Embed Size (px)
Citation preview
Classifying simulated wheat yield responses to changes in
temperature and precipitation across a European transect
S. Fronzek 1, N. Pirttioja1, T. R. Carter1, M. Bindi, H. Hoffmann, T. Palosuo, M. Ruiz-Ramos, F. Tao, M. Trnka, M. Acutis, S. Asseng, P. Baranowski, B. Basso, P. Bodin, S. Buis, D. Cammarano, P. Deligios, M.-F.
Destain, B. Dumont, F. Ewert, R. Ferrise, L. François, T. Gaiser, P. Hlavinka, I. Jacquemin, K. Christian Kersebaum, C. Kollas, J. Krzyszczak, I. J. Lorite, J. Minet, M. I. Minguez, M. Montesino, M. Moriondo, C. Müller, C. Nendel, I. Öztürk, A. Perego, A. Rodríguez, A. C. Ruane, F.
Ruget, M. Sanna, M. Semenov, C. Slawinski, P. Stratonovitch, I. Supit, K. Waha, E. Wang, L. Wu, Z. Zhao, R. P. Rötter
1Finnish Environment Institute (SYKE)
● Crop modelling experiment in MACSUR/CropM/WP4
Aims:
● To study crop model sensitivity to changes in precipitation and
temperature using a large ensemble of crop models across a
transect
● To quantify differences in winter and spring wheat yield
responses to changed climate across models
● By plotting results of the sensitivity analysis as impact
response surfaces (IRSs)
Aims
MATERIAL AND METHODS
Ensemble of 26 wheat models
Model Modelling groups
Contact person(s) Institute Country
AFRCWHEAT2 Manuel Montesino University of Copenhagen Denmark
APSIM-Nwheat Senthold Asseng, Davide Cammarano University of Florida USA
APSIM-Wheat Enli Wang CSIRO Land and Water Australia
AquaCrop Ignacio Lorite IFAPA Junta de Andalucia Spain
ARMOSA Alessia Perego University of Milan Italy
CARAIB Crop Julien Minet Université de Liège Belgium
CERES-wheat DSSAT v.4.6 Mirek Trnka, Petr Hlavinka Mendel University in Brno Czech Republic
CERES-wheat DSSAT v.4.5 Margarita Ruiz-Ramos Universidad Politecnica de Madrid Spain
CERES-wheat DSSAT v.4.5 Paola Deligios University of Sassari Italy
CropSyst Marco Moriondo,
Roberto Ferrise, Marco Bindi
CNR-IBIMET
University of Florence
Italy
Italy
DNDC Cezary Slawinski; Piotr Baranowski Polish Academy of Sciences Poland
Fasset Isk Öztürk Aarhus University Denmark
HERMES Chris Kollas, Christian Kersebaum Leibniz Centre for Agric. Landscape Research (ZALF) Germany
Lintul4 Iwan Supit Wageningen University Netherlands
LPJ-GUESS Per Bodin Lund University Sweden
LPJml Christoph Müller Potsdam Institute for Climate Impact Research Germany
MCWLA Fulu Tao Luke Natural Resources Institute Finland Finland
MONICA V1.2 Claas Nendel Leibniz Centre for Agric. Landscape Research (ZALF) Germany
SALUS Bruno Basso Michigan State University USA
SIMPLACE<Lintul2, Slim> Holger Hoffmann, Thomas Gaiser, Frank Ewert University of Bonn Germany
Sirius 2010 Mikhail Semenov, Pierre Stratonovitch Rothamsted Research UK
Sirius Quality Roberto Ferrise, Marco Bindi University of Florence Italy
SPACSYS Lianhai Wu Rothamsted Research UK
STICS Benjamin Dumont, Françoise Ruget, Samuel Buis Université de Liège & INRA EMMAH Belgium & France
WOFOST 7.1 Cezary Slawinski; Jaromir Krzyszczak Polish Academy of Sciences Poland
WOFOST 7.1 Taru Palosuo, Reimund Rötter Luke Natural Resources Institute Finland Finland
Locations of weather stations used in this study and
environmental zones of Metzger et al. (2005)
Study sites across a European transect
Mainly
precipitation
limited
High
current
suitability
Mainly
temperature
limited
30°E
30°E
20°E
20°E
10°E
10°E
0°
0°
10°W
10°W20°W
70°N 70°N
60°N 60°N
50°N 50°N
40°N 40°N
Environmental zones
Boreal
Continental
Atlantic Central
Mediterranean South
Other
Wheat cultivation area
0 500250 km
Jokioinen
NossenDikopshof
Lleida
Co-ordinate system: World Robinsoncentral meridian: 30°0'0''E
● Each group calibrated their model independently
● Limited data for calibration was provided (crop phenology and
yield; soil conditions; fertilisation, tillage and irrigation where available)
● Model simulations were performed
○ for water-limited yields
○ assuming optimal nutrients
● Error checking and model iteration
● Several output variables: annual grain yield, biomass,
phenology, cumulated water use, nitrogen content of yield
Simulation set-up (1/2)
Simulation set-up (2/2)
Sites Country Location N
Finland Jokioinen
Germany Dikopshof (winter wheat), Nossen (spring wh.) 3
Spain Lleida
Crops Crop /Cultivar type Cultivar
2 Spring wheat Different cultivar for each location
Winter wheat Different cultivar for each location
Baseline Harvest years 1981-2010 30
Perturbations Variable Min Max Interval
Precipitation (%) - 50 + 50 10 11
Temperature (°C) - 2 + 9 1 12
CO2 level 360 ppm (Year 1995) 1
Soils Clay loam 1
Management Fixed sowing date Location specific (observed) 1
Total number of simulations Sites x crops x years x P-changes x T-changes 23760
IRSs represent the sensitivity of modelled crop yield to incremental
changes in precipitation (vertical) and temperature (horizontal)
Impact response surface (IRS) of
a single crop model for spring
wheat yield, Germany, 2008
-2 0 2 4 6 8
-40
-20
020
40
DE, S_wheat, 2008, ARMOSA
Temperature change ( C)
Pre
cip
itatio
n c
hange (
%)
750
1500
2250
3000
3750
4500
5250
6000
6750
7500
8250
750 1500
2250
3000
3750
4500
5250
6000
6750
7500
8250
750
1500
2250
3000
3750
4500
5250
6000
67507500
8250
9000
9750
10500
11250
12000
12750
13500
14250
15000
15750
16500
17250
18000
18750
19500
Grain yield kg/ha
Baseline
Temperature change (°C)
Pre
cip
itation c
hange (
%)
RESULTS I:
ENSEMBLE AVERAGES
AND RANGES
Simulated yields for the baseline 1981-2010 1
2
3
4
Finland, Spring wheat
Harvest year
Yie
ld (
kg/h
a)
1985
1990
1995
2000
2005
2010
0
2000
4000
6000
8000
10000
12000
(a) Finland, Winter wheat
Harvest year
Yie
ld (
kg/h
a)
1985
1990
1995
2000
2005
2010
0
2000
4000
6000
8000
10000
12000
(b)
Germany, Spring wheat
Harvest year
Yie
ld (
kg/h
a)
1985
1990
1995
2000
2005
2010
0
2000
4000
6000
8000
10000
12000
(c) Germany, Winter wheat
Harvest year
Yie
ld (
kg/h
a)
1985
1990
1995
2000
2005
2010
0
2000
4000
6000
8000
10000
12000
(d)
Spain, Spring wheat
Harvest year
Yie
ld (
kg/h
a)
1985
1990
1995
2000
2005
2010
0
2000
4000
6000
8000
10000
12000
(e) Spain, Winter wheat
Harvest year
Yie
ld (
kg/h
a)
1985
1990
1995
2000
2005
2010
0
2000
4000
6000
8000
10000
12000
(f)
Individual model results
Ensemble median
Historical yields of wheat Finland: FAO Country level
statistics
Germany: Eurostat regional
statistics
Spain: provincial statistics for
northern Spain, Spanish
Ministry of Agriculture
Calibration data
Finland
Germany
Spain
Spain, Spring wheat
Harvest year
Norm
alis
ed y
ield
1985
1990
1995
2000
2005
2010
-4
-2
0
2
4
(b) Spain, Spring wheat
Harvest year
Norm
alis
ed y
ield
1985
1990
1995
2000
2005
2010
-4
-2
0
2
4
(b)
Winter wheat DM grain yields, Germany
1981-2010 mean
6500
7000
7500 8000
8500
9000
−1°C 1°C 3°C 5°C 7°C 9°C-50
%-3
0%
-10
%1
0%
30
%5
0%
1981
7000
70
00
7500
800
0 8
500
900
0
9000
9500
1982
5500 6000
6500
7000
7500
800
0 8500
1983
7000
70
00
75
00
80
00
8500
90
00
1984
7000
7500
8000
8500
9000
9500
1985
5500
600
0
6500
7500
8000
8500
900
0
9500
1986
6500 7000
75
00
8000
9000
95
00
1987
75
00
80
00
8500
90
00
9500
9500
9500
1988
70
00
7500
75
00
8000
8500
9000
9500
1989
6000
70
00
7500 8000
8000 8500
9000
950
0
1990
6000
7000
70
00
7500
8000 8500
900
0
1991
6500 7000
7500
8000 8500
900
0
9500
1992
7000
70
00
7500
8000
850
0 9
000
9500
1993
6000
6500
7000 7500
8000
8500
950
0
1994
60
00
6500
6500
70
00
7500
80
00
8500
8500
1995
75
00
8000 8000
85
00
1996
5500
6000
6500
7000
7500
8500
1997
6500
65
00
7000
75
00
8000
8500
9000
9000
9500
1998 5
50
0
600
0
65
00
7500
8000
85
00
90
00
1999
60
00
70
00
7500
80
00
8500
9000
95
00
2000
6000
650
0
7000
7500
8500
9000
2001
65
00
70
00
7
50
0
8500
85
00
9000 9500
2002
70
00
7500
75
00
8500
9000
2003
600
0
650
0
6500
7000
7500
8000
8500
9000
2004
6500
7000
7500
8000
8500
950
0
2005
60
00
7000
70
00
7500
8000
8500
2006
5000 5500
6500
7000 7500
8000
8500
2007
6500
7000
7500
800
0
8500
2008
7000
70
00
7500
80
00
9000 9
500
9500
2009
65
00
700
0
7500
75
00
8000
85
00
95
00
2010
5500 6000
650
0
7000
7500
8000 8500
4000
4500
5000
5500
6000
6500
7000
7500
8000
8500
9000
9500
kg/ha
Crosses in the 30-year mean
plot: changes in annual
temperature and precipitation
projected by the CMIP5
ensemble of 36 global climate
models for RCP8.5 over central
Europe by 2070-2099 relative to
1981-2010.
One crop model,
individual years 1981-
2010 (small sub-plots)
and 30-year mean
(larger sub-plot)
Yield changes relative to unperturbed baseline
30-year average change
in winter wheat DM
yields relative to
baseline climate (1981-
2010) in Germany
26 models (small sub-
plots) and ensemble
median (larger sub-plot)
By definition, the yield
change is 0% for the
baseline climate at the
intersection of the grey
lines.
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
10
20
30
40
%
Ensemble median
-50
-40
-30
-20
-10
0
−1°C 1°C 3°C 5°C 7°C 9°C-50
%-3
0%
-10
%1
0%
30
%5
0%
1
-30
-20
-10
0
2
-80
-70
-60
-50
-40
-20
-10
0
3
-70
-60
-50
-40
-30
-
20
-10
0
10
4
-20
-10
10
20
5
-40
-30
-10
0
6
-60
-50
-40
-
30
-20
-10
0
7
-90
-80
-70
-60
-40
-30
-20
-10
0
8
-50 -30
-20
-10
0
9
-40 -30
-20 -10
0
10
10
-50 -40 -30
-20
-10
0
20
11
-50
-40
-30
-20
-10
0
10
12
-30
-20
-10
0
13
-80
-60 -50
-40
-30
-20
-10
0
10
20
14
-90 -8
0
-70
-50
-40
-30
-20
0
10
20
15
-60
-50
-40
-30
-
20
-10
0
10
20
30
16
-70 -60
-50
-40
-30 -20
-10
0
10
20
30
17
-40
-30
-20
-10
0
10 20
18
-70 -50
-40
-30
-20
-10
0
19
-90
-80
-70
-60
-50
-40
-30
-20
0
10
20
-30
-20
-10
0
21
-30
-20
-10
0
22
-50 -40
-30 -20
-10
0
10
23
-50
-40
-30
-20
-10
0
24
0
25
-60 -50
-40
-30
-20
-10
0
10
26
-80
-70
-6
0
-50
-4
0
-30
-2
0
-10
0
Ensemble medians and IQR of yield changes
Winter wheat
Left: Median of yield
changes by 26 crop
models
Right: Inter-quartile range
(IQR) of relative
responses scaled to
100% at baseline
The ensemble median (Mbaseline) and
ensemble inter-quartile range
(IRQbaseline) of absolute yields for the
baseline are listed above each plot.
-2 0 2 4 6 8
-40
-20
020
40
(a) Finland, Mbaseline = 5155 kg/ha
Temperature change (°C)
Pre
cip
itation c
hange (
%)
-40
-30
-30
-20
-10
0
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
10
20
30
40
%
-2 0 2 4 6 8
-40
-20
020
40
(b) Finland, IQRbaseline = 1277 kg/ha
Temperature change (°C)
Pre
cip
itation c
hange (
%)
100 120
120
140
140
160
160
180
180
60
80
100
120
140
160
180
200
220
240
%
-2 0 2 4 6 8
-40
-20
020
40
(c) Germany, Mbaseline = 7995 kg/ha
Temperature change (°C)
Pre
cip
itation c
hange (
%)
-50 -40 -30
-20 -10
0
-2 0 2 4 6 8
-40
-20
020
40
(d) Germany, IQRbaseline = 1341 kg/ha
Temperature change (°C)
Pre
cip
itation c
hange (
%)
100
120
120
140 160
160
180 1
80
200
200
220
-2 0 2 4 6 8
-40
-20
020
40
(e) Spain, Mbaseline = 4005 kg/ha
Temperature change (°C)
Pre
cip
itation c
hange (
%)
-50 -40
-30
-20 -10
0
10
20
30
-2 0 2 4 6 8
-40
-20
020
40
(f) Spain, IQRbaseline = 2165 kg/ha
Temperature change (°C)
Pre
cip
itation c
hange (
%)
60
80
80
100
100
120
RESULTS II:
CLASSIFICATION OF
MODEL RESPONSES
● Different models and different conditions show different
behaviour
Can we group models according to their different
sensitivities to climate change?
○ Two approaches for grouping IRSs of mean yield
change:
• Clustering algorithm
• Rules defined by expert judgment
Can we find explanations for different model responses?
Classification of responses
Clustering of IRSs based on correlation
and Euclidian distance
● Distances between two IRSs are defined based on their pattern and magnitude:
○ Pearson correlation coefficient r *(-1) +1
○ Euclidian distance over all points of the IRS
and combined by taking the product of the two
● IRSs are clustered (per crop, for 3 locations) by hierarchical clustering that minimize the distances between members of each cluster:
○ agnes (agglomerative nesting) algorithm in R (Kaufman &
Rousseeuw 1990), using the average method to determine clusters
○ The number of clusters was set to 7 (according to a rule of thumb = sqrt(n/2), but after removing ”outlying” IRSs that were in a separate cluster
Winter wheat
30-yr mean change in yield IRSs
Index o
f dis
tance
Cluster 1 Cluster 2 Cluster 4 3, 5-7
01
00
20
03
00
40
05
00
60
07
00
W_wheat/product, Agglomerative Coefficient = 0.99
AF
RC
WH
EA
T2
_D
ES
IRIU
S2
01
0_
FI
SP
AC
SY
S_
DE
SP
AC
SY
S_
ES
FA
SS
ET
_D
EF
AS
SE
T_
FI
CA
RA
IB_
FI
SP
AC
SY
S_
FI
WO
FO
ST
/FI_
ES
WO
FO
ST
/PL
_E
SA
PS
IM_
DE
SA
LU
S_
FI
AP
SIM
_F
IW
OF
OS
T/P
L_
DE
DN
DC
_F
IA
PS
IM-N
WH
EA
T_
DE
SA
LU
S_
DE
CA
RA
IB_
DE
DN
DC
_D
ED
ND
C_
ES
EP
IC_
FI
LIN
TU
L2
_D
EC
ER
ES
/IT
_E
SE
PIC
_D
EA
FR
CW
HE
AT
2_
ES
MC
WL
A_
DE
SIR
IUS
QU
AL
ITY
_D
EM
CW
LA
_F
IS
IRIU
SQ
UA
LIT
Y_
FI
AR
MO
SA
_D
EA
PS
IM_
ES
CE
RE
S/C
Z_
DE
FA
SS
ET
_E
SA
PS
IM-N
WH
EA
T_
ES
SIR
IUS
QU
AL
ITY
_E
SS
TIC
S_
ES
CE
RE
S/C
Z_
ES
AR
MO
SA
_E
SM
ON
ICA
_E
SL
INT
UL
2_
ES
AP
SIM
-NW
HE
AT
_F
IS
IRIU
S2
01
0_
DE
CE
RE
S/IT
_D
EM
ON
ICA
_D
EM
ON
ICA
_F
IC
RO
PS
YS
T_
DE
CR
OP
SY
ST
_E
SW
OF
OS
T/F
I_D
ES
AL
US
_E
SH
ER
ME
S_
DE
HE
RM
ES
_E
SL
PJ-G
UE
SS
_E
SM
CW
LA
_E
SE
PIC
_E
SL
INT
UL
4_
FI
SIR
IUS
20
10
_E
SW
OF
OS
T/F
I_F
IC
AR
AIB
_E
SL
PJ-G
UE
SS
_D
EL
PJ-G
UE
SS
_F
IC
ER
ES
/ES
_E
SC
ER
ES
/ES
_D
EC
ER
ES
/ES
_F
IL
INT
UL
4_
DE
LIN
TU
L4
_E
SC
RO
PS
YS
T_
FI
HE
RM
ES
_F
IA
FR
CW
HE
AT
2_
FI
CE
RE
S/C
Z_
FI
ST
ICS
_F
IC
ER
ES
/IT
_F
IL
INT
UL
2_
FI
ST
ICS
_D
EA
QU
AC
RO
P_
DE
AQ
UA
CR
OP
_F
I
Cluster 1: strong temperature-sensitivity,
yield decreases with warming, no precip-
sens for high warming
1 AFRCWHEAT2_DE W_wheat
-30
-20
-10
0
1 APSIM-NWHEAT_DE W_wheat
-70
-60
-50
-40
-30
-20
-10
0
10
1 APSIM_DE W_wheat
-8
0
-7
0
-6
0
-50
-40
-20
-10
0
1 APSIM_FI W_wheat
-7
0
-6
0
-50
-30
-20
-20
-10
0
1 CARAIB_DE W_wheat
-60
-50
-40
-30
-20
-10
0
10
1 CARAIB_FI W_wheat
-60
-50
-40
-3
0
-20
-10
-10
0
1 CERES/IT_ES W_wheat
-40
-30
-20
-10
-10
0
0
1 DNDC_DE W_wheat
-50
-40
-30
-20
-10
0
10
1 DNDC_ES W_wheat -3
0
-20
-10
0
10
1 DNDC_FI W_wheat
-8
0
-70
-6
0
-5
0
-40
-30
-20
-20
-10
0
1 EPIC_DE W_wheat
0
10
20
30
40
1 EPIC_FI W_wheat
-30
-20
-10
1 FASSET_DE W_wheat
-30 -2
0
-10
0
1 FASSET_FI W_wheat
-4
0
-30
-20
-10
-10
0
1 LINTUL2_DE W_wheat
-30
-20 -10
0
1 SALUS_DE W_wheat
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
10
1 SALUS_FI W_wheat
-8
0
-50
-3
0
-20
-2
0
-10
0
1 SIRIUS2010_FI W_wheat
-30
-20
-10
0
1 SPACSYS_DE W_wheat
-50
-40
-30
-20
-10
0
1 SPACSYS_ES W_wheat
-40
-30
-20
-10
0
1 SPACSYS_FI W_wheat
-40
-30
-20
-10
0
1 WOFOST/FI_ES W_wheat
-60
-50
-40
-30
-20
-10
0
1 WOFOST/PL_DE W_wheat
-50
-30
-20 -10
0
1 WOFOST/PL_ES W_wheat
-4
0
-3
0
-20
-10
0
Temp. change (°C)
Pre
cip
ita
tio
n
ch
an
ge
(%
)
Central examples from each cluster
3 AFRCWHEAT2_FI W_wheat
-70
-6
0
0
20
30
40 4
0
3 CERES/CZ_FI W_wheat
-20
-10
0 0
10
20
30
3 STICS_FI W_wheat
-30
-20
-1
0
0
10
20
20
30
40
50
5 AQUACROP_DE W_wheat
-70 -40
-30 -10
0
10
20
5 AQUACROP_FI W_wheat
-60
-20
-10
0
10
20
30
40
50
Cluster 5: n=2
6 CERES/IT_FI W_wheat
-40
-30
-30
-20
-20
-10 -10
0
10
6 LINTUL2_FI W_wheat
-20 -
20
-10
-10
0
0
10
6 STICS_DE W_wheat
0
0
Cluster 6: n=3
7 CROPSYST_FI W_wheat
-80
-70
-60
-60
-50
-50
-40
-40
-30
-20
-10
0
10
7 HERMES_FI W_wheat
-70
-60
-60
-50
-50 -40
-30 -20
-10
0
10
20
Cluster 7: n=2 Separate: n=1
_ES _FI _FI
4 APSIM-NWHEAT_FI W_wheat
-50
-40
-30 -30
-20
-10
0
10
4 CARAIB_ES W_wheat
-70
-60
-50
-40
-30
-20
-10
0
10
20
30
4 CERES/ES_DE W_wheat
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
4 CERES/ES_ES W_wheat
-90 -8
0
-70
-60
-
50
-40
-
30
-20
-10
0
10
20
40
4 CERES/ES_FI W_wheat
-90
-
80
-70
-60
-50
-40
-30
-20
-10
0
10
20
4 CERES/IT_DE W_wheat
-50 -40
-30
-20
-10
0
0
4 CROPSYST_DE W_wheat
-50 -40
-30 -20
-10
0
10
20
4 CROPSYST_ES W_wheat
-50
-40 -30
-20
-10
0
10
20
4 EPIC_ES W_wheat
-60
-50
-40
-30
-20
4 HERMES_DE W_wheat
-80
-70 -60
-50
-40
-30
-20
-10
0
10
20
4 HERMES_ES W_wheat
-80
-70 -60
-50
-40
-30
-20
-10
0
10
20
30
40
4 LINTUL4_DE W_wheat
-90 -80
-70
-60
-50
-40
-30
-20
-10
0
10
20
30
40
4 LINTUL4_ES W_wheat
-90
-80
-70
-60
-50
-40
-30
-20
-10
0 1
0 2
0
30
40
50
60
100
4 LINTUL4_FI W_wheat
-80 -70
-60
-50
-40
-30
-20
-10
0
4 LPJ-GUESS_DE W_wheat
-60
-50
-40
-30
-20
-10
0
10
20
30
4 LPJ-GUESS_ES W_wheat
-70 -60
-50
-40
-30
-20
-10
0
10
20
30
4 LPJ-GUESS_FI W_wheat
-70
-60
-50
-40
-30
-20
-10
0
1
0
20
4 MCWLA_ES W_wheat
-80
-70
-60
-50
-40
-30
-20
-10
0
10
20
30
4 MONICA_DE W_wheat
-70 -60
-50
-40
-30
-20
-10
0
4 MONICA_FI W_wheat
-50 -40
-30
-20
-10
0
10
4 SALUS_ES W_wheat
-60
-50
-40
-30
-20
-10 0
10
20
4 SIRIUS2010_DE W_wheat
-30
-20
-10
0
4 SIRIUS2010_ES W_wheat
-50
-40
-30
-20
-10
0
10
4 WOFOST/FI_DE W_wheat
-50 -40
-30
-20
-10
0
10
20
4 WOFOST/FI_FI W_wheat
-50
-40
-30
-20
-10
0
10
2 AFRCWHEAT2_ES W_wheat
-50
-40
-30
-20
-10
0
10
2 APSIM-NWHEAT_ES W_wheat
-50 -50
-40
-30
-20
-10
0
10
20
30
40
2 APSIM_ES W_wheat
-40 -30 -30
-20
-10
0
10
20
2 ARMOSA_DE W_wheat
-50 -40
-30 -20
-10
0
2 ARMOSA_ES W_wheat
-80 -7
0
-60
-50
-40
-30
-20
-10
0
10
20
30
2 CERES/CZ_DE W_wheat
-40 -30
-20
-10
0
10
2 CERES/CZ_ES W_wheat
-60 -60 -50
-40 -30
-20
-10 0
10
20
30
40 50
2 FASSET_ES W_wheat
-40 -30 -20
-10
0
2 LINTUL2_ES W_wheat
-70
-60
-60
-50 -40
-30
-20 -10
0
10
20
30
40 50
2 MCWLA_DE W_wheat
-40
-30 -20
-10
0
10
20
2 MCWLA_FI W_wheat
-40 -30
-20 -10
0
10
2 MONICA_ES W_wheat
-70
-60
-50
-40 -30
-20
-10
0
10 20
30
40
50
2 SIRIUSQUALITY_DE W_wheat
-50 -50 -40
-30
-20
-10
0
10
2 SIRIUSQUALITY_ES W_wheat
-60 -50
-40
-30
-20
-10
0
10
20
30
2 SIRIUSQUALITY_FI W_wheat
-30 -30 -20
-10
0
10
2 STICS_ES W_wheat
-20 -10
0
10 20
Cluster 2: n=16
1 AFRCWHEAT2_DE W_wheat
-30
-20
-10
0
1 APSIM-NWHEAT_DE W_wheat
-70
-60
-50
-40
-30
-20
-10
0
10
1 APSIM_DE W_wheat
-8
0
-7
0
-6
0
-50
-40
-20
-10
0
1 APSIM_FI W_wheat
-7
0
-6
0
-50
-30
-20
-20
-10
0
1 CARAIB_DE W_wheat
-60
-50
-40
-30
-20
-10
0
10
1 CARAIB_FI W_wheat
-60
-50
-40
-3
0
-20
-10
-10
0
1 CERES/IT_ES W_wheat
-40
-30
-20
-10
-10
0
0
1 DNDC_DE W_wheat
-50
-40
-30
-20
-10
0
10
1 DNDC_ES W_wheat
-30
-20
-10
0
10
1 DNDC_FI W_wheat
-8
0
-70
-6
0
-5
0
-40
-30
-20
-20
-10
0
1 EPIC_DE W_wheat
0
10
20
30
40
1 EPIC_FI W_wheat
-30
-20
-10
1 FASSET_DE W_wheat
-30 -2
0
-10
0
1 FASSET_FI W_wheat
-4
0
-30
-20
-10
-10
0
1 LINTUL2_DE W_wheat
-30
-20 -10
0
1 SALUS_DE W_wheat
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
10
1 SALUS_FI W_wheat
-8
0
-50
-3
0
-20
-2
0
-10
0
1 SIRIUS2010_FI W_wheat
-30
-20
-10
0
1 SPACSYS_DE W_wheat
-50
-40
-30
-20
-10
0
1 SPACSYS_ES W_wheat
-40
-30
-20
-10
0
1 SPACSYS_FI W_wheat
-40
-30
-20
-10
0
1 WOFOST/FI_ES W_wheat
-60
-50
-40
-30
-20
-10
0
1 WOFOST/PL_DE W_wheat
-50
-30
-20 -10
0
1 WOFOST/PL_ES W_wheat
-4
0
-3
0
-20
-10
0
Cluster 4: n=25
Cluster 1: n=24
Cluster 3: n=3
Temp. change (°C)
Pre
cip
ita
tio
n
ch
an
ge
(%
)
Y+: T+, Y-: T++
Labels and cluster groups: winter wheat
3 AFRCWHEAT2_FI W_wheat
-70
-6
0
0
20
30
40 4
0
3 CERES/CZ_FI W_wheat
-20
-10
0 0
10
20
30
3 STICS_FI W_wheat
-30
-20
-1
0
0
10
20
20
30
40
50
5 AQUACROP_DE W_wheat
-70 -40
-30 -10
0
10
20
5 AQUACROP_FI W_wheat
-60
-20
-10
0
10
20
30
40
50
Cluster 5: n=2
6 CERES/IT_FI W_wheat
-40
-30
-30
-20
-20
-10 -10
0
10
6 LINTUL2_FI W_wheat
-20 -
20
-10
-10
0
0
10
6 STICS_DE W_wheat
0
0
Cluster 6: n=3
7 CROPSYST_FI W_wheat
-80
-70
-60
-60
-50
-50
-40
-40
-30
-20
-10
0
10
7 HERMES_FI W_wheat
-70
-60
-60
-50
-50 -40
-30 -20
-10
0
10
20
Cluster 7: n=2 Separate: n=1
_ES _FI _FI
4 APSIM-NWHEAT_FI W_wheat
-50
-40
-30 -30
-20
-10
0
10
4 CARAIB_ES W_wheat
-70
-60
-50
-40
-30
-20
-10
0
10
20
30
4 CERES/ES_DE W_wheat
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
4 CERES/ES_ES W_wheat
-90 -8
0
-70
-60
-
50
-40
-
30
-20
-10
0
10
20
40
4 CERES/ES_FI W_wheat
-90
-
80
-70
-60
-50
-40
-30
-20
-10
0
10
20
4 CERES/IT_DE W_wheat
-50 -40
-30
-20
-10
0
0
4 CROPSYST_DE W_wheat
-50 -40
-30 -20
-10
0
10
20
4 CROPSYST_ES W_wheat
-50
-40 -30
-20
-10
0
10
20
4 EPIC_ES W_wheat
-60
-50
-40
-30
-20
4 HERMES_DE W_wheat
-80
-70 -60
-50
-40
-30
-20
-10
0
10
20
4 HERMES_ES W_wheat
-80
-70 -60
-50
-40
-30
-20
-10
0
10
20
30
40
4 LINTUL4_DE W_wheat
-90 -80
-70
-60
-50
-40
-30
-20
-10
0
10
20
30
40
4 LINTUL4_ES W_wheat
-90
-80
-70
-60
-50
-40
-30
-20
-10
0 1
0 2
0
30
40
50
60
100
4 LINTUL4_FI W_wheat
-80 -70
-60
-50
-40
-30
-20
-10
0
4 LPJ-GUESS_DE W_wheat
-60
-50
-40
-30
-20
-10
0
10
20
30
4 LPJ-GUESS_ES W_wheat
-70 -60
-50
-40
-30
-20
-10
0
10
20
30
4 LPJ-GUESS_FI W_wheat
-70
-60
-50
-40
-30
-20
-10
0
1
0
20
4 MCWLA_ES W_wheat
-80
-70
-60
-50
-40
-30
-20
-10
0
10
20
30
4 MONICA_DE W_wheat
-70 -60
-50
-40
-30
-20
-10
0
4 MONICA_FI W_wheat
-50 -40
-30
-20
-10
0
10
4 SALUS_ES W_wheat
-60
-50
-40
-30
-20
-10 0
10
20
4 SIRIUS2010_DE W_wheat
-30
-20
-10
0
4 SIRIUS2010_ES W_wheat
-50
-40
-30
-20
-10
0
10
4 WOFOST/FI_DE W_wheat
-50 -40
-30
-20
-10
0
10
20
4 WOFOST/FI_FI W_wheat
-50
-40
-30
-20
-10
0
10
2 AFRCWHEAT2_ES W_wheat
-50
-40
-30
-20
-10
0
10
2 APSIM-NWHEAT_ES W_wheat
-50 -50
-40
-30
-20
-10
0
10
20
30
40
2 APSIM_ES W_wheat
-40 -30 -30
-20
-10
0
10
20
2 ARMOSA_DE W_wheat
-50 -40
-30 -20
-10
0
2 ARMOSA_ES W_wheat
-80 -7
0
-60
-50
-40
-30
-20
-10
0
10
20
30
2 CERES/CZ_DE W_wheat
-40 -30
-20
-10
0
10
2 CERES/CZ_ES W_wheat
-60 -60 -50
-40 -30
-20
-10 0
10
20
30
40 50
2 FASSET_ES W_wheat
-40 -30 -20
-10
0
2 LINTUL2_ES W_wheat
-70
-60
-60
-50 -40
-30
-20 -10
0
10
20
30
40 50
2 MCWLA_DE W_wheat
-40
-30 -20
-10
0
10
20
2 MCWLA_FI W_wheat
-40 -30
-20 -10
0
10
2 MONICA_ES W_wheat
-70
-60
-50
-40 -30
-20
-10
0
10 20
30
40
50
2 SIRIUSQUALITY_DE W_wheat
-50 -50 -40
-30
-20
-10
0
10
2 SIRIUSQUALITY_ES W_wheat
-60 -50
-40
-30
-20
-10
0
10
20
30
2 SIRIUSQUALITY_FI W_wheat
-30 -30 -20
-10
0
10
2 STICS_ES W_wheat
-20 -10
0
10 20
Cluster 2: n=16
1 AFRCWHEAT2_DE W_wheat
-30
-20
-10
0
1 APSIM-NWHEAT_DE W_wheat
-70
-60
-50
-40
-30
-20
-10
0
10
1 APSIM_DE W_wheat
-8
0
-7
0
-6
0
-50
-40
-20
-10
0
1 APSIM_FI W_wheat
-7
0
-6
0
-50
-30
-20
-20
-10
0
1 CARAIB_DE W_wheat
-60
-50
-40
-30
-20
-10
0
10
1 CARAIB_FI W_wheat
-60
-50
-40
-3
0
-20
-10
-10
0
1 CERES/IT_ES W_wheat
-40
-30
-20
-10
-10
0
0
1 DNDC_DE W_wheat
-50
-40
-30
-20
-10
0
10
1 DNDC_ES W_wheat
-30
-20
-10
0
10
1 DNDC_FI W_wheat
-8
0
-70
-6
0
-5
0
-40
-30
-20
-20
-10
0
1 EPIC_DE W_wheat
0
10
20
30
40
1 EPIC_FI W_wheat
-30
-20
-10
1 FASSET_DE W_wheat
-30 -2
0
-10
0
1 FASSET_FI W_wheat
-4
0
-30
-20
-10
-10
0
1 LINTUL2_DE W_wheat
-30
-20 -10
0
1 SALUS_DE W_wheat
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
10
1 SALUS_FI W_wheat
-8
0
-50
-3
0
-20
-2
0
-10
0
1 SIRIUS2010_FI W_wheat
-30
-20
-10
0
1 SPACSYS_DE W_wheat
-50
-40
-30
-20
-10
0
1 SPACSYS_ES W_wheat
-40
-30
-20
-10
0
1 SPACSYS_FI W_wheat
-40
-30
-20
-10
0
1 WOFOST/FI_ES W_wheat
-60
-50
-40
-30
-20
-10
0
1 WOFOST/PL_DE W_wheat
-50
-30
-20 -10
0
1 WOFOST/PL_ES W_wheat
-4
0
-3
0
-20
-10
0
Cluster 4: n=25
Cluster 1: n=24
Cluster 3: n=3
P-dominant (Y+: P+)
T-dominant (Y+: T+)
Y+: T+ & P+
T-dominant (Y-: T+)
V-shape 2-opt
Yield decrease (Y-) with
temperature increase (T+)
1 1
2 4
3 3
5 5
6 6
7 7
8 8
3 2 1
1 2 1
4 2 1
5 5 5
8 2 2
1 4 1
3 2 2
4 4 4
6 1 4
7 4 4
1 1 1
1 4 1
1 2 1
7 4 4
6 2 1
4 4 4
4 4 4
2 4 2
4 2 4
1 4 1
1 4 4
2 2 2
1 1 1
3 2 6
4 1 4
1 1
IRS clusters of the 3 locations for the
same model
FI ES DE # of different clusters
AFRCWHEAT2 Y+:T+ Y+:P+ Y-:T+ 3
APSIM Y-:T+ Y+:P+ Y-:T+ 2
APSIM-NWHEAT Y+:P+ Y+:P+ Y-:T+ 3
AQUACROP Y+:T+,P+ Y+:T+,P+ Y+:T+,P+ 1
ARMOSA 2-T-opt Y+:P+ Y+:P+ 2
CARAIB Y-:T+ Y+:P+ Y-:T+ 2
CERES/CZ Y+:T+ Y+:P+ Y+:P+ 2
CERES/ES Y+:P+ Y+:P+ Y+:P+ 1
CERES/IT Y+:T+,Y-:T++ Y-:T+ Y+:P+ 3
CROPSYST V-shape Y+:P+ Y+:P+ 2
DNDC Y-:T+ Y-:T+ Y-:T+ 1
EPIC Y-:T+ Y+:P+ Y-:T+ 2
FASSET Y-:T+ Y+:P+ Y-:T+ 2
HERMES V-shape Y+:P+ Y+:P+ 2
LINTUL2 Y+:T+,Y-:T++ Y+:P+ Y-:T+ 3
LINTUL4 Y+:P+ Y+:P+ Y+:P+ 1
LPJ-GUESS Y+:P+ Y+:P+ Y+:P+ 1
MCWLA Y+:P+ Y+:P+ Y+:P+ 2
MONICA Y+:P+ Y+:P+ Y+:P+ 2
SALUS Y-:T+ Y+:P+ Y-:T+ 2
SIRIUS2010 Y-:T+ Y+:P+ Y+:P+ 2
SIRIUSQUALITY Y+:P+ Y+:P+ Y+:P+ 1
SPACSYS Y-:T+ Y-:T+ Y-:T+ 1
STICS Y+:T+ Y+:P+ Y+:T+,Y-:T++ 3
WOFOST/FI Y+:P+ Y-:T+ Y+:P+ 2
WOFOST/PL ? Y-:T+ Y-:T+ 1
T-dominant (Y-: T+)
P-dominant (Y+: P+)
T-dominant (Y+: T+)
Y+: T+ & P+
Y+: T+, Y-: T++
V-shape
2-opt
Number of IRSs per cluster for spring and
winter wheat at the Finnish (FI), German
(DE) and Spanish (ES) sites
FI DE ES
Spring wheat
Num
ber
of
IRS
s0
510
15
20
1 2 3 4 5 6 7 8
FI DE ES
Winter wheat
Num
ber
of
IRS
s0
510
15
20
IRS differences of the same model calibrated
by different modelling groups (winter wheat)
Index o
f dis
tance
Cluster 1 Cluster 2 Cluster 4 3, 5-7
01
00
20
03
00
40
05
00
60
07
00
W_wheat/product, Agglomerative Coefficient = 0.99
AF
RC
WH
EA
T2
_D
ES
IRIU
S2
01
0_
FI
SP
AC
SY
S_
DE
SP
AC
SY
S_
ES
FA
SS
ET
_D
EF
AS
SE
T_
FI
CA
RA
IB_
FI
SP
AC
SY
S_
FI
WO
FO
ST
/FI_
ES
WO
FO
ST
/PL
_E
SA
PS
IM_
DE
SA
LU
S_
FI
AP
SIM
_F
IW
OF
OS
T/P
L_
DE
DN
DC
_F
IA
PS
IM-N
WH
EA
T_
DE
SA
LU
S_
DE
CA
RA
IB_
DE
DN
DC
_D
ED
ND
C_
ES
EP
IC_
FI
LIN
TU
L2
_D
EC
ER
ES
/IT
_E
SE
PIC
_D
EA
FR
CW
HE
AT
2_
ES
MC
WL
A_
DE
SIR
IUS
QU
AL
ITY
_D
EM
CW
LA
_F
IS
IRIU
SQ
UA
LIT
Y_
FI
AR
MO
SA
_D
EA
PS
IM_
ES
CE
RE
S/C
Z_
DE
FA
SS
ET
_E
SA
PS
IM-N
WH
EA
T_
ES
SIR
IUS
QU
AL
ITY
_E
SS
TIC
S_
ES
CE
RE
S/C
Z_
ES
AR
MO
SA
_E
SM
ON
ICA
_E
SL
INT
UL
2_
ES
AP
SIM
-NW
HE
AT
_F
IS
IRIU
S2
01
0_
DE
CE
RE
S/IT
_D
EM
ON
ICA
_D
EM
ON
ICA
_F
IC
RO
PS
YS
T_
DE
CR
OP
SY
ST
_E
SW
OF
OS
T/F
I_D
ES
AL
US
_E
SH
ER
ME
S_
DE
HE
RM
ES
_E
SL
PJ-G
UE
SS
_E
SM
CW
LA
_E
SE
PIC
_E
SL
INT
UL
4_
FI
SIR
IUS
20
10
_E
SW
OF
OS
T/F
I_F
IC
AR
AIB
_E
SL
PJ-G
UE
SS
_D
EL
PJ-G
UE
SS
_F
IC
ER
ES
/ES
_E
SC
ER
ES
/ES
_D
EC
ER
ES
/ES
_F
IL
INT
UL
4_
DE
LIN
TU
L4
_E
SC
RO
PS
YS
T_
FI
HE
RM
ES
_F
IA
FR
CW
HE
AT
2_
FI
CE
RE
S/C
Z_
FI
ST
ICS
_F
IC
ER
ES
/IT
_F
IL
INT
UL
2_
FI
ST
ICS
_D
EA
QU
AC
RO
P_
DE
AQ
UA
CR
OP
_F
I
x x x x x x x x x x x x x
2x WOFOST 3x CERES-wheat
IRS differences of model relatives (winter
wheat)
Index o
f dis
tance
Cluster 1 Cluster 2 Cluster 4 3, 5-7
01
00
20
03
00
40
05
00
60
07
00
W_wheat/product, Agglomerative Coefficient = 0.99
AF
RC
WH
EA
T2
_D
ES
IRIU
S2
01
0_
FI
SP
AC
SY
S_
DE
SP
AC
SY
S_
ES
FA
SS
ET
_D
EF
AS
SE
T_
FI
CA
RA
IB_
FI
SP
AC
SY
S_
FI
WO
FO
ST
/FI_
ES
WO
FO
ST
/PL
_E
SA
PS
IM_
DE
SA
LU
S_
FI
AP
SIM
_F
IW
OF
OS
T/P
L_
DE
DN
DC
_F
IA
PS
IM-N
WH
EA
T_
DE
SA
LU
S_
DE
CA
RA
IB_
DE
DN
DC
_D
ED
ND
C_
ES
EP
IC_
FI
LIN
TU
L2
_D
EC
ER
ES
/IT
_E
SE
PIC
_D
EA
FR
CW
HE
AT
2_
ES
MC
WL
A_
DE
SIR
IUS
QU
AL
ITY
_D
EM
CW
LA
_F
IS
IRIU
SQ
UA
LIT
Y_
FI
AR
MO
SA
_D
EA
PS
IM_
ES
CE
RE
S/C
Z_
DE
FA
SS
ET
_E
SA
PS
IM-N
WH
EA
T_
ES
SIR
IUS
QU
AL
ITY
_E
SS
TIC
S_
ES
CE
RE
S/C
Z_
ES
AR
MO
SA
_E
SM
ON
ICA
_E
SL
INT
UL
2_
ES
AP
SIM
-NW
HE
AT
_F
IS
IRIU
S2
01
0_
DE
CE
RE
S/IT
_D
EM
ON
ICA
_D
EM
ON
ICA
_F
IC
RO
PS
YS
T_
DE
CR
OP
SY
ST
_E
SW
OF
OS
T/F
I_D
ES
AL
US
_E
SH
ER
ME
S_
DE
HE
RM
ES
_E
SL
PJ-G
UE
SS
_E
SM
CW
LA
_E
SE
PIC
_E
SL
INT
UL
4_
FI
SIR
IUS
20
10
_E
SW
OF
OS
T/F
I_F
IC
AR
AIB
_E
SL
PJ-G
UE
SS
_D
EL
PJ-G
UE
SS
_F
IC
ER
ES
/ES
_E
SC
ER
ES
/ES
_D
EC
ER
ES
/ES
_F
IL
INT
UL
4_
DE
LIN
TU
L4
_E
SC
RO
PS
YS
T_
FI
HE
RM
ES
_F
IA
FR
CW
HE
AT
2_
FI
CE
RE
S/C
Z_
FI
ST
ICS
_F
IC
ER
ES
/IT
_F
IL
INT
UL
2_
FI
ST
ICS
_D
EA
QU
AC
RO
P_
DE
AQ
UA
CR
OP
_F
I
x x x x x x x x x x x x x
SUCROS family:
x: ARMOSA, HERMES, WOFOST
CERES family
x: CERES-wheat (3x)
c: APSIM-Nwheat, APSIM-wheat, SALUS
LINTUL family
L: SIMPLACE <LINTUL2, Slim>, LINTUL-4
c c c c c c c c c x x x x x L L L L L L
Ensemble distribution of simulated 30-
year averaged responses in the rate of
change of growing period length for spring
wheat (sowing to maturity)
FI
days / C
Fre
qu
en
cy
-12 -8 -6 -4 -2 0
02
46
81
01
21
4
DE
days / C
Fre
qu
en
cy
-12 -8 -6 -4 -2 0
02
46
81
01
21
4
ES
days / C
Fre
qu
en
cy
-12 -8 -6 -4 -2 0
02
46
81
01
21
4
Ratio of grain to above-ground dry matter at harvest
Harvest index for winter wheat in
Germany
Harvest index
Number of models in four range classes of the harvest index (HI; ratio of grain to above-ground dry
matter at harvest) for spring and winter wheat in Finland, Germany and Spain for the baseline climate, for
a large warming (T+9; temperature change = +9°C, precipitation at baseline) and large drying (P-50;
temperature at baseline, precipitation change = -50%). Thresholds for the HI ranges are based on
experimental data presented by Hay (1995) and Foulkes et al. (2011). The colours indicate if the number
of models remains the same as for the baseline (grey), decreases (blue) or increases (red).
Finland Germany Spain
HI class (range) Baseline T+9 P-50 Baseline T+9 P-50
Baseline T+9 P-50
Spring wheat
Low (<0.31) 2 3 2 1 2 3
2 3 7
Normal (0.31-0.50) 11 13 14 19 17 17
18 13 11
High (0.51 - 0.64) 9 5 7 3 4 3
3 7 6
Implausibly high (>0.64) 2 3 1 2 2 2
2 2 1
Winter wheat
Low (<0.43) 10 13 12 6 9 15
16 17 15
Normal (0.43-0.53) 12 9 10 15 13 4
8 6 9
High (0.54-0.64) 2 1 2 2 2 4
0 0 1
Implausibly high (>0.64) 2 3 2 3 2 3
2 3 1
1
Example of a particularly dry year at Nossen (DE) (spring wheat), year 2003
Analysis of extreme years
Relative to the 30-year mean
CONCLUSIONS
● Demonstration of using Impact Response Surfaces (IRSs) for a
systematic intercomparison of crop model behaviour under
conditions of changing climate
● Ensemble average yields decline with higher temperatures (3–7%
per 1°C) and decreased precipitation (3–9% per 10% decrease), but
benefit from increased precipitation (0-8% per 10% increase)
● Optimal temperatures for present-day cultivars are close to the
baseline under Finnish conditions but below the baseline at the
German and Spanish sites
Conclusions 1/2
● We have shown that clustering methods can be used to analyse
patterns of IRSs
● 30% of the models show consistent pattern for all 3 locations, 20%
end up in 3 different clusters (winter wheat)
● Diversity of patterns larger in Finland (temperature-limited) than in
Germany (close to optimum) and Spain (precipitation-limited)
● Different calibrations of the same model show similar IRS pattern (based on 2 WOFOST and 3 CERES-wheat calibration)
● Picture from model relatives less clear
● Next steps: comparing IRS clusters to harvest index and
phenology; clustering will also be tested for extreme years
● Follow-up to this study (IRS for T, P and CO2 sensitivity and
adaptations):
Ruiz-Ramos / An ensemble of projections of wheat adaptation to
climate change in Europe analyzed with impact response surfaces
– Wed 15:50, Session III
Conclusions 2/2
T-dominant response;
T-dominant response; optimal yield at
baseline T;
T-dominant response; optimal yield at
baseline T; strong decline with rising
T;
T-dominant response; optimal yield at
baseline T; strong decline with rising
T; baseline P-deficit TB+PD
Symbol Variants
T Temperature response dominates
P Precipitation response dominates
B Optimum yield close to baseline climate
C Optimum yield cooler than baseline T
W Optimum yield warmer than baseline T
D Precipitation deficit limits baseline yield
+ Strong response with large increase
relative to baseline
- Strong response with large decrease
relative to baseline
± Strong response with large increase and
large decrease relative to baseline
Grouping of IRSs with expert judgement
-2 0 2 4 6 8
-40
-20
020
40
Temperature change ( C)
Pre
cip
itation c
hange (
%)
-30
-20
-10
0 0
10
20
30
1 2
3
-90
-80
-70-60
-50
-40-30
-20
-10
010
20
3040
50
60
7080
90
100200
No.
1
2
3
4
5
6
Yield response behaviour
T-dominates; Insensitive to P
T-dominates; P has less effect
T and P have comparable effect
P-dominates; T has less effect
P-dominates; Insensitive to T
Unclassified
Winter wheat Results – Clustering of IRSs
”Subjective” method
applies expert judgement to
describe the climatic conditions
relative to the baseline and the
relative influence of temperature
and precipitation on yields away
from the baseline
Model FI ES DE No. clusters
AFRCWHEAT2 1 4 1 3
APSIM-Nwheat 3 4 2 3
APSIM 2 4 2 2
AquaCrop 2 4 4 2
ARMOSA 6 4 4 2
CARAIB 2 2 2 2
CERES-wheat DSSAT v.4.6/CZ 2 5 4 2
CERES-wheat DSSAT v.4.5/ES 2 2 2 1
CERES-wheat DSSAT v.4.5/IT 2 2 2 3
CropSyst 2 2 3 2
DNDC 2 2 2 1
EPIC 6 6 6 2
Fasset 1 4 1 2
HERMES 3 3 4 2
Lintul2 2 4 2 3
Lintul4 2 4 4 1
LPJ-GUESS 2 3 2 1
MCWLA 4 3 4 2
MONICA V1.2 2 4 3 2
SALUS 2 3 2 2
Sirius 2010 2 2 2 2
Sirius Quality 4 4 4 1
SPACSYS 2 1 1 1
STICS 2 4 1 3
WOFOST 7.1/FI 2 2 2 2
WOFOST 7.1/PL 2 2 2 2
Temperature dominates
yield response; decrease
with warming; low P-
sensitivity for strong
warming
1 AFRCWHEAT2_DE W_wheat
-30
-20
-10
0
1 APSIM-NWHEAT_DE W_wheat
-70
-60
-50
-40
-30
-20
-10
0
10
1 APSIM_DE W_wheat
-8
0
-7
0
-6
0
-50
-40
-20
-10
0
1 APSIM_FI W_wheat
-7
0
-6
0
-50
-30
-20
-20
-10
0
1 CARAIB_DE W_wheat
-60
-50
-40
-30
-20
-10
0
10
1 CARAIB_FI W_wheat
-60
-50
-40
-3
0
-20
-10
-10
0
1 CERES/IT_ES W_wheat
-40
-30
-20
-10
-10
0
0
1 DNDC_DE W_wheat
-50
-40
-30
-20
-10
0
10
1 DNDC_ES W_wheat
-30
-20
-10
0
10
1 DNDC_FI W_wheat
-8
0
-70
-6
0
-5
0
-40
-30
-20
-20
-10
0
1 EPIC_DE W_wheat
0
10
20
30
40
1 EPIC_FI W_wheat
-30
-20
-10
1 FASSET_DE W_wheat -3
0 -2
0
-10
0
1 FASSET_FI W_wheat
-4
0
-30
-20
-10
-10
0
1 LINTUL2_DE W_wheat
-30
-20 -10
0
1 SALUS_DE W_wheat
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
10
1 SALUS_FI W_wheat
-8
0
-50
-3
0
-20
-2
0
-10
0
1 SIRIUS2010_FI W_wheat
-30
-20
-10
0
1 SPACSYS_DE W_wheat
-50
-40
-30
-20
-10
0
1 SPACSYS_ES W_wheat
-40
-30
-20
-10
0
1 SPACSYS_FI W_wheat
-40
-30
-20
-10
0
1 WOFOST/FI_ES W_wheat
-60
-50
-40
-30
-20
-10
0
1 WOFOST/PL_DE W_wheat
-50
-30
-20 -10
0
1 WOFOST/PL_ES W_wheat
-4
0
-3
0
-20
-10
0
TB TC+PD- TB+PD TB+PB TCPD
TB+PB
U
TB+PD-
TB+PD
TC+PB TC+PD TC PD TB+PB
U TC TC+ TC PD
TB+PD TC PD TC+ TC+
TB+PD TB+PB TB+PD
Basic class
TB
TC
TW
TB PB
TB PD
TC PB
TC PD
TW PD
TB PD
TC PD
TB PD
TC PD
TW PD
PD
U
1
2
3
4
5
6
Results – Clustering of IRSs
”Subjective” method
”Objective” method, Cluster 1
2 AFRCWHEAT2_ES W_wheat
-50
-40
-30
-20
-10
0
10
2 APSIM-NWHEAT_ES W_wheat
-50 -50
-40
-30
-20
-10
0
10
20
30
40
2 APSIM_ES W_wheat
-40 -30 -30
-20
-10
0
10
20
2 ARMOSA_DE W_wheat
-50 -40
-30 -20
-10
0
2 ARMOSA_ES W_wheat
-80 -7
0
-60
-50
-40
-30
-20
-10
0
10
20
30
2 CERES/CZ_DE W_wheat
-40 -30
-20
-10
0
10
2 CERES/CZ_ES W_wheat
-60 -60 -50
-40 -30
-20
-10 0
10
20
30
40 50
2 FASSET_ES W_wheat
-40 -30 -20
-10
0
2 LINTUL2_ES W_wheat
-70
-60
-60
-50 -40
-30
-20 -10
0
10
20
30
40 50
2 MCWLA_DE W_wheat
-40
-30 -20
-10
0
10
20
2 MCWLA_FI W_wheat
-40 -30
-20 -10
0
10
2 MONICA_ES W_wheat
-70
-60
-50
-40 -30
-20
-10
0
10 20
30
40
50
2 SIRIUSQUALITY_DE W_wheat
-50 -50 -40
-30
-20
-10
0
10
2 SIRIUSQUALITY_ES W_wheat
-60 -50
-40
-30
-20
-10
0
10
20
30
2 SIRIUSQUALITY_FI W_wheat
-30 -30 -20
-10
0
10
2 STICS_ES W_wheat
-20 -10
0
10 20 Precipitation dominates yield response;
decrease with P-deficit, except for large
warming
TC PD TC PD- TW PD TB PD- TwPD-
Tw PD
TC PD
PD± TW PD TW+ PD± TC+PD-
TB PD TB+PD± TC PD- TC PD- TB PD
Basic class
TB
TC
TW
TB PB
TB PD
TC PB
TC PD
TW PD
TB PD
TC PD
TB PD
TC PD
TW PD
PD
U
1
2
3
4
5
6
Results – Clustering of IRSs
”Subjective” method ”Objective” method, Cluster 2
Comparison of clusters to model characteristics
37
