Warm sea water is what powers hurricanes. Usually, sea surface temperature (SST) in the Gulf of Mexico needs to exceed 29°C to intensify a hurricane, and every fraction of a degree above 29°C increases the chance — dramatically — of not just intensifying, but super-charging it, creating a “monster storm.”
Which makes one wonder … if a storm passes by, what are the odds the sea surface temperature (SST) will exceed 29°C? Or more? Have the odds changed over time? Of course SST isn’t the only factor at play, only fools say so, but only bigger fools deny its impact on tropical storms.
To learn more about the history of SST in the area visited by hurricane Ida during her journey from Cuba to New Orleans, I selected the region from longitude 92°W to 85°W, latitude 23°N to 30°N, and retrieved daily data for SST in that region from 1981 through Aug. 2 of this year, from the OI (optimal interpolation) v2 dataset:
There’s an obvious annual cycle (hotter in summer, colder in winter), and if we remove it to define anomaly we get this:
The red line is a trend estimate from least-squares regression, and indicates that SST has increased since 1981, by about 0.85°C. It’s overwhelmingly “statistically significant.”
For sea surface temperature and its impact on hurricanes, that amount of warming is HUGE. It’s easy to dismiss it as “quite small” with an offhand “less than 1°C.” But if you really know hurricanes, you know that “every fraction of a degree counts” is, if anything, an understatement. An increase of a “mere” 1°C is nothing like “quite small.”
Of course that’s anomaly, not temperature itself, and it’s for the entire year, not the summer months when temperatures are highest. I split the time span into four decades, from 1981 to 1991, 1991 to 2001, 2001 to 2011, and 2011 to 2021. Then I estimated the probability density function (pdf) during the summer months (Jun/Jul/Aug/Sep) for each decade by two methods: a histogram, and with a smoothed estimate (kernel smooth). I got this for the four decades:
It’s obvious; the chance of meeting or exceeding that 29°C threshhold has increased, particularly from the 1981-1991 decade to the 2011-2021 decade.
If we look at the survival function (1 minus the cumulative distribution function), not only does it give the probability of meeting or exceeding any given value, it’s straightforward to estimate uncertainty ranges for those probabilities (shaded regions surrounding solid lines):
Between the 1981-1991 decade, and 2011-2021, the chance of meeting (or exceeding) the 29°C threshhold went from about 42%, to over 70%. The chance of meeting (or exceeding) 29.5°C, offering much more energy to the storm, went from about 18% to just over 50%. As for the 30°C limit — big trouble in hurricane town — the odds go from a mere 3% chance to a whopping 21%, a 7-fold increase.
Next, I’ll look at SST in the Gulf of Mexico over a longer time span than just the past 40 years.
I chose the title for this post based on a quote from TV weather personality Al Roker,
“We are looking at the result of climate change … that’s what created this monster storm.”
Evidently, the comments by Mr. Roker and others annoy Cliff Mass, who feels compelled to lecture us about looking at real data.
Happy to oblige.
This blog is made possible by readers like you; join others by donating at My Wee Dragon.
This was all clear to me from the growth of Katrina in 2005. The way K rebuilt and strengthened over the Gulf after moving across FL was astounding. Ida and Katrina style hurricanes are the new normal for hurricanes that move over the Gulf and make landfall in the US.
The Met Office Atlantic Hurricane forecast used to be called JASON – due to those being the months of the Atlantic hurricane season (with a double J at the beginning).
So I’d have thought it would be worth including October and November in this sort of analysis.
I’m interested to see your longer analysis, and possibly one covering larger areas. But on the face of it, that is some very scary stuff.
Thx. Very convincing. For the influence of sea surface temperature on hurricane strength, there is a tweet of Dr. Robert Rohde (https://twitter.com/RARohde/status/1432341587636412423).
Not exactly the same definition of sea surface temp, but insightful nevertheless. At about 27 °C in the core location, not so much the mean wind speed increases, but the width of the distribution, i.e. the wind speed of the strongest storms. It grows with ~ 35 mph/°C. The energy content of a m² of air, going with v², increases from ~ 3.6 kJ to ~ 6.6 kJ per °C.
A similar relation holds for the force, that the air exerts on anything in its way.
An error: the energy values are o.c. to be multiplied by 1/2.
As a supplement to this, the floods, tornadoes, river overtoppings, and deaths which followed the storm’s “perfect” collection of (collision with) other meteorological effects on its way to drowning New York City last night, etc. etc., are also breaking civilization’s back. [That is, assuming we are still civilized; I’m afraid since 1980 there’s been a lot of encouragement to cast of the trappings of a sharing society in the name of greed, selfishness, and otherblaming.]
https://yaleclimateconnections.org/2021/09/idas-flooding-rains-shift-to-pennsylvania-new-england-as-tornadoes-threaten-mid-atlantic/ – one wxwannabe commenter put it well
Love the use of the pdfs and survival functions. Thanks, as always.
According to this graph,
Ida passed over extremely hot water, 31C, on its way to New Orleans. In addition, the warm water was deep so that to total energy content was very high.
Looking at your graph, in the 1981-91 decase hot water like this barely existed. So not only is the heat content much increased but Ida passed over the hottest water.
You would expect that with much increased heat content in the water that some hurricanes would pass over that hot water. Those hurricanes then become the monsters.
Sorry, the graph didn’t link like I thought it would, sorry.
It’s worth mentioning that the “SST threshold” for a hurricane to form is climate dependent, and is related to the need for the temperature of a parcel near the surface to be sufficiently buoyant to rise to the tropical tropopause. Since upper tropical troposphere temperatures are rather uniform (due to the low Coriolis parameter that prevails in the deep tropics), but this temperature increases in the future, the distribution of SSTs favorable for convection also needs to shift.
Dr Cliff Mass would have got as far as your first graph (with the massive annual cycle) then applied his amazing ocular trend detection technique and concluded “no trend”. :-)
I have to wonder if Cliff Mass ever reads posts like this and realizes that there are a few more steps for him to use, or if he just ignores it all and goes on his merry uninformed way.
” (due to the low Coriolis parameter that prevails in the deep tropics)”
There is no Coriolis force precisely at the equator, which removes one of the strongest vorticity generators, allowing stable standing wave phenomena such as ENSO to develop. The fascinating Tropical Instability Waves are higher order versions of ENSO that travel westward and may spawn off cyclones and hurricanes.
I am repeatedly, repeatedly reminded of the conclusion from your post from May 7, 2011: I’ll continue to do what I can come hell or high water. Expect both. And here we are as I write this from smoke choked Butte, Montana where we’ve suffered through smoke filled skies on and off since mid-June!