News and Views
First published in Nature 447 June 2007
Published online: 6 June 2007 | doi:10.1038/447647a
Climatology: Tempests in time
James B. Elsner1
The frequency of severe hurricanes in the North Atlantic has increased during the past decade. Scrutiny of the prehistoric record left by such storms helps to assess the factors contributing to hurricane activity.
A hurricane is a product of its environment: a warm ocean provides sustenance; calm atmospheric conditions nurture an infant storm; and a high-pressure cell in the subtropical atmosphere drives it in a given direction. Increases in oceanic heat from global warming will raise a hurricane's potential intensity, all else being equal. Yet increases in wind shear — in which winds at different altitudes blowing in different directions may tear apart the developing storm — could counter this tendency by dispersing the storm's heat.
In the long run, which effect will win out? Limited instrumental records of hurricanes and climate change make it difficult to answer this question. So researchers have turned to prehistoric 'proxy' data to uncover clues about what to expect in a warmer world. Two new papers, one published on 24 May1 and one on page 698 of this issue2, illustrate how the approach has been applied to hurricanes in the North Atlantic.
Palaeotempestology is the study of prehistoric storms from geological and biological evidence. Coastal wetlands and lakes are subject to 'overwash' during hurricanes, when barrier sand dunes are surmounted by storm surge. The assumption is that the waves and wind-driven storm surge reach high enough over the barrier to deposit a fan of sand in the lake3. A sediment core from the bottom of the lake shows that fan as a sand layer distinct from the fine organic mud that accumulates slowly under normal conditions.
Donnelly and Woodruff1 analysed sediment cores they extracted from a lagoon on the Puerto Rican island of Vieques. The lagoon is separated from the ocean by a stable barrier of sand. In the core, they found coarse-grained sand layers embedded in several metres of organic-rich silt. The layers are clearly the result of barrier and nearshore sediments that have been washed into the lagoon by strong hurricanes, the recent layers being correlated in time with known hurricane strikes. The authors calibrated the sensitivity of the site to storm surge by noting the intensity of known strikes that did not leave sand in the core.
Donnelly and Woodruff find more sand layers during the latter half of the Little Ice Age. This occurred between 300 and 150 years ago, and towards the end of this interval sea temperatures near Puerto Rico were 2 °C cooler than they are now. The authors say this is evidence that today's warmth is not needed for increased storminess. Not surprisingly, they find that intervals in which more hurricanes occurred correspond with periods of fewer El Niño events. El Niño events suppress hurricane activity in the North Atlantic by increasing the amount of wind shear and sinking air.
Nyberg et al.2 describe a different approach that has led them to the same conclusion — that, in the long run, shear is more important than ocean temperature in modulating hurricane activity. They use proxy records of shearing winds and ocean temperature to reconstruct a two-and-a-half century record of major hurricanes and wind shear. The proxies are based on luminescence banding in coral cores retrieved from sites in the northeastern Caribbean, and on a marine sediment core from further south.
However, studies relying on a spatially limited set of coring and proxy locations are not able to resolve changes in hurricane tracks. The northeastern Caribbean is in the direct path of hurricanes today, but has it always been? More hurricanes occurring locally could mean a shift in their direction rather than their abundance. Donnelly and Woodruff1 find that changes in hurricane frequency over the northeastern Caribbean seem to mirror the changes in frequency inferred from cores collected in New York, but the degree of correlation is not quantified. Proxy data from the Gulf coast show a pattern of frequent hurricanes between 3,800 and 1,000 years ago, followed by relatively few hurricanes during the most recent millennium, which has been explained in terms of the shifting position of the subtropical high-pressure zone4. Unravelling the causes of changes in local hurricane activity requires an understanding of the factors that influence what track they will take5. So further work is needed.
In addition, the assumption that hurricanes are simply passive responders to climate change should be challenged. Hurricane activity influences the observations and proxies used to compute mean quantities such as wind-shear and precipitation conditions, so the arguments can easily become circular. Reduced rainfall and greater mean shear are possible consequences of fewer hurricanes, not necessarily the causes. More importantly, a hurricane removes heat and water from the ocean and transports them upward and poleward, thereby modifying the environment that supports it. A strong hurricane cools the ocean surface beneath it as a result of evaporation and mixing of water layers. This makes the area less favourable for the next storm, but at depth adds heat to the ocean that can, in the long run, influence the climate system6.
Palaeotempestology is a valuable tool for answering questions on hurricane climatology. But more records are needed before localized prehistoric activity can be used to make sense of large-scale patterns of storminess. As Liu3 has pointed out, each record serves as a 'palaeoweather station', sensitive only to nearby hurricanes. At present, fewer than a dozen sequences that have been dated and validated are available in hurricane-prone regions of the United States and Caribbean. However, the new analyses1 and those of others4, 7 are a start.
When more palaeoweather stations have been established, a network can be constructed with links connecting sites that share similar periods of storminess. That network can then be compared to a network of storminess from modern records (Fig. 1) to better understand the evolving mechanisms responsible for changing hurricane risk.
Figure 1: One for the modern record.
Hurricane Katrina made its infamous assaults on the Bahamas, Cuba, south Florida and the Gulf coast in late August 2005.
GOES 12 SATELLITE/NASA/NOAA
References
Author affiliation
- James B. Elsner is in the Department of Geography, Florida State University, Tallahassee, Florida 32306-2190, USA.
e-mail: jelsner@fsu.edu
