The 2026 Atlantic Hurricane Season officially ends on November 30, but the outlook is quiet, so it very likely ended on October 22 this year. Overall, it was an exceptionally inactive season with a single hurricane which was also the single major hurricane.
This post marks the end of my 31st year writing these updates on tropical Atlantic activity. During that time, I have written approximately 1500 posts spanning 496 tropical cyclones including 229 hurricanes, 108 major hurricanes, 22 Category 5 hurricanes, and 57 retired storm names. As I announced at the end of the 2025 season, I am retiring from writing routine posts for this blog now.I began writing this blog in the summer of 1996 (prior to starting my junior year of college), and the internet, websites, and data availability looked very different then. I was honored to write for the New York Times' "Eye of the Storm" blog from 2007-2010 and then the Washington Post's "Capital Weather Gang" blog from 2012-2019.The 1996-2026 period spanned many historic storms and events, and I know that some of you reading this have been following along the entire time -- I sincerely appreciate your steadfast interest. And to the more recent followers and those who were not yet born when it began, I hope it has provided you with hype-free timely information and useful insights amidst the abundance of options out there now.
Now, on to the summary...
 |
| Track map of tropical cyclones during the 2026 hurricane season. Storm tracks are color-coded by intensity, and each storm's preliminary peak intensity, minimum central pressure, and Accumulated Cyclone Energy (ACE) is listed on the right. |
With a developing significant El Niño heading into the season, expectations were set for below-average activity, but even those expectations ended up being too bullish. The season had 10 named storms, 1 hurricane, and 1 major hurricane (Category 3+ on the Saffir-Simpson Scale). The climatological averages are 14, 7, and 2, so well below average by those counts.
The Accumulated Cyclone Energy, or ACE, is a metric that describes the overall intensity and duration of storms, not tied to a storm count. That ended up at just 16% of average, the lowest since 1983. The biggest contributor by far was Isaias, which made up 37% of the season's tally. Fay came in second place, and between Fay and Isaias, they made up 56% of the season's meager tally.
Of the ten named storms this season, most were brief, weak, and forgettable. The season started with Tropical Storm Arthur on June 17... that lasted less than a day. Then it was more than a month until the next storm, Bertha, which formed on July 19. Both of those storms lurked along the northern Gulf coast and made landfall.
Cristobal and Dolly were brief blips out in the central Atlantic, then Edouard was the third landfalling tropical cyclone on the U.S. this season on September 1 as a strong tropical storm.
 |
| Radar animation of Tropical Storm Edouard making landfall and moving inland on September 1-2. Available at https://bmcnoldy.earth.miami.edu/tropics/radar/. |
The sixth storm, Fay, was the longest-lived storm of the season... it was a named storm for five days, with a dissipated gap in the middle of its life. It was also VERY close to reaching hurricane intensity on the night of September 20-21. The NHC always conducts a post-season reanalysis on all storms to carefully scrutinize all available data when not under time constraints of the operational schedule. It would not be unrealistic if Fay is upgraded to a hurricane, which would of course, alter this season summary slightly.
Below I show two pieces of satellite-based evidence that will be very closely examined for Fay's potential upgrade. The left panel shows a Synthetic Aperture Radar (SAR) surface wind speed retrieval from 2025 UTC on September 20, and the right panel shows an ASCAT-C scatterometer wind speed retrieval from 2238 UTC, just two hours later. Without going into detail about what these platforms do and their limitations, I'll just point out that they independently suggest that Fay may have been a hurricane at some point -- the scrutiny will be on how reliable those specific hurricane-force wind retrievals are.
 |
| Two independent satellite-based retrievals of hurricane-force surface wind speeds in Fay on September 20: Synthetic Aperture Radar (left) and scatterometer (right). Available here (left) and here (right). |
Gonzalo and Hanna were the next two storms, and they were both short-lived weak tropical storms... Gonzalo near the Cabo Verde Islands, and Hanna in the remote north-central Atlantic. I'll throw the tenth and last storm, Josephine, into this paragraph too, because it was a brief subtropical storm off the northeast US coast. It formed and dissipated in just over a day, and the final advisory of this hurricane season was issued on the afternoon of October 22.
But, let's back up to the ninth storm, Isaias, to wrap up the discussion of the storms. Isaias was one of those storms that we saw coming a very long time in advance. It began as Tropical Depression 9 on October 6, but I wrote my first post on it back on September 30, including a track forecast that ended up looking pretty close to what happened.
It formed in the southern Gulf of Mexico and tracked toward the northeast then turned north toward the western Florida panhandle. It reached Category 1 hurricane intensity early on October 8, Category 2 later that day, and Category 3 early on October 9. That made Isaias not only the first and only hurricane of the season, but the first and only major hurricane of the season.
 |
| Enhanced infrared satellite imagery of Category 3 Hurricane Isaias on October 9. |
During the relatively reliable satellite era (mid-1960s onward), the latest first hurricane formation was September 11. Even if Fay ends up getting upgraded to a hurricane early on September 21 for example, that would be a new record. But going with the current data we have (Fay not being a hurricane), Isaias became the latest first hurricane formation by an even bigger margin: October 8!
Wind shear and dry air took their tolls on Isaias as it neared the coastline, and it weakened to a lop-sided Category 2 hurricane by the time it make landfall... which is still a powerful and destructive storm. As of now, it is responsible for 9 fatalities and over $2 billion in damage.
.gif) |
| Radar animation of Hurricane Isaias making landfall and moving inland on October 9-10. Available at https://bmcnoldy.earth.miami.edu/tropics/radar/. |
It is worth pointing out how rare it is to have just one hurricane in a season. Official records go back to 1851; only two seasons have zero hurricanes (1907 and 1914), and only two other seasons have one hurricane (1905 and 1925). Every other year has anywhere from 2 to 15 hurricanes. Prior to the weather satellite era, we know there are missing storms, we just don't know exactly how many are missing. We also know some known tropical storms were probably hurricanes at some point, we just don't know exactly how many or which ones. The historical record is certainly not perfect, but it's much better over the past 50-60 years. So is it possible that 1905, 1907, 1914, and 1925 all had more than one hurricane? Absolutely. At least during the more reliable satellite era, no year had only one hurricane, until now.
Here is a look at the preliminary track and intensity verification statistics for the entire season. I include a variety of models, and a forecast is only evaluated if all entries were present -- a homogeneous verification. Remarkably, because the storms were all so short-lived, there is not a single verifying 120-hour (5-day) forecast this season! And all six verifying 96-hour (4-day) forecasts are from Tropical Storm Fay. So, we're not dealing with a great sample size here to look at model performance, but it's what we've got!
There are a lot of cryptic acronyms here, so I'll break them down into groups. The National Hurricane Center's official forecasts are in red, labeled "OFCL". Then there are five conventional physics-based regional hurricane models, HWRF, HMON, HAFS-A, HAFS-B, and COAMPS-TC. Then we have the conventional physics-based American global model ("AVNI"), and its ensemble mean forecast ("AEMI"). The next two ("TVCN" and "HCCA") are consensus aides, so not really models by themselves, but tuned combinations of other models. Last but certainly not least is the Google DeepMind AI model ("GDMI").
Clearly, the DeepMind AI model was the most skillful model overall for track... even the NHC's forecasts were less accurate which is not common. That same model dominated the track guidance last season too. There is no doubt that AI models are rapidly changing the tropical cyclone forecasting scene.
For intensity, the list is nearly identical, but I swapped out the two global model entries and replaced them with routine statistical-dynamical models ("DSHP" and "LGEM"). Those models were designed for intensity guidance and were introduced 26 and 20 years ago. Because of their relatively course resolution, global models are historically not used for intensity guidance. And IVCN is an intensity consensus aide, similar to TVCN for track.
Here the outcome is more muddled. The low number of cases make these charts pretty noisy, but that's what we get in a quiet season. It's easy to see the two consensus aides standing out with low errors overall, and the Google DeepMind model is impressive here again, though not as clear of a leader.
September 18 was the 100th anniversary of the Great Miami Hurricane of 1926, and there was quite a lot of special coverage of that centennial across television, social media, and regular online media. I wrote a brief post on August 18, a month ahead of the anniversary, to introduce my new reconstruction of the surface wind field over the storm's lifetime. There's a webpage with lots more detail and maps at https://bmcnoldy.earth.miami.edu/tropics/miami26/, but here's an example:
I will end this post with a bit of personal nostalgia: the season that started this long journey, 1996. Sadly, my emails/posts from 1996-2001 have been lost to time, and I only have the season summaries from 1998-2001. I'm a digital packrat, so it annoys me terribly that I no longer have those.
First is a seasonal track map, made in the same fashion as the current one at the top of the post. You'll recognize a few of the same names that are still in use on this year's list.
I also have a stack of printed infrared satellite image hardcopies from The Weather Channel; here's a photo of one from October 17 showing Hurricane Lili heading for Cuba... who else remembers these? If you were watching Tropical Updates at :50 on TWC in 1996, you may also remember John Hope, Stu Ostro, Brad Edwards, Dave Schwartz, Jim Cantore, Mike Seidel, Jeff Morrow, and so many others. Some have passed on, and some are still in the business!
Three storm names were retired after the 1996 season: Cesar, Fran, and Hortense; they were replaced with Cristobal, Fay, and Hanna for the 2002 season, and all three were just used again in 2026 and will be back in 2032. [see How do storm names get retired?]
Another fun historical tidbit: six years into this blog, the NHC began using the "cone of uncertainty"... here's my post from July 15, 2002 that mentioned it. There are several good nuggets among the posts in the archive, you just have to hunt for them.
Looking ahead, I may use this blog and the mailing list occasionally, but just not for the quasi-daily updates of current activity that you've gotten for the past three decades. If anyone is interested in writing as a guest blogger on a specific relevant topic, let's talk!
Again, thank you all for your interest and support over the years, it definitely kept me motivated!
You will be missed! Thanks for being a trusted voice thru the years.
ReplyDeleteThank you for the kind message!
Delete