Dobson Stations Research Articles

Technical note: The US Dobson station network data record prior to 2015, re-evaluation of NDACC and WOUDC archived records with WinDobson processing software

Abstract. The United States government has operated Dobson ozone spectrophotometers at various sites, starting during the International Geophysical Year (1 July 1957 to 31 December 1958). A network of stations for long-term monitoring of the total column content (thickness of the ozone layer) of the atmosphere was established in the early 1960s and eventually grew to 16 stations, 14 of which are still operational and submit data to the United States of America's National Oceanic and Atmospheric Administration (NOAA). Seven of these sites are also part of the Network for the Detection of Atmospheric Composition Change (NDACC), an organization that maintains its own data archive. Due to recent changes in data processing software the entire dataset was re-evaluated for possible changes. To evaluate and minimize potential changes caused by the new processing software, the reprocessed data record was compared to the original data record archived in the World Ozone and UV Data Center (WOUDC) in Toronto, Canada. The history of the observations at the individual stations, the instruments used for the NOAA network monitoring at the station, the method for reducing zenith-sky observations to total ozone, and calibration procedures were re-evaluated using data quality control tools built into the new software. At the completion of the evaluation, the new datasets are to be published as an update to the WOUDC and NDACC archives, and the entire dataset is to be made available to the scientific community. The procedure for reprocessing Dobson data and the results of the reanalysis on the archived record are presented in this paper. A summary of historical changes to 14 station records is also provided.

An assessment of the stray light in 25 years of Dobson total ozone data at Athens, Greece

Abstract. In this study, we investigated the susceptibility of the Dobson spectrophotometer No. 118 to stray light interference. In this regard, a series of total ozone content measurements were carried out in Athens, Greece for air-mass values (μ) extending up to μ = 5. The monochromatic-heterochromatic stray light derived by Basher's model was used in order to evaluate the specific instrumental parameters which determine if this instrument suffers from this problem or not. The results obtained indicate that the measurements made by the Dobson instrument of the Athens station for air mass values up to 2.5, underestimates the total ozone content by 3.5 DU in average, or about 1 % of the station's mean total ozone content (TOC). The comparison of the values of the same parameters measured 15 years ago with the present ones indicates the good maintenance of the Dobson spectrophotometer No. 118. This fact is of crucial importance because the variability of the daily total ozone observations collected by the Athens Dobson Station since 1989 has proved to be representative to the variability of the mean total ozone observed over the whole mid-latitude zone of the Northern Hemisphere. This stresses the point that the Athens total ozone station, being the unique Dobson station in south-eastern Europe, may be assumed as a ground truth station for the reliable conversion of the satellite radiance observations to total ozone measurements.

Total ozone observations at Athens, Greece by satellite-borne and ground-based instrumentation

A comparison is made of total ozone (TOZ) content observations conducted by the Dobson spectrophotometer No. 118, the SCanning Imaging Absorption SpectroMeter for Atmospheric CHartographY (SCIAMACHY), the Total Ozone Mapping Spectrometer (TOMS) and the Ozone Monitoring Instrument (OMI) over Athens, Greece, during 1991–2008. Spearman's and Wilcoxon's tests were used to determine the measure of the agreement between the ground-based and satellite column ozone data. The correlation coefficient between Dobson and Nimbus-7, ADEOS, Earth Probe, OMI and SCIAMACHY observations was found to be 0.95, 0.96, 0.94, 0.93 and 0.87, respectively, while the correlation coefficient between total ozone observations of SCIAMACHY and Earth Probe-TOMS and OMI is 0.85 and 0.93, respectively. SCIAMACHY overestimates the column ozone with respect to Dobson, Earth Probe-TOMS and OMI by 10, 15 and 3 DU, respectively, while Dobson underestimates the column ozone with respect to Nimbus-7, ADEOS and OMI by 5, 10 and 8 DU. The results obtained confirm that the Athens Dobson station may continue to be considered as a ground-truth total ozone station for the validation of the satellite column ozone observations. In addition, linear regression analysis of the deseasonalized monthly mean column ozone, as derived from Dobson measurements, gives an increase of +0.33 ± 0.07% per year during 1991–2000 and a decrease of –0.33 ± 0.07% per year for the period 2001–2008.

Comparison of total column ozone data from ground-based MFRSR measurements taken in the southern U.S. region with OMI satellite retrievals

[1] A study of total column ozone (TCO) measurement data taken with a network of ground-based MultiFilter Rotating Shadowband Radiometers (MFRSR) compared with data retrieved from the satellite-based Ozone Monitoring Instrument (OMI) shows an average positive bias of 3.04% ± 1.7% under all sky days. The MFRSR station-to-station standard deviation was ±2.1%. In contrast, a recent 2-year comparison of ground-based Dobson and Brewer stations shows that OMI-TOMS TCO averages only 0.4% higher than the station average, with station-to-station standard deviation of ±0.6%. Annual seasonal variability in the TCO is manifested through an increase beginning in December and going through May before the levels decrease over the northern summer until reaching the lowest levels around October. In this study, positive biases were found for the 11 southern U.S. MFRSR monitoring sites after an analysis of 2 years of data from January 2006–December 2007. Overall, our comparisons show that there is good agreement between OMI-TOMS and MFRSR readings over the southern U.S.

Statistical reconstruction of daily total ozone over Europe 1950 to 2004

A statistical model is developed to reconstruct the total ozone field over Europe since 1 January 1950. Daily total ozone data are required for surface UV reconstruction. The model is trained on satellite data (National Institute of Water and Atmosphere Research (NIWA) assimilated total column ozone database) over the period 1979–2004. The model explains ∼80% of variance of the satellite total ozone. Standard deviation of the relative differences, observed minus model daily data in percent of observed data, is ∼4% and ∼6% in May‐September and October‐next year April season, respectively. The total ozone values are reconstructed backward in time using the regression constants derived as a result of the training, time series of various atmospheric indices of the atmospheric circulation, and the meteorological variables. The proxies for the ozone variations are taken from the National Centers for Enivronmental Prediction (NCEP)/National Center for Atmosphere Research (NCAR) reanalysis database. Optimal combination of the proxies is selected using MARS technique. The quality of the database is assured by a comparison of the reconstructed total ozone with the ground‐based data from several Dobson stations functioning in the early 1950s and 1960s. The model explains ∼70–80% variance of the ozone data collected before the satellite era. Bias and the long‐term drift between the reconstructed and measured Dobson ozone are within a range of ±2%. Standard deviation of relative differences between reconstructed and station Dobson daily data is 5–7% and ∼10% in May‐September and October‐next year April season, respectively. The long‐term tendency, comprising an anthropogenic trend and superposition of “natural” long‐term fluctuations, appears almost trendless in the period 1950–1980 with a clear ozone decline after 1980.

Total ozone influence on the surface UV‐B radiation in the late spring‐summer 1963–1997: An analysis of multiple timescales

Monthly means and minima of total ozone for the late springs and summers (May–August) of 1963–1997 have been examined for the European Dobson stations (Arosa, Belsk, Hohenpeissenberg, Hradec Kralove, Uccle). It is shown that long‐term tendencies in total ozone means were almost similar to those in the total ozone minima. Analyses of the late spring/summer means of UV daily doses, total ozone, and global solar radiation (proxy for the overall atmospheric transparency), measured at Belsk (52°N, 21°E) for the period 1976–1996, show that an importance of the total ozone changes for the UV‐B level increases with the timescale. Decadal variations in total ozone are the main source of the UV trend at Belsk. Frequency of appearance of extreme daily total ozone values in the selected late spring/summer season seems to be important for analyses of the ozone forcing in the interannual timescale. Regional and temporal differences in the number of days with extreme low ozone values are discussed using the total ozone extrema taken at Arosa, Belsk, and Hradec Kralove in the 1963–1997 period. A statistical model is developed for diagnosis of the next day value of the UV‐B level. The changes in the overall atmospheric transparency are essential for the UV‐B level when the day‐to‐day variations in the UV forcing factors are examined.

Validation of NOAA‐9 SBUV/2 total ozone measurements during the 1994 Antarctic Ozone Hole

The Solar Backscatter Ultraviolet (SBUV/2) instrument on the NOAA‐9 spacecraft made total ozone measurements over Antarctica during the 1994 Austral Spring depletion. These measurements continue those made by the SBUV/2 on NOAA‐11. In recent years NOAA‐9 drifted from a poor orbit to one where earth observations and calibration capabilities are not restricted. An interim calibration of the NOAA‐9 SBUV/2 instrument was established with data from June 1994 and applied to observations during September, October, and November 1994. To validate the NOAA‐9 ozone measurements, daily zonal ozone averages from NOAA‐9 and NOAA‐11 measurements in the Northern Hemisphere have been compared. Comparisons have also been made with ground‐based measurements from five Dobson stations dispersed on the Antarctic continent. The results show that, on average, the NOAA‐9 data agree to within 1–2 percent with the Dobson stations with standard deviations of the difference ranging from 5.3 to 7.7% and to within several percent with NOAA‐11 Northern Hemisphere data when restricted to solar zenith angles less than 80 degrees. This agreement makes possible not only a continuation of the Antarctic measurements without a large instrument‐related bias, but also establishes the NOAA‐9 data as a suitable transition data set during the replacement of NOAA‐11 by NOAA‐14.

Total ozone trends from quality‐controlled ground‐based data (1964–1994)

Seasonal and year‐round trend analyses of ground‐based total ozone data that have been carefully quality controlled are presented for 46 Dobson stations, for four regions of the former USSR and for the two polar regions. The trend model incorporated the 10.7‐cm solar flux and equatorial stratospheric wind as indicators for removal of the solar and quasi‐biennial components in ozone variations. The trends were calculated for two main time intervals: from January 1964 through March 1994, with slope since 1970, and from January 1979 through March 1994. The analyses show continuous year‐round ozone decline in middle latitudes (35°–60°) where the estimated values are −4.3 and −4.1% per decade for the period since January 1979 in the northern and southern hemispheres respectively. The northern hemisphere data permit estimation of the trends over vast regions such as North America −4.1±1.5%, Europe −5.1±2.0%, Siberia and Far East −5.6±1.8% (values and two sigma intervals are given in percent per decade for year‐round trend). The ozone decline over the Arctic is about −5.6±2.0% per decade for year‐round trend and −7.5±3.8% for the winter‐spring season. The most dramatic ozone decline occurs in spring over Antarctica, −22±7.2% per decade. The trends estimated for the 1979–1994 period are about 1.5% per decade stronger in northern and southern middle and high latitudes than the trends with slope since 1970, which shows a substantial acceleration of the rates of ozone decline. The sensitivity of the ozone trends to the extremely low ozone during the 1992–1993 period is discussed. Most of the ozone data and calibrations in this paper were the base for the Ozone Trends section of the World Meteorological Organization/United Nations Environment Program Ozone Assessment 1994. The results presented here in general concur with the assessment and provide further estimates of the ozone changes in the northern middle latitudes in view of another extremely low winter‐spring season of 1994–1995.

Variations of the total ozone trend over the northern hemisphere midlatitudes in winter‐spring seasons 1963–1964 through 1992–1993

Monthly means of total ozone for winter‐spring seasons (December through the next March) since 1963–1964 through 1992–1993 from a network of 19 Dobson stations are averaged to form a composite time series of the winter‐spring midlatitudinal total ozone. A multiple regression statistical model, comprising overall mean level, linear trend, quasi‐biennial, El Niño, solar (including both the 11‐year and its double length “geomagnetic” oscillations), and dynamics of the winter polar vortex terms, is applied to estimate interannual variations of total ozone over the northern hemisphere midlatitudes in winter‐spring seasons. The model used explains about 88% of total variance of the winter‐spring total ozone and gives the value of the anthropogenic trend in the northern hemisphere midlatitudinal total ozone of the order of −3.5%±0.4% per decade since winter‐spring of 1969–1970. The model's results support the hypothesis that the increase of the ozone loss during the 1980s and 1990s over the northern hemispheric midlatitudes may be partially caused by a superposition of “natural” long‐term oscillations in total ozone related to the climatic changes in the atmospheric dynamics. The initial evidences for a strong manifestation of the interannual changes of the polar vortex in the winter‐spring midlatitudinal total ozone data are found. The 22‐year oscillation in the total ozone data is disclosed; however, because of the limited length of the ozone data, one should be cautious to relate this fluctuation uniquely to the solar input.

On the phase propagation of extratropical ozone quasi‐biennial oscillation in observational data

Global column ozone data from total ozone mapping spectrometer (TOMS), backscattered ultraviolet (BUV) and Dobson stations are analyzed to determine the pattern and phase property of the ozone quasi‐biennial oscillation (QBO) signal. It is found that the ozone QBO signal is strongest in middle and high latitudes and is present mainly in the winter‐spring season in both hemispheres. The extratropical ozone QBO signal is out of phase with the equatorial ozone QBO, which is itself in phase with the QBO in equatorial zonal wind. There are three distinctive regions, namely tropical, midlatitudinal, and polar regions, in each of which the ozone QBO signal has a fairly constant phase with respect to latitude. There is a phase reversal (sign change) between the equatorial and the extratropical regions associated with the return branch of the equatorial QBO secondary circulation, and this sign reversal occurs at ±12° of latitude symmetric about the equator. In the northern hemisphere between the midlatitudinal and polar regions, there is another possible phase reversal in some (but not all) years possibly in connection with the presence or absence of midwinter sudden warming, which creates a positive or negative anomaly relative to the region outside the polar vortex. In the southern hemisphere polar latitudes, the ozone QBO signal is usually delayed until spring in connection with the final warming. These properties are found in all data sets analyzed by the same method. Evidence does not support a gradual phase propagation from the subtropical region to the high‐latitude region. Previous reported evidence for phase propagation is reexamined and is found to be artifacts of data processing.

Total ozone changes over Eurasia since 1973 based on reevaluated filter ozonometer data

Since the early 1960s, on the vast territory of the former USSR, 45 stations have been in continuous operation, utilizing the broadband filter M‐83 ozonometer. The quality of the ozone data during the first decade was unsatisfactory. After 1972 an improved version of the ozonometer was introduced together with improved quality control practices, including methodology of observations. The more reliable data of 1973 through March 1994 have been rigorously reexamined by applying variability analysis, comparison with lower‐stratosphere temperatures and/or nearby Dobson stations, and overpassing TOMS for identifying concurrence or discrepancies. These control procedures together with the information on instrument relocation and calibrations made it possible to reevaluate the record of all 45 stations. The accuracy of the improved ozonometer data is about 3% for direct Sun measurements and ∼5% for zenith sky observations; although not so good as that of the Dobson, in the long run it provides consistent ozone data sets. This data set is now made available to the World Ozone Data Center (WO3DC), Toronto. Thus for the first time, based on a 21‐year long record, information is deduced on the differences in the ozone annual cycle between Eastern Siberia and the European part, on the strong appearance of quasi‐biennial oscillation (QBO) signals especially pronounced as ozone deficiency during the western phase of the QBO, on the ozone variability, and on the long‐term changes over the huge territory from Central Europe to the Far East. The specifics of the ozone changes considered in concurrence with the prevailing general stratospheric circulation conditions permitted us to distinguish four broad regions with consistent ozone regimes. The appearance of the strongest northern hemisphere ozone maximum and a monthly mean of ∼470 matm cm over Siberia during winter‐spring, compared with ∼400 matm cm over Europe, the occurrence of the ozone annual minimums as early as August over Eastern Siberia, compared with October over the European part, was established and is probably related to the specifics of the atmospheric circulation patterns. The long‐term ozone changes are considered in relation to the stratospheric temperature at 100‐hPa. For each deviation from the monthly normal temperature by 1°C, there is a corresponding change by the same sign 5–6 matm cm in the monthly ozone deviations. The calculated long‐term ozone trends for 1973 through March 1994 (given in percent per decade ±2σ) are as follows: European part, −3.6±0.8; Central Asia, −2.0±0.6; Western Siberia, −3.5±0.8; and Eastern Siberia ‐ Far East, −3.2±0.8. They show a steady ozone decline similar to that deduced from the Dobson stations at the same latitudes. The decline during the past 15 years is stronger and, in general, concurs with Total Ozone Mapping Spectrometer (TOMS) trends which are, however, slightly more negative.

Statistical significance and pattern of extratropical QBO in column ozone

Statistical tests are performed on 30 years (1957–1986) of Dobson station data. The triple‐period (30, 20, and 8.6 months) extratropical QBO signal in ozone is found to be statistically significant. Earlier tests, based on 9 years of TOMS data, gave ambiguous results. We further suggest an objective method of extracting ozone QBO signal in the extratropics using a linear regression model.

Changes in ozone trends over the northern hemisphere middle latitudes during the period 1970–1990

Composite time series combining the results of total ozone measurements taken at Dobson stations located within the latitude band 30°N–60°N, in Europe, and North America, have been examined in order to detect any trends. Various regression trend models were used to identify any trend variations over the regions during the period 1970–1990. The results of fitting the models to the data imply that the model which assumes a linear trend provides precise information about the long-term ozone trends (trends during the period 1970–1990). The study identifies short-term summer trends in the 1980s that are evidently more strongly negative than trends that occur in the 1970s (the differences are statistically significant at the 2σ level). The year-round loss (in all analyzed regions) and the winter loss in total ozone (the belt 30°N–60°N) N. America, during the 1980s are about 2–3 times higher than the losses during the 1970s (the differences are statistically significant at the 1σ level).

A comparison of total ozone data from satellite and ground‐based observations at northern latitudes

Ozone measurements from the Total Ozone Mapping Spectrometer on the Nimbus 7 satellite are compared with ground‐based measurements from five Dobson stations at northern latitudes to evaluate the accuracy of the TOMS data, particularly in regions north of 50°N. The measurements from the individual stations show mean differences from −2.5% up to +8.3% relative to TOMS measurements and two of the ground stations, Oslo and Longyearbyen, show a significant drift of +1.2% and +3.7% per year, respectively. It can be shown from nearly simultaneous measurements in two different wavelength double pairs at Oslo that at least 2% of the differences result from the use of the CC' wavelength double pair instead of the standard AD wavelength double pair. Since all Norwegian stations used the CC' wavelength double pair exclusively, a similar error can be assumed for Tromsø and Longyearbyen. A comparison between the tropospheric ozone content in TOMS data and from ECC ozonesonde measurements at Ny‐Ålesund and Bear Island shows that the amount of tropospheric ozone in the standard profiles used in the TOMS algorithm is too low, which leads to an error of about 2% in total ozone. Particularly at high solar zenith angles (>80°), Dobson measurements become unreliable. They are up to 20% lower than TOMS measurements averaged over solar zenith angles of 88° to 89°.

Global average ozone change from November 1978 to May 1990

A recent recalibration and reprocessing of the total ozone mapping spectrometer (TOMS) data have made possible a new determination of the global average (69°S to 69°N) total ozone decrease of 3.5% over the 11‐year period, January 1, 1979, to December 31, 1989, with a 2σ error of 1.4%. The revised TOMS ozone trend data are in agreement, within error limits, with the average of 39 ground‐based Dobson stations and with the world standard Dobson spectrometer 83 at Mauna Loa, Hawaii. Superimposed on the 11‐year ozone trend is a possible solar cycle effect, quasi‐biennial oscillation (QBO), annual, and semiannual cycles. Using solar 10.7‐cm flux data and 30‐mbar Singapore wind data (QBO), a time series has been constructed that reproduces the long‐term behavior of the globally averaged ozone. Removal of the apparent solar cycle effect from the global average reduces the net ozone loss to 2.66±1.4% per decade. The precise value of the global average ozone trend depends on the latitude range selected, with ranges greater than ±69° emphasizing the larger variations at high latitudes.

Long‐term changes in the total ozone mapping spectrometer relative to world primary standard Dobson spectrometer 83

We have examined the stability of the calibration of the Nimbus 7 solar backscatter ultraviolet (SBUV) and total ozone mapping spectrometer (TOMS) instruments by comparing their ozone measurements with those made by a single, very stable Dobson instrument: the world primary standard Dobson spectrometer number 83. Measurements of ozone made with instrument 83 at Mauna Loa observatory in eight summers between 1979 and 1989 were compared with coincident TOMS ozone measurements. The comparison shows that relative to instrument 83, ozone measured by TOMS (and SBUV) was stable between 1979 and approximately 1983, had decreased by 3% by 1986, and had decreased by almost 7% by 1989. A similar time dependence is seen when data from an ensemble of 39 Dobson stations throughout the world is compared with TOMS over the period 1979–1987. The most likely reason for the relative drift is that the diffuser plate used by both SBUV and TOMS to measure solar flux has suffered an uncorrected wavelength dependent degradation, with most of the degradation occurring after 1983. The recently released version 6 TOMS data, corrected using the internal “pair justification” technique, show almost no drift relative to Dobson instrument 83. We conclude from these comparisons that accurate measurements of long‐term global ozone change will require a coherent system incorporating both ground based and satellite based ozone measurements.

An assessment of the long‐term drift in TOMS total ozone data, based on comparison with the Dobson Network

Using techniques very similar to those used in comparing SBUV total ozone with Dobson [Fleig et al. 1986A], seven years of total ozone data derived from the TOMS instrument on Nimbus‐7 are compared with results from 41 Dobson stations. In this study however, a significant improvement in fit is obtained using a model which assumes a change in the rate of drift between TOMS and Dobson around the middle of 1982. Results indicate that the TOMS measurements have drifted relative to the Dobson stations with a two‐part linear trend of −0.25 ± 0.17 % per year during the period from launch to 6/30/82, and −0.51 ± 0.21 % per year during 7/1/82 ‐ 10/31/85. As discussed in Fleig et al. [1986A], we are not yet able to definitively separate the causes of this drift between residual uncorrected drift in the TOMS instrument (a similar drift is apparent in the SBUV‐Dobson comparisons), limited sensitivity of the TOMS to increases in tropospheric ozone, and the effect of local increases in pollution levels on individual Dobson stations.

Comparison of ground‐based and total ozone mapping spectrometer measurements used in assessing the performance of the global ozone observing system

Uniformly reevaluated Nimbus 7 Total Ozone Mapping Spectrometer (TOMS) total ozone data (1978–1985) are compared with data from ground‐based stations during the daily overpasses of the satellite. Existence of a drift between the two systems is identified, showing the TOMS‐derived total ozone to be decreasing nonuniformly by an average of about 0.4% per year. After removing the individual monthly biases and using the satellite as a transfer standard, the differences between 92 currently operating stations of the Global Ozone Observing System (GO3OS) are analyzed. About two thirds of all Dobson stations show differences of less than 2%, while only half of the M‐83 filter stations fall into this good station category; close to 20% of the Dobson stations and nearly 30% of the filter stations show differences greater than 3%. Several examples of individual station discrepancies are discussed in order to encourage the complete reevaluation of the ground‐based ozone record and to forewarn potential users from taking the published ozone data at their face value.

An analysis of the 7‐year record of SBUV satellite ozone data: Global profile features and trends in total ozone

An analysis of Nimbus 7 solar backscattered ultraviolet (SBUV) monthly average total ozone data for changes over the 7‐year period from November 1978 through September 1985 is considered. Regression‐time series models, which include seasonal components, a linear trend term, an F10.7 solar flux term, and an autocorrelated autoregressive noise term to account for variations in total ozone, are estimated for both latitudinal zonal averages and a global series over this time period. The resulting linear trend and solar flux coefficient estimates are examined. A comparison between SBUV monthly average total ozone data near Dobson ground station locations and the corresponding Dobson station total ozone data is performed for a network of 35 Dobson stations. Linear drifts are estimated for the differences between SBUV and Dobson total ozone data at each station. The results show an average negative linear drift in SBUV data relative to Dobson data of about −0.4% per year. A linear trend estimate for the global SBUV series of (−0.74±0.26)% per year is obtained for the model which takes into account the association of ozone changes with solar flux variations. When this linear trend estimate for the global SBUV series is “corrected” for the negative drift between SBUV and Dobson data, the 95% confidence interval estimate of the linear trend component in the global SBUV series, exclusive of trend variations associated with solar flux variations, over this 7‐year period is (−0.35±0.28)% per year. The global estimate of the relation between total ozone and solar flux variations is (0.97±0.61)% per 100 solar flux units, which represents a change in total ozone over this 7‐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 data at the Dobson station network of locations over this 7‐year period are quite similar to trend estimates obtained from the global SBUV series. When comparing the above SBUV data global trend estimate for the 7‐year period with a linear trend estimate of (−0.04±0.07)% per year obtained from a network of Dobson station data over the longer period 1970–1985, we find that the trend estimate for the recent 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 the period 1982–1985, no firm conclusions can be drawn in terms of interpretation of the negative trend finding over the relatively short 7‐year time period. Such interpretations require a detailed understanding of the atmospheric dynamical and chemical mechanisms related to ozone variations during this period.

Seasonal Differences in the Trend of Total Ozone and Contributions from Tropospheric and Stratospheric Layers

Based on an average of the total-ozone changes determined by means of linear regression at individual Dobson stations within climatic zones, trends of total ozone for each of the four seasons have been evaluated for five climatic zones, and the world; intervals are 1960–84, 1965–84, 1970–84 and 1975–84. Based on ozonesonde data, trends of ozone by season in 2–8 km, 8–16 km, 16–24 km and 24–32 km layers have been evaluated for the four extratropical climatic zones for intervals 1965–84, 1970–84 and 1975–84. The decrease in September–October–November (SON) total ozone relative to other seasons in the south polar zone (Antarctica) is observed in the last 10–15 years also in south temperate, tropical and north polar zones, and for the world as a whole, but the seasonal differences in trend are significant only in the south polar zone. In the north temperate zone, where the data are most numerous, the total ozone has decreased in December-January-February (DJF) relative to other seasons, although in the last decade the trends in DJF and SON an comparable. The decrease in SON total ozone relative to other seasons is indicated to be due mainly to an ozone decrease in the 16-24 km layer of the south polar zone and the 24-32 km layer of south temperate and north polar zones, but once again the seasonal differences in trend are not significant. There are inconsistencies in the north temperate zone between relative seasonal changes in total ozone and layer-mean ozone, casting doubt on the representativeness of the ozonesonde data. There has been a substantial decrease of SON temperature in the 16–20 km layer of the south polar zone, presumably associated with the ozone decrease in this region in this season. The global surface temperature has also decreased in SON relative to other seasons, perhaps associated with the relative decrease of global total ozone in this season during the last decade.