May 27, 2026 Wednesday Tropical Update
GFS model with a known model bias spinning up bogus-canes
May 2026 was the 29th-warmest May in the 132-year national record (1895–2026), at 61.7°F against a 1991–2020 normal of 61.0°F (+0.6°F).
May 2026 was the 63rd-driest May in the 132-year national record (1895–2026), at 2.9in against a 1991–2020 normal of 3.1in (-0.2in).
NOAA/NCEI Climate at a Glance, contiguous United States. This is the observed record for the month, not a score of anything said in this piece — the forecast ledger begins 2026-08-17 and cannot reach earlier posts.
The NOAA GFS version 16 — flagship global model — has had a serious flaw for several years that causes tropical cyclones to spin up when strong Caribbean —> Gulf of Mexico flow patterns persist. The initial vorticity is shed from the mountains of Colombia and Venezuela (yes — that’s weird) and continues across/near Central America while gaining latitude. The cyclonic vorticity develops into an intense hurricane off the Yucatan and threatens Florida. This was the 12z cycle from earlier Wednesday afternoon.
This is ephemeral or bogus-canes — and will be a false alarm at this intensity. However, that does not discount the otherwise very moist tropical moisture plume.
You may see these forecast cycles with “monster hurricane” shared on social media (X, Facebook, TikTok) but don’t be fooled about GFS. It’s been doing this for years. We hope the next version 17 coming soon (later in 2026) solves this major problem.
Here is the updated 18z forecast cycle from GFS — similar large scale pattern, of course, but a different & less ambitious outcome through 10-12 days. All that being said, we should monitor the Gulf of Mexico and western Caribbean for weak tropical depression / storm systems into early June — that’s the “hot spot” for Atlantic storms in the early part of the season.
Just to set your mind at ease, the ECMWF 12z full-physics ensembles had a blank slate for the Gulf and Caribbean through Day 10.
And, AIFSv2 Ensembles at 18z showed 1 member with potential TS = 2% through 10-days.
My Atlantic Hurricane Season Forecast (2026)
From April 7, 2026: Atlantic Seasonal Hurricane Forecast: (11/5/2)
9-13 Named Storms (11 most likely)
4-7 Hurricanes (5 or 6 most likely)
1-3 Major Hurricanes (2 most likely)
Accumulated Cyclone Energy: 75-85 (well below normal)
NOAA 2026 Atlantic Hurricane Season Forecast
More details in Thursday, May 21, 2026 Tropical Update
Atlantic Basin Satellite Channels
Saharan dust has crossed into Caribbean … milky haze tracks back to the African coast.
GeoColor satellite shows the visible “true color” just as you’d see from space — in geostationary orbit (locked in place to look down) … aside from the puffy clouds and some storms, you can see the milky brown haze = African dust or Saharan Air Layers (SAL) that come off Western Africa into the Atlantic Ocean. SAL is bad for the hurricane business = stable/warm layer that limits instability and convection (storms)
Infrared (Channel 13) colorized —> showers and thunderstorms
Very limited convection over the tropical basin
Water Vapor (Mid-Level Ch. 9) - troughs + areas of descending/dry air
Massive upper-level trough from Florida extending through Bahamas along 25°N latitude to eastern Atlantic = VERY hostile environment in late May.
ECMWF 18z | Precipitation Type + MSLP | Next 6-days
The Intertropical Convergence Zone (ITCZ) in the tropics along 5°N to 10°N is where moist air can ascend and form thunderstorms or like a convective pot boiling over = convection. In the coming weeks, especially into late June, we’ll look for areas of disturbed weather — which means persistent convection — and potential formation of low pressure swirls along the ITCZ with tropical waves, in the Caribbean and Gulf of Mexico (America) with decaying frontal boundaries, and other pathways like tropical transition or subtropical cut-off lows. Lots of options/ways to form a tropical storm in the Atlantic in any given season.
The Bermuda High —> Azores High in the subtropics usually keeps things quiet in the latitude around 30°N and suppresses the ITCZ to the south in late May into June. On the periphery of the Bermuda High, moisture plumes out of the Caribbean into the Gulf and then Texas + Southeast.
NASA GEOS5 (12z) —> African dust plume forecast next 5-days
NASA runs a version of GFS internally that is configured to ingest / test various satellite based instruments. Their dust algorithm with GEOS-5 is very good, and I really like to show the evolution of African dust (Saharan Air Layers) over the Atlantic.
Right on schedule into early June, we have multiple plumes coming off the continent that will sweep into the Caribbean on a 5-7 day frequency.
North Atlantic Tropics | Quiet Next 7-days
3 Tropical Waves — none developing. The eastern most is embedded with the SAL and is heavily suppressed convectively.
Global Ocean Temperatures —> Sea Surface Temperature (SST)
Tropical storms need at least 26°C SST to form and intensify … and hurricanes most often become major with 28°C SST (at least) … the Atlantic has plenty of real estate that could support a TS or H, but nothing percolating as the basin is very hostile or unfavorable to formation. Note, it is the “actual” ocean temperatures that fuel hurricanes, not anomalies or deviations from climatology.
There are still rather cool (relative) areas in the Western Pacific and North Atlantic compared to the growing warmth along the Equator and off the western coast of Mexico with the oncoming El Niño. The recent 20-year climatology highlights much better than an older 1991-2020 baseline.
The Atlantic remains cooler than past 2-decades across the Main Development Region [MDR] from the Lesser Antilles (Caribbean) to the Azores and coast of Africa. This is NOT a positive signal you’d expect for an active season, but Atlantic SST varies on short-time scales (1-2 weeks) and a change in position of the Bermuda/Azores subtropical ridges can adjust trade winds to quickly warm up the Atlantic MDR.
The SAL + strong trade winds should keep the MDR cooler than normal into early June.
My favorite parameter —> Precipitable Water (PWAT) next 6-days
Tropical storms and hurricanes are bundles of high moisture air > 2” in the column (or more) so we look for areas of concentrated high PWAT air in the tropics —> and we also can view frontal boundaries or dry air input from the middle-latitudes into the subtropics or tropics + Saharan Air Layers (dry)
Boatloads of tropical moisture from the Caribbean and Gulf of Mexico surging into Texas to Florida. Heavy rain continues to add up — very welcome, too.
Integrated Vapor Transport —> IVT next 6-days
The transport of rivers of moist air —> IVT is great metric to show tropical storm development. The impressive plume of tropical moisture is easily tracked through the Caribbean into the Gulf of Mexico — with the subtropical ridge pushed well east to the Azores as the massive upper-level trough dominates the Western subtropical Atlantic.
Western Pacific Basin — Typhoon Season is Underway | Jangmi (06W)
Tropical Depression 06W —> Jangmi was named earlier Wednesday (30 knots)
Jangmi does not have an inner-core — but will eventually build one in the next few days.
Track of Jangmi is well east of Taiwan — with southern islands of Japan under most threat at storm expands in size and undergoes extratropical transition.
Ocean Temperatures (SST) very warm > 28°F over the “warm pool” plenty sufficient for any category of Typhoon, assuming one can form, but actually slightly cooler than normal along the northern side of the warm pool, with a sharp gradient north of the Kuroshio Current wall. Jangmi has 3-4 days over the warmest water before hitting the much cooler waters.
HAFS-A mesoscale model — much more ambitious with Jangmi —> Category 4 with minimum pressure in the mid-940s mb … I’m skeptical based upon the current disorganized state of the system, but who knows, maybe it can rapidly intensify.
This is very rapid intensification through 72-hours from HAFS-A. We’ll see!
ECMWF Full Physics Ensemble Tracks — 50 members next 10-days
I’ll note a difference between ECMWF full-physics and the Google Deep Mind … there’s almost no cross-track spread with the Google ensemble forecasts. We want more spread — to capture outliers especially at the outset of any storm development. Something to watch in the coming weeks/months of hurricane season.
ECMWF IVT shows the circulation and intensification of Jangmi (06W)
IVT is under-utilized for tropical storm formation and intensity forecasting
WPAC Precipitation + MSLP Next 6-days … ECMWF 18z in low 970s mb
ECMWF AIFS-Single v2 (Data Driven Forecast System)
ECMWF flagship A.I. weather modeling system — develops a 960 mb typhoon (Cat 3)
And, the ensembles from AIFS-ENS v2 (different than AIFS-Single) — similar track + intensity (18z update)
Eastern Pacific Tropics | Quiet next 7-days | Ghost Town
Thank you for reaching the end of this post! Hopefully you enjoyed the tour of the global tropics and what’s coming in the weeks/months ahead in the tropics here on Weather Trader on Substack.
Maps sourced from weathermodels.com designed and innovated by yours truly! Please subscribe there for real-time access to the newest maps, charts from all of the weather models including ECMWF and AIFS, and NOAA GFS/GEFS, HRRR/RRFS and Blend of Models (NBM). NOAA WPC and SPC for fronts/convective outlooks.
For 2026 Hurricane Season: Every 6 or 12 hours, we get an update including these model output
ECMWF EPS full-physics ensembles … 50 perturbed members + control/HRES
ECMWF AIFS-Ensemble (data driven) … another 50 members + control
ECMWF AIFS-Single … a special deterministic system that has exceptional skill
NOAA GEFSv12 Ensembles … 30 perturbed members + control
NOAA GFSv16 … 13-km high resolution global deterministic
UK Met Office MOGREPS lagged ensembles … global 14-km 72-members per day
Environment Canada ensembles + Germany ICON + Australia BOM Unified Model
New Google WeatherNext 1000-member ensemble system (!) more on that in future post
- Milepost
- 944 · laid 2026-05-28T02:33:45.834000Z
- Source
- https://weather.substack.com/p/may-27-2026-wednesday-tropical-update
- Digest
- 8d0f7a6c71978eb2d3e81b92d087a055292ad8e5ca8eb67434c8f25a6a5d1567
- Machine
- JSON