JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 93, NO. D2, PAGES 1689-1703, FEBRUARY 20, 1988 An Analysis of the 7-Year Record of SBUV Satellite OzoneData: Global Profile Features and Trends in Total Ozone GREGORY C. REINSEL, • GEORGE C. TIAO, 2 SUNG K. ABe, • MARIAN PUGH, • SABYASACHI BASU, • JOHN J. DELUISI, 3 CARL L. MATEER, 4 ALVIN J. MHLER, s PETER S. CONNELL, 6 AND DONALD J. WUEBBLES 6 An analysis of Nimbus7 solarbackscattered ultraviolet (SBUV) monthlyaverage total ozonedata for changes over the7-year period from November 1978 through September 1985 is considered. Regression-time series models, which include seasonal components, a linear trend term, anF 10.7 solar flux term, and an auto- correlated autoregressive noise termto account for variations in totalozone, areestimated for bothlatitudinal zonal averages anda global series overthistimeperiod.The resulting linear trend andsolar flux coefficient estimates areexamined. A comparison between SBUV monthly average totalozone data nearDobson ground station locations and thecorrespond• Dobson station totalozone data is performed for a network of 35 Dob- son stations. Linear drifts are estimated for the differences between SBUV and Dobson total ozone data at each station.The results show an average negative lineardrift in SBUV datarelative to Dobson dataof about -0.4% peryear. A linear trend estimate for theglobal SBUV series of (-0.74 + 0.26)%peryearis obtained for the model which takesinto account the association of ozonechanges with solar flux variations. When this lineartrend estimate for the globalSBUV series is "corrected" for the negative drift between SBUV andDob- son data, the 95% confidence interval estimateof the linear trend component in the global SBUV series, exclusive of trendvariations associated with solarflux variations, over this 7-yearperiodis (-0.35 + 0.28)% peryear. The global estimate of therelation between total ozone and solar fluxvariations is (0.97+ 0.61)%per 100 solar flux units, which represents a change in totalozone overthis7-year period associated with solar flux variations of about (-0.20 + 0.13)% per year. Comparisons are also performed which indicate that trends obtained from SBUV dataat the Dobson station network of locations over this7-yearperiod are quite similar to trend estimates obtained from the global SBUV series. When comparing the above SBUV dataglobal trend estimate for the7-yearperiod with a linear trendestimate of (-0.04 + 0.07)% per yearobtained from a network of Dobson station data overthelonger period 1970-1985,we findthatthetrend estimate for therecent 7-year period is considerably more negative but with a much larger standard error. In recognition of the rather unusual natural atmospheric fluctuations which occurred during theperiod 1982-1985, no firm conclusions can be drawnin termsof interpretation of the negative trendfindingover the relatively short 7-yeartime period. Such interpretations require a detailed understanding of the atmospheric dynamical andchemical mechanisms related to ozone variations during this period. 1. INq3•ODU•ON The solar backscattered ultraviolet (SBUV) instrument[Heath et al., 1975] on the Nimbus 7 satellite measuresthe solar ultra- violet radiation backscattered by the Earth's atmosphere and reflected by clouds andthe terrestrial surface.The measurements are made within 12 wavelength channels (1-nm width) from 256 to 340 nm. From these measurements the total column atmos- pheric ozoneand its v•nxical profile may be estimated.This also requiressolar irradiance measurements, which are obtained by the instrument through thedeployment of a diffuser platethatdif- fusely reflects the direct solar radiation into the instrument. These SBUV data are available on magnetic tape from the National SpaceScience Data Center. A detailed presentation of thenature andquality of the SBUV measurements is given by R. P. Cebula and D. F. Heath (unpublished report, 1986). Early results from the SBUV insmunent have beendescribed by Bhar- 1 University of Wisconsin-Madison. 2University of Chicago, Chicago, glinois. 3National Oceanic and Atmospheric Administration, Boulder, Colora- do. 4Scarborough, Ontario, Canada. 5 National Oceanic and Atmospheric Administration, Washington, D.C. 6Lawrence Livermore National Laboratory, Livermore, California. Copyright 1988 by the American Geophysical Union. Paper number 7D0867. 0148-0227/88/007D-0867 $05.00 tia eta/. [ 1984a] for totalozoneandby Bhartia et al. [1984b] for verticalprofile ozone. More recently, Fleig et al. [1986] have compared SBUV total ozone measurements with thosefrom 41 Dobson ground stations and foundan approximately linearnega- tive drift of about 0.4% per year in the SBUV data relativeto the groundmeasurements. It is importantto note that the SBUV data, as archived in the National Space Science Data Center, are not calibratedwith respeatto the Dobson measurements. The SBUV total ozone data have also previously been analyzed by Stolarski et al. [ 1986], particularly with regardto recentozone decreases in Antarctica. In thispaper we report on a trendanalysis of total ozone and profileozone datafrom theNimbus7 SBUV satellite experiment for the 7-year period from November 1978 throughSeptember 1985,with themainemphasis on the totalozone data wends. The data consist of measurements of the total column ozone amount and profile measurements of ozone for Umkehr layers 1-12. (SeeMcPeters et al. [1986] for a description of the Umkehrlayer system, and also, notethat approximate altitude ranges for layers 4-10 are indicatedin Figure 1.) In our analysis the trend esti- matesthat will be obtained represent a rate of linear change in ozone over thisparticular 7-year time period. The effects of solar flux variations on ozone changes over this period will also be considered, and in these cases the corresponding trendestimates represent the trendcomponent of ozonethat is not accounted for by the association of ozonechanges with solarflux variations. It must be acknowledged at the outsetthat the time span of the SBUV data set is relatively short, and hence results of any analysis of these data for longer-term trends must be interpreted with caution.Moreover, the possible influence on theozone data 1689