Mediterranean-Atlantic water exchange over the Miocene-Pliocene boundary

Resumen

The Mediterranean-Atlantic gateway connection plays a major role in ocean circulation. In the past during the Messinian Salinity Crisis (MSC; 5.33-5.97 Ma), water exchange between the Mediterranean Sea and the open Atlantic Ocean became restricted or intermittently absent, resulting in extensive evaporitic deposits in Mediterranean Basins. The Miocene-Pliocene boundary marks the re-establishment of Mediterranean-Atlantic exchange through the Gibraltar Strait and once exchange was initiated, Mediterranean Outflow water (MOW) developed. This thesis integrates multi-proxy records of MOW outside the Mediterranean Sea in the Gulf of Cadiz to study the termination, initiation and development of Mediterranean Outflow during the late Miocene to early Pliocene. The initiation and development of MOW was investigated in Integrated Ocean Drilling Program (IODP) Hole U1387C from 731.20 to 865.85 m below seafloor (mbsf), which is located on the upper slope in the Gulf of Cadiz (559 m water depth). Combining bio-, magneto-, and cyclostratigraphy, an age model was constructed with a time interval from about 5.0 to 5.7 Ma. The Miocene-Pliocene boundary at ~826 mbsf in this Hole is associated with a distinct and abrupt change in depositional environment. During the latest Messinian, hemipelagic sediments exhibit precession-induced carbonate cycles, in which the largest quantity of detrital input are found during Northern Hemisphere insolation minima. These cyclic patterns are likely to result from terrigenous input to the basin, while during insolation maxima, more widespread vegetation cover reduced riverine sediment load. There are no indications of bottom water currents or any other strong influence of MOW in stable isotopes or XRF scanning data in Hole U1387C during the late Miocene. Consequently, it is thought that MOW was absent or severely restricted during this time period. Just above the Miocene-Pliocene boundary, two contouritic bigradational sandy-beds and an increase and more variable XRF core scanning Zr/Al ratios with respect to the underlying sediment, indicate the activity of processes related to particle sorting. This provides the earliest evidence for the onset of bottom water current flow immediately at or just above the Miocene-Pliocene boundary. During the Pliocene, strong coupling between sedimentary variations and benthic stable isotope suggest the influence of bottom waters on the stable isotopes. When bottom water currents are stronger, stable isotopic expressions with similar values as in Hole U1387C are found in the Mediterranean benthic foraminiferal record from the Capo Rosello sections (Sicily, Italy), characterized by well-ventilated, warmer waters. However, bottom water currents are repeatedly interrupted, as indicated with more aluminosilicate rich beds with smaller grain sizes. During these intervals, benthic stable isotopes do not match the Mediterranean record, but record the influence of a colder and poorly ventilated water mass likely to derive from the Atlantic. Remarkably, sea surface d18O and relative abundances of planktic foraminifers Globigerinoides record coeval transitions, but in the opposite direction; when bottom waters become cold, warming is observed at surface. This pattern can be attributed to short episodes of estuarine circulation in the Mediterranean Sea, forcing Atlantic Inflow water over the sill of the Gibraltar Strait at depth while MOW exited at surface. Such a circulation pattern can be triggered by lower water densities at surface in the Mediterranean, what can be triggered by a heat balance that is less negative and a fresher net water budget. Alternatively, a three-layer flow through the gateway or a vertical shift of the MOW-plume towards shallower depths may have caused the observed pattern in the Gulf of Cadiz, although, the latter does not explain the quick synchronous shifts in the benthic stable isotopes with sea surface records. The origin of bottom water masses in the Gulf of Cadiz was also investigated through the use of authigenic neodymium (Nd) and lead (Pb) isotopes. Isotope ratios of Ferro-Manganese (Fe-Mn) oxyhydroxide sediment leachates of Hole U1387C are compared to ratios in sediments leachates of the Alboran Sea and in three Fe-Mn crusts of different water depths in the NE Atlantic (Abouchami et al., 1999; Muiños et al., 2008), to investigate endmember compositions of MOW and NE Atlantic water during the late Miocene to early Pliocene. Authigenic isotope ratios seem to correspond to oceanographic and deposistional settings, however, it remains unclear what exactly controls the incorporated authigenic signal. The isotopic Nd imprint does not match any endmember isotopic signatures as recorded in the Fe-Mn crusts. Therefore, this record is interpreted as mainly controlled by local processes, best explained by the release of rare earth elements with accompanied Nd isotopic compositions from suspended detrital sediments. Pb isotope ratios in the Alboran Sea sediments and NE Atlantic Fe-Mn crusts interpretations are incompatible. If however, the sediment leachate samples from the Alboran Sea provide reliable Pb isotopic compositions for MOW, then the leachate compositions match the stable isotope interpretations. This hints towards the usability of Fe-Mn oxyhydroxide Pb isotopic compositions as a reliable bottom water tracer in the marginal settings of the Gulf of Cadiz. Mediterranean-Atlantic water exchange before the MSC took place through the Betic Corridor in southern Spain and the Rifian Corridor in North West Morocco. These two gateways ceased during the late Miocene, but the actual timing of the closure of these Corridors has been under debate (Flecker et al. 2015). In the western Betic Cordilleras, upper Miocene basins nearby Ronda, Antequera and Arcos de la Frontera are adjacent to the late Miocene Guadalhorce Corridor, which has previously been described as the last remaining branch of the Betic Corridor (Martín et al., 2001). Sedimentary sequences in these basins are characterized by a transition from deeper marine sandy marls to shallow marine calcarenites or limestones. Biostratigraphic analyses of the marls indicate an age of deposition before 7.58 Ma. The overlying calcarenites and limestones were not reliably dated due to the lack of adequate material, however, based on the typical high sedimentation rates of these deposits, it seems unlikely that the calcarenites and limestones are much younger than the marls. This implies that the Guadalhorce Corridor is likely to have closed during the late Tortonian or early Messinian and thus cannot have supplied the necessary water to the Mediterranean to develop the extensive evaporite deposits. Considering similarities in timing and lithological successions, the shallowing upward sequences are linked to the same tectonic pulse that closed the remaining branches in the eastern Betics.