Ice Core Archives of Mineral Dust - Mineral Dust: A Key Player in the Earth System - page 467

18.4 Dust Provenance

18.4.1 Antarctica

The earliest investigations of insoluble particles in Antarctic snow and ice used
elemental ratios to determine the source of particles to be primarily crustal, rather
than volcanic or marine, in nature (Petit et al. 1981 ). Such an approach had been
originally used to determine the inputs of anthropogenic elements to polar ice
cores (Murozumi et al. 1969 ). The deserts of the Southern Hemisphere landmasses
were considered the most likely sources of such dust, with the contribution also of
continental shelves exposed during glacial periods of low-standing sea levels.
A definitive evaluation of Antarctic dust sources was not available until 1992,
with a study of Sr and Nd isotopes in Dome C and PSA samples. Grousset et al.
( 1992 ) were the first to identify that South America was the dominant source of
dust, despite their sampling of a limited range of PSAs and the availability of
only one, composite, LGM sample from the first Dome C ice core (the analytical
technique required 11 half-metre ice core sections for one measurement). Despite
a limited number of samples from deserts and arid areas in Argentina, South
Africa, Australia and Antarctica, they established a good agreement between the
Argentinean samples and the Dome C composite sample, dated to 17 1kaBP.

Exposure of the Argentine continental shelf during glacial periods was invoked
as a possible explanation for the change in glacial-interglacial dust fluxes. They
confirmed the continental source of Dome C dust by REEs; the shape of the REE
pattern of the Dome C and continental dust samples were very similar, and clearly
different to a volcanic sample from the South Shetland Islands.
In the last two decades, Sr and Nd isotopic provenance techniques have been
expanded and complemented, but the primary findings have remained unchanged.
Basile et al. ( 1997 ) expanded the initial study to Vostok and Dome C ice (sample
size reduced to 7 kg) from the dustiest time periods of the last 160 ka: MIS2 (Dome
C samples from 16-20 ka ago), MIS4 (Vostok samples from 60-62 ka ago) and
MIS6 (Vostok samples from 158-167 ka ago). The field of PSAs was also expanded,
with additional samples from Tierra del Fuego, the Argentine continental shelf, New
Zealand and Antarctic moraines. They found that the Argentinean continental shelf
was not the source of enhanced Antarctic glacial dust fluxes, and that the isotopic
signature of dust deposited at Dome C and Vostok was identical. Delmonte et al.
( 2004a ) expanded the PSA dataset to southern South America, southern Africa,
New Zealand and the Antarctic Dry Valleys. They found a persistent common dust
source for Dome C and Vostok over the last several glacial cycles, attributable to a
mixture of southern South American sources (most likely Patagonia and the Puna
Altiplano). Recent studies confirm the same source for glacial dust at Talos Dome
(Delmonte et al. 2010 ) and Berkner Island (Bory et al. 2010 ). Strontium and Nd
isotopes have been determined in only six interglacial samples (Delmonte et al.
2007 ), but these are clearly offset from glacial dust samples (Fig. 18.6 ). The isotopic
offset indicates that other dust sources are active during interglacial periods, most

˙