A measurement/model comparison of ozone photochemical loss in the Antarctic ozone hole using Polar Ozone and Aerosol Measurement observations and the Match technique

[1] The Polar Ozone and Aerosol Measurement (POAM III) instrument has provided 6 years ( 1998 to present) of Antarctic ozone profile measurements, which detail the annual formation of the ozone hole. During the period of ozone hole formation the measurement latitude follows the edge of the polar night and presents a unique challenge for comparing with model simulations. The formation of the ozone hole has been simulated by using a photochemical box model with an ensemble of trajectories, and the results were sampled at the measurement latitude for comparison with the measured ozone. The agreement is generally good but very sensitive to the model dynamics and less sensitive to changes in the model chemistry. In order to better isolate the chemical ozone loss the Match technique was applied to 5 years of data to directly calculate ozone photochemical loss rates. The measured loss rates are specific to the high solar zenith angle conditions of the POAM-Match trajectories and are found to increase slowly from July to early August and then increase rapidly until mid-September. The Match results are sensitive to the choice of meteorological analysis used for the trajectory calculations. The ECMWF trajectories yield the smallest, and perhaps most accurate, peak loss rates that can be reproduced by a photochemical model using standard JPL 2002 kinetics, assuming reactive bromine (BrOx) of 14 pptv based solely on contributions from CH3Br and halons, and without requiring ClOx to exceed the upper limit for available inorganic chlorine of 3.7 ppbv. Larger Match ozone loss rates are found for the late August and early September period if trajectories based on UKMO and NCEP analyses are employed. Such loss rates require higher values for ClO and/or BrO than can be simulated using JPL 2002 chemical kinetics and complete activation of chlorine. In these cases, the agreement between modeled and measured loss rates is significantly improved if the model employs larger ClOOCl cross sections ( e. g., Burkholder et al., 1990) and BrOx of 24 ppt which reflects significant contributions from very short-lived bromocarbons to the inorganic bromine budget.