A recent paper published by Gaertner et al. in Geophysical Research Letters (GRL) (in press as of 12 July 2007) explores the possibility for tropical cyclones to occur in the Mediterranean area in the future with global warming. It has been featured in the Times. The paper correctly points out that tropical cyclones have recently ventured into some unusual places with Catarina on the east coast of Brazil in March 2004 and Vince making landfall in Spain in 2005. Gaertner et al. use an ensemble of regional climate models to assess new locations of tropical cyclone occurrence. They find an increase in extremes of cyclone intensity over the Mediterranean Sea in regional model climate change scenario simulations. This increase, they claim, is clearly related to tropical cyclone formation, revealing for the first time a risk of tropical cyclone development over the Mediterranean Sea under future climate change conditions.
The regional model framework for this study is one limitation as it may not create the correct large-scale atmospheric circulation across the region. In particular, the Mediterranean climate in summer is one of clear skies and sunny days associated with the overall global monsoon circulation, such that the upward motion and rains in southern Asia are linked to the subsiding air over the Mediterranean that makes for a very stable environment unfavorable for storms. Unless that link is properly simulated (and it may not be, especially in a regional model) the vertical atmospheric temperature and wind structure are unlikely to be right.
The authors are aware of the summer difficulties and so they choose September as the time for the simulations. This has the advantage that the sea surface temperatures (SSTs) are highest then. The sun has been beating down all summer, but it tends to form a shallow layer of warm water and whether there is adequate heat below the surface is also a critical question. As cyclones form they churn up the ocean, bringing cold waters up from below and cool the ocean, creating a cold wake. In this work the SSTs were specified and fixed and this was not allowed to happen.
Moreover, by September, the mid-latitude westerlies have set in and the winds in the upper troposphere are typically over 20 m/s, so the wind shear environment is generally unfavorable for such storms. It is possible that the weather could change enough to relax the winds for just long enough to open a window for tropical cyclones but the odds are not high. This is the extratropics, after all.
All these questions remain. Nonetheless, with higher SSTs it does seem likely that cyclones would be more vigorous. But whether they are truly tropical or not is a key issue. The paper does not comment on water vapor: the Mediterranean Sea is surrounded by land where the air is drier and this can cause storms to run out of moisture and peter out. Nor does it mention rainfalls and flooding: a chief characteristic of tropical cyclones is the heavy rainfalls of several mm per hour. Nonetheless, warm core storms have been noted in the Mediterranean and are colloquially referred to as “Medicanes”.
For further discussion of tropical cyclones and climate change, see my recent article in Scientific American here.
Kevin E. Trenberth
Head of the Climate Analysis
US National Center for Atmospheric Research