"Why, sometimes I've believed as many as six impossible things before breakfast."
— one of which is that a warm patch of ocean can organise itself into the most powerful
engine on the planet, and then come looking for your house.
The machine
A hurricane is a heat engine. It takes warm water in at the bottom, throws cold air out of the top,
and the difference pays for the wind.
It needs sea water near 80 °F / 26.5 °C down to about 150 feet. A shallow warm
layer is not enough — the storm churns up cold water from below and starves itself.
It needs weak wind shear. If winds aloft blow at a different speed or direction than at the
surface, the chimney gets tilted and the engine falls apart. Shear is why most storms never become anything.
It cannot form within about 5° of the equator, because there is no Coriolis force there to
start the spin. A hurricane has never crossed the equator.
In the northern hemisphere it turns counterclockwise. Southern hemisphere, clockwise.
The structure is simple and brutal: a calm eye, a ring of the worst wind and rain called the
eyewall, and spiral bands feeding in from outside. The eye is warm, dry, sinking air — the
exhaust port. The eyewall is where the engine actually does its work.
The right side is the bad side. The storm's forward motion adds to its rotational wind on one
flank and subtracts on the other. In the northern hemisphere that makes the right-front quadrant
the strongest: worst wind, highest surge, most tornadoes. A storm doing 17 mph is roughly 34 mph
nastier on its right than its left. Which side of the track you end up on matters as much as how
strong it is.
How much energy are we talking about
NOAA's hurricane research division puts the latent heat released by an average mature hurricane at
roughly 5 × 1019 joules a day — about 200 times the entire world's electrical
generating capacity. Only a fraction of that becomes wind; most of it goes into making clouds and rain.
The kinetic energy of the wind alone still runs to about half the world's electrical output.
You cannot bomb one, cool one, or seed one. The United States spent twenty years trying
(Project Stormfury, 1962–1983) and concluded the storms were doing what they would have
done anyway.
The scale, and what it leaves out
The Saffir–Simpson scale was built around 1971 by Herbert Saffir, a structural
engineer who had been surveying hurricane damage for the UN, and Robert Simpson, then director of
the National Hurricane Center. Saffir did the wind-versus-damage half; Simpson added surge and flooding.
In 2010 the surge and flooding parts were removed. Not because they stopped mattering — because
they did not track the wind category reliably, and people were reading a number that implied a water
threat the scale could not actually promise.
Cat
Sustained wind
What it does
TS
39–73 mph
Branches down, power flickers, driving gets dangerous
1
74–95 mph
Roof, shingle, vinyl and gutter damage. Days without power
2
96–110 mph
Major roof and siding damage, shallow trees snapped. Weeks without power
3
111–129 mph
Major hurricane. Roof decking and gables removed. Water and power out for weeks
4
130–156 mph
Most of the roof and some exterior walls gone. Area uninhabitable for weeks to months
5
157+ mph
A high percentage of framed homes destroyed. Months
Those are sustained winds — a one-minute average. Gusts run meaningfully higher,
and gusts are what break things. Wind force scales with the square of speed: a 150 mph gust pushes about
four times as hard as a 75 mph one, not twice.
What actually kills people
Not the category. A study of US tropical cyclone deaths from 1963 to 2012 found about
half were from storm surge, and in the decades since, freshwater flooding — rainfall,
inland, often far from the coast — has become the leading cause. Wind is a long way down the list.
Hide from wind, run from water. It is the oldest advice on this coast and it survives because the
arithmetic has not changed. Six inches of moving water takes you off your feet. Twelve inches floats most
cars. Two feet moves a truck. Over half of US hurricane fatalities in vehicles are people who drove into
water they thought they could read.
The records
Record
Holder
Number
Lowest pressure, Atlantic
Wilma, 2005
882 mb
Strongest US landfall, pressure
Labor Day Hurricane, 1935
892 mb, Florida Keys
Strongest winds, Atlantic
Allen, 1980
190 mph
Strongest US landfall, wind
Labor Day 1935 & Dorian 2019*
185 mph
Deadliest US disaster of any kind
Galveston, 1900
8,000+ dead
Costliest US hurricane
Katrina, 2005
~$125B at the time
Most rain from a US tropical cyclone
Harvey, 2017
60.58 in, Nederland TX
Longest-lived Atlantic hurricane
San Ciriaco, 1899
28 days
Busiest Atlantic season
2020
30 named storms
*Dorian struck Great Abaco in the Bahamas, not the US mainland. Older records are
reconstructions — there were no satellites before 1960 and no aircraft before 1943, so anything from the
1935 era carries real uncertainty, and several have been revised decades later in reanalysis.
This coast
The Florida Panhandle and the Alabama shore have taken a particular beating, and the local memory here
is long and specific.
Storm
When
What happened
Camille
Aug 1969
Cat 5 into Mississippi, 900 mb. One of only four Cat 5 US landfalls on record
Frederic
Sep 1979
Cat 3 at the Alabama line; rebuilt Mobile's building codes
Opal
Oct 1995
Cat 3 near Pensacola Beach. Intensified fast overnight, caught the coast with little time
Ivan
Sep 2004
Cat 3 at Gulf Shores. Took out the I-10 Escambia Bay bridge and much of Pensacola's waterfront
Dennis
Jul 2005
Cat 3 near Navarre Beach, ten months after Ivan, into the same wreckage
Michael
Oct 2018
Cat 5 at Mexico Beach, 919 mb. Upgraded to Cat 5 months later in post-season analysis
Sally
Sep 2020
Cat 2 at Gulf Shores, crawling at 3 mph. The damage was the rain, not the wind
October storms behave differently. The September peak is driven by waves coming off Africa crossing
the open Atlantic. By October that conveyor shuts down and the action moves to the
western Caribbean and the Gulf itself — closer, warmer, less warning, and often sharper turns
northward into this coast. A late-season Gulf storm is a short-notice storm.
The cone is not the storm
The single most misread graphic in American life. The cone of uncertainty shows where the
centre is likely to go — drawn so that historically the centre stays inside it about
two-thirds of the time. That is all it says.
It is not the size of the storm. A storm's damaging winds routinely extend well outside it.
It is not a risk map. Being outside the cone is not being safe, and the edge is not a wall.
Roughly one time in three, the centre goes somewhere the cone did not cover.
The cone has been getting narrower for decades, because the forecasts got better. Its width is
rebuilt each year from the previous five years of real track errors.
Track forecasting has improved enormously — today's three-day forecast beats the one-day forecast of
1990. Intensity forecasting has improved far less, which is why storms still surprise people by
how strong they get, rarely by where they go.
The names
Six lists rotate, so 2026's names return in 2032. The World Meteorological Organization keeps them.
A name is retired when the storm was so deadly or costly that reusing it would be insensitive.
Katrina, Camille, Andrew, Ivan, Michael, Harvey, Maria, Ian, Sally — all gone from the lists forever.
Over 90 Atlantic names have been retired.
Storms were named after saints' days in the Caribbean long before meteorologists got involved —
San Ciriaco, Santa Ana. Modern naming started in 1950; women's names only from 1953; men's names were
finally added in 1979 after sustained criticism.
When a season ran out of names, forecasters used the Greek alphabet — in 2005 and again in 2020.
It was abandoned after 2020, partly because Zeta, Eta and Theta in the same season turned out to
be genuinely confusing when lives depended on hearing the right one. There is now a supplemental list of
ordinary names instead.
The people who fly into them
Since 1943 — when a US Army pilot flew into a hurricane essentially on a bar bet and came back — people
have flown aircraft into the eyewall on purpose, because there is still no substitute for being there.
The 53rd Weather Reconnaissance Squadron, the Air Force Reserve's "Hurricane Hunters", fly
WC-130J aircraft out of Keesler AFB in Biloxi, Mississippi — about 120 miles west along this
same coast.
NOAA flies two WP-3D Orions named Kermit and Miss Piggy, and a high-altitude Gulfstream IV
called Gonzo that samples the atmosphere steering the storm.
They release dropsondes — instrument tubes on little parachutes that radio back pressure,
temperature, humidity and wind the whole way down. The 959 mb you read on an advisory was very often
measured by a device falling through the eye a couple of hours ago.
Those flights measurably improve the forecast. The track error drops by a meaningful margin on storms
that get flown versus storms that do not.
Odds and ends
Rapid intensification is formally defined as a 35 kt (40 mph) increase within 24 hours.
It has become more common, and it is the single hardest thing to forecast.
The Atlantic season runs June 1 to November 30, and the statistical peak is almost exactly
September 10.
Same storm, different word: hurricane in the Atlantic and eastern Pacific, typhoon in the
western Pacific, cyclone in the Indian Ocean and near Australia.
The word comes from Huracán, a Taíno and Maya storm deity, by way of Spanish. One of the few
Caribbean indigenous words to make it into everyday English intact.
A hurricane's eye can be anywhere from 2 to over 200 miles wide. Small eyes often mean an
intense, tightly-wound storm — and small eyewalls can strengthen and weaken fast through
eyewall replacement cycles, where a new ring forms outside the old one and strangles it.
Barometric pressure is the better single number. Wind is estimated; pressure is measured, and
it is the one forecasters quietly trust most.
Hurricanes spawn tornadoes, usually in the outer bands on the right side, usually weak and
short-lived — but they arrive with no warning inside an already-howling night.
"If you don't know where you are going, any road will get you there."
— which is precisely the problem with a storm that has not decided yet.