GFS '002 behind CMC

1
Northern Hemisphere 500 last month
GFS .002 behind CMC .006 behind UKMO
.043 behind ECMWF
2
Southern Hemisphere 500 last month
GFS .008 behind UKMO .043 behind ECMWF
3
Northern Hemisphere 500 anomaly correlation day 8
Day 8 forecasts useful in winter Skill increased
in last 3 years
4
Southern Hemisphere day 5 anomaly correlation
Not as far behind ECMWF and Met Office As few
months ago
5
Southern Hemisphere day 8 anomaly correlation
Hard to see trend since 2001
6
Improvement last 3 years
NH rmse hgt day 1
7
Improvement last 4 years
NH rmse hgt day 3
8
Improvement last 4 years
NH rmse hgt day 5
9
Improvement last 4 years
NH rmse hgt day 8
10
Days 6-10 tied with ECMWF, explain just over
half the variance
11
Week 2 GFS just below .6, several periods of
useful skill Ensemble on average useful skill
12
Improvement last 3 years compared to CDAS, ECMWF
other centers improved more
SH rmse hgt day 1
13
SH rmse hgt day 3
Improvement relative to cdas last 10 years
14
SH rmse hgt day 5
Met Office, CMC, FNMOC Improved last few years
15
Forecasts useful out to day 8 in Northern
Hemisphere winter Improvement last few years in
Northern Hemisphere Improvement last few years
at days 1 and 3 Southern Hemisphere, Harder to
see at longer ranges in Southern Hemisphere
16
Feb. 2008
Feb. 2007
17
August 2007
GFS large bias, Relatively poor threat scores
February Relatively better threat scores August
18
Prs50h Jan.1March 25, 2008 Threat scores
slightly negative for parallel RMS height and
wind scores positive AC not significantly
different
19
(No Transcript)
20
(No Transcript)
21
(No Transcript)
22
(No Transcript)
23
(No Transcript)
24
(No Transcript)
25
(No Transcript)
26
(No Transcript)
27
(No Transcript)