supplement of evaluation of the cmip5 models in the aim of … · 2016-01-11 · by cpi scores for...
TRANSCRIPT
Supplement of The Cryosphere, 9, 2311–2321, 2015http://www.the-cryosphere.net/9/2311/2015/doi:10.5194/tc-9-2311-2015-supplement© Author(s) 2015. CC Attribution 3.0 License.
Supplement of
Evaluation of the CMIP5 models in the aim of regional modelling of theAntarctic surface mass balance
C. Agosta et al.
Correspondence to:C. Agosta ([email protected])
The copyright of individual parts of the supplement might differ from the CC-BY 3.0 licence.
NCEP-NCAR-v1
MIR
OC-ESM
-CH
EM
MPI-ESM-LRH
adG
EM
2-C
C
BCC-C
SM1-
1
FGOALS-g2
CESM
1-B
GC
CMCC-CESM
CMCC-CMS
GIS
S-E2
-RCC
GISS-E2-HCC
Nor
ESM
1-M
GFD
L-C
M3
JRA-5
5
NorE
SM
1-M
E
CESM
1-1-
FV2
HadGEM
2-ES
ERA-Interim
CSIR
O-M
k360CM
CC-CM
INM
-CM
4
CCSM4
MPI-ESM-MR
GISS-E2-H
IPSL-CM5B-LR
AC
CESS1-3
ACCESS1-0
NOAA-20CR-v2
CESM1-CAM5
GFDL-ESM2MBNU-ESM
MERRA-v1
MRI-ESM1
MRI-CGCM3
CN
RM
-CM
5
GFD
L-ESM2G
1
FIO-E
SM
Can
ESM
2
HadG
EM
2-A
O
IPSL-CM5A-MR
NCEP-DOE-v2
MIROC-ESM
BC
C-C
SM
1-1
-m
MIR
OC5
0
GISS-E2-R
EC
-EAR
TH
0.5
IPSL-
CM
5A-L
R
tos[sum]
HadG
EM
2-A
O
NorESM1-ME
CSIR
O-M
k360
GISS-E2-HINM-CM4
NCEP-NCAR-v1
NOAA-20CR-v2
JRA-55
CMCC-CM
CESM
1-1
-FV
2
CESM1-BGC
MRI-CGCM3
0.5
AC
CESS1-3
FIO-ESM
CanESM
2
CMCC-CESM
GFDL-ESM2M
MIROC5
NCEP-DOE-v2
CM
CC
-CM
S
Had
GEM
2-C
C
CCSM4
BNU-ESM
CESM
1-CA
M5
FGOALS-g2
MPI-E
SM
-MR
MIROC-ESM
BCC-
CSM
1-1-
m
1
MPI-ESM
-LR
Had
GEM
2-E
SG
ISS-E
2-H
CC
AC
CESS1-0
GIS
S-E
2-R
MRI-ESM1
IPSL
-CM
5A-L
R
EC
-EA
RTH
BCC-CSM1-1
NorES
M1-
M
IPSL-CM5B-LR
0
GFD
L-ES
M2G
ERA-InterimMERRA-v1
GFD
L-CM3
GIS
S-E
2-R
CC
MIR
OC-ESM
-CH
EM
IPSL-CM5A-MR
CNRM-C
M5
msie[win]
CM
CC
-CESM
AC
CESS1-3
1
NCEP-DOE-v2
NorESM1-ME
MIROC-ESM-CHEM
MRI-CGCM3
BC
C-C
SM
1-1
-m
EC-EARTH
MRI-ESM1
FGOALS-g2
CCSM4
Can
ESM
2
MERRA-v1
0.5
CMCC-CM
GFDL-ESM2G
ERA-Interim
CSI
RO
-Mk3
60
GFDL-CM3
MPI-E
SM
-LR
GIS
S-E2
-RCC
HadG
EM
2-E
S
IPSL-C
M5A
-MR
JRA-55
0
NCEP-NCAR-v1
CESM
1-1-
FV2
BNU-ESM
NOAA-20C
R-v2
MPI-ESM-MR
CESM1-BGC
GISS-
E2-R
CNRM-CM5
INM
-CM
4
GFD
L-ESM2M
BC
C-C
SM
1-1
MIR
OC5
NorE
SM
1-M
CES
M1-
CAM
5
IPSL-CM5A-LR
HadGEM2-CC
MIROC-ESM
FIO-ESM
HadG
EM
2-A
O
AC
CESS1-0
GIS
S-E
2-H
CC
IPSL-CM5B-LR
GIS
S-E
2-H
CM
CC
-CM
S
psl[win]
psl[sum]
psl[aut]
psl[spr]
MPI-ESM-M
R
CanESM2
GIS
S-E
2-H
CC
CMCC-CM
MIROC-E
SM-CHEM
HadGEM2-AO
GFD
L-CM
3
ACCE
SS1-
0
CSIRO-Mk360
MRI-ESM1
CESM1-CAM5
NC
EP-N
CA
R-v
1MPI-ESM-LR
HadG
EM
2-E
S
NorESM1-ME
MERRA-v1
ACCES
S1-3
GISS-E2-RCC
INM-CM
4
CESM
1-1
-FV2
CNRM-CM5
IPSL-C
M5
A-M
R
NorES
M1-
M
FGO
ALS
-g2
GFD
L-ESM
2G
ERA-Interim
HadG
EM
2-C
C
GIS
S-E2
-H
IPSL-CM5A-LR
BC
C-C
SM
1-1
FIO-ESM
JRA-55
0.5
IPSL-CM5B-LRGFDL-ESM2M
NCEP-D
OE-v2
MRI-CGCM3
EC-EARTH
BC
C-C
SM
1-1
-m
CM
CC
-CESM
CM
CC
-CM
S
GISS-E2-R
BN
U-E
SM
MIR
OC-ESM
0
CESM1-BGC
MIRO
C5
NOAA
-20C
R-v2
CCSM4
1prw[sum]
prw[win]
EC
-EA
RTH
CMCC
-CES
M
1
BNU-ESM
CESM1-CAM5M
PI-ESM-M
R
FIO
-ESM
NorESM1-M
E
MIROC-ESMMPI-ESM-LR
JRA-55
CNRM-C
M5
CSI
RO
-Mk3
60 MIR
OC
5
NCE
P-DOE-
v2
GFDL-E
SM2G
AC
CESS1-0
MR
I-ESM
1
CESM1-1-FV2
IPSL-CM5A-M
R
HadGEM2-AO
MERRA-v1
0
ERA-Interim
GFDL-CM3
CMCC-CM
ACCESS1-3
CanESM
2
0.5
MIROC-ESM-CHEM
GISS-E2-RCC
IPSL-CM5B-LR
GIS
S-E
2-R
NOAA-20C
R-v2
NorESM1-M
HadG
EM2-ES
INM-CM4
FGO
ALS
-g2
CC
SM
4
IPSL-CM5A-LR
GISS-E2-H
CC
CM
CC
-CM
S
NCEP-NCAR-v1
HadGEM2-CC
BC
C-C
SM
1-1
GISS-E2-H
MR
I-C
GC
M3
BCC-C
SM1-
1-m
CESM
1-B
GC
GFD
L-ESM
2M
ta850[sum]
ta850[win]
HadGEM2-CC
NorESM1-M MRI-ESM1
CNRM-CM5
GFDL-CM3
CES
M1-
BG
C
NOAA-20C
R-v2
MPI-ESM-LR
MIR
OC-E
SM
GFD
L-ESM
2G
HadG
EM
2-E
S
JRA-5
5
GFDL-ESM2M
GISS-E2-H
AC
CESS1-0
MIROC5
0.5
EC-EARTH
ERA-Interim
BC
C-C
SM
1-1
IPSL-CM5A-LR
IPSL-CM5B-LR
CanE
SM2
NorESM1-M
E
MIR
OC-ESM
-CHEM
HadG
EM2-A
O
0
BC
C-C
SM
1-1
-m
NCEP-NCAR-v1
1
CSIR
O-M
k360
CMCC-CM
S
CMCC-CM
CESM1-1-FV2
GIS
S-E
2-R
INM
-CM
4
CC
SM
4
FIO
-ESM
IPSL-CM5A-MR
MPI-ESM
-MR
CESM1-CAM5
MERRA-v1
FGO
ALS
-g2
BNU-ESM
NCE
P-DOE-
v2
GIS
S-E
2-H
CC
CM
CC
-CESM
MRI-C
GCM
3
GIS
S-E
2-R
CC
ACCESS1-3
ta850[sum]
Figure S1: CPI scores, computed as exp(-CPI2/2), range from 1 (best, external circle) to 0 (nulprobability of belonging to the ERA-Interim distribution, internal circle). Models are rankedby CPI scores for each variable: summer sea surface temperature, winter meridional sea-iceextent, annual sea-level pressure, summer/winter precipitable water, summer/winter 850hPaair temperature, and summer 850hPa air temperature. Markers are for seasonal CPI scoresand solid black line is for combined CPI scores. Blue lines are upper and lower bounds for thecombined score taking into account multi-decadal variabilities of seasonal components. The greycrown width is the combination of 90th percentiles of CMIP5 GCMs multi-decal variabilities.Models with obvious similarities in code or produced by the same institution are marked withthe same color (clusters), following Knutti et al. (2013).
2
Figure S2: Same as Fig. 1 but for summer sea surface temperature (in °C).
3
Figure S3: Same as Fig. 1 but for winter sea-ice concentration.
4
Figure S4: Same as Fig. 1 but for summer precipitable water (in kg m-2).
5
Figure S5: Same as Fig. 1 but for winter precipitable water (in kg m-2).
6
Figure S6: Same as Fig. 1 but for summer air temperature at 850hPa (in °C).
7
Figure S7: Same as Fig. 1 but for winter air temperature at 850hPa (in °C).
8
0.0 0.5 1.0 1.5 2.0 2.5 3.00.0
0.5
1.0
1.5
2.0
2.5
3.0
R2=0.95
y=1.09x+0.0
crmse psl[ann]
crm
se z
g500[a
nn]
0.0 0.5 1.0 1.5 2.0 2.5 3.00.0
0.5
1.0
1.5
2.0
2.5
3.0
R2=0.94
y=0.95x+0.1
crmse psl[ann]
crm
se z
g250[a
nn]
Figure S8: The spatial centered rmse (crmse) is computed for 1980-2010 mean values of modelsvs. ERA-Interim and scaled by spatially-averaged standard deviation of ERA-Interim annualvalues. Here we compare annual geopotential heights at 500hPa (zg500[ann], left) and at 250hPa(zg250[ann], right) without masking ice-sheet areas to annual sea-level pressure over the ocean(psl[ann], X-axis). See Fig. 3 for chart details.
9
18601880190019201940196019802000202020402060208021000123456789
msi
e[w
in]
rmse
186018801900192019401960198020002020204020602080210002468
1012
tos[
sum
]
18601880190019201940196019802000202020402060208021000.00.51.01.52.02.53.0
psl
[ann]
1860188019001920194019601980200020202040206020802100012345678
prw
[s/w
]
18601880190019201940196019802000202020402060208021000123456
ta8
50
[s/w
]
18601880190019201940196019802000202020402060208021000123456
ta8
50
[sum
]
Figure S9: 31-years time-averages running CPIs between 1950 and 2100. The field of referenceis ERA-Interim for the period 1980-2010. Thin light blue lines are for the 41 CMIP5 GCMs(historical+RCP85) and thick dark grey lines are for the reanalyses. Three GCMs are highlightedwith thick colored lines: ACCESS1-3 in purple, CESM1-CAM5 in dashed red and NorESM1-Min solid red. Each point of a line is the central year of the 31-year period.
10
0 1 2 3
0
1
2
3psl
[ann]
am
ip
psl[ann] historical
rmse
0 1 2 3 4 5
0
1
2
3
4
5
prw
[s/w
] am
ip
prw[s/w] historical
0 1 2 3
0
1
2
3
ta850[s
/w]
am
ip
ta850[s/w] historical
ACCESS1-0
ACCESS1-3
BCC-CSM1-1
BCC-CSM1-1-m
BNU-ESM
CCSM4
CESM1-CAM5
FGOALS-g2
NorESM1-M
CSIRO-Mk3-6-0
CMCC-CM
MPI-ESM-LR
MPI-ESM-MR
CNRM-CM5
EC-EARTH
GFDL-CM3
GISS-E2-R
INM-CM4
MRI-CGCM3
IPSL-CM5A-LR
IPSL-CM5A-MR
IPSL-CM5B-LR
MIROC5
Figure S10: Comparison of Amip (y-axis) versus Historical (x-axis) for the six selected metrics.For Amip, the averaged period is 1980-2008 (29 years) whereas for ERA-Interim and Historicalruns the averaged period is 1980-2010 (31 years). See Fig. 3 for chart details.
11