Faculty of Biology, Department Biology I, Anthropology and Human Genomics, Ludwig Maximilian University Munich
April 16, 2020 , Diamond 6
The definition of local bioavailable 87Sr/86Sr isotopic signatures is crucial for the detection of non-local skeletal finds and provenance analysis. A multi-isotope approach by use of additional isotopic ratios permits the establishment of an isotopic fingerprint that can be used for similarity search.
A mixing model was developed for the prediction of local 87Sr/86Sr signatures for selected sites along an archaeological important passage across the European Alps (Inn-Eisack-Adige passage across the Brenner Pass). This model was based on strontium concentrations and isotopic ratios of environmental samples (wood, water, soil) and correctly predicted the isotopic signatures of local vertebrates.
A multi-isotope fingerprint consisting of stable strontium, lead, and oxygen isotopes in the bioapatite of archaeological animals and human cremations (omitting the thermally unstable δ18Ophosphate in the latter) along the Alpine passage was forwarded to a Gaussian Mixture Model (GMM) clustering for the scope of similarity search. GMM clustering was capable of identifying groups of animals and humans that were spatially separated with a high probability (average p > 0.9). This way, local, non-local and also mixed isotopic signatures in the multi-isotope fingerprint were firmly detected.
GMM clustering was also successfully applied to a palaeoecological study in an ecological complex region at the Baltic coast (Viking Age Haithabu and medieval successor Schleswig). A multi-isotope fingerprint in vertebrate skeletal finds including humans that consisted of δ13Ccollagen, δ15Ncollagen, δ13Ccarbonate, δ18Ocarbonate, δ18Ophosphate, δ34Scollagen, and 87Sr/86Srapatite permitted the definition of e.g. fishing grounds, quantification of the “sea spray” effect, and provenance analysis.
German Science Foundation, grants no. GR 959/16-1,2 and GR 959/20-1,2