The study, published in the journal Climate of the Past, presents the first high-resolution, quantitative reconstruction of dust provenance in the EPICA Dome C ice core, one of the most celebrated climate archives on Earth. Led by Sibylle Boxho and Steeve Bonneville of the Université Libre de Bruxelles, the team analyzed rare earth element patterns in 279 samples of ancient dust spanning 33,700 to 2,800 years before present. Using a novel statistical unmixing algorithm called DEEPOT, they quantified for the first time exactly how much dust reached this remote inland site from each of the Southern Hemisphere’s major source regions, resolving changes on timescales of roughly a century.

The technical achievement lies in the chemistry of the lanthanoid series, the fourteen rare earth elements whose relative abundances act like a fingerprint for the rocks from which dust was derived. Traditional provenance work has relied on strontium, neodymium, and lead isotope ratios, but those methods demand relatively large sample masses, a serious problem at EPICA Dome C where dust fluxes plummeted by up to a factor of twenty-five from glacial to interglacial conditions. Rare earth element patterns, by contrast, capture multidimensional geochemical information from far smaller samples, and because the patterns of potential source areas differ subtly but consistently, a constrained least-squares model can decompose any measured dust sample into a weighted mixture of contributing sources.

To make the approach work, the researchers compiled a database of 245 rare earth element patterns from known dust source regions across the Southern Hemisphere: Patagonia and Tierra del Fuego, Central Western Argentina, the Puna-Altiplano Plateau, Southern Africa, Australia, New Zealand’s South Island, and the ice-free McMurdo Dry Valleys of Antarctica itself. Each ice core sample was then modeled as a linear combination of these end-members, with the constraint that all contributions must be positive, since physical proportions of dust cannot be negative. To guard against analytical uncertainty, the team ran 2,000 Monte Carlo simulations per sample, randomly perturbing measured concentrations within their error bars and recalculating the best-fit source mix each time. Only samples whose modeled patterns correlated strongly with the measurements, at a threshold corresponding to statistical significance, were retained, yielding 238 reliable reconstructions.

The headline result is strikingly clear. During the late Marine Isotope Stage 3, the Last Glacial Maximum, and Heinrich Stadial 1, roughly 65 to 75 percent of all dust deposited at Dome C came from Patagonia, with secondary contributions from New Zealand, Australia, Southern Africa, and the Puna-Altiplano. This dominance was remarkably stable, even as total dust flux began collapsing around 18,000 years ago. The explanation lies in the geography of the ice age world: the Patagonian Ice Sheet then stretched 2,090 kilometers along the Andes, grinding bedrock into fine glacial flour that braided rivers spread across vast outwash plains. With global sea level about 135 meters lower than today, an additional 763,000 square kilometers of continental shelf lay exposed, dry, and ready for the strong westerly winds to strip away.

New Zealand’s role emerged as a surprise. The country contributed an average of nearly 14 percent of dust at Dome C during the Last Glacial Maximum, making it the second-largest source, a finding that contrasts with earlier isotope-based studies that had argued against significant New Zealand input to East Antarctica. The researchers point to the extensive South Island ice sheet, an enlarged continental shelf of roughly 62,000 square kilometers, and cold, windy glacial conditions as factors that plausibly boosted New Zealand dust emissions to levels capable of reaching the deep interior of the continent. Modern trajectory modeling, they note, suggests New Zealand dust can still contribute between 14 and 32 percent of deposition at inland Antarctic sites.

Then, around 14,500 years ago, everything changed. The dust assemblage shifted decisively away from high-latitude glaciogenic sources toward lower-latitude ones. Patagonia’s share fell to an average of 43 to 53 percent during the Holocene, while Australia surged to become the second-most important supplier at 21 to 28 percent, the Darling Basin in particular. Southern Africa and the Puna-Altiplano each climbed to roughly 8 to 9 percent, and dust from the McMurdo Dry Valleys, essentially absent before 13,000 years ago, appeared as the Antarctic valleys deglaciated. The Holocene record also became far more variable, with source contributions fluctuating at high frequency in a way the stable glacial regime never did.

The drivers of this transition, the authors argue, were geological and hydrological rather than purely meteorological. The timing coincides with Meltwater Pulse 1A, a rapid episode of global sea-level rise that submerged vast tracts of the Patagonian and New Zealand continental shelves, permanently removing exposed, deflatable sediment from the wind’s reach. At the same time, the waning Patagonian Ice Sheet triggered a major reorganization of regional drainage: new channels opened, redirecting large portions of river catchments from the Atlantic to the Pacific, while rivers on the eastern flank of Patagonia shifted from braided to meandering forms. Together these changes starved the aeolian system of the fine sediment that had fed Antarctic dust for millennia. Crucially, the provenance shift was decoupled from the earlier, massive decline in total dust deposition between 18,000 and 15,000 years ago, which the researchers attribute instead to enhanced wet scavenging, rainout of particles along warmer, moister transport pathways.

Independent lines of evidence bolster the reconstruction. When the team converted their rare earth element-based source contributions into expected strontium and neodymium isotope ratios, the modeled values matched the sparse isotopic measurements from Dome C and other East Antarctic cores, clustering around less radiogenic signatures during the glacial period, consistent with fresh, mafic Patagonian and New Zealand lithologies, and drifting toward more weathered, radiogenic compositions in the Holocene as Australian and African inputs grew. Comparison with the EPICA Dronning Maud Land core, on the Atlantic side of the continent, revealed broadly synchronous millennial-scale changes but telling regional differences: the Atlantic-sector core received more Patagonian dust, while Dome C, deeper inland and in the Indian Ocean sector, received more Australian material, consistent with subtropical jet streams lofting low-latitude dust over the polar vortex during Rossby wave intrusions.

The broader significance extends beyond paleoclimate trivia. Mineral dust delivered iron to the Southern Ocean during glacial times, fertilizing marine productivity and helping draw down atmospheric carbon dioxide, so knowing precisely where that dust originated sharpens our understanding of the iron fertilization feedback that helped terminate ice ages. The study also demonstrates that sea-level rise and postglacial hydrological reorganization can durably rewire hemispheric dust cycles, a reminder that coastlines and river systems are active, threshold-sensitive components of the Earth system. As modern sea levels climb and hydrology shifts once again, the dust trapped in Antarctic ice offers a sobering preview of how quickly the planet’s connective tissue can be rearranged when the climate crosses a tipping point.

Subject of Research: Quantitative reconstruction of Southern Hemisphere dust provenance in Antarctic ice during the last glacial-interglacial transition

Article Title: Quantifying Southern Hemisphere dust sources during the Last Glacial-Interglacial Transition using rare earth elements in the EPICA Dome C ice core

Article References: Boxho, S., Vanderstraeten, A., Mattielli, N., Laruelle, G. G., Bory, A., Gabrielli, P., & Bonneville, S. (2026). Quantifying Southern Hemisphere dust sources during the Last Glacial-Interglacial Transition using rare earth elements in the EPICA Dome C ice core. Climate of the Past, 22(9), 1655-1674. https://doi.org/10.5194/cp-22-1655-2026

Image Credits: AI Generated

DOI: 10.5194/cp-22-1655-2026

Keywords: EPICA Dome C, ice core, dust provenance, rare earth elements, Last Glacial Maximum, Patagonia, Australia, Puna-Altiplano, sea-level rise, Southern Hemisphere circulation, last deglaciation, Climate of the Past