By Richard Black
Environment correspondent, BBC News
A New Orleans street battens down as Hurricane Gustav approaches
Hurricanes in the Atlantic are more frequent than at any time in the last 1,000 years, according to research just published in the journal Nature.
Scientists examined sediments left by hurricanes that crossed the coast in North America and the Caribbean.
The record suggests modern hurricane activity is unusual - though it might have been even higher 1,000 years ago.
The possible influence of climate change on hurricanes has been a controversial topic for several years.
Study leader Michael Mann from Penn State University believes that while not providing a definitive answer, this work does add a useful piece to the puzzle.
"It's been hotly debated, and various teams using different computer models have come up with different answers," he told BBC News.
"I would argue that this study presents some useful palaeoclimatic data points."
Hurricanes strike land with winds blowing at up to 300km per hour - strong enough to pick up sand and earth from the shore and carry it inland.
In places where there is a lagoon behind the shoreline, this leads to "overwash" - material from the shore being deposited in the lagoon, where it forms a layer in the sediment.
Researchers have studied eight such lagoons on shores where Atlantic hurricanes regularly make landfall - seven around the US mainland and one in Puerto Rico.
Over time, Dr Mann's team believes, the number of hurricanes making landfall on these sites will be approximately proportional to the total number of hurricanes formed - so these zones provide a long-term record of how hurricane frequency has changed over the centuries.
Wind shear at altitude can prevent a tropical storm's structure developing
The last decade has seen an average of 17 hurricanes and tropical storms in the Atlantic - earlier in the century, half that number were recorded.
But current levels were matched and perhaps exceeded during the Mediaeval Climate Anomaly (also known as the Mediaeval Warm Period) about 1,000 years ago.
"I think if there's one standout result (from this study), it's that the high storm counts we've seen in the last 10 to 15 years could have been matched or even exceeded in past periods," commented Julian Heming, a tropical storm specialist from the UK Met Office who was not involved in the new research.
"So it's worth feeding into the debate about whether what we're seeing now is exceptional or something related to multi-decadal or even multi-centennial variability; and it does tell us that the levels we're seeing at the moment are within the bounds of uncertainty."
Dr Mann's team also used a pre-existing computer model of hurricane generation to estimate activity over the same 1,500-year period.
The model includes three factors known to be important in determining hurricane formation: sea surface temperature in the tropical Atlantic Ocean, the El Nino/La Nina cycle in the eastern Pacific, and another natural climatic cycle, the North Atlantic Oscillation.
This analysis suggests, Dr Mann argues, that the hurricane peak 1,000 years ago and the current high activity are not produced by identical sets of circumstances.
Then, he says, an extended period of La Nina conditions in the Pacific - which aid hurricane formation - co-incided with relatively warm conditions in the Atlantic.
Now, the high number is simply driven by warming waters in the Atlantic - which is projected to increase in the coming decades.
"Even though the levels of activity are similar (between 1,000 years ago and now), the factors behind that are different," said Dr Mann.
"The implication is that if everything else is equal - and we don't know that about El Nino - then warming of the tropical Atlantic should lead to increasing levels of Atlantic tropical cyclone activity."