The Repository Hook Echo

Katrina 2005: The Record Surge No Gauge Recorded

October 12, 2026 · Runtime 15:06 · Watch on YouTube

On 29 August 2005 the sea rose about 25 feet over Waveland and Bay St. Louis, Mississippi. The NOAA tide gauge at Waveland sent its last reading at 4:12 in the morning: eight feet. Its neighbours at Biloxi and Ocean Springs went silent too, all before landfall. The highest storm surge ever recorded on the US coast was read off walls, weeks later. Military footage from Waveland and Biloxi shows what it left; we read the gauges and the 400 high-water marks ourselves.

Transcript

Waveland, ten days later 0:00

This is the coast at Waveland and Bay St. Louis, Mississippi, ten days after Hurricane Katrina. These are the pilings of what used to be piers. Behind the beach, whole streets are down to their foundations. On the morning of August 29th, 2005, the sea rose over this town to about twenty-five feet.

The gauge that said eight feet 0:20

Nobody saw that number on an instrument. The tide gauge at Waveland sent its last reading at twelve minutes past four in the morning. It said eight feet. Then it went silent, hours before the worst of the water arrived. The highest storm surge ever recorded on the coast of the United States was not recorded by any gauge. It was read off walls, weeks later.

So here is the question this episode answers. How do you measure a flood that switches off every instrument put there to measure it? And how much can you trust a number that someone read off a wall, two weeks after the water went down?

The coast and its gauges 0:57

First, the map. The Mississippi coast runs about a hundred kilometers in a straight line, from the Pearl River in the west to Pascagoula in the east. Waveland and Bay St. Louis sit on the west side of St. Louis Bay, Pass Christian on the east side, then Gulfport, Biloxi and Ocean Springs. Three of NOAA's permanent tide gauges were on this stretch. Remember where they are.

From the Bahamas to Category 5 1:22

Katrina began as a tropical depression over the Bahamas on August 23rd. Two days later it crossed southern Florida as a Category 1 hurricane. Then it reached the Gulf of Mexico, and the warm water did the rest. On August 28th it went from Category 3 to Category 5 in less than twelve hours, and peaked at one hundred and fifty knots, with a central pressure of nine hundred and two millibars.

Strength was only half of it. On August 27th Katrina nearly doubled in size. By the next day, hurricane-force winds reached about ninety nautical miles from the center, and tropical-storm winds about two hundred. The National Hurricane Center called it not only extremely intense, but exceptionally large.

The forecast: locally 28 feet 2:07

The forecast did not hold back. By ten in the morning on August 28th, the Hurricane Center's surge forecast for the coast was eighteen to twenty-two feet, locally as high as twenty-eight. Minutes later, the Weather Service office in New Orleans sent a message that most of the area would be uninhabitable for weeks, perhaps longer. Keep that twenty-eight in mind.

Landfall at Buras 2:30

Katrina weakened over the last night before landfall. At about ten past six in the morning on August 29th, the center came ashore near Buras, Louisiana, at the upper end of Category 3, with winds of a hundred and ten knots. This is the radar at Slidell, near New Orleans, watching the eye come north.

Category 3 sounds like a step down from Category 5, and for the wind it was. But the scale rates one thing only, the strongest sustained wind. It says nothing about the water. And the water had been building for a day, under a storm that was huge, and that only the day before had been a Category 5.

The radar archive stops 3:10

Watch the clock at the bottom of this loop. The Slidell radar's archive runs until the scan that began at eight fifty-four in the morning. After that, there is nothing in it until September 9th. Ten minutes after that last scan, the New Orleans office issued its final statement before its communications failed. The record does not say why the radar data stops. It only says that it does.

That final statement went out at four minutes past nine. It warned of extremely violent winds of a hundred and ten miles an hour, with gusts to a hundred and thirty-five, in the eyewall. From that moment, the Weather Service office in Mobile took over the warnings for southeast Louisiana.

Final landfall on the Mobile radar 3:51

The radar in Mobile, Alabama, kept recording. At about a quarter to ten, Katrina made its final landfall near the mouth of the Pearl River, on the border between Louisiana and Mississippi. Still a Category 3, a hundred and five knots, nine hundred and twenty-eight millibars. Three and a half hours earlier, at Buras, the pressure had been nine hundred and twenty, the third lowest of any landfall in the United States on record. Its radius of maximum winds was twenty-five to thirty nautical miles, and hurricane-force winds reached at least seventy-five nautical miles to the east, over Mississippi.

Even Mobile, about a hundred and fifty kilometers east of the landfall, was flooded from the sea. The surge in Mobile Bay reached eleven and a half feet, second only to the hurricane of July 1916, at eleven point six. That was on the weak edge of the storm.

Three gauges go silent 4:46

Now, the instruments that were supposed to measure the water. Along the Mississippi coast, NOAA ran permanent tide gauges that report every six minutes. Here is when each one sent its last reading. Biloxi, at two forty-eight in the morning. Waveland, at four twelve. Ocean Springs, at eight eighteen. All three stopped before landfall.

The Waveland record 5:11

This is the Waveland gauge, from our own pull of NOAA's archive. At midnight, local time, the water was about three and a half feet above mean sea level. Then it climbed fast, almost two feet in the last hour alone. The last point is eight point two feet. The archive's wind and pressure record for that station had already ended, the evening before.

Ocean Springs, fifty-five kilometers to the east, held on longest. Its last reading came at eight eighteen in the morning: twelve point four feet above mean sea level, and still rising. NOAA's own summary says it plainly. The sensor ceased transmissions and did not record the maximum water level.

The instruments that survived 5:54

The gauges that kept working were the ones farther from the center. Dauphin Island, Alabama, operated through the whole storm and peaked at five point eight feet above mean sea level. Pensacola, at six point one. On the Louisiana side, the anemometer at Grand Isle failed about two hours before the eye's closest approach. The station at Southwest Pass was knocked out by the surge itself.

Inland, the pattern repeated. The highest gust the Weather Service could document for Katrina was a hundred and thirty-five miles an hour, at the emergency operations center in Poplarville, Mississippi, about sixty kilometers inland. The assessment records it with four words attached: before the instrument failed.

A 55-foot sea 6:39

Offshore, a NOAA buoy in the open Gulf, south of the Alabama coast, was measuring waves. At seven in the evening on the 28th, the significant wave height was already thirty feet. By six the next morning it was fifty-five feet, equal to the largest significant wave height ever measured by a NOAA buoy. Another buoy, closer to the track, broke loose and went adrift.

Measuring it afterward 7:03

So when the water went down, the coast had a flood with no measurement. The measurement had to be made afterward, by hand. Under contract to FEMA, survey teams went out in pairs, six to nine teams at a time. In Mississippi, the field work did not begin until September 12th, two weeks after landfall.

The Hurricane Center's report describes the problem in two sentences. A precise measurement of the surge was complicated by many factors, including the widespread failures of tide gauges. And in many locations, most of the buildings along the coast were completely destroyed, leaving relatively few structures in which to find a mark of still water.

How a high-water mark is read 7:44

What they looked for was the line the water left behind. A mud line or a water line on a wall. A line of debris, or of dead grass and seaweed, pushed up against a fence or a bridge embankment. The preferred place was inside a building, where the water is calmer than outside. If there was no inside mark, an outside one. If there was no building, the ground.

The problem was obvious from the first day. In many places the buildings were simply gone, so there was no wall left to hold a mark. And coastal flooding comes in and goes out quickly, often before it leaves a line at all. Each mark that was found was surveyed to a quarter of a foot, against the national vertical datum.

A word on the scale, because it matters for what comes next. The tide gauges measure against their own local datum, and we converted every reading to mean sea level, using each station's published datums. The marks were surveyed in the national vertical datum of 1988. At the gauges on this coast that publish both, the two levels sit within a third of a foot of each other. Close enough to compare eight feet with twenty-four. Not close enough to argue over a tenth.

402 marks 8:57

In the three coastal counties of Mississippi, the survey flagged four hundred and two high-water marks. Three hundred and twelve were flagged by the contractor, URS, and ninety by the U.S. Geological Survey. Eighty-two in Hancock County, where Waveland and Bay St. Louis are. A hundred and ninety-one in Harrison County, from Pass Christian to Biloxi. A hundred and twenty-nine in Jackson County, out to Pascagoula.

The surge, coast profile 9:24

The marks are public today in the U.S. Geological Survey's database, and we pulled every one. Keeping only the marks classified as surge, without waves, here is the highest along each stretch of coast. About twenty-seven feet from Waveland to Pass Christian. Between twenty-five and twenty-six through Long Beach, Gulfport and Biloxi. Then it falls off to the east, to seventeen at Pascagoula.

The survey's own summary puts it in one line: the marks were noticeably highest in western Mississippi, near the border between Hancock and Harrison counties. That is where the east side of the eyewall, the side with the strongest onshore wind, came ashore. In Waveland and Bay St. Louis, the flooding ranged from twenty-four to twenty-eight feet.

27.8 feet, the record 10:11

The number that went into the record is twenty-seven point eight feet, at Pass Christian. The Hurricane Center's report calls it the maximum high-water mark observation of storm surge. Its guidance for mariners goes further: the highest storm surge ever recorded on the United States coast.

What the survey's table says 10:28

But open the survey's own table. The twenty-seven point eight foot mark is classified there as wave height, not surge only. It has waves on top of it. The highest mark in that table is thirty-four point nine feet, also a wave mark, less than a kilometer away. The highest mark classified as surge only is twenty-seven point two feet, about six kilometers to the east.

That does not break the record. Twenty-seven point two is still higher than any surge measured on this coast before. But it shows what kind of number this is. The survey itself warns that high-water marks rest on the teams' best judgment, and should be used to identify trends, not to extrapolate.

Gauge vs. wall 11:11

Now put the two kinds of measurement side by side. The Waveland gauge stopped at eight point two feet. The nearest surge mark, less than a kilometer away, is twenty-four point four. Biloxi stopped at three point three. Its nearest mark says twenty point nine. Ocean Springs stopped at twelve point four. Half a kilometer away, the mark is twenty-one point one.

None of the three recorded the peak. The best of them caught about sixty percent of the water. The worst, about a sixth. Not because they were wrong, but because they stopped. Everything between the last reading and the line on the wall happened with no instrument watching. The rise, the peak, and how long it lasted are not in any record.

Camille, 1969 11:59

Katrina's surge is always measured against one storm. In 1969, Hurricane Camille came ashore on almost the same stretch of coast, and the Weather Service gives its storm tide at Pass Christian as twenty-four point six feet. Camille was more intense at landfall than Katrina, nine hundred and nine millibars against nine hundred and twenty. It was also far more compact, and its high water covered a much narrower swath.

Why a Category 3 12:26

So why did a weaker Category 3 push more water than Camille? The Hurricane Center's answer is size. Katrina's wind field was huge, and a large storm moves a large volume of water. On top of that, during the day before landfall, while it was a Category 4 and 5, it had already sent large swells north, raising the water level along the coast with the waves themselves.

That volume did not stop at the beach. The surge reached at least six miles inland in many parts of coastal Mississippi, and up to twelve miles along bays and rivers. It crossed Interstate 10 in many places. The FEMA survey found the same thing, all along the coast: the flooding often reached more than five miles inland, from the Pearl River to the Alabama line.

New Orleans, in one minute 13:13

West of the state line, the surge reached fifteen to nineteen feet east of New Orleans. Levees and floodwalls around the city were overtopped or breached, and about eighty percent of New Orleans flooded, in places to about twenty feet. That is its own story, and it is a story about structures, not instruments.

The Hurricane Center's final count is one thousand three hundred and ninety-two deaths, direct and indirect. Five hundred and twenty were direct deaths, three hundred and forty-one in Louisiana and one hundred and seventy-two in Mississippi. The damage is estimated at one hundred and twenty-five billion dollars.

What the record supports 13:51

Here is what the record supports. The water on the Mississippi coast reached about twenty-four to twenty-eight feet across a twenty-mile swath centered on Bay St. Louis. We know that from four hundred marks read by hand, not from a single instrument. And the day before, the forecast had said locally as high as twenty-eight feet. The forecast was right. No gauge was left to confirm it.

Thirteen years later, in episode four of this channel, a sensor bolted to a pier in Mexico Beach, Florida, went under the water of Hurricane Michael. It was found in the debris, and it had recorded the whole surge. That is the measurement Katrina never got.

Next: Mayfield, 2021 14:32

Next, back to tornadoes, and to the radar that followed one storm across half a state. On the night of December 10th, 2021, a tornado crossed into western Kentucky from Tennessee shortly before nine, and kept going, through Mayfield, for almost three hours. The Weather Service in Paducah measured its path at a hundred and sixty-five point seven miles, among the longest in United States history. That is the next episode. Mayfield, 2021: the tornado that stayed down for a hundred and sixty-five miles.

Description and sources

Our pull of NOAA's archive: Waveland stopped at 8.2 ft above mean sea level (04:12 CDT), Biloxi at 3.3 ft (02:48), Ocean Springs at 12.4 ft (08:18), still rising. The Slidell radar's archive ends with the scan of 08:54 CDT. The FEMA survey began in Mississippi on 12 September and flagged 402 high-water marks; within about a kilometre of each silent gauge, the surge-only marks read 24.4, 20.9 and 21.1 ft. The record, 27.8 ft at Pass Christian, is a mark the survey classified as wave height; the highest surge-only mark is 27.2 ft -- still above Camille's 24.6 ft. The day before, the NHC had forecast 18 to 22 ft, locally 28.

PRIMARY SOURCES

Knabb, R. D., Rhome, J. R. & Brown, D. P. (2005; updated 2023). National Hurricane Center Tropical Cyclone Report: Hurricane Katrina (AL122005). https://www.nhc.noaa.gov/data/tcr/AL122005_Katrina.pdf
NOAA CO-OPS (2005). Hurricane Katrina Preliminary Storm Tide Summary; and 6-minute water levels, stations 8747766, 8744117, 8743281, 8735180, 8729840: https://tidesandcurrents.noaa.gov/
FEMA/URS (2006). High Water Mark Collection for Hurricane Katrina in Mississippi, FEMA-1604-DR-MS.
USGS Short-Term Network (STN), event 361 (Katrina), high-water marks: https://stn.wim.usgs.gov/STNServices/Events/361/HWMs.json
NOAA NWS (2006). Service Assessment: Hurricane Katrina, August 23-31, 2005. https://www.weather.gov/media/publications/assessments/Katrina.pdf
NHC (2018). Tropical Cyclone Concerns for Marine Interests. https://www.nhc.noaa.gov/marine/outreach/Marine_Interests.pdf
NWS Jackson, Hurricane Camille (1969): https://www.weather.gov/jan/1969_08_17_hurricane_camille
NOAA NWS NEXRAD Level II, KLIX and KMOB, 28-29 August 2005: https://unidata-nexrad-level2.s3.amazonaws.com/?prefix=2005/08/29/

Hook Echo reads a disaster through the instruments that recorded it: radar, sensor, survey and report, with the primary sources above.

About the material. Video: U.S. Army (Ohio Army National Guard, Joint Task Force Buckeye), U.S. Air Force (1st Combat Camera Squadron), Joint Task Force Katrina, U.S. Navy (Fleet Combat Camera Pacific, Joint Combat Camera) and U.S. Coast Guard footage from 2005, released through DVIDS, public domain. Every radar image is rendered by Hook Echo from NOAA's public archive of KLIX and KMOB Level II data (NEXRAD, unidata-nexrad-level2 on AWS), reflectivity only, in the National Weather Service colour scale. The tide-gauge series (NOAA CO-OPS) and the high-water-mark profile (USGS Short-Term Network, event 361) are our own measurements from public data. Every diagram is drawn in code from the sources listed. Nothing in this video is AI-generated imagery.

#HookEcho #HurricaneKatrina #StormSurge