Interannual Variability of the Antarctic Ozone Hole in a GCM. Part II: A Comparison of Unforced and QBO-Induced Variability

Abstract

Simulations were performed with the Goddard Institute for Space Studies GCM including a prescribed quasi-biennial oscillation (QBO), applied at a constant maximum value, and a physically realistic parameterization of the heterogeneous chemistry responsible for severe polar ozone loss. While the QBO is primarily a stratospheric phenomenon, in this model the QBO modulates the amount and propagation of planetary wave energy in the troposphere as well as in the stratosphere. Dynamical activity is greater in the easterly than in the unforced case, while westerly years are dynamically more quiescent. By altering zonal winds and potential vorticity, the QBO forcing changes the refraction of planetary waves beginning in midwinter, causing the lower-stratospheric zonal average temperatures at Southern Hemisphere high latitudes to be ?3?5 K warmer in the easterly phase than in the westerly during the late winter and early spring. Ozone loss varies nonlinearly with temperature, due to the sharp threshold for formation of heterogeneous chemistry surfaces, so that the mean daily total mass of ozone depleted in this region during September was 8.7 ? 1010 kg in the QBO easterly maximum, as compared with 12.0 ? 1010 kg in the westerly maximum and 10.3 ? 1010 kg in the unforced case. Through this mechanism, the midwinter divergence of the Eliassen?Palm flux is well correlated with the subsequent springtime total ozone loss (R2 = 0.6). The chemical ozone loss differences are much larger than QBO-induced transport differences in our model. Inclusion of the QBO forcing also increased the maximum variability in total ozone loss from the ?20% value found in the unforced runs to ?50%. These large variations in ozone depletion are very similar in size to the largest observed variations in the severity of the ozone hole. The results suggest that both random variability and periodic QBO forcing are important components, perhaps explaining some of the difficulties encountered in previous attempts to correlate the severity of the ozone hole with the QBO phase.