Late Miocene Freshwater Runoff Seasonality Inferred by LA-ICP-MS and TIMS Analyses on Eastern Mediterranean Corals

Distinct trace element ratios in corals (e.g., Sr/Ca, U/Ca, Ba/Ca) can be used to monitor environmental conditions such as sea surface temperature (SST), soil erosion, or river runoff. Only trace element ratios from diagenetically unaltered corals yield results suitable for environmental reconstructions. Generally, such material is only available from strata of Pliocene or younger age. However, we found exceptionally well-preserved Porites corals of Late Miocene age on Crete (Eastern Mediterranean) in a succession of various clastic sediments. Previous studies on stable isotope (δ18O, δ13C) compositions yielded SST seasonality patterns, similar to those found in present-day corals. In this study, trace element concentrations in coral skeletons of these samples were measured using a 213 nm the axis of maximum growth, corresponding to eight to eleven samples per year. Also, two 87Sr/86Sr profiles were measured using a multi-collector TIMS (ThermoFisher Triton) on powder samples gained by equidistant microdrilling (spherical 0.8 mm bit) at a resolution of 4 samples per year. Both 87Sr/86Sr profiles run parallel to the sampling transects of the LA-ICP-MS and previously published stable isotope analyses representing a time span of 7 years. Sr/Ca and U/Ca co-vary with δ18O, and thus closely reflect SST seasonality. Other element ratios (e.g. Al/Ca, Ba/Ca, REE/Ca) do not correlate with δ18O (and thus SST), but do, however, exhibit well- defined peaks in the winter months. This pattern is not considered to originate from diagenetic alteration or post- growth contamination. Rather, these element ratios most likely represent seasonal terrigenous input into the coral reef environment syn-genetically incorporated into skeletal porosity as non-lattice bound components during coral growth. This pattern could arise from seasonality in Saharan dust plumes. However, because of its winter dominance it more likely reflects variations in riverine suspended load from Crete into the near-shore environment of the corals caused by winter rainfall. This is in good agreement with our SEM results indicating terrigenous particles trapped and incorporated into skeletal parts grown during the winter. In this case, the non- lattice bound element ratios are proxies for the frequency and intensity of heavy winter rain events and associated plumes of local turbid fresh water in the Eastern Mediterranean during the Late Miocene. The 87Sr/86Sr ratios cluster closely around 0.708910, confirming an age of ~9 Ma for the corals, and suggest a working hydrographic working of the Eastern Mediterranean Basin to the open ocean at that time. Nevertheless, several Sr isotope excursions (down to 0.708676) occur in the summer seasons, requiring additional input of unradiogenic Sr. These excursions do not correspond straightforwardly to the trace element patterns, implying that 87Sr/86Sr variability is unrelated to winter sediment discharge events. Although the Nile has suitably unradiogenic 87Sr/86Sr, its low dissolved Sr concentration effectively rules it out as a viable source. At present, the origin of the Sr isotope excursions remains enigmatic.