Report of the SPARC Workshop on the Role of the Stratosphere in Tropospheric Climate
Whistler (BC), Canada, 29 April 2 May, 2003.
Nathan P. Gillett, University of Victoria (BC), Canada (email@example.com)
Mark P. Baldwin, Northwest Research Associates (WA), USA
David W.J. Thompson, Colorado State University (CO), USA
Emily F. Shuckburgh, University of Cambridge, UK
Warwick A. Norton, University of Oxford, UK
Jessica L. Neu, NOAA Climate Monitoring and Diagnostics Laboratory (CO), USA
Participants at the "Role of the stratosphere in troposphere climate" workshop, in Whistler, BC, Canada
Is the stratosphere important for predicting changes in weather and climate? Do perturbations to the stratosphere have a significant influence on the climate in the troposphere? These were the key questions addressed at a recent SPARC workshop in Whistler, British Columbia, sponsored by SPARC, NASA, NOAA, NSF, ESA and the Risk Prediction Institute. A total of 56 scientists participated, with 43 invited talks and 3 posters. Papers presented at the workshop explored observational evidence of stratosphere-troposphere coupling, the theory behind possible coupling mechanisms and the simulation of such coupling using a range of models, from simple mechanistic models to full GCMs. The meeting format allowed for ample discussion after each talk, with a summary/discussion session on the last day.
It has long been known that conditions in the stratosphere are controlled by wave driving from the troposphere, but traditionally it has been assumed that the stratosphere has little effect on the troposphere. Stratospheric variations, especially variations in the strength of the polar vortex, appear to be involved in feedback processes that in turn alter weather patterns in the troposphere. Stratospheric variations are largest during the winter season in the NH (and spring in the SH), and they are influenced by changes in solar irradiance, volcanic aerosols, changes in greenhouse gases, ozone depletion and the phase of the Quasi Biennial Oscillation (QBO).
A major focus of the meeting was the mechanisms by which stratospheric circulation anomalies affect the troposphere. Stratospheric circulation anomalies are caused mainly by wave forcing from the troposphere. Stochastic variations in the troposphere during NH winter lead to high-frequency changes in the planetary wave flux upwards into the stratosphere. When these waves break, they deposit momentum in the stratosphere, slowing the zonal-mean wind and weakening the polar vortex. The interaction of the waves with the mean flow tends to draw these zonal wind anomalies downward through the stratosphere. Our understanding of this process is incomplete, but in some ways it is analogous to the process by which zonal wind anomalies descend in the QBO.
The lowermost stratosphere has a long radiative timescale during winter, causing anomalies to persist there for several weeks. Thus, the lower stratosphere (LS) acts as an integrator of tropospheric variations, leading to longer-lasting anomalies in this region. Observational and modelling evidence suggests that LS anomalies are associated with circulation anomalies in the troposphere. This dynamical coupling occurs in the NH winter, when strong polar vortex anomalies in the stratosphere are followed, on average, by long-lived anomalies in the Northern Annular Mode (NAM) near Earths surface. The NAM is similar to the North Atlantic Oscillation, and NAM anomalies are associated with strong westerly winds in the mid-latitudes, and mild wet winters over Northern Europe and much of the U.S. In the SH the dynamical coupling occurs during spring, when the stratospheric polar vortex is most variable. Surface effects are seen in the Southern Annular Mode (SAM).
Synopsis of Presentations
Based on a cross-spectral analysis of annular mode variations at the surface and in the LS, D. Stephenson demonstrated that the surface annular mode shows stronger variations on timescales of ~60 days than would be expected if it were forced solely by high frequency weather noise, and that this increased variability on monthly timescales is associated with variations in the LS. This apparent downward influence may allow improved seasonal predictions beyond the ~10-day limit of deterministic forecasts. M. Baldwin showed that statistical forecasts of the monthly-mean surface NAM can be improved by using stratospheric data, and A. Charlton demonstrated that knowledge of stratospheric conditions gives some additional skill in forecasts of 15-20 days by using a medium-range weather forecasting model. W. Norton demonstrated that if stratospheric variability is artificially suppressed in a GCM, the timescale of the surface NAM is decreased, and D. Thompson showed a range of evidence for stratospheric influence on the tropospheric annular mode.
Observed stratosphere-troposphere links are not, however, limited solely to the extra-tropics: M. Hitchman and M. Giorgetta presented evidence that the equatorial QBO in stratospheric winds and temperatures can influence the surface via induced changes in tropopause conditions, which in turn can alter tropical convection. K.-K. Tung also presented statistical evidence that the influence of the QBO is manifested in tropospheric geopotential height. R. Quadrelli and K. Kodera presented evidence that the coupling between annular mode anomalies over the pole may be modulated by the phase of the El Niño/Southern Oscillation (ENSO), the warm ENSO phase being associated with stronger stratosphere-troposphere coupling in the extra-tropics. H. Graf also demonstrated that wave driving of the polar stratosphere is correlated with the phase of ENSO in observations.
Observational studies thus show that the strength of the zonal circulation in the troposphere is correlated with that in the stratosphere and based on the lag found between surface anomalies and those in the stratosphere, we might hypothesize that the influence is downwards. How might we test the direction of causality more conclusively, and find what mechanisms underlie the coupling? These questions were addressed by theoretical and model-based studies. Several presenters demonstrated that simple mechanistic models exhibit changes in the strength of the zonal circulation in the troposphere when conditions in the stratosphere are perturbed (P. Kushner, W. Robinson , M. Taguchi). Several other presenters pointed out that full GCMs show an apparent downward propagation of annular mode anomalies similar to that observed (R. Garcia, B. Boville, K. Hamilton). These studies generally concluded that stratospheric sudden warmings are preceded by an increase in the upward flux of planetary wave energy from the troposphere and they are followed, on a timescale of 1-2 months, by a weakening of the zonal circulation in the troposphere. P. Newman showed that the Antarctic polar vortex of 2002 was anomalously warm and disturbed due to stratospheric wave driving and not due to reduced photochemical ozone depletion. M. Salby demonstrated that anomalous wave driving of the stratosphere accounts for almost all the variability in polar stratospheric temperatures. D. Waugh further showed that much of the variability in the strength of the stratospheric vortex could be explained by changes in the upward wave flux near the tropopause. It remains an open question what determines the upward flux of wave activity into the stratosphere, but it is becoming clear that the configuration of the stratosphere itself is important and that tropospheric patterns by themselves are not sufficient to cause an upward flux of wave activity (R. Scott).
Several presenters discussed possible mechanisms by which anomalies in the zonal flow of the stratosphere might have a downward influence. The simplest and most direct is that anomalies in the LS have non-local dynamical effects in the troposphere (much as an electric charge has non-local effects on the surrounding electric field) (P. Haynes, R. Black ). However, this effect by itself is unlikely to be large enough to explain the observed downward influence. A consensus emerged that a positive feedback mechanism in the troposphere is required to explain the strength of the observed coupling. One likely mechanism involves positive feedbacks on the subtropical jets due to baroclinic eddies (A. Plumb, T. Shepherd, T. Dunkerton, W. Robinson). Baroclinic eddies extend into the LS, providing a region of overlap where stratospheric anomalies can influence tropospheric eddies. G. Vallis described a related mechanism underlying tropospheric variability, whereby a NAM-like mode is a natural consequence of stirring by baroclinic eddies in the mid-latitudes. The tropospheric and stratospheric flows may also be coupled by other mechanisms, as for example the reflection of upward-propagating planetary-scale Rossby waves from the stratosphere back to the troposphere (J. Perlwitz, N. Harnik, D. Ortland). Analagous wave-mean flow interactions were demonstrated in a laboratory setting by P. Rhines.
Coupling between the stratosphere and troposphere may have important implications for our understanding of the climatic response to greenhouse gas (GHG) increases and stratospheric ozone depletion. GHGs are expected to have a radiative cooling effect on the stratosphere, though in some models this effect is outweighed in the Arctic polar vortex by an increase in upward planetary wave flux, which has a warming effect (E. Manzini). Stratospheric ozone depletion cools the polar vortex in the spring due to the reduced absorption of UV radiation. Persistent changes in the strength of the stratospheric polar vortex of either hemisphere might be expected to influence the tropospheric circulation. B. Christiansen showed that a significant change has occurred in the frequency of occurrence of strong and weak stratospheric vortex states in the NH winter over the past 50 years. D. Karoly showed evidence that both greenhouse gas increases and ozone depletion may have contributed to a strengthening of the tropospheric zonal circulation in the SH, and D. Rind presented analogous results for the NH. J. Fyfe showed that the SAM response to GHGs increases is sensitive to the ocean mixing parameterization used, due to its effect on the meridional gradient of surface warming. N. Gillett showed that the observed trend in SH surface circulation can be simulated in a high resolution GCM forced solely with stratospheric ozone depletion.
Some volcanic eruptions cause large increases in stratospheric sulphate aerosols, which heat the stratosphere and persist for 2-3 years. Observations and modelling evidence (A. Robock, L. Oman) suggests that large volcanic eruptions can induce positive NAM anomalies at Earths surface, though the exact mechanism remains unclear. Some authors have argued that this effect comes about through changes in the strength of the stratospheric polar vortex, though other experiments appear to show a NAM-like response without such associated stratospheric changes (A. Robock). Likewise, changes in solar irradiance alter the thermal structure of the stratosphere, largely through induced changes in the stratospheric ozone distribution. K. Coughlin demonstrated a solar signal in geopotential observations that extends from Earths surface to the middle stratosphere. The signal is latitude-independent and, hence, unrelated to annular modes. J. Haigh demonstrated that solar changes may induce changes in the strength and position of the subtropical jets and in this case the suggested mechanism was tropical. L. Gray suggested that the solar cycle may play a role in determining the strength of the connection between the QBO and the polar vortices (the so-called Holton-Tan relationship). Her results indicated that the strength of the Arctic polar vortex is affected by the winds in the upper equatorial stratosphere and, thus, that the phase of the QBO indirectly affects the surface NAM.
The evidence presented at this workshop suggests that the stratosphere plays an important role in the climate system on timescales from weeks to decades. In the seasons when stratosphere-troposphere coupling is strongest its role may be of comparable importance to that of the tropical oceans. The participants reached a consensus that stratospheric circulation anomalies influence the troposphere, but the mechanisms underlying this coupling are not yet fully understood. Improving our understanding of these mechanisms may help us to better predict the climate response to greenhouse gas and ozone changes, as well as improving seasonal forecasts.
Meeting Web page (abstracts and presentations): http://www.atm.damtp.cam.ac.uk/shuckburgh/whistler/