Let me offer the following reason why NWS's GFS model has been so erroneous
this week.

It's concept of soil moisture is bad, and especially so now over the
Southwest quadrant of the U.S. where it has been telling itself all along,
that it is dry.

It is not nearly so dry there as it assumes.

Not then, until it brought fresh moisture into the Southwest this weekend,
could it come to think that it would be more moist there for a while, and
them begin to find better solutions for subsequent days.

If you wish to see how dry the GFS thinks it is, look at;
http://grads.iges.org/pix/soil1.html

Garbage in/garbage out.

David Salmon,

dba. Weather Derivatives

Purveyor of accurate soil moisture representations.

On Fri, 18 Mar 2005 06:13:31 -0600,
David Salmon , in
wrote:

+ Let me offer the following reason why NWS's GFS model has been so erroneous
+ this week.

Care to quantify this statement?

James
--
Consulting Minister for Consultants, DNRC
I can please only one person per day. Today is not your day. Tomorrow
isn't looking good, either.
I am BOFH. Resistance is futile. Your network will be assimilated.

I should quantify the fact that GFS has performed so poorly, or the "why"?

That it was so wrong; earlier in the week GFS was trying to warm up big
chunks of the U.S., especially in the middle and now it says it will be cool
to cold instead. With the warmth was to come increased precipitation, though
now it is to be dry--total bust.

I can't unfortunately, send a more accurated representation of the soil
moisture distribution across the U.S. to this newgroup because attachement
are banned. However, picture this; it is just as moist in West Texas as it
is in the western Midwest right now (compare that to what GFS thinks) and
big pieces of the interior West are nearly as moist too--GFS has them dry.

David Salmon wrote:
I should quantify the fact that GFS has performed so poorly, or the "why"?

That it was so wrong; earlier in the week GFS was trying to warm up big
chunks of the U.S., especially in the middle and now it says it will be cool
to cold instead. With the warmth was to come increased precipitation, though
now it is to be dry--total bust.

I can't unfortunately, send a more accurated representation of the soil
moisture distribution across the U.S. to this newgroup because attachement
are banned. However, picture this; it is just as moist in West Texas as it
is in the western Midwest right now (compare that to what GFS thinks) and
big pieces of the interior West are nearly as moist too--GFS has them dry.


....and how does that error in moisture representation translate
into bad forecasts, versus, say, poor representation of upstream
shortwaves in fast flow off the Pacific?

I don't recall any GFS forecasts that suggested warmth in the midwest,
'tho I confess to not seeing all the model runs. I do know that
the instant occlusion that's tracking across the midwest now was
fairly well forecast earlier this week, and the persistence of
below-normal temperatures in the midwest has also been well
represented in GFS forecasts for most of this month.

Scott

On Fri, 18 Mar 2005 15:09:04 -0600,
David Salmon , in
wrote:

+ I should quantify the fact that GFS has performed so poorly

Yes.

+ , or the "why"?

You're trying to convince us of the why.

Truthfully, tho, I'm not surprised that a model isn't handling soil
moisture quite right. Since I know next to nothing about the GFS, I'll
have to go look it up...

http://meted.ucar.edu/nwp/pcu1/ic4/f...navigate(3,7,1)

Under the soil moisture category, in the "Correction to forecast", we see:

Model has too little moisture

* Decrease surface temperature

* Increase evaporation and decrease sensible heat flux

* Increase surface RH

* Decrease near-surface winds, gustiness, and change direction
based on vertical directional and velocity wind shear in the lower
troposphere

* Decrease PBL depth

* Increase the probability of convective precipitation in summer,
subject to assessing instability based on modified skew-T and CAPE

Seems to agree with what you're saying.

You can read more on the models in general:
http://meted.ucar.edu/nwp/pcu2/index.htm

It's not obvious, tho, where they get their soil moisture *from*.

+ That it was so wrong; earlier in the week GFS was trying to warm up big
+ chunks of the U.S., especially in the middle and now it says it will be cool
+ to cold instead. With the warmth was to come increased precipitation, though
+ now it is to be dry--total bust.

Interesting, tho. Looks like there's a fairly nice storm in the middle
part of the USA when I looked at the weather earlier today (18 Mar
2005).

My uncle in Wisconsin is going to gripe that it's still freakin'
winter and he may get a shot of snow. Again. What a shock, it's still
winter in Wisconsin, even if it's "spring" in the Northern Hemisphere...

James
--
Consulting Minister for Consultants, DNRC
I can please only one person per day. Today is not your day. Tomorrow
isn't looking good, either.
I am BOFH. Resistance is futile. Your network will be assimilated.

"shortwaves" are only to a product of their environment. Indeed, if poorly
initialized out in the Pacific, they can't possibly be expected to perform
as modeled once they are inland. Likewise, if the shortwaves are forced over
some landmass thought to be dry (and therefore welcome to warming) they
won't perform as expected once they past downstream from the area that was
indeed moist by absolute measure, and relatively cooler than thought in
earlier stabs. Sooner or later, all characteristics within the atmosphere
are derived from the surface from whence the air came.

I R A Darth Aggie wrote:
My uncle in Wisconsin is going to gripe that it's still freakin'
winter and he may get a shot of snow. Again. What a shock, it's still
winter in Wisconsin, even if it's "spring" in the Northern Hemisphere...

James



I'm not your uncle, but I'll gripe about the weather here!

Someone shut off this snow! If it keeps up we'll get up
to normal for the season!

Ah well, at least I don't live 50 miles north where they'll
get *pounded* overnight.

scott

On Fri, 18 Mar 2005 16:30:29 -0600,
Scott , in
wrote:
+ I R A Darth Aggie wrote:
+ My uncle in Wisconsin is going to gripe that it's still freakin'
+ winter and he may get a shot of snow. Again. What a shock, it's still
+ winter in Wisconsin, even if it's "spring" in the Northern Hemisphere...

+ I'm not your uncle, but I'll gripe about the weather here!

+ Ah well, at least I don't live 50 miles north where they'll
+ get *pounded* overnight.

He lives in the Two Rivers area, about 100 miles north of Madison. But
if the low tracks to his south, like I'm thinking it will, he'll get a
significant on-shore wind for quite a while. Can you say "lake
effect"?

:-D

James
--
Consulting Minister for Consultants, DNRC
I can please only one person per day. Today is not your day. Tomorrow
isn't looking good, either.
I am BOFH. Resistance is futile. Your network will be assimilated.

The soil model for the GFS lives in the GFS--it's part of their model. It
just doesn't work......I think in large part because they assume much too
capacity for water in their "soil". The huge capacity then yeilds too little
relative recharge in the West and way, way, way too much total moisture
available long after dryness has set in, over the Central and East U.S.

And, there is no real data going into their soil model either. The input for
the soil model is model precipitation and temperatures. That would account
for a little drift over time, even if the soil capacity were correct and the
model made more accurate assumptions. They told me (back several years ago),
they have a very sophisticated model and were adverse to tinkering with it.
Too bad and so sad. More than 5 years ago I offered to give them a soil
model that WORKS. They didn't think they need any help from the likes of me.

Of course, if they were to put real data into the model's soil model, NWS
(along with the FAA) would have to make real-time precipitation data a
priority, and that apparently wasn't going to happen either. " 7/// " is and
acceptable precipitation report.

"They" are however, now going to implement a multi-stage soil model later
this spring, and perhaps they may improve their more immediate results on
feedback to the atmospher a little bit with that. I'll still contend though,
that so long as the total model is a stinker, the layers will remain
stinkers as well.