Google DeepMind shows a 33% chance of a tropical storm. ECMWF's tropical cyclone probabilities from EPS & AIFS exceed 80%. Surely that means this nor'easter will be a tropical storm, right?
Short answer... it's not that simple.
Breaking down why in this technical thread 🧵
Short answer... it's not that simple.
Breaking down why in this technical thread 🧵
September 25, 2026 at 1:24 AM
Google DeepMind shows a 33% chance of a tropical storm. ECMWF's tropical cyclone probabilities from EPS & AIFS exceed 80%. Surely that means this nor'easter will be a tropical storm, right?
Short answer... it's not that simple.
Breaking down why in this technical thread 🧵
Short answer... it's not that simple.
Breaking down why in this technical thread 🧵
Milton's track is so odd it's outsmarted ECMWF's GIS assumptions
There is a long technical explanation for why ECMWF’s tracker is having issues, but the short version is it partitions tropical cyclones into basins, and their Atlantic & Pacific basin border is delineated around 90W.
The problem? Milton is west of 90W and crossing that boundary from west to east.
The problem? Milton is west of 90W and crossing that boundary from west to east.
October 6, 2024 at 2:53 AM
Milton's track is so odd it's outsmarted ECMWF's GIS assumptions
Is the Gulf sea breeze the ECMWF’s Katrina?
Question: Did the European operational model change how it handles sea breeze convection? Because this "exploding" diurnal activity it has been showing for a few days is bizarre.
June 13, 2025 at 4:17 AM
Is the Gulf sea breeze the ECMWF’s Katrina?
2024 is now virtually certain to be the warmest year on record, and will be above 1.5C in most surface temperature datasets, including Copernicus/ECMWF's ERA5 shown here:
October 28, 2024 at 3:18 PM
2024 is now virtually certain to be the warmest year on record, and will be above 1.5C in most surface temperature datasets, including Copernicus/ECMWF's ERA5 shown here:
To illustrate this, let's look at the 12z AIFS - which ECMWF's tracker identified as a tropical cyclone at hour 48.
Looking at its model fields, this is clearly a baroclinic cyclone associated with synoptic forcing undergoing warm seclusion:
Looking at its model fields, this is clearly a baroclinic cyclone associated with synoptic forcing undergoing warm seclusion:
September 25, 2026 at 1:25 AM
To illustrate this, let's look at the 12z AIFS - which ECMWF's tracker identified as a tropical cyclone at hour 48.
Looking at its model fields, this is clearly a baroclinic cyclone associated with synoptic forcing undergoing warm seclusion:
Looking at its model fields, this is clearly a baroclinic cyclone associated with synoptic forcing undergoing warm seclusion:
All of the plots in the post above this are from objective algorithms that use global model & ensemble member data to estimate whether a cyclone is tropical or non-tropical.
This is no easy task, and can mis-identify systems often. Per ECMWF's own explanation:
This is no easy task, and can mis-identify systems often. Per ECMWF's own explanation:
September 25, 2026 at 1:24 AM
All of the plots in the post above this are from objective algorithms that use global model & ensemble member data to estimate whether a cyclone is tropical or non-tropical.
This is no easy task, and can mis-identify systems often. Per ECMWF's own explanation:
This is no easy task, and can mis-identify systems often. Per ECMWF's own explanation:
This is the time step when ECMWF's objective algorithm has AIFS peaking in intensity while cateogorized as a tropical cyclone.
A simple look at this map is a clear tell to any meteorologist or experienced forecaster that this is not a tropical cyclone:
A simple look at this map is a clear tell to any meteorologist or experienced forecaster that this is not a tropical cyclone:
September 25, 2026 at 1:26 AM
This is the time step when ECMWF's objective algorithm has AIFS peaking in intensity while cateogorized as a tropical cyclone.
A simple look at this map is a clear tell to any meteorologist or experienced forecaster that this is not a tropical cyclone:
A simple look at this map is a clear tell to any meteorologist or experienced forecaster that this is not a tropical cyclone:
I work at #ECMWF on improving the representation of radiative transfer in weather and climate models. This staff profile article was published today (thanks to my colleague Georg Lentze). www.ecmwf.int/en/about/med...
Accounting for solar and infrared radiation in ECMWF’s weather forecasts
Robin Hogan is responsible for the representation of solar and infrared radiation in the model used for ECMWF’s weather forecasts. He also co-chairs the mission advisory group for the EarthCARE satell...
www.ecmwf.int
November 18, 2024 at 12:10 PM
I work at #ECMWF on improving the representation of radiative transfer in weather and climate models. This staff profile article was published today (thanks to my colleague Georg Lentze). www.ecmwf.int/en/about/med...
This enormous Absorbed Solar Radiation and net-flux increase means that there's a lot more warming in the pipeline!
I estimate about 0.5°C additional regional warming over the next decade.
See regional 12-m temperature anomaly since 1940
From ECMWF's ERA5 data.
Nov-Oct to include 2024.
I estimate about 0.5°C additional regional warming over the next decade.
See regional 12-m temperature anomaly since 1940
From ECMWF's ERA5 data.
Nov-Oct to include 2024.
December 9, 2024 at 8:54 AM
This enormous Absorbed Solar Radiation and net-flux increase means that there's a lot more warming in the pipeline!
I estimate about 0.5°C additional regional warming over the next decade.
See regional 12-m temperature anomaly since 1940
From ECMWF's ERA5 data.
Nov-Oct to include 2024.
I estimate about 0.5°C additional regional warming over the next decade.
See regional 12-m temperature anomaly since 1940
From ECMWF's ERA5 data.
Nov-Oct to include 2024.
🌍 Explore how ECMWF’s 2024 #CodeforEarth projects leveraged #CopernicusAtmosphere and #CopernicusClimate data to develop innovative open-source coding solutions to Earth system modelling challenges.
Dive into the 2024 projects!
🔗 bit.ly/4gbLdwr
🔗 bit.ly/4ht4nyE
Dive into the 2024 projects!
🔗 bit.ly/4gbLdwr
🔗 bit.ly/4ht4nyE
January 23, 2025 at 1:23 PM
🌍 Explore how ECMWF’s 2024 #CodeforEarth projects leveraged #CopernicusAtmosphere and #CopernicusClimate data to develop innovative open-source coding solutions to Earth system modelling challenges.
Dive into the 2024 projects!
🔗 bit.ly/4gbLdwr
🔗 bit.ly/4ht4nyE
Dive into the 2024 projects!
🔗 bit.ly/4gbLdwr
🔗 bit.ly/4ht4nyE
50th anniversary celebrations are happening at ECMWF’s Reading site today.
As distinguished guests gather, take a look at what was happening in 1979 when His Majesty King Charles, as Prince of Wales, opened ECMWF’s building where it’s all going on today. 📸
As distinguished guests gather, take a look at what was happening in 1979 when His Majesty King Charles, as Prince of Wales, opened ECMWF’s building where it’s all going on today. 📸
December 3, 2025 at 3:03 PM
50th anniversary celebrations are happening at ECMWF’s Reading site today.
As distinguished guests gather, take a look at what was happening in 1979 when His Majesty King Charles, as Prince of Wales, opened ECMWF’s building where it’s all going on today. 📸
As distinguished guests gather, take a look at what was happening in 1979 when His Majesty King Charles, as Prince of Wales, opened ECMWF’s building where it’s all going on today. 📸
Google claims that in tests for 2023–24 storms, its 5-day track forecasts were ~85 miles closer to actual tracks than ECMWF’s ensemble and their AI model outperformed NOAA’s best intensity model – the HAFS – on intensity forecasts. www.wfla.com/weather/clim...
June 25, 2025 at 11:59 AM
Google claims that in tests for 2023–24 storms, its 5-day track forecasts were ~85 miles closer to actual tracks than ECMWF’s ensemble and their AI model outperformed NOAA’s best intensity model – the HAFS – on intensity forecasts. www.wfla.com/weather/clim...
Global temperatures are now available from
Copernicus/ECMWF's ERA5 reanalysis product for 2023.
They find it was the warmest year on record by a large margin, at 1.48C above preindustrial levels, higher than the 1.43C that JRA-55 reported earlier this week:
Copernicus/ECMWF's ERA5 reanalysis product for 2023.
They find it was the warmest year on record by a large margin, at 1.48C above preindustrial levels, higher than the 1.43C that JRA-55 reported earlier this week:
January 4, 2024 at 6:20 PM
Global temperatures are now available from
Copernicus/ECMWF's ERA5 reanalysis product for 2023.
They find it was the warmest year on record by a large margin, at 1.48C above preindustrial levels, higher than the 1.43C that JRA-55 reported earlier this week:
Copernicus/ECMWF's ERA5 reanalysis product for 2023.
They find it was the warmest year on record by a large margin, at 1.48C above preindustrial levels, higher than the 1.43C that JRA-55 reported earlier this week:
📣 𝗪𝗲’𝗿𝗲 𝗵𝗮𝗽𝗽𝘆 𝘁𝗼 𝗮𝗻𝗻𝗼𝘂𝗻𝗰𝗲 𝗮 𝘀𝗶𝗴𝗻𝗶𝗳𝗶𝗰𝗮𝗻𝘁 𝘂𝗽𝗱𝗮𝘁𝗲 𝘁𝗼 𝗘𝗖𝗠𝗪𝗙’𝘀 𝗸𝗲𝘆 𝗳𝗼𝗿𝗲𝗰𝗮𝘀𝘁𝗶𝗻𝗴 𝘀𝘆𝘀𝘁𝗲𝗺𝘀 𝗜𝗙𝗦 𝗮𝗻𝗱 𝗔𝗜𝗙𝗦.
All updates are live low. Read more and discover the key improvements ➡️ www.ecmwf.int/en/about/med...
All updates are live low. Read more and discover the key improvements ➡️ www.ecmwf.int/en/about/med...
May 13, 2026 at 8:56 AM
📣 𝗪𝗲’𝗿𝗲 𝗵𝗮𝗽𝗽𝘆 𝘁𝗼 𝗮𝗻𝗻𝗼𝘂𝗻𝗰𝗲 𝗮 𝘀𝗶𝗴𝗻𝗶𝗳𝗶𝗰𝗮𝗻𝘁 𝘂𝗽𝗱𝗮𝘁𝗲 𝘁𝗼 𝗘𝗖𝗠𝗪𝗙’𝘀 𝗸𝗲𝘆 𝗳𝗼𝗿𝗲𝗰𝗮𝘀𝘁𝗶𝗻𝗴 𝘀𝘆𝘀𝘁𝗲𝗺𝘀 𝗜𝗙𝗦 𝗮𝗻𝗱 𝗔𝗜𝗙𝗦.
All updates are live low. Read more and discover the key improvements ➡️ www.ecmwf.int/en/about/med...
All updates are live low. Read more and discover the key improvements ➡️ www.ecmwf.int/en/about/med...
📣ECMWF’s Real-time Catalogue is now open to all, supporting global efforts to improve early warnings and build resilience.
Find out more about what this means and the importance of attribution ➡️ www.ecmwf.int/en/about/med...
Find out more about what this means and the importance of attribution ➡️ www.ecmwf.int/en/about/med...
October 1, 2025 at 12:17 PM
📣ECMWF’s Real-time Catalogue is now open to all, supporting global efforts to improve early warnings and build resilience.
Find out more about what this means and the importance of attribution ➡️ www.ecmwf.int/en/about/med...
Find out more about what this means and the importance of attribution ➡️ www.ecmwf.int/en/about/med...
FYI: there are issues with ECMWF’s ensemble tropical cyclone tracker for Tropical Storm Milton. I activated my own in-house ensemble tracker, which is available on my website:
arctic.som.ou.edu/tburg/produc...
arctic.som.ou.edu/tburg/produc...
October 6, 2024 at 2:27 AM
FYI: there are issues with ECMWF’s ensemble tropical cyclone tracker for Tropical Storm Milton. I activated my own in-house ensemble tracker, which is available on my website:
arctic.som.ou.edu/tburg/produc...
arctic.som.ou.edu/tburg/produc...
Final set of additions to PolarWx for the time being - two AI emulator models (ECMWF's AIFS, and the GFS-based GraphCast) are now plotting in real time:
polarwx.com/models/?mode...
polarwx.com/models/?mode...
polarwx.com/models/?mode...
polarwx.com/models/?mode...
January 21, 2025 at 1:35 AM
Final set of additions to PolarWx for the time being - two AI emulator models (ECMWF's AIFS, and the GFS-based GraphCast) are now plotting in real time:
polarwx.com/models/?mode...
polarwx.com/models/?mode...
polarwx.com/models/?mode...
polarwx.com/models/?mode...
As summer fades in the Northern Hemisphere, the zonal-mean zonal winds at 10 hPa and 60°N have returned to westerly.
The 2025–2026 Arctic stratospheric polar vortex is born.
The 2025–2026 Arctic stratospheric polar vortex is born.
August 25, 2025 at 1:55 PM
As summer fades in the Northern Hemisphere, the zonal-mean zonal winds at 10 hPa and 60°N have returned to westerly.
The 2025–2026 Arctic stratospheric polar vortex is born.
The 2025–2026 Arctic stratospheric polar vortex is born.
It's currently just a building site I walk past on campus.
But soon this will be the European Centre for Medium-Range Weather Forecasts' (ECMWF's) new UK home.
I'm excited we're creating the world's largest concentration of weather and climate scientists, right here at the University of Reading.
But soon this will be the European Centre for Medium-Range Weather Forecasts' (ECMWF's) new UK home.
I'm excited we're creating the world's largest concentration of weather and climate scientists, right here at the University of Reading.
November 21, 2024 at 8:38 PM
It's currently just a building site I walk past on campus.
But soon this will be the European Centre for Medium-Range Weather Forecasts' (ECMWF's) new UK home.
I'm excited we're creating the world's largest concentration of weather and climate scientists, right here at the University of Reading.
But soon this will be the European Centre for Medium-Range Weather Forecasts' (ECMWF's) new UK home.
I'm excited we're creating the world's largest concentration of weather and climate scientists, right here at the University of Reading.
Collaboration & innovation, with people at the core, are key to ECMWF’s success, says new Director-General Florian Pappenberger.
With nearly a decade leading forecasting & four years as Deputy Director-General, he has guided major scientific & technical transformation ➡️ www.ecmwf.int/en/about/med...
With nearly a decade leading forecasting & four years as Deputy Director-General, he has guided major scientific & technical transformation ➡️ www.ecmwf.int/en/about/med...
January 8, 2026 at 11:35 AM
Collaboration & innovation, with people at the core, are key to ECMWF’s success, says new Director-General Florian Pappenberger.
With nearly a decade leading forecasting & four years as Deputy Director-General, he has guided major scientific & technical transformation ➡️ www.ecmwf.int/en/about/med...
With nearly a decade leading forecasting & four years as Deputy Director-General, he has guided major scientific & technical transformation ➡️ www.ecmwf.int/en/about/med...
ECMWF's HPC system has around 1 million processors and 248 Graphics Processing Units that can perform many calculations in parallel.
Around 1.5 exabytes (1,500,000,000,000,000,000 bytes) of data are currently stored, equivalent to roughly 175 billion music tracks.
www.ecmwf.int/en/about/med...
Around 1.5 exabytes (1,500,000,000,000,000,000 bytes) of data are currently stored, equivalent to roughly 175 billion music tracks.
www.ecmwf.int/en/about/med...
August 31, 2026 at 7:12 AM
ECMWF's HPC system has around 1 million processors and 248 Graphics Processing Units that can perform many calculations in parallel.
Around 1.5 exabytes (1,500,000,000,000,000,000 bytes) of data are currently stored, equivalent to roughly 175 billion music tracks.
www.ecmwf.int/en/about/med...
Around 1.5 exabytes (1,500,000,000,000,000,000 bytes) of data are currently stored, equivalent to roughly 175 billion music tracks.
www.ecmwf.int/en/about/med...
ECMWF’s data services are more than repositories: they are the platforms, interfaces, applications and support that enable users across Europe – and beyond – to turn Earth system information into decisions.
Find out more about ECMWF's data ecosystem
➡️ www.ecmwf.int/en/about/med...
Find out more about ECMWF's data ecosystem
➡️ www.ecmwf.int/en/about/med...
May 29, 2026 at 8:06 AM
ECMWF’s data services are more than repositories: they are the platforms, interfaces, applications and support that enable users across Europe – and beyond – to turn Earth system information into decisions.
Find out more about ECMWF's data ecosystem
➡️ www.ecmwf.int/en/about/med...
Find out more about ECMWF's data ecosystem
➡️ www.ecmwf.int/en/about/med...
The @wmo-global.bsky.social Global State of the Water Resources 2024 Report includes a contribution from #ECMWF. We provided data from both ECMWF’s ecLand and Copernicus EMS GloFAS, and 10 other global hydrological modelling systems, to build a global picture of changing river discharge in 2024.🧵
September 18, 2025 at 10:12 AM
The @wmo-global.bsky.social Global State of the Water Resources 2024 Report includes a contribution from #ECMWF. We provided data from both ECMWF’s ecLand and Copernicus EMS GloFAS, and 10 other global hydrological modelling systems, to build a global picture of changing river discharge in 2024.🧵
🌍 Have you read our #CodeforEarth 2024 articles?
See how ECMWF’s initiative tackled Earth system challenges with #CopernicusAtmosphere & #CopernicusClimate data.
🌟 Get inspired for the 2025 edition!
🔗 C3S: bit.ly/4gbLdwr
🔗 CAMS: bit.ly/4ht4nyE
See how ECMWF’s initiative tackled Earth system challenges with #CopernicusAtmosphere & #CopernicusClimate data.
🌟 Get inspired for the 2025 edition!
🔗 C3S: bit.ly/4gbLdwr
🔗 CAMS: bit.ly/4ht4nyE
January 31, 2025 at 9:06 AM
🌍 Have you read our #CodeforEarth 2024 articles?
See how ECMWF’s initiative tackled Earth system challenges with #CopernicusAtmosphere & #CopernicusClimate data.
🌟 Get inspired for the 2025 edition!
🔗 C3S: bit.ly/4gbLdwr
🔗 CAMS: bit.ly/4ht4nyE
See how ECMWF’s initiative tackled Earth system challenges with #CopernicusAtmosphere & #CopernicusClimate data.
🌟 Get inspired for the 2025 edition!
🔗 C3S: bit.ly/4gbLdwr
🔗 CAMS: bit.ly/4ht4nyE