Understanding differences in chemistry climate model projections of stratospheric ozone

Chemistry climate models (CCMs) are used to project future evolution of stratospheric ozone as concentrations of ozone-depleting substances (ODSs) decrease and greenhouse gases increase, cooling the stratosphere. CCM projections exhibit not only many common features but also a broad range of values for quantities such as year of ozone return to 1980 and global ozone level at the end of the 21st century. Multiple linear regression is applied to each of 14 CCMs to separate ozone response to ODS concentration change from that due to climate change. We show that the sensitivity of lower stratospheric ozone to chlorine change ΔO3/ΔCly is a near-linear function of partitioning of total inorganic chlorine (Cly) into its reservoirs; both Cly and its partitioning are largely controlled by lower stratospheric transport. CCMs with best performance on transport diagnostics agree with observations for chlorine reservoirs and produce similar ozone responses to chlorine change. After 2035, differences in ΔO3/ΔCly contribute little to the spread in CCM projections as the anthropogenic contribution to Cly becomes unimportant. Differences among upper stratospheric ozone increases due to temperature decreases are explained by differences in ozone sensitivity to temperature change ΔO3/ΔT due to different contributions from various ozone loss processes, each with its own temperature dependence. Ozone decrease in the tropical lower stratosphere caused by a projected speedup in the Brewer-Dobson circulation may or may not be balanced by ozone increases in the middle- and high-latitude lower stratosphere and upper troposphere. This balance, or lack thereof, contributes most to the spread in late 21st century projections.