Early-Middle Pliocene Sea-Level Elevations in the Western Mediterranean

Improving our knowledge of sea level variability during the Pliocene allows an enhanced understanding of the response of major ice sheets to warmer temperatures under higher than present CO2levels. Due to the limited information of the extent of Pliocene ice masses, the stability of major ice sheets remains unclear, and consequently, sea-level estimates vary greatly. In this study we investigate sea-level variability during early to middle Pliocene using phreatic overgrowths on speleothems (POS) from Coves d’Artà, a coastal cave in Mallorca (western Mediterranean). POS are carbonate encrustations that precipitate at the air-water interface in brackish waters onto preexisting supports, such as vadose speleothems and cave walls. Since Artà Cave is located within 10s of meters of the Mediterranean Sea, the cave water level is and was in the past, coincident with the sea level. These deposits can be dated accurately with U-Th and U-Pb methods. POS provide precise spatial geographic positioning, an accurate elevation, an absolute age, and have a clear indicative meaning, which makes them ideal sea-level index points (SLIP). The U-Pb isochron dates for six POS levels (+ 22.5, + 23.5, + 25, + 27.5, + 30.5, and + 32 m above present sea level; mapsl) in Artà provide the first absolutely dated sea-level elevations between 4.39 ± 0.39 and 3.27 ± 0.12 Myr. The oldest sea-level stand at +32 mapsl yields an age of 4.39 ± 0.39 Myr, whereas the youngest one formed 3.27 ± 0.12 Ma ago at 23.5 mapsl. The latter one is of particular interest since it documents the sea elevation during the Mid-Pliocene Warm Period, an interval that perhaps provides the best analogue for ice sheet sensitivity in a slightly warmer than today’s climate. In order to relate local sea level to global (ice equivalent) mean sea level, the elevation of local SLIP need to be corrected for post-depositional deformation, such as glacial isostatic adjustment (GIA) and dynamic topography (DT). GIA simulations indicate that a downward correction of a few meters is necessary when using the local elevation to infer global mean sea level, while DT simulations are significantly less constrained. Due to its temporal extent, the POS record can put unique constraints on models of dynamic topography and the evolution of global mean sea level during the Pliocene