Alkenone and boron based Oligocene pCO2 records

The Pliocene period is the most recent time when the Earth was globally significantly (~3°C) warmer than today. However, the existing pCO2 data for the Pliocene are sparse and there is little agreement between the various techniques used to reconstruct palaeo-pCO2. Moreover, the temporal resolution of the published records does not allow a robust assessment of the role of declining pCO2 in the intensification of the Northern Hemisphere Glaciation (INHG) and a direct comparison to other proxy records are lacking. For the first time, we use a combination of foraminiferal (delta11B) and organic biomarker (alkenone-derived carbon isotopes) proxies to determine the concentration of atmospheric CO2 over the past 5 Ma. Both proxy records show that during the warm Pliocene pCO2 was between 330 and 400 ppm, i.e. similar to today. The decrease to values similar to pre-industrial times (275-285 ppm) occurred between 3.2 Ma and 2.8 Ma - coincident with the INHG and affirming the link between global climate, the cryosphere and pCO2.

Orbital-resolution Pliocene-Pleistocene simulations of CO2, sea level, global temperature and benthic d18O, and d11B-based proxy-CO2 data

During the past five million yrs, benthic d18O records indicate a large range of climates, from warmer than today during the Pliocene Warm Period to considerably colder during glacials. Antarctic ice cores have revealed Pleistocene glacial-interglacial CO2 variability of 60-100 ppm, while sea level fluctuations of typically 125 m are documented by proxy data. However, in the pre-ice core period, CO2 and sea level proxy data are scarce and there is disagreement between different proxies and different records of the same proxy. This hampers comprehensive understanding of the long-term relations between CO2, sea level and climate. Here, we drive a coupled climate-ice sheet model over the past five million years, inversely forced by a stacked benthic d18O record. We obtain continuous simulations of benthic d18O, sea level and CO2 that are mutually consistent. Our model shows CO2 concentrations of 300 to 470 ppm during the Early Pliocene. Furthermore, we simulate strong CO2 variability during the Pliocene and Early Pleistocene. These features are broadly supported by existing and new d11B-based proxy CO2 data, but less by alkenone-based records. The simulated concentrations and variations therein are larger than expected from global mean temperature changes. Our findings thus suggest a smaller Earth System Sensitivity than previously thought. This is explained by a more restricted role of land ice variability in the Pliocene. The largest uncertainty in our simulation arises from the mass balance formulation of East Antarctica, which governs the variability in sea level, but only modestly affects the modeled CO2 concentrations.

Geochemical data derived from an absolutely dated shell-based chronology for the Irish Sea

The identification in various proxy records of periods of rapid (decadal scale) climate change over recent millennia, together with the possibility that feedback mechanisms may amplify climate system responses to increasing atmospheric CO2, highlights the importance of a detailed understanding, at high spatial and temporal resolutions, of forcings and feedbacks within the system. Such an understanding has hitherto been limited because the temperate marine environment has lacked an absolute timescale of the kind provided by tree-rings for the terrestrial environment and by corals for the tropical marine environment. Here we present the first annually resolved, multi-centennial (489-year), absolutely dated, shell-based marine master chronology. The chronology has been constructed by detrending and averaging annual growth increment widths in the shells of multiple specimens of the very long-lived bivalve mollusc Arctica islandica, collected from sites to the south and west of the Isle of Man in the Irish Sea. The strength of the common environmental signal expressed in the chronology is fully comparable with equivalent statistics for tree-ring chronologies. Analysis of the 14C signal in the shells shows no trend in the marine radiocarbon reservoir correction (DR), although it may be more variable before ~1750. The d13C signal shows a very significant (R**2 = 0.456, p < 0.0001) trend due to the 13C Suess effect.