Northern Hemisphere blocking simulation in current climate … · 2020. 6. 15. · Schiemann et...
Transcript of Northern Hemisphere blocking simulation in current climate … · 2020. 6. 15. · Schiemann et...
Supplement of Weather Clim. Dynam., 1, 277–292, 2020https://doi.org/10.5194/wcd-1-277-2020-supplement© Author(s) 2020. This work is distributed underthe Creative Commons Attribution 4.0 License.
Supplement of
Northern Hemisphere blocking simulation in current climatemodels: evaluating progress from the Climate ModelIntercomparison Project Phase 5 to 6 and sensitivity to resolutionReinhard Schiemann et al.
Correspondence to: Reinhard Schiemann ([email protected])
The copyright of individual parts of the supplement might differ from the CC BY 4.0 License.
Schiemann et al.: Blocking in GCMs (Supplement) 1
Supplement
Northern Hemisphere blocking simulation in current climatemodels: evaluating progress from CMIP5 to CMIP6 and sen-sitivity to resolution, Weather and Climate Dynamics, 2020,by Reinhard Schiemann et al.5
List of Figures
S1 Blocking frequency bias for the ANOM in-dex, boreal winter, and (a) high-resolutionforced, (b) high-resolution coupled, (c) low-resolution forced, (d) low-resolution coupled10
PRIMAVERA simulations, and (e) CMIP5,(f) CMIP6 simulations. Stippling showsagreement on the sign of the bias by at least(a,c) 6 of 6, (b,d) 6 of 7, (e) 19 of 29, and (f)10 of 13 simulations. . . . . . . . . . . . . . 415
S2 As Fig. S1 but for boreal summer. (Blockingfrequency bias for the ANOM index and (a)high-resolution forced, (b) high-resolutioncoupled, (c) low-resolution forced, (d) low-resolution coupled PRIMAVERA simula-20
tions, and (e) CMIP5, (f) CMIP6 simula-tions. Stippling shows agreement on the signof the bias by at least (a,c) 6 of 6, (b,d) 6 of7, (e) 19 of 29, and (f) 10 of 13 simulations.) . 5
S3 Metrics of blocking performance (a,b -25
blocking frequency, c,d - spatial correla-tion, e,f - root-mean-square error) for theANOM index and boreal winter, for theATL domain (-90-90E, 50-90N). The left-hand side of each panel shows metrics for30
PRIMAVERA simulations at different gridspacings (resolutions). Boxplots on the right-hand side show distributions of the metricacross CMIP5 and CMIP6 simulations. The’*’ symbol in the column ’ERA/IV’ shows35
the reanalysis estimate and the boxplot is anestimate of the expected agreement given in-ternal variability (see text). . . . . . . . . . . 6
S4 As Fig. S3 but for boreal summer. (Met-rics of blocking performance (a,b - block- 40
ing frequency, c,d - spatial correlation, e,f- root-mean-square error) for the ANOMindex and for the ATL domain (-90-90E,50-90N). The left-hand side of each panelshows metrics for PRIMAVERA simulations 45
at different grid spacings (resolutions). Box-plots on the right-hand side show distribu-tions of the metric across CMIP5 and CMIP6simulations. The ’*’ symbol in the column’ERA/IV’ shows the reanalysis estimate and 50
the boxplot is an estimate of the expectedagreement given internal variability (see text).) 7
S5 As Fig. S3 but for the AGP index and thePAC domain (90-270E, 50-75N). (Metricsof blocking performance (a,b - blocking fre- 55
quency, c,d - spatial correlation, e,f - root-mean-square error) for boreal winter. Theleft-hand side of each panel shows met-rics for PRIMAVERA simulations at dif-ferent grid spacings (resolutions). Boxplots 60
on the right-hand side show distributionsof the metric across CMIP5 and CMIP6simulations. The ’*’ symbol in the column’ERA/IV’ shows the reanalysis estimate andthe boxplot is an estimate of the expected 65
agreement given internal variability (see text).) 8
S6 As Fig. S3 but for the AGP index, the PACdomain (90-270E, 50-75N) and boreal sum-mer. (Metrics of blocking performance (a,b- blocking frequency, c,d - spatial correla- 70
tion, e,f - root-mean-square error). The left-hand side of each panel shows metrics forPRIMAVERA simulations at different gridspacings (resolutions). Boxplots on the right-hand side show distributions of the metric 75
across CMIP5 and CMIP6 simulations. The’*’ symbol in the column ’ERA/IV’ showsthe reanalysis estimate and the boxplot is anestimate of the expected agreement given in-ternal variability (see text).) . . . . . . . . . . 9 80
2 Schiemann et al.: Blocking in GCMs (Supplement)
S7 As Fig. S3 but for the PAC domain (120-240E, 40-90N). (Metrics of blocking perfor-mance (a,b - blocking frequency, c,d - spatialcorrelation, e,f - root-mean-square error) forthe ANOM index and boreal winter. The left-5
hand side of each panel shows metrics forPRIMAVERA simulations at different gridspacings (resolutions). Boxplots on the righ-hand side show distributions of the metricacross CMIP5 and CMIP6 simulations. The10
’*’ symbol in the column ’ERA/IV’ showsthe reanalysis estimate and the boxplot is anestimate of the expected agreement given in-ternal variability (see text).) . . . . . . . . . . 10
S8 As Fig. S3 but for the PAC domain (120-15
240E, 40-90N) and boreal summer. (Met-rics of blocking performance (a,b - block-ing frequency, c,d - spatial correlation, e,f -root-mean-square error) for the ANOM in-dex. The left-hand side of each panel shows20
metrics for PRIMAVERA simulations at dif-ferent grid spacings (resolutions). Boxplotson the righ-hand side show distributions ofthe metric across CMIP5 and CMIP6 sim-ulations. The ’*’ symbol in the column25
’ERA/IV’ shows the reanalysis estimate andthe boxplot is an estimate of the expectedagreement given internal variability (see text).) 11
S9 Persistence of blocking events (a,b - me-dian, c,d - 90th percentile) for the ANOM30
index and boreal winter, for the ATL domain(-90-90E, 50-90N). The left-hand side ofeach panel shows metrics for PRIMAVERAsimulations at different grid spacings (res-olutions). Boxplots on the right-hand side35
show distributions of the persistence metricacross CMIP5 and CMIP6 simulations. The’*’ symbol in the column ’ERA/IV’ showsthe reanalysis estimate and the boxplot is anestimate of the expected agreement given in-40
ternal variability (see text). . . . . . . . . . . 12
S10 As Fig. S9 but for boreal summer. (Per-sistence of blocking events (a,b - median,c,d - 90th percentile) for the ANOM indexand for the ATL domain (-90-90E, 50-90N).45
The left-hand side of each panel shows met-rics for PRIMAVERA simulations at differ-ent grid spacings (resolutions). Boxplots onthe right-hand side show distributions of thepersistence metric across CMIP5 and CMIP650
simulations. The ’*’ symbol in the column’ERA/IV’ shows the reanalysis estimate andthe boxplot is an estimate of the expectedagreement given internal variability (see text).) 13
S11 As Fig. S9 but for the AGP index and for 55
the PAC domain (90-270E, 50-75N). Per-sistence of blocking events (a,b - median,c,d - 90th percentile) for boreal winter. Theleft-hand side of each panel shows metricsfor PRIMAVERA simulations at different 60
grid spacings (resolutions). Boxplots on theright-hand side show distributions of the per-sistence metric across CMIP5 and CMIP6simulations. The ’*’ symbol in the column’ERA/IV’ shows the reanalysis estimate and 65
the boxplot is an estimate of the expectedagreement given internal variability (see text).) 14
S12 As Fig. S9 but for the AGP index, for thePAC domain (90-270E, 50-75N), and for bo-real summer. (Persistence of blocking events 70
(a,b - median, c,d - 90th percentile). Theleft-hand side of each panel shows metricsfor PRIMAVERA simulations at differentgrid spacings (resolutions). Boxplots on theright-hand side show distributions of the per- 75
sistence metric across CMIP5 and CMIP6simulations. The ’*’ symbol in the column’ERA/IV’ shows the reanalysis estimate andthe boxplot is an estimate of the expectedagreement given internal variability (see text).) 15 80
S13 As Fig. S9 but for the PAC domain (120-240E, 40-90N). (Persistence of blockingevents (a,b - median, c,d - 90th percentile)for the ANOM index and boreal winter. Theleft-hand side of each panel shows metrics 85
for PRIMAVERA simulations at differentgrid spacings (resolutions). Boxplots on theright-hand side show distributions of the per-sistence metric across CMIP5 and CMIP6simulations. The ’*’ symbol in the column 90
’ERA/IV’ shows the reanalysis estimate andthe boxplot is an estimate of the expectedagreement given internal variability (see text).) 16
S14 As Fig. S9 but for the PAC domain (120-240E, 40-90N) and boreal summer. (Persis- 95
tence of blocking events (a,b - median, c,d- 90th percentile) for the ANOM index. Theleft-hand side of each panel shows metricsfor PRIMAVERA simulations at differentgrid spacings (resolutions). Boxplots on the 100
right-hand side show distributions of the per-sistence metric across CMIP5 and CMIP6simulations. The ’*’ symbol in the column’ERA/IV’ shows the reanalysis estimate andthe boxplot is an estimate of the expected 105
agreement given internal variability (see text).) 17
Schiemann et al.: Blocking in GCMs (Supplement) 3
S15 Decomposition of blocking bias into (i) aterm due to the bias in the number of events(E), (ii) a term due to the bias in persis-tence (P), and (iii) a cross term (C). Boxplotsshow the ensemble spread across the CMIP5,5
CMIP6, and all coupled PRIMAVERA mod-els, respectively. The AGP index with 5-daypersistence is used and results are shown forthe ATL and PAC domains, and for borealwinter (DJF) and summer (JJA). The block-10
ing bias is normalised by the reanalysis esti-mate of blocking frequency. . . . . . . . . . . 17
List of Tables
S1 CMIP5 simulations used in this study. Ther1i1p1 ensemble member from the historical15
experiment is used for each simulation. . . . 18S2 CMIP6 simulations used in this study. All
simulations are for the historical experimentduring 1950–2014. . . . . . . . . . . . . . . 18
4 Schiemann et al.: Blocking in GCMs (Supplement)
PRIMAVERA HF(x6) ANOM
a DJF
PRIMAVERA HC(x7) ANOM
b DJF
PRIMAVERA LF(x6) ANOM
c DJF
PRIMAVERA LC(x7) ANOM
d DJF
CMIP5(x29) ANOM
e DJF
CMIP6(x13) ANOM
f DJF
-0.06 -0.04 -0.02 0.00 0.02 0.04 0.06
Figure S1. Blocking frequency bias for the ANOM index, boreal winter, and (a) high-resolution forced, (b) high-resolution coupled, (c) low-resolution forced, (d) low-resolution coupled PRIMAVERA simulations, and (e) CMIP5, (f) CMIP6 simulations. Stippling shows agreementon the sign of the bias by at least (a,c) 6 of 6, (b,d) 6 of 7, (e) 19 of 29, and (f) 10 of 13 simulations.
Schiemann et al.: Blocking in GCMs (Supplement) 5
PRIMAVERA HF(x6) ANOM
a JJA
PRIMAVERA HC(x7) ANOM
b JJA
PRIMAVERA LF(x6) ANOM
c JJA
PRIMAVERA LC(x7) ANOM
d JJA
CMIP5(x29) ANOM
e JJA
CMIP6(x13) ANOM
f JJA
-0.06 -0.04 -0.02 0.00 0.02 0.04 0.06
Figure S2. As Fig. S1 but for boreal summer. (Blocking frequency bias for the ANOM index and (a) high-resolution forced, (b) high-resolution coupled, (c) low-resolution forced, (d) low-resolution coupled PRIMAVERA simulations, and (e) CMIP5, (f) CMIP6 simulations.Stippling shows agreement on the sign of the bias by at least (a,c) 6 of 6, (b,d) 6 of 7, (e) 19 of 29, and (f) 10 of 13 simulations.)
6 Schiemann et al.: Blocking in GCMs (Supplement)
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IVatm. grid spacing at 50°N (km)
0.00
0.02
0.04
0.06
0.08
0.10
0.12
frequ
ency
of b
lock
ed d
ays
ANOM, ATL, DJF, coupled
a
ANOM, ATL, DJF, coupled
a
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.00
0.02
0.04
0.06
0.08
0.10
0.12
frequ
ency
of b
lock
ed d
ays
ANOM, ATL, DJF, AMIP
b
ANOM, ATL, DJF, AMIP
b
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.2
0.4
0.6
0.8
1.0
1.2
spat
ial c
orre
latio
n
ANOM, ATL, DJF, coupled
c
ANOM, ATL, DJF, coupled
c
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.2
0.4
0.6
0.8
1.0
1.2sp
atia
l cor
rela
tion
ANOM, ATL, DJF, AMIP
d
ANOM, ATL, DJF, AMIP
d
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.00
0.01
0.02
0.03
0.04
0.05
0.06
root
-mea
n-sq
uare
erro
r
ANOM, ATL, DJF, coupled
e
ANOM, ATL, DJF, coupled
e
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.00
0.01
0.02
0.03
0.04
0.05
0.06
root
-mea
n-sq
uare
erro
r
ANOM, ATL, DJF, AMIP
f
ANOM, ATL, DJF, AMIP
f
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
Figure S3. Metrics of blocking performance (a,b - blocking frequency, c,d - spatial correlation, e,f - root-mean-square error) for the ANOMindex and boreal winter, for the ATL domain (-90-90E, 50-90N). The left-hand side of each panel shows metrics for PRIMAVERA simu-lations at different grid spacings (resolutions). Boxplots on the right-hand side show distributions of the metric across CMIP5 and CMIP6simulations. The ’*’ symbol in the column ’ERA/IV’ shows the reanalysis estimate and the boxplot is an estimate of the expected agreementgiven internal variability (see text).
Schiemann et al.: Blocking in GCMs (Supplement) 7
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IVatm. grid spacing at 50°N (km)
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
frequ
ency
of b
lock
ed d
ays
ANOM, ATL, JJA, coupled
a
ANOM, ATL, JJA, coupled
a
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
frequ
ency
of b
lock
ed d
ays
ANOM, ATL, JJA, AMIP
b
ANOM, ATL, JJA, AMIP
b
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.2
0.4
0.6
0.8
1.0
1.2
spat
ial c
orre
latio
n
ANOM, ATL, JJA, coupled
c
ANOM, ATL, JJA, coupled
c
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.2
0.4
0.6
0.8
1.0
1.2
spat
ial c
orre
latio
nANOM, ATL, JJA, AMIP
d
ANOM, ATL, JJA, AMIP
d
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.035
0.040
root
-mea
n-sq
uare
erro
r
ANOM, ATL, JJA, coupled
e
ANOM, ATL, JJA, coupled
e
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.035
0.040
root
-mea
n-sq
uare
erro
r
ANOM, ATL, JJA, AMIP
f
ANOM, ATL, JJA, AMIP
f
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
Figure S4. As Fig. S3 but for boreal summer. (Metrics of blocking performance (a,b - blocking frequency, c,d - spatial correlation, e,f -root-mean-square error) for the ANOM index and for the ATL domain (-90-90E, 50-90N). The left-hand side of each panel shows metricsfor PRIMAVERA simulations at different grid spacings (resolutions). Boxplots on the right-hand side show distributions of the metric acrossCMIP5 and CMIP6 simulations. The ’*’ symbol in the column ’ERA/IV’ shows the reanalysis estimate and the boxplot is an estimate of theexpected agreement given internal variability (see text).)
8 Schiemann et al.: Blocking in GCMs (Supplement)
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IVatm. grid spacing at 50°N (km)
0.000
0.005
0.010
0.015
0.020
0.025
0.030
frequ
ency
of b
lock
ed d
ays
AGP, PAC, DJF, coupled
a
AGP, PAC, DJF, coupled
a
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.000
0.005
0.010
0.015
0.020
0.025
0.030
frequ
ency
of b
lock
ed d
ays
AGP, PAC, DJF, AMIP
b
AGP, PAC, DJF, AMIP
b
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.70
0.75
0.80
0.85
0.90
0.95
1.00
spat
ial c
orre
latio
n
AGP, PAC, DJF, coupled
c
AGP, PAC, DJF, coupled
c
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.70
0.75
0.80
0.85
0.90
0.95
1.00
spat
ial c
orre
latio
nAGP, PAC, DJF, AMIP
d
AGP, PAC, DJF, AMIP
d
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.0000
0.0025
0.0050
0.0075
0.0100
0.0125
0.0150
0.0175
0.0200
root
-mea
n-sq
uare
erro
r
AGP, PAC, DJF, coupled
e
AGP, PAC, DJF, coupled
e
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.0000
0.0025
0.0050
0.0075
0.0100
0.0125
0.0150
0.0175
0.0200
root
-mea
n-sq
uare
erro
r
AGP, PAC, DJF, AMIP
f
AGP, PAC, DJF, AMIP
f
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
Figure S5. As Fig. S3 but for the AGP index and the PAC domain (90-270E, 50-75N). (Metrics of blocking performance (a,b - blockingfrequency, c,d - spatial correlation, e,f - root-mean-square error) for boreal winter. The left-hand side of each panel shows metrics forPRIMAVERA simulations at different grid spacings (resolutions). Boxplots on the right-hand side show distributions of the metric acrossCMIP5 and CMIP6 simulations. The ’*’ symbol in the column ’ERA/IV’ shows the reanalysis estimate and the boxplot is an estimate of theexpected agreement given internal variability (see text).)
Schiemann et al.: Blocking in GCMs (Supplement) 9
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IVatm. grid spacing at 50°N (km)
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.035
frequ
ency
of b
lock
ed d
ays
AGP, PAC, JJA, coupled
a
AGP, PAC, JJA, coupled
a
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.035
frequ
ency
of b
lock
ed d
ays
AGP, PAC, JJA, AMIP
b
AGP, PAC, JJA, AMIP
b
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.60
0.65
0.70
0.75
0.80
0.85
0.90
0.95
1.00
spat
ial c
orre
latio
n
AGP, PAC, JJA, coupled
c
AGP, PAC, JJA, coupled
c
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.60
0.65
0.70
0.75
0.80
0.85
0.90
0.95
1.00sp
atia
l cor
rela
tion
AGP, PAC, JJA, AMIP
d
AGP, PAC, JJA, AMIP
d
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.000
0.005
0.010
0.015
0.020
0.025
root
-mea
n-sq
uare
erro
r
AGP, PAC, JJA, coupled
e
AGP, PAC, JJA, coupled
e
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.000
0.005
0.010
0.015
0.020
0.025
root
-mea
n-sq
uare
erro
r
AGP, PAC, JJA, AMIP
f
AGP, PAC, JJA, AMIP
f
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
Figure S6. As Fig. S3 but for the AGP index, the PAC domain (90-270E, 50-75N) and boreal summer. (Metrics of blocking performance (a,b- blocking frequency, c,d - spatial correlation, e,f - root-mean-square error). The left-hand side of each panel shows metrics for PRIMAVERAsimulations at different grid spacings (resolutions). Boxplots on the right-hand side show distributions of the metric across CMIP5 and CMIP6simulations. The ’*’ symbol in the column ’ERA/IV’ shows the reanalysis estimate and the boxplot is an estimate of the expected agreementgiven internal variability (see text).)
10 Schiemann et al.: Blocking in GCMs (Supplement)
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IVatm. grid spacing at 50°N (km)
0.00
0.02
0.04
0.06
0.08
0.10
0.12
frequ
ency
of b
lock
ed d
ays
ANOM, PAC, DJF, coupled
a
ANOM, PAC, DJF, coupled
a
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.00
0.02
0.04
0.06
0.08
0.10
0.12
frequ
ency
of b
lock
ed d
ays
ANOM, PAC, DJF, AMIP
b
ANOM, PAC, DJF, AMIP
b
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.2
0.4
0.6
0.8
1.0
1.2
spat
ial c
orre
latio
n
ANOM, PAC, DJF, coupled
c
ANOM, PAC, DJF, coupled
c
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.2
0.4
0.6
0.8
1.0
1.2
spat
ial c
orre
latio
nANOM, PAC, DJF, AMIP
d
ANOM, PAC, DJF, AMIP
d
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.035
0.040
root
-mea
n-sq
uare
erro
r
ANOM, PAC, DJF, coupled
e
ANOM, PAC, DJF, coupled
e
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.035
0.040
root
-mea
n-sq
uare
erro
r
ANOM, PAC, DJF, AMIP
f
ANOM, PAC, DJF, AMIP
f
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
Figure S7. As Fig. S3 but for the PAC domain (120-240E, 40-90N). (Metrics of blocking performance (a,b - blocking frequency, c,d -spatial correlation, e,f - root-mean-square error) for the ANOM index and boreal winter. The left-hand side of each panel shows metrics forPRIMAVERA simulations at different grid spacings (resolutions). Boxplots on the righ-hand side show distributions of the metric acrossCMIP5 and CMIP6 simulations. The ’*’ symbol in the column ’ERA/IV’ shows the reanalysis estimate and the boxplot is an estimate of theexpected agreement given internal variability (see text).)
Schiemann et al.: Blocking in GCMs (Supplement) 11
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IVatm. grid spacing at 50°N (km)
0.00
0.01
0.02
0.03
0.04
0.05
0.06
frequ
ency
of b
lock
ed d
ays
ANOM, PAC, JJA, coupled
a
ANOM, PAC, JJA, coupled
a
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.00
0.01
0.02
0.03
0.04
0.05
0.06
frequ
ency
of b
lock
ed d
ays
ANOM, PAC, JJA, AMIP
b
ANOM, PAC, JJA, AMIP
b
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.2
0.4
0.6
0.8
1.0
1.2
spat
ial c
orre
latio
n
ANOM, PAC, JJA, coupled
c
ANOM, PAC, JJA, coupled
c
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.2
0.4
0.6
0.8
1.0
1.2
spat
ial c
orre
latio
nANOM, PAC, JJA, AMIP
d
ANOM, PAC, JJA, AMIP
d
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.000
0.005
0.010
0.015
0.020
0.025
0.030
root
-mea
n-sq
uare
erro
r
ANOM, PAC, JJA, coupled
e
ANOM, PAC, JJA, coupled
e
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5CM
IP6
ERA/
IV
atm. grid spacing at 50°N (km)
0.000
0.005
0.010
0.015
0.020
0.025
0.030
root
-mea
n-sq
uare
erro
r
ANOM, PAC, JJA, AMIP
f
ANOM, PAC, JJA, AMIP
f
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
Figure S8. As Fig. S3 but for the PAC domain (120-240E, 40-90N) and boreal summer. (Metrics of blocking performance (a,b - blockingfrequency, c,d - spatial correlation, e,f - root-mean-square error) for the ANOM index. The left-hand side of each panel shows metrics forPRIMAVERA simulations at different grid spacings (resolutions). Boxplots on the righ-hand side show distributions of the metric acrossCMIP5 and CMIP6 simulations. The ’*’ symbol in the column ’ERA/IV’ shows the reanalysis estimate and the boxplot is an estimate of theexpected agreement given internal variability (see text).)
12 Schiemann et al.: Blocking in GCMs (Supplement)
0 25 50 75 100
125
CMIP
5
CMIP
6
ERA/
IV
atm. grid spacing at 50°N (km)
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
med
ian
bloc
king
per
siste
nce
(day
s)
ANOM, ATL, DJF, coupled
a
ANOM, ATL, DJF, coupled
a
ANOM, ATL, DJF, coupled
a
ANOM, ATL, DJF, coupled
a
ANOM, ATL, DJF, coupled
a
ANOM, ATL, DJF, coupled
a
ANOM, ATL, DJF, coupled
a
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5
CMIP
6
ERA/
IV
atm. grid spacing at 50°N (km)
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
med
ian
bloc
king
per
siste
nce
(day
s)
ANOM, ATL, DJF, AMIP
b
ANOM, ATL, DJF, AMIP
b
ANOM, ATL, DJF, AMIP
b
ANOM, ATL, DJF, AMIP
b
ANOM, ATL, DJF, AMIP
b
ANOM, ATL, DJF, AMIP
b
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5
CMIP
6
ERA/
IV
atm. grid spacing at 50°N (km)
6
8
10
12
14
16
90th
per
cent
ile b
lock
ing
pers
isten
ce (d
ays) ANOM, ATL, DJF, coupled
c
ANOM, ATL, DJF, coupled
c
ANOM, ATL, DJF, coupled
c
ANOM, ATL, DJF, coupled
c
ANOM, ATL, DJF, coupled
c
ANOM, ATL, DJF, coupled
c
ANOM, ATL, DJF, coupled
c
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5
CMIP
6
ERA/
IV
atm. grid spacing at 50°N (km)
6
8
10
12
14
16
90th
per
cent
ile b
lock
ing
pers
isten
ce (d
ays) ANOM, ATL, DJF, AMIP
d
ANOM, ATL, DJF, AMIP
d
ANOM, ATL, DJF, AMIP
d
ANOM, ATL, DJF, AMIP
d
ANOM, ATL, DJF, AMIP
d
ANOM, ATL, DJF, AMIP
d
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
Figure S9. Persistence of blocking events (a,b - median, c,d - 90th percentile) for the ANOM index and boreal winter, for the ATL domain(-90-90E, 50-90N). The left-hand side of each panel shows metrics for PRIMAVERA simulations at different grid spacings (resolutions).Boxplots on the right-hand side show distributions of the persistence metric across CMIP5 and CMIP6 simulations. The ’*’ symbol in thecolumn ’ERA/IV’ shows the reanalysis estimate and the boxplot is an estimate of the expected agreement given internal variability (see text).
Schiemann et al.: Blocking in GCMs (Supplement) 13
0 25 50 75 100
125
CMIP
5
CMIP
6
ERA/
IV
atm. grid spacing at 50°N (km)
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
med
ian
bloc
king
per
siste
nce
(day
s)
ANOM, ATL, JJA, coupled
a
ANOM, ATL, JJA, coupled
a
ANOM, ATL, JJA, coupled
a
ANOM, ATL, JJA, coupled
a
ANOM, ATL, JJA, coupled
a
ANOM, ATL, JJA, coupled
a
ANOM, ATL, JJA, coupled
a
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5
CMIP
6
ERA/
IV
atm. grid spacing at 50°N (km)
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
med
ian
bloc
king
per
siste
nce
(day
s)
ANOM, ATL, JJA, AMIP
b
ANOM, ATL, JJA, AMIP
b
ANOM, ATL, JJA, AMIP
b
ANOM, ATL, JJA, AMIP
b
ANOM, ATL, JJA, AMIP
b
ANOM, ATL, JJA, AMIP
b
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5
CMIP
6
ERA/
IV
atm. grid spacing at 50°N (km)
6
8
10
12
14
16
90th
per
cent
ile b
lock
ing
pers
isten
ce (d
ays) ANOM, ATL, JJA, coupled
c
ANOM, ATL, JJA, coupled
c
ANOM, ATL, JJA, coupled
c
ANOM, ATL, JJA, coupled
c
ANOM, ATL, JJA, coupled
c
ANOM, ATL, JJA, coupled
c
ANOM, ATL, JJA, coupled
c
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5
CMIP
6
ERA/
IV
atm. grid spacing at 50°N (km)
6
8
10
12
14
16
90th
per
cent
ile b
lock
ing
pers
isten
ce (d
ays) ANOM, ATL, JJA, AMIP
d
ANOM, ATL, JJA, AMIP
d
ANOM, ATL, JJA, AMIP
d
ANOM, ATL, JJA, AMIP
d
ANOM, ATL, JJA, AMIP
d
ANOM, ATL, JJA, AMIP
d
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
Figure S10. As Fig. S9 but for boreal summer. (Persistence of blocking events (a,b - median, c,d - 90th percentile) for the ANOM indexand for the ATL domain (-90-90E, 50-90N). The left-hand side of each panel shows metrics for PRIMAVERA simulations at different gridspacings (resolutions). Boxplots on the right-hand side show distributions of the persistence metric across CMIP5 and CMIP6 simulations.The ’*’ symbol in the column ’ERA/IV’ shows the reanalysis estimate and the boxplot is an estimate of the expected agreement given internalvariability (see text).)
14 Schiemann et al.: Blocking in GCMs (Supplement)
0 25 50 75 100
125
CMIP
5
CMIP
6
ERA/
IV
atm. grid spacing at 50°N (km)
1.8
2.0
2.2
2.4
2.6
2.8
3.0
med
ian
bloc
king
per
siste
nce
(day
s)
AGP, PAC, DJF, coupled
a
AGP, PAC, DJF, coupled
a
AGP, PAC, DJF, coupled
a
AGP, PAC, DJF, coupled
a
AGP, PAC, DJF, coupled
a
AGP, PAC, DJF, coupled
a
AGP, PAC, DJF, coupled
a
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5
CMIP
6
ERA/
IV
atm. grid spacing at 50°N (km)
1.8
2.0
2.2
2.4
2.6
2.8
3.0
med
ian
bloc
king
per
siste
nce
(day
s)
AGP, PAC, DJF, AMIP
b
AGP, PAC, DJF, AMIP
b
AGP, PAC, DJF, AMIP
b
AGP, PAC, DJF, AMIP
b
AGP, PAC, DJF, AMIP
b
AGP, PAC, DJF, AMIP
b
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5
CMIP
6
ERA/
IV
atm. grid spacing at 50°N (km)
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
9.0
90th
per
cent
ile b
lock
ing
pers
isten
ce (d
ays) AGP, PAC, DJF, coupled
c
AGP, PAC, DJF, coupled
c
AGP, PAC, DJF, coupled
c
AGP, PAC, DJF, coupled
c
AGP, PAC, DJF, coupled
c
AGP, PAC, DJF, coupled
c
AGP, PAC, DJF, coupled
c
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5
CMIP
6
ERA/
IV
atm. grid spacing at 50°N (km)
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
9.0
90th
per
cent
ile b
lock
ing
pers
isten
ce (d
ays) AGP, PAC, DJF, AMIP
d
AGP, PAC, DJF, AMIP
d
AGP, PAC, DJF, AMIP
d
AGP, PAC, DJF, AMIP
d
AGP, PAC, DJF, AMIP
d
AGP, PAC, DJF, AMIP
d
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
Figure S11. As Fig. S9 but for the AGP index and for the PAC domain (90-270E, 50-75N). Persistence of blocking events (a,b - median, c,d- 90th percentile) for boreal winter. The left-hand side of each panel shows metrics for PRIMAVERA simulations at different grid spacings(resolutions). Boxplots on the right-hand side show distributions of the persistence metric across CMIP5 and CMIP6 simulations. The ’*’symbol in the column ’ERA/IV’ shows the reanalysis estimate and the boxplot is an estimate of the expected agreement given internalvariability (see text).)
Schiemann et al.: Blocking in GCMs (Supplement) 15
0 25 50 75 100
125
CMIP
5
CMIP
6
ERA/
IV
atm. grid spacing at 50°N (km)
2.0
2.2
2.4
2.6
2.8
3.0
3.2
med
ian
bloc
king
per
siste
nce
(day
s)
AGP, PAC, JJA, coupled
a
AGP, PAC, JJA, coupled
a
AGP, PAC, JJA, coupled
a
AGP, PAC, JJA, coupled
a
AGP, PAC, JJA, coupled
a
AGP, PAC, JJA, coupled
a
AGP, PAC, JJA, coupled
a
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5
CMIP
6
ERA/
IV
atm. grid spacing at 50°N (km)
2.0
2.2
2.4
2.6
2.8
3.0
3.2
med
ian
bloc
king
per
siste
nce
(day
s)
AGP, PAC, JJA, AMIP
b
AGP, PAC, JJA, AMIP
b
AGP, PAC, JJA, AMIP
b
AGP, PAC, JJA, AMIP
b
AGP, PAC, JJA, AMIP
b
AGP, PAC, JJA, AMIP
b
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5
CMIP
6
ERA/
IV
atm. grid spacing at 50°N (km)
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
9.0
90th
per
cent
ile b
lock
ing
pers
isten
ce (d
ays) AGP, PAC, JJA, coupled
c
AGP, PAC, JJA, coupled
c
AGP, PAC, JJA, coupled
c
AGP, PAC, JJA, coupled
c
AGP, PAC, JJA, coupled
c
AGP, PAC, JJA, coupled
c
AGP, PAC, JJA, coupled
c
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5
CMIP
6
ERA/
IV
atm. grid spacing at 50°N (km)
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
9.0
90th
per
cent
ile b
lock
ing
pers
isten
ce (d
ays) AGP, PAC, JJA, AMIP
d
AGP, PAC, JJA, AMIP
d
AGP, PAC, JJA, AMIP
d
AGP, PAC, JJA, AMIP
d
AGP, PAC, JJA, AMIP
d
AGP, PAC, JJA, AMIP
d
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
Figure S12. As Fig. S9 but for the AGP index, for the PAC domain (90-270E, 50-75N), and for boreal summer. (Persistence of blockingevents (a,b - median, c,d - 90th percentile). The left-hand side of each panel shows metrics for PRIMAVERA simulations at different gridspacings (resolutions). Boxplots on the right-hand side show distributions of the persistence metric across CMIP5 and CMIP6 simulations.The ’*’ symbol in the column ’ERA/IV’ shows the reanalysis estimate and the boxplot is an estimate of the expected agreement given internalvariability (see text).)
16 Schiemann et al.: Blocking in GCMs (Supplement)
0 25 50 75 100
125
CMIP
5
CMIP
6
ERA/
IV
atm. grid spacing at 50°N (km)
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
med
ian
bloc
king
per
siste
nce
(day
s)
ANOM, PAC, DJF, coupled
a
ANOM, PAC, DJF, coupled
a
ANOM, PAC, DJF, coupled
a
ANOM, PAC, DJF, coupled
a
ANOM, PAC, DJF, coupled
a
ANOM, PAC, DJF, coupled
a
ANOM, PAC, DJF, coupled
a
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5
CMIP
6
ERA/
IV
atm. grid spacing at 50°N (km)
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
med
ian
bloc
king
per
siste
nce
(day
s)
ANOM, PAC, DJF, AMIP
b
ANOM, PAC, DJF, AMIP
b
ANOM, PAC, DJF, AMIP
b
ANOM, PAC, DJF, AMIP
b
ANOM, PAC, DJF, AMIP
b
ANOM, PAC, DJF, AMIP
b
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5
CMIP
6
ERA/
IV
atm. grid spacing at 50°N (km)
6
8
10
12
14
16
90th
per
cent
ile b
lock
ing
pers
isten
ce (d
ays) ANOM, PAC, DJF, coupled
c
ANOM, PAC, DJF, coupled
c
ANOM, PAC, DJF, coupled
c
ANOM, PAC, DJF, coupled
c
ANOM, PAC, DJF, coupled
c
ANOM, PAC, DJF, coupled
c
ANOM, PAC, DJF, coupled
c
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5
CMIP
6
ERA/
IV
atm. grid spacing at 50°N (km)
6
8
10
12
14
16
90th
per
cent
ile b
lock
ing
pers
isten
ce (d
ays) ANOM, PAC, DJF, AMIP
d
ANOM, PAC, DJF, AMIP
d
ANOM, PAC, DJF, AMIP
d
ANOM, PAC, DJF, AMIP
d
ANOM, PAC, DJF, AMIP
d
ANOM, PAC, DJF, AMIP
d
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
Figure S13. As Fig. S9 but for the PAC domain (120-240E, 40-90N). (Persistence of blocking events (a,b - median, c,d - 90th percentile) forthe ANOM index and boreal winter. The left-hand side of each panel shows metrics for PRIMAVERA simulations at different grid spacings(resolutions). Boxplots on the right-hand side show distributions of the persistence metric across CMIP5 and CMIP6 simulations. The ’*’symbol in the column ’ERA/IV’ shows the reanalysis estimate and the boxplot is an estimate of the expected agreement given internalvariability (see text).)
Schiemann et al.: Blocking in GCMs (Supplement) 17
0 25 50 75 100
125
CMIP
5
CMIP
6
ERA/
IV
atm. grid spacing at 50°N (km)
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
med
ian
bloc
king
per
siste
nce
(day
s)
ANOM, PAC, JJA, coupled
a
ANOM, PAC, JJA, coupled
a
ANOM, PAC, JJA, coupled
a
ANOM, PAC, JJA, coupled
a
ANOM, PAC, JJA, coupled
a
ANOM, PAC, JJA, coupled
a
ANOM, PAC, JJA, coupled
a
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5
CMIP
6
ERA/
IV
atm. grid spacing at 50°N (km)
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
med
ian
bloc
king
per
siste
nce
(day
s)
ANOM, PAC, JJA, AMIP
b
ANOM, PAC, JJA, AMIP
b
ANOM, PAC, JJA, AMIP
b
ANOM, PAC, JJA, AMIP
b
ANOM, PAC, JJA, AMIP
b
ANOM, PAC, JJA, AMIP
b
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5
CMIP
6
ERA/
IV
atm. grid spacing at 50°N (km)
4
6
8
10
12
14
90th
per
cent
ile b
lock
ing
pers
isten
ce (d
ays) ANOM, PAC, JJA, coupled
c
ANOM, PAC, JJA, coupled
c
ANOM, PAC, JJA, coupled
c
ANOM, PAC, JJA, coupled
c
ANOM, PAC, JJA, coupled
c
ANOM, PAC, JJA, coupled
c
ANOM, PAC, JJA, coupled
c
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
0 25 50 75 100
125
CMIP
5
CMIP
6
ERA/
IV
atm. grid spacing at 50°N (km)
4
6
8
10
12
14
90th
per
cent
ile b
lock
ing
pers
isten
ce (d
ays) ANOM, PAC, JJA, AMIP
d
ANOM, PAC, JJA, AMIP
d
ANOM, PAC, JJA, AMIP
d
ANOM, PAC, JJA, AMIP
d
ANOM, PAC, JJA, AMIP
d
ANOM, PAC, JJA, AMIP
d
AWICMCCCNRMEC-EarthECMWFMOHCMPI-M
Figure S14. As Fig. S9 but for the PAC domain (120-240E, 40-90N) and boreal summer. (Persistence of blocking events (a,b - median,c,d - 90th percentile) for the ANOM index. The left-hand side of each panel shows metrics for PRIMAVERA simulations at different gridspacings (resolutions). Boxplots on the right-hand side show distributions of the persistence metric across CMIP5 and CMIP6 simulations.The ’*’ symbol in the column ’ERA/IV’ shows the reanalysis estimate and the boxplot is an estimate of the expected agreement given internalvariability (see text).)
E P C E P C E P C E P C E P C E P C E P C E P C E P C E P C E P C E P C0.8
0.6
0.4
0.2
0.0
0.2
0.4
0.6
0.8
rel.
bloc
king
bia
s
ATL, DJF ATL, JJA PAC, DJF PAC, JJA
CMIP5CMIP6PRIMAVERA
Figure S15. Decomposition of blocking bias into (i) a term due to the bias in the number of events (E), (ii) a term due to the bias inpersistence (P), and (iii) a cross term (C). Boxplots show the ensemble spread across the CMIP5, CMIP6, and all coupled PRIMAVERAmodels, respectively. The AGP index with 5-day persistence is used and results are shown for the ATL and PAC domains, and for borealwinter (DJF) and summer (JJA). The blocking bias is normalised by the reanalysis estimate of blocking frequency.
18 Schiemann et al.: Blocking in GCMs (Supplement)
Table S1. CMIP5 simulations used in this study. The r1i1p1 ensemble member from the historical experiment is used for each simulation.
No. Centre Model Period
1 CSIRO-BOM ACCESS1-0 1950–20052 CSIRO-BOM ACCESS1-3 1950–20053 BCC BCC-CSM1-1 1950–19974 BCC BCC-CSM1-1-M 1950–20055 BNU BNU-ESM 1950–20056 CCCma CanESM2 1950–20057 CMCC CMCC-CESM 1950–20058 CMCC CMCC-CM 1950–20059 CMCC CMCC-CMS 1950–200510 CNRM-CERFACS CNRM-CM5 1950–200511 EC-Earth-Consortium EC-EARTH 1950–200512 LASG-CESS FGOALS-g2 1950–200513 NOAA-GFDL GFDL-CM3 1950–200514 NOAA-GFDL GFDL-ESM2G 1950–200515 MOHC HadCM3 1960–200516 MOHC HadGEM2-CC 1950–200517 MOHC HadGEM2-ES 1981–200518 IPSL IPSL-CM5A-LR 1950–200519 IPSL IPSL-CM5A-MR 1950–200520 IPSL IPSL-CM5B-LR 1950–200521 MIROC MIROC5 1950–200522 MIROC MIROC-ESM-CHEM 1950–200523 MIROC MIROC-ESM 1950–200524 MPI-M MPI-ESM-LR 1950–200525 MPI-M MPI-ESM-MR 1950–200526 MPI-M MPI-ESM-P 1950–200527 MRI MRI-CGCM3 1950–200528 MRI MRI-ESM1 1950–200529 NCC NorESM1-M 1950–2005
Table S2. CMIP6 simulations used in this study. All simulations are for the historical experiment during 1950–2014.
No. Centre Model Ensemble member
1 BCC BCC-CMS2-MR r1i1p1f12 BCC BCC-ESM1 r1i1p1f13 CCCma CanESM5 r1i1p1f14 CNRM-CERFACS CNRM-CM6-1 r2i1p1f25 CNRM-CERFACS CNRM-ESM2-1 r3i1p1f26 IPSL IPSL-CM6A-LR r1i1p1f17 MOHC HadGEM3-GC31-LL r1i1p1f38 MOHC UKESM1-0-LL r1i1p1f29 MRI MRI-ESM2-0 r1i1p1f110 NASA-GISS GISS-E2-1-G r1i1p1f111 NCAR CESM2 r1i1p1f112 NCAR CESM2-WACCM r2i1p1f113 NOAA-GFDL GFDL-CM4 r1i1p1f1