The snowball stratosphere

Published in Journal of Geophysical Research-Atmospheres, 2019

Recommended citation: Graham, R. J., T. A. Shaw and D. S. Abbot, 2019: "The snowball stratosphere", Journal of Geophysical Research: Atmospheres, 10.1029/2019JD031361. https://doi.org/10.1029/2019JD031361

Download paper here

In this paper, we carry out the first GCM simulations of the stratosphere of the snowball Earth. I conducted this research with Dorian Abbot and Tiffany Shaw while pursuing my Bachelor’s in Geophysical Sciences at the University of Chicago.

A sharp, anomalous change in the ratio of oxygen-17 and oxygen-18 isotopes preserved in Neoproterozoic rocks is often used as evidence that CO2 partial pressure was elevated during the Marinoan snowball, since increased CO2 changes the partitioning of oxygen-17 and oxygen-18 in photochemical reactions between O2, O3, and CO2 that occur in the upper stratosphere. However, quantitatively mapping these changes in oxygen isotopes onto atmospheric pCO2 relies intimately on assumptions about the cross-tropopause mass exchange rate and the circulation and mixing of the stratosphere during the snowball, since any photochemical signal preserved in rocks at Earth’s surface will have been diluted to some degree during its journey from stratosphere to rock. Previous studies of these oxygen isotope anomalies assumed that the stratospheric mixing efficiency and cross-tropopause mass exchange rate during the snowball were the same as those of the modern atmosphere, despite the enormous differences in boundary conditions between the two climate states. We decided to test this hypothesis with explicit simulations of the snowball’s stratosphere. We found that the cross-tropopause mass exchange rate during the snowball is smaller than that of the modern atmosphere, and the stratospheric mixing efficiency is larger – but, since interpretation of the isotopic proxy for pCO2 actually relies on the product of these two quantities, the differences resulted in minimal changes to the interpretation of the geochemical evidence for high pCO2.

From a general dynamics point of view, in our simulations the stratospheric circulation during the snowball Earth was considerably weaker than that of the modern Earth, which we found was due to reduced wave driving in the stratosphere, likely because of the reduction in land-sea contrast on a planet with glaciers covering both land and sea. For the same reason (reduced wave driving), we found that the polar vortex in a snowball stratosphere with modern ozone levels is much more vigorous than on the modern Earth, and the vortex does not display sudden stratospheric warming events. In the simulation without ozone, since there is minimal solar heating in the stratosphere, there is no polar vortex at all.