Model Diagnostic Discussion
NWS Weather Prediction Center College Park MD
336 AM EDT Sun Apr 28 2019

Valid Apr 28/0000 UTC thru May 01/1200 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: General model blend through Day 1 
            UKMET/CMC blend Day 2/3 across SW/Central Plains
Confidence: Above average except on Day 3: Average

07z update: The 00z ECMWF/UKMET/CMC showed more agreement with the
GFS/NAM on Day 3 with the shortwave ejecting out into the central
Plains, with the ECMWF trending faster. This is giving higher
confidence and less model spread. A UKMET/CMC compromise blend for
mass fields still seems to be a good approach but if trends hold,
then a general model blend would be a good approach.

---previous discussion---
The two systems of concern across the CONUS through the next 3
days are one across the northern US associated with a strong
shortwave trough while another system entering the Four Corners
region will eject into the central Plains by Day 3. In general,
the upper level pattern will transition from fast zonal flow to
broad troughing over the western US and ridging over the southeast
US by the end of the forecast period.

For the northern stream shortwave on day 1/2, there has been
better model agreement in the last couple model cycles in terms of
the mass fields. In general, placement of the major features
spatially and temporally are in good agreement and much of the
spread in the models is fairly minor. As such, a general model
blend is preferred with above average confidence.

A closed low over southern California on Monday is expected to
gradually open up and eject into the central Plains by Tuesday and
Tuesday  night. Here, model differences are bit more significant
with temporal and spatial differences as well as QPF convective
issues to contend with. The GFS and NAM remain the faster
solutions, opening up the wave early and allowing it to progress
into the central Plains a good 12-18 hours ahead of the rest of
the guidance. The ECMWF meanwhile is on the slower end of the
model envelope, but has been trending faster toward the GFS/NAM
solutions the last 2 runs. In the middle lie the CMC/UKMET
solutions, which appear to be a fairly good compromise solution in
terms of the mass fields. Given the typical biases seen in the
deterministic models, and looking at the ensembles, a preference
for something between the fast GFS/NAM and slow ECMWF is preferred
(similar to CMC/UKMET mass fields). QPF differences however are a
bit more complicated due to surface boundary differences and
convective feedback issues, where is a large spread spatially in
the highest amounts. As such would favor the mass fields to
utilize the timing of the CMC/UKMET but the QPF/mass fields of the
GFS/ECMWF are usable but have to be shifted toward a more central
location near the CMC/UKMET axis.

All in all, while a general model blend can be used in the first
36-48 hours, particularly across the northern stream energy,
forecast confidence drops to average or slightly below average for
the central Plains systems on Day 3. Mass field blend toward the
CMC/UKMET will be probably help wash out the fast GFS, slow ECMWF
and be a good compromise at this point.

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

Taylor