Editor’s Note: The 2009 Atlantic Hurricane Season got off to a surprisingly slow start. Then, all within a week, mid-August brought us Ana, Bill, and Claudette. Bill, with maximum sustained winds of 135 mph, reached Category 4 on the Saffir-Simpson scale and became the first major hurricane threat of the 2009 season. Hurricane Bill, however, threatened not Florida or the Gulf States, but the coast of New England. The next named storm, Danny, followed a week later almost directly in Bill’s wake. AIR’s Director of Atmospheric Science, Peter Dailey, describes Hurricane Bill’s development and the red flag it raised: the U.S. Northeast remains subject to hurricane risk. As companies consider their Northeast hurricane risk with renewal season approaching, events like Hurricane Bill should prompt questions about how catastrophe models reflect these storms.

By Dr. Peter Dailey

We are now just past the halfway mark of the official 2009 Atlantic hurricane season and six named storms have developed. Two have reached hurricane strength. The first, Hurricane Bill, put the entire U.S. Northeast coastline on high alert. Hurricane Bill’s passage provides an opportunity to discuss several aspects of fundamental hurricane risk in a region where such risk is often ignored—if not dismissed.

A Hurricane in the Northeast?

Figure 1 below shows the “ensemble” of possible storm tracks projected for Hurricane Bill by the Global Forecast System (GFS)¹ on August 20. Each of the pink lines represents a track forecast initialized with the environmental conditions in place about three days before Hurricane Bill actually approached the Northeast. Several of the possible paths take Bill directly over some of the most densely populated areas of the New England coastline.

Can a hurricane make landfall in the Northeast United States? The answer is a resounding “Yes!” The GFS model has demonstrated significant forecasting skill over periods of at least five days. The tracks in Figure 1, based as they are on the actual environmental conditions of a particular moment, are not simply realistic simulations; they display very plausible outcomes for Hurricane Bill that could have had real consequences. Bill’s near miss only serves as a reminder that Northeast hurricane risk is both real and significant.

Figure 1: GFS model ensemble of August 20 showing Hurricane Bill’s potential for making landfall along the Northeast U.S. coastline. (Source: Weather Underground)

Could Hurricane Bill Have Made Landfall as a Major Hurricane?

The Northeast United States last experienced a landfall by a major hurricane 70 years ago. The Great New England Hurricane of 1938 was a powerful storm—likely Category 5 when it passed Puerto Rico—that began to ”transition” when it moved north into the oceanic and meteorological environment of the mid-latitudes. Not all Northeast hurricanes exhibit this phenomenon. “Transition” tends to make a storm’s wind field less intense and more diffuse. Nevertheless, despite undergoing transition, the 1938 hurricane still made landfall over Long Island and Connecticut as a Category 3 hurricane. It was responsible for 600 deaths and produced significant damage. Were the so-called Long Island Express to happen today, AIR estimates that it could cause insured losses of as much as USD 38 billion.

Figure 2 below shows the track of the 1938 hurricane, which is strikingly similar to the path taken by Hurricane Bill. Where they differ is that as the 1938 hurricane moved up the eastern seaboard it was swept westward—closer to the U.S. coastline—by environmental steering conditions that did not similarly materialize for Hurricane Bill. However, Bill’s forecast paths and the historical path of the 1938 hurricane—and that of the Great Atlantic Hurricane of 1944, Hurricanes Carol and Edna of 1954, Hurricane Donna in 1960, Hurricane Agnes in 1972, Bob in 1991, and Floyd in 1999—demonstrate that tropical cyclones can and do track toward the Northeast.

The important question for insurers is how strong such storms will be when they arrive—and thus how much damage they can produce in a coastal region of particularly high population density.

Figure 2: Track of the Great New England Hurricane of 1938. (Source: NHC)

When the National Hurricane Center (NHC) released its forecast for Hurricane Bill at 11:00 am EDT on August 20th, insurers took note. The NHC Advisory not only said that the potential for a Northeast landfall was very real, it also indicated that a landfall at major hurricane intensity (Category 3 or higher) was quite possible.

That NHC August 20th forecast appears in the left-hand panel of Figure 3 below. The white cone represents the range of potential tracks for Hurricane Bill (usually called the “cone of uncertainty”). “M” designates the positions at which the NHC expected Bill to maintain major hurricane intensity (where its maximum sustained winds would be of at least 111 mph). The illustration on the right shows only the left-most side of the NHC cone, together with the NHC’s forecast maximum wind speeds along that potential track.

While the NHC did expect Hurricane Bill to weaken once it moved into cooler waters, wind shear was quite low through that part of the Atlantic basin at the time, and the expectation was that Bill’s maximum winds would diminish only to 115 mph (Category 3 level) by the time it reached the Northeast. Had this scenario actually played out, Hurricane Bill would have struck the Massachusetts coastline with stronger winds than Hurricane Ike hit the Texas coast last year.

In the actual event, Hurricane Bill followed a track farther from shore and also dissipated more quickly than forecast. This fact does not negate, however, the underlying risk that hurricanes pose for New England residents and insurers alike.

Figure 3: NHC Forecast “Cone of Uncertainty” (on the left), and (on the right) the NHC left-most scenario showing the potential track and intensity. The NHC August 20th forecast showed a potential major hurricane landfall near Cape Cod, Massachusetts.

What Conditions Were Necessary for Bill to Have Made Landfall at Hurricane Strength?

Hurricanes in the Northeast exist in a very different environment from the one in which they originate and mature. Most hurricanes develop within a favorable combination of warm sea surface temperatures (SSTs), high levels of humidity, and low levels of wind shear. The prevalence of these conditions from June to November in the Atlantic defines the hurricane season. However, in the late summer months these favorable conditions can extend well into the middle latitudes. This was the case when Hurricane Bill threatened to maintain its strength well outside of the tropics, as shown in Figure 3.

Figure 4, below, displays the SST conditions that were in place on Friday, August 21st. The diagram is a contour map of 27˚ Celsius SSTs (80.6˚ Fahrenheit)—the generally accepted temperature at which hurricanes can intensify—over which are superimposed regions of SST anomalies (temperatures higher or lower than the long-term average for this time of year), which are indicated by different colors. The diagram shows both that ocean temperatures off the Northeast coast were anomalously warm (the red- and orange-colored areas), and that they were warm enough to sustain hurricane intensity. This extended region of warm Atlantic waters represents the northern sector of the Gulf Stream, which has been quite robust this season.

Figure 4: Atlantic Sea Surface Temperatures (SSTs) and Anomalies on 8/21. Just south of New England the northern reaches of the Gulf Stream were warm enough (>27˚C) to support a hurricane. (Source: AIR)

Figure 5 shows the wind shear conditions in place at the same time (regions of low shear are colored blue). Wind shear less than 10 mph is conducive for hurricane development. Between Hurricane Bill and the Northeast coast is a large area of low wind shear. Although it is fairly unusual for the Northeast U.S. to have this combination of warm ocean temperatures and benign wind shear, it is exactly the type of environment that can sustain a hurricane’s intensity through to landfall.

Figure 5: Wind shear conditions on 8/21, the same day shown in Figure 4. Ahead of Hurricane Bill wind shear was very low (<10 mph), which supports hurricane intensity. (Source: AIR)

How Winds Are Distributed in Northeast Hurricanes

It has sometimes been suggested that all hurricanes have transitioned to extratropical cyclones by the time they reach the Northeast—and that this always reduces wind risk. Such misperceptions are encouraged by the fact that the last landfall of a major hurricane in the Northeast took place over 70 years ago when wind measurements were sparse and wind engineering research was in its infancy.

Today, however, we know that while most Northeast hurricanes do undergo some degree of transition from a tropical structure to an extratropical one, this change does not mean that risk can be discounted. For example, just last year Hurricane Ike produced record-breaking insured loss in the state of Ohio and substantial damage throughout the Midwest during its transition phase. Having become a weak disturbance once it was over land, Ike was re-energized as it moved north of Arkansas, where it interacted with a pre-existing disturbance moving out of Michigan. Ike produced wind gusts of hurricane force—nearly 1,000 miles from landfall.

In discounting risk, some point to the fact that transitioning hurricanes become asymmetric; that is, their damaging winds shift largely to the right of the storm track. This observation might indeed comfort residents of the Northeast who know that most hurricanes pass to the east of the region. If damage is likely to occur only on the east side of the track, then most Northeast hurricanes should be of little concern.

Such arguments have three flaws.

First, even when hurricanes do transition, their wind fields do not immediately shift to the east. In fact, transitioning by definition is a process that occurs over hours to days, and it is only when hurricanes have fully transitioned that their wind fields take on the features of extratropical storms. Also, Northeast residents very well know that blizzards (which are “pure” extratropical cyclones) can produce widespread wind damage over land—not just offshore and not just to the right of the track.

Second, transitioning storms can intensify. Thus, despite the rightward shift in its wind field, a transitioning storm might actually become stronger, making the shift in the position of its winds much less important than the strengthening of damaging winds close to the storm’s center. Hurricane Ike is again a good example of a storm that produced damaging winds on both sides of its track well after it had entered the mid-latitude environment and transitioned.

Third, not all Northeast hurricanes undergo transition. Hurricane Bill itself provides a good example. The upper diagram of Figure 5 below shows the wind field of Hurricane Bill on Sunday, August 23rd, well after Bill had entered the mid-latitude environment. As stated earlier, the Northeast environment was able to support Bill’s intensity well beyond a latitude where many hurricanes would either have transitioned or dissipated completely.

At a significant distance north, when it was tracking beyond the Canadian border, Bill was still considered to be a “pure hurricane” that exhibited the expected hurricane wind field structure. Figure 5 shows damaging winds (> 40 mph) extending well to the west of Bill’s storm track, and indicates even hurricane force winds to the left of the track.

In fact, because Bill was such a large storm, had it made landfall along the Cape Cod shore, damage would have been expected well into upstate New York and as far away as Montreal. Had Bill followed a track similar to the Great New England Hurricane of 1938 (shown in the lower diagram below), wind damage would have been expected over large parts of New England.

Figure 6: Upper diagram: Bill is a hurricane well into the mid-latitudes and hurricane- and tropical storm-force winds (colored maroon and orange respectively) persist well to the left of the track. Lower diagram: The overlay shows the same wind field but assuming a track similar to that of the Great New England Hurricane of 1938. (Source: NOAA/NHC).

What was the likelihood that Bill would follow a more westerly path? Hurricanes generally move with the atmospheric flow. In the tropical latitudes, storms generally follow the trade winds, which carry them from east to west. When hurricanes enter the mid-latitudes, they are likely to begin moving with the jet stream (a permanent belt of strong upper level winds), which turns storms to the north, then the northeast, and eventually to the east.

However, this is not always the case. Disturbances embedded in the jet stream can allow storms to track north without making a northeasterly turn (as in the case of the Great New England Hurricane), or they may even retrograde and move back to the northwest. Hurricane Ivan actually turned to the south when it exited the Delaware coast in 2004. The climatological mean, or expected, track for a U.S. hurricane at these latitudes is oriented to the northeast. However, just as storms in the tropical latitudes sometimes take unusual turns and loops, northeast hurricanes don't always follow expectation.

Conclusion

Bermuda and the Northeast United States are fortunate to have escaped significant damage by this year’s second named storm. Even as we breathe a collective sigh of relief, however, Hurricane Bill should serve as a reminder of the potential for hurricane risk outside the tropical latitudes—specifically with respect to three realities.

First, Northeast hurricanes do exist. And while Northeast landfalls are rare, their lower frequency should not lead to an underestimation of their potential to produce damage and loss. Even when Northeast hurricanes do not actually make landfall, they can still cause significant wind damage along the highly exposed Northeast corridor.

Second, Northeast hurricanes can persist and sometimes can even intensify in the mid-latitude environment. Last year’s Hurricane Ike is a recent example of this kind of underlying risk and also demonstrates that hurricanes can strengthen when they enter an otherwise harsh extratropical environment. The sustained winds the NHC projected in its three-day forecast for Hurricane Bill that showed a potential Cape Cod landfall were stronger than the winds experienced in Galveston from Hurricane Ike. Fortunately, this scenario did not play out. But it emphasizes the sure possibility of an intense hurricane event in the Northeast U.S.

Third, Northeast hurricanes do often transition as they move into the mid-latitudes. But this fact does not by itself mean that risk is reduced. The conditions in place in the Northeast at the end of this August could have supported the intense winds of Hurricane Bill, and even though Bill became weaker than was originally forecast, its swath of damaging winds still extended more than 100 miles to the west of its track. The perception that Northeast hurricanes produce damage only to the right of their tracks may be difficult to disprove conclusively using the sparse data of the 1930s, but the belief simply does not match the realities of recent experience—the most recent being Hurricane Bill.

Before Hurricane Bill is committed to the history books, notice should be taken of “what might have been.”

__________________

¹One of the National Oceanic and Atmospheric Administration (NOAA) forecasting models.