ATL: DORIAN - Models

[SouthFLTropics]

I think it’s safe to assume a further South track adjustment will be warranted at 11:00AM.

[lando]

6z GFS OPERATIONAL LANDFALL- cape canaveral 102-108 hr

[ronjon]

Looks like 06z GFS moving toward its ensemble mean. The operational was an NE outlier of its ensemble runs.

[ronjon]

6z GFS OPERATIONAL LANDFALL- cape canaveral 102-108 hr

Well this will make the NHC staff very happy today. Finally getting our two major globals closer.

[KWT]

GFS is being disregarded by the NHC so not worth using as it’s weakening the ridge to quickly.

Where do you get this information? Seriously, every time I come in here I hear the craziest stuff.

The GFS
is a northern outlier from the rest of the guidance, showing a
weaker ridge and bringing Dorian close to the Florida/Georgia
border, while the UKMET and ECMWF models show stronger ridges
and remain the southernmost solutions near South Florida. Given the
spread in the guidance, the new NHC forecast blends the previous
forecast with the simple multi-model consensus aids and the HCCA and
Florida State Superensemble aids, and it actually ends up in a
position very close to the tightly clustered global model ensemble
means. The most notable change in the new forecast is that it's a
little bit slower than the previous one as Dorian approaches
Florida.

They're slowing it down because now a recurve very near the coast is in the cards. The GFS is part of the multi-model consensus.

To be fair, the NHC doesn't say that either, they just mention the models are a little bit slower, you've added up own subjective view into that part of the NHC discussion

Anyway the GFS suite is also slower this run, along with the 06z ICON, so makes sense for them to slow it all down.

[STRiZZY]

GFS is being disregarded by the NHC so not worth using as it’s weakening the ridge to quickly.

Where do you get this information? Seriously, every time I come in here I hear the craziest stuff.

The GFS
is a northern outlier from the rest of the guidance, showing a
weaker ridge and bringing Dorian close to the Florida/Georgia
border, while the UKMET and ECMWF models show stronger ridges
and remain the southernmost solutions near South Florida. Given the
spread in the guidance, the new NHC forecast blends the previous
forecast with the simple multi-model consensus aids and the HCCA and
Florida State Superensemble aids, and it actually ends up in a
position very close to the tightly clustered global model ensemble
means. The most notable change in the new forecast is that it's a
little bit slower than the previous one as Dorian approaches
Florida.

They're slowing it down because now a recurve very near the coast is in the cards. The GFS is part of the multi-model consensus.

I think what you meant was the re-curve near the coast is in the cards because it's slowing down. Not the other way around.

The ECMWF has consistently modeled the northern turn to occur around Wednesday. That didn't change with this latest run.

What did change is the movement of Dorian westward was modeled much slower, thus the northern turn occurring over the Florida peninsula instead of the GOM.

[chris_fit]

GFS Legacy still no landfall and It's Monday afternoon...

[jlauderdal]

icon, euro and gfs all slow it around the same period, that much they agree on....miami to jax still in play but a clear south trend with this new gfs, nhc will come a little farther south, not much...everyone close to 95 in florida needs to prepare for a major

[SFLcane]

Sorry guys about my comment about the GFS.. I misread I guess that’s what happens when your on 4-5 hr sleep.

[KWT]

The big difference between the 06z GFS and the legacy is the impact the legacy has on the Bahamas compared with the 06z GFS operational, which misses alot of it to the north.

[tolakram]

GFS is being disregarded by the NHC so not worth using as it’s weakening the ridge to quickly.

Where do you get this information? Seriously, every time I come in here I hear the craziest stuff.

The GFS
is a northern outlier from the rest of the guidance, showing a
weaker ridge and bringing Dorian close to the Florida/Georgia
border, while the UKMET and ECMWF models show stronger ridges
and remain the southernmost solutions near South Florida. Given the
spread in the guidance, the new NHC forecast blends the previous
forecast with the simple multi-model consensus aids and the HCCA and
Florida State Superensemble aids, and it actually ends up in a
position very close to the tightly clustered global model ensemble
means. The most notable change in the new forecast is that it's a
little bit slower than the previous one as Dorian approaches
Florida.

They're slowing it down because now a recurve very near the coast is in the cards. The GFS is part of the multi-model consensus.

To be fair, the NHC doesn't say that either, they just mention the models are a little bit slower, you've added up own subjective view into that part of the NHC discussion

Anyway the GFS suite is also slower this run, along with the 06z ICON, so makes sense for them to slow it all down.

I worded it poorly but the comment stands. The curve to the north is very close to the coast and it's possible it happens before the coast. It's slower because the ridge is not as strong and as it breaks down the forward speed will slow. I'm not saying the NHC is slowing it down because they're unsure.

The latest GFS is well south, good for the track accuracy, bad for Florida, but there's still some hope the models have the ridge too strong or breaking down too late.

[lando]

KEY TAKE AWAYS FROM THE NIGHT RUNS

The GFS and euro are now down to landfalls about 120 miles apart. (Cape canaveral vs West Palm)

The avg margin of error 5 days out is 200 miles

The intensity forecasted by the NHC is conservative by their own statements, at a high CAT 3

The likely landfall is somewhere between northern Broward County and northern Brevard county.

The time to prepare is now if you live within this zone, and be prepared for a slow moving, extremely powerful cyclone capable of loss of life and serious injury

[Bocadude85]

Where do you get this information? Seriously, every time I come in here I hear the craziest stuff.

The GFS
is a northern outlier from the rest of the guidance, showing a
weaker ridge and bringing Dorian close to the Florida/Georgia
border, while the UKMET and ECMWF models show stronger ridges
and remain the southernmost solutions near South Florida. Given the
spread in the guidance, the new NHC forecast blends the previous
forecast with the simple multi-model consensus aids and the HCCA and
Florida State Superensemble aids, and it actually ends up in a
position very close to the tightly clustered global model ensemble
means. The most notable change in the new forecast is that it's a
little bit slower than the previous one as Dorian approaches
Florida.

They're slowing it down because now a recurve very near the coast is in the cards. The GFS is part of the multi-model consensus.

To be fair, the NHC doesn't say that either, they just mention the models are a little bit slower, you've added up own subjective view into that part of the NHC discussion

Anyway the GFS suite is also slower this run, along with the 06z ICON, so makes sense for them to slow it all down.

I worded it poorly but the comment stands. The curve to the north is very close to the coast and it's possible it happens before the coast. It's slower because the ridge is not as strong and as it breaks down the forward speed will slow. I'm not saying the NHC is slowing it down because they're unsure.

The latest GFS is well south, good for the track accuracy, bad for Florida, but there's still some hope the models have the ridge too strong or breaking down too late.

I’m getting flashbacks of Frances with the recent model trends, steering currents collapsed right before the coast and she slowly chugged inland while curving NW

[ronjon]

06z Legacy now landfalls near Melbourne and slides NW in Orlando as a 977 mb hurricane.

[AdamFirst]

Really dont like that the time table is moving back either. Giving it more time over water...yeesh

[Bocadude85]

I am curious why the GFS doesn’t show the wsw dip prior to landfall, almost every other global model has a wsw dip in the NW Bahamas.

[USTropics]

It's a matter of timing between the ECMWF and GFS model on how quickly the ridge begins to retrograde back east and to what extent it extends westward. You can see this even 72 hours from now. The GFS moves it out earlier, which allows Dorian to round the periphery of the ridge a bit quicker than the ECMWF:


Both models show a collapse in steering currents and a break in the ridge for this to turn north. There's a reason such a huge spread in the ECMWF ensembles exist from 12z to 00z, how quickly the ridge retreats and how far west Dorian gets will determine when Dorian moves north. You can see this in the ECMWF ensembles, here is an image of them just out to 120 hours:



Now add another 48 hours to those ensemble members, and the ones that have reached the GOM at 120 hours turn north towards panhandle, the ones off the Florida coast turn north to a recurve, and the one's over Florida move north up the spine:



The UKMET ensembles show a very similar evolution:

[gatorcane]

Looks like the GFS FV3 is still a Northern outlier with all of the other models into SE Florida and the legacy GFS close to SE Florida. What is really interesting is how the models are moving back the timing. The chance of a turn just before landfall seems to have increased some.

[ronjon]

Great analysis UStropics. That's the steering challenges in 4-5 days.

[tolakram]

I’m getting flashbacks of Frances with the recent model trends, steering currents collapsed right before the coast and she slowly chugged inland while curving NW

That's what I'm thinking as well, and in my opinion we've seen this play out before. Everyone focuses on the GFS outlier and how it tracks, expecting it to 'cave' to the euro. The euro keeps it's more accurate track past 72 hours but starts breaking down the ridge earlier and earlier so the storm ends up further north due to an earlier recurve. Neither has the correct location past 72 hours but past verification shows the euro to always be 40 to 50 miles closer to an accurate location than the GFS (I linked to the verification pdfs a few pages ago).