Large amplitude variations in carbon cycling and terrestrial weathering during the latest Paleocene and earliest Eocene

Global temperatures rose ~6°C from the late Paleocene ca. 58 Ma to the Early Eocene Climatic Optimum \n(EECO) ca. 52–50 Ma. Superimposed, were certainly two geologically brief (< 200 kyr) intervals of \nextreme warming, the Paleocene-Eocene thermal maximum (PETM) and Eocene thermal maximum 2 \n(ETM-2 or H-1); at least four more events, H-2, I-1, I-2 and K/X, probably occurred. Both the longterm \nrise and short-term “hyperthermals” have been linked to massive injections of 13C-depleted carbon \ninto the ocean-atmosphere system and greater continental weathering. However, relationships remain \nuncertain, principally because detailed and coupled proxy records do not extend across the entire \ninterval of interest. \nMead Stream, New Zealand, exposes a ~650 m-thick sequence of limestone originally deposited on an \nupper continental slope from the late Cretaceous to the middle Eocene. Previous work has provided \naccurate ages for this expanded section, and has shown that the PETM, ETM-2, H-2, I-1 and I-2 \nhyperthermals are marked by pronounced negative carbon isotope excursions (CIEs) and clay-rich \nhorizons (marls), the latter caused by excess terrigenous dilution. 283 new samples were collected, \nmostly above the I-2 event; these were analyzed for carbonate content, lithology, and bulk carbonate \ncarbon isotopes. A greatly expanded (100 m-thick) unit represented by a series of marl beds lies above I-2 \nand correlates to the EECO. Carbonate contents are generally 60–90% throughout the studied interval, \nwith lows being marls. The δ13C is generally low but, in detail, represents a series of negative CIEs with \nmagnitudes ranging between 0.2–0.6‰. Of these, the K/X/ETM-3 event is the most pronounced (0.6‰). \nThe late Paleocene-early Eocene δ13C record at Mead Stream is remarkably similar to that generated \nat Site 1262 at Walvis Ridge (south Atlantic), except that lows in δ13C span intervals of relatively high \nsedimentation (terrigenous dilution) rather than intervals of relatively low sedimentation (carbonate \ndissolution). We suggest that over ~6 million years, there was a series of short-term climate perturbations, \neach characterized by massive carbon input and greater continental weathering. The suspected link \ninvolves global warming and enhanced seasonality in precipitation. We are evaluating this model further \nby identifying hyperthermal events in DSDP sites from the Indian Ocean, and assessing whether they \nare related to carbonate dissolution or siliciclastic dilution