ATL: BERYL - Models

[toad strangler]

 https://twitter.com/andyhazelton/status/1810119506066375036

[Steve]

Let the king model for Beryl take a bow. Unreal model performance vs the suspects

https://www.tropicaltidbits.com/analysi ... 70800&fh=6

[capNstorms]

July 7th 1100PM CDT Update: National Weather Service Doppler radar near Houston, Texas and
@53rdWRS aircraft data indicate #Beryl has strengthened and is again a hurricane with maximum sustained winds of 75 mph.

[cycloneye]

In my estimation, HAFS depictions of the dry air muddling the core in GOM were correct and was the best delaying the RI. ICON also did well.

[LarryWx]

Just for the record, both the 12Z June 30th ICON and JMA had a landfall at Galveston! And many later runs of both have been similar. Are those two going to score a huge win?

Answer: absolutely yes, a big win for ICON and JMA for the W GOM track/TX landfall portion!

[LarryWx]

12Z JMA hooks it right into Freeport. Just for fun...

Not because it is a good model (it isn’t at all), but for the record just in case the highly unlikely landfall at the upper TX coast were to occur:

these 12Z JMA runs landfall at the upper TX coast: 6/30, 7/2, 7/3, 7/4…so 4 of the last 5 12Z JMA runs

Actual landfall wasn’t up at upper TX coast but still the 6/30 and 7/2-4 JMA runs into Galveston were closer to actual landfall than GFS/Euro/UKMET. Only the ICON was overall about as close. So, kudos to the JMA for the W GOM/TX landfall!

[LarryWx]

As noted above, ICON was best of the more respected globals with the W GOM landfall in C TX. This goes back to the 12Z 6/30 run into Galveston! It then had runs up into LA or near the TX/LA border as late as the 6Z on 7/3, which weren’t good. However, starting on 12Z 7/3 (when ICON hit C TX while GFS, Euro, CMC, and UKMET were into NE MX) onward ICON was always between Galveston and Corpus. So, major kudos! CMC wasn’t consistently TX til 0Z of 7/6. UKMET wasn’t til 12Z of 7/5. Euro/GFS wasn’t til 0Z of 7/5.

(ICON also did well with Ian for FL and SC landfalls (2nd best to UKMET)).

 JMA gets a well deserved honorable mention as it was similar to the ICON with its 12Z 6/30 run and had several more runs (12Z runs on 6/2, 3, 4) into Galveston followed by 6/5-6 at CC. Actual landfall wasn’t up at upper TX coast but still the 6/30 and 7/2-4 JMA runs into Galveston were much closer to actual landfall than GFS/Euro/UKMET then. Only the ICON was overall about as close.

  So, my ratings of globals for W GOM portion of track:

Best: ICON and JMA

Next best but not close: GFS/Euro

Further back: UKMET

Worst: CMC

[got ants?]

12Z JMA hooks it right into Freeport. Just for fun...

Not because it is a good model (it isn’t at all), but for the record just in case the highly unlikely landfall at the upper TX coast were to occur:

these 12Z JMA runs landfall at the upper TX coast: 6/30, 7/2, 7/3, 7/4…so 4 of the last 5 12Z JMA runs

Actual landfall wasn’t up at upper TX coast but still the 6/30 and 7/2-4 JMA runs into Galveston were closer to actual landfall than GFS/Euro/UKMET. Only the ICON was overall about as close. So, kudos to the JMA for the W GOM/TX landfall!

This was crazy! Sunday before last, whike tshe was just sout, southeadt of Barbados, I wish-cadted it hitting Carriacou, and was told too early, but I had a feeling. Then on approach to Jamaica, while calls for a turn south, and shear, I again, couldn"t see that action, and wish-cadted it full on landfall on Kingston (thank God it didn't at cat 5). But as far back as that, somecwere already calling for a Texas hit, with NHC saying nope. Cray Cray, was thus storm

[tolakram]

So the question that has to be asked is did these models win due to skill, or did they win due to a consistent right bias? It seems to me a right bias usually wins in the gulf, which is why the Euro, with a consistent left bias, usually loses to the GFS, but only in the gulf.

[LarryWx]

So the question that has to be asked is did these models win due to skill, or did they win due to a consistent right bias? It seems to me a right bias usually wins in the gulf, which is why the Euro, with a consistent left bias, usually loses to the GFS, but only in the gulf.

Great Q!
1. JMA:
I’m pretty sure JMA doesn’t have a consistent right bias. If anything, I’d lean toward left bias because I recall many times when it threatened the SE coast for TCs that ended up staying well offshore.

The JMA having done well in the Gulf is by no means a reason by itself to suddenly promote it to respectable as it has more often than not proven to be highly inaccurate. A step above NAVGEM would likely still be as far as I’d now go. The most likely way I’d possibly consider promoting it to respectable in the near future is if I discover it has recently had a major upgrade and no longer is the semi-joke it has been overall. Anyone know?

But even that possible path to promotion to respectability would require consistently decent results vs major models with near future storms.

2. ICON:
A totally different story from JMA as regards already being pretty respectable. For example, it was easily 2nd best to UKMET for the Ian FL/SC landfalls as documented here. Also, it did the best with the historic 1/3/2018 SE coastal winter storm as it on many early runs had the surface low off FL move NNE instead of NE like the other major models thus keeping left side of main precip shield onshore. So, in general unlike JMA, I consider ICON to be a serious model.

Does ICON have a consistent right bias? Not that I’m aware of but I’m not 100% sure. It sure didn’t with the 1/3/18 winter storm. But IF it does, that likely would have helped it with the Ian FL landfall further S.