Modelling shi s in ag oclima e and c op cul i a esponse
unde clima e change
Reimund P. R€
o e
1
, Jukka Ho
¨hn
2
, Mi ek T nka
3,4
, S e an F onzek
5
, Timo hy R.Ca e
5
&
Helena Kahiluo o
1
1
Plan P oduc ion Resea ch, MTT Ag i ood Resea ch Finland, L€
onn o inka u 5, FI-50100 Mikkeli, Finland
2
Plan P oduc ion Resea ch, MTT Ag i ood Resea ch Finland, Vakolan ie 55, FI-03400 Vih i, Finland
3
Ins i u e o Ag osys ems and Bioclima ology, Mendel Uni e si y in B no, Zemedelska 1, 61300 B no, Czech Republic
4
Global Change Resea ch Cen e, Academy o Science o he Czech Republic, Belidla 986/4a, 60300 B no, Czech Republic
5
Clima e Change P og amme, Finnish En i onmen Ins i u e (SYKE), P.O. Box 140, FI-00251 Helsinki, Finland
Keywo ds
Adap a ion, ag oclima ic indica o , ba ley,
c op simula ion model, cul i a esponse
di e si y.
Co espondence
Reimund P. R€
o e , Plan P oduc ion
Resea ch, MTT Ag i ood Resea ch Finland,
L€
onn o inka u 5, FI-50100 Mikkeli, Finland.
Tel: +358403534506; Fax: +35815226578;
E-mail: [email p o ec ed]
Funding In o ma ion
This wo k was suppo ed by he Academy o
Finland, p ojec s A-La-Ca e, Ma e ic and
NORFASYS (decision nos. 140846, 128043
and 268277, espec i ely), and by he
Minis y o Ag icul u e and Fo es y o
Finland, p ojec s Adacapa and FACCE-
Macsu .
Recei ed: 31 May 2013; Re ised: 20 Augus
2013; Accep ed: 21 Augus 2013
Ecology and E olu ion 2013; 3(12): 4197–
4214
doi: 10.1002/ece3.782
Abs ac
This pape aims: (i) o iden i y a na ional scale a eas whe e c op yield o ma-
ion is cu en ly mos p one o clima e-induced s esses, (ii) o e alua e how
he se e i y o hese s esses is likely o de elop in ime and space, and (iii) o
app aise and quan i y he pe o mance o wo s a egies o adap ing c op cul i-
a ion o a wide ange o (unce ain) clima e change p ojec ions. To his end
we made use o ex ensi e clima e, c op, and soil da a, and o wo modelling
ools: N-Ag iCLIM and he WOFOST c op simula ion model. N-Ag iCLIM
was de eloped o he au oma ic gene a ion o indica o s desc ibing basic
ag oclima ic condi ions and was applied o e he whole o Finland. WOFOST
was used o simula e de ailed c op esponses a ou ep esen a i e loca ions.
N-Ag iCLIM calcula ions ha e been pe o med na ionally o 3829 g id boxes
a a 10 910 km esolu ion and o 32 clima e scena ios. Ranges o p ojec ed
shi s in indica o alues o hea , d ough and o he c op- ele an s esses
ac oss he scena ios a y widely –so do he spa ial pa e ns o change. O e all,
unde e e ence clima e he mos isk-p one a eas o sp ing ce eals a e ound
in sou h-wes Finland, shi ing o sou h-eas Finland owa ds he end o his
cen u y. Condi ions o g ass a e likely o imp o e. WOFOST simula ion esul s
sugges ha CO
2
e iliza ion and adjus ed sowing combined can lead o small
yield inc eases o cu en ba ley cul i a s unde mos clima e scena ios on
a ou able soils, bu no unde ex eme clima e scena ios and poo soils. This
in o ma ion can be aluable o app aising al e na i e adap a ion s a egies. I
acili a es he iden i ica ion o egions in which clima ic changes migh be apid
o o he wise no able o c op p oduc ion, equi ing a mo e de ailed e alua ion
o adap a ion measu es. The esul s also sugges ha u ilizing he di e si y o
cul i a esponses seems bene icial gi en he high unce ain y in clima e change
p ojec ions.
In oduc ion
Ag icul u al p oduc ion is sensi i e o a ia ions in
wea he and clima e and can be expec ed o be in luenced
ma kedly by clima e change (Rosenzweig and Hillel 1998;
R€
o e and an de Geijn 1999; Pa y e al. 2004; Fische
e al. 2005; God ay e al. 2010). Global wa ming is
expec ed o lead o apid inc eases in empe a u e, espe-
cially in no he ly la i udes (Be s e al. 2011; Ruos eenoja
e al. 2011). Fu u e p ojec ions o p ecipi a ion mainly
show inc eases in no he n Eu ope, which a e usually
la ges in win e (F onzek e al. 2012), bu wi h conside -
able a ia ion be ween clima e models (Slo h Madsen
e al. 2012). P ojec ed changes in mean clima ic condi-
ions ha e gene ally been conside ed bene icial o ag i-
cul u e in he No dic egion (e.g., Ca e e al. 1996).
Howe e , ecen ly doub s ha e been aised whe he ha
also holds ue i clima ic a iabili y inc eases ma kedly
and p og ess in plan b eeding and ag onomy canno
keep pace ensu ing e ec i e adap a ion (R€
o e e al.
ª2013 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d.
This is an open access a icle unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s use,
dis ibu ion and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed.
4197
2011a). E en ually, implemen a ion o e ec i e adap a ion
migh also be hinde ed by oo high unce ain ies in
clima e change p ojec ions. The cu en s udy aims a a
de ailed na ional assessmen o clima e change isks o
c op p oduc ion ha , o he i s ime, sys ema ically
combines an ag oclima ic indica o app oach wi h c op
g ow h simula ion using he same daily inpu da a.
Mos s udies o clima e change impac s on c op yields
apply ei he s a is ical models (Lobell and Bu ke 2010) o
p ocess-based c op simula ion models (R€
o e e al. 2011b;
Whi e e al. 2011; Osbo ne e al. 2013). Mos p ocess-based
models a e also capable o simula ing, in addi ion, e ec s
o enhanced CO
2
concen a ion and managemen p ac ices
on biomass, seed yields and wa e use o c ops (Rosenzweig
and Pa y 1994; Nelson e al. 2009; Ewe e al. 2011; Ang-
ulo e al. 2013). Howe e , e en he mo e complex p ocess-
based c op simula ion models canno ake all impo an
in e ac ions be ween he en i onmen and managemen
(E 9M) in o accoun , such as e ec s o hea y ain all on
ha es ed yield. Nei he do hey include all in e ac ions
be ween geno ype and en i onmen (G 9E) such as yield
educ ion due o wea he -induced pes and/o disease
occu ence. On he o he hand, c op g ow h simula ion is
he only meaning ul p ac ical way o analysing he in e ac-
ions be ween he many op ions o combining di e en
c op cul i a s wi h di e se managemen p ac ices unde a
wide ange o possible new en i onmen al condi ions
(Semeno and Hal o d 2009; R€
o e e al. 2011a,b; Rosenz-
weig e al. 2013). Usually, c op-clima e models do no
co e all impo an c ops and soils in a egion. Fo his ea-
son, ag oclima ic indica o app oaches a e some imes
applied o p o ide a mo e comp ehensi e pic u e o he ag-
oclima e o la ge a eas and i s shi s unde clima e
change (Ha ison and Bu e ield 1996; T nka e al. 2011).
Knowledge o he b oad-scale ag oclima e can also p o ide
a use ul basis o upscaling si e speci ic c op simula ion
esul s, o e ing a s ong a gumen o combining he wo
app oaches. Such a combina ion can p o ide in o ma ion
on shi s in he sui abili y and po en ial o c op p oduc-
ion in a egion o coun y unde clima e change (Ca e
and Saa ikko 1996; Challino 2011).
This s udy demons a es he bene i o combining ag o-
clima ic indica o s calcula ed wi h g idded wea he da a
o Finland wi h mo e de ailed c op g ow h simula ions.
I co e s one o he ew egions in Eu ope whe e he
changing clima e is expec ed o imp o e o e all ag ocli-
ma ic condi ions (Ca e e al. 1996; T nka e al. 2011),
bu whe e conce ns s ill emain abou he abili y o u ilize
he po en ial due o speci ic soil condi ions, inc eased pes
and disease isks, he apid a e o he change and possi-
ble inc easing clima e a iabili y.
The speci ic objec i es o his pape a e: (i) o iden i y
a eas in Finland whe e c op yield o ma ion is cu en ly
mos p one o clima e-induced s esses, (ii) o e alua e
how he se e i y o hese s esses is likely o de elop in
ime and space unde a wide ange o u u e clima e
p ojec ions, and, based on such isk assessmen , (iii) o
app aise and quan i y he pe o mance o wo al e na i e
s a egies o adap ing c op cul i a ion o unce ain
p ojec ions o u u e clima e. To exempli y his, we use
sp ing ba ley (Ho deum ulga e L.) as es c op and daily
wea he da a o he baseline pe iod (1971–2000) and a
wide ange o p ojec ed u u es (32 clima e scena ios) up
o yea 2100, a a spa ial esolu ion o 10 910 km o
he en i e coun y. Ba ley (see, pho o) is he mos widely
g own ield c op in Finland - i s cul i a ion a ea is shown
in Fig. 1. Resul s o he s udy a e expec ed o p o ide
undamen al knowledge o a ge -o ien ed plan b eeding
and ag onomic ad ancemen s designed o enhance he
esilience o ag icul u al sys ems unde a changing clima e
in Finland.
Ruukki
Laukaa
So kamo
Mikkeli
Pälkäne
Maaninka
Ylis a o
Kokemäki
Mie oinen
Tohmajä i
Anjalankoski
U i
Oulu
Seinäjoki
Jokioinen
BOR4
BOR3
BOR1
BOR6
ALN1
BOR1
NEM1
BOR8
Oulu
Tu k u
Vih u
Lah i
Van aa
Kuopio
U sjoki
Joensuu
Sodankylä
Ro aniemi
Jy äskylä
Ma iehamn
Ba ley Cul i a ion [ha]
> 0 - 500
500 - 1000
1000 - 1500
1500 - 2000
> 2000
Selec ed g id cell o
c op yield simula ion
MTT Va ie y T ial Si es
Wea he S a ion
Figu e 1. Ba ley cul i a ion, wea he s a ions, majo MTT o icial
a ie y ial si es and En i onmen al Zones (EnZs) o Finland acco ding
o Me zge e al. (2005). T iangles indica e loca ions o MTT o icial
a ie y ial si es o ba ley. Filled la ge squa es indica e selec ed g id
used o c op yield simula ion in his s udy (small illed ci cles indica e
long- e m wea he s a ions).
4198 ª2013 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d.
Shi s in Ag oclima e and Cul i a Response R. P. R€
o e e al.
Ma e ials and Me hods
Se -up o he s udy
To assess shi s in he ag oclima ic sui abili y o majo
c ops and in he yield po en ial o cu en cul i a s o
sp ing ba ley (as a key c op) in Finland, we applied a com-
bina ion o wo impac assessmen me hods ha a e usu-
ally applied sepa a ely. Fi s , he Ag iCLIM so wa e o
calcula e ag oclima ic indica o s (T nka e al. 2011) was
ex ended o include indica o s ele an o highe la i udes
in a e sion called N-Ag iCLIM. A desc ip ion o how
hese indica o s we e selec ed is gi en in he Da a S1. The
ool was applied o assess shi s in ag oclima ic sui abili y
o cul i a ing c op- and g assland, and iden i y a eas mos
p one o clima ic isks unde a wide ange o clima e
change scena ios. Second, he p ocess-based dynamic c op
simula ion model WOFOST ( e sion 7.1; an Diepen e al.
1989; Boogaa d e al. 1998) was applied o quan i y
impac s o clima e change on yields o di e en cu en ly
a ailable ba ley cul i a s and o a la ge ensemble o cli-
ma e change scena ios.
Bo h N-Ag iCLIM and WOFOST we e un wi h he
same daily wea he da a on a 10 910 km
2
g id basis o
he pe iod 1971–2100. While N-Ag iCLIM was un o
he whole o Finland, WOFOST simula ions we e con-
duc ed only o selec ed g id cells (see, Fig. 1), and wi h
soil da a o ep esen a i e soil ypes. C op da a applied
in N-Ag iCLIM we e based on cha ac e is ics o he
popula ba ley cul i a Sca le , while he mo e comp e-
hensi e c op da a equi ed o c op modelling we e
ex ac ed and p ocessed om MTT o icial a ie y ial
da abases (e.g., Kangas e al. 2006).
N-Ag iCLIM, de eloped om Ag iCLIM (T nka e al.
2011) ha had been used o calcula e ag oclima ic indica-
o s selec ed on he basis o a p e ious Eu ope-wide
s udy, was applied o unde ake subsequen s a is ical
analysis o he ela ionships be ween yield o sp ing ba ley
cul i a s and wea he a iables in Finland (Hakala e al.
2012) (see, Table S1). Ou o ha analysis a inal se o
10 ag oclima ic indica o s was selec ed, which we e
deemed mos ele an o Finnish ag icul u e, cap u ing
condi ions ha ha e he mos p onounced in luence on
g ow h and yield o ma ion o majo Finnish c ops. These
comp ise: (i) he sum o e ec i e global adia ion (Eg ),
(ii) numbe o e ec i e g owing days (Egd), (iii) da e o
he las os (Las F os ), (i ) ela i e sowing da e
(DelayS) and ( ) p opo ion o sui able days o sowing
in sp ing (Sowing), ( i) numbe o days wi h wa e de ici
du ing he pe iod om Ap il o June (D yAJ) and ( ii)
June o Augus (D yJA), ( iii) o al p ecipi a ion du ing
he pe iod om 3 o 7 weeks a e sowing (RainAS), (ix)
numbe o days wi h maximum empe a u e o 28°Co
highe 1 week be o e o 3 weeks a e heading (S essE)
and (x) empe a u e sum accumula ion a e du ing
pe iod be ween an hesis and physiological ma u i y, ha is,
g ain illing (TempHRA g); de ini ions a e p o ided in
Table 1. As discussed o example, by T nka e al. (2011),
indica o s i- a e much mo e impo an o g ass and pe en-
nial c ops han hey a e o annual ield c ops such as ce eals.
Ag oclima ic indica o s we e calcula ed o each cell o
he 10 910 km
2
g idded da abase and mapped. De ails
on N-Ag iCLIM, o example, on de e mining ele an
c op phenological s ages o wa e de ici s, a e desc ibed in
he me hods sec ion o he (Da a S1).
Mul iple eg ession analyses o ag oclima ic
indica o s on yield
Obse ed yields om ba ley ials (be ween 1971–2009)
conduc ed a h ee loca ions (Jokioinen, Ylis a o and
Ruukki) (Fig. 1) we e collec ed o pe o m mul iple
eg ession analyses on he ela ionship be ween yield and
he a ious ag oclima ic indica o s used in his s udy. The
deg ee o i ness o he models di e s by loca ion (Table
S2) (see also, Resul s sec ion).
WOFOST c op simula ion model
The c op model WOFOST (WO ld FOod S udies, e sion
7.1, Boogaa d e al. 1998), de eloped in he amewo k o
an in e disciplina y s udy on wo ld ood p oduc ion
po en ials o annual c ops, was applied. I s p incipal com-
ponen s, p ocess o mula ions, and a ious applica ions
ha e been desc ibed by an Diepen e al. (1989) and an I -
e sum e al. (2003). The model p o ides a dynamic
desc ip ion o phenological de elopmen , CO
2
assimila-
ion, espi a ion, pa i ioning o assimila es o a ious
plan o gans, g ow h and yield o ma ion and (e apo-)
anspi a ion o a c op om eme gence un il ma u i y (a a
daily ime s ep), on he basis o c op gene ic cha ac e is ics,
en i onmen al condi ions and managemen p ac ices
(G 9E9M in e ac ions). WOFOST had been calib a ed
and applied o di e en Finnish and Eu opean ba ley cul-
i a s (R€
o e e al. 2011b, 2012) wi h daily wea he , soil
and c op da a es ablished o Finnish condi ions. Yield sim-
ula ions we e pe o med o selec ed g id cells ha a e close
(wi hin 10 km dis ance) o long- e m a ie y ial si es, and
ep esen he mos impo an ba ley cul i a ion a eas and
he majo en i onmen al zones (Me zge e al. 2005) ele-
an o ag icul u e (Fig. 1).
Inpu da a: c op, soil and cu en wea he
Fi s we g ouped a ailable mode n ba ley cul i a s
( eleased a e 1985) as g own by Finnish a me s, in o
ª2013 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. 4199
R. P. R€
o e e al. Shi s in Ag oclima e and Cul i a Response
h ee g oups, depending on hei ma u i y class, naming
hem a e widely known indi idual cul i a s: Annabell
(la e ma u ing), Kus aa (medium), and Kunna i (ea ly).
As a s a ing poin , we used c op pa ame e s o sp ing
ba ley based on mul i-loca ional ield expe imen s o
indi idual cul i a s (R€
o e e al. 2011b). MTT’s o icial
a ie y ial da a (Kangas e al. 2006) we e used o adjus
phenology- ela ed c op pa ame e s o he medium
(Kus aa) and ea ly (Kunna i) ma u ing g oups (see, Table
S6). Fu he mo e, we modi ied c op pa ame e s o
accoun o he enhanced ne pho osyn hesis and
inc eased wa e use e iciency (R€
o e and an de Geijn
1999) due o h ee di e en le els o ele a ed a mosphe ic
CO
2.
The alues o hose pa ame e s a ec ed by he le el
o a mosphe ic CO
2
concen a ion (see, Table S7) di e ed
sligh ly om a p e ious s udy (R€
o e e al. 2011a) due o
small di e ences in he CO
2
concen a ions conside ed.
A mosphe ic CO
2
o he nex decade is expec ed o
inc ease a a es be ween 2 and 4 ppm pe annum
(Ande son and Bows 2008). This implies ha by 2025
Table 1. The 10 selec ed ag o-clima ic indica o s gene a ed by N-Ag iCLIM (T nka e al. 2011) (as p esen ed in Fig. 6).
Ag oclima ic indica o Indica o name (uni s) De ini ion Symbol
Po en ial biomass and
c op de elopmen
Sum o e ec i e global
adia ion (MJ m-2 season -1)
Sum o global adia ion o days
wi h daily mean empe a u e
>5°C, daily minimum empe a u e
>0°C, ETa*/ET a io >0.4 and no
snow co e
Eg
Time pe iod sui able o
c op g ow h
Sum o e ec i e g owing
days (days)
Numbe o days wi h daily mean
empe a u e >5°C, daily minimum
empe a u e >0°C, ETa*/ET a io
>0.4 and no snow co e
Egd
Low empe a u e limi a ions Da e o he las os
(da e om Janua y 1s )
Las occu ence o a daily minimum
empe a u e o <0.1°C in he
gi en season be o e June 30 h
Las F os
Sowing condi ions ha will
a ec he g owing season
Delayed sowing (day)
1
Day o he yea when 10-day mo ing
a e age o daily mean empe a u e
exceeds h eshold empe a u e o
8°C exp essed as de ia ion om
May 1s
DelayS
P opo ion o sui able days o
sowing in ime window Ap il
26
h
h ough May 20 h
(la e sp ing)
All days wi h soil-wa e con en in he
op 0.1 m be ween 10% and 70%
o he maximum soil wa e -holding
capaci y (SWC), mean daily
empe a u e on he gi en day and
on he p eceding day >5°C, wi hou
snow co e and wi h p ecipi a ion on
he gi en day <= 1 mm and
p ecipi a ion on he p eceding
day <= 5mm
Sowing
Wa e de ici du ing g owing
season ha may esul in
d ough
Numbe o days wi h wa e
de ici s om Ap il o
June (days)
All days wi hin he gi en pe iod
wi h ETa/ET o <0.4
D yAJ
Numbe o days wi h wa e
de ici s om June o
Augus (days)
All days wi hin he gi en pe iod wi h ETa/ET o <0.4 D yJA
Rain a e sowing (mm) Sum o ain 3–7 weeks a e sowing RainAS
Po en ial g ain numbe o ma ion
and yield po en ial de e mina ion
2
Ve y high empe a u e
s ess (days)
Numbe o days wi h maximum empe a u e
o 28°C o highe 1 week be o e o
2 weeks a e heading
S essE
Mean daily empe a u e
sum accumula ion a e
a g ain illing
Ra e o Tsum abo e 0°C accumula ion (pe day)
om heading o yellow ipeness
TempHRA g
ETa and ET s and o ac ual e apo anspi a ion and e e ence e apo anspi a ion espec i ely calcula ed acco ding o FAO me hods (Allen e al.
1998) conside ing sp ing ba ley as a co e c op.
1
Ca e and Saa ikko (1996).
2
Hakala e al. (2012).
4200 ª2013 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d.
Shi s in Ag oclima e and Cul i a Response R. P. R€
o e e al.
(midpoin o 2011–2040) we may each le els o app oxi-
ma ely 420–450 ppm ; o 2055 his would be 480–
570 ppm and 540–690 ppm by 2085. Acco dingly, we
adjus ed c op pa ame e s o concen a ions o 435, 525
and 615 ppm , espec i ely, using es ablished p ocedu es
(Table S7). Soil and opog aphic da a comp ised ield
da a o olume ic soil mois u e (SM) con en a sa u a-
ion (SM0 o o al po e space), a ield capaci y (SMFC)
and a wil ing poin (SMW) and a ansmission zone pe -
meabili y pa ame e (SOPE) o he oo zone. Run-o
was assumed o be absen . Plan a ailable soil mois u e
(PASM) con en is calcula ed as he ac ual amoun
a ailable a ield capaci y minus he con en a wil ing
poin (SMFC-SMW); da a we e de i ed o a clay loam
and a sil y sand soil wi h PASM alues o 0.18 and
0.22 (cm
3
/cm
3
), espec i ely.
Daily wea he da a in e pola ed om s a ions o a
egula 10 910 km g id we e ob ained om he Finnish
Me eo ological Ins i u e co e ing he pe iod 1971–2009
o he ollowing a iables: minimum and maximum
(Tmax) nea -su ace empe a u e, global adia ion, p e-
cipi a ion and apou p essu e (Ven€
al€
ainen e al. 2005;
upda ed). As mean daily wind speed was no a ailable
om his da a se , we bi-linea ly in e pola ed he daily
mean alues o 10-m wind speed om wo e-analysis
da a p oduc s p o ided by he Eu opean Cen e o
Medium-Range Wea he Fo ecas s (ECMWF) om hei
o iginal, coa se spa ial esolu ion o he 10 910 km g id.
The e-analysis da a se s ERA-in e im (Dee e al. 2011)
and ERA-40 (Uppala e al. 2005) gi e a high empo al, bu
ela i e low spa ial esolu ion wi h 0.75°g id cell size o
ERA-in e im and 2.5° o ERA-40. ERA-40 was used o he
1971–1978 and ERA-in e im o 1979–2010. The esul ing
ime se ies o in e pola ed alues p o ides a gene al spa ial
pa e n o di e ences in wind speed o example be ween
coas al and inne -land a eas and compa ed ela i ely well
wi h da a om selec ed s a ions.
Clima e scena ios
Da a om he CMIP3 a chi e (Meehl e al. 2007) was
downloaded ep esen ing mon hly mean alues o ou pu
om Gene al Ci cula ion Models (GCMs). Simula ions
om ou expe imen s we e used, one wi h g eenhouse
gas concen a ions as obse ed o he 20 h cen u y and
h ee o cing scena ios o he 21s cen u y, SRES A2
(high emission), SRES A1B (mode a e) and SRES B1
(low) (Nakiceno ic e al. 2000). Da a ha e been down-
loaded o all GCM-SRES combina ions o which he
equi ed se o a iables (Table S4) was a ailable. This
esul ed in an ensemble o 36 simula ions, 13 o ced by
SRES B1, 14 by A1B and 9 by A2 (Table S4), ou o
which a sub-se o 11 scena ios has been selec ed
spanning he ange o unce ain y. Figu e 2 shows o
en i e Finland his o ical anomalies as well as p ojec ed
changes in empe a u e and p ecipi a ion. Clima e change
scena ios we e calcula ed in h ee s eps:
1Calcula ion o mon hly long- e m mean changes o h ee
u u e pe iods, 2011–2040, 2041–2070 and 2071–2100,
ela i e o he baseline pe iod 1971–2000 on he o iginal
GCM g id o empe a u e, p ecipi a ion (as ela i e
change), wind speed (calcula ed om i s zonal and me idi-
onal componen s), apou p essu e change and global
adia ion. Change in apou p essu e was es ima ed using
Tempe a u e anomaly [°C]
B1
A1B
A2
P ecipi a ion anomaly [%]
1960 1980 2000 2010−2039 2040−2069 2070−2099
1960 1980 2000 2010−2039 2040−2069 2070−2099
−2 0 2 4 6
−20−100 102030
B1 A1B A2
(A)
(B)
Figu e 2. (A–B) Obse ed anomalies in a e age annual (A) ai
empe a u e and (B) p ecipi a ion and p ojec ed changes du ing he
21s cen u y o Finland o SRES (Special Repo on Emissions
Scena ios (Nakiceno ic e al. 2000) scena ios B1, A1B and A2
simula ed wi h 11 GCMs selec ed o his s udy (s a s, see Table S4)
and o a la ge ensemble o 24 GCMs (boxplo s).
ª2013 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. 4201
R. P. R€
o e e al. Shi s in Ag oclima e and Cul i a Response

an app oxima e ela ion o sea le el ai p essu e and
speci ic humidi y (Mi chell e al. 2004, p. 9).
2Bi-linea in e pola ion o he 10 910 km g id o he
obse ed da a.
3Fo each 10 km g id cell, linea in e pola ion o
mon hly changes o daily es ima es o leap- and non-
leap yea s. The daily del as we e hen added o he
obse ed ime-se ies 1971–2000 o each g id cell o he
h ee u u e pe iods. The esul ing scena ios a e he e-
o e ep oducing he obse ed in e annual and daily
a iabili y and only changes in mean condi ions a e
conside ed.
Adap a ion measu es
Apa om examining he pe o mance o di e en
cu en ba ley cul i a g oups (cha ac e ized in Table S6)
unde changed clima ic condi ions, we also ook in o
accoun he CO
2
e iliza ion e ec (Table S7). Mo eo e ,
we adjus ed sowing da es based on es ablished empe a-
u e c i e ia (Ca e and Saa ikko 1996), bu wi h co ec-
ions o di e ences obse ed be ween calcula ed op imal
and ac ually obse ed sowing o a me s, who in he
majo i y show a mo e conse a i e beha iou in adjus ing
sowing (on a e age 1 week la e ) han would be possible
in esponse o empe a u e condi ions alone.
Resul s
Ag oclima ic indica o s and hei p ojec ed
shi s by 2025, 2055 and 2085
Mul iple eg ession analyses on he ela ionship be ween
yield and he a ious ag oclima ic indica o s showed ha
esul s a e loca ion-speci ic. The highes coe icien o
a ia ion (adjus ed R
2
) is obse ed o Ruukki ial si e,
explaining up o 46.6% o he a ia ion o g ain yields,
whe eas he equi alen alue is only 33.3% o Jokioinen
and 23.8% o Ylis a o. The impo ance o p edic o
a iables also a ies om si e o si e (see, Tables S2
and S3).
F om he 10 ag oclima ic indica o s ( esul s p esen ed
in Fig. 6), h ee we e selec ed o p esen a ion in o m o
maps (Figs. 3, 4 and 5): (i) RainAS, (ii) S essE, and (iii)
TempHRA g. The selec ion was based on a li e a u e
e iew (e.g., Ca e and Saa ikko 1996; Hakala e al.
2012) and mul iple eg ession analysis o sp ing ba ley.
Since ba ley is a good indica o o many o he de e mi-
na e (sp ing) ce eal c ops (R€
o e and an de Geijn 1999),
we can assume ha unde ecen pas and p esen -day
clima e, he h ee indica o s men ioned abo e, can be
conside ed mos impo an : They indica e known phe-
nomena such as ea ly season d ough (R€
o e e al. 2012),
hea empe a u e s ess du ing mos sensi i e phase
a ound lowe ing (e.g., Po e and Gawi h 1999), and
high empe a u es du ing g ain illing pe iod has ening
ma u i y and, he eby, educing yield po en ial (e.g., Ha-
kala e al. 2012). We hypo hesized ha hese isks a e
likely o be u he exace ba ed unde u u e clima ic con-
di ions. In o de o p o ide clima e isk in o ma ion o
o he c ops, including pe ennials such as g ass, we also
p esen esul s o he se en o he indica o s (Fig. 6).
Figu e 3 shows he p ojec ed changes in he h ee indi-
ca o s o one clima e scena io (IPSL-CM4/ A2 –wa m
and d y, see Table 2). Resul s o a con as ing scena io
(MIROC3.2(med es)/A1B –wa m and we ) ha e been gen-
e a ed (shown in Fig. S2). Figu e 4 illus a es he spa ial
pa e ns o he mos isk p one a eas o each o hese indi-
ca o s o scena io IPSL-CM4/A2 using p e-de e mined
h esholds (Hakala e al. 2012), as well as, hei o e lay o
h ee u u e ime slices (2011–2040, 2041–70 and 2071–
2100, espec i ely) (Fig. S3 shows his o scena io MI-
ROC3.2(med es)/ A1B). Figu e 5 uses wo indica o s (Ra-
inAS and S essE) o illus a e indica o disc epancies
esul ing om di e ences in he clima e p ojec ions o
IPSL-CM4/A2 and MIROC3.2(med es)/A1B.
Figu e 4 shows o scena io IPSL-CM4/A2, pe iod
2011-40, ha he mos isk p one a eas a e ound along
he wes and sou h coas , mainly due o ea ly d ough
s ess alling below he c i ical alue ( h eshold 39.4 mm).
Towa ds he middle o he cen u y (2041–2070), isk-
p one a eas expand inland om he wes and sou h,
whils in some smalle a eas o sou h-eas e n Finland
bo h ea ly d ough and educed yield po en ial isk com-
bine, ende ing hese a eas (nea U i, see Fig. 1) he mos
isk p one in his scena io. By he end o he cen u y
(2071–2100), highe isk a eas a e widesp ead, co e ing
mo e han 70% o he coun y due o exceedance o
h esholds o bo h speci ic hea s ess and educed yield
po en ial in mos a eas, whils a eas whe e all h ee
isk ac o s exceed he h eshold a e ound mainly in
sou h-eas e n Finland. The pic u e di e s dis inc ly o
MIROC3.2(med es)/ A1B (Fig. S3).
Figu e 6 (colou ed ables wi h a design modi ied om
T nka e al. 2011) conside s a se o 10 ag oclima ic indi-
ca o s, which ha e a ied ele ance ac oss a ange o
c ops including ba ley. Resul s a e shown o a sample o
i e clima e change scena ios (ou o 32) ha span he
ange o clima e scena io ealiza ions, hus e ealing
he unce ain y ange in impac p ojec ions. Resul s o
he median changes o 10 ag oclima ic indica o s a e gi en
o h ee ime slices (a,b,c) i e clima e scena ios, and
eigh loca ions. The able illus a es conside able di e -
ences in indica o alues o he di e en clima e scena -
ios, and also shows la ge di e ences be ween loca ions.
O e all, esul s sugges ha condi ions o pe ennial c ops
4202 ª2013 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d.
Shi s in Ag oclima e and Cul i a Response R. P. R€
o e e al.
1971 - 2000 2011 - 2040 2041 - 2070 2071 - 2100
DelayS20412070
Days
<= –30
–29 - –20
–19 - –10
–9 - 0
1 - 10
11 - 20
> 20
DelayS20712100
Days
<= –30
–29 - –20
–19 - –10
–9 - 0
1 - 10
11 - 20
> 20
DelayS19712000
Days
<= –30
–29 - –20
–19 - –10
–9 - 0
1 - 10
11 - 20
> 20
DelayS20112040
Days
<= –30
–29 - –20
–19 - –10
–9 - 0
1 - 10
11 - 20
> 20
RainAS19712000
mm
10.1 - 20
20.1 - 30
30.1 - 40
40.1 - 50
50.1 - 60
60.1 - 70
70.1 - 80
> 80
<= 10
RainAS20112040
mm
10.1 - 20
20.1 - 30
30.1 - 40
40.1 - 50
50.1 - 60
60.1 - 70
70.1 - 80
> 80
<= 10
RainAS20412070
mm
10.1 - 20
20.1 - 30
30.1 - 40
40.1 - 50
50.1 - 60
60.1 - 70
70.1 - 80
> 80
<= 10
RainAS20712100
mm
10.1 - 20
20.1 - 30
30.1 - 40
40.1 - 50
50.1 - 60
60.1 - 70
70.1 - 80
> 80
<= 10
S essE19712000
Days
0
1
2
3
4
5
6
>= 7
S essE20112040
Days
0
1
2
3
4
5
6
>= 7
S essE20712100
Days
0
1
2
3
4
5
6
>= 7
S essE20412070
Days
0
1
2
3
4
5
6
>= 7
TempHRa g197100
°C
<= 10
10.1 - 11
11.1 - 12
12.1 - 13
13.1 - 14
14.1 - 15
15.1 - 16
> 16
TempHRa g207100
°C
<= 10
10.1 - 11
11.1 - 12
12.1 - 13
13.1 - 14
14.1 - 15
15.1 - 16
> 16
TempHRa g204170
°C
<= 10
10.1 - 11
11.1 - 12
12.1 - 13
13.1 - 14
14.1 - 15
15.1 - 16
> 16
TempHRa g201140
°C
<= 10
10.1 - 11
11.1 - 12
12.1 - 13
13.1 - 14
14.1 - 15
15.1 - 16
> 16
(D) Mean daily empe a u e accumula ion a e (pe day) du ing g ain illing
(C) Ex eme high empe a u e s ess (Tmax >= 28 °C) a ound heading
(A) Sowing da e (de ia ion om May 1s )
(B) Rain 3-7 weeks a e sowing indica ing d ough
Figu e 3. P ojec ed changes o (A) sowing da e (DelayS, de ia ions ela i e o ixed da e 1s May) and h ee ag oclima ic indica o s: (B) ea ly
d ough s ess (RainAS), (C) speci ic hea s ess (S essE), and (D) mean daily empe a u e accumula ion a e a g ain illing (TempHRA g, highe
alue signals highe likelihood o yield educ ion), o clima e scena io IPSL-CM4/A2. The legend cap ion con ains he abb e ia ion o he indica o
(see Table 1) and he obse ed ime pe iod (e.g., DelayS1140 =sowing da e exp essed as de ia ion om May 1s o he ime pe iod 2011–2040).
ª2013 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. 4203
R. P. R€
o e e al. Shi s in Ag oclima e and Cul i a Response
like g ass a e gene ally likely o become mo e a ou able
(g eene shading, as shown especially o indica o s 1–5)
excep o ex eme scena io IPSL-CM4/A2. On he o he
hand, o annual ield c ops like sp ing ce eals, oilseeds
and oo c ops, condi ions end o de e io a e ( edde
shading, as shown o indica o s 6–10). Fo sp ing-sown
annual ield c ops, howe e , he pic u e a ies mo e
and whe he condi ions become mo e o less a ou -
able depends a lo on he clima e scena io. The secula
a iabili y o he indica o “ea ly d ough s ess” is shown
o he cen al co-o dina es o ou di e en g id cells in
Fig. S4 – o h ee di e en clima e change scena ios up
o he end o he cen u y (30 yea ime slices 2011–2040,
2041–2070 and 2071–2100, espec i ely).
2011 - 2040
2041-2070
2071-2100
Indica o O e lay
Risk
High isk
Mode a e isk
Low isk
No isk
Indica o O e lay
Risk
High isk
Mode a e isk
Low isk
No isk
Indica o O e lay
Risk
High isk
Mode a e isk
Low isk
No isk
RainAS20712100
mm
>= 39.4
< 39.4
<= 14.5
S essE20712100
days
< 6
>= 6
S essE20112040
days
< 6
>= 6
TempHRa g20112040
°C
<= 14.5
> 14.5
TempHRa g20412070
°C
<= 14.5
> 14.5
TempHRa g20712100
°C
> 14.5
RainAS20412070
mm
>= 39.4
< 39.4
S essE20412070
days
< 6
>= 6
Regions wi h ain a e
sowing < 39. 4 mm
Regions wi h ex eme high
empe a u e s ess
(Tmax >= 28 °C a ound heading)
Regions wi h mean daily
empe a u e accumula ion
a e > 14.5 °C a g ain illing
Risk indica o o e la
y
RainAS20112040
mm
>= 39.4
< 39.4
Figu e 4. Spa ial pa e ns o he mos isk p one a eas o each o hese indica o s using p e-de e mined h esholds, as well as, he o e lay o all
h ee isk ac o s –IPSL-CM 4/A2 - o each o he h ee u u e ime slices (2011–2040, 2041–2070 and 2071–2100).
4204 ª2013 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d.
Shi s in Ag oclima e and Cul i a Response R. P. R€
o e e al.
Simula ed c op cul i a esponses o
changes in clima e, a mosphe ic CO
2
and
sowing da es
D awing om he 11 selec ed clima e scena ios (Table
S4), we ocused ini ially on an analysis o a “wo s -case”
scena io, which p ojec s he lowes p ecipi a ion wi h high
wa ming du ing summe (Ma ch–Augus ) by mid cen u y
- IPSL-CM4/A2 (Fig. S1). Simula ion esul s wi h he
WOFOST model o his scena io a e illus a ed o ou
g id cells (Fig. 7A-D), ep esen ing clima ic condi ions o
he ou main concen a ion a eas o ba ley cul i a ion
(Fig 1.). Resul s, p esen ed o h ee cu en ba ley cul i-
a g oups assuming po en ial p oduc ion (i.e., no limi a-
ions o nu ien s o wa e ), and o wo soil ypes unde
wa e -limi ed ( ain ed) p oduc ion, indica e some com-
mon ea u es, bu also dis inc di e ences among he
ou loca ions. A common cha ac e is ic is he ela i ely
mino a ia ion in simula ed po en ial yield le el ( ange
6.4–7.2 /ha) o e he en i e simula ion pe iod, 1971–2100.
One e ec o u u e wa ming is a has ening o pheno-
logical de elopmen , sho ened g ow h cycle and a educ-
ion o biomass and g ain yield. Coun e ac ing his a e
he posi i e e ec s o CO
2
e iliza ion and ea lie sowing
on yield o ma ion, bu o po en ial p oduc ion hese a e
no su icien o compensa e o de elopmen - ela ed
losses a all g ow h s ages and loca ions. Only g id cell
“Ylis a o” shows a sligh inc ease in po en ial yields in he
second hal o he cen u y compa ed o he baseline. In
con as , changes in yields a ainable unde ain ed
condi ions (a ainable yield) show mo e he e ogenei y
o esponse, wi h he main a ia ion in yield decline
a ibu able o soil ype, hough loca ion also has a mino
con ibu ion o his a ia ion.
Fo bo h soil ypes, clay loam and sil y sand, he gap
be ween po en ial yield and a ainable yield widens wi h
ime a all ou loca ions. Fo clay loam, ha gap is small-
es o Ylis a o, whe e i is negligible unde baseline
clima e wi h cul i a Annabell, bu wi h a clea di e en i-
a ion among cul i a s. A he end o he cen u y, he e is
a gap o abou 1 /ha bu ha dly any di e ence in yield
esponses among cul i a s. A he same loca ion, he e is
also qui e a small yield gap o sil y sand du ing he base-
line pe iod 1971–2000. Howe e , his ollows a apid yield
decline du ing i s hal o his cen u y, also showing
some di e ences among cul i a s a he end o he
cen u y. The simula ed yield pa e n o e ime ound o
Ylis a o mos esembles ha ound a g id cell “Oulu”,
hough po en ial yields a Oulu a e somewha lowe .
Fo he Jokioinen and U i g id cells, loca ed u he
sou h, he decline o a ainable yield is mo e linea . The
yield decline and gap o po en ial yield a e sligh ly la ge
a Jokioinen han a U i, whils di e ences in cul i a
esponses on he clay loam disappea wi h ime.
Gene ally, yield gaps be ween simula ed po en ial and
a ainable yields g ow om 1 /ha (Ylis a o, clay loam)
and 2.3 /ha (U i, sil y sand) unde baseline clima e, up
o 4.2 /ha, o abou 45% o po en ial yield (Jokioinen
and Ylis a o, sil y sand) a he end o he cen u y.
Fo one loca ion we conside ed a wide di e si y o
(ele en) clima e change scena ios (ou o 32) in simula -
ing he esponse o cu en ba ley cul i a s o changes cli-
ma e and a mosphe ic CO
2
du ing 2071–2100. Fo
simpli ica ion, we assumed an a mosphe ic CO
2
concen-
a ion o 615 ppm o all clima e scena ios ( o de ails,
see Table S7). To examine how he h ee di e en cul i a
g oups espond, we chose a a ou able clay loam and sin-
gle g id cell (U i) loca ed in he a ea mos p one o high
empe a u e and d ough isk (see Fig. 4).
Figu e 8 illus a es ( o he clay soil a U i) how
di e ences in clima e change impac s on ba ley yields
we e g ea e be ween clima e scena ios han be ween
cul i a g oups. Fo he la e ma u ing g oup (Annabell),
a e age yields o he wo s case clima e scena io (i.e.,
IPSL-CM4/A2) a e 1.2 /ha lowe han o he baseline
clima e (6.4 /ha), bu hey a e mo e han 1.8 /ha highe
o he bes case clima e scena io (i.e., GISS-ER/B1)
(Table 2).
Eigh ou o he o he nine clima e scena ios esul in
highe a e age yields and yield a iabili y. Fo he med-
ium ma u ing (Kus aa) and ea ly ma u ing (Kunna i)
cul i a s, yield educ ions o IPSL-CM4/A2 a e less p o-
nounced, while yield inc eases unde GISS-ER/B1 a e o
he same o de o magni ude (nea ly 2 /ha) as o
Annabell. Fo hese wo cul i a g oups, as o Annabell,
BOR4
BOR3
BOR1
BOR6
ALN1
BOR1
NEM1
BOR8
BOR4
BOR3
BOR1
BOR6
ALN1
BOR1
NEM1
BOR8
RainAS20712100
–5 - 5
5 - 15
> 15
–15 - –5
S essE20712100
0
1
2
–2
–1
Figu e 5. Di e ences in p ecipi a ion sum 3–7 weeks a e sowing
(RainAS) in mm, and e y high empe a u e s ess (S essE) in days,
be ween MIROC3.2(med es)/A1B and IPSL-CM4/A2 scena io. The wo
di e ence maps show he de ia ion alues o MIROC3.2 ela i e o
IPSL-CM4.
ª2013 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. 4205
R. P. R€
o e e al. Shi s in Ag oclima e and Cul i a Response
Fische , G., M. Shah, F. N. Tubiello, and H. an Vel huizen.
2005. Socio-economic and clima e change impac s on
ag icul u e: an in eg a ed assessmen , 1990–2080. Philos.
T ans. R. Soc. B 360:2067–2083.
F onzek, S., T. R. Ca e , and K. Jylh€
a. 2012. Rep esen ing wo
cen u ies o pas and u u e clima e o assessing isks o
biodi e si y in Eu ope. Glob. Ecol. Biogeog . 21:19–35.
God ay, H. C. J., J. R. Bedding on, I. R. C u e, L. Haddad,
D. Law ence, J. F. Mui , e al. 2010. Food secu i y: he
challenge o eeding 9 billion people. Science 327:812–818.
G aux, A.-I., G. Bellocchi, R. La dy, and J.-F. Soussana. 2013.
Ensemble modelling o clima e change isks and
oppo uni ies o managed g asslands in F ance. Ag ic. Fo .
Me eo ol. 170:114–131.
Hakala, K., L. Jauhiainen, S. J. Himanen, R. R€
o e , T. Salo,
and H. Kahiluo o. 2012. Sensi i i y o ba ley a ie ies o
wea he in Finland. J. Ag ic. Sci. 150:145–160.
Ha ison, P. A. and R. Bu e ield. 1996. E ec s o clima e
change on Eu ope-wide win e whea and sun lowe
p oduc i i y. Clim. Res. 7:225–241.
Himanen, S., E. Ke oja, K. Hakala, R. R€
o e , T. Salo, and
H. Kahiluo o. 2013. Cul i a di e si y has g ea po en ial o
inc ease yield o eed ba ley. Ag on. Sus ainable De .
33:519–530.
an I e sum, M. K., P. A. Le elaa , H. an Keulen,
M. J. K op , L. Bas iaans, and J. Goud iaan. 2003. On
app oaches and applica ions o he Wageningen c op
models. Eu . J. Ag on. 18:201–234.
Kangas, A., A. Laine, M. Niskanen, Y. Salo, M. Vuo inen,
L. Jauhiainen, e al. 2006. Vi aliis en lajikekokeiden ulokse
1999–2006 (Resul s o o icial a ie y ials 1999–2006),
MTT:sel i yksi€
a 132 (in Finnish and English language).
MTT, Jokioinen, Finland.
Lobell, D. B. and M. B. Bu ke. 2010. On he use o s a is ical
models o p edic c op yield esponses o clima e change.
Ag ic. Fo . Me eo ol. 150:1443–1452.
Mand yk, M., P. Reidsma, and M. K. I e sum. 2012. Scena ios
o long- e m a m s uc u al change o applica ion in
clima e change impac assessmen . Landscape Ecol.
27:1–19.
Maye , K. F. X., R. Waugh, P. Lang idge, W. B own,
A. Schulman, M. Pla ze , e al. 2012. A physical, gene ic and
unc ional sequence assembly o he ba ley genome. Na u e
491:711–717.
Meehl, G. A., C. Co ey, K. E. Taylo , T. Delwo h,
R. J. S ou e , M. La i , e al. 2007. The WCRP CMIP3
mul imodel da ase - A new e a in clima e change esea ch.
Bull. Am. Me eo ol. Soc. 88:1383–1394.
Me zge , M. J., R. G. H. Bunce, R. H. G. Jongman,
C. A. M€
uche , and J. W. Wa kins. 2005. A clima ic
s a i ica ion o Eu ope. Glob. Ecol. Biogeog . 14:549–563.
Mi chell, T., T. R. Ca e , P. Jones, M. Hulme, and M. New.
2004. A comp ehensi e se o high- esolu ion g ids o
mon hly clima e o Eu ope and he globe: he obse ed
eco d (1901–2000) and 16 scena ios (2001–2100). Tyndall
Cen e Wo king Pape 55. Tyndall Cen e, No wich, UK.
Nakiceno ic, N., J. Alcamo, G. Da is, B. de V ies, J. Fenhann,
S. Ga in, e al. 2000. Emissions scena ios. A special epo
o wo king g oup III o he in e go e nmen al panel on
clima e change. Camb idge Uni . P ess, Camb idge, U.K.
Nelson, G. C., M. W. Roseg an , J. Koo, R. Robe son,
T. Sulse , T. Zhu, e al. 2009. Clima e change: impac on
ag icul u e and cos s o adap a ion. Food Policy Repo .
In e na ional Food Policy Resea ch Ins i u e, Washing on
DC.
Olesen, J. E., M. T nka, K. C. Ke sebaum, A. O. Skjel

ag,
B. Seguin, P. Pel onen-Sainio, e al. 2011. Impac s and
adap a ion o Eu opean c op p oduc ion sys ems o clima e
change. Eu . J. Ag on. 34:96–112.
Osbo ne, T., G. Rose, and T. Wheele . 2013. Va ia ion in he
global-scale impac s o clima e change on c op p oduc i i y
due o clima e model unce ain y and adap a ion. Ag ic.
Fo . Me eo ol. 170:183–194.
Pa y, M. L., C. Rosenzweig, A. Iglesias, M. Li e mo e, and
G. Fische . 2004. E ec s o clima e change on global ood
p oduc ion unde SRES emissions and socio-economic
scena ios. Global En i on.Change 14:53–67.
Pel onen-Sainio, P., L. Jauhiainen, and K. Hakala. 2009. A e
he e indica ions o clima e change induced inc eases in
a iabili y o majo ield c ops in he no he nmos
Eu opean condi ions? Ag ic. Food Sci. 18:206–222.
Po e , J. R. and M. Gawi h. 1999. Tempe a u es and he
g ow h and de elopmen o whea : a e iew. Eu . J. Ag on.
10:23–36.
Reye , C. P. O., S. Leuzinge , A. Rammig, A. Wol ,
R. P. Ba holomeus, A. Bon an e, e al. 2012. A plan ’s
pe spec i e o ex emes: e es ial plan esponses
o changing clima ic a iabili y. Glob. Change Biol. 19:
75–89.
Rosenzweig, C. and D. Hillel (1998) Clima e change and he
global ha es : po en ial impac s o he g eenhouse e ec on
ag icul u e. Ox o d Uni . P ess, New Yo k, Pp. 336.
Rosenzweig, C. and M. L. Pa y. 1994. Po en ial impac o
clima e change on wo ld ood supply. Na u e 367:133–138.
Rosenzweig, C., J. W. Jones, J. L. Ha ield, A. C. Ruane,
K. J. Boo e, P. Tho bu n, e al. 2013. The Ag icul u al Model
In e compa ison and Imp o emen P ojec (AgMIP):
p o ocols and pilo s udies. Ag ic. Fo . Me eo ol. 170:166–182.
R€
o e , R. P. and S. C. an de Geijn. 1999. Clima e change
e ec s on plan g ow h, c op yield and li es ock. Clim.
Change 43:651–681.
R€
o e , R. P., T. Palosuo, N. K. Pi ioja, M. Dub o sky,
T. Salo, S. F onzek, e al. 2011a. Wha would happen o
ba ley p oduc ion in Finland i global wa ming exceeded
4°C? A model-based assessmen . Eu . J. Ag on. 35:205–214.
R€
o e , R. P., T. R. Ca e , J. E. Olesen, and J. R. Po e .
2011b. C op-clima e models need an o e haul. Na . Clim.
Change 1:175–177.
4212 ª2013 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d.
Shi s in Ag oclima e and Cul i a Response R. P. R€
o e e al.

R€
o e , R. P., T. Palosuo, K. C. Ke sebaum, C. Angulo,
M. Bindi, F. Ewe , e al. 2012. Simula ion o sp ing ba ley
yield in di e en clima ic zones o No he n and Cen al
Eu ope. A compa ison o nine c op models. Field C ops
Res. 133:23–36.
Rummukainen, M. 2010. S a e-o - he-a wi h egional clima e
models. WIREs Clim. Change 1:82–96.
Ruos eenoja, K., J. R€
ais€
anen, and P. Pi inen. 2011. P ojec ed
changes in he mal seasons and he g owing season in
Finland. In . J. Clima ol. 31:1473–1487.
Semeno , M. A. and N. G. Hal o d. 2009. Iden i ying a ge
ai s and molecula mechanisms o whea b eeding unde
a changing clima e. J. Exp. Bo . 60:2791–2804.
Slo h Madsen, M., C. Fox Maule, N. MacKella , J. E. Olesen,
and J. Hesselbje g Ch is ensen. 2012. Selec ion o clima e
change scena io da a o impac modeling. Food Addi .
Con am. Pa A 29:1502–1513.
Smi h, P. and J. E. Olesen. 2010. Syne gies be ween he
mi iga ion o , and adap a ion o, clima e change in
ag icul u e. J. Ag ic. Sci. 148:543–552.
Teixei a, E. I., G. Fische , H. an Vel huizen, C. Wal e , and
F. Ewe . 2013. Global ho -spo s o hea s ess on
ag icul u al c ops due o clima e change. Ag ic. Fo .
Me eo ol. 170:206–215.
T nka, M., J. E. Olesen, K. C. Ke sebaum, A. O. Skjel

ag,
J. Ei zinge , B. Seguin, e al. 2011. Ag oclima ic condi ions
in Eu ope unde clima e change. Glob. Change Biol. 17:
2298–2318.
Uppala, S. M., P. W. K

allbe g, A. J. Simmons, U. And ae,
V. Da Cos a Bech old, M. Fio ino, e al. 2005. The ERA-40
e-analysis. Qua . J. R. Me eo ol. Soc. 131:2961–3012.
Ven€
al€
ainen, A., H. Tuomen i a, P. Pi inen, and A. D ebs.
2005. A basic Finnish clima e da a se 1961–
2000-desc ip ion and illus a ion. Finnish Me eo ological
Ins i u e Repo s 5. Finnish Me eo ological Ins i u e,
Helsinki, Finland, Pp. 27.
Whi e, J. W., G. Hoogenboom, B. A. Kimball, and
G. W. Wall. 2011. Me hodologies o simula ing impac s o
clima e change on c op p oduc ion. Field C ops Res.
124:357–368.
Ylh€
aisi, J. S., H. Tie €
a €
ainen, P. Pel onen-Sainio,
A. Ven€al€ainen, J. Eklund, J. R€ais€anen, e al. 2010. G owing
season p ecipi a ion in Finland unde ecen and p ojec ed
clima e. Na . Haza ds Ea h Sys . Sci. 10:1563–1574.
Suppo ing In o ma ion
Addi ional Suppo ing In o ma ion may be ound in he
online e sion o his a icle:
Da a S1. Me hods, Resul s and Re e ences
Me hods
Table S1. Ag oclima ic indica o s gene a ed by N-Ag iC-
LIM.
Table S2. Goodness-o - i o he models o p edic ing
ba ley yields om ag oclima ic indica o s and o de o
a iables’ en y (using s epwise selec ion) p esen ed o
each ial si e.
Table S3. E ec s o he es ed ag oclima ic indica o s on
ba ley yields a Jokioinen, Ylis a o and Ruukki.
Table S4. Lis o Gene al Ci cula ion Model (GCM) sim-
ula ions downloaded om he CMIP3 a chi e (Meehl
e al. 2007) o h ee SRES emission scena ios (B1, A1B,
A2) (Nakiceno ic e al. 2000) o which all a iables
equi ed o cons uc scena io da a o c op modeling
we e a ailable.
Table S5. Obse ed and scena io a iables used o c op
modeling (c . Wol e al., 2012).
Table S6. The mal equi emen s [°C day] o h ee mod-
e n ba ley cul i a g oups om eme gence o lowe ing
(TSUM1), and om lowe ing o physiological ma u i y
(TSUM2); assuming a common base empe a u e
(TBASE) o 0°C - and indica ion o di e ences in o he
c op pa ame e s.
Table S7. Changes o c op pa ame e se (uni o m o all
sp ing ba ley cul i a g oups) o di e en CO
2
le els:
Speci ic lea a ea (SLA), maximum CO
2
assimila ion a e
(AMAX occu ing o e indica ed de elopmen s age
(DVS)) and co ec ion ac o o po en ial e apo anspi a-
ion (CFET) unde e e ence clima e (350 ppm ) and
enhanced (435, 525 and 615 ppm ) a mosphe ic CO
2
concen a ion (wi h % changes in ela ion o cu en
le el).
Resul s
Figu e S1. P ojec ed changes in mean empe a u e and
p ecipi a ion du ing Ma ch-Augus ela i e o he baseline
clima e (1971–2000) p esen ed o he ime pe iods 2011–
2040, 2041–2070 and 2071–2100 o selec ed loca ions
(Tu ku (1), Jokioinen (2), U i (3), Ylis a o (4), Oulu (5),
Ro aniemi (6)) ep esen ing he en i onmen al zones
mos ele an o ag icul u al p oduc ion in Finland (see
Fig. 1).
Figu e S2. P ojec ed changes o (A) sowing da e (de ia-
ions ela i e o ixed da e 1s May) and h ee ag oclima ic
indica o s: (B) ea ly d ough s ess, (C) speci ic hea s ess,
and (D) yield po en ial educ ion isk, o clima e scena io
2 (wa m and we ), combining SRES emissions scena io
A1B wi h MIROC3.2 (med es) (see Table S4).
Figu e S3. Spa ial pa e ns o he mos isk p one a eas
o each o hese indica o s using p e-de e mined h esh-
olds, as well as, he o e lay o all h ee isk ac o s –MI-
ROC3.2(med es)/A1B - o each he h ee u u e ime
slices (2011–2040), (2041–2070) and (2071–2100).
Figu e S4. Ea ly d ough s ess (Rain sum 3–7 weeks a e
sowing) p esen ed as 10-yea mo ing a e age unde cu -
en (1971–2009) and p ojec ed u u e clima e condi ions
(2011–2040, 2041–2070, 2071–2100) applying del a change
ª2013 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d. 4213
R. P. R€
o e e al. Shi s in Ag oclima e and Cul i a Response
me hod and p ese ing he a iabili y o he e e ence cli-
ma e (1971–2000) o g id cells (A) Jokioinen, (B) U i,
(C) Ylis a o, (D) Oulu, ep esen ing he en i onmen al
zones mos ele an o ag icul u al p oduc ion in Finland
(see Fig. 1) Clima e change p ojec ions based on h ee
GCMs 9SRES combina ions: GISS-ER/B1, CCCMA-
CGCM3.1 (T63)/A1B and IPSL-CM4/A2 (see, Table S4).
Re e ences Addi ional e e ences o Da a S1
4214 ª2013 The Au ho s. Ecology and E olu ion published by John Wiley & Sons L d.
Shi s in Ag oclima e and Cul i a Response R. P. R€
o e e al.