Science Introduction

ER-2 & DC-8 Aircraft Mission To Investigate Antarctic Ozone In Late Winter 1987


Recent observations have shown since 1979 a dramatic and unexpected downward trend in the overhead column abundance of ozone during late winter and early spring over Antarctica, at the Halley Bay and Argentine Islands stations (76°S, 27°W and 65°S, 64°W). The reduction, amounting by 1985 to about 40% of the historical October monthly mean, has been confirmed and given a geographically mapped perspective by observations from NASA's Total Ozone Mapping Spectrometer (TOMS) on Nimbus 7. Ozonesonde ascents from Syowa station (69°S, 40°E) in 1982 and 1983 have shown that in October, before the final warming, ozone is depleted by between 10% and 50% at altitudes between about 10 and 22 km, compared to values observed in the late 1960's and early 1970's. The chemical data base has been considerably enhanced by the observations taken from late August to the beginning of November 1986 from McMurdo Base (78°S, 167°E) by the National Ozone Expedition (NOZE), which was organized by NSF. Ozone profiles, 33 in number distributed fairly evenly between 25 August and 6 November, confirmed the picture suggested by the Syowa data, and showed considerable vertical structure in the mixing ratio, particularly during October. It was clear that the ozone loss over McMurdo developed during September. Further, as yet unpublished, observations of the column abundances of O3, NO2, and OClO by the NOAA Aeronomy Lab, of HCl, ClONO2 and HNO3 (inter alia) by JPL and observations of N2O and ClO by SUNY which contain some coarse information about altitude distribution, should all have a substantial impact on knowledge of the photochemical balance, and its interpretation. The NOAA data Show evidence of very unusual odd nitrogen chemistry (very low abundance and very small diurnal variation of NO2 and of unusual chlorine chemistry (high abundance of OClO).

In any instance where spatial or temporal change of a stratospheric chemical species is observed, the first question to be asked is whether it can be attributed to chemical or meteorological processes, or both. Correlations of the Antarctic total ozone amount during the relevant time of year with stratospheric temperature have been reported by five sets of authors in the special issue of Geophysical Research Letters (November Supplement, 1986). Temperature correlation with ozone does not resolve this issue, since it could be cause or effect. As yet unpublished work from a collaborative study between the UK Meteorological Office, the ECMWF, and the NOAA Aeronomy Lab reveals clear correlations of the area bounded by TOMS total ozone contours in October and the area bounded by the Antarctic sea ice limit in late winter, from 1979-86. In addition, the Halley Bay total ozone amount during October is highly anticorrelated with Southern Hemisphere sea surface temperatures, averaged for July-August-September, over the period 1957-1985, with r = -0.74 (significant at better than 1% level). The angular momentum from 1000 to 100 mb, in the latitude belt 51.5°S to 65.9°S has shown a marked increase during the period 1976-1985, a result consistent with an equatorward shift in the contours of zonal mean wind in September, as analyzed by NASA Ames meteorologists from radiosonde data.

Although correlations have been established between tropospheric variables and total ozone during late Antarctic winter, no causal mechanism is immediately evident. The question then arises as to whether there is a corresponding signal of change in any stratospheric dynamical variables. The objectively analyzed cross-section of wind and temperature produced at NASA Ames from radiosonde data show that, for example, the 20 ms-1 September zonal mean wind contours enclose a monotonically increasing area; at 50 mb the increase from 1976 to 1980 amounts to a factor of two. Examination of stratospheric potential vorticity fields, calculated with a geostrophic wind assumption from satellite soundings of temperature and pressure, does not show evidence of systematic change in the lower Antarctic stratosphere during late winter. In the upper stratosphere, however, there is evidence of variation in the areas bounded by potential vorticity contours defining the edge of the vortex during much of the winter and spring. During Septembers 1979-86, the area sharply increased by ~20% from 1979 to 1980, and then decreased steadily until by September 1986 it was about 5% below the 1979 value. On the other hand, the October monthly means show an increase in area of about 50% from 1979 to 1985, with a partial recovery in 1986. This could be interpreted as a tendency to a later final warming in October, or as an expanded vortex, or some combination of the two. In either case, the effect would be to produce a downward trend in the total ozone column via the 15-20% of the total column which is above about 30 km. It cannot, however, account for changes Of 40-50% in the total column; most of such large changes can only arise from losses where most of the column is, in the lower stratosphere. Finally, it has been observed that polar stratospheric clouds occur in the same region as the Antarctic ozone depletion, and at about the same time; they have also been observed over small regions and less persistently, during the Arctic winter.

Summarizing the chemical and meteorological data, there is evidence both for an unusual chemical composition in the lower Antarctic stratosphere of late winter, and for changes in the Southern Hemisphere circulation which are correlated with the ozone change.


There are very broadly two categories of hypothesis which have been advanced to explain the late winter loss of Antarctic ozone. One is that it is essentially chemical in origin, the other that it is caused by a change in the circulation leading to enhanced ingress of ozone-poor air from the troposphere. The chemical theories require some means of maintaining a large positive departure of the mixing ratio of chlorine monOxide from that expected due to standard gas phase stratospheric photochemistry; most hypotheses have appealed to an intervention by heterogeneous phase chemistry occurring on polar stratospheric clouds. One suggestion along these lines is that active chlorine is released from HCl and especially ClONO2 by reactions on p.s.c., polar stratospheric clouds, surfaces; another is that nitric acid vapor is condensed into p.s.c.'s, having the effect of removing NOx from the gas phase and so allowing high Clx amounts because of the resulting inhibition of ClONO2 formation. The loss of HNO3 from the gas phase could permit high HOx amounts, which in turn can liberate Clx from HCl.

There are four chain reactions by which chlorine can destroy ozone:

ClO + O → Cl + O2 (1)
Cl + O3 → ClO+ O2 (2)
net: O + O3 → O2 + O2
ClO + BrO → Cl + Br + O2 (3)
Cl + O3 → ClO + O2 (4)
Br + O3 → BrO + O2 (5)
net: O3 + O3 → O2 + O2 + O2

Note that there is another channel in reaction (3)

ClO + BrO → Br + OClO (3b)

which does not cause catalytic O3 loss, and which produces the OClO molecule.

ClO + ClO + M → (ClO)2 + M (6)
(ClO)2 + hν → Cl + ClOO (7)
ClOO + M → Cl + O2 + M (8)
Cl + O3 → ClO + O2 (2)
Cl + O3 → ClO + O2 (2)
net: O3 + O3 → O2 + O2 + O2

ClO + HO2 → HOCl + O2 (9)
HOCl + hv → OH + Cl (11)
OH + O3 → HO2 + O2 (10)
Cl + O3 → ClO + O2 (2)
net: O3 + O3 → O2 + O2

One common feature to all four chain reactions is that ClO is the product of the reaction destroying O3, so measurement of it must be a high priority. Measurement of BrO and OClO would decide the extent of occurrence of the 2nd mechanism, (3a-4-5). None of the above mechanisms can be made to work with purely homogeneous gas phase chemistry, because the reaction

ClO + NO2 + M → ClONO2 + M (12)

locks up the chlorine in the reservoir species ClONO2, which it is thus obviously useful to measure. Another chlorine reservoir is HCl, so this too should be measured. HCl may be in either the gas phase or condensed on particle surfaces.

It is not certain how reactive nitrogen is partitioned between NO, NO2, N2O5, HNO3 and the aqueous phase (in p.s.c's), so measurements of these species and/or their sum total in the gas phase is also important. The nitric acid content of p.s.c. particles would provide a crucial constraint.

There are a number of other chemical theories, such as one suggesting a solar cycle induced enhancement of NOy that would affect ozone. Although these currently seem unlikely, evidence concerning them will be sought.

Measurements of species which have a source at the surface and a sink in the middle and upper stratosphere (for example N2O, CF2Cl2, CFCl3) should provide valuable data on how long the air sampled has been in the stratosphere, and whether it has descended during the period of winter circulation (March - October) from higher levels, or whether it has ascended from the troposphere below. Such measurements, if taken horizontally (or isentropically) across the vortex, could also test another possible mechanism, not so far aired in the literature, that increased frequency in recent years of isentropic transport of air from the lower midlatitude stratosphere to high latitudes would appear as an ozone reduction there; below about 60 mb, the isentropic ozone gradient from 35°S to 90°S appears to be equatorward in late winter.

Excerpt from Science and Objectives in Airborne Antarctic Ozone Experiment (AAOE) PDF file, 71p, National Aeronautics and Space Administration (NASA), Ames Research Center, 1987.