Stable isotope ratios and paleoceanaographic reconstructions from sediment cores 80-548 and 161-978A

Largely continuous millennial-scale records of benthic d18O, Mg/Ca-based temperature, and salinity variations in bottom waters were obtained from Deep Sea Drilling Project Site 548 (eastern Atlantic continental margin south of Ireland, 1250 m water depth) for the period between 3.7 and 3.0 million years ago. This site monitored mid-Pliocene changes in Mediterranean Outflow Water (MOW) documented by continuously high Nd values between -10.7 and -9. Site 978 (Alboran Sea, 1930 m water depth) provides a complementary record of bottom water variability in the westernmost Mediterranean Sea, which is taken to represent MOW composition at its source. Both sites are marked by a singular and persistent rise in bottom water salinities by 0.7-1.4 psu and in densities by ~1 kg m-3 from 3.5 to 3.3 Ma, which is matched by an average 3 °C increase in bottom water temperatures at Site 548. This event suggests the onset of strongly enhanced deep-water convection in the Mediterranean Sea and a related strengthened MOW flow, which implies a major aridification of the Mediterranean source region. In harmony with model suggestions, the enhanced MOW flow has possibly intensified Upper North Atlantic Deep Water formation.

(Table 3) Mediterranean core-top d18O values for various planktonic foraminiferal species

A box model is presented to simulate changes in Mediterranean long-term average salinity and d18O over the past 20,000 years. Simulations are validated by comparison with observations. Sensitivity tests illustrate robustness with respect to the main assumptions and uncertainties. The results show that relative humidity over the Mediterranean remained relatively constant around 70%, apparently narrowly constrained to the lower end of the range observed globally over sea surfaces by the basin's land-locked character. Isotopic depletion in run off, relative to the present, is identified as the main potential cause of depletions in the Mediterranean d18O record. Also, slight increases in relative humidity (of the order of 5%) might have caused very pronounced isotopic depletions, such as that in sapropel S5 of the penultimate interglacial maximum. The model shows distinctly non proportional responses of d18O and salinity to environmental change, which argues against the use of isotope residuals in Mediterranean paleosalinity reconstructions.

Expanded GHOST database

In this study we review a global set of alkenone- and foraminiferal Mg/Ca-derived sea surface temperatures (SST) records from the Holocene and compare them with a suite of published Eemian SST records based on the same approach. For the Holocene, the alkenone SST records belong to the actualized GHOST database (Kim, J.-H., Schneider R.R., 2004). The actualized GHOST database not only confirms the SST changes previously described but also documents the Holocene temperature evolution in new oceanic regions such as the Northwestern Atlantic, the eastern equatorial Pacific, and the Southern Ocean. A comparison of Holocene SST records stemming from the two commonly applied paleothermometry methods reveals contrasting - sometimes divergent - SST evolution, particularly at low latitudes where SST records are abundant enough to infer systematic discrepancies at a regional scale. Opposite SST trends at particular locations could be explained by out-of-phase trends in seasonal insolation during the Holocene. This hypothesis assumes that a strong contrast in the ecological responses of coccolithophores and planktonic foraminifera to winter and summer oceanographic conditions is the ultimate reason for seasonal differences in the origin of the temperature signal provided by these organisms. As a simple test for this hypothesis, Eemian SST records are considered because the Holocene and Eemian time periods experienced comparable changes in orbital configurations, but had a higher magnitude in insolation variance during the Eemian. For several regions, SST changes during both interglacials were of a similar sign, but with higher magnitudes during the Eemian as compared to the Holocene. This observation suggests that the ecological mechanism shaping SST trends during the Holocene was comparable during the penultimate interglacial period. Although this "ecology hypothesis" fails to explain all of the available results, we argue that any other mechanism would fail to satisfactorily explain the observed SST discrepancies among proxies.

(Table 2) Grain composition of sediments from ODP Hole 161-974B and from beach samples

Miocene to Pleistocene sand and sandstone were recovered at Ocean Drilling Program Site 974 in the Tyrrhenian Basin and Sites 976 and 977 in the Alboran Basin. Sand detrital modes were determined for 45 samples from these sites, as well as 10 samples of Spanish beach sand. At Site 974, the Pleistocene section includes a number of volcaniclastic (vitric ash) and terrigenous sand layers; the latter are heterogeneous and contain sedimentary and metamorphic lithic fragments. Submarine canyon and onshore drainage patterns suggest that the most likely source of this sediment is the Tiber River drainage basin in central Italy, where a Pleistocene volcanic field is superimposed on Apennine orogenic rocks. In contrast, the Miocene sand in Unit III at Site 974 may have been derived from local basement highs. The quartzolithic composition and preponderance of metamorphic and sedimentary lithic debris in sand samples from Unit II at Site 976, Unit I at Sites 977 and 978, and Unit I at Site 979 are consistent with derivation from metamorphic rocks and sedimentary cover sequences that crop out in the Betic Cordillera of southern Spain (976-978) and in the Rif of Northern Africa (979). The sedimentary to metamorphic lithic fragment ratios in these samples reflect the relative proportion of metamorphic and sedimentary rocks exposed in onshore source terranes. In contrast, the source of the few quartzose Pleistocene sands at Site 976 was likely the Flysch Trough Units that crop out near Gibraltar. The significant volcanic component in certain intervals at Sites 976 (upper Miocene) and 977 (lower Pliocene to Miocene) is consistent with widespread volcanic activity during basin inception and development. Mean sand detrital modes for sand subgroups from both the Alboran and Tyrrhenian Basin sites plot in the Recycled Orogenic and Magmatic Arc compositional fields of Dickinson et al. (1983, doi:10.1130/0016-7606(1983)94<222:PONAPS>2.0.CO;2), reflecting the hybrid tectonic histories of these basins.

Alkenone unsaturation index and alkenones in ODP Hole 161-977A

A number of essential elements closely related to each other are involved in the Earth's climatic system. The temporal and spatial distribution of insolation determines wind patterns and the ocean's thermohaline pump. In turn, these last two are directly linked to the extension and retreat of marine and continental ice and to the chemistry of the atmosphere and the ocean. The variability of these elements may trigger, amplify, sustain or globalize rapid climatic changes. Paleoclimatic oscillations have been identified in this thesis by using fossil organic compounds synthesized by marine and terrestrial flora. High sedimentation rate deposits at the Barents and the Iberian peninsula continental margins were chosen in order to estimate the climatic changes on centennial time resolution. At the Barents margin, the sediment recovered was up to 15,000 years old (unit ''a'', from latin ''annos'') (M23258; west of the Bjørnøya island). At the Iberian margin, the sediment cores studied covered a wide range of time spans: up to 115,000 a (MD99-2343; north of the Minorca island), up to 250,000 a (ODP-977A; Alboran basin) and up to 420,000 a (MD01-2442, MD01-2443, MD01-2444, MD01-2445; close to the Tagus abyssal plain). At the northern site, inputs containing marine, continental and ancient reworked organic matter provided a detailed reconstruction of climate history at the time of the final retreat of the Barents ice sheet. At the western Barents continental slope, warm climatic conditions were observed during the early Holocene (~from 8,650 a to 5,240 a ago); in contrast, an apparent long-term cooling trend occurred in the late Holocene (~from 5,240 a to 760 a ago), in consistence with other paleoarchives from northern and southern European latitudes. The Iberian margin sites, which were never covered with large ice sheets, preserved exceptionally complete sequences of rapid events during ice ages hitherto not studied in such great detail: during the last glacial (~from 70,900 a to 11,800 a ago), the second glacial (~from 189,300 a to 127,500 a ago), the third ice age (~from 278,600 a to 244,800 a ago) and the fourth (~from 376,300 a to 337,500 a ago). In this thesis, crucial research questions were brought up concerning the severity of different glacial periods, the intensity and rates of the recorded oscillations and the long distance connections related to rapid climate change. The data obtained provide a sound basis to further research on the mechanisms involved in this rapid climate variability. An essential point of the research was the evidence that, over the past 420,000 a, at the whole Iberian margin, warm and stable long periods similar to the Holocene always ended abruptly in few centuries after a gradual deterioration of climate conditions. The detailed estimate of past climate variability provides clues to the natural end of the present warm period. Returning to an ice age in European lands would be exacerbated by a number of factors: a lack of differential solar heating between northern and southern north Atlantic latitudes, enhanced evaporation at low latitudes, and an increase in snowfall or iceberg discharges at northern regions. It must be emphasized that all climatic oscillations observed in this thesis were caused by forces of nature, i.e. the last two centuries were not taken into consideration.

(Table 1) Oxygen isotopic composition of interstitial waters from ODP Hole 161-976

Interstitial waters recovered from Ocean Drilling Program, Leg 161, site 976 in the western Mediterranean Sea are used in conjunction with a numerical model to constrain the delta18O of seawater in the basin since the Last Glacial Maximum, including Sapropel Event 1. To resolve the oxygen isotopic composition of the deep Mediterranean, we use a model that couples fluid diffusion with advective transport, thus producing a profile of seawater delta18O variability that is unaffected by glacial-interglacial variations in marine temperature. Comparing our reconstructed seawater delta18O to recent determinations of 1.0 per mil for the mean ocean change in glacial-interglacial delta18O due to the expansion of global ice volume, we calculate an additional 0.2 per mil increase in Mediterranean delta18O caused by local evaporative enrichment. This estimate of delta18O change, due to salinity variability, is smaller than previous studies have proposed and demonstrates that Mediterranean records of foraminiferal calcite delta18O from the last glacial period include a strong temperature component. Paleotemperatures determined in combination with a stacked record of foraminiferal calcite depict almost 9°C of regional cooling for the Last Glacial Maximum. Model results suggest a decrease of ~1.1 per mil in seawater delta18O relative to the modern value caused by increased freshwater input and reduced salinity accompanying the formation of the most recent sapropel. The results additionally indicate the existence of isotopically light water circulating down to bottom water depths, at least in the western Mediterranean, supporting the existence of an 'anti-estuarine' thermohaline circulation pattern during Sapropel Event 1.