Combined Sr-Nd-Hf isotopic data of two reference materials (AGV-1/BCR2) and 50, 10 and 5 mg aliquots of carbonate-free fine grain (<10 µm) separates of three loess samples (Central Europe/NUS, China/BEI, USA/JUD) are presented. Good agreement between measured and reference Sr-Nd-Hf isotopic compositions (ICs) demonstrate that robust isotopic ratios can be obtained from 5-10 mg size rock samples using the ion exchange/mass spectrometry techniques applied. While 87Sr/86Sr ratios of dust aluminosilicate fractions are affected by even small changes in pretreatments, Nd isotopic ratios are found to be insensitive to acid leaching, grain-size or weathering effects. However, the Nd isotopic tracer is sometimes inconclusive in dust source fingerprinting (BEI and NUS both close to epsilon-Nd(0) -10). Hafnium isotopic values (<10 µm fractions) are homogenous for NUS, while highly variable for BEI. This heterogeneity and vertical arrays of Hf isotopic data suggest zircon depletion effects towards the clay fractions (<2 µm). Monte Carlo simulations demonstrate that the Hf IC of the dust <10 µm fraction is influenced by both the abundance of zircons present and maturity of crustal rocks supplying this heavy mineral, while the <2 µm fraction is almost unaffected. Thus, epsilon-Hf(0) variations in the clay fraction are largely controlled by the Hf IC of clays/heavy minerals having high Lu/Hf and radiogenic 176Hf/177Hf IC. Future work should be focused on Hf IC of both the <10 and <2 µm fractions of dust from potential source areas to gain more insight into the origin of last glacial dust in Greenland ice cores.

Monte Carlo simulation were done in MATLAB to model zircon depletion effects on the Hf isotopic composition of hypothetical 5 mg amount of dust (<10 and <2 micron fractions). All simulation parameters as defined in the original publication.

Supplement to: Újvári, Gábor; Wegner, Wencke; Klötzli, Urs; Horschinegg, Monika; Hippler, Dorothee (2018): Sr-Nd-Hf Isotopic Analysis of <10 mg Dust Samples: Implications for Ice Core Dust Source Fingerprinting. Geochemistry, Geophysics, Geosystems, 19(1), 60-72