Application of the Sm-Nd Isotope System to the Late Quaternary Paleoceanography of the Yermak Plateau (Arctic Ocean)

By storing and transporting vast amounts of energy derived from
solar insolation, the oceans play an important role in shaping
Earth’s climate. On the largest scale, ocean currents smooth the
temperature gradients between the equator and the poles by
redistributing excess energy from the tropics to higher latitudes.
Much of this excess heat is transported by the so-called Ocean
Conveyor Belt (Broecker, 1991), a global network of ocean currents
driven by thermohaline convection. Changes in the pattern and
strength of thermohaline circulation affect the redistribution of
heat, and thereby significantly influence climate on local to
global scales. The reconstruction of paleocurrents has long been a
subject of paleoceanographic research. Among the various methods
employed in tracing paleocurrents (and modern currents), the Sm-Nd
isotope system is experiencing ever increasing attention. First
applied in an oceanographic context by O’Nions et al. (1978), it is
by now established as a standard tool, as shown by numerous recent
publications (e.g. Rutberg et al., 2000; Tütken et al., 2002;
Weldeab et al., 2002; Benson et al., 2003; Farmer and Barber, 2003;
Piotrowski et al., 2004; Bayon et al., 2002, 2003, 2004; Lacan and
Jeandel, 2001, 2004, 2005, and many more). Two lines of application
of the Sm-Nd isotope system to oceanography/paleoceanography can be
distinguished, both of which were followed for this thesis. The
first approach uses the isotopic composition of Sm and Nd hosted in
detrital minerals to infer the provenance of terrigenous sediments.
This information can be used to draw conclusions about the
direction and distance of sediment delivery. The second approach
uses the isotopic signature of Nd as a tracer of different water
masses. Due to the oceanic residence time of Nd being shorter than
the global turnover rate of seawater (500-1000 years vs ~1000
years; Tachikawa et al., 2003), different bodies of water acquire
distinct Nd isotopic signatures as a function of the age of
adjacent continents. Apart from directly analyzing the Nd isotopic
compositions of water samples to trace the modern distribution of
different watermasses (e.g. Lacan and Jeandel, 2001, 2004),
suitable archives of seawater-derived Nd can be employed to study
paleocurrents. Possible archives are fossil remains of marine
organisms (e.g. foraminifers; Burton and Vance, 2000), or, most
widely used for the recent geological past, Fe-Mn nodules and
crusts (e.g. Frank et al., 2002). With slow growth rates on the
order of mm/Ma, however, Fe-Mn nodules do not offer the high
temporal resolution necessary to study Late Quaternary climate
change. Attention has therefore recently turned to authigenic Fe-Mn
oxyhydroxides finely dispersed throughout the sediment column (e.g.
Rutberg et al., 2000; Bayon et al., 2002, 2003, 2004; Piotrowski et
al., 2004). For this thesis, both lines of application of the Sm-Nd
isotope system to paleoceanography were followed. The samples were
taken from a sediment core collected from the Yermak Plateau in the
north-eastern Fram Strait. Situated between Greenland and the
Svalbard Archipelago, the Fram Strait is the only deep connection
between the Arctic Ocean and, via the Greenland-Iceland-Norwegian
(Nordic) Seas, the North Atlantic. The Nordic Seas are an area of
deep-water formation important for the global thermohaline
circulation. There, the processes of deep-water formation are in a
state of equilibrium that is most sensitive to changes in surface
water salinity, which, in turn, is strongly influenced by the
outflow of water of low salinity from the Arctic Ocean. This makes
the history of water exchange between the Atlantic and the Arctic
Ocean through the Fram Strait a subject of key interest for climate
research. In particular, it was attempted to reconstruct the
provenance of sediments deposited on the western Yermak Plateau
over the last 129 000 years. This was done by analyzing samples
from the sediment core and from potential source areas for their
Sm-Nd isotopic compositions. The current understanding is that
under present interglacial conditions sediment is delivered to the
Yermak Plateau by ice drift from the Siberian shelf areas (Kara-
and Laptev Sea) and as suspended load of Atlantic water advected
from the south. To resolve these assumed differences in provenance
and transport mechanism, the majority of the samples was split into
the grain-size fractions clay, fine silt, coarse silt, and sand for
Sm-Nd analyses. The position of the investigated core on the upper
slope of the western Yermak Plateau limits delivery of sand-size
(or coarser) material to ice rafting. The sand fractions of the
core samples were therefore interpreted to be exclusively of ice
rafted origin, and thus used as an indicator of changes in the
pattern of surface currents. Clay- to silt-size material, on the
other hand, yields a mixed signal of ice rafting and suspended-load
delivery. Based on a comparison of the isotopic compositions of the
core samples with those of the samples from potential source areas,
a number of conclusions can be drawn: Most core sample show only
little isotopic variation between their constituent size fractions
(mostly less than analytical uncertainty). Only sand fractions show
considerable differences. This can probably be explained by the
sand samples’ small sample size relative to their coarse grain
size; as a result, most sand fractions probably are not
representative. The generally good agreement between the isotopic
compositions suggests a common origin of ice rafted detritus (IRD)
and suspended load. The possibility of suspended particulate matter
transport from the Siberian shelf areas of the Kara- and Laptev
Seas to the Yermak Plateau in significant amounts can be excluded.
An origin of IRD in the Kara and Laptev Sea is therefore equally
unlikely. Instead, a common provenance of IRD and suspended
particulate matter from the Svalbard/Barents Sea area is a
plausible scenario, supported by isotope-independent data from the
literature (e.g. grain-size distribution, mineralogical
composition, faunal abundance, etc.). The moderate downcore Nd
isotopic variation suggests that, despite repeated large-scale
glaciations in the Svalbard/Barents Sea area, the general
modern-type circulation in the Fram Strait area has been active for
most of the last 129 000 years. The largest deviation from modern
conditions is indicated for the peak of the last glacial phase,
approximately 20 000 years ago. Then, large amounts of IRD were
delivered to the Yermak Plateau by icebergs calving from the
Scandinavian ice sheet. Moreover, the occurrence of chalk fragments
confirms iceberg drift from as far south as the North Sea. A
similar finding has previously been reported for samples from the
southern Fram Strait by Spielhagen (1991). Regarding the second
analytical approach, i.e. the Nd isotopic analysis of finely
dispersed authigenic Fe-Mn oxyhydroxides, implementation of the
experimental technique was targeted first. The method of Fe-Mn
oxyhydroxide extraction by means of leaching with a mixed reagent
(acetic acid and hydroxylamine-hydrochloride) largely is based on
the work of Chester and Hughes (1967). Modifications of their
method have been reported in Tessier et al. (1979), Chao and Zhou
(1983), and Hall et al. (1996), and have recently been compared by
Bayon et al. (2002). Based on the experimental protocol described
by Bayon et al. (2002), five core samples were processed and
analyzed for their rare earth element (REE) concentrations by
ICP-MS at the European Union Large Scale Geochemical Facility at
the University of Bristol, England, financed by the EU. In
addition, nine core samples were processed and the leachates
analyzed for their Nd isotopic composition in Munich. The REE
patterns of the leachates show an enrichment of the middle REE that
is atypical for authigenic Fe-Mn phases. The isotopic analysis also
yielded controversial results: downcore, the Nd isotope curves for
the leachates and the detrital phases run approximately parallel,
suggesting a systematic genetic relationship between the analyzed
Nd fractions. A similar relationship appears to exist between data
reported in Rutberg (2000), Rutberg et al. (2000), and Piotrowski
et al. (2004) for a sediment core from the south-eastern Atlantic.
To answer the questions raised by these controversial results, a
sequential leaching experiment was designed. Several aliquots of
one core sample were treated for different durations with different
concentrations of the leaching reagents, and at intermediate steps
were analyzed for their Sm-Nd isotopic composition. The results of
this leaching experiment point towards a conceptual weakness of the
method. In order to avoid contamination by non-authigenic sediment
components, all experimental methods described in the literature
focus on adjusting the concentration of the
hydroxylamine-hydrochloride used to reduce Fe and Mn to their
soluble states. This approach, however, does not take into account
the dissolution of acid-soluble phases by acetic acid, which in all
cases is used at a strength of 4.4 mol·l-1. Consequently, the
leaching reagent is sufficiently corrosive to attack easily-soluble
detrital minerals and release non-seawater-derived Nd (Hannigan and
Sholkovitz, 2001; Dubinin and Strekopytov, 2001). Phosphatic phases
are therefore a likely source of nonseawater-derived Nd. Apatite,
for instance, is a common component of clastic sedimentary rocks,
is easily dissolved by weak acids, and can account for the middle
REE enrichment in the leachates. Its high Nd concentrations would
mask any seawater signal. To conclude, it appears as though the
available extraction techniques are not yet sufficiently refined to
reliably determine the Nd isotopic composition of finely dispersed
Fe-Mn oxyhydroxides as a proxy for paleoseawater composition.