The Deadliest, Costliest, and Most Intense United States Tropical Cyclones From 1851 to 2004

This version of the Deadliest, Costliest, and Most Intense United States Tropical Cyclones extends the work of Jarrell et al. (2001) to begin with the year 1851. New updates include data from the period 1851-1899 provided by the best track reanalysis project headed by Chris Landsea, some significant revisions to the period 1900-1914 and a revised intensity of Hurricane Andrew [Landsea et al. (2004)]. A new feature for this update is a list of landfalling hurricanes during this era, updating and supplementing information provided in Neumann et al. (1999). The paper continues the methodology of Jarrell et al. (2001) in producing an estimate of the monetary loss that historical hurricanes could exact on the current property-at-risk in the same location.

During 1995, the National Meteorological Center, which included the National Hurricane Center, was re-organized into the National Centers for Environmental Prediction (NCEP). Under NCEP, the National Hurricane Center became the Tropical Prediction Center (TPC), a name which more accurately reflects the broad scope of its responsibilities, and more formally publicizes that the majority of its operational products were for tropical weather events exclusive of hurricanes. The name “National Hurricane Center” was retained to apply to the hurricane operations desk at TPC. We will follow the convention where “NHC” refers to the previous National Hurricane Center, “TPC” refers to the current center and “TPC/NHC” refers to the hurricane operations desk of TPC.

This technical memorandum lists the deadliest and costliest tropical cyclones in the United States during 1851-2004. The compilation ranks damage, as expressed by monetary losses, in three ways: 1) contemporary estimates; 2) contemporary estimates adjusted by inflation to 2004 dollars; and 3) contemporary estimates adjusted for inflation and the growth of population and personal wealth [Pielke and Landsea, 1998] to 2004. In addition, the most intense (i.e., major¹ ) hurricanes to make landfall in the United States during the period are listed. Some additional statistics on United States hurricanes of this and previous centuries, and tropical cyclones in general, are also presented.

1. INTRODUCTION

The staff of the Tropical Prediction Center receives numerous requests for statistical information on deaths and damages incurred during tropical cyclones affecting the United States. Information about their intensity is also frequently of interest. Estimates of these measures vary in the literature. Our hope is to present the best compilation of currently available estimates. In some instances, data in our lists represent revised estimates based on more complete information received following earlier publications including previous versions of this technical memorandum.

There are other frequently asked questions about hurricanes, such as: What is the average number of hurricanes per year? What year(s) had the most and least hurricanes? What hurricane had the longest life? On what date did the earliest and latest hurricane occur? What was the most intense Atlantic hurricane? What was the largest number of hurricanes in existence on the same day? When was the last time a major hurricane or any hurricane hit a given community directly²? Answers to these and several other questions are provided in Section 3.

Many of the statistics in this publication depend directly on the criteria used in preparing another study, “Hurricane Experience Levels of Coastal County Populations Texas to Maine"” [(Jarrell et al. (1992)]. The primary purpose of that study was to demonstrate, county by county, the low hurricane experience level of a large majority of the population. Statistics show that the largest loss of life and property occur in locations experiencing the core of a category 3 or stronger hurricane.

The Saffir/Simpson category is defined by pressure, wind, and storm surge. In nature, however, there is not a one-to-one relationship between these elements. Therefore, in practice, the TPC uses the maximum wind speed to establish the category. Operationally, however, the central pressure is often used to make a first estimate of the wind. Thereafter, available surface wind reports, aircraft reconnaissance flight-level winds (from which surface wind speed can be estimated), and dropsonde data are used to anchor the wind estimate. In post-storm analysis, the central pressure ranges of hurricanes on the Saffir/Simpson Hurricane Scale will usually agree fairly well with the wind ranges in that category. On the other hand, the storm surge is strongly dependent on the slope of the continental shelf (shoaling factor). This can change the height of the surge by a factor of two for a given central pressure and/or maximum wind.

Heavy rainfall associated with a hurricane is not one of the criteria for categorizing.

The process of assigning a category number to a hurricane is subjective, as is TPC’s estimate of a cyclone’s impact . It is made on a county by county basis. In this study, we continue to use criteria for direct hit and indirect hit described in the work by Jarrell et al. (1992):

Direct Hit Using "R" as the radius of maximum winds in a hurricane (the distance in miles from the storm's center to the circle of maximum winds around the center), all or parts of counties falling within approximately 2R to the right and R to the left of a storm's landfall point were considered to have received a direct hit. (This assumes an observer at sea looking toward the shore. If there was no landfall, the closest point of approach was used in place of the landfall point). On average, this direct hit zone extended about 50 miles along the coastline (R15 miles). Of course, some hurricanes were smaller than this and some, particularly at higher latitudes, were much larger. Cases were judged individually, and many borderline situations had to be resolved.

Indirect Hit In general, areas on either side of the direct hit zone which received wind gusts of hurricane force and/or tides of at least 4 to 5 feet above normal were considered to have had an indirect hit. The evaluation subjectively incorporated a hurricane's strength and size, and the configuration of county coastlines.

The authors acknowledge that there are limitations to this technique. For example, the effect of an indirect hit by a large category 4 hurricane can be greater than that by a direct hit from a small category 1 hurricane.

Neumann et al. (1999) gives the variation in tropical cyclone frequency along the United States coastline for all tropical storms and hurricanes, hurricanes only, and major hurricanes (category 3 or greater). In that study, counts were made of the number of tropical cyclones or hurricanes whose center passed within 75 nautical miles of the coastal location. This counting method thus includes near-misses, as well as direct and indirect hits as defined above.

Statistics on tropical storm and hurricane activity in the North Atlantic Ocean (which includes the Gulf of Mexico and the Caribbean Sea) can be found in Neumann et al. (1999). A stratification of hurricanes by category which have affected coastal counties of the Gulf of Mexico and North Atlantic Ocean can be found in Jarrell et al. (1992). Additional information about the impact of hurricanes can be found in annual hurricane season articles in Monthly Weather Review , Storm Data and Mariner’s Weather Log.

3. DISCUSSION

Part I

The remainder of this memorandum provides answers to some of the most frequently asked questions about the characteristics and impacts of the tropical cyclones to affect the United States from 1851-2004.

(1) What have been the deadliest tropical cyclones in the United States? Table 2 lists the tropical cyclones that have caused at least 25 deaths on the U.S. mainland 1851-2004. The Galveston Hurricane of 1900 was responsible for at least 8000 deaths and remains #1 on the list. The death total from the Lake Okeechobee Hurricane of 1928 has been revised to include work from Pfost (2003) to reflect that the hurricane killed at least 2500 people. Tropical Storm Allison in 2001 caused torrential flooding in the Houston area and is the most significant addition since 2000 to the list. However two powerful hurricanes that struck in 1893 are now #3 and #4 on the list. A tropical storm which affected southern California in 1939 and the deadliest Puerto Rico and Virgin Islands hurricanes are listed as addenda.

(2) What have been the costliest tropical cyclones in the United States? Table 3a lists the 30 costliest tropical cyclones to strike the U.S. mainland 1900-2004. No monetary estimates are available before 1900 and figures are not adjusted for inflation. The 2004 hurricane season had the second, third, fourth and sixth most-costly systems to strike the United States. Table 3b re orders the first list and adds several other hurricanes after adjusting to 2004 dollars³. Hawaiian, Puerto Rican and Virgin Island tropical cyclones are listed as addenda to Tables 3a and 3b. Table 3b also lists the thirty costliest hurricanes 1900-2004 assuming that a hurricane having the same track, size and intensity as noted in the historical record would strike the area with today’s population totals and property-at-risk. See Pielke and Landsea (1998).

(3) What have been the most intense hurricanes to strike the United States? Table 4 lists the 60 most intense major hurricanes to strike the U.S. mainland 1851-2004. Hurricanes are ranked by estimating central pressure at time of landfall. Hawaiian, Puerto Rican and Virgin Island hurricanes are listed as addenda to Table 4.

A look at the lists of deadliest and costliest hurricanes reveals several striking facts: (1) Fourteen out of the fifteen deadliest hurricanes were the equivalent of a category 3 or higher. (2) Large death totals were primarily a result of the 10 feet or greater rise of the ocean (storm surge) associated with many of these major hurricanes. About three-quarters of the deadliest hurricanes were major hurricanes. (3) A large portion of the damage in four of the fifteen costliest tropical cyclones (Table 3a) resulted from inland flooding caused by torrential rain. (4) One-third of the deadliest hurricanes were category four or higher, but only one-seventh of the costliest hurricanes met this criterion. (5) Only five of the deadliest hurricanes occurred during the past twenty five years in contrast to three-quarters of the costliest hurricanes (this drops to one-half after adjustment for inflation and about one-third after adjustment for inflation, population, and personal wealth).

Addenda to tables 2 through 4 include some noteworthy storms from the U.S. West coast and the Hawaiian Islands, as well as in the U. S. Caribbean Islands. The rank represents the position they would occupy if included in the main table.

Table 5 summarizes the direct hits on the U. S. mainland since 1851. The data indicate that an average of 3 major hurricanes every 5 years made landfall somewhere along the U.S. Gulf or Atlantic coast. (All categories combined average about 5 hurricanes every 3 years.) Note that not all areas of the U.S. were settled before 1900 and there could be substantial gaps in landfall data coverage, especially in South Florida. For more details see Landsea et al. (2004b).

One of the greatest concerns of the National Weather Service's (NWS) hurricane preparedness officials is that the statistics in Table 2 will mislead people into thinking that no more large loss of life will occur in a hurricane because of our advanced technology. Max Mayfield, spokesman for the NWS hurricane warning service and Director of TPC, as well as former NHC Directors, have repeatedly emphasized the great danger of a catastrophic loss of life in a future hurricane if proper preparedness plans for vulnerable areas are not formulated, maintained and executed.

The study by Jarrell et al. (1992) used 1990 census data to show that 85% of U.S. coastal residents from Texas to Maine had never experienced a direct hit by a major hurricane. This risk is higher today as an estimated 50 million residents have moved to coastal sections during the past twenty five years. The experience gained through the landfall of Charley, Ivan, Jeanne, Andrew and Hugo has not lessened an ever-growing concern brought by the continued increase in coastal populations.

Table 6, which lists hurricanes by decades since 1851, shows that during the forty year period 1961 2000 both the number and intensity of landfalling U.S. hurricanes decreased sharply! Based on 1901 1960 statistics, the expected number of hurricanes and major hurricanes during the period 1961 2000 was 75 and 28, respectively. But, in fact, only 55 (or 74%) of the expected number of hurricanes struck the U.S. with only 20 major hurricanes or 71% of that expected number. Even the very active late 1990s showed below average landfall frequencies. It could be noted that of the most recent four decades, only the 70's and 80's were significantly below normal in terms of overall tropical cyclone activity.

During the past 35 years, the United States has experienced three Category 4 or stronger hurricanes: Charley in 2004, Andrew of 1992 and Hugo of 1989. However, on the average, a category 4 or stronger hurricane strikes the United States once every 6 or 7 years. This suggests we have seen fewer exceptionally strong hurricanes than an expected 35 year average of about 5 or 6. Fewer hurricanes do not necessarily mean a lesser threat of disaster, however. Records for the most intense U.S. hurricane in 1935, and the costliest, Andrew in 1992, occurred in years which had much below-average hurricane activity.

A large death toll in a U.S. hurricane is still possible. The decreased death totals in recent years could be as much a result of lack of major hurricanes striking the most vulnerable areas as they are of any fail proof forecasting, warning, and observing systems.

Continued coastal growth and inflation will almost certainly result in every future major landfalling hurricane (and even weaker hurricanes and tropical storms) replacing one of the current costliest hurricanes. For example, 4 out of 6 hurricane landfalls of 2004 made the top 30 list.

If warnings are heeded and preparedness plans developed, the death toll can be reduced. In the absence of a change of attitude, policy, or laws governing building practices (codes and location) near the ocean, however, large property losses are inevitable.

Part II

This section answers some frequently asked questions about tropical storm and hurricane activity.

(1) What is the average number of hurricanes per year?
Table 7 gives the average number of tropical cyclones which reached tropical storm, hurricane and major hurricane strength during selected time periods. A total of eleven tropical systems reaching storm strength with six of these becoming hurricanes and two attaining major hurricane status are the best averages to use based on the past 40 year time period of routine satellite surveillance.

(2) What year(s) have had the most and least hurricanes?
Table 8a shows the years of maximum and minimum tropical storm and hurricane activity for the Atlantic hurricane basin. Table 8b lists the years of maximum United States hurricane landfalls. The only times that the U.S. mainland has gone as long as two years without a hurricanes are 1862-64, 1930-31, 1981-82 and 2000-01. Note there is considerable uncertainty before 1900 because significant areas of the Gulf and Southeast Atlantic coasts were unpopulated and uninstrumented. The largest number of hurricanes to strike in one year was seven (1886), with six occurring in 1916, 1985, and 2004, plus five in 1893, 1909 and 1933. Three or four hurricanes have struck the U.S. in one year a total of 37 times. Eleven U.S. hurricanes were recorded in the two-year period 1886-87 with 15 recorded from 1886-1888.

(3) When did the earliest and latest hurricanes occur?
The hurricane season is defined as June 1 through November 30. An early hurricane can be defined as occurring in the three months prior to the start of the season, and a late hurricane can be defined as occurring in the three months after the season. With these criteria the earliest observed hurricane in the Atlantic was on March 7, 1908, while the latest observed hurricane was on December 31, 1954, the second “Alice” of that year which persisted as a hurricane until January 5, 1955. The earliest hurricane to strike the United States was Alma which struck northwest Florida on June 9, 1966. The latest hurricane to strike the U. S. was late on November 30, 1925 near Tampa, Florida.

(4) What were the longest lived and shortest lived hurricanes?
The third system of 1899 holds the record for most days as a tropical storm (28) and major hurricane (11.5), while Ginger in 1971 holds the record for the most days as a hurricane (20). There have been many tropical cyclones which remained at hurricane intensity for 12 hours or less.

(5) What were the strongest and weakest hurricanes?
In terms of central pressure (and probably winds), the strongest observed hurricane in the Atlantic basin was Gilbert in 1988 with a pressure of 888 millibars in the northwestern Caribbean with estimated sustained winds of 185 mph. The 1935 Labor Day hurricane in the Florida Keys, with a pressure of 892 millibars, was the most intense hurricane to strike the United States. Numerous hurricanes have reached only the minimum wind speed near 74 miles per hour and struck the United States.

(6) How many hurricanes have there been in each month?
Table 9, adapted from Neumann et al. (1999), shows the total and average number of tropical storms, and those which became hurricanes, by month, for the period 1851 2004. It also shows the monthly total and average number of hurricanes to strike the U. S. since 1851 (updated from Jarrell et.al. (2001).

(7) What was the largest number of hurricanes in the Atlantic Ocean at the same time?<>br Four hurricanes occurred simultaneously on two occasions. The first occasion was August 22, 1893, and one of these eventually killed 1,000-2,000 people in Georgia South Carolina. The second occurrence was September 25, 1998, when Georges, Ivan, Jeanne and Karl persisted into September 27, 1998 as hurricanes. Georges ended up taking the lives of thousands in Haiti. In 1971 from September 10 to 12, there were five tropical cyclones at the same time; however, while most of these ultimately achieved hurricane intensity, there were never more than two hurricanes at any one time.

(8) How many direct hits by hurricanes of various categories have affected each state?
Table 10, updated from Jarrell et al. (2001), shows the number of hurricanes affecting the United States and individual states, i.e., direct hits. The table shows that, on the average, close to seven hurricanes every four years (~1.75 per year) strike the United States, while about three major hurricanes cross the U.S. coast every five years (0.60 per year). Other noteworthy facts, updated from Jarrell et al. (2001), are: 1.) Forty percent of all U.S. hurricanes hit Florida; 2.) Eighty-three percent of category 4 or higher hurricanes strikes have hit either Florida or Texas; 3.) Pennsylvania’s only hurricane strike between 1851-2004 was 1878.

(9) When are the major hurricanes likely to strike given areas?
Table 11 shows the incidence of major hurricanes by months for the U.S. mainland and individual states. September has as many major hurricane landfalls as October and August combined. Texas and Louisiana are the prime targets for pre-August major hurricanes. The threat of major hurricanes increases from west to east during August with major hurricanes favoring the U.S. East Coast by late September. Most major October hurricanes occur in southern Florida.

(10) How long has it been since a hurricane or a major hurricane hit a given community?
A chronological list of all hurricanes to strike the United States 1900 through 1990 including month, states affected by category of hurricane, and minimum sea level pressure at landfall can be found in Jarrell et al. (1992). Appendix A extends that publication to cover the entire database from 1851-2004. Table 12 summarizes the occurrence of the last hurricane and major hurricane to directly hit the most populated coastal communities from Brownsville, Texas to Eastport, Maine. In addition, if a hurricane indirectly affected a community after the last direct hit, it is listed in the last column of the table. In order to obtain the same type of information listed in Table 12 for the remaining coastal communities, the reader is again referred to Jarrell et al. (1992) or NOAA Coastal Services (http://hurricane.csc.noaa.gov/hurricanes/index.htm). There are many illustrative examples of the uncertainty of when a hurricane might strike a given locality. After nearly 70 years without a direct hit, Pensacola, Florida was hit directly by Hurricane Erin in 1995 and major Hurricane Ivan in 2004 within 10 years. Miami, which expects a major hurricane every nine years, on average, has been struck only once since 1950 (in 1992). Tampa has not experienced a major hurricane for 84 years. Many locations along the Gulf and Atlantic coasts have not experienced a major hurricane during the period 1851-2004 (see Table 12).

(11) What is the total United States damage (before and after adjustment for inflation) and death toll for each year since 1900?
Table 13a summarizes this information. Table 13b ranks the top 30 years by deaths, by unadjusted damage and by adjusted damage. In most years the death and damage totals are the result of a single, major hurricane. Gentry (1966) gives damages adjusted to 1957 59 costs as a base for the period 1915 1965. For the most part, death and damage totals for the period 1915 1965 were taken from Gentry's paper, and for the remaining years from Monthly Weather Review. Adjusted damages were converted to 2004 dollars by the factors used in Table 3a.

(12) What are the deadliest and costliest hurricanes to affect Hawaii, Puerto Rico and the U.S. Virgin Islands since 1900?
Table 14, provided by Hans Rosendal and Raphael Mojica of the Weather Service Forecast Offices in Honolulu and San Juan, respectively, summarizes this information. Iniki in 1992 is the deadliest and costliest hurricane to affect Hawaii while Georges of 1998 is the costliest hurricane to affect Puerto Rico. The notorious San Felipe hurricane of 1928 was the deadliest hurricane in Puerto Rico since 1900.

(13) Are there hurricane cycles?
Figures 1 through 16 show the landfalling portion of the tracks of major hurricanes that have struck the United States between 1851-2004. The reader might note the tendency for the major hurricane landfalls to cluster in certain areas during certain decades. Another interesting point is the tendency for this clustering to occur in the latter half of individual decades in one area and in the first half of individual decades in another area. During the very active period of the thirties this clustering is not apparent.

A comparison of twenty year periods beginning in 1851 indicates that the major hurricanes tended to be in Gulf Coast states before 1891, then favored Florida and the W. Gulf until 1911, shifting to the eastern Gulf Coast states and Florida during the next twenty years, then to Florida and the Atlantic Coast states during the 1940s-1950s, and back to the western Gulf Coast states in the following twenty-year period.

CONCLUSIONS

In virtually every coastal city from Texas to Maine, the present Tropical Prediction Center Director (Max Mayfield) former National Hurricane Center Directors have stated that the United States is building toward its next hurricane disaster. The population growth and low hurricane experience levels indicated in Hebert et al. (1984), together with updated statistics presented by Jarrell et al. (1992) form the basis for their statements. The areas along the United States Gulf and Atlantic coasts where most of this country’s hurricane related fatalities have occurred are also now experiencing the country’s most significant growth in population. This situation, in combination with continued building along the coast, will lead to serious problems for many areas in hurricanes. Because it is likely that people will always be attracted to live along the shoreline, a solution to the problem lies in education and preparedness as well as long-term policy and planning.

The message to coastal residents is this: Become familiar with what hurricanes can do, and when a hurricane threatens your area, increase your chances of survival by moving away from the water until the hurricane has passed! Unless this message is clearly understood by coastal residents through a thorough and continuing preparedness effort, disastrous loss of life is inevitable in the future.

Acknowledgments: Richard Pasch and Max Mayfield made helpful suggestions. Paul Hebert, J.G. Taylor and R.A. Case, co authors of previous versions of this paper, are recognized for their enduring contributions to this work. Lenworth Woolcock drafted the 19th century figures and Joan David drafted the 20th and 21st century figures.

REFERENCES

Notes:

States Affected and Category by States Affected: The impact of the hurricane on individual U.S. states based upon the Saffir-Simpson Scale (through the estimate of the maximum sustained surface winds at each state). TX S-South Texas, TX C-Central Texas, TX N-North Texas, LA-Louisiana, MS-Mississippi, AL-Alabama, FL NW-Northwest Florida, FL SW-Southwest Florida, FL SE-Southeast Florida, FL NE-Northeast Florida, GA-Georgia, SC-South Carolina, NC-North Carolina, VA-Virginia, MD-Maryland, DE-Delaware, NJ-New Jersey, NY-New York, PA-Pennsylvania, CT-Connecticut, RI-Rhode Island, MA-Massachusetts, NH-New Hampshire, ME-Maine. In Texas, south refers to the area from the Mexican border to Corpus Christi; central spans from north of Corpus Christi to Matagorda Bay and north refers to the region from north of Matagorda Bay to the Louisiana border. In Florida, the north-south dividing line is from Cape Canaveral [28.45N] to Tarpon Springs [28.17N]. The dividing line between west-east Florida goes from 82.69W at the north Florida border with Georgia, to Lake Okeechobee and due south along longitude 80.85W.)

Highest U.S. Saffir-Simpson Category: The highest Saffir-Simpson Hurricane Scale impact in the United States based upon estimated maximum sustained surface winds produced at the coast.

Central Pressure: The observed (or analyzed from peripheral pressure measurements) central pressure of the hurricane at landfall.

Maximum Winds: Estimated maximum sustained (1-min) surface (10 m) winds to occur along the U. S. coast. Winds are estimated to the nearest 10 kt for the period of 1851 to 1885 and to the nearest 5 kt for the period of 1886 to date. (1 kt = 1.15 mph.)

* - Indicates that the hurricane center did not make a U.S. landfall (or substantially weakened before making landfall), but did produce the indicated hurricane force winds over land. In this case, central pressure is given for the hurricane's point of closest approach.

& - Indicates that the hurricane center did make a direct landfall, but that the strongest winds likely remained offshore. Thus the winds indicated here are lower than in HURDAT.

# - Indicates that the hurricane made landfall over Mexico, but also caused sustained hurricane force surface winds in Texas. The strongest winds at landfall impacted Mexico, while the weaker maximum sustained winds indicated here were conditions estimated to occur in Texas. Indicated central pressure given is that at Mexican landfall.

Additional Note: Because of the sparseness of towns and cities before 1900 in some coastal locations along the United States, the above list is not complete for all states. Before the Gulf of Mexico and Atlantic coasts became settled, hurricanes may have been underestimated in their intensity or missed completely for small-sized systems (i.e., 2004’s Hurricane Charley). The following list provides estimated dates when accurate tropical cyclone records began for specified regions of the United States based upon U.S Census reports and other historical analyses. Years in parenthesis indicate possible starting dates for reliable records before the 1850s that may be available with additional research:

Texas-south – 1880, Texas-central – 1851, Texas-north – 1860, Louisiana – 1880, Mississippi – 1851, Alabama < 1851 (1830), Florida-northwest – 1880, Florida-southwest – 1900, Florida-southeast – 1900, Florida-northeast – 1880, Georgia < 1851 (1800), South Carolina < 1851 (1760), North Carolina < 1851 (1760), Virginia < 1851 (1700), Maryland < 1851 (1760), Delaware < 1851 (1700), New Jersey < 1851 (1760), New York < 1851 (1700), Connecticut < 1851 (1660), Rhode Island < 1851 (1760), Massachusetts < 1851 (1660), New Hampshire < 1851 (1660),and Maine < 1851 (1790).