Dr Steve Smith explores the climatic factors affecting hurricane activity and landfalls

Intense work by meteorologists on climate patterns within the Atlantic basin is allowing us to understand the nature of hurricanes better, and thereby is helping us to produce better exposure models.

Outlooks for the 2005 hurricane season

A summary of the most recent forecasts (as of 28 June 2005) is presented in Table 1. "Intense Hurricanes" refers to hurricanes of Saffir-Simpson category 3 or higher, and the ACE (Accumulated Cyclone Energy) is a general measure of tropical cyclone activity in the Atlantic basin. These forecasts are based on statistical relationships between hurricane activity in the future and climate variables observed at the present time.

Above average activity predicted

All the forecasts predict above-average activity during the 2005 season.

The NOAA forecast indicates that the likelihood of an above-average season is 70%. The general climactic conditions driving these forecasts arise from an expectation that there will be a neutral to weak El Nino, coupled with a continuation of the warm Atlantic sea-surface temperatures. A point should be noted: the 2005 season forecasts are slightly higher than those made at the same time for the 2004 and 2003 seasons.

Insurers and reinsurers should note, however, that there is no guarantee that "above average" means an above average incidence of landfall. In fact, the number of landfalls is only weakly correlated with the number of storms in the Atlantic basin.

The 2004 hurricane season

2004 saw four significant hurricanes make landfall. Three of those made landfall as major hurricanes, causing the majority of their damage to the State of Florida. The level of losses for the season, well over $20bn in insured losses, represents the most costly hurricane season ever, surpassing even 1992 (unadjusted for inflation and societal shifts). As predicted in previous ReAdvisory papers, insurance claims for the 2004 season have gradually increased. The increases are due to four factors:

- demand surge;

- lack of available loss adjusters;

- waiving of second and third event deductibles; and

- the effect of prior damage on subsequent storm losses.

One unusual characteristic was the seriality of storms, as they occurred within a six-week period and followed similar tracks across the Atlantic.

At present, the available catastrophe models do not include the effects of serial storms. Risk Management Solutions have estimated in preliminary work that including seriality would increase aggregate losses by approximately five percent, although some tests have indicated an increase of 15% - 20%. If this becomes reality, we predict insurers will critically evaluate their vertical and horizontal appetites for risk retention versus exposure.

Based on a simple statistical analysis, the insurance industry should expect another season with losses comparable to 2004, with a return period of roughly 50-70 years. However, as the activity rates for hurricanes in the Atlantic basin are in an enhanced phase (due to a number of climatic factors), the return period might be shorter, at least in the medium term (20-40 years).

While some of the statistics for the 2004 season are impressive, they are not unprecedented. Three hurricanes made landfall in Florida in 1964 and two major hurricanes made landfall in Florida in 1950. Eight named storms made landfall in 1916.

Climatic factors affecting hurricane activity and landfalls

Hurricanes are affected by a number of climatic variables, which influence both activity and landfall. The earlier years of this decade saw few hurricane landfalls, even though hurricane activity in the Atlantic was distinctly above average. Figure 1 highlights the total number of hurricanes in the Atlantic basin and those that made landfall on the US East Coast during the main hurricane season (May through October).

This paucity of landfalling hurricanes was somewhat unexpected given the increased activity. Several academic papers, have noted that the recent period (since 1994) has seen an increase in hurricane activity. The reason given is two-fold:

- An increase in North Atlantic sea-surface temperatures which makes more energy available for hurricanes to form.

- A decrease in vertical wind shear, which provides a "safe" environment in which hurricanes can grow.

Since these conditions can persist for decades, this period of enhanced hurricane activity can persist for anywhere from 10 to 40 more years.

The increased activity in the Atlantic due to this cycle in sea-surface temperature represents a general tendency. The actual activity and landfalls, year-on-year, will be influenced by several other factors, described below.

The Bermuda High and North Atlantic Oscillation (NAO)

The Bermuda High is a persistent weather feature over the Atlantic during the summer - it is a stable area of high pressure which has a corresponding area of low pressure over the continental US. Between the two sits a trough which, when fortuitously positioned, protects the US East Coast from hurricane landfalls. Research by ReAdvisory has quantified this situation - by developing a Bermuda High Index (BHI). ReAdvisory has shown that there have been landfalls along the US East Coast when the trough has been 'out of position' (or the BHI has a high value). Figure 2 shows the correlation of the BHI with East Coast landfalls. From ReAdvisory's work, the position of the Bermuda High emerges as the most important factor in explaining the lack of hurricane landfalls from 2000 to 2003. The Bermuda High is a weather system and, as such, is predictable some 5-10 days in advance.

However, prediction for the behaviour of the Bermuda High over longer (ie seasonal) periods is beyond current knowledge.

The Bermuda High is seen as an expression of a larger scale atmospheric cycle, the North Atlantic Oscillation (NAO) which is currently the subject of intense study by academia. It is hoped that the NAO (and the Bermuda High by extension) will gain a greater element of predictability in the wake of this research.

El Nino Southern Oscillation (ENSO)

The cycle of tropical atmosphere and ocean events are most evident in the anomalously warm and cold ocean temperatures in the central and eastern equatorial Pacific Ocean. The anomalously warm state is referred to as El Nino; the corresponding cold state as La Nina. This cycle can manipulate the weather over much of the globe and has a significant effect on hurricane activity:

- Hurricane activity in the Atlantic basin is reduced dramatically during the El Nino phase. This is caused by an increase in vertical shear over the Atlantic due to El Nino.

- During El Nino, hurricanes in the Atlantic basin are weaker and shorter-lived.

- Hurricane tracks in the Atlantic shift northward during El Nino.

La Nina (cold) events show the opposite effects on hurricanes. For example, during a cold event, the Atlantic basin experiences more and stronger hurricanes. It should be noted that this is the average behaviour. Hurricane Andrew occurred during an El Nino event when fewer, weaker hurricanes would be expected in the Atlantic basin.

There are other climate variations which can have an impact on hurricane activity and landfalls, such as the Quasi-Biennial Oscillation and the Madden-Julian Oscillation. These are, however, beyond the scope of this article.

Dr Smith is vice president of ReAdvisory, a service of Carvill, and spearheads ReAdvisory's catastrophe modelling efforts. Dr Smith has been researching weather phenomena for 11 years including the last five years in the insurance and reinsurance market.