Evidence for sea spray effect on oxygen stable isotopes in bone phosphate - Approximation and correction using Gaussian Mixture Model clustering.

Palaeobiodiversity research based on stable isotope analysis in coastal environments can be severely hampered by the so-called "sea spray" effect. This effect shifts the isotopic signal of terrestrial individuals towards too marine values. It is commonly agreed upon that sea spray influences sulphur stable isotopes. However, we were able to approximate a remarkable sea spray effect also in carbon and oxygen stable isotopes of bone carbonate previously. In the present study we could approximate a minimum sea spray effect of about 13.9% even present in oxygen isotope values of bone phosphate, which was validated by Gaussian Mixture Model (GMM) clustering. This approximated value is by some magnitudes smaller than the minimum sea spray effect approximated for both δ13Ccarb and δ18Ocarb, and quite close to the sea spray detected for δ34Scoll in a previous study. It may therefore be interpreted as purer minimum sea spray signal compared to the approximation in bone carbonate. Furthermore, detection of sea spray in δ18Ophos can serve as additional validation of the effect present in bone carbonate, which is more prone to diagenetic alteration compared to bone phosphate. Moreover, the presence of the sea spray effect in both δ18Ocarb and δ18Ophos demonstrates that sea spray can be taken up by terrestrial mammals not only via food (δ18Ocarb) but also via drinking water (δ18Ophos). Finally, this study once more confirmed that calculation of δ18Ophos from δ18Ocarb values using a fixed oxygen isotope spacing (Δδ18O) can be highly misleading, especially in coastal environments affected by sea spray.