Model Diagnostic Discussion
NWS Weather Prediction Center College Park MD
1252 PM EDT Fri Apr 12 2019

Valid Apr 12/1200 UTC thru Apr 16/0000 UTC

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

12Z Model Evaluation...with Preferences and Confidence
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

...Overall Pattern Across the CONUS...
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Preference: Non-CMC Blend, less weight on 12Z GFS
Confidence: Average

Most of the precipitation across the CONUS in the next few days
will be delivered by a developing cyclone that will move from the
Arklatex region Saturday, to the Northeast by Monday. Models are
in reasonably good agreement with the evolution and timing of this
cyclone, and with precipitation patterns across the rest of the
CONUS. Although there are some detail differences, a general model
blend seems to provide a good consensus forecast. There are some
exceptions worth noting, however.

The 00Z CMC is fastest with the progression of the primary cyclone
into the Northeast, despite not being considerably faster with the
progression of the trough aloft. From a conceptual model
standpoint, this does not seem particularly likely; a deepening
and eventually occluding low should generally be closer to the
trough axis with time rather than running well out ahead of it.
Therefore, the CMC is excluded from the preference. It has some
other mass field differences that deviate from the clustering of
the other deterministic models. For example, the upstream ridge
behind the Eastern U.S. trough is much more pronounced in the CMC,
and it eventually has a more amplified pattern in the West.

Other than the CMC, less weight was also placed on the 12Z GFS. It
shows a faster progression to the trough in the Eastern U.S. It is
joined somewhat by the 00Z UKMET, but both of the ensemble means
(ECMWF and GEFS) as well as the 12Z NAM and 00Z ECMWF are slower.
Given the steady deepening of the surface low and strength of the
trough (500mb heights in the lowest percent of the April
climatological distribution in Texas), the preference is to place
a greater weight on the ensemble means and slower NAM and ECMWF
with the expectation of a more dynamic system. The GFS also shows
some differences further west. For example, it is on the southern
end of model spread with the track of a surface low in the
Canadian Prairies on Sunday Night and Monday. However, the GFS
forecast is similar enough to the other models that it is
incorporated into the preferred blend, just with less weight than
the ECMWF, NAM and ensemble means.


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

Lamers