.-ll1l1all oJ (;Iaciolog), 2 1 1995
C Int e rn a ti o na l Gl acio logical Soci e t\'
Sensitivity experiIllents to sea surface teIllperatures,
sea-ice extent and ice-sheet reconstruction, for the
Last Glacial MaxiDluDl
G. R ,\\ISTEL\
Laboraloire de .Iiodel/salioll d" Chl1wl el de l'EIIl'irollllelllenl CE \ . Sad(~)' . Bal . 709. GiFl'1Ir-Jioelle 9 119 1, Frallce
S. JO USSA Ui\IE
La/Joraloire de .l/odel/salioll dl! Cl/owl 1'1 de l'EIIl'irollllclI7en l alld LOD1 (;, C \ RSjClliz'ersile P. & ill . Cll rie jORS TO .I/,
Parij , Prallce
ABSTRACT. Fo r th e L as t Gl ac ia l .\l ax imum, ( LG:'l:
2 1 000 BP ). simul a ti o ns
using atmosph e ri c ge nera l- circ ul a ti o n models \ AG C \l s a rc I'er y se nsilil'C to th e
presc ribed bo und a ry co nditi ons. \l os t o f th e rece nt num e ri ca l ex p erim e nls ha lT used
rh e CL I \IAP ( 198 1) d a ta se t fo r ice-shee t topogra ph y. sea -i ce extenl a nd sea surface
tempera tures (SSTs). T o d e m o nstra te th e impa ct o f ice-sheet reco nstru ct io n o n th e
LG \l clim a te, \\'e perfo rm ed tl\'O simul a ti o ns: o nc usin g CLI\IAP ( 198 1) ice-shee t
to pogra ph y, th e o th e r using th e nel\' reco nstru c ti o n pro"id ed by Pe lti er. " 'e sho l\' th a l,
a lthoug h th e geog ra phi ca l structure of the annu a ll y a\'C'J'aged temperature is no t
modified , th ere a re impo rtant seaso nal and regio nal impac ts on th e temp era ture
distribution. In a seco nd step. to a na lyze th e efl(~c ts of coo le r SSTs a nd sea-i ce ext e nt,
\\'e pe rform ed Cl simul a ti o n using CLr~IAP ( 198 1) fo r th e ice-shee t to pog raph y. but
I\'ith prese nt SSTs. W e find th a t th e coo lin g du e to ice shee ts fo r th e L G i\1 c lim a te is
o ne-third of th e g lo ba l a nnu a ll y a\T raged coo ling , a nd th a t th e so uth\\'a rd shift of th e
~ o rth Atl a nti c 10 1\' in \I'inter is not du e to sea -i ce ex tent , but is a n o rogra phi c e flec t
du e to th e L a urelllid e ice shee t. Thi s se t o f se nsitil'it y ex pe rim e nt s a ll o \\'s us a lso to
di scrimin a te be t\l'een th erm a l a nd QI'ogra phi c fo rcin gs a nd to sho \,' th e impac t of th e
ice-sheet topogra ph y a nd cooler SSTs o n th e pa tt ern of pl a ne ta r y \I'a l'es durin g th e
LGi\I clim a te,
1. INTRODUCTION
Atm os ph e ri c ge ne ra l-c irc ul a ti o n mod els (AGC:'I[s i a re
impo rt a nt lOo ls [o r reco ll structin g pas t a tm os pheri c
circ ul a ti o ns a nd stud yin g th e ph ys iea l m ec h a nisms of
clim a te c ha nge. T o illl'Cs tiga te th e ro le of th e difTe re nt
co mpo nents (or bo und a r y co nditi o ns of'th e Las t Gl ac ia l
\I ax imum (LG\I ) clim a te, we perfo rm ed a se t of (Q UI'
num eri ca l expe rim ents, using th e L a bo rato ire d e i\I etcoro logie D yna miqu e ( L~ID ) AGCi\1. i\I a n y simul a tion s
ha l'C bee n d o ne \\'ith difTe re nt !\GCi\ Is to qu a ntify th e
c flec t o f' bo und a r y co nditi o ns o n th e g lo ba l tempe ra ture
d ec rease durin g th e LG?lI. Fo r exa mpl e, Brocco li and
~l a n a b e \ 1987 ). using a n AG C :-- I co upled \\'ith a slab
ocea n m od el, emph as ized th e ro le of lo\\' C O 2 partial
press ure, es pec ia ll y fo r th e So uth e rn H e mi sphere. \I'hil e
Rind ( 1987 ) sho \\'ed th e impac t of sno\l' feed-b ac k a nd o f
th e to pog raph y of th e ice shee ts,
Our approac h fo c uses o n th e ro le of t\l'O m aj o r
bo und a r y co ndition s:
s urf ~l ce
tempera tures (SSTs) as gil'e n b y CLI(\lclnt yre a nd C lin e, 198 1; he rea ft er ca ll ed
C Ll ~lAP ( 198 1)) in \lin ter a nd summ er fo r the LG "'I
a nd m od ern clim a tes ,
I , Sea
~I:\P.
2, lee-shee t recon stru c ti o ns (ISRs) fo r whi ch (\\' 0 d a ta
se ts a re Cl I'a il a bl e, CLl\IAP ( 198 1) a nd a rel'ised o ne
prm 'id ed bl' Peltier ( 199+),
So me se llsiti \'ity ex perim ents \I'ith SST a nd ISR ha lT
a lread y bee n pe rfo rm ed, Rind ( 1987 ) studied th e impac t
o rCLI\ IAP ( 1981 ) LG\] SSTs under prese nt clim a ti c
co nditi o ns, a nd Rind a nd Petee t ( 1985 ) m a d e a sensitil'ity
expe riment uniformll' lowe rin g th e CLIMAP ( 198 1)
LGi\l SST b y 2 c C to impro' T th e co nsistency o f' 1011' la titud e te mperatures d edu ced fr o m d a ta and fr o m
m odels; bu t th ere is still no st ud y of presen t-c1im a te
SST impac t o n LG \ ! d yna mi cs ,
On th e o th er ha nd . Shinn a nd Ba rro n 1989 i ha l'e
studi ed th e se nsitil'it y of th e LG\I clim a te to extrem e
co ntin e ntal in>-sh ee l size a nd co nfig ura ti o n, In th eir
ex pe rim e nl , th e difTe re nces be t\\' ee n m ax imum a nd
minimulll ice sheets a re hu ge in co mpa riso n \I'jth o ur
simul a ti o n ; ye t so m e inte res tin g co mpa ri so ns. es p ec ia ll y
abo ut th e splittin g of the polar j et du e to th e L a urentid e
ice shee t, can be m a d e, ~ I o l 'Co \' e r. th e fa ct that all th e
mod ell e rs il1\'o h 'Cd in th e Pa leoc lim a te \[od eling Interco mpariso n Proj ect (perso na l co mmuni catio n fr o m S,
J o ussa um e a nd K. T aylo r, 1994 ) will use th e ne\\' iceshee t reco nstru c ti o n prOl'id ed by Pel ti er 199+) reinfo rces
343
Ralllsleill al/d ]ollSSallll7e: SfIIsilil'if)' njJerilllelllsJor Ihe Lasl Glacial .I/rl\imulI1
inte lTs t in \\'hether c ha ng in g the ISR has a n importa nt
impact o n the LC\( c lim a te,
2. MODEL AND SIMULATION DESCRIPTION
Al l th e simulation s were p e rfo rm ed w ith th e Li-.1D \'e rsio n
Tre ut a nd Li , J99 1) which is a gr id -point
m ode l. \\'e used the low-reso luti o n \'e rsio n, corresponding
to 48 po ints regu la rl y sp aced in lo ng itud e a nd 36 points
regu larly spaced in sin e or lat itud e , At th e equ aror. th e
grid spac in g is 830 km o f'l o ng itud e x 270 km of'l a titude .
a nd a t 45 of latitud e it co rrespon d s to 590 km lo ng itud e
x 530 km la tirud e. Th e m od e l h as II !c\'e!s in n orm a li zed
press ure coordinate (a ) , The m ode! so h-cs the d yn a mi c
and th e rm od y n a mi c equations, a nd th e co ntinuity
equat io n fo r m ass and water \'a p o ur (Sa dou r n y ane!
L a\'a l, 1984) , Th e predicted clo ud s a rc co nsid ered fo r the
ra di a ti\'e-tra nsfe r ca lcu la ti o n s ( Le Treut a nd L a \' a l,
1984 ). Each ex perim ent in clud es th e [u ll seaso n a l cycle
a nd lasts 6 yea rs; a\'e rages a rc m a d e o\'e r th e las t 5 yea rs.
The first of th e fo ur simu lations ( E~PI ) is th e co ntrol
run lI'hi ch simula tes th e present clim ate. Th e second
(EXP2 ) is a simul ation of th e LC:' I using th e CLli\ [AP
( 198 1) d a ta se t [or SST a nd l SR. Th e third (EXP3 ) is a
sc nsiti\it )' ex perim ent with th e ice-shee t tOpography in
wh ich 11 (' rep laced th e CUl\lAP ( 198 1) reconstruction b y
the nell' reconstruction prO\'ided by Peltier (1994): a ll rhe
oth er bo undary co nditio ns are identical to EXP2, The
fo u rt h simu lat ion (EXP4 ) is a sensiti\'it)' ex perim ent with
SST in II'hich th e CLl \ IAP ( 198 1) reco nstru cted SST [or
th e Last Glacial :-' I ax imum has been replaced by the prese nt
clim a te SSI' of the cO lltrol run , lI'ith a ll th e o th er boundan'
conditi ons rema ining id en ti cal to EXP2 (T able I ) ,
Co mp a rin g EXP3 a nd E~P2 lI' ill gi\'e u s th e
se nsiti vit y of the LCJ\ I clim a te to th c differences in ice sh ee t reco nstructi o n, Comparing EXP4 a nd EXPI which
bo th have the sa me presc rib ed SST wil l g i\'e us th e impact
of' th c ice-shee t to p ograp h )'. Fin a ll y, co mparing EXP+
a nd EXP2, \\'hi ch both ha\'(' the sa m e CLIl\IAP ( 198 1)
[SR , will g i\'e us th e SST impact on t1w LC \l climate,
+ J\CCi\ L (L e
3. CHANGES IN BOUNDARY CONDITIONS
3.1 Ice-sheet reconstruction
A maj o r difTe rence betwee n present a nd LGJ\ I clim a te is
th e hu ge ice sheets co \'C rin g n o rthw es t e rn Europe
(Fennoscandia ) a nd th e no rth ern part of' North America
(La urentid e ), Th e , hape a nd sizc o r th ese ice shee ts ha\'e
a n impo rt an t impac t o n th e a tm osp heri c d ynam ics O\'C r
the Northern H emisph e re durin g th e LC .\!. D elll on a nd
H ug hes ( 1981 ) pro\' id ed fo r CUJ\ [ i\ P (\\'0 reconstru ctions
refe renced as J\IIi\' and J\ IAX, co rres p o ndin g to a e Llsta ti c
sea-b'c1 ri se or 127 a nd 163 m , res pec ti\'e! y. Pelti er ( 1994 )
sholl's th a t e\'Cn th e 1\ 11 N reco nst run ion was excessi \'e, and
proposes a new reco nstru ctio n b ascd on a g ra \'i ta ti o na ll y
se lf~ co n s i s t e llt th eo ry o r relat i\'e sea -Ie\'C1 changes \\'hi c h
co rres po nd s to a e usLa ti c sea-l e\'c1 rise o f 105 m. and to a
55'10 redu ctio n of the ice \'o lum c in co mparison lI'ith t he
CL TJ\ IAP J\IAX reco nstru c ti o n . Fi g ure I a a nd b sholl' th e
ch a nge in ele\'a tio n betll'een th e LCJ\ 1 and th e p resen t
climate [or th e CLI J\ [AP a nd Pclti er reconstructio ns. Th e
ex tent oj' ice sh ee ts is no t \'Cry difTcrent , but th e ele\'ation is
1500 111 h igh er O\'e r Lau re ntidc and abou t 1000 m hi g he r
ove r Fennosc andia in th e CL I J\ IAP reco nstr uc tion. Th e
~
j
·Ieo
· 120
~
a
b
0
60
120
180
60
120
180
LONGITUDE
·' 80
· 120
~
0
LONGITUDE
Fig. I, DiJfereJlces in elel'alion ( 111111 ) belu.'een /he presenl
( EXP l) alld Ihe Le:,I J, using ( a) CLL ILIP (1981) Jar
U;.I / ice- sh eellojJograJ)/~) ' ( EXP2) , ( b) Pellier ( 199-1 )
Ior I,G, I! ice-sllfl'l lojJogmjJII)' ( '~·. \'P3) , Dark gr~) '
(orresj)ollds 10 elevalioll differences qf 1I10re Ihall 2000 Ill,
l/IediuIII grl!)' 10 elel'alioll differences oJ more /h all 1000 Ill,
figh l grq 10 Ihe emerged lallds al Ihe Le:,\[, /sol/lies Jar
1500 alld 500 m are also drm('II,
Table 1, Jlain jJarameler differences Jar Ihe SI'/ oJ 1/l/llIerica/ el/Jerill1enls
co:?
[molalioll
SST dala sel
[SR dala
.1'1'/
ppm
EXPl co ntro l run
EXP2-LG:-.[ CLL\l AP ( 1981 )
EXP3-LG:-'I P elti er ( 1994)
EXP4-LCl\l (SST present )
34+
320
200
200
200
1950
21 ka BP
2 1 ka BP
2 1 ka BP
Prese nt
Prese nt
CLIJ\ I AP ( 198 1) C LIi\IAP ( 1981 )
C LTMAP ( 198 1)
P e lti er ( 1994 )
Prese nt
eLl .\ [A P ( I 98 I )
Ralllsleill alld ] Ollssallllle: Smsilil'il), el'periJl1f17lsfor Ihe Lasl Glacial JIrlli1ll1l17l
10\l'e r sea-l evel rise in Pelti er's reco nstru cti o n. co mpa red
Il'ith CLE1AP, leads to a redu c ti o n of th e emerged la nd s.
\\·C ta ke this int o acco unt a nd use difTerent sea /land m as ks
ro r th e lll'O simul a ti o ns,
3.2 Sea surface te:rnperatures
l\ I os t or th e recent simu la tio ns of' th e L G l\I cli m a te.
perlo r m ecl \I'ith AGC l\ Is a nd presc ribed SSTs. hal'c used
th e reco nstru ctio n 0 (' th e sea -i ce e" tent a nd o r th e sea
surface tempera tures, fo r summ er a nd winte r seaso ns.
prO\'id ed b y CUl\ l A P ( 1981 ) , T hese d a ta a re sti ll
co ntrO\'e rsia l, e.g , in t he tro pi cal a rea (\\'ebster and
Stre ten . 1978; R ind a nd Pe tee t. 1985 ), bu t no glo ba l
re\'isio n o r LG'\I SST is I'e l al'aila ble . a ncl \I'e use
C LIl\ LAP ( 198 1) to rcconstru c t d a il y cI'o lutio n 0 1' SST
and se a-i ce e" ten t.
Th e SST differen ces be tween th e LG 1\1 a nd th e
prese nt climate a re a b o ut 4° to 6"C fo r So uth ern and
N o rth ern H emi spher e mid -l a tit ud es . a nd reac h a ho ut
10- C fo r th e N o nh Atla ntic, durin g \I'inte!'. The coo lin g
o l'e r tro pica l areas is redu ced to 1 or 2"C. with la rge
a reas of th e Pa cifi c O cea n w here te mpera tures el'e n
c
increase sli g h tl y ( + 1° to 2 C),
Because of th e Berin g Stra it , th e difference in sea -Ice
e"te nt be twee n th e LGl\ r a nd th e prese nt cli m a te is sm a ll
in hi gh la titud es [or th e N o rth Pac iIi e. but th e re is a hu ge
equ a to[,lI'a rd shirt o r th e sea -ice ex tent rro m 70 lO 50
la titud e fo r th e N o rth Atlanti c a nd SOllth ern H emi sph ere.
Th e seaso na li ty of th e sea ice is \Veakened durin g th e
LG1\ l , and th e sea- ice ex tent is only slig htl y redu ced
dur ing th e summ er season,
0
4. DISCUSSION
4.1 Orographical i:rnpact of the ice sheet for the
LGM cli:rnate
As with mos t o r th e AC C l\ (s' simu la ti o ns of' th e LG'\l
(Rind , 1987 ; J o ussa um e, 1993 ), th e m a in fea tures a re
reproclu ced in EXP2 a nd EXP3: shin oCth e storm tra c ks
so uthwa rd ol'th e ice shee t, d eep ening of'th e N orth Pacifi c
10\1' a nd southll'<lrd shirt o r th e N o nh Atlantic 10\1'. a nd
spliltin g or th e j e t aro und th e L a ure ntid e ice shee l.
C o mp a ri so n o r EXP4 a nd EXPI , whi c h bo th ha l e th e
sam e present SSTs a nd sea- ice ex tent. di sp lays the impac t
of' th e o l'ogra phi ca l fo rcing of th e ice shee ts, I n EX P+,
th e re is no enh a nce m ent of th e A le utia n lo w (no t shown ).
thus prm' in g th a t t his e nh a nce m e nt. durin g th e LGl\ l, is
m os tl v du e to c han ges in SS']'s. I n contras t. th e
so uthll'a rd shift o r th e At la nti c lo ll' is o bse n 'ecl in
EXP2, EXP3 a nd EXP4; it is c lea rl y re la ted to th e
presen ce o f' the La ure n tid e ice shee t.
In Fig ure 2a (EXP4 m inu s EXP] ), we ill ustrate th e
e fTec t of o rogra phi ca l fo rcing o n th e geo po te n tia l heig h t a t
500 h P a, fo r th e lI'i nte r seaso n. Th e imp ac t o f' th e
orographical f()JT in g is ra t her loca l, leadin g to a nti cyc lo nic circ ul a tio n wes twa rd o r th e ice ca ps, in ag ree m ent
\lith H eld ( 1983 ) a nd Coo k a nd H eld ( 1988 ), T hese
m aj o r c h a nges in a tm os ph e ri c c irc ul a ti o n a re a lso
d e pe nd e n t o n th e size a nd config ura ti o n of' th e ice shee l.
as il lustra ted in neX l pa rag ra ph,
a
-180
-120
· 80
0
80
120
180
~
120
100
120
180
LONGITUDE
§
I::
j
b
c
- 180
- 120
~o
LONGITUDE
-110
-120
~
0
LONGITUDE
&0
Fig , 2, (;eojJolflllial heighl ( ill 11/ ) al 500h Pa dijprmre5
Jor December . j allllalJ alld Febmal), (a) belu'eeJl EXP.f.
alld EXPI . (b) belu'CClI E.\P 2 alld EXP3 , al/d ( e)
bell('el'll r;;XP2 alld t:_\.P .f. , i ll ( a) {(lid ( b) dark g r~]' is
liSI'd Ior dijJi'rence" /Iighl'/ tltall 200111 , light gr~]' Jor
difji'/"(' II!'e.l higltn I//{{II 100 Ill; i_Iolilles {(re drawlI Ior
differellm eqllal la 50 alld l jO 1/1.. i ll (() dark gr~l' is IIsed
Jor differellces lower Ihall - 200 111 . lighl gre),Jor dijJerfllces
lower lIiall lOO171 ; i,wlille,l (Ire drau'IIJor dijjereJlccs eqllal
lo50alld - 150 m,
4.2 Sens itivity to different ice-sheet
reconstructions for the LGM cli:rnate
Th e co mpa ri so n bc t\lTcn EXP2 and EXP3 re\Ta ls th e
c ha nges in LCl\ ( a tm os ph eri c c irc ula ti o n du e to th e
difTe re nccs in ice-shee t reco nstru ctio ns, A lt ho ug h th e
g loba ll\' a nci a nnu a ll y a \T ragecl coo lin g is identical fo r
th e t\l'O LG;\ 1 simulati ons ( l},T = -3.3 "C ) a nd th e
geog ra phi cal stru c ture o r th e a nnll a l m ea n coo lin g is
a lso l'e rY similar. th e re a rc im po rt a nt diffe re nces in th e
seaso na l c)T les. In Fi g ure 3a, \l'e d e pi ct th e tempera ture
difle re nces be t\l'ee n EXP2 a nd EXPl fo r th e SUlllm e r
season, A \I'a rm er regio n a pp ea rs cl ea rl y in EXP2 a nd
ex tend s so uthward o r Fenn osca ncli a, from th e no rth easte rn .\leclit crra nea n coas t to ce ntra l Ru ssia , Thi s
\I'a rmin g is se\'e re ly d a mped in EXP3. as sl1 o \l'n in
Fig ure 3 b. \I'hi ch d em o nst ra tes th al thi s \I'a rmin g is \'C ry
3+5
Ralllsteill alld ] oussallllle:
Sensitil'i~)'
nperimentsJor the Lasl Glacial .IJarimum
IX>
eo
30
Jol
§
""j::
...-<
·30
.eo
.go
· 1B()
a
·120
-60
0
eo
120
eo
120
LONGITUDE
IX>
eo
30
Jol
c:.
P
""j::j
·30
.eo
-iIO
b
· 1B()
·120
-60
0
180
LONGITUDE
e\c n lti o n o r th e L a ure ntid e in EXP2 intensifi es th e
p la neta ry \\'a \'es a nd hea t tra nspo rt , leadin g to a n
in c rease in tempera ture a nd geo po tenti a l heig ht. As
m e nti o n ed in th e prev io us pa r ag ra ph , th e hi g he r
eleva ti o n of th e CLlr-lAP [SR indu ces a n enh a nce m en t
o f th e a nti cycl o nic circ ul a ti o n wes twa rd of' th e ice shee t.
Th e wind a t 85 0 hPa durin g winte r (no t sho \yn ) co nfirm s
th e e nhan ced d ee pe ning of th e A le uti a n low, leadin g to
g rea ter tra nspo rt of wa rm er a ir rrom th e N orth P acifi c to
Alas ka rol' CLIl\1AP ISR (EXP2), whi ch ex pl a ins th e
d evelo pm ent o r a wa rm e r regio n oye r Alas ka in winte r.
Ano th er \\'ay to illu stra te th e impac t of ice-shee t
ele\'a ti o n o n \\'inte r a tm os ph e ri c circ ul a ti o n fo r th e LG;\[
is th e splittin g o r th e j e t. As Kutzbac h a nd \ \'ri g h t ( 1985 )
no ted , this splitting is a fllll cti o n of ice-ca p size a nd
di stributi o n. In this se nsiti vit y ex perim ent , th e ice-ca p
co nfi g ura ti o n m odifies th e stru c ture of the sp[i ttin g , \V e
find , fo r th e simul a ti o n usin g C LIMAP ice -sh ee t
reco nstru ctio n (EXP2 ), th e sam e res ults d ese ri bed b y
C OH]\lAP (A nderso n a nd o th ers, 1988 ): a splittin g o rth e
j e t w ith a no rth ern bra nc h a t th e ed ge o rth e ice ca p a nd a
so uth ern bra nc h o n th e so uth e rn part o r l\o rth Am e ri ca
(Fi g. 4a ) , With P elti er reco nstru c ti o n ( EXP3), I\T find
th e sa m e intensity but a mo re zo na l circ ul a ti o n fo r th e
..,
00
'"~
C
·1B()
·120
-60
0
eo
j
120
·00
LONGITUDE
-60
Fig. 3. .-1ir temperature differences ( C) Jar J une, J u£l'
all d AllgllS/ (aJ be/ween E.\P2 and EXPt , (b) be/wee/{
E .\P3 alld E. tPJ . (c J be/wem EXP4 a/{ d E. \P J. Dark
grr:)' is llsed Jar temjJerature difJerences Izigl/er than 3"(;,
Light gr~JI Jar tem/malllre difJerences higher thall re.
IsoLilles are drawn Jar T = 2° and 4°C.
se nsiti\'C to th e Fenn osea ndi a n rec0l1S tru Clio n. Thi s broad
wa rmer regio n still ex ists in EXP4, as shown in Fig ure 3c,
whi ch pro\'es th a t it is no t indu ced by th e C LDIAP
SSTs. Th e wa rming o btain ed usin g th e C LI?dAP lSR
co rres po nd s to a n in crease in te mpera ture o r + 3 to SaC,
co ntradi c tin g th e data sy nth es is prov id cd b y Pe terso n a nd
o th ers ( 19 79) a nd m o re rece n tly by 'vV e bb a nd o th ers
( 1993 ) . Th ese d a ta a re in be tte r ag reem ent with th e
res ults obtain ed with th e Pelti er ( 1994) reconstru c ti o n . [n
th e wint e r seaso n , th e sa m e ph eno m e no n is o bsen 'ed OW l'
A las ka wes tward o r th e L a ure ntid e (no t sho wn ). In bo th
cases, th e differe nces in te mpe ra ture pa tterns d em o nstra te
th e impac t o r th e ice-shee t reco nstru c ti o ns on th e midla titud e a tm osph e ri c d yna mi cs. W e now illustra te thi s
rea ture with th e winter circul a ti o n.
Fi g ure 2b shows th e difference o r geo po te nti a l heig ht
a t 500 hP a, in w inter , be t wee n EXP2 a nd EXP 3.
Isoco n to urs indi ca te cha nges in th e m ea n Oo w. Th e
wa rmin g O\'Cf Al as ka cann o t be a direc t o rogra phica l
effec t: th e re is no diffe rence in ice-shee t ele\'a tio n O\'e r
Al as ka be tween th e two reco nstru c ti o ns. Th e hig her
·00
- 180
· 120
0
..,
120
100
..,
120
180
LONGITUDE
0
346
-<10
a
· 100
·120
.eo
b
0
LONGITUDE
._-
IX>
eo
30
Jol
§
..
t:
j
·00
~
-iIO
·180
C
·120
-eo
0
eo
120
180
LONGITUDE
Fig, -I . a. W ind speed a/ 200 It PaJor EXP2 ill December,
J al111al), and Febn(([I)" b, Tile same JOT EXP3. (.
DifJerences ill wind sjJeed between a and b,
Ramsleill alld J oussaume: Sellsililli!), e:rjJerimel1ls /or Ihe L asl Glacial ,Ha,lill/ uIII
no rth e rn branch oC th e j et (Fig, 4b ) , Beca use o[ the lo wer
ele\'a ti on of th e L a urentid e. the wes terli es are streng the n ed O\'e r th e ice cap . as sh own in Fi g ure 4c ,
co rrespond ing to th e wind-speed differ ences be tw ee n
EXP2 and EXP3, Th e impac t of the ice-shee t topography is, in ou r case, weaker than in Shinn an d Barron 's
(1989 ) ex trem e-se nsiti\'it y experim ent, [or whi ch the y
found a more defined structure of the splittin g using th eir
maximum reco nstruction.
4.3 Cha nge in SST for the LGM
Co mpa rin g EXP2 a nd EXP4 , lI'e ca n ana lyze the effect of
th e cha nge of SSTs on th e LG\I clim a te. The globa l
a nnu a ll>' a\'t'l'aged coo lin g reac hes on ly 1.1 cC in EXP4.
impl yin g that the ice shee t is responsibl e [o r 33% o[ th e
g lo ba l coo lin g durin g the LG;\l. Th e diITere nce of
geopotential heig ht at 500 hPa between E:\:P2 a nd EXP4
[or th e \I'inter seaso n. shown in Fi g ure 2c, demonstra tes the
ex tensi\'e th ermal for cing du e to th e SST \'ariations, A
st rong tro ug h d e\'e lops o\'Cr the co ld l'\ ort h Atl an tic ,
co\'e red b v sea ice , as a rcsult of' th e d ecrease of the surface
hea t Ou xes, H owe\'e r, the m ajo r difference in winter sea-ice
ex tent betwee n th e LGM a nd th e prese nt climate is not
responsibl e for the so uthwa rd shift of the Atlantic low
II'hi ch is \'ery simil a r in both simul a tions.
5. CONCLUSIONS
The ap proac h d e\'eloped 111 thi s p a per a Il O\<\·s a better
und erstand in g 0 (' the o rog rap hi ca l a nd thermal fOl-cings.
These se nsi ti\·i ty experi men to sho\\' the im pact oC ice-sh ee t
orog ra ph y and SSTs with empha sis o n th e eITect o[
diffcren t ice-shee t reco nstruction s on th e LGf-1 c lim a te.
\ \ ' e fi nd th at the tempera tures si m ul a ted fo r P el tier
reconstruction d o not show the impo rta nt wa rm er regio ns
for the LGi\I th at many o th er mod els found usin g
CLI1\IAP ( 198 1) recon stru c tion (Rind , 1987 ) , and
\I'hi ch co ntra di c ted to data . \ Vc a lso sho\\' th a t th c
d eepe nin g of th e Ale uti a n low is due to th e LG"' I SSTs
but that th e so uthw a rd shift of th e Atl a ntic low is due to
the circu la ti o n indu ced bl' the La urentid e. Fin a lh', we
illu strate th e diITerent co ntributions of SSI' a nd ISR o n
thc patt e rn of plan e tary wa\'cs o\'Cr th c :\f o rth e rn
H emisph ere durin g the LG1\1.
ACKNOWLEDGEMENTS
\ , .e g ra tefull y ac kn owled ge the La bora toire d e ~li: t eor ­
o logie D ), na miqu e (CNRS , Pari s, France ) [or pro\'iding
us with th e \'C rsion 4 of their GCM, Dr Pelti er [or
providing th e new ice-shee t reconstructi on , ~1. D o u tria ux
(L1\1D ) for participating in th e simul a ti o ns a nd J. Y.
Pete rsc hmitt (LMCE ) for the grap hi c o utputs.
REFERENCES
,\nd erso n. P,~1. alld 32 olhers , C lim atic ch a nges or th e la,t 13.000 ' Tars :
o bse rs'ations a nd model si mul at io ns, Srinlle. 24 1 (·fH69 , 10+3- 1052 .
Broccoli . A, ,J . an e! S, :'sl a n a bc. 1987, Th e innu encc or cont inclll al ice,
atmospheric CO 2 , a nd la nd albedo 0 11 th e c li ml1H' of t he last g la c ia l
maximum, Clilll(/Ie D)'II.. 1 (2 ), 87 99.
Cook , K , H, and l. id , H eld. 1988. Sta ti onars' IVal'e, or th e ice age
climate, J. Clilllale .. 1 (8 ), 807- 8 19 ,
Delllo n , G, H, and 1'.,J , H ug hes. el!, . 1981. Th e IIHI greal ifl' shedS. Ne,,'
Y ork. etc.. ,J o hn \I' i1 c, and Sons,
H eld. I :'sI. 1983, S ta tion ars' and qua si stati onar\' ed di es in the
ex tratrop ica l troposp here: th eory . [ 1/ I-I os kin s. D . .1 an d R. P.
P earcc . er!' . LIII;ge swle dl'lIallliwl jJrOCl'SJel ill Ihe allllo,phaf, London.
Academic Press, 127- 159,
J o ussaulllc, S. 1993, Paleocl im il ti c tracers: an illl 'es ti ga li on mi ng
atmospheric general circulati o n mode l und er ir(' a gc co nditi o ns
I. Dcsert dust. J. Geo/Jh)',. Re,l .. 98 (0 2 ), 2767 2805.
hu tLbach, J, E, and 1-1 , E, \I' ri g ht. 1985. Simula ti o n or th e climate of
18,OOO,'Cars BP: resu lts ro r t lw Nort h American / Nort h Atl an ti c l
Europl"(l n sector and co mpa ri so n ,,"ith th e gCQ iog"lc reco rd
~on h
Am er ica , Qpal, Sci. Ret .. 4 (3 ), 1+7- 187,
or
Le Tre ul. H, and K , L m ·a l. 198+. Th e impo rt ance of cloud rad iati o n
tnt e ractl o n [o r the simul ation of c li mate. [ 11 Berger , .\. and C .
Ni co li s. edJ . ,\ elt ' /}frJj)('rli vec, ill dim ate model/ing . . \nl s terdalll. 1~~ I "i ('\ ' i f" rl
199- 221.
Le 'Trc llt. H. and L•. Li. 1991. Scnsitj yit y of a n atlllosphcric gene ra l
c irc ulati o n mode l to prescrib ed SST chan ges: ICed bac k cfl('Ct
associatecJ wit h th e simulati o n of clo ud optica l prope rties. Clima/e
D 1'I1 .. 5 ,171 - 185.
:'slcl nLyre. A. ane! R, Cline, pds, 1981 . Seasollal recom/ruclioll of Ihl' Earth ',
'lIIj{/ce at Ih e la,1 glacial mIll illlUIII , Bou ld e r , CO, Geological Society or
,\ mer ica. \ :'s l ap and C h a rt Series 1\1 C -36, ).
Pe lti c r. R, \\', 1994. I ce a ge paleotopograp h y. Seima. 265 5169 ),
195 201,
Pct crso n , G. :'sI.. T. \I'ebb, I I I , J, E. E-;. utLbac h , T. , 'an d er I-Ia tllm e n,
'J'. :\, \I' ijm stra and F, A, Street. 1979. Th e con tin enta l record of
e n\'ironme nt a i concJitio ns at 18,000)'1' BP: an initial cI'<t1 uari o n , 01111,
Re", 12 ( I ) . + 7 82,
Rind, O . 1987, Components or the ice age c ircul ation. ]. (;eoph n . RI" .,
92 (0+, -12+1 +281,
Rind. D . a nd D . Petre l. 1985 . T errestrial con ditioll s at tll C' la s t g lac ia l
maximulll (lnci CLL\l AP sea surfa ce tem p erature cstinl alcs: arc they
co nsistent' 01(11, R eI " 24 11 i . I 22,
Sadoul'!l\', R, a nd K , L al·a l. 198-k Janu ary and J uly performance of the
L:'sID ge nera l circ ul at ion mode l. III Bcrger, .-\. and C. Nicoli s. er/."
. I 'elL' jJers/,et'lil'eJ ill ciilllllit' 1II(j(Mlillg. Amstc rd am. E lse,· i"r. 173 198.
S hinn, R. A, a ncl E . .1 . B a r ron , 1989, Climate se nsiti,'it y to co n tin ental
ice sh ee t size ancl config uratio n. J. . 1111 . .l/eleorol. SO( .. 2 12 . I jl7
1537.
\\'c bb , \1' " I II alld 7 IIlhO's, 1993. Climate c h a nge durin g th e past 18.000
vea rs: regio nal sy n theses. m ec hanism s, and ca uscs . [ 1/ \I' ri g h t. H.E .,
J r. ed. Global clillla/es sillce the last giaciaimarillllllll , ?dinn ea polis. Id~ ,
Uni'Trsitl or :'slinn esota Press. 51+-536.
\I'c bstcr. P . .1 , ancl:\f . ,\ . Strclen. 1978, La tc Qu a terna ry icc agc clim ates
oC lrop ica l .\uqralasia: intcrpreta tions and rccon ~ tru c tion s . QJlal.
Re> .. 10(3 1. 279-3 09.
347