Model Diagnostic Discussion
NWS Weather Prediction Center College Park MD
303 PM EDT Fri Apr 05 2019

Valid Apr 05/1200 UTC thru Apr 09/0000 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: Blend of 12Z GFS, NAM, ECMWF and UKMET
            Less weight on NAM and UKMET beyond 08.00Z
Confidence: Average

---18Z UPDATE---
No change to the preliminary preference with the new ECMWF, UKMET
and CMC runs. Reasoning remains similar.

---PREVIOUS DISCUSSION---
Models show a continuation of a broadly zonal flow pattern over
the next several days, with a lack of complete phasing of the
northern and southern streams. Model spread appears fairly limited
over the first couple days of the forecast, with more significant
divergence beginning Sunday afternoon and night. In general, a
broad-based model blend should be acceptable through 07.12Z or
even 08.00Z depending on the region.

Analysis of model spread through those times shows the most
significant variability to exist with the amplitude of a southern
stream trough in the vicinity of Texas, and with the trough in the
Pacific Northwest. These differences grow by the Day 3 time frame,
and begin to have more notable impacts on the forecast.

For the entire forecast period, the 00Z CMC is the model that
shows the greatest differences from the average of the other
deterministic models. By Day 3 this leads to considerable
variability in the placement of QPF maxima across the country.
Therefore, the CMC is excluded from the overall model preference
at this time. Beyond 08.00Z, the NAM's height fields begin to
depart considerably from model consensus as well, and therefore it
is excluded beyond that time.

The 00Z UKMET does offer a reasonable forecast over the weekend,
but by Monday it has a few important differences. With the
southern stream trough, it develops a much stronger surface low
than any of the other deterministic models and all GEFS+ECMWF
ensemble members. Positionally, it is also further to the south.
With a northern stream shortwave pushing toward the Northeast, the
UKMET is more progressive than all the other models. Therefore,
the UKMET is excluded from the preference beyond Sunday evening.

The GFS and ECMWF, meanwhile, are fairly similar in most aspects
of their forecasts and are also fairly close to the ensemble
means. Over time, the preferred blend is increasingly weighted
toward those two models. However, there are still some differences
in the placement of QPF maxima, particularly on Day 3. WPC QPF
opts to generally maintain the QPF maxima, but find a compromise
with the location.


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

Lamers