Posted by Olive Heffernan on behalf of Kevin Trenberth

The 2007 hurricane season is about to get officially underway. Never mind that nature has already provided the first named storm in the North Atlantic: Andrea. Several forecasts suggest that the 2007 season in the North Atlantic will be well above average. Sea surface temperatures (SSTs) are above normal and atmospheric conditions look likely to be favorable for tropical storm activity.

In 2005, the record breaking year in the North Atlantic, record high SSTs in the critical region from 10 to 20 degrees N in the North Atlantic provided ample fuel for the 28 named storms (normal about 15, previous record 21 in 1933) and 15 hurricanes (normal about 6, previous record 12). Atmospheric conditions were favorable with weak wind shear (that otherwise tends to tear a vortex apart) and the absence of stable layers that prevent convection from developing. In contrast in 2006, SSTs were much lower and closer to the long term normal, and the atmospheric conditions were not favorable, as a developing El Niño in the Pacific created an atmospheric circulation that increased the wind shear in the Atlantic. This year, in 2007, there is no El Niño in the Pacific, and SSTs and the upper ocean heat content are more favorable for Atlantic storms.

There seems to be general agreement on these points, yet the whole issue of Atlantic hurricanes is mired in controversy over the role of global warming. It is not a disagreement that SSTs are higher but rather whether the warming is due to natural processes such as the Atlantic Multi-decadal Oscillation or global warming. To the public, the result is the same for now. To me this is obvious: global warming is “unequivocal” to quote the recent IPCC Working Group I report and global SSTs have increased about 0.6 degrees C. In the last half century this warming is associated with increased greenhouse gases in the atmosphere. It is not possible that the Atlantic has escaped from this warming.

The latest shot fired on this issue was by Chris Landsea in Eos on the long term Atlantic hurricane record. The Atlantic has the best record of tropical storms owing in part to aircraft surveillance since 1944. It is widely recognized that the older records are less reliable as the reports depend on shipping or landfall of the storms. Landsea makes this case quite well, but he goes further. He assumes that the percentage of landfalling storms should be constant over time and this provides a basis for adjusting the older records and increasing the numbers of storms by 2.2 per year from 1900 to 1965, when the satellite era began. The critical assumption is that this fraction should remain the same, yet it is abundantly clear that SSTs have increased and so other things have not remained the same. Given the dependence of hurricanes on high SSTs above about 26 degrees C, shouldn’t increased SSTs increase the scope for storms to develop farther to the east over what were colder waters and for the tracks to become longer? An example may be Vince in October 2005, which was the first tropical storm to make landfall in Spain and Portugal. Indeed shouldn’t one expect a decreasing fraction of landfalling storms in a warming climate?

The fact is that SSTs have increased and so there is every reason to believe that activity has increased also. Activity can be manifested in several ways:

1) increased number

2) increased intensity

3) increased size

4) increased duration.

What we expect from the dynamics is that the storms will experience increased intensity because there is more fuel for the storms (higher SSTs and more water vapor) but stronger storms take more energy out of the ocean and leave behind a colder wake, and so the expectation is that numbers could actually decrease. There is some evidence to suggest that size, and thus damage, may increase, but there are no statistics on size at all. A key issue for the Landsea paper is what about duration? As long as the storms keep moving to a new piece of ocean they do not get affected by the cold wake, and so in recent times the storms may be developing farther to the east and this may mean there are fewer storms missed than Landsea claims. Moreover, storms that were east of 55 degrees W that might not have been subject to aircraft surveillance tend to move west into the area where they are tracked, and so all that may be missing is a bit at the start of the storm?

These uncertainties argue that there should be a reanalysis of tropical storms in the satellite era since about 1970 or even 1966, to look not just at numbers but other statistics as well.

Kevin Trenberth

Head of the Climate Analysis Section

National Center for Atmospheric Research, USA