06z Model Disc. 12/7/10

[Ellinwood]

They be coming in.

[Ji]

Looks like stronger -nao is supressing Storm track

[Dino]

Looks like a 100 mile shift SOUTHEAST.

The QPF is lower and more evenly split on this run, but I'd rather have a dusting then see it rain; so much cold air around it would be pitiful.

[A-L-E-X]

Just saw a post stating that the verification scores of all the model runs of the GFS are very similar, since there is new data assimilation on all the model runs now-- so we cant short change this development just because its the 6z GFS.

[SnowGoose69]

Looks like stronger -nao is supressing Storm track

This track is more believable than the 00Z run given it runs up the Hudson Valley and not the mountains....and its extremely close on the NAO and 50/50 low...the NAO remains a bit too east based and hence the 50/50 low a bit too far north and north-east.

[isnice]

It's probably just a blip.

But believe it or not, ALL of the ensembles fall in line with the 06z GFS.

[Stebo]

This track is more believable than the 00Z run given it runs up the Hudson Valley and not the mountains....and its extremely close on the NAO and 50/50 low...the NAO remains a bit too east based and hence the 50/50 low a bit too far north and north-east.

Which would tend to favor the track being on the otherside of the mountains, plus the bias of the GFS diving energy too far south is certainly in play here, as no other model except maybe the Ukie 00z is close on how far south this energy dives in. Plus you throw in the projected MJO movement from the 4 to 5 and eventually 6 zone.

[JustinBerk]

Still remember the southern bias with most storms last year. Even this weekend's clipper ended up north of forecast track. There does seem to be a faster arrival of system in Baltimore... so I would expect an overnight Sat into Sunday morning start as mix or ice... then rain ending with some snow with brief wrap around Sunday night.

[MJO812]

The GFS still has a cold bias? I thought they upgraded the model.

[Stebo]

The GFS still has a cold bias? I thought they upgraded the model.

Yes it does, DT has mentioned it many times.

[SnowGoose69]

Yes it does, DT has mentioned it many times.

Every model has a cold bias in the winter it seems when it comes to long range cold blasts....the Euro often overdoes 850 temps at day 7 or 8 in signiificant arctic outbreaks as well.

[Stebo]

Every model has a cold bias in the winter it seems when it comes to long range cold blasts....the Euro often overdoes 850 temps at day 7 or 8 in signiificant arctic outbreaks as well.

True, its all in the extent of how strong the bias is, the GFS's is clearly the strongest bias at this time range.

[atownwxwatcher]

The GFS still has a cold bias? I thought they upgraded the model.

I thought i remembered back on eastern that the GFS was verifying to warm after it was upgraded and that one of the things it improved on was the cold bias?

[atownwxwatcher]

I thought i remembered back on eastern that the GFS was verifying to warm after it was upgraded and that one of the things it improved on was the cold bias?

GFS

The major upgrade to the GFS in August 2010 introduced three biases:

-2 m temp warm bias

-cold bias in the stratosphere

-low wind bias

Corrective changes are expected to be implemented in late February 2011

Source

[Dr No]

What an amazing difference between the GFS 6Z run and the 0Z Euro.

GFS is about 50 miles away from a major storm here in the DC area... we want the H5 LP to track from near Atlanta to anywhere SE of DCA.

Notice the 700 MB LP moves from about Staunton VA to the ENE from there.

Of course, we all know it will be well to the west at 12Z.

[Alpha5]

What an amazing difference between the GFS 6Z run and the 0Z Euro.

GFS is about 50 miles away from a major storm here in the DC area... we want the H5 LP to track from near Atlanta to anywhere SE of DCA.

Notice the 700 MB LP moves from about Staunton VA to the ENE from there.

Of course, we all know it will be well to the west at 12Z.

Euro has no support from any other model, even its ensembles are much further east. I've heard it struggles in nina patterns....

[okie333]

GFS

The major upgrade to the GFS in August 2010 introduced three biases:

-2 m temp warm bias

-cold bias in the stratosphere

-low wind bias

Corrective changes are expected to be implemented in late February 2011

Source

I believe there's a pretty strong dry bias too... surprised that NOAA didn't mention it

[psuhoffman]

I believe there's a pretty strong dry bias too... surprised that NOAA didn't mention it

not sure it has a dry bias now...before it had a wet bias so maybe it just corrected that.

[psuhoffman]

Which would tend to favor the track being on the otherside of the mountains, plus the bias of the GFS diving energy too far south is certainly in play here, as no other model except maybe the Ukie 00z is close on how far south this energy dives in. Plus you throw in the projected MJO movement from the 4 to 5 and eventually 6 zone.

You are assuming of course, that those other factors are modeled correctly. Your logic is good but some of your posts recently smack of a kind of concrete thinking towards this situation that I have learned to be dangerous. While it is probable that the track is incorrect as shown, it is also likely that some of the factors that will determine the eventual track are also not yet being shown accurately by the models. As these factors come into better focus the track will shift and the exact setup may end up completely different then you are assuming it to be at this time. Sometimes the pattern is very stable with large scale h5 features locked in and the basic synoptic setup can be determined at long lead times. Some of the storms last year fell into that category with the basic outcome becoming clear several days in advance. Other times, like now, things are more fluid and being stuck on a solution this far in advance is a bad idea IMHO.

[okie333]

not sure it has a dry bias now...before it had a wet bias so maybe it just corrected that.

I thought I read on Eastern somewhere that the model way over-corrected the wet bias, and there's a dry bias of 20% or so (that is, the averaged-out QPFs are only 80% of verification). Could be wrong though.