Active hurricane seasons do not automatically mean massive losses, though tropical storms are quite capable of widescale damage, says Richard Dixon.
Early predictions are suggesting another active hurricane season in the Atlantic Basin. Prior to the start of the 2001 season, the forecasts from Bill Gray's team at Colorado State University predicted above-average hurricane activity in the Atlantic Basin with 12 named storms and seven hurricane-strength storms. The season's tally of 14 tropical storms, eight hurricanes and four intense (greater than category 3) hurricanes confirmed these expectations as a well-above-average hurricane season. However, the 2001 hurricane season should remind us that high hurricane activity need not mean high losses.
As with other recent years, 2001 forecasts revealed a significant shortfall in the forecasts for landfalling storms. The US, for the second year running, did not have a single landfalling hurricane, despite Gray's forecast of a 69% probability for a major (category 3-5 hurricane) hitting the US coast. In the US, the main losses in the 2001 season were linked with near-stationary Gulf Coast tropical disturbances that resulted in flooding and little wind damage, underscoring the imperfect correlation between hurricane forecasts and hurricane losses, and the recent trend of increased basin activity with decreased number of landfalls.
Figure 1 shows the tracks of the named tropical systems from 2001. The three named storms (Allison, Barry and Gabrielle) that struck the US mainland in 2001 only reached tropical storm status. However, Allison stood out as by far the most devastating storm of the 2001 season. Springing to life within the first five days of the season, Allison dropped record-breaking rains as it stalled over Texas for six days and then moved over Louisiana, Georgia, the Carolinas and up the eastern seaboard. In all, Allison caused 50 deaths, over $6bn in total damages, and more than $2.5bn in insured losses.
The metropolitan region of Houston, Texas, accounted for 90% of the losses from Tropical Storm Allison. In just six days, Allison dropped as much as 37 inches (94 cm) of rain, or 80% of the average annual rainfall, on the Houston area, a large proportion of this precipitation falling in one 12-hour period. Houston's bayous and storm water drainageways were unable to cope with the heavy rainfall. Insured losses were heaviest in the Texas Medical Center and Downtown areas, where water travelled via underground pedestrian tunnel systems, flooding subterranean parking garages and basements, some containing expensive equipment.
Although exceptional for Houston, Allison is not a record for Gulf Coast rainfall. Tropical Storm Claudette in 1979 dropped 45 inches of rain in a day (a US 24-hour precipitation record) at Alvin, Texas. Other towns along the Gulf Coast with similar extreme rainfall risk to Houston include New Orleans, Beaumont, Mobile and St Petersburg.
Four storms in 2001 reached category 3 status or greater, but none came close to the US. Iris (October 4-9) and Michelle (October 29-November 6) developed over the warm waters of the Caribbean, intensifying into category 4 storms before making landfall in Belize and Cuba, respectively. When Iris made landfall in Eastern Belize, housing in several coastal towns was almost totally destroyed and 18 people were killed. Michelle dropped heavy rains on Honduras and Nicaragua and then strengthened as it moved out to sea before making landfall in Southern Cuba as a category 4 hurricane, the worst to hit the island since Hurricane Fox in 1952. More than 90,000 homes were damaged and 45,000 destroyed, and electrical and telecommunications systems were crippled in Cuba, where property damage is estimated at more than $2bn. Michelle then moved east, weakening to a category 1 hurricane before crossing the Bahamian islands, causing a further $150m in damage. Erin (September 1-15) and Felix (September 7-19), both category 3 systems, developed over the Atlantic, with Felix staying well away from land during its northward track, and Erin only scraping past Newfoundland as it transitioned into an extra-tropical cyclone.
Why did so few hurricanes make landfall? Figure 1 also shows the region where the June-October upper level winds were more than 3.3 feet/sec (1 metre/sec) stronger than the average from 1970-2000. This region of stronger winds highlights a branch of a strong upper-level trough that appears to have existed over mainland US in recent hurricane seasons. Although the past seven years have seen higher temperatures in the tropical Atlantic basin, leading to more hurricane activity, it has been suggested that the large-scale south-westerly flow from this trough has acted to curve any potential landfalling systems away from land. Meteorologists have seen the position of the trough as somewhat fortuitous; without its presence, the chances of an intense landfalling hurricane would increase, as there is no large-scale forcing to direct the systems away from land.
The current prospects for this season are similar to the predictions made prior to the start of the 2001 season. Bill Gray's team at Colorado State University forecast late last year that the 2002 season will see 13 named storms, eight hurricanes and four intense hurricanes - another year well above the 1951-2000 average. Despite last year's relative inactivity regarding landfalling systems, the forecast, as for 2001, still assigns a strong chance (86%) of an intense (category 3-5) hurricane landfalling somewhere along the US coastline.
As shown in figure 2, in the 30 years prior to 1995, hurricane seasons were generally less active, with an average of only one or two intense hurricanes per season (indicated by B). However, this did not preclude the chance of intense landfalling hurricanes - the 1992 season serves as a reminder. Only five systems developed during that season, however, the one landfalling system turned out to be Andrew, which devastated southern Florida. The forecasting team at Colorado State University has emphasised that since 1995 (and also pre-1965), a change in the large-scale circulation in the Atlantic Ocean led to anomalously warm temperatures in the regions where hurricanes spawn - leading to more frequent formation of hurricanes (indicated by A).
Historical evidence has demonstrated that timescales for such changes in the regional oceanic temperatures are in the order of decades, and the forecasting team feels that the early part of the 21st century may be dominated by active hurricane seasons, as was the spell prior to 1965. What remains to be seen is whether the upper-level winds that have spared the US from major landfalling hurricanes in the last few years are prominent again this season.
By Richard Dixon
Richard Dixon is a senior catastrophe analyst with catastrophe model provider Risk Management Solutions.