Model Diagnostic Discussion
NWS Weather Prediction Center College Park MD
238 AM EDT Fri Apr 19 2019

Valid Apr 19/0000 UTC thru Apr 22/1200 UTC

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

00Z Model Evaluation...with Preferences and Confidence
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

...Overall Pattern Across the CONUS...
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Preference: Blend of the 00Z GFS and 12Z ECMWF, weighted
            differently for several synoptic systems
Confidence: Average

---06Z UPDATE---
No change to the preliminary preference. The 00Z ECMWF (and CMC)
did trend further north with the position of the closed low in the
Eastern U.S., in line with the preliminary preference that leaned
toward the GFS for that system. Out West, the ECMWF remained
consistent with the ensemble means.

---PREVIOUS DISCUSSION---
The model preference is for a blend of the GFS and ECMWF. These
model forecasts are reasonably similar, and do not exhibit the
larger deviations from ensemble means that are seen with some of
the other deterministic models. However, there are still important
differences with two of the primary synoptic systems.

First, with the closed low that will be affecting the Eastern
U.S., greater weight is placed on the 00Z GFS. The GFS is in
greater agreement with ensemble means (both 18Z GEFS and 12Z ECMWF
ENS) regarding the latitudinal position of the closed low. It is
further north, while the ECMWF is further south, joined only by
the 12Z CMC. This does have practical effects on the position of a
QPF maximum in the Great Lakes to the northwest of the low.

Second, with the trough moving into the Western U.S., greater
weight is placed on the 12Z ECMWF. In this case, the GFS and UKMET
seemed to be far too flat with the mid-upper level height pattern
given the expectation of a strong trough digging into the region.
The 00Z NAM, however, seemed too amplified with the trough, which
is a typical bias. In this case, the 12Z ECMWF seemed most
reasonable, supported to a certain extent by the 12Z CMC, 12Z
ECMWF Ensemble, and even the 18Z GEFS Mean (closer to the ECMWF
than the GFS).

The most significant deviations from model consensus were with the
12Z UKMET. By 22.12Z it was nearly outside of the envelope of
ensemble spread for mid-upper level height patterns in some places
in the Southwest (with the primary trough), and in the Northern
Plains (with a secondary trough). Although the NAM and CMC did
differ from the GFS and ECMWF in some important respects, they may
be reasonable to incorporate in a model blend, particularly on Day
1 or 2 (prior to 21.12Z) when they remained fairly close to
ensemble means.


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

Lamers