Texas A&M Expert: Rainfall From Harvey Shattered Every Record

By Keith Randall, Texas A&M Marketing and Communications

Chris Ginter wades through deep floodwaters on September 6, 2017 in Houston, Texas. Ginter, a Houston resident, has been taking local residents to their flooded homes in his monster truck which can drive through waters up to 4 feet deep. (Photo by Justin Sullivan/Getty Images)

Hurricane Harvey was billed as a once-in-every-500-year event, and it more than lived up to its billing. It produced rainfall amounts that will re-write the weather books in Texas and the United States, says a Texas A&M University expert.

John Nielsen-Gammon, who is a Regents Professor of Atmospheric Sciences at Texas A&M and also serves as Texas State Climatologist, says Harvey set new standards for historic rainfall and flooding.

“Harvey is head and shoulders above all previous multi-day storms ever recorded in the continental United States,” says Nielsen-Gammon.

“I examined 18 different combinations of storm lengths and area sizes, from two days long to five days long, and standard areas from 1,000 square miles to 50,000 square miles. According to the preliminary data, Harvey was the worst in all but one.”

Nielsen-Gammon said that the most amazing record is for the five-day total over an area of 10,000 square miles.

“For Harvey to average 34.72 inches over five days across that large an area is ridiculous,” Nielsen-Gammon says. “The previous all-time United States record, set in Texas back in 1899, was estimated at 21.39 inches. Harvey exceeded that record by 62 percent.”

Ten other records were shattered by 20 percent or more, he confirms.

“To examine a storm’s flooding potential, scientists and engineers look at total rainfall over a large area over an extended period of time,” Nielsen-Gammon notes. “Extreme rainfall in one location causes problems, but when that rain falls over a much broader area, you get a disaster.”

To analyze Harvey, Nielsen-Gammon and Texas A&M colleague Brent McRoberts used the daily analyses produced by the West Gulf River Forecast Center in Fort Worth. Those detailed analyses are based on a combination of radar and rain gauge data, and are designed to help the National Weather Service with flood monitoring and prediction.

For historical context, Nielsen-Gammon compared Harvey’s totals to those of major United States storms of the past, as analyzed by Applied Weather Associates, a consulting firm based in Colorado that works extensively on extreme rainfall. He supplemented those analyses with reports prepared by the Army Corps of Engineers.

Nielsen-Gammon included all of the largest storms anywhere in the lower 48 states, as well as smaller storms east of the continental divide. He included storms as recent as last year and as distant as the 19th century.

Attached is his spreadsheet with the numbers that Dr. Nielsen-Gammon has provided via direct email – Anthony

Harvey Rainfall Data Top5co.xlsx (Excel)


34 thoughts on “Texas A&M Expert: Rainfall From Harvey Shattered Every Record

  1. Gee, does that include all the rainfall over the ocean and NOT just on land? If 10,000 square miles why not include all the hurricanes as they pass over the gulf? If this is for only CONUS then shouldn’t it also include the 20 mile costal zone of economic interest?

  2. How would Harvey’s total rainfall compare to Hurricane Flora’s in 1963? Is there enough data to make an accurate comparison?

  3. By itself, Harvey was not all that unique or unusual. What made it so were the circumstances which had it stalled over one area, and still able to tap into energy from the Gulf.

  4. So the real scientific question is, “Is the total quantity of precipitation from Harvey “head and shoulders” above similar Cat 3-4 hurricanes?” Confining it to a “precip per unit area” metric

    That is because Harvey basically stalled for 3-4 days over a coastal region rather than continuing to move and disperse its precip over a wider pathlength.

    P.S. If you didn’t read it, Joe Bastardi quite adequately destroyed Mann’s dishonest attempt at hurricane meteorology in claiming the one-off stalling behavior of Harvey is predicted by climate change models.

    • As a comparison, some hurricane stalled over Cuba for 4 days and dropped 100″ of rain in one spot. If that same scenario had happened with Alvin, that conceivably could have resulted in 168″ of rain. For those who don’t know, Alvin is part of the Houston metro area.

    • “Claudette produced torrential rains in both Texas and Louisiana when it made landfall. The highest one-day total was reported near Alvin, Texas where 42 inches (1,100 mm) of rain fell. This remains the twenty-four-hour rainfall record for any location in the United States. Two other towns also reported rainfall totals exceeding 30 inches. There was only one death from drowning and Louisiana received only minor damage from up to 15 inches (381 mm) of rainfall. Texas was hard hit by Claudette, with flooding reported in southeast Texas from up to 45 inches (1143 mm) of rainfall. Many residents had to be rescued from low-lying areas that were flooded.”


      This may or may not exceed Harvey’s statistics (34.72″ for 5 days over 10000 sq miles).

  5. I don’t suppose this fellow has also compared the historical storm tracks in his study, there is no mention, so the stalling out due to high pressure NW of it is not factored in. Or is that the 500-year event?

  6. 10,000 sq miles? That it actually a very tiny area, say 100 x 100 miles.

    What conclusions would he reach for say a million sq miles? I suspect very different.

    As we know, the only unusual thing about Harvey was that it stalled for five days. Many other similar storms have gradually moved around, dumping probably similar amounts over a much wider area.

    • Katia is certain to do something very similar to the Veracruz and the surrounding area as it is also almost stationary for the last 4 days and slowly growing.

  7. I wonder what the total rainfall was for the 1900 Galverston Hurricane which tracked from Galverston through Dodge, turned noorth east and all the way to Montreal in just 4 days.

  8. I wonder if this is a matter of perceived larger impact due to increased surveillance. I’ve seen this scenario play out in the public health sector where there is an appearance of increased disease occurrence concurrent with an increased surveillance capability. In other words, it’s always been there, but the incidents appear to be larger now because of the more expansive diagnostic capability.

  9. I had storm cell move here other day (mid maine) got 4.5 inches (measured in multiple locations with 2k feet between them) in 2 hours, almost 3 in first hour. good thing didn’t stall out for 24 hours.
    course I found a small roof leak (amazing how a pinhole in a vent pipe can allow so much water in) then and am awaiting for rain to end before fixing.
    damn suv driving bastards caused me 35$ work…..

  10. I know about the pause in hurricanes reaching land, about Harvey being 14th equal, about the precipitation being severe because Harvey stalled, about Joe vs Mann on why it stalled.

    The one claim I have trouble with is this. The BBC keeps banging on about warmer air having a greater capacity for moisture and that is why global warming contributed to record precipitation. They claim the waters of the Gulf are warmer too and that made things worse.

    I think linking AGW to more extreme weather is just propaganda but the points in my second paragraph are worth debate. What do others think? I would love to prove the BBC wrong, but ultimately, I want the right explanation, regardless of which way it goes. Our arguments against AGW are worthless if we also reject the truth.

    • It all hinges on whether Harvey (or Irma) are outlier events and if so, how far an outlier.

      Human operational memory is short. No one alive today remembers or can go on a newsbroadcast to retell the human side of the 1935 Cat 5 hurricane in the Florida Keys devastation. So the propagandasensationalist newcaster finds people who have lived 50-60 in an affected area and describe how “they’ve never seen anything like this before.”

      They make a true, honest humanistic statement. But the disinformation-propaganda occurs though the unspoken attempt to leave the false impression such an extreme event has never happened before.

      Such it is with 1930 US heatwaves, the Little Ice Age of the 18th Century, the Mideval Warm Period, Roman Warm Period. Etc.

      We all want to believe we live in exceptional times, all the time. Climate change propagandists play strongly to this human weakness.

      • In the Warner Brothers film “Key Largo,” Lionel Barrymore tells the story of the 1935 hurricane to gangster Edward G. Robinson like it was a ghost story. The film was directed by John Huston and starred Humphrey Bogart and Lauren Bacall.

  11. I am going to repost a comment I made on another thread, but this time without the snark and editorializing:

    What was the impact of global warming on Harvey rainfall? Let’s crunch some numbers. First of all, according this article in Wired magazine:

    Long term, the sea surface temperature of that region has risen about 1 degree over the past few decades—from roughly 86 to 87 degrees Fahrenheit, according to Michael Mann, a climatologist at Penn State


    Converting to metric, 86°F is 30°C and 87°F is 30.55°C. It takes one calorie to heat one gram of water one degree Celsius. Likewise the heat of vaporization of water is 540 calories per gram. Thus, starting at 30.55°C, it would take 69.45 calories to heat one gram of water (ignoring that it is salt water) to 100°C plus 540 calories to evaporate it for a total of 609.45 calories. 609.45 calories per gram equals 609,450,000 calories per cubic meter.

    According to this article in the Washington Post, 1 trillion gallons of water fell on Harris County alone in four days from Saturday (Sep. 2) through Tuesday. According to Wikipedia, the area of Harris County is 4,602 sq. km. One trillion gallons is 3,781,998,708 cubic meters. Dividing by 4,602,000,000 sq. meters, the height of the water column that fell on Harris County would be 0.8225579713 meters or about 32.3841720984 inches.

    If we take the 609,450,000 calories per cubic meter it takes to evaporate that water and multiply it by the 3,781,998,708 cubic meters that fell on Harris County, we obtain a total of 2.3070192117588E+018 calories which is the energy that Harvey used to evaporate the water that fell on Harris County. Assuming that all the water that Harvey evaporated also was condensed, then we can use this energy to calculate how much water Harvey would have dropped on Harris County if there had been no global warming (i.e. if the SST had been 86°F instead of 87°F).

    Starting at 30°C, it would take 70 calories to heat one gram of water (ignoring that it is salt water) to 100°C plus 540 calories to evaporate it for a total of 610 calories per gram or 610,000,000 calories per cubic meter. Dividing 2.3070192117588E+018 calories by 610,000,000 calories per cubic meter, we obtain a new estimate of the volume of water that Harvey would have evaporated sans global warming: 3,781,998,708 cubic meters. Again, we divide by the area of Harris County and we obtain a water column of 0.8218163207 meters or about 32.3549732559 inches, which is what would have fallen on Harris County sans global warming.

    Subtracting, the latter water column from the former, we obtain an estimate of the purported impact:

    0.0291988425 inches or about 29/1000 of an inch!!!!!!

    Of course, my estimate might be wrong. If somebody spots an error, please do not hesitate to point it out. It is surprising that it would make so little difference (honestly).

    Implicit (and simplifying) assumptions are:

    1. Harvey’s energy remains constant.
    2. Rainfall distribution remains constant.
    3. Harris County is representative of the entire rainfall extent.
    4. The difference in heat capacity of salt water vs. pure water is disregarded.
    5. Changes in the boiling point of water and latent heat of vaporization due to electrolyte levels and/or atmospheric pressure are disregarded.
    6. All the evaporated water falls as rainfall (no differences in rainfall due to varying atmospheric conditions).
    7. Storm path remains constant.

    Basically, this exercise seems to show that a 1°F difference in SST would make very little difference given the total energy of the storm implied from the amount of water that was evaporated from the Gulf of Mexico and dropped over Harris County.

    • Harvey was also pretty much stalled over the area for 4 days. If it had been moving like most of these storms do, it would have only rained on the area for 1/4 the time and only dropped 1/4 the ammount of water over the area. So they would have received 8.88 inches of water before the storm had moved through and only 07.25/1000 of an inch was from global warming effect.

    • ” …. starting at 30.55°C, it would take 69.45 calories to heat one gram of water (ignoring that it is salt water) to 100°C plus 540 calories to evaporate it for a total of 609.45 calories. 609.45 calories per gram equals 609,450,000 calories per cubic meter.”

      Water doesn’t have to be heated to 100°C to evaporate:

      “Heat (energy) is necessary for evaporation to occur. Energy is used to break the bonds that hold water molecules together, which is why water easily evaporates at the boiling point (212° F, 100° C) but evaporates much more slowly at the freezing point. Net evaporation occurs when the rate of evaporation exceeds the rate of condensation. A state of saturation exists when these two process rates are equal, at which point the relative humidity of the air is 100 percent.”


      • Enthalpy is a state function, not path related. It takes the same amount of energy, it’s just easier to calculate it the way Phil did.

  12. I’d like to reiterate here something that I brought up in another thread – the possibility of “recycled rain”. Very interesting, if true.

    From https://www.scientificamerican.com/article/hurricane-harvey-why-is-it-so-extreme :

    How can Harvey produce such extreme rainfall even though it is no longer over the ocean?

    The answer to this is fascinating. Normally a hurricane pulls moisture up from the ocean and releases it as rain all around the storm’s area, particularly in the northeastern quadrant. But Harvey has dropped so much water over such a large area of southeastern Texas that the storm is pulling that water back up into itself and dumping it again as more rain. The flood area is so far and wide that it is acting like part of an ocean, feeding warm moisture up into Harvey. “You only need about 50 percent of the land to be covered with water for that to happen,” Masters says. “Obviously we have more than that in Texas.”

    • I think it is called the “Brown Ocean Effect” which explains the addition of water to a cyclone when it is no longer strictly over the ocean.

  13. Are we talking total rainfall from the hurricane? Hazel in 1954 streaked inland and flooded Toronto, killing 81 people in that city as well as devastating others along the path. Had it been stalled in one place, what would have been rainfall?

  14. Eh? What’s that, you say? 500 years? Indigenous coastals used which sticks to measure rainfall year over year? And wrote it down where? Inquiring minds, and all …

    • When looking at a distribution with whatever statistic and using available data, you can figure out how likely some level of storm is per year. whatever measure of storm strength is on the Y axis, and time between events of that strength are on the X axis.You can figure out which distribution fits the data the best.

      As long as your assumed distribution does not drop down to zero within 500 years, you have that long tail to the right for projecting however long in time the selected distribution would report.

      Historically, geologically, knowing the planet changes across time, it would be reasonable to say that we cannot estimate 500-year flood, 1,000 year flood, etc., the way we maybe can say 100-year flood.

      Unless a youthful statistician or modeler is in charge. Then, common sense is not involved.

      I laughed out loud recently, hearing that Harvey was a 10,000 year event. It is preposterous to make that statement, even if your stats support it.

      • Back in 2007 and 2008 this area had back to back 500 year floods. One town got so tired of it they moved the whole town up the hill.
        I figure we’re good for a thousand years now – right?

  15. I’m guessing that the parameters were carefully chosen so that an unprecedented claim could be made.
    Clarksville had the record for most rain in a year in 1873. It seems to have stayed dry for those 5 days ( just north of Dallas). The greatest rainfall in year for the whole state was in 1941. 10 000 square miles is not even 5% of the state.
    And the greatest rainfall in one day was not broken.
    It was a big storm that was unlike any other but only because they are never exactly alike.

Comments are closed.