Model Diagnostic Discussion
NWS Weather Prediction Center College Park MD
227 PM EDT Sun May 12 2019

Valid May 12/1200 UTC thru May 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: 12Z GFS, 00Z ECMWF, 00Z CMC
Confidence: Slightly above average

---18Z UPDATE---
No change to the preliminary preference. The 12Z UKMET did
converge slightly to the other global models, but the preference
remains a blend of the GFS, ECMWF and CMC.

---PREVIOUS DISCUSSION---
Model mass field spread was generally pretty low across the
country. ECMWF ensemble spread was below the previous 30-day
average over large swaths of the CONUS. There were two exceptions
where model spread was a bit larger. First, the trough ejecting
through the Northeast by mid-week. The 12Z NAM and 00Z UKMET were
slightly more amplified and slower with the trough, which tended
to shift them slightly out of phase (to the west) relative to
other models and ensemble means upstream. Therefore, the NAM and
UKMET are not included in the preference. Second, spread was
larger with the amplitude of the trough approaching the West
Coast. The 00Z ECMWF and its ensemble members are more amplified
than most of the other deterministic models and NAEFS members. The
trend on the GFS over the past several model cycles has been
toward lower heights over much of the northeast Pacific, which
might support slightly more weight on the ECMWF. A compromise with
a lean toward the ECMWF is the preference at this point.


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

Lamers