Below average does not mean safe. It never has.
Each June, the Atlantic basin opens a season that carries within it both statistical probability and the memory of exceptions. In 2026, forecasters project a quieter-than-usual stretch through November, shaped by the moderating hand of El Niño — yet the science and the history agree on one enduring truth: a storm does not need company to be catastrophic. The ghost of Hurricane Andrew, born in a similarly calm year and buried in the rubble of sixty-four thousand Florida homes, reminds us that nature reserves the right to contradict its own averages.
- NOAA's below-average forecast for 2026 — six hurricanes, fourteen named storms — offers statistical comfort that experts are careful not to let harden into complacency.
- The 1992 hurricane season wore the same quiet forecast before Andrew obliterated Homestead, killing 65 people and erasing $27 billion in a single landfall, a wound the Gulf Coast still carries in its building codes.
- Flooding, not wind, now accounts for more than half of all hurricane deaths, yet public attention still gravitates toward wind speed categories — a dangerous mismatch between perception and reality.
- NOAA is deploying drones, AI-assisted models, and autonomous ocean vehicles this season, aiming to sharpen intensity forecasts by roughly 10% and expand the storm track cone's accuracy from 67% to 90%.
- New combined flood maps will embed hydrological data directly into hurricane warnings, a structural shift in how the agency communicates the true shape of the threat.
The 2026 Atlantic hurricane season arrived with numbers that seemed to invite relief. NOAA projected six hurricanes and fourteen named storms through November — well below historical averages — crediting the expected onset of El Niño, a Pacific weather pattern known to suppress Atlantic storm formation. But the meteorologists presenting those figures were careful to attach a caveat, and they reached back to 1992 to make it.
Hurricane Andrew formed late in a season that had looked almost dormant, then made landfall near Homestead, Florida, with a ferocity that redrew the map of American disaster response. Sixty-five people died. Roughly sixty-four thousand homes were destroyed. More than one point four million households lost power. The damage reached twenty-seven billion dollars — a record at the time — and Florida's decades-long complacency about major storms ended overnight. Building codes were rewritten. Emergency protocols were overhauled. The season's below-average label became a bitter footnote.
Florida International University meteorologist Haiyan Jiang invoked Andrew directly when discussing 2026, underscoring that storm frequency and storm danger are separate questions. The science adds another layer of concern: satellite records from the past thirty years show hurricanes are producing significantly more rainfall than they once did, and flooding now accounts for more than half of all hurricane-related deaths. Wind speed categories, the traditional shorthand for public risk, increasingly miss the point.
To close that gap, NOAA is fielding new tools this season — drones gathering close-range storm data, expanded AI integration in forecast models, and autonomous surface and underwater vehicles. The agency is also replacing its storm track uncertainty cone with a redesigned version intended to capture roughly ninety percent of possible center positions, up from sixty-seven. Perhaps most significantly, new combined flood maps will weave hydrological forecasting directly into hurricane warnings, reflecting a hard-won institutional acknowledgment that water, not wind, is where the danger most often lives.
The 2026 season will likely pass more quietly than most. The eastern Pacific, by contrast, is forecast to be unusually active. But the lesson Andrew left behind has not expired: forecasters can measure the odds, and they cannot predict which quiet season will produce the exception. That irreducible uncertainty is not a flaw in the science. It is the condition under which the science must be used.
The Atlantic hurricane season opened on Monday with a forecast that should have felt reassuring. The National Oceanic and Atmospheric Administration projected six hurricanes and fourteen named storms through the end of November—well below the historical average. Meteorologists attributed the quieter outlook to the near-certain arrival of El Niño, a weather pattern that suppresses hurricane formation across the Atlantic basin. But the experts offering these numbers came with a warning attached, one rooted in a specific memory: Hurricane Andrew, 1992.
Andrew arrived during a year that looked, on paper, exactly like 2026 was supposed to look—below average, calm, nothing to lose sleep over. The storm did not even form until August, arriving late to a season that had seemed almost drowsy. Then it made landfall near Homestead, Florida, and the word "below average" became meaningless. The hurricane killed sixty-five people. It destroyed roughly sixty-four thousand homes. It cut electricity to more than one point four million households across the state. The damage bill reached twenty-seven billion dollars, making it the costliest natural disaster the United States had recorded up to that point. Andrew shattered Florida's long complacency—the state had gone decades without a direct hit from a Category 5 hurricane—and forced a complete overhaul of building codes and emergency protocols.
Haiyan Jiang, a meteorologist at Florida International University, invoked Andrew's example when discussing the 2026 forecast. "We have some examples from previous years," she said, "like 1992, when Hurricane Andrew hit south Florida, specifically in Homestead, and caused enormous damage. That was a below-average year." Her point was simple but essential: fewer storms do not mean weaker storms. The science shows something more troubling. Satellite data spanning the last three decades reveals that hurricanes are dumping more rain than they used to. That intensification of precipitation is the primary driver of flooding, and flooding has become the deadliest component of the hurricane threat. More than half of all hurricane deaths in the past decade came from water, not wind.
The forecasting tools available in 2026 are vastly superior to what existed in 1992. Jiang acknowledged the improvement while noting an irreducible fact: uncertainty remains embedded in every prediction of where a hurricane will go. To narrow that margin of error, NOAA is deploying new systems this season. Drones will fly closer to developing storms, gathering data that should improve intensity forecasts by roughly ten percent. The agency is expanding its use of artificial intelligence in prediction models and testing autonomous surface and underwater vehicles. A redesigned uncertainty cone—the cone that shows where a storm's center might track—is being rolled out. The old version captured the actual center location about sixty-seven percent of the time. The new model is designed to account for roughly ninety percent of the possible positions.
Ken Grant, director of the National Weather Service, highlighted another innovation: combined flood maps that integrate hydrological forecasting directly into hurricane warnings. This matters because the data is unambiguous. Flooding kills. The shift toward predicting not just where a hurricane will go but how much water it will drop represents a recognition that the traditional focus on wind speed misses much of the actual threat.
Neil Jacobs, the NOAA administrator, presented these technological advances as incremental but meaningful progress. The improvements are real. But they exist within a constraint: forecasters cannot predict which below-average seasons will produce an Andrew. That uncertainty is not a failure of science. It is the nature of the problem. The 2026 Atlantic season will likely be quieter than average. The Pacific, by contrast, is expected to see more activity—between fifteen and twenty-two named storms in the eastern Pacific, with nine to fourteen hurricanes. But quiet does not mean safe. It never has.
Citações Notáveis
We have some examples from previous years, like 1992, when Hurricane Andrew hit south Florida and caused enormous damage. That was a below-average year.— Haiyan Jiang, meteorologist, Florida International University
The new uncertainty cone approach accounts for roughly 90% of possible positions where a hurricane's center could be located, compared to 67% with the previous model.— Ken Grant, director of the National Weather Service
A Conversa do Hearth Outra perspectiva sobre a história
Why does NOAA keep bringing up 1992? It was thirty-four years ago.
Because Andrew proved that the forecast itself can be a false comfort. It arrived in a year that looked calm on the spreadsheet. That's the lesson they're trying to keep alive.
But forecasting is better now, right? The drones, the AI—doesn't that change the equation?
It changes how accurately we can predict intensity and track. But it doesn't change the fact that a below-average season can still produce a catastrophe. Better tools help you see the storm coming. They don't prevent the storm.
The article mentions that more than half of hurricane deaths now come from flooding, not wind. That's a shift, isn't it?
It is. And it's partly because we've gotten better at building structures that withstand wind. But we haven't solved the water problem. Hurricanes are dumping more rain than they used to. That's what the satellite data shows over the last thirty years.
So the forecast of six hurricanes—is that actually good news?
It's good news if you interpret it as fewer storms. But the meteorologists are careful not to let people think that means less danger. Andrew is the exhibit A for why that thinking fails.
What's the most important new tool NOAA is deploying?
Probably the combined flood maps. If you can predict where the water will go, not just where the wind will hit, you can actually save lives. That's the shift happening this season.