How Did Models Do?

February 2, 2011

Prompted by an email, I figured I should see how the the UW’s Mesoscale models did with respect to reality with our mini cold snap.  I know, not exactly a cold snap, but that’s relative to the month of January I suppose.

Here is data from the NAM-MM5 and WRF-GFS models for three locations, Portland, Salem, and The Dalles.  The number in parenthesis is the departure from the actual temp.  So a negative number means the model was too cold, zero is spot on, and positive means the model was too warm:

PORTLAND MON HI TUES LO TUES HI WED LO
ACTUAL TEMP 44 34 42 29
NAM-MM5 FCST 39 (-5) 21 (-13) 37 (-5) 25 (-4)
WRF-GFS FCST 48 (+4) 34 (0) 42 (0) 36 (+7)

 

SALEM MON HI TUES LO TUES HI WED LO
ACTUAL TEMP 46 32 46 26
NAM-MM5 FCST 41 (-5) 21 (-11) 38 (-8) 21 (-5)
WRF-GFS FCST 51 (+5) 30 (-2) 48 (+2) 29 (+3)

 

THE DALLES MON HI TUES LO TUES HI WED LO
ACTUAL TEMP 38 24 38 18
NAM-MM5 FCST 13 (-25) 3 (-21) 12 (-26) 5 (-13)
WRF-GFS FCST 41 (+3) 25 (+1) 39 (+1) 25 (+7)

The NAM was too cold in the end, especially east of the Cascades.  Actually it was pathetic over there.  A quick check of Redmond showed a huge cold bias there too.  The WRF-GFS was slightly warm but did very well here in the Metro area, until this morning.    Note the NAM forecast highs only in the 30s to maybe 40 up and down the Valley both Monday and Tuesday, but that was a good 5-10 degrees too cold.

This morning’s lows are interesting because the WRF-GFS was too quick to warm things up (big surprise…no) in the cold air coming through the Gorge.  I think the NAM did a little better this morning due to holding onto the cold air longer, even if it was far to cold to start with.  I should also point out that the WRF-GFS was too warm (maybe keeping it too windy?) for outlying areas.  Although I think we all know that dewpoints forecast in the teens and clear/calm conditions means lots of 20s away from the wind.  Even we forecast mid-upper 20s here in the city with colder outlying areas.

Let’s talk about the long range maps…ah, well, nevermind…maybe another time?

Chief Meteorologist Mark Nelsen