The Eocene-Oligocene transition (EOT), ~34 Ma, marks a tipping point in the long-term Cenozoic greenhouse to icehouse climate transition. Paleorecords reveal stepwise rapid cooling and ice growth across the EOT tightly coupled to a transient benthic δ13C excursion and a major and permanent deepening of the carbonate compensation depth (CCD). Based on biogeochemical box modeling, Merico et al. (2008) suggested that a combination of (1) glacioeustatic sea level fall-induced shelf-basin carbonate burial fractionation and (2) shelf carbonate weathering can account for the carbon cycle perturbation, but this finding has been questioned. Alternative proposed mechanisms include increased ocean ventilation, decreased carbonate burial, increased organic carbon burial, increased silicate weathering, and increased ocean calcium concentration. Here we use an improved version of the biogeochemical box model of Merico et al. (2008) to reevaluate these competing hypotheses and an additional mechanism, the expansion of "carbon capacitors" such as permafrost and peatlands. We find that changes in calcium concentration, silicate weathering, and carbonate or organic carbon burial each yield a response that is fundamentally at odds with the form and/or sign of the paleorecords. Shelf-basin carbonate burial fractionation (CCD change), plus shelf carbonate weathering, sequestration of 12C-enriched carbon into carbon capacitors, and possibly increased ocean ventilation (δ13C excursion), offers the best fit to the paleorecords. Further work is needed to understand why the EOT carbon cycle perturbation is so unique when the forcing mechanisms hypothesized to be responsible (cooling and ice growth) are not peculiar to this event.
