How reliable is the GFDL model?

[logybogy]

It's showing a northern turn the other models are not.

[LCfromFL]

I copied this from another post:

I've been staining our cedar siding all day and was thinking "what can I do that's more fun than this?". So I ran a verification though 18Z on Frances.

Through 72 hours, the overall best performance has been turned in by the lowly LBAR

Second place: TPC (they are #1 at 48 hrs)

UKMET is second to last so far...only thing worse than the UKie to date has been the A98E (barf).

GFDL not done well inside 72 hrs. GFS was better last year with Carib/GOM storms than thus far with Frances so we'll see.

The "unofficial" stands (EC is not included)

1. LBAR
2. Official TPC/NHC
3. GFDL
4. BAMD
5. GFS
6. CMC
7. NOGAPS
8. BAMM
9. UKMET
10 A98E

chug-chug lil' LBAR...you're the model that could.

UKMET and LBAR

Scott

...so it seems that the GFDL isn't doing all that badly (at least outside of 72 hours). This will be interesting to watch...and to use that old euphamism...."only time will tell".

[logybogy]

What I'm interested in is why the GFDL is showing the northern turn and the other models are not.

What does this model do to the ridge that the others don't?

[CalmBeforeStorm]

It's showing a northern turn the other models are not.

Interesting, though, that if you run the GFDL depiction of Gaston, the model shows a more southerly course for Frances, maybe because it's stronger in that run.