Model Diagnostic Discussion
NWS Weather Prediction Center College Park MD
101 PM EDT Tue Aug 27 2019

Valid Aug 27/1200 UTC thru Aug 31/0000 UTC

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

12Z Preliminary Model Evaluation with Preferences and Forecast
Confidence
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

...Overall Pattern Across the CONUS...
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Preference: General model blend
TD 06L NHC Proxy: 00z UKMET/00Z GEFS blend
Confidence: Slightly above average

For TD6 moving northeast off the eastern coast of the CONUS
through the period, the NHC track is a bit slower than the
majority of the available global suite. However, a blend of the
UKMET and GEFS mean produces something reasonably close to the
track as these are on the slow edge of the envelope. The trend in
the ECMWF has been towards this slower solution over the past 3
runs, so some weight on the ECMWF is usable, but will end up with
something that gradually gets more away from the official track.
For this reason, a UKMET/GEFS blend is suggested. This will also
play a role in the QPF, especially across eastern New England
Thursday as interaction with the approaching front has the
potential to produce heavy rainfall. The UKMET is the most
aggressive with QPF, and despite its typical wet bias, it being
the best deterministic model towards the track suggests an
increasing likelihood for heavy rainfall, so from a QPF standpoint
a general model blend is acceptable here.

Elsewhere across the CONUS, there is generally good agreement in a
western ridge/eastern trough generally becoming more zonal by day
3. While subtle differences exist in timing and intensity of
features which will eventually describe the synoptic shift, the
spread appears minimal. The biggest discrepancies lie in the
amplitude of the western ridge early in the forecast period, but
overall the QPF along the stalled front in the south and height
anomalies by day 3 feature a small envelope, so a general model
blend should be sufficient.

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

Weiss