an analysis of subseasonal variability in the ncep cfs and nasa nsipp coupled gcms

An Analysis of Subseasonal Variability in the NCEP CFS and NASA NSIPP

Coupled GCMs

Myong-In Lee1,2, Siegfried Schubert2, Max Suarez2, Phil Pegion2, Ben Kirtman3, Kathy Pegion3, Arun Kumar4, Bhaskar Jha4, and Duane Waliser4

1 Goddard Earth Sciences and Technology Center/ UMBC2 NASA/GSFC Global Modeling and Assimilation Office

3 COLA/George Mason University4 NCEP/ Climate Prediction Center

5 NASA/Jet Propulsion Laboratory/CalTech

The 30th Annual Climate Diagnostics & Prediction Workshop

The Pennsylvania State University

October 24-28, 2005

Questions ??Questions ??

How well do the current coupled models reproduce How well do the current coupled models reproduce the leading patterns of extratropical subseasonal the leading patterns of extratropical subseasonal variability?variability? CGCM intercomparisons ( CFS and NSIPP CGCM)CGCM intercomparisons ( CFS and NSIPP CGCM) Coupling versus prescribed (comparison with AMIP)Coupling versus prescribed (comparison with AMIP)

How well do the current coupled models reproduce How well do the current coupled models reproduce the changes in subseasonal variability associated the changes in subseasonal variability associated with ENSO ?with ENSO ? ENSO simulation in CGCMsENSO simulation in CGCMs Subseasonal variance changesSubseasonal variance changes

Model DescriptionsModel Descriptions

NCEP CFS T62L64 NCEP CFS T62L64 NCEP Global Forecast System (GFS) for the

atmospheric component GFDL Modular Ocean Model version 3 (MOM3)

for the ocean component

NASA NSIPP CGCM v.1NASA NSIPP CGCM v.1 NSIPP1 AGCM (1x1.25)NSIPP1 AGCM (1x1.25) Poseidon v4 (1/3x5/8xL27) OGCM Poseidon v4 (1/3x5/8xL27) OGCM

DatasetsDatasets 50 years of NCEP/NCAR Reanalysis-150 years of NCEP/NCAR Reanalysis-1

1951-2000 1951-2000 Daily and monthly GPHDaily and monthly GPH200200 Monthly UMonthly U200200

50 years of monthly HadISST50 years of monthly HadISST

50 years coupled runs50 years coupled runs NCEP CFS (T62L64)NCEP CFS (T62L64) NSIPP CGCM (1x1.25)NSIPP CGCM (1x1.25)

50 years of AMIP (1951-2000) Runs50 years of AMIP (1951-2000) Runs NCEP GFS T62L64 (T62L64)NCEP GFS T62L64 (T62L64) 9-member NSIPP AGCM ensemble runs 9-member NSIPP AGCM ensemble runs

(2x2.5)(2x2.5)

Principal patterns of monthly 200 mb height (Rotated EOFs)

meters per STDV

ENSO AO

AAO

PNA

ENSO AO

AAO

PNA

Variance of Leading PatternsVariance of Leading Patterns

(Unit: meter2)

0

50

100

150

200

250

300

350

400

ENSO AO AAO PNA NAO RMN(*0.1) TOTAL(*0.1)

Mode

variance

Reanalysis

GFS AMIP

NSIPP AMIP

CFS CGCM

NSIPP CGCM

Reanalysis (var.=341 m2)

ENSO

CFS (coupled)

NSIPP (coupled) (var.=134 m2)

(var.=252 m2) GFS AMIP

NSIPP AMIP

(var.=340 m2)

(var.=364±9 m2)

Reanalysis(var.=314 m2)

GFS AMIP

NSIPP AMIP

(var.=371 m2)

(var.=308±33 m2)

CFS(coupled)

NSIPP(coupled) (var.=339 m2)

(var.=217 m2)

AO


AAO

GFS AMIP

NSIPP AMIP

(var.=283 m2)

(var.=231±25 m2)

CFS

NSIPP (var.=282 m2)

(var.=242 m2)


GFS AMIP

NSIPP AMIP

(var.=193 m2)

(var.=176±14 m2)

CFS

NSIPP (var.=197 m2)

(var.=180 m2)

PNA


NAO

GFS AMIP

NSIPP AMIP

(var.=164 m2)

(var.=128±26 m2)

(var.=109 m2)CFS

(var.=109 m2)NSIPP

Ensemble spreads in NSIPP AMIP runsEnsemble spreads in NSIPP AMIP runs(spaghetti diagram of EOFs)(spaghetti diagram of EOFs)

- contours from each ensemble members- contours from each ensemble members

- compared with the reanalysis (shading)- compared with the reanalysis (shading)

ENSO response from NSIPP 9 AMIPs

AO from NSIPP 9 AMIPs AAO from NSIPP 9 AMIPs

PNA from NSIPP 9 AMIPs NAO from NSIPP 9 AMIPs

ENSO simulations in the CGCMsENSO simulations in the CGCMs

Time-Longitude SSTA (5S-5N)

Reanalysis

CFS

NSIPP

Nino3 SSTA (5S-5N, 150-90W)

Warm SST Composite(> 1σ)

Cold SST Composite(< -1σ)

Subseasonal Variance AnalysisSubseasonal Variance Analysis

- GPH- GPH200200 daily daily

- remove seasonal cycle - remove seasonal cycle (0-3 harmonics of 50 year-averaged daily climatology)(0-3 harmonics of 50 year-averaged daily climatology)

- 10-60 day band-pass filtered - 10-60 day band-pass filtered

- variance in NH winter (DJF)- variance in NH winter (DJF)

22warmcold

Subseasonal VarianceChanges (La Nina-El Nino)

contour: 200mb u-wind difference

Rotated EOFs from daily band-pass (10-60d) filtered 200 mb height

NSIPP AMIP

0

100

200

300

400

500

600

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41

mode

m^2

NSIPP CGCM

0

100

200

300

400

500

600

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41

mode

m^2

GFS AMIP

0

100

200

300

400

500

600

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41

mode

m^2

CFS CGCM

0

100

200

300

400

500

600

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41

mode

m^2

Reanalysis

0

100

200

300

400

500

600

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41

mode

m^2

Variance of PCs (daily 200 mb height)

Reanalysis

GFS AMIP

NSIPP AMIPNSIPP (coupled)

CFS (coupled)

AAOAONAOPNA

AAOAO

NAO

PNA

AAO

AO

NAOPNA

AAOAONAOPNA

AAOAONAO

PNA

NH pattern SH patternTotal variance(*0.1)

Reanalysis

Pattern 1 Pattern 2 Pattern 3 Pattern 4

GFS AMIP

NSIPP AMIP

CFS (coupled)

NSIPP (coupled)

NAOPNA

r=0.88 r=0.80 r=0.73 r=0.97

r=0.97 r=0.97 r=0.94 r=0.83

r=0.96 r=0.85 r=0.95 r=0.83

r=0.96 r=0.79 r=0.72 r=0.94

#1#2 #20 #6

#7

#8

#8#13

#2

#5#1

#1

#1

#3

#4

#4


Reanalysis

AO

GFS AMIP

NSIPP AMIP

CFS (coupled)

NSIPP (coupled)

r=0.94 r=0.95 r=0.68 r=0.81

r=0.94 r=0.92 r=0.93 r=0.78

r=0.95 r=0.91 r=0.80 r=0.71

r=0.90 r=0.89 r=0.47 r=0.96

#5#4 #13 #9

#12

#7

#2#3

#9

#6#3

#11

#5

#5

#9

#2

r=0.92 r=0.87 r=0.96 r=0.80

r=0.92 r=0.96 r=0.94 r=0.95

r=0.95 r=0.94 r=0.82 r=0.89

r=0.84 r=0.92 r=0.96 r=0.94

Reanalysis

GFS AMIP

NSIPP AMIP

CFS (coupled)

NSIPP (coupled)

#7#10 #8 #14

#10

#22

#11#14

#4

#13#9

#13

#6

#6

#7

#8


Reanalysis

GFS AMIP

NSIPP AMIPNSIPP (coupled)

CFS (coupled)

Variance difference (cold-warm) – reconstructed from EOFs

(*1000 m2)

Variance difference (cold-warm) - Reanalysis

p1 p2 p3 p4

p5 p6 p7 p8

p9 p10 p11 p12

PNA NAO

AO

Variance difference (cold-warm) – GFS AMIP

p1 p2 p3 p4

p5 p6 p7 p8

p9 p10 p11 p12

PNA NAO

AO

Variance difference (cold-warm) – NSIPP AMIP

p1 p2 p3 p4

p5 p6 p7 p8

p9 p10 p11 p12

PNA NAO

AO

Variance difference (cold-warm) – CFS coupled

p1 p2 p3 p4

p5 p6 p7 p8

p9 p10 p11 p12

PNA NAO

AO

Variance difference (cold-warm) – NSIPP coupled

p1 p2 p3 p4

p5 p6 p7 p8

p9 p10 p11 p12

PNA NAO

AO

SummarySummary

1.1. Current models reproduce the leading wintertime extra-Current models reproduce the leading wintertime extra-tropical patterns of tropical patterns of monthlymonthly variability reasonably well variability reasonably well

REOFs identify the patterns of ENSO, AO, AAO, PNA and NAOREOFs identify the patterns of ENSO, AO, AAO, PNA and NAO

an assessment of the spread of the NSIPP AMIP ensemble an assessment of the spread of the NSIPP AMIP ensemble shows these patterns to be robust in samples of 50 yearsshows these patterns to be robust in samples of 50 years

there are, however, large differences in the variance of there are, however, large differences in the variance of individual patterns individual patterns

total monthly variance is in general weaker in the simulations total monthly variance is in general weaker in the simulations

2.2. An assessment of the full subseasonal (10-60 day) variance shows the An assessment of the full subseasonal (10-60 day) variance shows the following:following:

the variance is weak in the coupled runs, whereas it is comparable to the the variance is weak in the coupled runs, whereas it is comparable to the reanalysis in the AMIP runs reanalysis in the AMIP runs

but, all models have a sign of variance increase over the northern reach of Pacific but, all models have a sign of variance increase over the northern reach of Pacific and Alaska region and decrease over the arctic and northern Atlantic regionand Alaska region and decrease over the arctic and northern Atlantic region

REOFs from daily band-pass 200 mb height show a much richer spectrum of REOFs from daily band-pass 200 mb height show a much richer spectrum of patterns compared with the monthly results, apparently a variation of several patterns compared with the monthly results, apparently a variation of several leading principal patternsleading principal patterns

interannual changes in the subseasonal variance associated with ENSO have interannual changes in the subseasonal variance associated with ENSO have realistic patterns (but are weak, especially in the coupled runs)realistic patterns (but are weak, especially in the coupled runs)

changes in the PNA and NAO are robust, though not so for other leading patternschanges in the PNA and NAO are robust, though not so for other leading patterns

3.3. Future work will focus on furthering our understanding of the nature of the Future work will focus on furthering our understanding of the nature of the various subseasonal patterns and their underlying dynamicsvarious subseasonal patterns and their underlying dynamics

Summary Summary (continued)(continued)

Thank You !

an analysis of subseasonal variability in the ncep cfs and nasa nsipp coupled gcms

Documents