I am pleased to present information to your sub-committee on the outlook for the 1999 Atlantic basin hurricane season and the prospects for increased hurricane related damage on the U.S. coast in coming decades. Trends in global oceanic and atmospheric observations during recent years indicate that we are entering (or reverting to) a multi-decadal period of increased intense or ``major" hurricane activity. The latter include hurricanes of Saffir/Simpson intensity classes 3, 4, and 5, the salient qualities of which include maximum sustained winds greater than 110 mph. Major hurricanes, when normalized by coastal population, wealth per capita, and inflation over the last century typically cause about 85 percent of the total U.S. tropical cyclone destruction even though they only account for one-fourth of the total number of ``named" tropical storms. If our interpretation of the current direction of multi-decadal climate trends proves valid and we are indeed entering a period of intense hurricane activity more typical of the 1940s and 1950s, then the cost of U.S. hurricane-spawned destruction will most certainly rise to unprecedented magnitudes. This assessment is based on the large increase in population and property values along the U.S. Southeast Coast since the last active period for major hurricane landfall which occurred between the mid-1940s through the mid-1960s. Note, however, that the last four years (1995-1998) have been the most active consecutive four-year period of hurricane activity on record; this despite the large downturn during 1997 due to the effects of the strongest El Nino event on record. The characteristic time frame of prior active periods suggests that this new active period will likely persist for the next few decades.  

Table 1 provides a summary and comparison of Atlantic basin hurricane activity of the last four years versus the level of activity in other recent periods. The West Atlantic had 15 major hurricanes and 36 major hurricane days during 1995-1998. As shown, the average for these four-seasons are 250 and 419 percent of the annual average of hurricane activity during the recent quarter century period 1970-1994. We attribute the recent upturn to the large scale effects of a major reconfiguration of the "thermohaline" circulation of the Atlantic Ocean leading to alterations in surface pressure and sea surface temperature throughout much of the Atlantic basin. Various historical and geological records suggest that the thermohaline circulation varies on multi-decadal scales as well as on century--millennium time scales. (See Appendix B for a discussion of the physical factors which cause Atlantic Ocean thermohaline circulation trends this time scale). We emphasize that these are "natural" changes with numerous prior occurrences in the historical and geological climate data which are unrelated to any type of hypothesized human-induced global warming.

Table 1: Comparison of recent 1995--1998 hurricane activity with activity in other recent periods.

A sharp downturn in Atlantic basin major hurricane activity began in the late 1960s. This downturn in relation to earlier levels of activity is especially obvious in the large decrease of major hurricanes affecting the U.S. East Coast and Caribbean during 1970--1994. Comparative values for hurricane activity during these two recent periods include 15 major U.S. hurricane strikes along the U.S. East Coast during the 23-year period between 1947--1969 versus only three such strikes during the more recent 25 years between 1970--1994 (Figure 1).      

Concurrent with these changes in the Atlantic Ocean are global scale alterations of Sea-Surface Temperature (SST) patterns wherein somewhat cooler SSTs tended to develop over much of the Northern Hemisphere during 1970-1994 while much of the Southern Hemisphere ocean surfaces warmed. A generally opposite trend in global SST patterns was present during the earlier 1947-1969 period. In addition to the 1970-1994 decrease of major hurricane activity, other notable climate trends during the 1970--1994 period also appear to be associated with the changes in the Atlantic Ocean circulation. These include the long running (1970-present) West African Sahel drought, increased frequency and intensity of El Nino events and generally stronger middle latitude westerly wind patterns; these among numerous additional climatologically important changes. In general, the global circulation trends which prevailed during 1970--1994 are known to inhibit major hurricane activity on season-to-season and longer time frames.  

The prospect of returning to a period of enhanced major hurricane activity more typical of the mid-1940s to late-1960s accentuates the vastly increased vulnerability of U.S. coastal areas which are now much more developed in comparison with earlier periods. Consequently, I view the hurricane threat to be our country's greatest natural hazard; more than earthquakes (see Appendix D), and much more than floods, tornadoes, extreme temperatures, global warming, etc. Sadly, the large threat posed by landfalling major hurricanes has yet to be fully recognized, either by the general public or by most U.S. government officials. In fact, there is no U.S. government or private industry program of which I am aware that specifically addresses relationships between the U.S. landfalling hurricanes and global climate variability. This situation is likely the combined result of (1) Key Federal agencies having little knowledge or appreciation of the compounding effects of population growth and climate variability in increasing the magnitude of the hurricane threat and (2) of the grossly disproportioned diversion of attention and resources to global warming issues; issues that have no relevance to the potential human peril in lives and property loss expected from the increase of major hurricane activity now apparently underway. This out-of-balance funding allocation needs to be addressed.

Economic Considerations

When we account for increased U.S. coastal population, the costs of durable goods per capita, inflation and loss of economic gain during the reconstruction period following a hurricane, it is possible to make cost-normalized estimates of U.S. damage which previous (historical) hurricanes would have wrought, if they had crossed the U.S. coastline and impacted on today's more developed conditions.

Pielke (Jr.) and Landsea (1998) have estimated that landfalling intense hurricanes cause 80-85 percent of the normalized (eg, by population, inflation, and wealth) hurricane generated damage. Analysis of their recently issued damage survey (plus additional analyses by the author) indicate that a category 4-5 hurricane making landfall (18 during this century) typically causes about 64 times the amount of damage associated with the average landfalling category-1 hurricane. A category 3 cyclone coming ashore causes about 16 times the damage of a category-1 storm; a category-2 cyclone causes about 4 times the damage of a category-1 hurricane. A typical Tropical Storm (TS) causes about half the damage of a category-1 storm.  

Table 2 provides a summary on the number of U.S. landfalling tropical cyclones for various intensity categories during the 100 year period 1899-1998 plus the author's estimated relative destruction potential by cyclone intensity class. Note how rapidly potential hurricane-spawned destruction rises in going from intensity category 1 to category 4-5. Using this standard of destruction, a relative measure of destruction potential can be computed for the entire U.S. coastline. For example, during the last 100 years, South Florida has experienced 35 times the (potential) destruction as has the coastal region from Fort Pierce, Florida to the Georgia-South Carolina border to the north. No major hurricanes made landfall along this coastal strip in the last 100 years.  

Table 2: Comparative statistics for andfalling named storms in U.S. (1899-1998) showing normalized damage typically associated with each intensity class.

Comparison of the 1950s with the 1970s provides an illustration of the long-period climate-linked variability that can occur for major hurricane landfall frequency. The climate specified mean annual rate of major hurricane landfall in the Florida Peninsula and the U.S. East Coast during the 1970s was 0.025 which is only 3.5 percent of the value of the 1950s (.728 per year), a ratio of 29 to 1. Seven major hurricane landfall events occurred on the Florida and East Coast during the 1950s whereas no major landfalling hurricanes occurred during the 1970s.

Annual hurricane-spawned damage of the 1990s has been estimated to average about $5 billion a year. Thirty years ago the figure that was about $0.5 billion per year. If we extrapolate the current trends of U.S. coastal population and property values into the future, it follows that this figure will rise to $25-50 billion per year by the year 2030. Most hurricane related destruction during the last century came as intermittent catastrophic events every 5-15 years when a major hurricane made landfall in an especially vulnerable U.S. coastal area. The 10 most destructive U.S. hurricanes (normalized to today's dollar amounts) of the last century caused about 60 percent of all hurricane linked damage during this period. As coastal property values keep increasing, we will likely be looking at occasional individual major hurricanes coming ashore at a vulnerable coastal location and causing damage amounts of $50-100 billion or more.    

Predictions for 1999 Atlantic Basin Activity

Information obtained through March 1999 indicates that 1999 Atlantic hurricane activity is likely to be above the average for the 1950--1990 period with 9 hurricanes (average is 5.8), 14 named storms (average is 9.3), 4 intense (category 3-4-5) hurricanes (average is 2.2). Collectively, net tropical cyclone activity for 1999 is expected to be about 160 percent of the long term average. Consequently, the 1999 season should have hurricane activity comparable to the recent busy hurricane seasons of 1996 and 1998.      

Landfall Probabilities for 1999

A new initiative in our research involves efforts to develop extended range forecasts of the probability of hurricane landfall in a given season along the U.S. coastline. Whereas specific individual hurricane landfalls can never be accurately forecast months in advance, the seasonal probability of landfall for an individual year can be forecast with statistical skill. Statistical aspects of hurricane landfall are functions of varying year-to-year and longer term climate signals and can thus be specified. Probability specification can be accomplished by statistical studies of U.S. hurricane and named storm landfalls for the last 100 years (1899--1998). Specific landfall probabilities can be estimated for all cyclone intensity classes for each of 11 U.S. coastal regions. Landfall probability has been found to be (statistically) related to the overall Atlantic basin Net Tropical Cyclone Activity (NTC) in any year and to climate related trends linked to multi-decadal variations of the Atlantic Ocean thermohaline circulation. Distinctive landfall characteristics occur for the Gulf Coast area versus the rest of the U.S. coast (Fig. 2). Table 3 lists current values for landfall probabilities for a range of Tropical Storms (TS), Cat 1-2, and Cat 3-4-5 hurricanes impacting the whole U.S. coastline, the Gulf Coast and Florida and the East Coast for 1999.        

In terms of the comparative rate or number of landfall events, our analysis for 1999 indicates a probability of major hurricane landfall for the Florida Peninsula and East Coast which is 205 percent of the average for the last 100 years and 145 percent (of the 100 year average) on the Gulf Coast. The expectation for major hurricane landfall for Florida and the East Coast by this method for 1999 are 5.5 times larger than the average year during the recent quarter century quiescent period of 1970-1994.

Thank you for inviting me to present this information.

Table 3: Estimated percent probability of one or more U.S. landfalling Tropical Storms (TS), category 1-2 hurricanes, and category 3-4-5 hurricanes, total hurricanes and total named storms along the entire U.S. coastline, along the Gulf Coast, and along the Florida Peninsula and the East coastline for 1999.

The attached Appendices for more detailed information.

REFERENCES  

Gray, W. M., C. Landsea, P. Mielke, Jr., and K. Berry, 1999: Early April forecast of Atlantic seasonal hurricane activity and U.S. landfall strike probabilities for 1999. Dept. of Atmos. Sci. Paper, Colo. State Univ., Ft. Collins, CO, 80523 [available on the World Wide Web: http://tropical.atmos.colostate.edu/forecasts/index.html].  

Pielke, R. and C. W. Landsea, 1998: Normalized Atlantic hurricane damage, 1925--1995. Weather and Forecasting, 13, 621-631. baselineskip 4mm    

APPENDICES

For reference and more background information on this topic please see the attached papers which has been sent to the U.S. House Sub-committee.

• A) Preliminary Outline of Some Current Research at Colorado State University on Seasonal Forecasting the Probability of U.S. Hurricane Landfall

• B) Climate Trends Associated With Multidecadal Variability of Atlantic Hurricane Activity

• C) Normalized Atlantic hurricane damage, 1925--1995 (by Roger A. Pielke, Jr. and Christopher W. Landsea.

• D) Comparison of U.S. Hurricanes Versus Earthquake Damage

• E) Most recent forecast for the 1999 Atlantic basin hurricane season with landfall probabilities.