The Man With the Right Answer
On the morning of September 20, 1938, a junior meteorologist named Charles H. Pierce walked into the Weather Bureau's Washington office with a forecast his superiors did not want to hear. Pierce was 28 years old. He had been with the Bureau less than a year. And he believed a Category 5 hurricane that had formed near the Cape Verde Islands eleven days earlier was about to strike New York and southern New England directly.
The conventional wisdom said otherwise. For decades, Weather Bureau doctrine held that hurricanes do not hit New England. The historical record seemed to support it. The last major direct strike on the region had been the Great September Gale of 1815 — more than a century earlier. Any storm approaching the Eastern Seaboard, the thinking went, would be deflected northeastward by the Bermuda High pressure system, curve safely out to sea, and dissipate over cold Atlantic waters. It was not a theory built on atmospheric physics. It was a pattern based on memory and habit.
Pierce had studied the data more carefully. The Bermuda High, he found, had shifted unusually far to the north. That shift meant the storm would not be deflected. It would be funneled straight up the coast, squeezed between the high pressure to the east and a deep trough of low pressure to the west. The storm was not going to curve away. It was heading for Long Island.
A noon meeting was called. Pierce presented his analysis. Chief forecaster Charles Mitchell, his senior staff, and the accumulated weight of bureau doctrine were all in the room. Mitchell overruled him. The 2 p.m. advisory that day described conditions consistent with a dissipating offshore storm. No hurricane warning was issued for Long Island or New England. None would be.
A Storm Like No Other
What made the Long Island Express genuinely different from any previous Atlantic hurricane on record was not its intensity. It was its speed.
A typical hurricane moving north along the Eastern Seaboard travels at 20 to 30 miles per hour. By the morning of September 21, this storm was moving at nearly 60 miles per hour. Some estimates place peak forward speed closer to 70. The mechanism was the jet stream. The hurricane had been captured by a powerful upper-level trough of low pressure and was being driven northward as if on rails. The extraordinary speed had two consequences that made everything worse.
First, the storm had almost no time to weaken over the colder waters of the North Atlantic. Hurricanes draw energy from warm sea surface temperatures. Normally, as a storm moves north, colder water starves it of fuel and it loses strength. The Long Island Express moved too fast for that process to operate. It arrived at New England's coastline still carrying the full force of a major hurricane.
Second, the speed compressed the warning window to nearly nothing. Even if the Weather Bureau had issued a hurricane alert at noon on September 21, when the storm was still off the Carolina coast, it would have given coastal communities roughly two hours to act. At its actual forward speed, the storm covered 500 miles in less than ten hours. Communities that were not watching the sky simply did not know it was coming.
September 21, 1938
At 9 a.m. on September 21, the storm was reported off Cape Hatteras, North Carolina. By 1 p.m. it had reached Atlantic City, New Jersey, where part of the boardwalk was torn away. At 2:30 p.m., the eye made landfall near Bayport on Long Island's south shore. The barometric pressure at Bellport, New York registered 27.94 inches — among the lowest ever recorded on the Eastern Seaboard. Peak sustained winds were 120 miles per hour. Maximum gusts exceeded that considerably.
The storm hit within hours of high tide on a new moon. Astronomical high tide is at its greatest during a new moon. The timing was nearly the worst possible. Storm surge does not simply add to the existing tide level — it rides on top of it. Along Long Island's south shore, the ocean rose 14 to 16 feet above the mean tide line. In some locations, the combined wave and surge height reached 30 feet. Fire Island saw 153 of its 179 houses destroyed. Westhampton Beach lost all but 26 structures on the dune line.
The storm did not slow when it crossed Long Island Sound. It kept moving north at the same extraordinary speed, delivering a wall of water ahead of it into the shallow funnel of Narragansett Bay. Providence, Rhode Island sat at the head of that funnel. By 5 p.m., downtown Providence stood under 13 feet and 8 inches of water — surpassing the previous flood record set by the hurricane of 1815 by more than two feet. City Hall was submerged. The streets were navigable only by boat.
The Blue Hill Number
Blue Hill Observatory sits on a hill in Milton, Massachusetts, ten miles south of Boston. It is one of the oldest continuously operating weather stations in the United States, with records stretching back to 1885. At 6:11 p.m. on September 21, as the Long Island Express swept through Massachusetts, the observatory's anemometer registered a wind gust of 186 miles per hour.
That figure has never been exceeded at any surface weather station in New England. It remains the highest hurricane-related wind gust ever recorded in the continental United States east of the Rockies. The sustained wind speed at the same time was 121 miles per hour, maintained for a five-minute period between 6:11 and 6:16 p.m. Both numbers were authenticated and have stood for 87 years.
By 8 p.m. the storm had crossed Vermont and was approaching Lake Champlain. By 11 p.m. it had entered Canada. It had traversed the entire length of New England in under nine hours.
The Scale of the Damage
Final death toll estimates vary by source and method. The most commonly cited figure is approximately 700 killed, the majority in Long Island, Connecticut, and Rhode Island. The storm rendered an estimated 63,000 people homeless. It destroyed or severely damaged more than 57,000 homes and buildings. Roughly 3,300 boats were sunk or wrecked. Twenty thousand miles of power and telephone lines were downed on Long Island alone. Two billion trees were destroyed across New England.
New London, Connecticut burned for ten hours. Electrical short circuits caused by the flooding ignited fires along the waterfront that firefighters could not reach because the streets were under water. The city's central business district was gutted.
The storm also permanently changed the physical geography of Long Island. New inlets were punched through barrier islands from Fire Island to East Hampton by the surge, some of them remaining open for years. Napatree Point in Westerly, Rhode Island — a spit of land with a summer colony of 39 homes — was completely swept clean. Every structure was gone.
What People Believed Could Not Happen
The deadliness of the Long Island Express was not simply the product of a strong storm and bad timing. It was the product of a belief system.
New England in 1938 had not experienced a direct major hurricane strike in living memory. The 1815 storm was a family story at best, a forgotten history at worst. The last hurricane to seriously threaten the region before 1938 had passed offshore in 1903 without causing significant damage. This created a functional certainty among coastal residents that hurricanes were someone else's problem. They belonged to Florida, the Carolinas, the Gulf Coast. Not here.
That certainty was reinforced by the Weather Bureau's institutional doctrine. Mitchell and his colleagues were not acting in bad faith when they overruled Pierce. They were acting in accordance with a framework that had never been tested against a storm of this particular character. The doctrine about New England being immune to direct strikes was not written anywhere. It was assumed, and in the absence of contrary evidence, it had calcified into fact.
The result was an almost complete absence of preparation. Summer resort communities along Long Island's south shore were still occupied on September 21 — the season ran into late September. Residents saw the weather deteriorating and assumed it was a strong northeaster. They went to the beach to watch the waves. They stayed in the cottages facing the ocean. When the surge arrived, there was no high ground to reach and no warning to act on.
"The failure to adequately warn for the fast-moving storm spurred the Weather Bureau to adopt the air mass and polar front theories that had revolutionized European meteorological operations."
— NOAA Atlantic Oceanographic and Meteorological Laboratory, 80th Anniversary Summary
The Forecast That Was Right
Charles Pierce's noon meeting analysis proved correct in every particular. The Bermuda High was displaced northward. The storm did not curve out to sea. It struck Long Island between 2:30 and 3:00 p.m. on September 21, within the window he had projected. His mechanism — the interaction of the displaced high pressure with the low-pressure trough to the west — was the accurate physical description of what drove the storm northward at unprecedented speed.
Charles Mitchell resigned in the storm's aftermath. Pierce was promoted. These two facts are sometimes presented as the institutional reckoning, a kind of rough justice. They are not. Mitchell's resignation and Pierce's advancement did not bring back the 700 dead or rebuild the 57,000 structures. They are markers of how thoroughly the bureau's framework had failed, and how suddenly that failure became undeniable.
What was actually at fault was not a single chief forecaster's stubbornness. It was the Weather Bureau's theoretical foundation. American meteorology in 1938 had not incorporated the Bergen School's air mass and polar front analysis, which had transformed European forecasting in the 1920s and 1930s. European forecasters understood the jet stream's role in steering surface weather systems. American forecasters were still largely working from surface pressure maps and pattern-matching from historical records. Pierce saw what the Bergen framework would have predicted. Mitchell's doctrine had no framework for seeing it at all.
The Institutional Aftermath
The 1938 hurricane forced a reckoning that reshaped American meteorology over the following decade. The Bergen School's air mass and polar front theory was formally integrated into Weather Bureau training and forecasting practice. The idea that the United States could maintain a modern meteorological service while ignoring two decades of European theoretical advances became indefensible after September 21.
The broader institutional lesson was harder to write down. Forecasting doctrines are not scientific laws. They are patterns extracted from incomplete historical records. The Long Island Express was not an impossible storm. It was a storm outside the documented range of a 120-year memory. The doctrine that New England was immune to direct hurricane strikes was derived from records going back to the mid-nineteenth century. It had never been tested against a storm moving at 60 miles per hour on a jet stream track during an anomalous displacement of the Bermuda High. The historical record was simply too short to have captured it.
This remains the core lesson the storm left behind. The absence of a precedent is not the same as an impossibility. When the observation base is narrow and the physical system is capable of rare extremes, the rarest outcomes will eventually arrive without historical warning. The forecaster who relies on what has happened before will always be surprised by what has not happened yet.
The storm also forced practical improvements in warning dissemination. By the time hurricanes began threatening the Northeast again in the 1940s and 1950s, the Weather Bureau had established dedicated storm warning networks and communication protocols for coastal populations that simply had not existed in 1938. The radio broadcasts that reached no one in time on September 21, because no one had thought to make them, were replaced by systems designed specifically for the speed at which information had to travel.
What Popular Accounts Get Wrong
The standard telling of the Long Island Express centers on Charles Pierce as a lone genius ignored by hidebound superiors. There is truth in that framing. Pierce was right and Mitchell was wrong. But the simpler story obscures the more important one.
Mitchell's failure was not personal stubbornness. It was an institutional failure of theory. Pierce reached the correct answer partly through good judgment, but also because he was thinking in terms that the bureau's dominant framework did not yet support. The forecast he made was not a radical departure from available knowledge. It was an application of physical reasoning that European meteorologists had been using routinely for fifteen years. The bureau's failure was in not having absorbed that reasoning into its standard practice.
The other common error in popular accounts is the death toll. Figures in news coverage and commemorative histories range from 564 to over 700. The variation reflects genuine uncertainty. Many deaths were not definitively attributed to the storm until months afterward. Drownings, injuries that became fatal weeks later, people whose bodies were carried out to sea and never recovered — the accounting was never precise, and it has never been closed. The most conservative careful estimates land near 682. The figure most widely used in official NOAA documentation is approximately 700.
The storm's forward speed is also routinely underreported. Popular accounts frequently cite 47 miles per hour, which is the average forward speed during landfall on Long Island. By the time the storm was crossing Massachusetts, it was moving considerably faster. The 60-mile-per-hour figure cited by NOAA's Atlantic Oceanographic and Meteorological Laboratory for peak forward speed is the more accurate characterization of why the storm was so difficult to track, warn against, and escape.
The Record That Has Stood for 87 Years
Blue Hill Observatory still operates. It sits on the same Milton hilltop, still logging wind speed, temperature, barometric pressure, and precipitation without interruption since 1885. The 186-mile-per-hour gust registered on September 21, 1938 remains in its record as an outlier — a single data point that has not been approached in 87 years of continuous measurement.
The storm itself has no official name. The 1938 hurricane was not named, because the system of naming Atlantic storms was not adopted until 1953. It is known by its nickname, the Long Island Express, which came from the newspapers covering its extraordinary speed. It is sometimes called the Great New England Hurricane or the Great Hurricane of 1938.
Charles Pierce went on to a distinguished career in the Weather Bureau and later the National Weather Service. He was not, in the years after 1938, presented as a hero or a prophet. He was a junior forecaster who had read a day's worth of data correctly. The more important figure in the long run was the bureau itself, which took the storm's lesson seriously enough to rebuild its theoretical foundation on the basis of what it had failed to see coming.
That rebuilding took years. It was uneven. It required confronting an uncomfortable truth: that a century of American meteorological observation had not been enough to anticipate what the atmosphere was physically capable of doing. The Long Island Express did not break the laws of nature. It broke the assumptions that had been allowed to substitute for them.