Print Send to Printer

The Weather Prediction Center

Model Diagnostic Discussion

[Abbreviations and acronyms used in this product]
Geographic boundaries:    Map 1- [Color] [B/W Print Version]      Map 2 - [Color] [B/W Print Version]

Model Diagnostic Discussion
NWS Weather Prediction Center College Park MD
157 PM EST Mon Feb 19 2018

Valid Feb 19/1200 UTC thru Feb 23/0000 UTC

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

...12Z Model Evaluation with Final Preferences/Confidence...

...Amplified Western US Trough with Primary Shortwave Lifting
Northeast from the Four Corners Today to Ontario by Tuesday
...Associated Weak Surface Low Tracking into the Great Lakes...
...Trailing, Slow-moving Front Stretching from the Great Lakes to
the Southern Plains through Wednesday...
Preference: Non-NAM, Non-CMC Blend
Confidence: Slightly Above Average

---18Z UPDATE---
The 12Z CMC has been removed from the model preference. Its
heights aloft are much lower than other models over the
South-Central US, while still having one of the strongest upper
level ridges on the east coast. The remaining global models' 12Z
runs were relatively consistent with their previous runs.

From a synoptic scale, the models are in very good agreement on
the big picture, which is to be expected with a relatively stable
1-wave pattern over the CONUS -- an amplified trough in the West
and an amplified ridge in the East. There are still some small
detail differences on the mesoscale, or with timing of individual
shortwaves embedded in the broader flow pattern, that could make
an impact on the forecast.

The 12Z NAM kicks the primary shortwave out of the base of the
trough a little earlier than the other models, and thus is about
six hours faster with the progression of the wave and the
associated surface low. With a lack of support for such a scenario
from other models or ensemble members, the preference is to
exclude the NAM at this time. The 12Z GFS, meanwhile, continues to
be on the slower end of the model spread, which was also the case
yesterday. However, the difference between the GFS and the other
global models has narrowed sufficiently, that it is a reasonable
possibility. The 06Z GEFS was slightly faster with the forward
progression of the surface low, indicating that the GFS may
converge further to the other models over the next couple cycles.

Otherwise, the preference is for a broad-based blend to account
for mesoscale uncertainties. Along the southern extent of the
front in particular, convection may be the primary driver of
certain things, such as how quickly the front advances to the
southeast. A broader blend would account for the possibility of
small mass field variations having an effect on the sensible
weather impacts.

...Secondary Shortwave Dropping South Along West Coast of North
America through Wednesday, Reinforcing the Larger Trough over the
Western CONUS...
...Shortwave Kicking into the Four Corners on Thursday with Front
Possibly Returning North Across the South-Central US...
Preference: General Model Blend
Confidence: Average

---18Z UPDATE---
No change to the preliminary preference. 12Z models remained
relatively consistent with their previous model runs.

Models have come into better agreement with the shortwave digging
along the West Coast on Tuesday and Wednesday. The GFS had
previously been much more amplified with the shortwave trough
aloft, and had a much stronger surface low. While the other models
have trended slightly more amplified than they were a couple model
cycles ago, the GFS has trended substantially less amplified with
the trough aloft. The resulting model mass fields are relatively
similar. The primary difference appears to be timing, with the 12Z
GFS and NAM slightly slower than the 00Z ECMWF, CMC and UKMET. The
GFS and ECMWF both have support from their respective ensemble
means. The differences through the short term period would be
relatively subtle, but a faster solution (closer to the ECMWF)
would produce height falls over the Southern Plains sooner, with a
corresponding earlier low-level mass field response. Thus, the
return flow and surface front may begin to push north from the
northwestern Gulf a little sooner. That being said, there is not a
clear preference between the small timing difference, so a general
model blend is preferred with this system.

...Third Shortwave Also Digging South Along West Coast Wednesday
Night and Thursday and Further Reinforcing the Trough...
Preference: Blend of 12Z GEFS Mean, ECMWF
Confidence: Slightly Below Average

---18Z UPDATE---
Substantial differences remain on the 12Z model cycle for this
system, although the GFS does still seem to be the only model that
digs the shortwave so far south. The 12Z CMC also maintained a
more concentrated vorticity maximum relative to the other models,
despite not closing off a low any more. The preference will remain
with the GEFS Mean and ECMWF as they remained relatively

More significant differences emerge with the mid-week shortwave
digging into the West Coast. The 12Z GFS digs the primary
shortwave into southern California and has the primary mid-level
vorticity maximum in that area by 23/00Z. The other models have it
further north, and this is supported by the vast majority of the
GEFS/ECMWF Ensemble clusters. The 06Z GEFS mean has lower heights
near WA/OR, with higher heights in the Desert Southwest, so the
GEFS mean seems like a reasonable alternative to the GFS. The 00Z
CMC closes off an upper level low very quickly and has a much more
concentrated vorticity maximum relative to the other models (which
have an elongated/sheared appearance along the base of the
trough). These differences in the GFS and CMC seem substantial
enough to not be included in the preference at this time.
Therefore, will lean more in the direction of a blend of the

...Large Mid-Upper Level Anticyclone just off the Southeast Coast
with East Coast Ridge Axis...
Preference: Blend of 00Z ECMWF, ECMWF ENS Mean
Confidence: Average

---18Z UPDATE---
No change to the preliminary preference. 12Z models remained
relatively consistent with their previous model runs.

The 00Z CMC and UKMET have one of the stronger depictions of the
upper level ridge in the first 24 hours of the forecast, and this
eventually grows over time by mid-week. The 00Z UKMET eventually
has higher heights aloft from the Tennessee Valley into the Upper
Midwest, and has the biggest deviation from the model average.
Meanwhile, the 12Z GFS is much flatter with its depiction of the
Ridge by Wednesday and Thursday, and has lower heights over much
of the Northeast. The preference is for a depiction in the middle
of these extremes, and closest to a blend of the 00Z ECMWF and 00Z
ECMWF Ensemble Mean.

Model trends at
500 mb forecasts at


Last Updated: 157 PM EST Mon Feb 19 2018