Coupled chemistry-climate model simulations for the period 1980 to 2020: Ozone depletion and the start of ozone recovery

Two simulations of a coupled chemistry-climate model are completed for the period 1980 to 2020, covering the recent past during which extensive satellite ozone and temperature data exist, and covering the near future when ozone levels are expected to begin to recover. In the first simulation, Rayleigh friction is used to decelerate the polar night jet. In the second simulation, a parametrized spectral gravity-wave forcing scheme is included. This has the effect of considerably reducing the model temperature bias in the polar regions and weakening the polar night jet. In the simulations the concentrations of chlorine, bromine and the well-mixed greenhouse gas concentrations are specified in accordance with past observations and future projected values. The calculated trends in temperature and ozone in the two runs are similar, indicating that model internal variability does not have a significant impact and suggesting that the trends arise largely from changes in external parameters. Typically, after about the year 2000, the trend in the modelled annually averaged ozone changed from a decrease to a small increase. The change was found to be statistically significant in the upper stratosphere and in the lower stratosphere over Antarctica, which are the regions most affected by halogen chemistry. Globally averaged temperature results suggest that the best place to look for future atmospheric change is in the upper stratosphere. Decadally averaged statistics are used to estimate the timing of the start of recovery of total ozone. The simulations indicate no significant further ozone loss from the current atmosphere with minima typically occurring in the years from 2000 to 2005, except in the spring Arctic where ozone values continued to decrease slowly until the end of the integrations. One major problem with the detection of the start of ozone recovery, is that the concentrations of halogens are expected to reduce only slowly from their peak value. Hence, no substantial recovery is simulated before the year 2020. The difficulty in detecting the start of ozone recovery suggests the need to continue the model simulations until the second half of this century which would also help to establish the timing of complete ozone recovery.