Figure 9.30: Storm track activity averaged over north-west Europe (6°W to 20°E, 40° to 70°N) in the ECHAM4/OPYC greenhouse gas scenario run (Unit: gpm). A 4-year running mean is shown for smoother display. The grey band indicates the variability of this index in the control run as measured by one standard deviation. The non-linear climate trend optimally obtained from quadratic curve fitting is marked by the dashed line; y-axis is activity in gpm (geopotential metres) and x-axis is time in calendar years. From Ulbrich and Christoph (1999).
Storms not only have obvious effects on extremes of temperature and precipitation, but also have severe impacts associated with wind, ocean waves, etc. Due to model limitations in previous generations of global climate models, until recently there have been few studies examining changes in extra-tropical cyclones in a future climate. With the improved recent generation of global climate models (see Chapter 8), such studies are now becoming more credible. An analysis of an ensemble of four future climate change experiments using a global coupled model with increased CO2 and sulphate aerosols showed an increase in the number of deep low pressure systems in Northern Hemisphere winter, while the number of weaker storms was reduced (Carnell and Senior, 1998). Studies using different models show a similar change for both hemispheres (Sinclair and Watterson, 1999) or for a study region limited to the North Atlantic (Knippertz et al., 2000).
The reasons given for this common result are still under discussion. Carnell and Senior (1998) ascribe it to a decrease in the mean meridional temperature gradient in the future climate, with high latitudes warming more than low latitudes (producing fewer storms), and greater latent heating in the moister atmosphere (resulting in deeper lows). Sinclair and Watterson (1999) point to the reduced mean sea level pressure and emphasise that vorticity as a measure of cyclone strength does not increase. Knippertz et al.(2000) consider the increasing upper tropospheric baroclinicity to be an important indicator of the change in surface cyclone activity. They also detect an increasing number of strong wind events in their simulation that can be assigned to the increasing number of deep lows. Upper air storm track activity (defined as the standard deviation of the band pass filtered 500 hPa height and related to the surface lows) has been found to increase over the East Atlantic and Western Europe with rising greenhouse gas forcing (such as seen in Figure 9.30 from Ulbrich and Christoph, 1999).
They related this increase to a change in the NAO (see discussion of possible NAO changes in Section 18.104.22.168). Several studies have tried to look at mechanisms of changes (e.g., Lunkeit et al., 1998). For example, Christoph et al. (1997) identify a mid-winter suppression of the North Pacific storm track in present day climate which they attribute to very strong upper level winds at that time of year. In a 3xCO2 climate model experiment, they note that very intense upper level winds occur more often, thus producing a more pronounced mid-winter suppression of the Pacific storm track.
Longer time-series from models have made the statistics more robust (e.g., Carnell and Senior, 1998). High-resolution models may improve the representation of storms, but the present experiments are mainly too short to provide indications of significant changes (e.g., Beersma et al., 1997). As can be seen, there are now a growing number of studies addressing possible changes in storm activity, but in spite of an emerging common signal there remains uncertainty with respect to the governing mechanisms.
Other reports in this collection