The so-called sea spray effect influences animals and humans living in coastal regions. As a consequence, δ13Ccarbonate, δ18Ocarbonate, δ18Ophosphate, and δ34Scollagen isotope values of affected individuals are more positive than otherwise expected. However, the effect is hidden in the case of humans who actually might have consumed marine food what would (partly) explain their isotopic signature. In order to correct for the sea spray effect in humans the dietary proportions were calculated based on the δ13Ccollagen and δ15Ncollagen isotope values using stable isotope mixing models. Four different programs (SISUS, simmr, IsotopeR, MixSIAR) were applied which resulted in quite different calculated diets. Each individual human can be corrected for the sea spray effect using the calculated proportion of terrestrial food (e.g. domesticated mammals, plants) and the approximated sea spray effect for each isotopic system. The differences in the calculated food proportions detected for the different mixing model programs, however, lead to differences in the correction procedure. We suggest using the dietary proportions as obtained by probabilistic SISUS rather than those of the Bayesian programs (simmr, IsotopeR, MixSIAR). The correction against the sea spray effect using the dietary proportions calculated by SISUS was supported by Gaussian Mixture Model (GMM) clustering which also enables the identification of probably non-local individuals in the dataset.
Keywords: Aerosol; Archaeometry; Bone; Cluster analysis; Diet.
Copyright © 2020 Elsevier B.V. All rights reserved.