The Storm No One Believed Was Coming

On the morning of September 21, 1938, a powerful hurricane was moving up the Atlantic seaboard. The Weather Bureau knew it existed. Forecasters in Miami had been tracking it for days. And yet when it made landfall on Long Island at 3:30 p.m. that afternoon — moving at 70 miles per hour, with winds exceeding 180 mph at its core — virtually no one in its path had received any warning. By the time the storm had crossed New England and dissipated over Canada, 682 people were dead, 1,700 were injured, and the landscape of coastal New England had been permanently altered.

The Great New England Hurricane of 1938 — called the Long Island Express for the speed with which it raced up the coast — remains the most powerful hurricane to strike New England in the 20th century. It is also one of the most consequential forecasting failures in American meteorological history: a case study in institutional bias, overconfidence, and the cost of dismissing evidence that contradicts established assumptions.

The Storm’s Origins and Track

The hurricane that would become the Long Island Express developed from a tropical wave off the African coast in early September 1938 and tracked westward across the Atlantic on a path that, by mid-September, had brought it into the Bahamas as a powerful Category 5 storm. It then turned northward along the Eastern Seaboard — a track that, if continued, would bring it directly toward the densely populated Northeast.

The Weather Bureau’s chief forecaster in Washington, Charles Mitchell, monitored the storm’s progress but concluded on the morning of September 21 that it would recurve out to sea before reaching the coast — a track that most Atlantic hurricanes follow as they encounter the westerlies and curve northeast away from land. This recurvature pattern was the historical norm, and Mitchell’s forecast reflected that historical norm.

What Mitchell discounted was the specific atmospheric setup that was in place on September 21: a blocking high-pressure system positioned over New England that effectively closed the door on recurvature and instead funneled the storm directly northward along the coast. A young meteorologist at the Jacksonville Weather Bureau office, Charles Pierce, had run calculations suggesting exactly this scenario and telephoned Washington to warn of a possible New England landfall. His warning was dismissed.

The storm continued north.

The Speed That Made It Lethal

What distinguished the 1938 hurricane from other powerful Atlantic storms was not just its intensity but its extraordinary forward speed. Most Atlantic hurricanes move northward at 10 to 20 miles per hour. The Long Island Express was accelerating into the westerly flow and moving at 60 to 70 miles per hour by the time it reached Long Island — faster than many passenger trains of the era.

This speed had two critical consequences. First, it dramatically compressed the warning time available. Even if warnings had been issued that morning, the storm’s rapid approach left only hours — perhaps two or three — between the time it could be identified as a New England threat and the time it made landfall. Second, the forward speed added directly to the destructive potential of the storm’s right-front quadrant winds: the storm’s own wind speed combined with its forward motion produced a wind field on the right side of the track that was among the most intense ever recorded at landfall in the Northeast.

The Long Island Express crossed Long Island and Connecticut in approximately two hours, reaching the Massachusetts coast by late afternoon and Vermont and New Hampshire by evening. The speed that prevented adequate warning also meant that the storm’s most destructive phase — the landfall and immediate inland penetration — was of short duration, but the concentrated violence of those hours was catastrophic.

What It Did to the Landscape

The storm’s impacts on Long Island and coastal New England were of a severity that had no precedent in living memory across a region where major hurricanes were vanishingly rare. Communities that had been established for three centuries with no serious hurricane experience were struck by a storm that reached Category 3 intensity at landfall with a 14 to 17-foot storm surge in some coastal areas.

On Long Island’s south shore, the storm surge overwhelmed beach communities within minutes of landfall. In Westhampton Beach, 29 of 42 oceanfront homes were destroyed and 29 people died. In New Haven, Providence, and other coastal cities, the storm surge moved inland for blocks, flooding streets and buildings that had never experienced flooding in their recorded histories.

The forests of New England — dense, old-growth timber that had never been managed for wind resistance — were devastated. An estimated 2 billion board feet of timber were blown down across New England, more than the annual timber harvest of the entire region. The mountain forests of Vermont and New Hampshire were stripped of significant portions of their canopy. The fallen timber created fuel conditions that produced some of the largest forest fires in New England history over the following two years as the downed wood dried.

The coastline itself was permanently altered. The storm broke through barrier beaches, created new inlets, and closed others, redistributing the coastal geography of Long Island and Rhode Island in ways that are still visible today. Napatree Point in Watch Hill, Rhode Island — a barrier spit with 39 summer cottages — was completely stripped of every structure; only the foundations remained.

The 39 Summer Cottages of Napatree Point

Among the stories that emerged from the 1938 hurricane, the fate of Napatree Point became one of the most documented. The narrow spit extending into Little Narragansett Bay held 39 summer cottages occupied by families enjoying the last weeks of the summer season. When the storm surge arrived, the spit was inundated so rapidly that escape was impossible.

Residents who survived described the water rising from ankle height to over their heads in minutes, then the cottages themselves beginning to move as the surge undermined their foundations. Those who survived generally did so by clinging to floating debris — pieces of their own homes — as the surge carried them inland across the bay. Eight people died on Napatree Point. The 39 cottages were reduced to nothing. When the water receded, the spit was bare sand.

Stories like Napatree Point’s illustrated the specific danger of low-lying coastal geography in the face of storm surge — the same danger that the hurricane myths piece addressed and that the Galveston Hurricane of 1900 had demonstrated on a far larger scale 38 years earlier. The lessons of Galveston had not penetrated the consciousness of New England coastal communities that had no experience of hurricane surge and no cultural framework for imagining it.

The Forecasting Failure and What It Changed

The failure to warn coastal New England of the approaching hurricane on September 21, 1938 was immediately recognized as a significant institutional failure. The dismissal of Charles Pierce’s warning, the reliance on historical recurvature patterns without adequate accounting for the specific atmospheric blocking that prevented recurvature, and the absence of any rapid communication network that could have pushed last-minute warnings to coastal communities all contributed to the death toll.

The 1938 hurricane directly accelerated the development of hurricane reconnaissance and warning infrastructure in the United States. The Weather Bureau’s capacity to track and forecast hurricanes was substantially expanded in the years following 1938. The first dedicated hurricane reconnaissance aircraft flights — dropping into storms to directly measure their intensity and position — began in the early 1940s, partly in response to the demonstration that ship-based and shore-based observation networks were inadequate to track a fast-moving storm along the heavily populated East Coast.

The concept of issuing hurricane watches and warnings for specific coastal areas — rather than general advisories that left the public without specific geographic guidance — was refined in the aftermath of 1938. The storm demonstrated that general knowledge of a hurricane’s existence was insufficient protection; what coastal communities needed was specific, timely guidance about whether their location was in the path of a storm.

New England’s Hurricane Vulnerability

The 1938 hurricane also forced a reckoning with a geographic reality that coastal New England had implicitly denied: the region is genuinely vulnerable to powerful hurricanes, and the rarity of such events in historical memory was a function of relatively short recorded history rather than of any geographic protection.

The climatological record, examined carefully, shows that powerful hurricanes have struck New England regularly over longer timescales — in 1635, 1675, 1815, and 1869, among other events. The 19th and 20th centuries before 1938 had been relatively quiet, creating a cultural assumption of safety that the Long Island Express demolished.

That assumption has been partially corrected but remains imperfect. New England’s extensive coastline, shallow bays that amplify storm surge, and dense coastal development create a vulnerability to storm surge from a major hurricane that emergency managers and meteorologists consider one of the more serious underappreciated hazards on the Eastern Seaboard. The 1938 storm is the benchmark — and since 1938, the region has been fortunate.

The Speed of the Warning and the Speed of the Storm

The Long Island Express left a specific legacy in hurricane forecasting: the recognition that storm forward speed is a critical variable in both impact severity and warning time, and that the historical tendency of Atlantic hurricanes to recurve offshore cannot be assumed for any specific storm. The blocking pattern that funneled the 1938 storm directly north occurs with some regularity, and the atmospheric conditions that produce it — identifiable in modern model output days in advance — must be actively checked for rather than assumed absent.

Charles Pierce, whose warning was dismissed on the morning of September 21, 1938, was right. The tools available to him were cruder than modern forecast systems by orders of magnitude, and his warning arrived too late to do much even if it had been heeded. The lesson drawn from his experience — that unconventional forecasts based on careful analysis of specific atmospheric conditions deserve serious consideration even when they contradict historical norms — is one that the atmospheric science community worked to institutionalize in the decades that followed.

The Long Island Express reached New England in September 1938. The warning systems and forecasting practices it helped build have been protecting that coastline, imperfectly but meaningfully, ever since.