What the heck is going on with the global models re Chris?

[x-y-no]

Unbelievable ...

The GFS, NAM, ECMWF are still initializing Chris as an open wave ... I just don't get it.

The UKMET at least initializes a closed low, but only 1016mb.


As long as this massive error in initalization continues, I don't see that any of the models can be relied on, since they all run on some global model grid.

[Josephine96]

Maybe they think it'll go over Hispanola or Cuba and get destroyed

[deltadog03]

Models are absolutely crappy right now. Global models have never been very good on tropical wx but, this is down right crappy.

[x-y-no]

Maybe they think it'll go over Hispanola or Cuba and get destroyed

No, I'm talking about the initialization - the 0 hour frame.

Chris has been a 60mph TS with a 1003mb central pressure since last night - yet all the 0Z model runs initialized an open wave with the exception of the UKMET which initialized a 1016mb low.

That's incomprehensible to me. Guess I really don't understand how the model initialization really works (or rather doesn't).

[Derek Ortt]

50% of the initialization is the previous forecast

If the previous forecast is an open wave, chances are, the new initialization will be an open wave

[Eyewall]

i was wondering the same thing like 2 days ago...

makes no sense to me

[x-y-no]

50% of the initialization is the previous forecast

If the previous forecast is an open wave, chances are, the new initialization will be an open wave

OK ... but how long should that go on? This was a TD 36 hours ago, and a TS for 12 hours before last night's 0Z runs were started. It seems ridiculous to me that an error like this could persist so long.

There's some real problem in how they initialize models if this can happen, IMHO.

[bucman1]

How will this affect the projected track?

[DESTRUCTION5]

50% of the initialization is the previous forecast

If the previous forecast is an open wave, chances are, the new initialization will be an open wave

OK ... but how long should that go on? This was a TD 36 hours ago, and a TS for 12 hours before last night's 0Z runs were started. It seems ridiculous to me that an error like this could persist so long.

There's some real problem in how they initialize models if this can happen, IMHO.

I have to ask the same question...How the heck is this possible for none of these models to not be able to pick up on a near Cane..

[Derek Ortt]

because none of the previous forecasts has Chris, and Chris has not passed over any data bouys. FL data is not assimilated into the models

[DESTRUCTION5]

because none of the previous forecasts has Chris, and Chris has not passed over any data bouys. FL data is not assimilated into the models

Ahh makes sense but Im thinking it needs to be revamped for that reason alone...

[x-y-no]

because none of the previous forecasts has Chris, and Chris has not passed over any data bouys. FL data is not assimilated into the models

Given the manifestly nonsensical result, maybe they ought to reconsider this policy.

[Derek Ortt]

the model would almost certainly reject the FL data since it is only at one level. Would probably be filtered out as noise

Now, dropsondes would be a different story

[wxwonder12]

At what point will the models have better initialization and have a better grasp on forecast path??

[SouthFloridawx]

Well looking at the gfs and nogaps it does look like they take it more northerly.

[x-y-no]

the model would almost certainly reject the FL data since it is only at one level. Would probably be filtered out as noise

Now, dropsondes would be a different story

So maybe that would be the way to go. I'm not arguing for any given methodology, since I don't have the knowledge to do so. I'm just saying there's a big problem if the globals are still initializing an open wave when we've had a cyclone for over a full day.

[AJC3]

the model would almost certainly reject the FL data since it is only at one level. Would probably be filtered out as noise

Now, dropsondes would be a different story

So maybe that would be the way to go. I'm not arguing for any given methodology, since I don't have the knowledge to do so. I'm just saying there's a big problem if the globals are still initializing an open wave when we've had a cyclone for over a full day.

There's also the concept of "vortex bogussing", where a symmetric vortex representing the TC is is actually added to the model init. I'm, not sure if this is or has ever been done operationally with any of the globals. I seem to recall that it has, but I could be wrong. If you have time google the phrase and see what you come up with.

[curtadams]

Well, it passed over a buoy during its formation and they should have wind readings from the islands showing Nly winds. There's quickscat too, which showed a clearly closed low as of yesterday morning my time at the latest. How hard do they have to get whacked over the head? It really seems like something's wrong that they still see nothing. FWIW, most seem to think it will get somewhat stronger over the next few days so IMO they're doing better with its evolution, at least, than they did prior to TD status.

[MWatkins]

The GFS is at it again...poor initialization without a well defined circulation at 500MB:

http://www.nco.ncep.noaa.gov/pmb/nwprod ... 0_000m.gif

Watch the upper features for the track because once again the model depiction of it will be worthless.

MW

[vacanechaser]

lets also remember the gfs could not see a major hurricane, fabian headed for bermuda for several days... well, maybe it saw it, but certainly not a major hurricane in 2004


Jesse V. Bass III
http://www.vastormphoto.com
Hurricane Intercept Research Team