Disentangling variations of dust concentration in Greenland ice cores over the last glaciation: An overview of current knowledge and new initiative

Abstract

Greenland ice cores are high-resolution archives of atmospheric dust loadings, and a better understanding of the driving factors for Greenland ice core dust concentration changes across different timescales provides valuable insights into changes in atmospheric circulation pattern and benefits accurate prediction of the future climate change. In this review, based on the dust record of the NGRIP ice core, we systematically characterize the Greenland ice core dust concentration changes during the last glacial, and explore the major driving factors.Based on the previously published and own isotope geochemical data, we estimated the contributions of provenances for the Greenland ice core dust using Bayesian mixing model. The results show that the dust was primarily derived from the Asian sources (71.5%) during the last glacial. Dust contributions from Africa include 10% (6.9%–12.0%) for North Africa and 6.9% (4.6%–8.3%) for West Africa; East Central & Eastern European sources yield a contribution of 11.7% (8.3%–14.2%). Subsequently, we revealed the contributions of multiple influencing factors to the Greenland dust concentration changes on orbital scale using the Lindeman-Merenda-Gold method. Overall, the findings highlight a crucial role of the aeolian activity in sources in controlling the Greenland dust concentration changes during the last glacial (63%). A greater effect of the atmospheric transport efficiency is also found (31%); by contrast, the spilt of jet stream contributes little (5%). We further argue that the variations in contribution of intensity of aeolian activity in sources through MIS2–5 may be related to glacial grinding by high-altitude mountain glaciers, with more significant influences exerted on variations in the Greenland dust concentration during MIS4. Inversely, more contributions from the atmospheric transport efficiency (wet deposition and reduced atmospheric residence time of dust in transport pathways) are observed during MIS2 and MIS3 than those during MIS4 and maybe MIS5. The results about changes of the contributing proportions of intensities of the different source areas throughout MIS2–5 reveal the major roles of sediment availability, displacement of the westerlies, and the Northern Hemisphere ice volume. The variations in contribution of the split of jet stream through MIS2–5 also demonstrate its minor role in influencing the Greenland dust concentration changes. In addition, it is suggested that the 21-ka precessional period found in the Greenland dust records possibly reflects the integrated response of paleoclimate changes to Milankovitch forcing (northern summer insolation). As for variations in the last glacial maximum (LGM) dust concentration, the enhanced dust emissions induced by changes of the North Atlantic jet stream over the Central and Eastern Europe and the Tarim Basin promoted the sharp dust concentration increases during the early-LGM. Such a scenario was intimately related to more frequent Rossby-wave breakings under negative NAO/AO phases. Furthermore, loss of dust aerosol due to wet deposition en route and ice accumulation rates could also have affected the Greenland dust concentration changes during the LGM. The findings are thought to enable interpretation of the observed decoupled relationship between the CLP loess accumulation rates and the Greenland dust concentrations during the LGM.