CLIMATE
VARIABILITY
RESEARCH
The Impact of the Stratosphere on
Tropospheric Climate
The aim of this work is to answer the following questions: (a) Is an accurate representation of the stratosphere in global circulation models important for climate modelling? and (b) What is the impact of changes in the stratosphere on tropospheric climate?
General Circulation Model (GCM) Studies
CGAM.
Several multi-year simulations of stratosphere-troposphere configurations of the UK Meteorological (UKMO) Unified Model (UM) have been done. Results from a 49-level (L49) version have been published (Swinbank et al. 1998); an AMIP-II run with an L58 version has finished. This run carries a number of idealised passive tracers to study stratospheric transport and stratosphere-troposphere exchange.
The climate and variability of this L58 run is being analysed by comparison to the ERA-15 data (see Figure 1), and by comparison against UKMO runs from versions with the same troposphere as L58, but with a degraded stratosphere (L30). Initial investigations indicate that, in general, model performance is good. Some of the features in L58: tropospheric cold bias everywhere and a cold summer pole at ~200 hPa (simulations suggest that this error is associated with an incorrect water vapour distribution and its impact on the radiation scheme) are present in L30 (not shown). However, there are differences between the climate of L58 and L30 in the UTLS region (e.g. a stronger warm bias in the winter pole in L30), which suggest that the stratosphere may have an impact on tropospheric climate.
The UM has been used for short-term deterministic forecasts to study the effect of vertical resolution and location of model lid on predictive skill (Lahoz 1999). Results show that both flow history and vertical resolution are important, and that a better stratospheric representation does not necessarily give better predictive skill in the lower stratosphere.
Ensemble forecasts with L58 are being performed to determine whether a stratosphere-troposphere configuration of the UM has advantages as a climate prediction model at the seasonal time-scale. This work is being done in collaboration with the UKMO and the EU-funded PROVOST project. Preliminary results suggest that a well-resolved stratosphere has an impact on seasonal prediction.
An L74 version of the UM with very high stratospheric resolution (~0.8 km) has been run for 5 years in an attempt to simulate a realistic QBO. The model does not capture the QBO, but has a better representation of the tropical stratosphere than L49 or L58. The wave spectrum of this run is being analysed. To study the impact of a lack of a QBO in the UM, an imposed QBO from the UKMO stratospheric analyses will be included in the UM.
Parametrization Studies
Imperial College.
The Edwards-Slingo radiation code in the UM has been improved for use throughout the lower and middle atmosphere. Tests show good agreement with more accurate calculations (Zhong and Haigh, this issue). The improved code will now be implemented in a vertically extended UM.
Oxford.
The Hines gravity wave drag scheme has been implemented in a L54 (lid at 0.01 hPa) version of the UM. In comparison with runs performed with Rayleigh friction, the scheme produced more realistic middle atmosphere jets; in particular, these now tilt equatorward with height. The wintertime cold bias at 10 hPa is also reduced. There is also a notable improvement in the simulation of the break-up of the southern hemisphere (SH) polar vortex.
Mechanistic Model Studies
Edinburgh.
Simulations of SH / NH winters have been done using the COSMIC (coupled USMM/TOMCAT) model (see MacKenzie, this issue, for details).
RAL.
Mechanistic studies are being done using the UKMO SMM to study the sensitivity of NH winter to equatorial processes, including the QBO. Observations suggest a link between the phase of the QBO and the timing of stratospheric warmings, which could then impact the development of tropospheric disturbances. Ensemble runs have been performed to characterise the natural variability of the modelled NH winter; this compares well with observations. An imposed QBO has been included to study the influence of the QBO on this variability. In good agreement with observations, there is a statistically significant link between the QBO phase and the flow type, with cold undisturbed winters occurring more often during the westerly QBO phase than in the easterly phase. Cases when this simple relationship does not hold are being studied to elucidate other factors which might have a significant impact on the winter stratosphere evolution.
References
Lahoz, W. A. (1999). Predictive skill of the UKMO Unified Model in the lower stratosphere. Q. J. Roy. Meteorol. Soc., 125, in press
Swinbank, R., et al. (1998). Middle atmosphere variability in the UK Meteorological Office Unified Model. Q. J. Roy. Meteorol. Soc., 124, pp 1485-1525.
Dr. W. Lahoz
CGAM, University of Reading
wal@met.rdg.ac.uk
Dr. L. Gray, RAL
l.gray@rl.ac.uk
