Stratospheric temperature trends: impact of ozone variability and the QBO

Abstract

In most climate simulations used by the Intergovernmental Panel on Climate Change 2007 fourth assessment report, stratospheric processes are only poorly represented. For example, climatological or simple specifications of time-varying ozone concentrations are imposed and the quasi-biennial oscillation (QBO) of equatorial stratospheric zonal wind is absent. Here we investigate the impact of an improved stratospheric representation using two sets of perturbed simulations with the Hadley Centre coupled ocean atmosphere model HadGEM1 with natural and anthropogenic forcings for the 1979–2003 period. In the first set of simulations, the usual zonal mean ozone climatology with superimposed trends is replaced with a time series of observed zonal mean ozone distributions that includes interannual variability associated with the solar cycle, QBO and volcanic eruptions. In addition to this, the second set of perturbed simulations includes a scheme in which the stratospheric zonal wind in the tropics is relaxed to appropriate zonal mean values obtained from the ERA-40 re-analysis, thus forcing a QBO. Both of these changes are applied strictly to the stratosphere only. The improved ozone field results in an improved simulation of the stepwise temperature transitions observed in the lower stratosphere in the aftermath of the two major recent volcanic eruptions. The contribution of the solar cycle signal in the ozone field to this improved representation of the stepwise cooling is discussed. The improved ozone field and also the QBO result in an improved simulation of observed trends, both globally and at tropical latitudes. The Eulerian upwelling in the lower stratosphere in the equatorial region is enhanced by the improved ozone field and is affected by the QBO relaxation, yet neither induces a significant change in the upwelling trend.

Appendices

Appendix A: an observed ozone dataset for the 1979–2003 period for use in modelling studies

As most coupled atmosphere-ocean climate models do not include an interactive chemistry scheme, ozone concentrations have to be externally imposed. Here, we present a zonal mean ozone distribution timeseries for the 1979–2003 period derived mainly from satellite observations.

For the stratosphere, the SAGE-II, MLS, HALOE and SBUV satellite instruments are the main data source. Their combination achieves data coverage over the region 215–0.1 hPa. Prior to 1985, the only data included in the time series are from SBUV and TOMS. SBUV consists of measurements from several different satellites, and those measurements have been combined by taking into account drifts and differences between satellites (details are given in Stolarski and Frith 2006). The retrievals have been examined more recently than for SAGE-I. Hence the SBUV choice for the early period. Further, SBUV provides global coverage where there is sunlight, as opposed to one latitude for a day in a month as from SAGE (which is a solar occultation instrument). From 1985 to 1991, SAGE-II data are also included. After 1991, MLS and HALOE are added. For the period after the El Chichón eruption, SBUV data processing excluded data that was deemed aerosol contaminated, which means that some tropical data are missing for about a year after the eruption, but they are not missing globally. To make a continuous timeseries, a climatology had to be used to fill in missing data. For a 2-year period following the eruption of Mt. Pinatubo, the SAGE-II data drops out, as does some of the tropical HALOE data.

The SBUV and TOMS data are from the version 8 merged dataset (http://code916.gsfc.nasa.gov/Data_services/merged/) which have been adjusted to account for satellite differences over the period from late 1978 to the present. At heights above 0.1 hPa SME data are available for the period 1981–1989. Only data for the height region 0.14 hPa to about 0.0015 hPa were employed (SME data are available for the stratosphere but they are of poorer quality). Above 0.0015 hPa the UK Universities Global Atmospheric Modelling Programme (UGAMP) ozone climatology (Li and Shine 1995, see http://badc.nerc.ac.uk/data/ugamp-o3-climatology) which is valid for the period 1985–1989 was employed. It was scaled to match SME at the top of the SME range. Note that no diurnal effects were included in the data compilation.⁵

The first step in developing the ozone time series was the construction of a monthly mean zonal mean climatology for the period 1993–2000 with a 5° resolution. Prior to merging the SAGE-II, HALOE, MLS and SBUV data to form the stratospheric component of the climatology, the data were scaled to match HALOE observations, based on comparisons when coincidences were available. Reliable satellite data go up to 0.1 hPa, essentially everything above that was based on the SME and UGAMP climatologies. Data were scaled to match HALOE because HALOE had the best overlap with all the assorted datasets. Published accuracy for all the satellite instruments is on the order of 5–10%. The resulting changes to the SAGE values were less than 5% throughout the bulk of the stratosphere but were as much as 20% in the mesosphere at 0.1 hPa. Changes to MLS were less than 5%, except at 0.22 and 0.32 hPa where they were 8%. Changes to SBUV were less than 3%, except near 2 hPa where the change was 10–20%. The sign of the correction was positive for SAGE-II and MLS, and the sign varied with altitude for SBUV. This combination of SAGE-II, HALOE, MLS and SBUV data provides coverage from 215 to 0.1 hPa. SBUV data was only used as high as 1 hPa. Because SME data were not available for the 1990–2000 period, the available pre-1990 data were used and scaled to match the SAGE-II + HALOE + MLS + SBUV average at 0.1 hPa on a globally-averaged basis before appending to the top of the zonal climatology. Similarly, the UGAMP climatology which is valid for the period 1985–1989 was scaled to match SME at the top of the SME data range, before being appended. The UGAMP climatology used SBUV, SAGE II, SME airglow and TOMS, and ozonsondes, and was valid for 1985–1989. Shifts were just applied to make the two averages consistent. The SME yearly average climatology was based on all the SME data available (both airglow and UV instruments), the satellite ran from 1981–1989. Below 215 hPa, a scaled version of the KNMI (The Royal Dutch Meteorological Institute, Dutch: Koninklijk Nederlands Meteorologisch Instituut) climatology (Fortuin and Kelder 1998) was appended and then additionally adjusted to match the climatological total column ozone for the 1990s from the TOMS instrument.

Once the monthly mean climatology was developed, the time series could be constructed. Firstly, the SAGE-II, HALOE, SBUV and MLS data were binned on a 5° latitude grid, and then averaged on a monthly basis. After doing this, there were many points with missing data. Initial filling was done via interpolation across latitude or time if only one point was missing between two points with data. With more missing points, the climatology was used to fill in, scaled based on the ratio between the endpoints of the existing data and the climatology. For example, if March and April 1993 were missing, then the climatology was multiplied by the ratio of February and May 1993 with the February and May climatology and the resulting values used. Polar regions were filled in using the previously constructed climatology scaled by the observed total ozone for any given month from TOMS. Above 1 hPa, the climatology was scaled according to the ratio of the observed column between 1 and 2 hPa and the climatological column between 1 and 2 hPa. The troposphere was filled in by scaling the mean climatology to match the deficit in ozone column needed to match the TOMS data. The whole dataset was then checked on a profile-by-profile basis looking for discontinuities, and those found were adjusted, keeping the column within 3% of the observed TOMS column. As for the vertical coordinate system, the original coordinate for the ozone measurements varies according to the instrument. Initially there were pressure data for MLS, HALOE and SBUV, and altitude data (including pressure information) for SAGE II. Before combining datasets, everything was put on a set of standard pressure levels that corresponded with the UARS (Upper Atmosphere Research Satellite) standard pressure levels used by MLS and HALOE. The altitude range is from the surface to 0.01 hPa for the total combined set. Pressure ranges vary for the different instruments used.

Appendix B: preparation of the improved ozone dataset

For the perturbed sets of simulations, the tropospheric component of the dataset described in Appendix A below the WMO tropopause has been overwritten using the corresponding values employed in the baseline simulations (described in Sect. 2.2 and in Stott et al. 2006). This ensures that only changes in the stratosphere were imposed.

The resulting dataset was then converted to the height-based hybrid vertical coordinate of the model (see Table 2 of Martin et al. 2006), using the SPARC temperature climatology (Randel et al. 2004, available online on http://www.sparc.sunysb.edu/html/temp_wind.html) and finally the data were sampled at the appropriate model levels.

Note that vertical interpolations make use of the hydrostatic equation and require knowledge and/or assumptions about the temperature profile. There are simpler approximations than the one employed as e.g. assuming an isothermal atmosphere, which implies a constant scale height. In a standard atmosphere, temperature varies only with height but not with latitude or season. Here, we used the SPARC temperature profile climatology. Possible errors introduced by the simpler approach of using the International Standard Atmosphere (ISA, see http://en.wikipedia.org/wiki/International_Standard_Atmosphere) for the January climatology reach about +40% near the equator in the lowermost stratosphere (not shown) and −40% at high latitudes in the NH (winter hemisphere) in the upper troposphere and lower stratosphere. Potentially, biases may be introduced also by neglecting large trends in the vertical temperature profile. Such errors can induce differences in diabatic warming rates and hence temperature and wind perturbations. Although we do not test the resulting impact of applying any simplifying assumption here, the above figures call for caution in choosing an appropriate method for vertical interpolation of ozone distributions and suggest that the description of a new dataset should report the assumptions made.

The ozone amounts residing above the top of the model lid at about 5 hPa were discarded, which accounts for about 10–20 DU contribution to the total column of about 300 DU. While this treatment follows current practice, and indeed a similar treatment was carried out in the preparation of the ‘baseline’ ozone treatment, the discarding of ozone amounts will inevitably cause errors in the diabatic warming estimates, which are not straightforward to assess since ozone is active at both short and long wavelengths. The authors are not aware of any investigation of the error caused by this omission. While one might consider redistributing the discarded ozone amounts within the top few levels of the model, the height profile of such a redistribution would need careful consideration as it is likely to significantly perturb the diabatic warming profile. Indeed, this redistribution would lead to a systematic overestimate of the solar heating, because the ozone would lead to absorption of more solar radiation at altitudes that it would not ordinarily be absorbed, as it would not be “sheltered” by the absorption of solar radiation by ozone in the upper stratosphere. Also, by replacing the stratospheric component of the imposed ozone fields with the new improved dataset, the total column amounts in the baseline and the perturbed simulations will not be identical.