Information compiled by the BADC team.
The UGAMP ozone climatology has been set up by Dingmin Li and Keith P. Shine at the Department of Meteorology of the University of Reading. It is a four-dimensional distribution of atmospheric ozone that has been built up from the combination of several observational data sets.
The climatology was originally conceived to replace the old ozone distributions used as an input in the UGAMP models, namely the built in ECMWF parameterization, hereafter called the internal climatology, and a two-dimensional climatology based on satellite data (Thuburn, 1992), hereafter called external climatology. The weakness of the ECMWF parameterization was an overestimation of the low polar stratosphere ozone content, leading in the UGAMP general circulation model (UGCM) to a biased calculated zonal mean dynamics in polar regions, while the 2-D climatology, although realistic, naturally missed the longitudinal variation. Moreover, the new climatology would ideally have to give account for the ozone interannual variability (e.g. the QBO signal). Therefore, a multi-year three-dimensional climatology was required, based on the most comprehensive observations available. The approach that has been adopted in this endeavour is summarized in the next sections. More details can be found in Li and Shine, 1995.
The resulting climatology should be useful not only as a realistic prescribed ozone distribution for global circulation models, but also as a reference for validation of GCM that treat ozone as a prognostic variable.
The data consists in ozone column (1) monthly means above the grid levels, expressed in Dobson Units (1) (DU). A routine is provided that converts ozone columns into volume mixing ratios.
The geographical coverage of the climatology is global, in spite of the existence of sparse observation areas in the original data sets, that have been the object of a particular treatment (troposphere and polar night). The climatology horizontal resolution is of 2.5 by 2.5 degrees (144 points in longitude and 73 points in latitude from pole to pole). Its vertical coverage extends over 47 levels defined by isobaric surfaces, from the 0.0011 mb level down to the standard pressure level. The vertical resolution is irregular it is close to 2 km in the stratosphere.
These choices have been made according to the resolution adopted in the UGCM. However, a routine is provided that converts the climatological data onto any other desired grid and/or vertical scale.
The present state of the climatology includes five years (1985 to 1989), with one 3-dimensional data set per month.
At the time of its construction, the need of a multi-year climatology put a constraint on the possible data sets to be used. On the other hand, the latitude and height ranges of the designed climatology made it necessary to combine different sources of data, which would have to be homogenized in some way. As a result, the climatology is based mainly on satellite data from SBUV and SAGE II. Supporting data comes from ozone-sondes (Atmospheric Environment Service of Canada) and other satellite instruments, such as SME and TOMS. Data from the UARS were not included, as they were not available at the time when the project began.
SBUV (Fleig et al., 1990; WMO, 1988) measures the solar irradiance and the radiance backscattered by the Earth's atmosphere in 12 selected wavelength bands in the ultraviolet. From these radiances and irradiances, ozone columns (1) and ozone profiles are obtained, using retrieval algorithms. SBUV views the nadir with a field of view of 200 x 200 kilometers. Every day, the observation covers 200 km wide strips, 26° apart in longitude. There are 13 or 14 orbits per day. Data is available from November 1987 to June 1990. The SBUV data used in the current climatology is version 6.
SAGE II (McCormick, 1992; Zawodny and McCormick, 1991) is a seven-channel spectrometer covering selected wavelengths in the near UV, visible and near IR portion of the spectrum. The wavelengths were selected so that determination of aerosol extinction, ozone, nitrogen dioxide and water vapour profiles can be issued throughout the stratosphere and upper troposphere. SAGE II measures the Earth's limb extinction via the solar occultation technique during each spacecraft sunrise and sunset. This technique is inherently self-calibrating in that the instrument makes a direct measurement of the extraterrestrial solar irradiance before and after each sunrise and sunset occultation event. The 15 orbits per day yield 15 sunrise and 15 sunset profiles. The use of the sun as an irradiance source allows for a small instrument field of view and gives a very high vertical resolution. The data in a given day is uniformly spaced in longitude and closely spaced in latitude, which periodically ranges from 80° South to 80° North. Data is available for October 1984 to early 1990.
SBUV and SAGE II use very different spectral measurement techniques: nadir-viewing backscattered ultraviolet versus limb-viewing solar occultation. The vertical resolution of the two measurements is accordingly quite different. SBUV has an approximately 8-km resolution in the upper stratosphere but only 15-km resolution at and below the ozone maximum, whereas SAGE II gives a near 1-km vertical resolution. By virtue of its limb-view self-calibration for every profile, SAGE II may, in principle, be regarded as a benchmark standard against which the coarse resolution SBUV profiles may be judged (McPeters et al., 1994). The SAGE II data was used as the a priori climatology in the SBUV retrieval in the low to middle stratosphere layers, leading to retrieved SBUV profiles that follow the shape of the SAGE II profiles in this region though they do not necessarily agree with SAGE there. However, SAGE II measures ozone at a given latitude only once every month, providing very poor spatial and temporal coverage for the purpose of building a global climatology; whereas SBUV provides daily measurements on a global basis. SBUV has therefore been chosen as the main data set to establish the present climatology, at pressures between 1 and 70 mb, with SAGE II used as a supplement in the low stratosphere (at pressures higher than 70 mb) and in the mesosphere (at pressures lower than 1 mb). SAGE II is also used as a standard against which the quality of the climatology is checked. The comparison of profiles between SBUV and SAGE II by McPeters et al. (1994) shows that both observations agree well in most stratospheric layers, with the exception of the high stratosphere (between 40 and 50 km) and low stratosphere.
The ozone in SME data (Rusch et al., 1983; 1984) is measured by both UV and near infrared airglow instruments; the latter gives ozone measurement from 50 to 90 km. Observation is made at the Earth's limb with an altitude resolution of about 4 km. Only the near infrared data has been used in the climatology as both the UV and near IR techniques provide similar results and as UV ozone data is available over a smaller vertical range.
TOMS is an ozone-mapping instrument that was mounted next to the SBUV instrument on the Nimbus-7 satellite (Heath et al., 1975). It provided continuous measurements of total ozone (1) on a longitude-latitude grid. Like SBUV, TOMS uses a nadir-view spectral measurement technique and works on the principle of backscattered ultraviolet radiation. However, whereas SBUV observes solar radiation backscattered only in the nadir, TOMS scans across the orbital track, sampling radiation backscattered from swaths that pass from side to side through the nadir. In principle, TOMS can obtain the signal due to the presence of ozone in both the stratosphere and the troposphere, but in practice it only measures the ozone column (1) above the cloud top, with a correction from climatological ozone below the cloud during the retrieval. The uncertainty induced by this correction is found to be small (about 5 DU) in the monthly mean measurement (Thompson et al., 1993). The accuracy of the TOMS data is considered very high in comparison to the ground based Dobson data (WMO, 1988), with errors in the monthly mean measurements less than 1.5%. The TOMS data version 6 used in this climatology is from the BADC. It includes a correction of the long term drift (1978-1987) due to the degradation of the diffuser plate in the instrument (Watson et al., 1988).
The ozone-sonde data acquired from the Atmospheric Environment Service of Canada (AESC) consists in a set of ozone profiles at each station of the network, at times when the observation was made. There is a total of 289 stations, some of which stopped operations long ago or have been in operation during a few years only, when not only a few months. These make the data discontinuous in both time and space. Three-dimensional ozone values could however be derived for the troposphere and part of the low stratosphere, based on the 2-dimensional (latitude-height) AESC ozone-sonde climatology carefully established by London and Liu (1992), and using a technique described in next section, that was developed for this purpose.
As all data sets were supplied on different vertical grids and had different vertical resolutions, an accurate interpolation scheme had to be established in order to translate them onto the climatology grid whilst conserving the total ozone amount. As SAGE data have a high vertical resolution (of about 1 km), the error is negligible when these data are degraded to lower vertical resolution. However, for SBUV, which was chosen as the main data set for this climatology, the vertical resolution is poor, as mentioned in previous section. An interpolation scheme was therefore needed which must ensure the interpolated vertical profiles agree with the observed ones whatever the resolution of the latter. It was also required that the interpolated profiles should be smooth and agree with the observations both in terms of local mixing ratio and of ozone column (1). In this endeavour, SAGE II profiles were regarded as a benchmark standard and were used to investigate the interpolation scheme first. The SAGE II profiles were then degraded to the 12 levels of SBUV and the scheme was tested again to check its meeting with the requirements listed above. After comparison of various methods, the cubic spline interpolation of the ozone column was found to be the best scheme (linear schemes, in particular, all gave unrealistic wiggles in the mixing ratio profiles).
Given the level of 100 mb as a boundary, the difference between TOMS total ozone column (1) and the ozone column (1) above the 100 mb level as determined from the combination of SME, SBUV and SAGE II provides a value for the remaining slab S1 between the surface and the 100 mb level at the centre (λ,φ) of every 2.5° × 2.5° grid cell:
On the other hand, there are 6 layers between the surface and 100 mb in the climatology of London and Liu, 1992. Summing up the ozone amounts in these 6 layers gives a value S2(φ) of the same surface-100 mb slab, on a latitudinal grid from pole to pole.
Each horizontal grid point is then attributed a value of the ratio
Multiplying each ozone mixing ratio from the 2-D ozone sondes climatology by its associated γs provides a 3-dimensional tropospheric ozone field with profile shapes consistent with the ozone sondes and vertically integrated amounts consistent with TOMS. The adjustment performed in this way is generally small (0.8<γ<1.8) and the transition of the derived ozone profiles from troposphere to low stratosphere is smooth.
|above 0.24 mb||SME near infrared airglow data|
|0.24 - 1 mb||SAGE II|
|1 - 70 mb||SBUV|
|70 - 100 mb||SAGE II|
|below 100 mb||AESC Ozone Sondes|
The climatological ozone zonal mean has also been compared with the existing UGAMP external climatology. The general features in both fields agree very well (see Li and Shine, 1995).
The first experiment (Run 1 versus Run 2) was a 240-day integration of the model for perpetual January conditions : Run 1 used the climatological ozone of January 1986. The second experiment (Run 3 versus Run 4) was a calculation performed in seasonal cycle mode: Run 3 used the climatological ozone of Year 1986. The characteristics of the performed calculations are summarised in Table 2.
|3-D combined climatology||Internal ECMWF parameterization
|Mode||Perpetual January||Run 1||Run 2|
|Seasonal cycle||Run 3||Run 4|
In the case of perpetual January, the two ozone climatologies show significantly different values in the low stratosphere: the ECMWF parameterized ozone is larger than its climatological counterpart in both polar regions (by up to 3 ppmv in the summer hemisphere), shows an unrealistically pronounced vertical gradient in the tropical region and is lower at northern mid-latitudes. As a result, by comparison to Run 1, the low stratosphere in Run 2 is warmer by up to 10 K at summer high latitudes and cooler above the tropical tropopause and at Northern mid-latitudes.
The seasonal cycle experiment reveals strong discrepancies between Run 3 and Run 4 in terms of zonal mean temperature: the polar low stratosphere in Run 4 is warmer than in Run 3 by up to 10 K, in both summer and winter. Since the internal variability in the two runs is very similar both in distribution and in amplitude, the difference in the calculated temperature fields must be entirely attributed to the difference in the ozone amounts.
These experiments underline the sensitivity of the calculated temperature to the adopted ozone amount and emphasize the importance of choosing a realistic ozone climatology as an output to GCM.
Fishman, J., E. C. Watson, J. C. Larsen and J. A. Logan, Distribution of tropospheric ozone determined from satellite data, J. Geophys. Res., 95, D4, 3599-3617, 1990.
Fleig, A. J., R. D. McPeters, P. K. Bhartia, B. M. Schlesinger, R. P. Cebula, K. F. Klenk, S. L. Tayoy and D. F. Heath, Nimbus 7 Solar Backscatter Ultraviolet (SBUV) ozone products user's guide, NASA Reference Publication 1234, 1990.
Heath, D. F., A. J. Krueger, H. A. Roeder and B. D. Henderson, The solar backscatter ultraviolet and total ozone spectrometer (SBUV/TOMS) for Nimbus G, Opt. Eng., 14, 323-331, 1975.
London, J. and S. C. Liu, Long-term tropospheric and lower stratospheric ozone variations from ozonesonde observations, J. Atmos. and Terres. Phys., 54, 5, 599-625, 1992.
Li, D. and K. P. Shine, A 4-Dimensional Ozone Climatology for UGAMP Models, UGAMP Internal Report No. 35, April 1995.
McCormick, M. P., Stratospheric ozone profile and total trends derived from the SAGE I and SAGE II data, Geophys. Res. Lett., 19, 3, 269-272, 1992.
McPeters, R. D., T. Miles, L. E. Flynn, C. G. Wellemeyer and J. M. Zawodny, Comparison of SBUV and SAGE II ozone profiles: Implications for ozone trend, J. Geophys. Res., 99, D10, 20513-20524, 1994.
Rusch, D. W., G. H. Mount, C. A. Barth, G. J. Rottman, R. J. Thomas, G. E. Thomas, R. W. Sanders, G. M. Lawrence and R. S. Eckman, Ozone densities in the lower mesosphere measured by a limb scanning ultraviolet spectrometer, Geophys. Res. Lett., 10, 241-244, 1983.
Rusch, D. W., G. H. Mount, C. A. Barth, R. J. Thomas and M. T. Callan, Solar Mesosphere Explorer ultraviolet spectrometer: Measurements of ozone in the 1.0 to 0.1 mb region, J. Geophys. Res., 89, 11677-11687, 1984.
Thompson, A. M., D. P. McNamara, K. E. Pickering and R. D. McPeters, Effect of marine stratocumulus on TOMS ozone, J. Geophys. Res., 98, D12, 23051-23057, 1993.
Thuburn, UGAMP Internal Report No. 16, 1992.
Watson, R. T. and Ozone Trends Panel, M. J. Prather and Ad Hoc Theory Panel, and M. J. Kurylo and NASA Panel for Data Evaluation, Present state of knowledge of the atmosphere 1988: an assessment report, NASA Ref. Publ. 1208, 201 pp, 1988.
WMO: Report of the International Ozone Trends Panel, 1988.
Zawodny, J. M. and M. P. McCormick, Stratospheric Aerosol and Gas Experiment II measurements of the quasi-biennial oscillations in ozone and nitrogen dioxide, J. Geophys. Res., 96, 9371-9377.
Each file contains a ligne of text followed by the variable itself, in free format.
Every single three-dimensional field var is stored as
(((var(i, j, k), i=1, 144), j=1, 73), k=1, 47)where
If you are not familiar with this Fortran syntax, please see the note (3) below.
Every two-dimensional field (zonal means) is stored as
((var(j, k), j=1, 73), k=1, 47)with the same conventions as above.
Two Fortran programs are provided by the authors of the climatology:
Information about where to find the available software is given in Section 4 below. We invite you to provide the BADC with any software that you would judge relevant. A special directory is accessible for this purpose.
In any case, please refer to the instructions below to find the appropriate directories and files.
After having transferred them to your home machine, you will have to uncompress the files (2) in order to be able to read the data. For information about the size of compressed and uncompressed files, please refer to Subsection 4.2, Table 5. You will need a total of about 87 Mbytes to store the whole climatology in its gzipped version, including 5-year and zonal averages. When all files are gunzipped, their total volume is of about 435 Mb.
The ASCII data files are in subdirectories below the data directory. The names of these subdirectories and the nature of the ozone data they contain are given in Table 3 below.
|Subdirectory name||Type of data|
|y85||Year 1985 monthly means|
|y86||Year 1986 monthly means|
|y87||Year 1987 monthly means|
|y88||Year 1988 monthly means|
|y89||Year 1989 monthly means|
|yr5||5-year average of the monthly means|
Programs to read and treat the data can be found in the software directory. The three subdirectories below software correspond to their respective providers and are listed in Table 4. For details about the UGAMP software, please refer to Section 3 above. You are welcome to export any relevant software into subdirectory 3rdparty.
|Subdirectory name||Origin of software|
Table 5 also gives the approximate size of the files, in their compressed (2) and uncompressed (2) versions.
|Type of data||File name
|Approximate file size (Kbytes)|
|Yearly monthly means
|mmmyym.gz||1100 - 1200||6000|
|mmm5ym.gz||1100 - 1200||6000|
The August 1988 monthly mean of the ozone column is to be found
in the file
(1) Ozone column. Dobson Units. The ozone column (above level l) is the number of ozone molecules in a vertical column of unit section above l. It is expressed in number of molecules per unit area or in Dobson Units (height of the ozone-filled segment of the air column, in mm, if this was brought back to standard pressure and temperature and divided into its individual chemical components). The total ozone column is the ozone column above the ground level.
(2) GZip compressing package. The UGAMP O3 data files held at the BADC have been compressed using the gzip command (their names are of the form filename.gz). After having transferred the files to your home machine, you will need to use the reverse command in order to be able to read them. If your local machine does not support GZip, it is possible to retrieve this tool from a number of public domain archives. Typing gunzip filename.gz will issue an uncompressed ASCII file filename while deleting filename.gz. To compress it again, type gzip filename.
Embedded READ/WRITE loops. According to Fortran
conventions, the components of a multi-dimensional variable written
(((var(i, j, k), i=1, imax), j=1, jmax), k=1, kmax) are stored in the following order.
|k = 1||j = 1||i = 1, ..., imax|
|j = jmax||i = 1, ..., imax|
|k = kmax||j = 1||i = 1, ..., imax|
|j = jmax||i = 1, ..., imax|