Causes of dust size variability in central East Antarctica (Dome B): Atmospheric transport from expanded South American sources during Marine Isotope Stage 2

Highlights

• South American provenance of coarse- and fine-mode dust events at Dome B (Antarctica).

• Atmospheric circulation controls dust size variability in Antarctic ice cores.

• The Patagonian shelf became an important dust source during MIS 2 sea level minimum.

• Importance of Southern Patagonian glacial dust sources.

• Raman determination of micron-size mineral grains.

Abstract

We here investigate the spatial and temporal variability of eolian dust particle sorting recorded in the Dome B (77° 05′ S, 94° 55’ E) ice core, central East Antarctica, during Marine Isotope Stage (MIS) 2. We address the question whether such changes reflect variable transport pathways from a unique source area or rather a variable apportionment from diverse Southern Hemisphere sources transported at different elevation in the troposphere. The Sr-Nd radiogenic isotope composition of glacial dust samples as well as single-particle Raman mineralogy support the hypothesis of a single dust provenance both for coarse and fine mode dust events at Dome B. The southern South American provenance of glacial dust in Antarctica deduced from these results indicate a dust composition coherent with a mixture of volcanic material and minerals derived from metamorphic and plutonic rocks. Additionally, Dome B glacial samples contain aragonite particles along with diatom valves of marine benthic/epiphytic species and freshwater species living today in the northern Antarctic Peninsula and southern South America. These data suggest contribution from the exposed Patagonian continental shelf and glacial outwash plains of southern Patagonia at the time when sea level reached its minimum. Our results confirm that dust sorting is controlled by the relative intensity of the two main patterns of tropospheric dust transport onto the inner Plateau, i.e. fast low-level advection and long-range high-altitude transport including air subsidence over Antarctica.

Introduction

Ice cores provide direct and highly-resolved records of climate and aerosol load of the atmosphere over different timescales (EPICA Community Members, 2004, Kawamura et al., 2017, Petit et al., 1999). They record atmospheric parameters as well as forcing factors of global significance such as greenhouse gases, and of more regional significance such as mineral dust aerosol. The eolian dust record from EPICA Dome C (Lambert et al., 2008) along with the paleo-temperature record inferred from stable isotopes of water (Jouzel et al., 2007) allowed assessment of climate and atmospheric circulation changes in the Southern Hemisphere over the past ∼800 kyrs. By showing a significant correlation between dust flux and temperature during cold glacial periods, which is absent during interglacial periods, the EPICA Dome C dust record provided robust evidence for a progressive coupling of Antarctic and southern Hemisphere climate as temperature became colder (Lambert et al., 2008).

The first-order covariance among ice core dust records from different parts of East Antarctica (Petit and Delmonte, 2009) suggests a broad uniformity in the dust input onto the Plateau. During MIS 2 (∼19–30 kyr BP), the last glacial period, the general uniformity of glacial dust flux contrasts with the regionally-variable character of eolian mineral dust size (Delmonte et al., 2004a), a parameter linked to transport processes. Such differences were interpreted as the expression of different atmospheric pathways for dust windborne to the polar area (Delmonte et al., 2004a).

Eolian dust reaching central East Antarctica is micron-sized and well-sorted, as a consequence of long-distance atmospheric transport (Petit and Delmonte, 2009). The spherical-equivalent diameter of background dust particles is smaller than 5 μm and the modal value of dust mass-size distribution is generally around 2 μm. A previous study on the Dome B ice core (Delmonte et al., 2004a), drilled in the Antarctic interior (Fig. 1), revealed clear oscillations in the modal diameter of particles (spanning from 2.0 to 2.7 μm) during MIS 2 and a subsequent decrease throughout the last termination until the Holocene, when particles became smaller (∼1.8 μm). This glacial/interglacial pattern of changes observed at Dome B is opposite to what is observed in other areas of the Antarctic Plateau such as Dome C and Komsomolskaya (Fig. 1). Superimposed on glacial/interglacial trends, dust size oscillations occur over a wide range of periodicities (Delmonte et al., 2004a, Delmonte et al., 2005, Wegner et al., 2015). The underlying mechanism proposed for the interpretation of dust grading in central East Antarctic ice cores associates small particles to long-range high-altitude transport including mass convergence in the middle troposphere above Antarctica and subsequent air subsidence. Conversely, large particles are associated to lower-level advection events linked to the presence of cyclonic systems off the Antarctic coast including meridional, short-cut advections from the South Atlantic Ocean (Krinner and Genthon, 2003, Krinner et al., 2010). Interestingly, model-based investigations show that both during the Last Glacial Maximum (LGM) and today South American dust is transported over Antarctica at lower levels with respect to Australian dust (Albani et al., 2012, Krinner et al., 2010, Li et al., 2008).

Previous studies based on Sr and Nd radiogenic isotopes on several East Antarctic ice core sections, integrated to obtain a few large samples, concluded that southern South America was the major dust supplier for central East Antarctica during MIS 2 (Basile et al., 1997, Delmonte et al., 2004a, Delmonte et al., 2004b, Grousset et al., 1992), as supported by Pb isotope data (Vallelonga et al., 2010) and rare earth elements patterns (Gabrielli et al., 2010). According to these studies, source regions include Patagonia and possibly the Pampas region and generally the southern part of Central Western Argentina at lower latitude (Basile et al., 1997, Gaiero et al., 2007, Gaiero, 2007, Gili et al., 2016, Gili et al., 2017). During glacial-climate conditions, the so called source intensity changes in South America include: (I) increased dust production and deflation resulting from enhanced aridity and reduced continental vegetation (Basile et al., 1997, Mahowald et al., 1999); (II) intensified surface winds (Werner et al., 2002) and foehn winds on the lee side of the Andes resulting from the increased volume of the Patagonian ice sheet (Kaiser and Lamy, 2010 and references therein); (III) increased amount of fine sediments resulting from rock-flour production and deposition on the Patagonian glacial outwash plains (Sugden et al., 2009); and finally, (IV) expansion of deflation areas following exposure of the expanded Argentine/Patagonian continental shelf during glacio-eustatic low stand (Grousset et al., 1992). The 120–130 m sea-level drop during the last glacial (Spratt and Lisiecki, 2016 and references therein) approximately doubled the modern continental surface available for deflation around southern South America (Fig. 1). This source expansion contributed to the exacerbation of the continental character of climate both in the Pampas and in Patagonia. The role of the Patagonian continental shelf as dust source to Antarctica during MIS 2 is however still controversial (Basile et al., 1997, Gaiero et al., 2003, Gaiero et al., 2007, Kaiser and Lamy, 2010). Dust entrained from the shelf displays a geochemical signature similar to sedimentary and volcanic detritus produced in the adjacent continent (Gaiero et al., 2003, De Mahiques et al., 2008). Moreover, the sharp dust decrease at the end of the last glacial period was not in phase with sea level rise (Wolff et al., 2006) (Fig. 2A). Whether the areal expansion of the South American source actually translated into more dust transported to East Antarctica remains as an open question.

In this work we address the issue raised originally by Krinner et al. (2010) whether dust particle size variability recorded in Antarctic ice cores may reflect atmospheric transport from a single source area exclusively or a variable dust apportionment from different sources (e.g., Australia and South America), modulated in turn by source intensity changes and/or atmospheric transport. We approached this issue by analyzing the Sr and Nd isotopic composition of well-selected coarse-mode and fine-mode dust events from the glacial (MIS 2) and early deglacial portion of the Dome B ice core (Fig. 1). The strontium and neodymium radiogenic isotope fingerprint of dust, commonly used in Antarctic ice core research to discriminate among dust sources (Grousset and Biscaye, 2005), is here complemented by single-grain Raman mineralogical analyses performed on a subset of dust samples. This innovative technique allows us to identify all mineral species and polymorphs constituting the sample, along with their relative abundances (Andò et al., 2011, Andò and Garzanti, 2014), providing crucial information to draw conclusions about parent rocks and conditions under which the sediment formed (Godoi et al., 2006, Villanueva et al., 2008). By coupling these two powerful complementary approaches with microscopic observations we can assess the origin of glacial dust size changes recorded in the Antarctic ice cores, provide further evidence for South American provenance of dust deposited in inner Antarctica during MIS 2, and clarify the role of the Patagonian continental shelf and southernmost continental regions as dust sources at times of minimum sea level reached during MIS 2.