Nineteen named storms, 12 hurricanes, and five major hurricanesranked Category 3 and higher on the Saffir-Simpson scale.

That was the 2010 Atlantic hurricane season.

It was the second most active season on record in number ofhurricanes, and the third most active with respect to the number ofnamed storms (tropical cyclones with winds of at least 39 mph).Yet, not a single hurricane made landfall in the United States.

This exception is unprecedented.

In over a century–since at least 1900–there has been no otherAtlantic hurricane season that has had ten or more hurricanes inwhich not one has struck the United States.

As a result, many insurers may be asking how this reflects onthe catastrophe models they rely on to manage hurricane risk.

AN ACTIVE SEASON

In June, Hurricane Alex became the first Category 2 hurricane inmore than four decades to form so early. In September, eight namedstorms developed, tying 2002 and 2007 for the most tropicalcyclones on record ever to have developed in that month. In fact,from August 21 through September 26–36 days–not a single day passedin which there was not at least one tropical cyclone active in theAtlantic basin. And for only the second time in more than acentury, two Category 4 hurricanes, Igor and Julia, were active inthe Atlantic at the same time (also in September).

Just how active was it?

The 2010 season was about twice as active as the “average”season in terms of storm production.

o In a (statistically) average year, 10.7 named storms develop;2010 produced 19.

o In the average season, six storms develop into hurricanes;2010 saw 12.

o On average, three (2.8 to be precise) major hurricanes developin a season. In 2010, five major hurricanes developed.

WHY SO MANY STORMS?

A number of environmental factors, both long-term and short, mayhave contributed to the conditions that together produced the 2010hurricane season's unusual activity. One was a warm Atlantic.

A hurricane's “engine” is fueled by the heat and moisture of theocean surface. The warmer the ocean, the more readily available isfuel; the warmer the air, the more moisture it can hold forclouds–and eventually precipitation–to form and fall.Understandably, warm sea surface temperatures (SSTs) are positivelycorrelated with hurricane activity.

During the 2010 hurricane season North Atlantic sea surfacetemperatures were significantly warmer than average. While the juryis still out on the cause–be it the result of a naturalmultidecadal cycle as some have argued, or global warming–SSTs inthe Atlantic have been consistently above the long-term averageevery year since 1995. At the same time, while each individualseason has been unique, over this 15-year period of warm SSTs,counts of U.S. landfalling hurricanes, major hurricanes andaggregate annual losses fall in line with landfall frequency andaverage annual loss (AAL) encapsulated by the AIR U.S. HurricaneModel.

A second large-scale climatological factor is the “El Ni?o/LaNi?a-Southern Oscillation,” or ENSO, a climate cycle that occursacross the tropical Pacific Ocean and typically lasts for three toseven years.

ENSO is characterized by temperature fluctuations of thetropical eastern Pacific–a warming of the sea surface known as ElNi?o and a cooling known as La Ni?a. (Many years are neither warmor cool enough to classify as an ENSO event and are deemed“neutral.”)

Just before the onset of the 2010 Atlantic hurricane season, thePacific basin entered a new La Ni?a ENSO phase. La Ni?a isassociated with a seasonal reduction in destructive (to hurricanes)wind shear over the Atlantic, thereby providing a more favorableenvironment for tropical cyclone formation and development.

WHY SO FEW LANDFALLS?

While large-scale, long-term climate conditions had a strongeffect on the formation and development of tropical cyclones in2010, shorter-term weather patterns strongly influenced where thestorms tracked. For the most part, the most threatening storms weresteered away from the North American coastline.

In 2010, the “Bermuda High”–a large, semi-permanent center ofhigh atmospheric pressure in the eastern Atlantic near the Azores(and roughly at the latitude of Bermuda)–was at times situatedfurther east than usual and/or was weaker than usual. Additionally,for much of the core of the 2010 hurricane season, a low pressuresystem was in place over the U.S. East Coast. Because of thispattern, storms tended to track away from the United States.

Additionally, the jet stream's position over North America actedas a barrier that kept many storms over open water.

Similarly, since many storms formed in the extreme easternAtlantic (because of the favorable ocean/wind shear conditionsthere), they often began their turning to the north and east wellbefore ever reaching land. Storms that form off the coast of Africahave a relatively high probability of intensifying into hurricanesowing to their long trek over warm tropical waters, but arelatively low probability of making landfall along the U.S.coastline because of the Coriolis force (an artifact of the Earth'srotation), which inevitably deflects storms northward.

A LOOK BACK AT THE FORECASTS

The several government agencies, academics and privateorganizations that forecast the Atlantic hurricane season everyyear predicted that 2010 would be “an active year”–and they provedto be correct.

With respect to forecasting specific numbers of named storms orhurricanes that would form in 2010, most of the early predictionswere actually conservative. Nonetheless, in May, just before theformal start of the season on June 1, the National Oceanic andAtmospheric Administration forecast a very successful range ofcounts: 14-23 named storms (there were 19), 8-14 hurricanes (therewere 12), and 3-7 major hurricanes (there were 5).

NOAA does not forecast landfalls. Indeed, many forecasters donot, for the very reason that the factors that determine how stormstrack–atmospheric steering currents–are more closely associatedwith short-term weather variability that cannot be anticipated morethan a few weeks in advance.

Most forecasters that did venture a guess on the number ofhurricane landfalls expected a higher than average count. Theaverage is about two U.S. landfalls per year. Yet in 2010 onlyBonnie, which was barely a tropical storm at the time, came ashore,and with little adverse effect.

ONE NEAR MISS

The passage of one of the 2010 season's hurricanes elicited aloud and collective sigh of relief. Hurricane Earl–the season'sthird hurricane and second major hurricane–approached Puerto Ricoand the U.S. Virgin Islands at the end of August as a Category 4hurricane with winds near 150 mph. As the Labor Day weekendapproached, Earl's projected path was becoming clear: north alongthe U.S. East Coast with a possible landfall near downtownBoston!

Hurricane landfalls in the Northeast are relatively rare,occurring on average once every ten years. The last was HurricaneBob in 1991; the last major hurricane landfall in the Northeast–aneven more rare event–was the Great New England Hurricane of1938.

On September 2, the National Hurricane Center distributed itsforecast track and accompanying “cone of uncertainty” for HurricaneEarl. AIR estimates that had the left-most side of the cone ofuncertainty been realized and the storm had indeed cut up throughRhode Island and eastern Massachusetts, total insured losses fromEarl would have been about $4 billion.

A few days later Earl made a more northerly turn than projectedand eventually passed well east of Cape Cod. All the same,Earl–like Hurricane Bill last year–serves as a not so gentlereminder of the very real hurricane risk faced by the denselypopulated Northeast.

AFTERTHOUGHTS

The 2010 Atlantic hurricane season was unusually active andbroke or matched several records of hurricane formation, number andbehavior. All the same, not one hurricane made a U.S. landfall.

Climate conditions such as warm sea surface temperatures and theEl Ni?o/La Ni?a-Southern Oscillation can significantly influencetropical cyclone formation and development. However, even underfavorable conditions driven by a warm Atlantic, the number ofhurricane landfalls can be average or even well below average.

This was in fact the case in 2010.

It serves as another reminder that neither warm sea surfacetemperatures nor elevated activity in the Atlantic basin willnecessarily produce elevated losses. Loss projections depend on ahost of factors, many of which are not easily forecast.

What matters–in costs and in lives–are the storms that strikewhere people live and work. When, where and how often that willhappen is uncertain. Given the low accuracy of forecasting thenumber and location of hurricane landfalls and therefore theconsequent volatility in hurricane losses, there is currentlyinsufficient skill to predict losses in advance of a season.

The extreme variability of hurricane losses has led to the useof catastrophe models, which include thousands of scenarios thatrepresent the full range of plausible outcomes. Catastrophe modelsshould be expected to capture seasons like 2010–however unusualthey may be. Through the use of catastrophe models, estimatedprobabilities can be determined for different levels of loss, whichare essential input to robust catastrophe risk management.

Peter Dailey is assistant vice president,director of atmospheric science at AIR Worldwide.