Chinook Posted October 18 Share Posted October 18 Supertyphoon Champi earlier with over 120kt of base velocity near Guam Link to comment Share on other sites More sharing options...
Chinook Posted October 18 Share Posted October 18 Link to comment Share on other sites More sharing options...
WxWatcher007 Posted October 18 Share Posted October 18 Narrowly escaped a big time hit. 1 Link to comment Share on other sites More sharing options...
vortex95 Posted October 18 Share Posted October 18 On 10/18/2026 at 1:40 AM, WxWatcher007 said: Narrowly escaped a big time hit. Sometimes the tightness/RI *saves* and area from a bigger hit. Melissa over Jamaica was similar. If Melissa had been a long-tracked Cabo Verde, it would have been much larger when striking Jamaica. Sure, it may not have been as intense since multiple ERCs would preclude its sub-900 mb intensity, but a large Cat 4 impacting the entire island vs. only 1/3 of the islands getting Cat 3+ winds is not the same. Also, the wrn 1/3 of Jamaica is significantly less populated, and one of its main airports went largely unscathed, and the other being on the N coast on the wrn side of the island had sig less impact than if it were on the S side. Compare to Gilbert at Cat 4 in Sep 1988 that cut down the full length of the Jamaica E-W w/ the entire land devastated. 80% of home lost their roofs. At least w/ Melissa on about 1/2 of the island, one could find relief, relatively speaking, short-term before full aid arrived and was distributed. 2 1 Link to comment Share on other sites More sharing options...
olafminesaw Posted October 19 Share Posted October 19 Incredible model battle here. I have a feeling I know which one is going to win Link to comment Share on other sites More sharing options...
Chinook Posted October 19 Share Posted October 19 category 5 Link to comment Share on other sites More sharing options...
LongBeachSurfFreak Posted October 19 Share Posted October 19 Incredible model battle here. I have a feeling I know which one is going to win I am assuming that the AI models use historical tracks and climatology more so then the physics models. From my understanding of AI. Not your typical analogs but everything that’s possible. If anyone has more information on the subject I would love to hear it. . Link to comment Share on other sites More sharing options...
olafminesaw Posted October 19 Share Posted October 19 27 minutes ago, LongBeachSurfFreak said: I am assuming that the AI models use historical tracks and climatology more so then the physics models. From my understanding of AI. Not your typical analogs but everything that’s possible. If anyone has more information on the subject I would love to hear it. . Yeah it's mostly machine learning data correlations and such as far as I've heard, but I'm not sure that the specifics are really well understood by the weather community anyway. Actually there was a little segment about the deep mind and Isiais on NPR this morning and they were saying how mostly the project is worked on by non-meteorologist, not even programers with meteorology experience, mostly neural network engineers 2 Link to comment Share on other sites More sharing options...
Chinook Posted October 19 Share Posted October 19 2 hours ago, olafminesaw said: Yeah it's mostly machine learning data correlations and such as far as I've heard, but I'm not sure that the specifics are really well understood by the weather community anyway. Actually there was a little segment about the deep mind and Isiais on NPR this morning and they were saying how mostly the project is worked on by non-meteorologist, not even programers with meteorology experience, mostly neural network engineers People are now talking about how Google Deep-mind ensembles beat the other ones on Hurricane Isaias. 1 Link to comment Share on other sites More sharing options...
olafminesaw Posted Tuesday at 09:41 AM Share Posted Tuesday at 09:41 AM Textbook truck tire Link to comment Share on other sites More sharing options...
LongBeachSurfFreak Posted Tuesday at 09:47 AM Share Posted Tuesday at 09:47 AM Textbook truck tireJTWC has it at 135kt currently. I would say it should be 155kt. This has been one of the most underestimated TCs I have seen. . Link to comment Share on other sites More sharing options...
Windspeed Posted Tuesday at 11:30 AM Share Posted Tuesday at 11:30 AM Impressive large eye that continues to warm. Deep surrounding band of coldest CTs. If you'll notice the geometric pattern within the low-level clouds in the eye, most likely the result of multiple rotating mesovorticies currently riding around the eyewall. I'd imagine this is Cat 5 bound again. 1 Link to comment Share on other sites More sharing options...
olafminesaw Posted Tuesday at 11:47 AM Share Posted Tuesday at 11:47 AM Obviously Taiwan is used to heavy rainfall in the higher terrain on the East side of the island, but 20+40" of rainfall is going to cause some serious flooding I can only imagine.Also check out that rain shadow! Link to comment Share on other sites More sharing options...
WxWatcher007 Posted Tuesday at 05:24 PM Share Posted Tuesday at 05:24 PM 1 Link to comment Share on other sites More sharing options...
Chinook Posted Tuesday at 07:33 PM Share Posted Tuesday at 07:33 PM JTWC had this at 155kt this morning, and so that's an incredible situtation, it's much above the 137kt cutoff for Category 5. 1 Link to comment Share on other sites More sharing options...
olafminesaw Posted Tuesday at 07:39 PM Share Posted Tuesday at 07:39 PM 4 minutes ago, Chinook said: JTWC had this at 155kt this morning, and so that's an incredible situtation, it's much above the 137kt cutoff for Category 5. Pressure data from a buoy on the edge of the eye sugguests sub-900mb pressure Link to comment Share on other sites More sharing options...
LongBeachSurfFreak Posted Tuesday at 07:40 PM Share Posted Tuesday at 07:40 PM Based on old recon finding lots of sub 900 storms in that basin and this having that very high end look it likely was below 890. . 2 Link to comment Share on other sites More sharing options...
olafminesaw Posted Wednesday at 12:56 PM Share Posted Wednesday at 12:56 PM 1 Link to comment Share on other sites More sharing options...
olafminesaw Posted Wednesday at 06:25 PM Share Posted Wednesday at 06:25 PM 1 Link to comment Share on other sites More sharing options...
Chinook Posted Friday at 02:06 PM Share Posted Friday at 02:06 PM gust to 89 knots at Yonaguni Island, Japan (outside of Taiwan, which is on the left) Link to comment Share on other sites More sharing options...
NorthHillsWx Posted Friday at 05:32 PM Share Posted Friday at 05:32 PM On 10/19/2026 at 2:20 PM, LongBeachSurfFreak said: I am assuming that the AI models use historical tracks and climatology more so then the physics models. From my understanding of AI. Not your typical analogs but everything that’s possible. If anyone has more information on the subject I would love to hear it. . Huge fail for AI models. Champi just made landfall on Taiwan 1 Link to comment Share on other sites More sharing options...
NorthHillsWx Posted Friday at 05:35 PM Share Posted Friday at 05:35 PM https://www.cwa.gov.tw/V8/E/W/OBS_Radar.html Wild radar loop showing land interaction heavily influencing track (normal for cyclones near Taiwan) 2 Link to comment Share on other sites More sharing options...
vortex95 Posted 22 hours ago Share Posted 22 hours ago On 10/23/2026 at 1:32 PM, NorthHillsWx said: Huge fail for AI models. Champi just made landfall on Taiwan Yes, so this is a lesson about AI models. They are not infallible and can be off quite a bit. Just I see some putting them up on pedestal and worshiping as to having the final say. And look at what a physics-bases model (ECMWF ENS) did. It even caught the odd motion from terrain influences so common for when a TY approaches Tawain! 1 Link to comment Share on other sites More sharing options...
vortex95 Posted 22 hours ago Share Posted 22 hours ago On 10/23/2026 at 1:35 PM, NorthHillsWx said: https://www.cwa.gov.tw/V8/E/W/OBS_Radar.html Wild radar loop showing land interaction heavily influencing track (normal for cyclones near Taiwan) Complete archived loop here! 1 1 Link to comment Share on other sites More sharing options...
vortex95 Posted 1 hour ago Share Posted 1 hour ago On 10/20/2026 at 3:40 PM, LongBeachSurfFreak said: Based on old recon finding lots of sub 900 storms in that basin and this having that very high end look it likely was below 890. . Odds are high that there have been at least 6 other STYs stronger than Tip. This short paper from 2004 notes 2, and there have been several since (Haiyan comes to mind). stronger.pdf Link to comment Share on other sites More sharing options...
vortex95 Posted 1 hour ago Share Posted 1 hour ago On 10/20/2026 at 1:24 PM, WxWatcher007 said: That structure is likely an artifact from precip in this case. Other analyses showed Champi had completed an ERC earlier and there was no concentric eyewalls or more than one wind max. Link to comment Share on other sites More sharing options...
vortex95 Posted 52 minutes ago Share Posted 52 minutes ago On 10/19/2026 at 5:50 PM, Chinook said: People are now talking about how Google Deep-mind ensembles beat the other ones on Hurricane Isaias. I find this a bit misleading in the larger pix. Of course a newer model is going to be pushed and pointed out more often -- b/c it is new! Makes for better news stories. Also, beware of cherry-picking. For any given fcst, you often can find a model (AI or otherwise) that did much better than the others. Par for the course. Just is it pointed out a lot in the MSM or not? It's like the ECMWF praise that was big after Sandy. Sure, it nailed far out a mega event, but the ECMWF overall skill perception was somewhat tainted by this very fact. It did not mean that the ECMWF was the GOAT and panacea for all out fcst woes going forward! I recall ppl using the ECMWF blindly for fcst solutions b/c it was the ECMWF! That's not good fcst science. One needs to see stats for all fcsts for the model, not just one that happens to get a lot of attention and had high sensible wx impact! That does not mean GDMI is useless or should not be praised, but let's not get carried away. Link to comment Share on other sites More sharing options...
vortex95 Posted 15 minutes ago Share Posted 15 minutes ago On 10/19/2026 at 2:54 PM, olafminesaw said: Yeah it's mostly machine learning data correlations and such as far as I've heard, but I'm not sure that the specifics are really well understood by the weather community anyway. Actually there was a little segment about the deep mind and Isiais on NPR this morning and they were saying how mostly the project is worked on by non-meteorologist, not even programers with meteorology experience, mostly neural network engineers "were saying how mostly the project is worked on by non-meteorologist, not even programmers with meteorology experience, mostly neural network engineers." That's a big caveat, and something that I think is lost in the translation often. Example in the general pix. One can educate themselves about the wx and claim, "I've observed and experienced all sorts of wx all my life!" But that does not make one an expert on the subject by default. It would be like saying, "I've known and driven cars most of my life!" but that does not make one an expert on cars! Knowing is not enough. You need to know how to *apply* the knowledge w/ insight and wisdom only acquired by deep interest and study of a given subject. Having a formal education is key b/c if provides a lot of the basics and shows how other sciences connect well w/ meteorology, as an example. You then take that formal education, and then apply it in job in the field, gaining experience and countless tidbits over time that round out thing nicely. You also learn all the slang/jargon and the little nuances/idiosyncrasies. The fine details count! Stuff that you can often use to call out posers/hustlers/AI-generated content from the real deal! Don't get me wrong, to be a wx expert, you do not need a full 4-year degree. The Internet changes all that and if you put your mind to it and have the passion, you can get all the info ever needed from the web! Plus, forums like this you get the details and nuances that round things out nicely, as I mentioned earlier. So circling around back to the main topic at hand, when people that are not experts in a given field do work on products/services for the same field, there are caveats, as I said. Programming a model but not understanding exactly what the model does for practical application is non-trivial, and can lead to things said about the model from such programmers not in-line w/ the fcst meteorology and its limits! They can get idealistic w/o regard to practically. AI models are still bound by fcst chaos theory and in their case, the number of historical cases they drawn upon, so that becomes an issue when you get rare or outlier events. "but I'm not sure that the specifics are really well understood by the weather community anyway." Good point. I think most only understand the basics from a bird's eye view., including myself. It is a learning curve, as w/ all new tech. Unfortunately, some put the wagon in front of the horse, getting way ahead of themselves simply for content (hype) or shill purposes AI as a concept is has long been ingrained in out culture from movies like The Terminator and The Matrix, making us wary, cautions, and fearful, but that's only Hollywood! That does not mean it is reality or how things actually work! Link to comment Share on other sites More sharing options...
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now