Variability of ozone loss during Arctic winter (1991–2000) estimated from UARS Microwave Limb Sounder measurements

[1] A comprehensive analysis of version 5 (V5) Upper Atmosphere Research Satellite (UARS) Microwave Limb Sounder (MLS) ozone data using a Lagrangian transport (LT) model provides estimates of chemical ozone depletion for the 1991–1992 through 1997–1998 Arctic winters. These new estimates give a consistent, three-dimensional picture of ozone loss during seven Arctic winters; previous Arctic ozone loss estimates from MLS were based on various earlier data versions and were done only for late winter and only for a subset of the years observed by MLS. We find large interannual variability in the amount, timing, and patterns of ozone depletion and in the degree to which chemical loss is masked by dynamical processes. Analyses of long-lived trace gas data suggest that the LT model sometimes overestimates descent at levels above ∼520 K, so we have most confidence in the results at lower levels. When the vortex is shifted off the pole and the cold region is near the vortex edge (e.g., late winter 1993 and 1996), most rapid ozone depletion occurs near the vortex edge; when the vortex and cold region are pole-centered (e.g., late winter 1994 and 1997), most ozone loss takes place in the vortex core. MLS observed the most severe ozone depletion in 1995–1996, with about 1.3 ppmv cumulative loss for the winter at 465 K by 3 March 1996; ∼1.0 ppmv cumulative loss is seen at 465 K by mid-March 1993. Analyses of MLS data show significant ozone loss during January in most years, ranging from ∼0.3 to 0.6 ppmv at 465 K. A modified LT model used with the limited MLS data in 2000 gives rough estimates of ∼0.04 and 0.006–0.012 ppmv/day during 2–12 February and 12 February–29 March 2000, respectively, broadly consistent with other studies of the 1999–2000 winter. Estimates of depletion in MLS column ozone above 100 hPa are considerably smaller than other reported column loss estimates, primarily because many estimates include loss below 100 hPa and because MLS does not continuously observe the Arctic after early spring. Our results from analyses of MLS data confirm previous conclusions of broad overall agreement between many ozone loss estimates in the Arctic lower stratosphere near 450–480 K.