The 1992 global average total ozone, measured by the Total Ozone Mapping Spectrometer (TOMS) on the Nimbus-7 satellite, was 2 to 3 percent lower than any earlier year observed by TOMS (1979 to 1991). Ozone amounts were low in a wide range of latitudes in both the Northern and Southern hemispheres, and the largest decreases were in the regions from 10deg.S to 20deg.S and 10deg.N to 60deg.N. Global ozone in 1992 is at least 1.5 percent lower than would be predicted by a statistical model that includes a linear trend and accounts for solar cycle variation and the quasi-biennial oscillation. These results are confirmed by comparisons with data from other ozone monitoring instruments: the SBUV/2 instrument on the NOM-11 satellite, the TOMS instrument on the Russian Meteor-3 satellite, the World Standard Dobson Instrument 83, and a collection of 22 ground-based Dobson instruments.
The Nimbus-7 TOMS (Total Ozone Mapping Spectrometer) has observed the amount and distribution of atmospheric total column ozone since November 1978. From 1979 to 1991, the amount of total column ozone has decreased over most of globe (1-3). Small (3 to 5%) losses at mid-latitudes, larger (6 to 8%) losses at high latitudes, and no losses near the equator were reported (2, 3). The most dramatic ozone decease has been observed each year in the springtime Antarctic ozone hole region (4) and the 1992 ozone amounts there were about 50% of the 1979 amounts.
For latitudes between 65deg.S and 65deg.N, the average area-weighted ozone loss rate for all seasons (1979 to 1991), after correction for solar cycle and quasi-biennial oscillation (QBO) effects has been estimated to be 2.7 +/- 1.4% per decade (1-3). Analysis of the 13-year ozone data shows that most of the ozone depletion has occurred at mid- and high latitudes (2, 3). In this report we examine the decrease in the global daily average ozone amount from 1992 into 1993. We show that the observed decrease is consistent with measurements from other satellite and ground-based instruments.
The TOMS data show that the 1992 daily global average (65deg.S to 65deg.N) total ozone amount is significantly lower than in any of the earlier 13 years. The daily global average ozone value during 1992 falls 2 to 3% below the range of ozone values observed in any of the preceding 13 years (Fig. 1). The amounts were outside the earlier range starting in March, and reached a maximum of 4.7% below the daily mean of the earlier measurements near the end of December. The low values continued into January 1993 (Fig. 1). For the last 9 months of 1992, the daily ozone values were more than two standard deviations below the 12-year mean. In 1985 and 1987 the ozone amounts fell briefly, for 2 weeks, to the 2 standard deviation level, but the 1992 ozone decrease was much larger and of considerably longer duration.
The largest decreases in ozone amounts were observed in the regions from 10deg.S to 20deg.S and from 10deg.N to 60deg.N (Fig. 2). Outside of these bands, except near the equator, the amounts of ozone were at the lower edge of the range of values measured during earlier years. In the equatorial region, ozone amounts were near or above their climatological average. In contrast, during fall of 1991, ozone levels were unusually low in the equatorial region, and near the 12-year minimum between 45deg.N and 65deg.N. These low values may have been the result of changes in stratospheric circulation induced by the presence of aerosols from the Mount Pinatubo eruption (5). After November 1991, ozone amounts near the equator returned to normal values.
Some decrease in the 1992 global average ozone amount relative to the 1991 amount would be expected because the solar output has been declining during the current phase of the solar cycle, and because of the phasing of alternating high and low values of ozone associated with the QBO cycle (3). In addition to these cycles, there is a linear trend that has been estimated to be -2.7 +/- 1.4% per decade (1-3). A statistical model including the effects of the seasonal cycle, QBO, 10.7-cm solar flux cycle, and a linear trend accurately reproduces the observed ozone variation from 1979 to 1991 (1-3). For 12 years of data, 1979 to 1990, this model accurately predicts the 13th year, 1991. However, when we fit the data through 1991 (13 years) and used the model to predict amounts through the end of 1992, the observed values were 1 to 2% lower than the predicted values (Fig. 3). Thus, this statistical model that adequately describes the first 13 years of TOMS data failed to predict accurately the observed data in 1992.
We evaluated several possible sources of error in the TOMS data. Nimbus-7 is in a nearly sun-synchronous polar orbit that has slowly drifted from crossing the equator at noon to 10:45 a.m. over a period of 14 years. This relatively stable orbit means that observing conditions for Nimbus-7 TOMS have not changed significantly from year to year. The consistent observing conditions, and recent recalibration of Nimbus-7 TOMS (6), have permitted the determination of long-term trends in ozone amounts and relative year-to-year variation with high precision (1-3). The slow drift of the Nimbus-7 orbit, however, caused the user plate used for in-flight instrument calibration to be partially shaded between February 1992 and September 1992. On 30 September 1992, the Nimbus-7 TOMS instrument reacquired the sun, and normal solar calibration procedures were resumed. For the period between February 1992 and September 1992 the TOMS instrument calibration was interpolated between the February and September solar measurements, and the interpolated calibration was verified by comparison with other satellite and ground-based instruments (see below).
TOMS ozone data were affected by the stratospheric aerosol layer created by the eruption of Mount Pinatubo in June 1991. Radiative transfer calculations (7) show that a TOMS measurement at a single scan angle could have errors as large as 2% in low and mid-latitudes and up to 10% in high latitudes. However, these errors vary in both sign and magnitude over the range of scan angles. Their net effect is to cause a less than 1% overestimation of the zonal mean ozone in the low and mid-latitudes. At latitudes greater than 60deg., the aerosol effect is less than 1% in summer, but it can cause a 5% underestimation of ozone amounts near the time of the winter solstice. We estimate that the overall effect of the aerosols on the global mean ozone shown in Fig. 1 is about 0.3%. We conclude that there has not been an undetected change in the instrument calibration or sensitivity to ozone amounts and that the observed ozone decrease is real and not caused by Nimbus-7 TOMS instrument error or artifact.
To confirm the long-term stability of the Nimbus-7 TOMS ozone data, we compared the data with similar ozone data sets from the NOAA-11 SBUV/2 (solar backscattered ultraviolet), from the Meteor-3 TOMS, and with data from the groundbased Dobson network. The SBUV/2 instrument, onboard the NOAA-11 polar orbiting satellite, measures ozone by the same backscattered ultraviolet technique as the TOMS instrument. This instrument, which has been operational since 1 January 1989, observes 12 ultraviolet wavelengths in order to measure both the ozone altitude profile and total column ozone amounts. From 1990 to 1992, evidence from the onboard calibration system suggests that the relative error of the SBUV/2 total ozone measurements is less than 1%.
As with the TOMS data, the SBUV/2 data are 2 to 3% lower in 1992 than in 1990, the previous lowest year (Fig. 4). In addition, the latitudinal distribution of the decrease in ozone amounts in 1992 observed by SBUV/2 is similar to the decrease observed by the TOMS instrument. This similar ozone decrease indicates that the TOMS calibration has not drifted significantly relative to the SBUV/2 calibration.
A second TOMS instrument is on the Russian Meteor-3 spacecraft, launched on 15 August 1991. Unlike the Nimbus-7, the Meteor-3 is in a precessing orbit with a 212-day period. As a result, the Meteor-3 TOMS data can be obtained with solar zenith angles comparable to those for Nimbus-7 TOMS during periods when the Meteor-3 crosses the equator between 9 a.m. and 3 p.m. There have been five such periods of about 2 months each since its launch. During these periods (Fig. 5), the Nimbus-7 TOMS and Meteor-3 TOMS data show a consistent 1% bias. (In practice, the Meteor-3 TOMS data compare well with the Nimbus-7 TOMS data over a much wider range of equator crossing times.) The Meteor-3 comparison confirms that the interpolated calibration for the period February 1992 to September 1992 is accurate, and that the calibration did not shift in 1992.
The stability of TOMS is monitored relative to ozone observations made each summer at the Mauna Loa observatory with the World Standard Dobson Instrument 83 (I83) (8). The calibration of I83 has been maintained since 1962 to an accuracy of +/-0.5% (9). These comparisons indicate that the TOMS ozone measurements have been stable relative to I83 to approximately 0.5% (Fig. 6). The TOMS-I83 comparison in July and August 1992 is consistent with the comparison in previous years. The I83 comparison is a good test of the aerosol sensitivity of the TOMS measurements. Dobson measurements, because they are done with a pair of wavelength pairs, are not sensitive to contamination by the presence of atmospheric aerosols. The consistency of the I83-TOMS comparison confirms our radiative transfer calculations that show that the aerosols have a small net effect on the TOMS measurement (6).
We also compared the TOMS data with a summer (June-August) average of 22 Dobson stations for which data were available through 1992 (Fig. 6). Only direct-sun Dobson observations were used, and all the data were adjusted to use the Bass and Paur ozone cross sections. Although there was a small trend in the TOMS-Dobson difference (possibly caused by increasing tropospheric ozone), there was not a significant change in 1992. The absolute offsets of 4.5% relative to 183 and of 3% relative to the 22-station average most likely result from an error in the original ground calibration of TOMS. Both the World Standard Dobson Instrument and the 22-station average confirm that the TOMS calibration in 1992 was consistent with that for earlier years and that the low observed ozone values are not a result of a shift in the instrument calibration.
In summary, the 1992 global average amount of ozone is 2 to 3% lower than the lowest values observed in earlier years. The largest 1992 decreases occurred during November and December, and were 3 to 4 standard deviations below the 12-year daily mean. The largest decreases occurred from 10deg.S to 20deg.S and 10deg.N to 60deg.N. Only in the equatorial region are the ozone values well within the envelope data from earlier years. It is significant that 1992 is the first time that ozone amounts observed by TOMS showed a simultaneous sustained decrease over a wide latitude range in both hemispheres.
The cause of the 1992 low ozone values is uncertain. Although the mechanism for ozone decrease is unknown, the understandable first guess would be that the decrease is related to the continuing presence of aerosol from the Mount Pinatubo eruption. There are three possibilities related to the presence of the aerosol: (i) direct chemical loss through increased heterogeneous processing (10); (ii) an aerosol-induced change in radiative heating which can directly affect ozone transport; or (iii) changes in photochemical production or loss rates caused by the temperature changes resulting from the aerosol heating. The size and timing of these potential effects of heterogeneous processing have been modeled by several groups (11). In general, these model simulations have not predicted the size or the timing of the observed ozone decreases for 1992 to 1993. Transport effects caused by aerosol-induced radiative heating were proposed (5) as the cause of the short-term tropical ozone loss observed immediately after the Mount Pinatubo eruption, but the mechanism responsible for the long-term ozone changes observed more than 1 year after the eruption remains unknown.
J. F. Gleason, University Space Research Association, and National Aeronautics and Space Administration, Goddard Space Flight Center (NASA/GSFC), Code 916, Greenbelt, MD 20771.
P. K. Bhartia, J. R. Herman, R. McPeters, P. Newman, R. S. Stolarski, Laboratory for Atmospheres, NASA/GSFC, Greenbelt, MD 20771.
L. Flynn, Software Corporation of America, Lanham, MD 20706.
G. Labow, D. Larko, C. Seftor, C. Wellemeyer, Hughes-STX Corporation, Lanham, MD 20706
W. D. Komhyr, Cooperative Institute for Research in Environmental Sciences, University of Colorado and National Oceanographic and Atmospheric Administration (NOAA), Boulder, CO 80303.
A. J. Miller, Climate Analysis Center National Weather Service, NOAA, Camp Springs MD 20031
W. Planet, Satellite Research Laboratory, National Environmental Satellite, Data, and Information Service. NOAA. Camp Springs, MD 20031
1. R.S. Stolarski, P. Bloomfield, R.D. McPeters, J.R. Herman, Geophys. Res. Lett. 6, 1015 (1991); J.R. Herman, R. McPeters, R. Stolarski, D. Larko, R. Hudson, J. Geophys. Res. 96, 17279 (1992).
2. R.S. Stolarski et al., Science 256, 342 (1992).
3. R.S. Stolarski et al., World Metoerol. Org Ozone Rep. 25 (1992).
4. First reported by J.C. Faman, B.G. Gardiner, and J.D. Shanklin [Nature 315, 2017 (1985)] and confirmed as a regional phenomena by R.S. Stolarski et al. [ibid., 322, 808 (1986)].
5. S. Kinne, O. Toon, M.J. Prather, Geophys. Res. Lett. 19, 1927 (1992); S. Chandra, ibid. 20, 33 (1993); M.R. Schoeberl, P.K. Bahrtia, E. Hilsenrath, O. Torres, ibid., p. 29.
6. J.R. Herman et al., J. Geophys, Res. 96. 7531 (1991).
7. P.K. Bhartia, J.R. Herman, R.D. McPeters, O. Torres, in preparation.
8. R.D. McPeters and W.D. Komhyr, J. Geophys. Res. 96, 2987 (1991).
9. W.D. Komhyr et al., ibid. 94, 9847 (1989).
10. D.J. Hoffman and S. Solomon, ibid. p. 5029.
11. M.J. Prather, ibid. 97, 10187 (1992); G. Brasseur and C. Granier, Science 257, 1239 (1992). C. Granier and G. Brasseur, J. Geophys. Res. 97, 18015 (1992).
12. We thank J. Kerr for providing the Dobson data from the World Ozone Data Centre, Atmospheric Environment Service, Downsview, Ontario. The support of J.H. Lienech and H.D. Bowman and the NOAA Climate and Global Change Program in developing the SBUV/2 data set is gratefully acknowledged.
11 March 1993; accepted 30 March 1993