We are currently experiencing one of the worst periods of biodiversity loss in Earth history due mainly to the ongoing pressure of human development. The effects of anthropogenic land use change on ecosystem composition and diversity are expected to be further compounded by greenhouse gas (i.e. CO₂,CH₄) induced climate changes over the coming two centuries. The pre-Quaternary fossil record cannot resolve decadal or century scale vegetation responses to climate change necessary to refine future predictions on the magnitude of biodiversity loss. However, the fossil record affords great potential for identifying thresholds of abiotic change, which have driven major ecosystem reorganizations in the past and may therefore have application to the future.

This talk will focus on macroecological responses of vegetation to a major carbon cycle perturbation at the Triassic-Jurassic mass extinction boundary, the third greatest extinction event in Earth history. Paleoatmospheric-CO₂ estimates using the stomatal proxy method demonstrate a four-fold increase in atmospheric CO₂ spanning the T-J boundary and 3 to 4 °C global warming. Presence-absence fossil plant data indicate that a 95% plant species-level extinction occurred coincidently with peak atmospheric CO₂ levels and global temperatures. However paleoecological analyses of 5386 identified fossil leaves from 8 isotaphonomic fossil plant beds spanning Rhaetian-Liassic aged Kap Stewart Formation, in East Greenland suggest that ecosystem destabilization may have been in place well before peak CO₂ and global temperatures were reached. These results suggest a gradual rather than catastrophic causal mechanism for Triassic-Jurassic extinctions and that relatively minor abiotic forcings can result in marked ecosystem destabilization. The ecological and physiologically selectivity of the T-J extinction event will be discussed with reference to changes in fossil leaf generic richness, evenness and plant community structure. Feedbacks between fossil plant community composition and landscape stabilization/destabilization (as reflected in microsedimentological data) will also be explored.