@compterendumili.bsky.social revue HCCA
January 26, 2025 at 7:33 PM
@compterendumili.bsky.social revue HCCA
How is the Google DeepMind model doing so far this Atlantic hurricane season?
My *preliminary* homogeneous verification shows that the DeepMind ensemble mean "GDMN" is dominating in both track and intensity, beating the revered "HCCA" consensus and the NHC ("OFCL") at all/most lead times.
[1/2]
My *preliminary* homogeneous verification shows that the DeepMind ensemble mean "GDMN" is dominating in both track and intensity, beating the revered "HCCA" consensus and the NHC ("OFCL") at all/most lead times.
[1/2]
October 22, 2025 at 5:50 PM
How is the Google DeepMind model doing so far this Atlantic hurricane season?
My *preliminary* homogeneous verification shows that the DeepMind ensemble mean "GDMN" is dominating in both track and intensity, beating the revered "HCCA" consensus and the NHC ("OFCL") at all/most lead times.
[1/2]
My *preliminary* homogeneous verification shows that the DeepMind ensemble mean "GDMN" is dominating in both track and intensity, beating the revered "HCCA" consensus and the NHC ("OFCL") at all/most lead times.
[1/2]
The best overall is the HCCA consensus, followed by the Google DeepMind ensemble mean (GDMI). The NHC's forecasts (OFCL) come in third overall. The Navy's COAMPS-TC model (CTCI) is doing well at the 48- and 60-hour lead times. The rest... not so much.
September 23, 2026 at 1:43 PM
The best overall is the HCCA consensus, followed by the Google DeepMind ensemble mean (GDMI). The NHC's forecasts (OFCL) come in third overall. The Navy's COAMPS-TC model (CTCI) is doing well at the 48- and 60-hour lead times. The rest... not so much.
Another thing that's gratifying to me is to see HCCA (the corrected consensus based on past model trends) beating IVCN and TVCN. I worked on the fringes of that effort for many years and pleased to see HCCA adding value. HFIP supported that work.
October 31, 2025 at 5:00 PM
Another thing that's gratifying to me is to see HCCA (the corrected consensus based on past model trends) beating IVCN and TVCN. I worked on the fringes of that effort for many years and pleased to see HCCA adding value. HFIP supported that work.
The HCCA consensus aide is leading, the Google DeepMind ensemble mean (GDMI) continues to do well, and the Navy COAMPS-TC model (CTCI) is doing comparatively well at 2 and 3-day lead times.
The National Hurricane Center (OFCL) is somewhere in the mix between them, depending on lead time.
The National Hurricane Center (OFCL) is somewhere in the mix between them, depending on lead time.
September 24, 2026 at 12:20 PM
The HCCA consensus aide is leading, the Google DeepMind ensemble mean (GDMI) continues to do well, and the Navy COAMPS-TC model (CTCI) is doing comparatively well at 2 and 3-day lead times.
The National Hurricane Center (OFCL) is somewhere in the mix between them, depending on lead time.
The National Hurricane Center (OFCL) is somewhere in the mix between them, depending on lead time.
Et une belle revue HCCA comme on les aime 🥰
November 3, 2025 at 3:28 PM
Et une belle revue HCCA comme on les aime 🥰
In terms of overall accuracy, unlike with intensity the NHC managed to beat the HCCA, albeit barely, on track. Unlike the IVCN, the TVCN was also able to hold a candle to them. HAFS-A/B again failed to convincingly beat HWRF/HMON but were at least not clearly outmatched either.
September 30, 2026 at 12:06 PM
In terms of overall accuracy, unlike with intensity the NHC managed to beat the HCCA, albeit barely, on track. Unlike the IVCN, the TVCN was also able to hold a candle to them. HAFS-A/B again failed to convincingly beat HWRF/HMON but were at least not clearly outmatched either.
Pour les non initiés HCCA = habillement, couchage, campement et ameublement.
Son inventaire est toujours un des grands moments de la vie militaire.
Son inventaire est toujours un des grands moments de la vie militaire.
January 26, 2025 at 7:40 PM
Pour les non initiés HCCA = habillement, couchage, campement et ameublement.
Son inventaire est toujours un des grands moments de la vie militaire.
Son inventaire est toujours un des grands moments de la vie militaire.
The intensity prospects of Rachel may be quite sensitive to her position relative to Polo's cold wake. That being said the NHC sides with the upper-end of the guidance envelope, which is to be fair populated by the HCCA and most FNV3 members, very skilfull guides indeed.
September 29, 2026 at 9:59 PM
The intensity prospects of Rachel may be quite sensitive to her position relative to Polo's cold wake. That being said the NHC sides with the upper-end of the guidance envelope, which is to be fair populated by the HCCA and most FNV3 members, very skilfull guides indeed.
In terms of overall accuracy, the NHC beat every phyiscs based model except one. The venerable HCCA tended to be tad better. Also emberrasing for NCEP is that HAFS-A/B were not able to clearly outperform the legacy models. In fact HWRF was overall the best TC specific model. (2/4)
September 30, 2026 at 12:06 PM
In terms of overall accuracy, the NHC beat every phyiscs based model except one. The venerable HCCA tended to be tad better. Also emberrasing for NCEP is that HAFS-A/B were not able to clearly outperform the legacy models. In fact HWRF was overall the best TC specific model. (2/4)
This is why we are voting for Liz Walsh for Howard County Executive!! Please read, share, and give her your vote!
"I am the only Executive candidate in this race who does not take money from special interests...I am not for sale. I do not owe favors. I serve—exclusively and always—the public good."
"I am the only Executive candidate in this race who does not take money from special interests...I am not for sale. I do not owe favors. I serve—exclusively and always—the public good."
May 12, 2026 at 4:19 PM
This is why we are voting for Liz Walsh for Howard County Executive!! Please read, share, and give her your vote!
"I am the only Executive candidate in this race who does not take money from special interests...I am not for sale. I do not owe favors. I serve—exclusively and always—the public good."
"I am the only Executive candidate in this race who does not take money from special interests...I am not for sale. I do not owe favors. I serve—exclusively and always—the public good."
Been a lot of storms since I ran 2026 AL/EP/CP verifications in early September. Results haven't changed much: GDMI still beating all the guidance for track and slightly better than OFCL on D4-5. For intensity, HCCA best, followed by OFCL and GDMI. Reminder that ECMWF is not in public ATCF decks.
September 30, 2026 at 7:36 PM
Been a lot of storms since I ran 2026 AL/EP/CP verifications in early September. Results haven't changed much: GDMI still beating all the guidance for track and slightly better than OFCL on D4-5. For intensity, HCCA best, followed by OFCL and GDMI. Reminder that ECMWF is not in public ATCF decks.
For track, it's a mixed bag, aside from NHC (OFCL), the top three models were Google DeepMind (GDMI), HWRF (HWFI), and GFS (AVNI).
For intensity, NHC (OFCL) was best overall. Two consensus "models" (HCCA and IVCN) did well, and Google DeepMind was the best individual model overall.
[2/2]
For intensity, NHC (OFCL) was best overall. Two consensus "models" (HCCA and IVCN) did well, and Google DeepMind was the best individual model overall.
[2/2]
July 23, 2026 at 7:00 PM
For track, it's a mixed bag, aside from NHC (OFCL), the top three models were Google DeepMind (GDMI), HWRF (HWFI), and GFS (AVNI).
For intensity, NHC (OFCL) was best overall. Two consensus "models" (HCCA and IVCN) did well, and Google DeepMind was the best individual model overall.
[2/2]
For intensity, NHC (OFCL) was best overall. Two consensus "models" (HCCA and IVCN) did well, and Google DeepMind was the best individual model overall.
[2/2]
トムとジェリー | スゴ技大特集 | WB Kids
YouTube video by WB Kids International
www.youtube.com
December 15, 2025 at 10:54 PM
HCCA is also an early model.
October 22, 2025 at 7:05 PM
HCCA is also an early model.
Yesterday I showed Google DeepMind errors for Erin. Thought I'd repeat the analysis for 2025 as a whole, and add the HCCA corrected consensus to the mix. Sample is mainly Erin, of course. GDMI has been impressive so far, especially through 72 h and for intensity. (EMXI not in live NHC FTP data.)
August 26, 2025 at 8:55 PM
Yesterday I showed Google DeepMind errors for Erin. Thought I'd repeat the analysis for 2025 as a whole, and add the HCCA corrected consensus to the mix. Sample is mainly Erin, of course. GDMI has been impressive so far, especially through 72 h and for intensity. (EMXI not in live NHC FTP data.)
This questionaire is from Howard County Citizens Association - HCCA www.facebook.com/TheHCCA
May 12, 2026 at 4:24 PM
This questionaire is from Howard County Citizens Association - HCCA www.facebook.com/TheHCCA
Two other things stand out. First, the tight clustering of the primary dynamical track model errors. I presume the model tracks themselves aren't clustered, but the mean performance sure seems to be. Second is how much better HCCA is than TVCN. Quite a year for statistics over dynamics, so far.
September 6, 2026 at 8:47 PM
Two other things stand out. First, the tight clustering of the primary dynamical track model errors. I presume the model tracks themselves aren't clustered, but the mean performance sure seems to be. Second is how much better HCCA is than TVCN. Quite a year for statistics over dynamics, so far.
Google DeepMind (GDMI) has done real well this year. When NHC today mentioned the GDM ensembles showing most reaching C5 intensity, I was curious how GDMI was doing for RI. Using the HFIP RI metric, turns out only so-so. HMON, HCCA, and OFCL are the standouts for RI so far. Sample still small.
October 25, 2025 at 6:40 PM
Google DeepMind (GDMI) has done real well this year. When NHC today mentioned the GDM ensembles showing most reaching C5 intensity, I was curious how GDMI was doing for RI. Using the HFIP RI metric, turns out only so-so. HMON, HCCA, and OFCL are the standouts for RI so far. Sample still small.
Tropical storm Rachel is ragged today and only gradual development is expected, especially at first. However the intensity prognosis is interesting. While NOAA Hurricane Models show high C1 to mid C2 peaks, some guidance from HCCA, CTCX and Google DeepMind suggests potential for C3.
September 28, 2026 at 9:24 PM
Tropical storm Rachel is ragged today and only gradual development is expected, especially at first. However the intensity prognosis is interesting. While NOAA Hurricane Models show high C1 to mid C2 peaks, some guidance from HCCA, CTCX and Google DeepMind suggests potential for C3.
So I decided to look at the same sort of plot for Google DeepMind (GDMI) and the HFIP Corrected Consensus (HCCA). These models were the best performers for Lowell and had lower overall forecast errors than OFCL did.
September 9, 2026 at 2:00 AM
So I decided to look at the same sort of plot for Google DeepMind (GDMI) and the HFIP Corrected Consensus (HCCA). These models were the best performers for Lowell and had lower overall forecast errors than OFCL did.
Jesse Welles performing his anti MAGA song "Red" at FarmAid in Minneapolis.
Neil Young also performed.
youtube.com/watch?v=hcCa...
Neil Young also performed.
youtube.com/watch?v=hcCa...
Jesse Welles “Red” | Farm Aid Mpls 9/20/25
YouTube video
youtube.com
October 19, 2025 at 3:11 AM
Jesse Welles performing his anti MAGA song "Red" at FarmAid in Minneapolis.
Neil Young also performed.
youtube.com/watch?v=hcCa...
Neil Young also performed.
youtube.com/watch?v=hcCa...