Model Diagnostic Discussion
NWS Weather Prediction Center College Park MD
140 AM EST Tue Feb 12 2019

Valid Feb 12/0000 UTC thru Feb 15/1200 UTC

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

00Z Model Evaluation...with Final Preferences and Confidences
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

...Overall Pattern Across the CONUS...
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Preference: Blend of 00Z ECMWF, GFS; 12Z ECMWF ENS Mean
Confidence: Average

---06Z UPDATE---
No change to the preliminary preference. The 00Z UKMET has shifted
to be more in line with the ECMWF and GFS with respect to the West
Coast trough, and could also be useful in a model blend.

---PREVIOUS DISCUSSION---
The most substantial model spread over the past several model
forecast cycles has been in relation to the trough along the West
Coast. This continues to be the location of the largest spread in
the ensembles. However, the 00Z GFS has arrived with a forecast
that is much more similar to the 12Z deterministic ECMWF. This
leads to a much stronger low closing off over the eastern Pacific
(around 135W) and a slower approach to the coast. The 12Z ECMWF
ensemble mean was slightly faster, but still relatively close to
its operational forecast, with the 00Z NAM and 12Z UKMET slightly
faster than that, and the 12Z GFS and 12Z CMC considerably faster.
Given the notable trend in the GFS, and reasonably strong support
from the ECMWF ensemble members, the preference is to lean toward
the forecast cluster represented by the new GFS run and the 12Z
ECMWF run. The NAM and UKMET are slightly faster, but not far off,
and represent reasonable scenarios. The 12Z CMC, by contrast,
appears far too fast.

Elsewhere, there is fairly good model agreement with respect to
the mass field forecasts. The notable exception would be
downstream effects from the CMC being much faster with the West
Coast trough than other models, leading to more substantial
cyclogenesis over the Plains by late Thursday. Otherwise, the
models are fairly similar with the strong low dropping down toward
the Pacific Northwest, and the cyclone affecting the Great Lakes
and Northeast. A general model blend could conceivably be used in
these other areas, but greatest weight was still placed on the
ECMWF, GFS and ECMWF Ensemble Mean.


Model trends at www.wpc.ncep.noaa.gov/html/model2.shtml
500 mb forecasts at www.wpc.ncep.noaa.gov/h5pref/h5pref.shtml

Lamers