The Great New England Hurricane of September 21, 1938 on its 75th anniversary

September 20, 2013

Revisiting the “Long Island Express”

long-island-1 Surface weather map, 9 a.m., September 21, 1938 (National Weather Service)

In September 1938, the country was weathering two very different types of storms: the first was the Great Depression, now in its ninth year, and the second was the gathering storm in Europe, soon to engulf the whole world.

Ominously, on September 20, 1938, one of the headline stories of The New York Times read:


As serious as those economic and political storms were, within 24 hours it would be another kind of storm that would totally dominate the attention of Long Islanders  and New Englanders.  That day’s Times alluded to it: tucked away on an inside page was a story that read, in part:


It was a statement, no doubt, prompted by dispatches from (the then) U.S. Weather Bureau, which would echo those same words for more than another 24 hours.  Even the next day, as the storm turned northward after sideswiping the Carolinas and New Jersey, there was still no hurricane warning.

Only at the last minute, about 2:00 p.m. on September 21,when Hurricane No. 3* (sequence controversial) of 1938 was just off the New Jersey Coast did the forecaster issue a final advisory to include “whole gale force winds” (up to 73 mph).  By that time, the storm was racing northward at forward speeds of up to 70 mph.

Track of 1938 “Long Island Express” hurricane (National Weather Service)

Still, there was no mention of a hurricane, as chief forecaster Charles Mitchell and his senior staff refused to believe that landfall would occur. (“In Boston meteorologist E.B. Rideout told his WEEI radio listeners to the skepticism of his peers that the hurricane would hit New England.”) Unfortunately, by then, any warnings would have been futile, as most forms of communication—except wireless, which many people didn’t have anyway–were already destroyed.

Communications were destroyed because between 2:30 and 3:30 p.m. (reported times vary), the storm’s full force slammed into the central Long Island, NY shoreline as a Cat 3 (previously a Cat 5) hurricane, with 111-129 mph winds (lower to the west, higher to the east).   At that point, the storm was “racing” northward at the amazing speed of 60-70 mph, which contributed to the exceptional winds to the east of the hurricane’s eye.***

Monstrous 40-foot-tall waves and a storm surge 14-18 feet above normal tide level inundated parts of Long Island, and later the southern New England coastline (around 4:00 p.m.), with even higher tides toward Cape Cod.  Contributing to the onslaught was the fact that landfall occurred near the highest astronomical tide of the year and there was nearly a new moon, sometimes reported as a full moon.

Waves Hitting Sea Wall at Battery Park, NYC
Waves Hitting Sea Wall at Battery Park, NYC

Admittedly, in 1938, neither weather satellites, doppler weather radar (in the modern sense), nor weather buoys existed, so for the most part, forecasters relied on reports from ships at sea for up-to-date information on ocean weather systems.   Although most forecasters at the U.S. Weather Bureau, as it was then known, were probably aware that hurricanes had previously made landfall** in New England, they believed such events were extremely rare, so rare that they were reluctant, in fact, to even predict one.

In the case of Hurricane No. 3 of 1938, that was a fatal error, which was even more remarkable for two reasons.  The first was that, by coincidence, as late as the very morning of the hurricane’s landfall, a New York Times editorial, entitled Hurricane, gave the Bureau high praise for its forecasting prowess:

“Every year an average of three such whirlwinds sweep the tropical North Atlantic between June and November. In 1933, there was an all-time record of twenty. If New York and the rest of the world have been so well informed about the cyclone, it is because of an admirable organized meteorological service” (The New York Times, 9/21/38).

The second reason that the Bureau’s “fatal error” was so remarkable was that just five years earlier, as described in my recent post, when the Chesapeake-Potomac Hurricane struck the mid-Atlantic on August 23, 1933, the forecast had been near perfect, preparing the public for widespread devastation.  Of course, based on previous experience, forecasters had much greater confidence in a Virginia landfall than one on Long Island.

The reluctance of the Weather Bureau to issue hurricane warnings was not unlike the mentality of the 1920s, when the 1925 Tri-State Tornado went on a 219-mile Midwest killing rampage, resulting in the loss off 700 lives from the Missouri Ozarks to Indiana. Since authorities didn’t want to alarm the public (even if warranted), the word “tornado” was not used in that, or any other, weather forecast, even if the forecaster had contemplated doing so.  In fact, at the time, authorities believed that just mentioning the word “tornado“ would cause such widespread panic that the word had been banned from U.S. weather forecasts since the late 19th century.  After the Tri-State Tornado, that kind of thinking began to change.

As discussed, the Weather Bureau (at least officially) hadn’t predicted a hurricane landfall.  But that decision was not without controversy, as a junior forecaster (Charlie Pierce) believed the storm would curve into Long Island and southern New England, due to blocking high pressure to the northeast and a trough of low pressure which would guide the storm inland.

Mr. Pierce, of course, was overruled.   Afterward, the Weather Bureau defended itself by insisting that a better forecast wouldn’t have mattered anyway because New Englanders “aren’t hurricane-minded”.  In any event, shortly thereafter, chief forecaster Charles Mitchell resigned and Charlie Pierce was promoted.

Damage on Long Island (Frank Markus;, via National Weather Service)

On its rampage through Long Island and New England, the hurricane caused hundreds of casualties and enormous property damage on a scale never before seen by the oldest inhabitants.  Near Southampton, on the south shore of Long Island, the hurricane created many inlets, or breaches, in the 100-mile-long outer barrier, allowing the sea access to the inland bays.  The largest of these was Chinnecock Inlet.  Except for Chinnecock, all the other inlets filled up with wrecked cars, broken trees, structural debris, and millions of tons of sand.  My recent post about the Chesapeake-Potomac Hurricane of 1933 describes how the same thing happened to create the Ocean City, MD inlet.

Storm damage, Pawtuxet Village, Rhode Island (

Over the years, many short film clips and documentaries#, including one shown on PBS, entitled The Hurricane of ‘38, were done about the storm.  One of the earliest and most famous of these, Shock Troops of Disaster, was created by the old Works Progress Administration (WPA), and is now considered a vintage relic of the hyperbole associated with government-made disaster films.  In one of the other documentaries, members of the Geoffrey Moore family, of Napatree Point, RI, tell a harrowing tale:

Catherine Moore: “At the height of the storm, my father was trying to hold back the wind and rain from coming through the front door, but finally gave up, when it became apparent that he was eventually trying to hold back the ocean.”

Catherine’s sister, Anne: “We had gone upstairs, when the house broke apart and the whole family found themselves in the water, with part of the structure acting as a sail.  We all thought we were being carried out to sea.”  Authorities later declared them dead, only to find out that they had been “blown clear across Narragansett Bay to Connecticut” and were alive and well.

Storm flooding/damage, Norwich, Ct (Norwich Times via
Storm flooding/damage, Norwich, Ct (Norwich Times via

In the modern era, “The Great New England Hurricane of 1938,“ a Cape Verde Islands-type storm (frequently the strongest), has become the benchmark against which all New England hurricanes are measured. As a Cat 3 storm at landfall, it wasn’t the most powerful or the deadliest hurricane to strike U.S. shores (although the peak gust was comparable to that of some Cat 5 storms), but it did strike one of the most populated regions of the country (just 75 miles east of New York City).

Including losses at sea, over 600 people perished, possibly making the storm the 7th deadliest to strike the U.S. mainland in the last 162 years, according to the National Hurricane Center. In addition, total damages exceeded $5 billion (adjusted for inflation). By comparison, in 2005 Hurricane Katrina resulted in about 1200 deaths and $108 billion in damages, while last year’s Superstorm Sandy caused 286 deaths and $68 billion in damages.

Other statistics from the storm:

  • Maxumum sustained winds – 121 mph
  • Peak wind gust – 186 mph at Blue Hill Observatory, MA.
  • Lowest pressure – 27.94 in (946.2 mb) at Bellport, NY
  • Peak storm surge – 17 ft. above normal high tide (RI)
  • Peak wave heights – 50 ft. at Gloucester, MA
  • Homeless – 63,000
  • Homes, buildings destroyed – 8,900
  • Vessels lost at sea – 3,300
  • New York, New England trees destroyed –over 2 billion

* Hurricanes were not given names until 1950.

** such as the “Great Colonial Hurricane of 1635,” which crossed southeastern New England and was regarded by some as New England’s most catastrophic weather disaster.

*** In the northern hemisphere,  a hurricane’s winds to the right of the storm’s eye are generally much higher because overall wind speeds are the combined total of the storm’s forward speed and the wind speed of the circulation itself.

# Such as this and this  and this and this and this and this and many others.

Continue reading
Show Comments
Most Read Local