TOTAL OZONE FROM THE TIROS OPERATIONAL VERTICAL SOUNDER DURING THE FORMATION OF THE 1987 OZONE HOLE

Total ozone maps obtained from the infrared radiances measured by the TOVS/HIRS2 instrument on board the NOAA 10 satellite are used to study the formation of the 1987 Antarctic "ozone hole." We use in this study an improved version of the retrieval algorithm described by Muller and Cayla (1983), now calibrated for middle and high latitudes and taking into account the unusually depleted ozone profiles of the Antarctic spring. Error analysis indicates that our method has an accuracy of the order of 5-7% in clear sky conditions. Values determined from the TIROS operational vertical sounder (TOVS) are in good agreement with Dobson measurements in the mid-latitudes and with the ozonesondes launched from the Antarctic stations during the Airborne Antarctic Ozone Experiment (AAOE). The agreement with the total ozone mapping spectrometer (TOMS) data at mid-latitudes is also good, but significant differences are found in early September in the high latitudes. In particular, the large-scale zonally symmetric minimum representative of the ozone hole appears later in the TOVS maps than in the TOMS data. The ozone hole was already apparent in the TOMS map on the first days of September, while TOVS detected only localized ozone deep minima associated with optically thick polar stratospheric clouds (PSCs) and did not observe any circular depletion structure until September 17. This discrepancy seems to be the consequence of high solar zenith angles and climatological errors in the TOMS algorithm, which tends to underestimate the ozone content in late winter. It is only in mid-September that TOVS data show a rapid ozone decrease affecting the whole vortex. The low ozone amounts are first recorded in the vicinity of the PSCs detected in the ozone field and then spread into the vortex. TOVS observations suggest that a rapid ozone decrease might take place during or just after the formation of major water ice PSCs. Since the vortex is exposed to UV sunlight in mid-September, this could be the direct consequence of both a sudden increase of free chlorine and an efficient denitrification occurring during type 2 PSC events. It is concluded that since the algorithm presented in this paper allows reliable ozone determinations in middle and high latitudes and accurate type 2 PSC detection, measurements from TOVS could play an important role in the ozone layer monitoring, especially in the wintertime polar regions where UV techniques are ineffective or affected by the lack of intense sunlight.