2014 GSA Annual Meeting in Vancouver, British Columbia (19–22 October 2014)

Paper No. 106-2
Presentation Time: 8:15 AM

A NEW STOMATAL-BASED METHOD FOR RECONSTRUCTING ATMOSPHERIC CO2


FRANKS, Peter J.1, ROYER, Dana L.2, BEERLING, David J.3, VAN DE WATER, Peter K.4, CANTRILL, David J.5, BARBOUR, Margaret M.1 and BERRY, Joseph A.6, (1)Faculty of Agriculture and Environment, University of Sydney, Sydney, 2015, Australia, (2)Earth and Environmental Sciences, Wesleyan University, 265 Church St., Middletown, CT 06459, (3)Department of Animal and Plant Sciences, University of Sheffield, Sheffield, S10 2TN, United Kingdom, (4)Department of Earth & Environmental Sciences, California State University, Fresno, CA 93740, (5)Royal Botanic Gardens Victoria, South Yarra, 3141, Australia, (6)Department of Global Ecology, Carnegie Institution of Washington, Stanford, CA 94305

Estimates of atmospheric CO2 concentration for the Phanerozoic Eon are based on a variety of proxies, but quantitative reconstructions for the pre-Cenozoic come overwhelmingly from one approach, the carbon isotopic composition (δ13C) of pedogenic carbonate. One proxy for reconstructing CO2 during the late Cretaceous and Cenozoic relies on an empirical relationship between stomatal distributions and CO2, but estimates are usually unbounded above ~500-1000 ppm and are limited to only a handful of calibrated species. Here, we calculate atmospheric CO2 using a well-validated equation for leaf gas exchange, with key variables obtained directly from the δ13C and stomatal anatomy of fossil leaves. This mechanistic approach avoids the unbounded errors in extrapolative methods and is applicable to most fossil material bearing stomata, therefore opening up much of the 400 m.y. paleobotanical record for analysis. Our new approach, validated against ice cores and direct measurements, produces CO2 estimates that are less than 1000 ppm for most of the post-Devonian, coincident with the appearance and global proliferation of forests. Uncertainties, obtained from Monte Carlo simulations, are typically less than for CO2 estimates from other approaches. These results provide critical new empirical support for the emerging view that large (~2000-3000 ppm), long-term swings in CO2 do not characterize the post-Devonian and that Earth’s long-term climate sensitivity to CO2 is greater than originally thought.