Skip Navigation Links
NOAA NOAA United States Department of Commerce

The Deadliest Atlantic Tropical Cyclones, 1492-1996


  1. Introduction
  2. Tropical Cyclone Terminology
  3. Casualty Information
  4. Storm Lists and Statistics
  5. Conclusions
  6. Acknowledgements and References
  7. Appendix 1: Cyclones with 25+ deaths
  8. Appendix 2: Cyclones that may have 25+ deaths
  9. Notes to the Appendices
  10. References to the Appendices

3. Casualty Information

Losses over open waters - An ocean of trouble

The period under study saw a large and widespread increase in Atlantic coastal population. Available records, however, suggest that the population on the Atlantic was the most vulnerable to storms through the 18th century. These shipborne explorers, emigrants, combatants, fishermen, traders, pirates, privateers, slaves, and tourists made up the crews and passengers on an uncounted, but enormous number of local and transatlantic sailings. Most of the ships travelled to or from the ports of Spain, France, Great Britain, and the Netherlands. They usually proved no match for the intense inner-core region of a severe tropical cyclone.

It is doubtful if any sailing ship or any man aboard survived in this sector of a really great hurricane. (Tannehill 1955)

In fact, to 1825,

more than five percent of the vessels in the (West) Indies navigation were lost due to shipwrecks; the biggest part due to bad weather... (Marx 1981).

The total number of ship-related casualties associated with Atlantic tropical cyclones is unknown, but there are clues. Some perspective on the magnitude of ship losses worldwide is gained by realizing that on the coast of England alone there have been a minimum of 250,000 wrecks (Cameron and Farndon 1984)! On the other side of the Atlantic, near New England, it was estimated that three out of every five sailors drowned during the period 1790 to 1850 (Snow 1943). Of course, many of these disasters were unrelated to the weather, while others are attributable to the brutal, cold storms of the North Atlantic winter rather than to tropical cyclones. Still, an account of one 17th century hurricane indicates the great magnitude of some losses blamed on tropical cyclones:

By these kind of Tempests the King of Spain hath lost at several times near 1000 sail of ships. (Ludlum 1963)

Similar disasters continued for another two centuries. Even as late as the 1830's,

...the annual loss of life, occasioned by the wreck or foundering of British vessels at sea, may, on the same grounds (i.e., 'the boisterous nature of the weather and the badness of the ships'), be fairly estimated at not less than One Thousand persons in each year... (Parliament Select Committee 1839).

Steamship voyages contributed increasingly to the number of lost ships during the latter half of the 19th century. In 1875-76, "heavy weather" was blamed for the loss of 176 steamships. Over a longer period, 1840 to 1893, 7,523 people perished in 125 North Atlantic steamship disasters of all types (Garrett 1986).

The large number of ship losses was partially a consequence of the great number of ships that inadvertently encountered storms. Redfield's (1846) analysis of an 1845 hurricane off the U. S. mid-Atlantic coast contains, on one weather map, information from the logs of more than 50 ships within about 450 miles of the storm's center. There were likely other vessels in that area. Redfield suggested that the then-expanding electric telegraph could be used in the Atlantic ports of the United States to alert mariners of approaching bad weather. Unfortunately, occasional ship disasters related to Atlantic tropical cyclones continued into the early 1900's. Further technological advances in meteorology, communication, navigation, and the seaworthiness of ships makes such losses infrequent today.

Reference materials about specific ship losses range from non-existent to overwhelming. In some instances, where the sea claimed a lone ship or even an entire fleet, record of the cause and location of the catastrophe went down with the ship(s). Moreover, for centuries there were virtually no official records on lost ships (Cameron and Farndon 1984). On the other hand, Marx (1983) wrote that:

if a team of one hundred researchers spent their whole lives searching through the more than 250,000 large legajos (bundles) in the Archive of the Indies (at Seville), I doubt that they could locate all the important documents concerning Spanish maritime history in the New World.

Either way, we learned little or nothing about many lost or missing crews and the circumstances behind their disappearance. For this compilation, lacking contrary evidence, the crews and passengers of ships lost over open waters in tropical cyclones were counted as fatalities.

Coastal deaths

While losses over open waters have decreased of late, rapid growth of coastal communities over the past 500 years has meant an ever-increasing population at risk to tropical cyclones. As at sea, relatively primitive communication methods increased the possibility of disaster near the shoreline. Not until 1909 was the first in situ ship report of hurricane conditions received in time to assist coastal preparations (Garriott, 1910).

There are two primary components to the danger near the shore, coastal ship losses and storm surge disasters. It is estimated that 98% of the ships lost in the Western Hemisphere to 1825 wrecked in waters no deeper than 30 feet (Marx 1983). Proper disposition of many of these cases is uncertain. Undoubtedly, many mariners lost their lives while staying with their vessel until it was too late to reach safety. This seems especially true early on, as noted in the following examples, with the first passage about non-tropical cyclones:

for four winters after my appointment to the charge of the barracks at the above named place (Yarmouth in Norfolk, England) in 1803, I witnessed the loss of vessels with all their crews within a few yards from the shore....I witnessed His Majesty's gun-brig Snipe, stranded within 50 yards of the beach at the back of the pier, having 67 persons on board, who all perished... (Parliament Select Committee 1839)

Came to anchor in St. Thomas's harbour, and landed the mails. Here the hurricane of the 2nd (August 1837) appeared to have concentrated all its power, force, and fury; for the harbour and town were a scene that baffles all description. Thirty-six ships and vessels totally wrecked all around the harbour, among which about a dozen had sunk or capsized at their anchors; some rode it out by cutting away their masts, and upwards of 100 seamen drowned... (Reid, 1841).

In contrast, today's early warning system usually results in little or no loss of life aboard vessels that wreck on a coast or in a marina. In 1992 Hurricane Andrew, for example, only two boating-related deaths occurred in southeast Florida despite boat damage estimated at $0.5 billion (Mayfield et al. 1993). For purposes of this work, cases with ships lost on the coast or in port were excluded from the casualty lists unless explicit documentation of sufficient loss of life was found.

Storm surge, occasionally reaching heights of 20 to 30 feet, has been responsible for some of the largest losses of life associated with tropical cyclone at the coastline. Storm surge is the rise of water caused by the wind and pressure forces of a hurricane. These forces induce currents in the water. While the hurricane is in deep water, these currents produce little storm surge because converging water and the subsequent piling up is compensated by currents at greater depths moving water away. However, as the hurricane moves onto the continental shelf and makes landfall, the compensating currents are eliminated by the slope of the shelf and the shoreline, and the converging water rises. This rising water may over-top barrier islands or be funneled into bays and estuaries. In many cases, maximum storm surge heights measured relative to mean sea level have been recorded at the head of bays or even inland away from the shoreline. Generally, storm surge gradually rises to a peak and returns to normal, all in 6 to 12 hours. However, in intense or rapidly-moving hurricanes, rapid rises and falls on the order of minutes to an hour have been reported. Riding on top of the storm surge are waves which cause major damage when they break against structures.

Poor communication for many years left coastal communities virtually without warning of storm surge. In the United States, storm surge is blamed for 90% of hurricane-related fatalities (AMS 1973). Even with the many technological advances, much of the burgeoning coastal population of the Americas remains vulnerable to storm surge (Sheets 1990).

Inland deaths

Inland communities are also susceptible to tropical cyclone catastrophes. There, fresh-water flooding from excessive rainfall can lead to large numbers of deaths by drowning.

The number of inland deaths, indeed those near the coast and offshore as well, were only estimated by many of the references. Numerous entries in Appendix 1 appear rounded to the nearest ten, hundred, or even thousand. In addition, the data from many references suggest that the listed total is likely a lower threshold. For example, Millas (1968) indicates that there were 60 deaths in Dominica during a 1788 hurricane. He also presents a contemporary remark about Martinique from The Gentleman's Magazine:

...the number of persons who have lost their lives is so great, that we dare not mention what report estimates it at, for fear of exaggeration.

Furthermore, there is evidence that casualty statistics were intentionally withheld by government officials on occasion (Perez). Hence, in some cases the actual number of deaths could be many multiples of the total shown in Appendix 1.

We also note that in the past several years the National Hurricane Center has distinguished explicitly between deaths directly related to the forces of tropical cyclones (e.g., drowning due to storm surge) and those attributable only indirectly to the weather (e.g., due to a traffic accident on a rain-slickened road). For those systems, this study used only the direct death toll.

Next: Storm Lists and Statistics