Clobal warming could have a nasty surprise in store tor Europe

Transcription

Clobal warming could have a nasty surprise in store tor Europe
lce-cold
in
I
Clobal
warming
could
have
a nasty
surprise
instore
torEurope,
warns
ilefonRohmstof
.Instead
ofbringing
year_
round
warmth,
it may
heraTd
aneraoffreezing
winters
sTO NoRTHERNEuropeansshjverinsin
m e g L r p o t a n x n u s L l a t t tr r w i n r e f ,
g r . b d lw a i m i n gm i g h rs u d d e ! t ys e e ma n
arracnve proposirion. How pleasant |o
bask in a balny Medireffanein c]jmate
wrthort ever leaving home.
Drear on. E' iLlen.etror^ emersins re
v e a l " a r i s h r h a r s t o b a t* " n - g c . , l a
p r r n g e . m o s to i r u r o p e I n r o a b r g c h i l l
r : s n . g n u r a r e d so t \ e J r s ,b r g r n g u , r l h
rt ctrectsth:t cu!td be fclr
Bhr dr^dad
. Whar E!fopeans tend to forger is Ltar
|ne \rLdriL Ocernkeeps rhcnrrct.ti!etv
warni. B,\'nshtx, norfiem EL!.opesholtd
eiperience the same chilty ctjnate as
tne northem US and Canada,siDce
they are all at more or
less rhe same laritude.
But warm surface waiers
onginating in rhe tropics
are draan norrhwards by
ggantrcxnde$ea pumps".
which lull
_ the bainy
('arer towards
the Euro
pear conhnenr. Ir is this
vital pumpingprocessdrat
could be under rhrear
iiom Slobal wamins.
So. how does Europe's
cosyneahng sysiemwork?
The r\ra]m su ace waters
anive through rhe culf
Stream, on rhe wesren
side ofahuBe eddy knor,n
io. oceanographers as a
srDrroprcat gl,re. Simitaf
glres are tound in all
oceans and are driven by
the winds and rhe E.rrh\
rotadon- Normalll, this
gtre and |he cult Sir€am
would have litrle effect on
the climate of Eurotc. Bur the Norih
Atlnntic ls home to t\{o maDm.rtr
oceanicpunps-one east of creentard
and one in the southernLabrador sei_
s.hich exe an er1ratug onth€ warmed
srrr.ce water. Just as barhil.ater is
sucked down inro the Dtushole. rhe
lunips puli culf Srrean water from rhe
surfaced o\a,nrtlto rhe deepocean, drasglrg 1t rar enough north to hear Europe
lsee uiagram on pase 281.
The Gulf Srream, and the North
Adantrc pumps forn part of the global
ocean crcutahon system thar has been
dlbhedrhe Lo elerbetr byoc.anos
r a p h e r s .\ ^ r : r m s u i t a c e $ d r e r i ;
orawn nofth throughout the
Atlantic ata now rate more
than a huDdred rimes rhar
of the Amazon tuver. The,v
then sink ro the deeps
of
_t.Et)
the Creenland and
i .ador Seas,ard return to
tile Southe.n Ocean at 2
io 3 kilometres below thc
sudace as the so-ca ed
Nofh Arlandc DeepWaier.
'Ihe
warers .elease h€ar
rnto the cold nor$ern
atmosphere ar a rzre of a
tdllion kilowatts (10,: \\l,
an amount equFalent ro a
hundred times the worldk
energy consumptjon. This
energy warrls the a o1.er
Europe by abour 5 .C-a
fre€ hearing service thar
has operat€d reliably over
the past 10 000 yearsor so.
But rhere is jncr€asins
el.jdenceofabrupt and dr:matic cha.ges in Eurole,s
climate throuahout thelasr
rce ag€. Even rhe walm D€riod thar Drec e d e dr h e 1 . "a g . m a v h a i e h : d n r i l o r
c n m a u cl p s a n d d o w n s \ e r a g e t e r n
perarlres seemto have swuns by5 "C o.
n r o r eh r i h i n a t e r 1 e a r s ,l e a d r n er u r c v
-at
s p e l l s a t l a < r e df o r c e f r u i e s
ntoiogistsha\.e come !o view tie Dast10
000 years foilowin8 the end of the tast
rceageas an exceptionj. cLimarehistoll.
and sotre ar€ sal ing rhal human jnrcr
rerence r{ rflr $e climate s}stem mrghl
trtgger a rew pe.iod of instabiliw.
Mrch is alreadyLnown about the basjc
work]ngs oirh€ conveyorbett. Iis flo1r is
driven by differenccsin water densiiv ai
dilferenr points in the Atlanttc. rfilie
North Adantic DeepWarer k ro push irs
war solrhrvardsout of rhe Adanticbasin
rt needs to be densef than l,ater in rh€
sourh, near Sorth Africa, here itjoins
rl1e Antarcdc Cir.umpotar C!1renr,
\ r n r c l l j r c l e s i h € l l l n e r f t o m h . e s rr o
east. Some of this deep warer then rises
back to the surfacenearAnrarctica, atd
L
somet.arels inio tire other ocern basins
if n. nr 6.1.r I r: L i.r.r,rc rlr.
North Pa.ilic airer a thousaDdlears.
The latl DJ,i'hich ihls h'ater therl
.ctLLrrs ro rhc Atlanric is sLill hod,l
delared. There are i{., possibl€roLltes:
a wes rard " r!arm warer roure" passing
be veen the islands of indoresia and
aro.lnd Souft 11iic,r, and aq eashvard
''cold (ater rouie arolnd ihe southern
iit oi Souih Am€ica !ja D.ake Fassage
l s e . D l a g r a n ro . f a g r 3 0 1 .
l | . d e n s i r lo l r h e! . a r e r l n l h e l . { o l t
Arlantic is dcrcfirincd b,l lts salinitv and
Its temperature,so theselaclols also deiermire the acdnn ot dre con\'e,vorbelt.
uihen .he lfarm surface i{ater .ome!
in.o.ontac|diih lhe cold sub Arctic aif,
It cools. This irlcreaseslts denslN and
encoLrageslt to sinl. Oi the oiher haIld.
rr.r
r lrwrr,l
r F n I rfFnfr
fl
nu!
f
,-r,f-l
n F ( h , \r i . . . . . i r "
n..l
make-i il:e .xater less dense and hence
mofc buoyart. Lefi to 1tsel]. rhe licsh
r r a l e f l l o u l d w i n t h e c o n i e s t .B u t L l .
conre!or ielrlnngs a lerpetual flow oI
rew sajtl su ace witer iiom regjons to
ihe s.rLith, whilh ke€ps ihe seav,ar.f
,lFn.F
cn.'
r-ul, F.
'f
f|F
r.
. rL
..c-n.
rF
lri
!Lfh
.L f.
iri
:,1
l\rl, -llL
Lare more er.xDoratrorLhanireshr\ater
Detectiv€ work
'1
s seLf-nairraxing !os:ti!e leedb.ck
sysren has at leasr cne gladng hich: ii
\.npfl,
r,
lnf.rr,.f-
ir
rF,l
ff.
,.n
1,eyorbeli ginlis to a halt, it remains
s h r t d o w n [ s e e ' o t e r | h e e d g e ' ,p 3 0 1 .
Th; eifect a,as s€en in one of ihe fi t
.
im:f.
rr..J
li..
, !..rin.ni(
:.
in
clLldeborh oce.rnand amospher€. lrl dre
la.e l9E0s, Suki Manabednd Forr Srouli ' , i i h e c F , r i \ . . . l F l r r ' , n1 . n m i .
J . a b . f a r o f la l P f j f c e l o f i n N . J e r s e l
iound ihat iheir climai. nodel had ta.o
. Frr;
ffc,Frf
.r.
i..
tc!!
!rrfti
:.-fc-
rtr,l i qrf^n.f.
r'Fnr!,.r.
nFl
.o
flr.
i .n d a c o m c o l v e v o rb e l r i n ! h . A r L a ! r r a
t i r l r a l l eF ! r o l e a r c l i m a r c l n t h e o { h . f ,
lio'dcr.r drc co.!.t'or b€lt had shul
down, aird temqeraturesin northlvestefn
alrote !.rere u! to 10 'C colder than
today. The enisienceof rhcseFro slarcs
has since been conllrmed br marrr.
But ol\er quesiionsrenaine.l. Could
ihe cllnate $ritch betlvee! .hese differ
'rf
\.;
r,rf.!
u:ti-i
tsr.l
rf rl^nF
miotif
f.rocFr
r.
s r.f
I
flrF nrrrr
s.
r, Fl
l.uclil_.,,thc Eanh itself containssevenl
archircs of !ast climaiic condilions
lr'hich,with abli oiderc.tjr. rork.l,ield
nanl cltres..{moi1g the most uselid ot
rhcscafc thc sno,r lale6 thathave tiled
up on rh. Gfc.nland lce Sheet an.Lthe
la-versof sedimenr thar hale accumu
lated ai the botrom of the Arlanrlc.
T h r s e r e c o r d ss l r o a ' d r a t . a p i d a n d
s e r e r e c l n n a r i cj u m p s . c c u r r e d e v e r !
N Ew
\^/zte. works: the norrhward
part ofthe Atlanti. conveyor
belt p6ses t'nrough the
BenguelaCurrent, the Gllf
Str€m and the Norti
simply slifted sordr.
Three years ago, marine
seolosist Michel Samrhein
iiom the University of Kiel
n C€mrany, wirh coleaales
ftom France and the Netherlands, published reconstiuc,
tions of deep water flow in
the Atlantic at differenr
rimes in the past, based on
a large number of ocear
sedimenr cores. They follnd
ttuee cirfllation modes. The
fi$t was a warm conveyor
belt mode ihat has opemled
over ihe past 10 000 r€ars or
so. The second node was
a "glacial" conveyor bett,
which was sha ower and
did not o\rend norrh into the
creenland Sea, bur ended
soroewheresouth of Iceland.
Finally, they foud periods
r^rhere rhe convevor belr
was very weak becaus€
large imourts of meltwater
had entered the Adantic,
cappins otr oceanic convection with a sudace ,,lens,'
Greenland dd tab.addr s6.
North Atlantic Dcep Warer
(NADW) rrows sourh at a
depth of ? to 3 kilometres
thousand years or so dudng
the last ic€ age, in sharp
contrart to the stable conditions of the past 10 000
years. The last of rhese
jrmps is theYo nger Dryas
event which took place as
the Earrh emerged from
lhe last ice age. cradual
cLimaticwaming was caus
ing the huge continenrat ice
sheets to meli and disinte
8rate, but rhen, witlin a
decade, ice-age condirions
In 1989, a nodelling
expenment by Emst MaierReimer and Uwe Mikola,
jewicz al the Max Planck
institute in Hamb1lfg uncoverecl a near explanation
for the Younger Dryas
event. It showed that a massiveinflow of
meltwater from the Laurenride Ice Sheer
couid have led to a sudden collapseof
fie Adznric conveyorbell, firowinA rhe
Atlantic re$on b.ckin|o ttr€ fteez&.
Flickering switch
Now, researchers are askinq whether
todals global walmhs, L\t resutt of
accumdated carbon dioxide and other
Sreennouse gases in the armosDhere.
could have dle same etrect as the;e.iod
of natural warming at the end of rhe ice
age. Global warming is, for instance,
expectecl to warm the surhce water in
the northem high latirudes. rt shorld also
incleare the amornt of rain and snowfall
over rhe ocean, and speed uD rhe metting of hish latitude iie, wtiich would
make the water fresher. Borh the warrlt
ing and the fieshening wolrtd ma1{en\e
sutace water less dense, which cortd pur
Drakes on the pumpins mechanism.
in 1993, Manabeand Storffer studied
the effects of CO, concenrations on
global dimare in a nodel that coupted
the ocean, the armosphereand sea ice.
As tne at]Eospheric COz concentration
slowly increased ro lou rimes Lts pre
industrial level, rhe oceans deep c;cu_
ladon came to a complete standstill,
28
S cIENTIsT
However, this change required fairty
drastic amounrs of CO, to be released
nto rhe atmosphere,tie
Inrerqovern_
mental Panel on Climate chansadoes
not expect srch levels ro be
reached before 2100. A1so,
lhe deep circulation il rhe
model ground to a halt
slowly-over
centuriesunlike the abrupt climate
shifts shown by rhe crcenland ice core record.
Despite these caveatr.
tlere is nrounting evi
dence. both ftom rhe
past dimate record and
fuom more r€cent ocean
modeling, that the dnnare
system could be Eore
\dnemble than Manabe
and Stouffels fhdings
imply. Over the past few
years, as researchers have
driled and analysed nore
ocean sedimentcores,lhe
picture has been gettirg
more complex. The new
evideDce shows that during some cold spells, the
conveyor belt may rlot
have switched off brii
At that time, I was a
researche. ar the Institrte
of Marine Sciences at t}le
University of Kiel nnning a se.ies of
modelling experiments investigating
how the conveyor bek would respond if
a lot of freshwater suddeniy flowed
into die Adantic, and how fiis
would affect surface temperatures. SDrpnstl)giy,as
wel as &e then famili,r.ll
matic modes of operarion
wlrh the conveyor belt
switched either "on,, or
"otr", we too found a third
possibility of a cold conveyor belt extending not
flearly as far north as
at present.Alrhough this
conveyorbelt was almostas
vigorous,ir hardly wamed
the northem Atlantic
region, as its waterssank
and retumedsourhbefore
releasingmuchheat tu the
atmosphere.So a shifr in
the ocean cunents could
havedrown the rcgion into
a cold spell wirlout rhe
completecoilapseof rhe
conveyor
beh.Alsoin 1994.
erperiments by Andrew
Weaverand T€rtia Hushes
from the Unive$ity of Vic-
\r!r'Sainr-r1sT
tora tn Canada showed thal rv;lh in
creasedp.€cipitation in the Adartic the
conveyof b€li cin slarr to ,,flicker,,be
n{een iis dilferenr modes, leadins to
sirong cliniatjc oscilations over Eur;De.
The possibilirr that an iriux of ireih
wate-rinto ihe Atlantic cordd hal€ such
an ertect$ 1{orrying. Alrhough the Lind
lrocess ar one of rhese sites,
rhe tncomiojt fieshr{ater
would staft ro accumulateat
lhe surtace.This woutd make
the surface wat€.s nore ind
mofe buoyani and it woutd
lecom€ harder and harder to
restart the pum!. The modets
suggestthat in this r€y the
pumprnS at one of the sites
could shut dowr and the con
veyor belt coi d rhe! change
rts route wirhjn a lew years.
rhis bfeakdown couid be
triggered by a retativelv
small change in rhe amolnt
or resrlwaier ent€.ing the
ocean, Decausettre tlvo pump
j sites are very locatisert,eaci
i beinSjxsr a fela/hundred kno
i metl$. acrcss. Nor surpris" rngl\ rheseco.\.criveDunlDs
; ha\e beendubbedrhei.h,t[s
heel of the conveler beh.
Bur we do not kn.rw
wnerhe. &is wil rea|v hap
p€n. ExlstinSmodets're simply incapable ol quarrifung
now muchwamingis needcd
to swirch off convectio. ar
on€ of rhe pump sites.Though
someor rhe tas|estcomputers
'n the world are usedforrhesc
srnmladons,lack ofcomp ier
power is still forcing n;dets
ro use a very coarsemesh jn
thei. calculations.This means
mat they cannorrePresentthe
lrird of rcgional delails that
could turn our ro be cnciat.
The models are also Lnre
alistic in that thev oversrare
the role of saft and hear diffil
$on. Reducingthe amount of
ro make th.
o c e a n . . c r r r e n t sl e s s s i a b t e ,
accordng to comlarisonsthat
StouUerand I havenade over
the past six moorhs. Whart
more, mosr modeis siiu work
g with ad hoc tues, known
' as flux adjusrments,
at rhe
inrcr1aceDerweenocean and
atmosphere_ Uirhout these
iixes, lhe simutarcd ctimate
and ocean cifculation drifts
oi suddenclimarjc swing sho,,1/n
in these ro a less realisricstate_
modclsworlc through positivefeedback,
So we can rake no confori hom rhe
1r rs pnmarily associatednot wirh s.k
currenrglobaiwarmng scenarios,wbich
transport in rhe conveyorbeh, but with
tend to slow a smooth giaduat warirLing
the do,/vnwardmi\ing of wate. in the over rne
next century. we simPty dont
tuo punps. tf sornetting perhaps rhe moa, lvny
our present ctimate is much
e r r e . i o r d r o b a l w d r m i n g i n t e r m p r e d more stable
than the ciimate of the Dast.
the do!\nward mixing, of conv€ctlon, and lyherher thls
stabitity wil continue
in |he face of globai
srmuiarions-rhe Lhird
\4'amjng_ Though the
line olanack. More mea
moders suSgesi the
surementswill also help
etreci of global wa
ns
ro decide whether con
might be lessdrasticana
vection in rhe creentand
,}
npid thar pastchanges,
Sea.ha.salready weaki! could sniply be that
eneo stnce rhe early
presenrmodek donl yer
1940s due to gtobal
capture the physics of
warmrngjassomeoceanabmpr climate chairge.
oSraphers have sugResearchers af€ at
gested (see Thjs Week.
tacking these gaps it
19 March 1994, p 4), or
oxr l{Ilowledge on tlfee
1lwetlef$anatur2l
tronis- First, rhey are
climaricfllctuarion.
lookrng to see if past
so long as the jury
swngs n ocean circLrla
stals out o. just how rul,
non rook ptace onty
FlowihA hot and cotd:th€ gtobal conveyor bett is
.erable dle conveyor bett
dflv€n by dense, satt,
during the last ice sge,
surla.e water sinkina in rh€ North aflantic
h, there js still a ve4/ real
or wnetner the ocean
possibiliry
thar .rre will
was also unstable in the
unwittjngly
dismpt it
Eemian lnrerglacial Stagej a wafm inter_ the Eemianclimate.
The former showed aq Erggera catamito!\ (ooting rhroLrsn
lude between113 000 and 125 000 r€ars a record offrequen!
fluctuation but was o!'ruope
Lheconsequences
foi ccosls
ago. lf the first of rhese is tue , we ioutd
pmbably disturbed b,\'morioi! of rhe jce,
rems,
agnculture
and sociery could be
be safe. It implies dlat s\\dngsin the con
whlle the latter.evealed a Deriod ofstasevere.And jik not jusr a EuropeanDrobveyor belt aroseoily becauseof meltwa
bility Drilling has iust sraied on a new ]em. The effectsofpasrcoolins
epis;d€s,
ter or rcebergnodbc. ftom rh€ t\armlllc
ice core, known as Nodh czuq which
such as the Younger Dryas, have bee!
of tlc lar ge amo! ntSof l: nd rce rhat ha;
snourd heip io serrte the issue (This seenn the clrrate record
from the US.
fomed in the precedins slacial Dedod.
Week,6JLy1996,p7).
chil€ and evenNew Zealand.Geoct€his;
On rhc orher land, ctmadL <wrnssin
Researchis ako openirg up on ihe
Waily Broedrer of colunbia Unive$i$ in
tbe EemiarsuggcsL
r l a t t a r s e a m o ! n r s oceanographic fionr with expeditions
by
N e w Y o r kl s a b t u n r \ a ! v o f p u i L i n ; r :
or rano rce ale notnecessarytocausethe
buropean,Amencanand Canadianreams ' ' w e a re p l a | n g R l s s j a n
L o u l e t tre i & . l i
clrangcs.ThJr couldpresagea rolgh ride s r a r r i n S
L h j cx n r e f . T h e o c c a r o g r r p h e r s mate an.l no one knows what
ties in the
r o r u s I n L h et u t u r e ,w i r e r g t o b a t w a m
wr Destud)ng L[e Lon\ectionDrocesses chamberof ihe g!n.,.
tr
mg nas set n. Tvvorecent Greenlandice ard rheirlint ro dim:'ic
co.din;6 rn rhe
cores,the EuropeancRIp core talen jn
nonhen Atlantic. These m€asuremenrs
stelan Rahmstorf ,s d. oceon modelef dt de
1992 and tlte AnrericancISp2 cor€taken will in rxm provide
imporrant data Pats.lah tnstnute Cljmate tnpad Reseorch
fot
'n 1993. provided conflicting views of
in
neecledto validate and improve model