sampling interval) is found to be optimal, and a value of rrun < 0.5 is proposed as an indicator of sufficiently poor sorting to invalidate a section of mean size record. More than 40 grainsize records determined by laser particle sizers from over 30 core sites have been processed and examined for evidence of sorting. As expected, there is a tendency for poor sorting and unreliable records at points where the flow speed has decreased to very low values. There is no consistent relationship between the sorting of the fine fraction and the content of coarse ice-rafted debris (as long as the IRD fraction is not > 50%) because the two are not related. End member (EM) decomposition of several records yields variable results in terms of the relationship between EM ratios and grainsize parameters. Although such an approach can generate fine sediment parameters it does not provide a basis for deciding whether or not a record is acceptably current sorted and thus contains a valid flow speed proxy. Our proposed discrimination between current-sorted and unsorted fine fractions is applicable to all fine grained deposits, not only high-latitude deposits with coarse IRD.Examples from East Greenland, Faroe Bank Channel, Gardar Drift show mainly well sorted signatures. Amounts of coarse IRD range up to 60% with only those >50% having a consistent impact on sortable silt mean size. With the exception of a Southern Ocean site on the Antarctic continental rise where half the record is poorly sorted, the silt mean data are sufficiently well sorted to provide credible flow speed histories. This bodes well for the extraction of such histories from climatically important high-latitude flows such as the East Greenland Current.