Prognostic Meteorological Discussion
Issued by NWS

Home | Current Version | Previous Version | Text Only | Print | Product List | Glossary On
Versions: 1 2 3 4
FXUS10 KWNH 241844

Model Diagnostic Discussion
NWS Weather Prediction Center College Park MD
244 PM EDT Sat Mar 24 2018

Valid Mar 24/1200 UTC thru Mar 28/0000 UTC

...See NOUS42 KWNO (ADMNFD) for the status of the upper air

12Z Model Evaluation...Including Model Preferences and Confidence

...Pair of shortwaves moving into the Intermountain West this
weekend, emanating from a devolving closed low...
...Further amplification of the Western U.S. trough via a wave
digging sharply from the Aleutians, down the West Coast and into
the base of the trough...
...Lee cyclogenesis over the Plains Sunday and Monday...
Preference: 12Z ECMWF, 12Z UKMET, 12Z CMC
Confidence: Average

---18Z UPDATE---
No change to the preliminary preference. The ECMWF, CMC, and UKMET
continue to be slower with the trough than the GFS, although they
have notably sped up slightly from their previous positions. This
makes the 00Z ECMWF Ensemble Mean the slowest at this time, and
given the trends on the other global models, that may speed up on
its 12Z run.

Models continue to show limited differences through Sunday Night,
and diverging forecasts after that. In general, the same model
spread that existed yesterday continues to show up today with the
12Z GFS developing a broader trough over the West and spilling
height falls further to the east over the Plains than other
models. At 500mb, there are still 60m height differences over the
Southern Plains between the GFS and ECMWF. The 12Z NAM shows a
slightly sharper trough than the GFS, but still has lower heights
over the Plains (like the GFS), and amplifies the downstream ridge
more than any other model over the Eastern CONUS.

On the other end of the model spread, the 00Z operational ECMWF
and ECMWF ensemble mean are the slowest, and hold the base of the
trough back to the west the longest. The 00Z CMC and UKMET are
closer to the ECMWF than the GFS, but do offer an intermediate
solution. With an amplifying pattern and a shortwave digging
sharply into the base of a positively tilted trough, the
expectation would be for the trough to slow down -- a scenario
better depicted by the ECMWF, CMC, and UKMET. This also represents
continuity from the past several WPC model preferences, and is
close to the consistent ECMWF ensemble mean. Therefore, the
preference will be to lean toward a blend of those three global

By 27/12Z on the 00Z model cycle, the GEFS members and ECMWF
ensemble members 558-570 DM isoheights showed almost no overlap in
the Southern Plains. In other words, the slowest GEFS member was
just about faster than the fastest ECMWF ensemble member.

...Shortwave traversing the Midwest now and becoming sheared out
through the Ohio Valley and Carolinas...
...Weakening associated surface low transferring to the Carolinas
coastline by Sunday morning...
Preference: Non-CMC Blend
Confidence: Above Average

The 12Z CMC deamplifies the shortwave much faster than the other
models, which seems to occur because its initial representation of
the shortwave was weaker. This leads to a flatter wave, and a
slightly weaker surface low that ends up slower and further south
than model consensus. Otherwise, models are in very good
agreement. The preference is for a broad consensus with this
system with the exception of the CMC.

...Arctic shortwave digging sharply south through Quebec on
Saturday with closed low developing off the Northeast Sunday and
...More accelerated development of the Carolinas coastal low well
offshore on Monday and Tuesday...
...Associated strong surface high settling into New England...
Preference: 12Z NAM, 12Z UKMET, 12Z ECMWF
Confidence: Average

---18Z UPDATE---
The NAM, UKMET, and ECMWF continue to be most tightly clustered
with the surface low over the Atlantic, and maintain reasonable
continuity with their previous runs. Therefore, no change to the
preliminary preference.

Within the broader trough situated over the Northeast and eastern
Canada, an embedded shortwave is situated over the North Atlantic
around 40N/60W today. Most models are in good agreement with this
shortwave, except the 00Z CMC which barely indicates a shortwave,
with a much flatter height pattern aloft relative to the other
models. This initial difference, along with the faster
deamplification of the Ohio Valley shortwave (discussed above),
affect the eventual forecast over the Northwest Atlantic, and as a
result the CMC is not included in the preference.

The 12Z NAM, as well as the 00Z UKMET and ECMWF, seems to be the
most consistent with the position and intensity of the eventual
upper level low. They all have very similar timing of the digging
wave, while the 12Z GFS is still a little faster and kicks the
trough and low a little further east. The preference is to remain
closest to the more consistent models and this is also consistent
with previous WPC model preferences.

...Shortwave rounding a Pacific ridge through the Gulf of Alaska
on Monday and digging into the Northern Rockies Tuesday...
...Surface low developing over Alberta and Saskatchewan Tuesday...
Preference: 12Z NAM, 00Z UKMET, 00Z ECMWF
Confidence: Slightly above average

---18Z UPDATE---
No change to the preliminary preference. The GFS continues to be
well south of all the other global models with the surface low
position in southern Canada.

The models are in fairly good agreement overall with this system,
but there are subtle differences. This may be due to the variation
in how they handle the (much larger scale) trough that will be
gradually moving from the West out into the Plains. Most notably,
the 12Z GFS is further south with the surface low by the end of
the diagnostic period than any other model (even quite a bit
further south as compared to the 06Z GEFS Mean). The 00Z CMC is
also further south, but still slightly north of the GFS.
Meanwhile, the 12Z NAM and 00Z ECMWF and UKMET are clustered
fairly close to the ECMWF Ensemble Mean. The preference is to use
a blend of those models, given the lean toward the ECMWF Ensemble
Mean with the larger scale trough.

Model trends at
500 mb forecasts at


$$ is the U.S. government's official web portal to all federal, state and local government web resources and services.