w w w. n at ur e. c o m/ s ci e ntifi cr e p ort s
Atl a nti c S S Ts c o ntr ol r e gi m e s hifts
i n f or est fir e a cti vit y of N ort h er n
S c a n di n a vi a
I g or Dr o b ys h e v
1 ,2, Y v es B er g er o n
2, A n n e d e V er n al
3, A n d ers M o b er g
4, A d a m A. Ali
5&
M ats Ni kl ass o n
1U n d erst a n di n g t h e dri v ers of t h e b or e al f or est fir e a cti vit y is c h all e n gi n g d u e t o t h e c o m pl e xit y of t h e i nt er a cti o ns dri vi n g fir e r e gi m es. W e a n al y z e d dri v ers of f or est fir e a cti vit y i n N ort h er n S c a n di n a vi a ( a b o v e 6 0 N) b y c o m bi ni n g m o d er n a n d pr o x y d at a o v er t h e H ol o c e n e. T h e r es ults s u g g est t h at t h e c ol d cli m at e i n n ort h er n S c a n di n a vi a w as g e n er all y c h ar a ct eri z e d b y dr y c o n diti o ns f a v o ur a bl e t o p eri o ds of r e gi o n all y i n cr e as e d fir e a cti vit y. W e pr o p os e t h at t h e c ol d c o n diti o ns o v er t h e n ort h er n N ort h Atl a nti c, ass o ci at e d wit h l o w S S Ts, e x p a nsi o n of s e a i c e c o v er, a n d t h e s o ut h w ar d s hift i n t h e p ositi o n of t h e s u b p ol ar g yr e, r e dir e ct s o ut h w ar d t h e pr e ci pit ati o n o v er S c a n di n a vi a, ass o ci at e d wit h t h e w est erli es. T his d y n a mi cs str e n gt h e ns hi g h pr ess ur e s yst e ms o v er S c a n di n a vi a a n d r es ults i n i n cr e as e d r e gi o n al fir e a cti vit y. O ur st u d y r e v e als a pr e vi o usl y u n d o c u m e nt e d t el e c o n n e cti o n b et w e e n l ar g e s c al e cli m at e a n d o c e a n d y n a mi cs o v er t h e N ort h Atl a nti c a n d r e gi o n al b or e al f or est fir e a cti vit y i n N ort h er n S c a n di n a vi a. C o nsist e n c y of t h e p att er n o bs er v e d a n n u all y t hr o u g h mill e n ni u m s c al es s u g g ests t h at a str o n g li n k b et w e e n Atl a nti c S S T a n d fir e a cti vit y o n m ulti pl e t e m p or al s c al es o v er t h e e ntir e H ol o c e n e is r el e v a nt f or u n d erst a n di n g f ut ur e fir e a cti vit y a cr oss t h e E ur o p e a n b or e al z o n e.
Fire is a pri m ar y dri vi n g f act or of t he ec os yste m d y n a mics i n t he b ore al f orest, directl y a ffecti n g gl o b al c ar b o n
b al a nce a n d at m os p heric c o nce ntrati o ns of trace gases i ncl u di n g c ar b o n di o xi de1 , 2. L ar ge are as of t he b ore al f orest
t h at b ur n a n n u all y i n N ort h A meric a ( pre d o mi n a ntl y i n C a n a d a) a n d E ur asi a ( pre d o mi n a ntl y i n R ussi a) e xe
m-plif y a f u n d a me nt al li n k b et we e n l ar ge s c ale at m os p heric circ ul ati o n pr o cess es a n d fire re gi mes3 , 4. Fires i n t he
b ore al re gi o n are esti mate d t o re prese nt ar o u n d 9. 1 % of t he a n n u al t ot al bi o mass b ur ne d gl o b all y5. St u dies i n t he
E ur asi a n b ore al f orest s h o w t hat fire rele ases 4. 8 t o 1 5. 4 T g C/ ha de pe n di n g u p o n size a n d se verit y 6. Cli m atic all y
a n d h u ma ni n d uce d c ha n ges i n fire re gi mes i m p act t he f u ncti o ni n g of t he b ore al ec os yste m b y a ffecti n g re ge ner
-ati o n a n d gr o wt h c o n diti o ns f or d o mi na nt tree s pecies, f orest c o m p ositi o n, a n d its s uccessi o nal p at h w a ys7.
U n der nat ural c o n diti o ns, cli mate is t he maj or fact or c o ntr olli n g fire acti vit y as l ar ge-sc ale circ ul ati o n p atter ns deter mi ne p eri o ds of fire-c o n d uci ve we at her a n d s y nc hr o nize fire re gi mes acr oss l a n dsc a p es a n d re gi o ns of t he
b ore al z o ne8 , 9. H u m a n acti vities h a ve als o i m p acte d f orest fire re gi mes1 0 t hr o u g h t he us e of fire f or a gric ult ur al
p ur p oses s uc h as sl as h-a n d- b ur n a gric ult ure a n d, m ore rece ntl y, t hr o u g h fire s u p pressi o n. M o der n h u ma n i m p act res ults i n a decre ase i n nat ur al fire acti vit y, e ve n t h o u g h i ncre asi n g h u ma n de nsities are li n ke d t o a hi g her de nsit y of i g niti o ns. Alt h o u g h i n S c a n di na vi a f orest fires bee n hist oric all y o ne of t he pri nci p al f act ors s ha pi n g t he f orests’
st ate a n d f u ncti o n1 1, 1 2, f orest fire acti vit y i n N ort her n E ur o pe has bee n decre asi n g si nce t he l ate 1 9t h ce nt ur y1 3.
The E ur o p e a n cli m ate is l ar gel y deter mi ne d b y he at a n d h u mi dit y tr a nsfer fr o m t he a dj ace nt N ort h Atl a ntic
O c e a n, n ot a bl y t hr o u g h t he w ar m a n d s ali ne N ort h Atl a ntic C urre nt ( N A C) fl o wi n g n ort he ast w ar d 1 4 – 1 6. The
i n fl ue nce of t he str o n g N A C res ults i n rel ati vel y hi g h wi nter te m p er at ures a n d l o wer se as o n al te m p er at ure c o n-trasts as c o m p are d t o t he re gi o ns at si mil ar l atit u des i n N ort h A meric a. L ar ge sc ale circ ul ati o n a n o malies a ffecti n g t he e ner g y b al a nc e of t he n ort her n N ort h Atl a ntic O c e a n a n d ass o ci ate d at m os p heric press ure p atter ns h a ve
1S o ut h er n S w e dis h F or est R es e ar c h C e ntr e, S w e dis h U ni v ersit y of A gri c ult ur al S ci e n c es, P. O. B o x 4 9, 2 3 0 5 3
Al n ar p S w e d e n. 2C h air e i n d ustri ell e C R S N G- U Q A T- U Q A M e n a m é n a g e m e nt f or esti er d ur a bl e, U ni v ersit é d u
Q u é b e c e n A biti bi- T é mis c a mi n g u e ( U Q A T), 4 4 5 b o ul. d e l ′ U ni v ersit é, R o u y n- N or a n d a, Q u é b e c, J 9 × 5 E 4, C a n a d a.
3G E O T O P, U ni v ersit é d u Q u é b ec à M o ntr é al, C P 8 8 8 8, M o ntr é al, Q u é b ec, H 3 C 3 P 8, C a n a d a. 4D e p art m e nt of P h ysic al
G e o gr a p h y, St o c k h ol m U ni v ersit y, S v a nt e Arr h e ni us v ä g 8, 1 0 6 9 1 St o c k h ol m, S w e d e n. 5C e ntr e f or Bi o- Ar c h e ol o g y
a n d E c ol o g y ( U M R 5 0 5 9 C N R S), U ni v ersit é M o nt p elli er 2, 1 6 3 r u e A u g ust e Br o uss o n et, F- 3 4 0 9 0 M o nt p elli er, Fr a n c e. C orr es p o n d e n c e a n d r e q u ests f or m at eri als s h o ul d b e a d dr ess e d t o I. D. ( e m ail: i g or. dr o b ys h e v @sl u.s e or i g or. dr o b ys h e v @ u q at. c a)
R e c ei v e d: 2 5 A u g ust 2 0 1 5 A cc e pt e d: 1 7 F e br u ar y 2 0 1 6 P u blis h e d: 0 4 M ar c h 2 0 1 6
b e e n re p e ate dl y s h o w n t o a ffe ct E ur o p e a n we at her1 7 – 1 9. Hist oric all y, h o we ver, m ost of t he a n al ys es h a ve b e e n
f oc use d o n t he wi nter cli m ate, si nce it is t he peri o d w he n oce a n- we at her fee d b ac k a p pe are d t o be t he m ost pr
o-n o u o-nce d2 0. It is re as o n a ble t o h y p ot hesize t h at t he pr ocess a ffecti n g t he h u mi dit y a n d e ner g y tr a ns p ort o ver t he
N ort h Atl a ntic Oce a n s h o ul d b e e q u all y i m p ort a nt f or t he s u m mer cli m ate a n d, t heref ore, re gi o n al fire acti vit y. I n dee d, t he s u m mer te m p er at ures a n d preci pit ati o n d y n a mics o ver S c a n di n a vi a h a ve b ee n s h o w n t o b e l ar gel y
c o ntr olle d b y t he p ositi o n of t he westerl y tr ac ks a n d t he stre n gt h of N A C2 1. St u dies i n di ffere nt p arts of t he w orl d
ha ve s h o w n str o n g li n ka ges bet wee n re gi o nal fire acti vities a n d c ha n ges i n se a s urface te m perat ures ( S S T) a n d t he
ass oci ate d press ure p atter ns c o ntr olli n g preci pit ati o n o ver a dj ace nt l a n ds2 2 – 2 7.
I n t his p a p er, we h y p ot h esiz e d t h at o c e a n c o n diti o ns i n t h e n ort h er n N ort h Atl a nti c pl a y a d eter mi n a nt
r ole o n re gi o n al fire acti vit y i n n ort her n S c a n di n a vi a2 8. We pr o p os e t h at a we a ke ni n g of t he N A C a n d str o n ger
re gi o nal wi nter c o oli n g, re flecte d b y c ol der S S T i n t he N ort h Atl a ntic, w o ul d res ult i n i ncre ase d fire acti vit y i n t he f oll o wi n g fire se as o n. We teste d t his hy p ot hesis usi n g m o der n f orestr y d at a o n a n n u all y b ur ne d are as i n n ort her n S c a n di na vi a ( n ort h of 6 0° N), m ultice nt ur y a n d s p ati all ye x plicit de n dr oc hr o n ol o gic al rec o nstr ucti o ns of a n n u -all y b ur ne d are as, rec o nstr ucti o ns of s urf ace air te m p er at ure ( S A T) a n d fire acti vit y, a n d rec o nstr ucti o ns of s e a s urf ace c o n diti o ns ( S S T a n d se a ice c o ver) o ver t he n ort he aster n N ort h Atl a ntic d uri n g t he H ol oce ne.
T h e st u d y r e gi o n.
fie m ai n st u d y re gi o n c o vere d t he are a of n ort her n S we de n b et we e n 6 0° N a n d 7 0° N. It lies i n t he b ore al cli m atic z o ne a n d recei ves air m ass es fr o m t he Atl a ntic re gi o n, es p eci all y d uri n g t he wi nter se as o n, w hic h m a kes it p artic ul arl y se nsiti ve t o t he stre n gt h a n d t her m al pr o p erties of N A C. He nce, des pite t he hi g h l atit u d es, t he me a n Ja n u ar y te m p er at ure r a n ges b et we e n − 4 t o − 1 6 °C a n d t he me a n J ul y te m p er at urer a n ges fr o m 1 2 t o 1 6 ° C. fie n u m ber of d a ys wit h a me a n te m per at ure a b o ve 5 ° C r an ges bet wee n 1 0 0 a n d 1 6 02 9.
fie t ot al a n n u al preci pit ati o n is 6 0 0 – 7 0 0 m m o n a vera ge, b ut wit h ver y hi g h vari a bilit y at hi g her ele vati o ns w here it r a n ges b et we e n 4 0 0 a n d 1 4 0 0 m m. S n o w acc o u nts f or 3 0 % t o 5 0 % of t he a n n u al preci pit ati o n a n d c o vers t he
gr o u n d f or a n a vera ge of 1 7 0 – 2 2 5 d a ys/ ye ar2 9. fie maj orit y of m o der n f orest fires i n n ort her n S we de n occ ur e arl y
i n t he fire s e as o n ( Ma y t hr o u g h J u ne), alt h o u g h t he t ot al b ur ne d are a is d o mi n ate d b y A u g ust fires1 3. N ort her n
S we de n e ncl oses t he al pi ne z o ne, n ort h, mi d- a n d s o ut h b ore al f orests3 0. S c ots pi ne (Pi n us sylvestris L.), N or w a y
s pr uc e (Pice a a bies ( L.) H. K arst), a n d birc h (Bet ul a p u besce ns E hr h. a n d B. pe n d ul a R ot h) are t he d o mi n a nt s pecies i n t he f orest ve get ati o n c o ver. Fire has bee n t he pri mar y dist ur b a nce f act or i n t his b ore al ec os yste m wit h
st a n d-le vel fire ret ur n i nter v al i n mi xe d pi ne f orests r a n gi n g bet wee n 3 0 t o 1 0 0 ye ars1 2. fie m o der n (sec o n d h alf
of t he 2 0t h ce nt ur y) fire c ycle i n t he n ort her n p art of S we de n is 1 03– 1 04 ye ars1 3, w hic h is attri b ute d t o e fficie nt fire
s u p pressi o n p olicies.
M et h o ds
R ati o n al e a n d d at as et s el e cti o n.
fie st arti n g p oi nt i n t he a n al ys es w as a p atter n s h o w n i n a pre vi o us st u d y, w hic h de m o nstr ate d a rel ati o ns hi p bet wee n maj or fire ye ars i n n ort her n S c a n di na vi a a n d re gi o n al p ositi vete m per at ure a n d ne g ati ve preci pit ati o n a n o malies2 8. We hy p ot hesize d t hat t he fire p atter n arises fr o m l ar ge sc ale
a n o malies i n at m os p heric circ ul ati o n, e xte n di n g f ar o utsi de t he st u d y re gi o n a n d be y o n d t he fire se as o n. fie f act t hat t he o bser ve d preci pit ati o n a n d te m perat ure a n o malies e xte n de d o ver m ost of t he s u m mer s u g geste d t hat t he y were li kel y rel ate d t o c ha n ges i n e ner g y st ates of t he at m os p here a n d oce a n s urf ace. Acc or di n gl y, we e xte n de d t he ge o gr a p hic al fr a me of o ur a nal yses o ver t he N ort h Atl a ntic a n d i ncl u de d t he wi nter a n d s pri n g m o nt hs prece di n g t he st art of t he fire se as o n. We teste d t he li n k bet wee n t he s urf ace oce a n te m per at ures a n d fire acti vit y at a n n u al, ce nte n ni al a n d mille n ni al s c ales. D et aile d a n al ys es of s p ati al rel ati o ns hi ps were m a de usi n g m o der n d at a fr o m
t he 2 0t h ce nt ur y. We als o e xte n de d t he ti me i nter v al f or t he a n al ysis t o t he l ast 8 0 0 0 ye ars t o a n al ys e l o n g-ter m
c ha n ges b ut still a v oi di n g cli mate pert ur b ati o ns rel ate d t o t he gl aci al-i nter gl aci al tr a nsiti o n. Ei g ht t h o us a n d ye ars a g o als o r o u g hl y c oi nci de d wit h t he st art of p ale o fire rec or ds i n t he st u d y are a ( S u p ple me nt ar y I nf or mati o n Ta ble SI 1).
I n o ur l o n g-ter m a n al ys es, we l ar gel y c a pit alize d o n tre e-ri n g a n d c hir o n o mi d- b as e d s u m mer te m p er at ure rec o nstr ucti o ns e ve n t h o u g h we re alize d t h at ari dit y rec or ds (e. g. preci pit ati o n) m a y pr o vi de e q u all y i m p ort a nt i nf or mati o n a b o ut c ha n ges i n hist oric al b ur n r ates. Reli a ble a n d s u fficie ntl y l o n g p ale o- ari dit y rec or ds are, h o w-e vw-er, n ot yw-et a v ail a blw-e f or o ur st u d y rw-e gi o n. S w-e as o n all y-rw-es ol vw-e d prw-e ci pit ati o n rw-e c or ds fr o m trw-e w-e ri n gs arw-e o nl y
a v ail a ble f or t he p ast fe w ce nt uries3 1, 3 2 a n d d o n ot c o nt ai n ce nte n ni al si g n al3 3. S o me preci pit ati o n rec or ds fr o m
l a ke se di me nt d at a are res ol ve d wit h dec a d al res ol uti o n, b ut c o ver o nl y a p art of t he H ol oce ne3 4, 3 5. B y c o ntr ast, air
te m p er at ure rec or ds fr o m n ort her n S we de n are a v ail a ble wit h se as o n al (e. g.3 6) t o dec a d al-ce nte n ni al res ol uti o n
f or t he e ntire H ol oce ne3 7.
Fir e r e c or ds wit h a n n u al r es ol uti o n.
F or t he a n al ys es of fire at a n n u al s c ale, we us e d m o der n f orestr y st atistics a n d de n dr oc hr o n ol o gic al fire rec o nstr ucti o ns. F orestr y st atistics pr o vi de d d at a o n t he a n n u all y b ur ne df oreste d are as i n n ort her n S we de n fr o m 1 9 4 2 t o 1 9 7 5 a n d fr o m 1 9 9 6 t o 2 0 1 41 3. fie c o m p osite fire rec or d di d n ot
e x hi bit te m p or al a ut oc orrel ati o n (|r| < 0. 1 0 f or ti me l a gs u p t o 1 5 ye ars).
De n dr oc hr o n ol o gic al fire rec o nstr ucti o ns, w hic h ai me d at e xte n di n g a n n u all yres ol ve d fire rec or ds o ver ce n -t uries, were b ase d o n assi g ni n g -t he e xac-t c ale n d ar ye ar (cr oss d a-ti n g) -t o fire sc ars i n -t he w o o d of Sc o-ts pi nes (Pi n us sylvestris L.). We use d t w o de n dr oc hr o n ol o gic al d at asets. First, we de vel o pe d s p ati all y e x plicit rec o nstr ucti o ns of fire acti vit y ( a m o u nt of are a b ur ne d a n n u all y) at t w o sites i n n ort her n S we de n, o ne si nce 1 3 2 0 A D (site Bj ur h ol m, are a 6 0 8 0 0 ha, Fi g. 1, Ta ble SI 1) a n d t he ot her fr o m 1 3 7 1 A D (site Ti ve de n, are a 1 2 0 0 ha). F or b ot h are as, we use d t he net w or k of sites wit h i n de pe n de ntl y rec o nstr ucte d fire hist ories fr o m tree fire sc ars. We c o nsi dere d pr o perties of m aj or fire bre a ks (e. g. l a kes a n d ri vers) a n d distri b uti o n of b ur ne d a n d u n b ur ne d ( a n d rec or di n g) sites acr oss
t he l a n dsc a p es t o deter mi ne t he size of e ac h fire (f or t he det aile d met h o d ol o g y, s ee3 8). S ec o n d, we us e d a pre
vi-o usl y p u blis he d list vi-of ye ars w hic h were s h vi-o w n t vi-o e x hi bit hi g h fire acti vit y at t he re gi vi-o n al s c ale (s vi-o-c alle d l arge fire ye ars , L F Ys), i de nti fie d b y c o nti n ge nc y a n al ys es o n a n ort her n S we dis h net w or k of fire hist or y sites ( Fi g. 1;
w w w. n at ur e. c o m/ s ci e ntifi cr e p ort s/
Ta ble SI 1 3 9). L F Ys acc o u nt f or t he m aj orit y of t he f orest b ur ne d are a at dec a d al-ce nte n ni al ti me s c ales4 0. L F Y
were i de nti fie d usi n g b ot h t he perce nt a ge of sites b ur ne d i n a ye ar a n d t he t he oretic al pr o b a bilities of o bser vi n g a p artic ul ar n u m ber of are as b ur ne d i n a ye ar. We t he n assesse d t he t he oretic al pr o b a bilities t h o u g h c o nti n ge nc y a nal ysis, c o nsi deri n g fire acti vit y as a re aliz ati o n of a bi n o mi nal pr ocess:
= − − p X N X N X p q ( ) ! ! ( ) x N X, ( 1)
w here N w as t he t ot al n u m b er of rec or di n g are as i n t he a n al ysis of a s p eci fic p eri o d; X – t he n u m b er of b ur ne d are as i n a si n gle ye ar; p – t he pr o b a bilit y of a n are a b ur ni n g i n a ny ye ar, a n d q – t he i n verse of t his pr o b a bilit y. fie di ffere nces bet wee n e x pecte d a n d o bser ve d fre q ue ncies were esti m ate d wit h t he C hi-s q u are test ( S o k al a n d R olf
1 9 9 5). We refer t o2 8 f or t he det ails of t his a nal ysis.
Fir e r e c or ds at a b o v e- a n n u al r es ol uti o n.
As a pr o x y of f orest fire acti vit y i n t he p ast, we us e d c h ar-c o al -c o n-ce ntr ati o n d at a fr o m t hree 1 4C d ate d l a ke s e di me nt c ores fr o m N ort her n S we de n ( Fi g. 1, Ta ble SI 1 4 1).
Fi g ure 1. M a p of t h e m ai n c urre nt p at h w a ys i n t h e n ort h er n N ort h Atl a ntic a n d l oc ati o n of sites pr o vi di n g pr o x y d at a. L C L a br a d or c urre nt, W G C Wester n Gree nl a n d C urre nt, E G C E aster n Gree nl a n d C urre nt, I C Ir mi n ger C urre nt, EI C E aster n Icel a n dic C urre nt. fie gre y p ol y g o n of t he u p per p a nel refers t o t he m a p are a of t he l o wer p a nel.
fie c ores were c ali br ate d o n a ge- de pt h m o dels a n d were tr a nsf or me d t o t he s a me a n d c o nst a nt ti me res ol uti o n
( 2 0 ye ars4 1). We a ver a ge d t he d at a fr o m t hes e sites t hr o u g h a st a n d ar d t w o-ste p s m o ot hi n g pr o ce d ure usi n g R
p ac k a ge p ale o fire4 2 a n d c o n verte d t he d at a i nt o c harc o al i n fl u x bi n ne d i nt o 2 0 0 ye ar i nter v als. Pri or t o a ver a gi n g,
t he site c hr o n ol o gies were st a n d ar dize d t o e ns ure t h at e q u al wei g hts were attri b ute d t o e ac h site i n t he m aster c hr o n ol o g y.
S urf a c e air t e m p er at ur e ( S A T) p al e o r e c o n str u cti o n s.
As a s o urc e of a n n u all y res ol ve d s u m m er ( Ma y- A u g ust) te m per at ure rec or ds o ver t he l ast mille n ni u m, we use d a rec o nstr ucti o n fr o m t he Tor netr äs k are ai n n ort her n S we de n ( Fi g. 1), w hic h i ncl u de d tree ri n g- wi dt h a n d m a xi m u m de nsit y me as ure me nts3 6. To e xte n d
b ac k t he te m per at ure rec or ds o ver m uc h of t he H ol oce ne, we use d rec o nstr ucti o ns of s u m mer ( Ma y- S e pte m ber) te m p er at ures b as e d o n t he a n al ysis of c hir o n o mi d a n d di at o m ass e m bl a ges re c o vere d fr o m s e di me nt c ores of
t w o l a kes i n A bis k o Valle y ( L a ke 8 3 0 a n d L a ke V u os k k u), n ort her n S we de n3 7. Si mil ar t o t he tre at me nt of c h arc o al
d at a, t he i niti al c hr o n ol o gies were st a n d ar dize d a n d c o n verte d t o t he s a me te m p or al res ol uti o n ( 2 0 0 ye ars) pri or t o a ver a gi n g, t o o bt ai n a c o m p osite te m per at ure rec or d.
M o d er n r e c or ds a n d pr o xi es of at m os p h eri c a n d o c e a n v ari a bilit y.
We use d s atellite- b ase d se a-icec o nce ntr ati o n d at a fr o m t he Nati o n al S n o w a n d Ice D at a C e nter, B o ul der C ol or a d o U S4 3, f or t he i nter v al o
ver-l a p pi n g of t he m o der n fire d at a ( 1 9 9 6 – 2 0 1 4). S e a-s urf ace te m p er at ures ( S S T) si nce 1 8 7 0 were o bt ai ne d fr o m
t he U K Met O ffice Ha dle y C e ntre o bs er v ati o ns d at as ets4 4. We o bt ai ne d m o nt hl y ge o p ote nti al 5 0 0 h P a press ure
hei g hts ( ~ 2 0 0 k m2 gri d s c ale) fr o m t he N C E P/ N C A R R e a n al ysis pr o d uct fr o m 1 9 4 8 t o 2 0 1 44 5, 4 6. To re pres e nt
fire-c o n d uci ve we at her c o n diti o ns o ver t he peri o ds l o n ger t ha n t h ose c o vere d b y fire st atistics, we use d a m o nt hl y
0. 5° gri d de d d at aset of self-c ali br ate d P D SI ( versi o n 3. 2 14 7) o ver a n ort her n S we de n are a ( 6 0 – 6 4 ° N a n d 4 – 1 4 ° E),
w hic h are a v ail a ble f or 1 8 8 0 – 2 0 1 0. I n a d diti o n, we use d m o nt hl y 0. 5° gri d de d preci pit ati o n a n d te m per at ure d at a
( 1 9 0 1 – 2 0 1 3) fr o m t he C R U T S 3. 2 2 d at aset4 8.
As a p ale o N A C stre n gt h i n dic at or, we use d a n al ke n o ne- b ase d S S T rec o nstr ucti o n fr o m t he Vøri n g Pl ate a u,
N or we gi a n S e a (c ore M D 9 5 2 0 1 1 4 9). We als o use d t w o rec o nstr ucti o ns of N ort h Atl a ntic dri fi ice, a v ail a ble fr o m
c ori n g sites i n t he n ort h wester n (c ore K N 1 5 8- 4 G G C 2 2) a n d n ort he aster n (c ore V M 2 9- 1 9 1) N ort h Atl a ntic5 0.
Perc e nt a ges of Ic el a n dic gl ass a n d p erc e nt a ges of he m atite-str ai ne d gr ai ns were us e d as pr o xi es f or ic e r a fie d de bris (I R D) a v ail a ble fr o m t hese sites. fiese c hr o n ol o gies were c o n verte d t o t he s a me te m p or al res ol uti o n ( 2 0 0 ye ars), st a n d ar dize d a n d a ver a ge d t o o bt ai n a site-s peci fic rec or d of dri fi ice si nce 8 0 0 0 c al ye ars B P. We als o use d
a di n oc yst- b ase d rec o nstr ucti o n of se as o nal se a ice c o ver e xte nt i n t he L a br a d or S e a (c ore H U 8 4- 0 3 0- 0 2 1)5 1, 5 2.
St atisti c al a n al ys es.
We c orrel ate d ( b y usi n g Pe ars o n c orrel ati o n) t he c hr o n ol o g y of 2 0 t h ce nt ur y a n n u all yb ur ne d are as i n N ort her n S we de n wit h e ac h p oi nt i n t he S S T a n d 5 0 0 h Pa press ure fiel ds f or t he are a li mite d b y 2 0 – 8 0° N a n d 8 0° W – 3 0° E, a n d f or t he p eri o ds 1 9 4 8 – 1 9 7 5 a n d 1 9 9 6 – 2 0 1 4. Pri or t o t he a n al ys es, t he fire d at a were l o g- a n d s q u are-r o ot tr a nsf or me d t o e ns ure t he n or m alit y of t he distri b uti o n. St atistic al si g ni fic a nce of t he c orrel ati o ns was teste d t hr o u g h a t w o-si de d St u de nt t-test, wit h a ut oc orrel ati o n i n t he ti me series bei n g t a ke n i nt o
acc o u nt t hr o u g h a dj usti n g t he critic al v al ues f or t he t-test5 3. F or e ac h c orrel ati o n m a p, we als o c alc ul ate d fiel d
si g ni fic a nce, a me as ure of c orrel ati o n stre n gt h at t he sc ale of t he st u die d ge o gr a p hic al are a5 4.
A n n u al b ur ne d are as rec o nstr ucte d fr o m de n dr oc hr o n ol o gic al d at a were tr a nsf or me d t o c u m ul ati ve a m o u nts o ver a m o vi n g a ver a ge of 1 0 0- ye ars wit h 1 0- ye ar s hi fis, c alc ul ate d o ver t he peri o d fr o m 1 3 5 0 ( 1 4 0 0) t o 1 9 0 0 A D. fiis ste p res ulte d i n s m o ot he d fire c hr o n ol o gies at 1 0 ye ar res ol uti o n t o b e us e d i n re gi me s hi fi a n al ysis. fie te m p or al li mits of t he a nal yses were de fi ne d b y d at a a v ail a bilit y (t he st art of t he peri o d) a n d t he o nset of fire s u p-pressi o n (t he e n d of t he peri o d). The list of L F Ys o ver t he a n are a N ort her n S we de n, de fi ne d a n L F Y o n t he b asis
of c o nti n ge nc y a nal yses a n d t he s y nc hr o nicit y t hres h ol d o ver 1 2 7 0 – 1 9 1 4 A D3 9 w as use d i n c o nj u ncti o n wit h t he
n or malize d tree ri n g- b ase d rec o nstr ucti o n of s u m mer te m perat ure3 6. To t his e n d, we c o m p ute d t he me a n n u m ber
of L F Ys d uri n g “c ol d” a n d “ w ar m” peri o ds, as i de nti fie d b y re gi me s hi Th a nal yses (see bel o w), a n d t he n r a n d o ml y s h u ffle d t he res ult wit h t he re pl ace me nt of t he L F Y occ urre nces 1 0 0 0 ti mes, t h us o bt ai ni n g a b o otstr a p p e d n ull distri b uti o n of r a n d o m L F Y occ urre nce d uri n g e ac h of t he w ar m/c ol d peri o ds. We t he n c o nstr ucte d a 9 5 % c o n fi -de nce i nter v al ar o u n d me a ns a n d e v al u ate d t he st atistic al si g ni fic a nce of t he -de vi ati o n of t he o bser ve d e m piric al v al ue as c o m p are d t o t he n u m ber e x pecte d u n der t he ass u m pti o n of r a n d o m L F Y occ urre nce.
To e v al u ate t he s p ati al p atter n of de vi ati o ns i n t he Atl a ntic S S T a n d press ure fiel ds d uri n g 2 0t h ce nt ur y L F Ys,
we us e d s u p er p os e d e p o c h a n al ysis ( S E A 5 5). We s electe d t he fi ve L F Ys k n o w n fr o m t he m o der n f orestr y st
atis-tics ( ye ars 1 9 0 1, 1 9 3 3, 1 9 5 9, 1 9 6 9, 2 0 1 4)1 3. Alt h o u g h t he re plic ati o n f or t he S E A a n al ys es w as l o w, o ne s h o ul d
ac k n o wle d ge a n i ne vit a ble c o m pr o mis e b et we e n t he n oti o n of l ar ge fire ye ar (i.e. h o w i nte ns e t he fire acti vit y s h o ul d be t o q u alif y a ye ar as a l arge fire ye ar, L F Y) a n d t he n u m ber of s uc h ye ars i n a nal ysis. Ta ki n g a m ore c o n-s er v ati ve a p pr o ac h t o w ar dn-s t he de fi niti o n of L F Y we t heref ore a d o pte d a m ore o p p ort u nin-stic a p pr o ac h t o w ar dn-s n u m ber of ye ars, c o nsi dere d acce pt a ble f or t he a nal ysis. S E A a nal yses were d o ne wit h t he K N MI Cli mate E x pl orer
t o ol5 3, 5 6. S E A is b ase d o n a g gre g ati n g d at a fr o m ye ars wit h a p artic ul ar o utc o me (e. g. ye ars wit h l ar ge f orest are a
b ur ne d, L F Y), a vera gi n g t hat d at a, a n d assessi n g si g ni fic a nce of t he de p art ure of t hat a vera ge val ue fr o m t he me a n
c alc ul ate d o n t he c o m plete s a m ple d distri b uti o n of t he v ari a ble i n q uesti o n5 7. The met h o d is desi g ne d t o test f or
c o nsiste nc y i n t he res p o nse of a v ari a ble u n der or d uri n g a discrete e ve nt or f act or, wit h si g ni fic a nce bei n g teste d t hr o u g h b o otstr a p pi n g or e v al u ati n g distri b uti o n of t he c o m plete s a m ple d distri b uti o n.
To i n vesti g ate t he a gre e me nt of te m p or al c h a n ges i n fire acti vit y a n d t h os e i n re c o nstr ucte d air a n d o ce a n
te m p er at ure c h a n ges, we us e d a re gi me dete cti o n al g orit h m5 8. The al g orit h m f o c us es o n t he dete cti o n of ti me
p oi nts of c ha n ges i n t he me a n v al ues of t he res pecti ve ti me series a n d uses t he Re gi me S hi ft I n de x ( R SI)5 9, w hic h
w w w. n at ur e. c o m/ s ci e ntifi cr e p ort s/
∑
σ = = + ⁎ R S I x l , ( 2) i c c m i lw here m = 0, … , l – 1, wit h l as a c ut- o ff le n gt h of e v al u ate d re gi mes; σ l is t he a ver a ge st a n d ar d de vi ati o n f or t he
c o m plete set of o ne- ye ar i nter v als i n t he ori gi nal ti me series; a n d xi* is a n or malize d de vi ati o n fr o m t he hy p ot
het-ic al me a n le vel f or a ne w re gi me.
To test f or st atistic al si g ni fic a nc e i n t he R SI c h a n ges, t he met h o d reli es o n s e q u e nti al t-tests, s u bs a m pli n g ( M o nte C arl o tec h ni q ue) a n d bi as c orrecti o n, w hic h i ncl u des le ast s q u ares esti m ates of t he s eri al c orrel ati o n t o
pr o vi de r o b ust esti m ates of t he si g ni fic a nce of c h a n ges i n t he me a ns6 0. A si g ni fic a nce le vel of 0. 1 a n d H u b er’s
wei g ht p ar a meter of 1, w hic h c o ntr olle d f or wei g hts assi g ne d t o o utliers, were a d o pte d f or all of t he a n al ys es. fiis setti n g c a n be vie we d as a t hres h ol d bel o w w hic h t he se nsiti vit y of t he al g orit h m t o detect p ote nti al c h a n ges pr o gressi vel y decli nes. Re d n oise esti m ati o n w as d o ne t hr o u g h or di n ar y le ast s q u ares esti m ati o n wit h d at a pre -w hite ni n g -wit h t he le n gt h of s u b-s a m ple size bei n g e q u al t o 5 0 % le n gt h of t he de fi ne d c ut- o ff.
We a p plie d t he al g orit h m i n de pe n de ntl y f or e ac h of t he a nal yze d ti me series: t he n ort her n S we dis h fire acti v -it y, s u m mer S A T rec o nstr ucti o ns, e aster n N ort h Atl a ntic S S Ts, I R D a n d L a br a d or se a ice c o ver c hr o n ol o gies. fie a n al ysis w as p erf or me d f or t w o ti me s c ales. F or t he de n dr o c hr o n ol o gic al re c o nstr ucti o ns of a n n u all y b ur ne d are as (sites Bj ur h ol m a n d Ti ve de n) a n d n ort her n S c a n di n a vi a n s u m mer S A T, we use d a 1 0 0 ye ar c ut- o ff le n gt h. F or t he S S Ts, l a ke te m p er at ure esti m ates a n d c h arc o al- b ase d fire rec o nstr ucti o ns o ver t he H ol oce ne, t he c ut- o ff w as set at 8 0 0 ye ars. Pri or t o t he a n al yses, S A T, S S T, a n d I R D rec o nstr ucti o ns were detre n de d b y s u btr acti n g t he l o n g-ter m li ne ar tre n d i n or der t o s u p press t he i n fl ue nce fr o m a ny tre n d o ver t he e ntire a nal ysis peri o d.
R es ults
R el ati o ns hi p b et w e e n m o d er n fir e a cti vit y, S S T, a n d 5 0 0 h P a pr ess ur e h ei g hts.
O ver t he 2 0 t hce nt ur y, t he a n n u all y b ur ne d are a i n n ort her n S we de n c orrel ate d ne g ati vel y wit h S S T v ari ati o ns i n p arts of t he e xtr atr o pic al Atl a ntic Oce a n ( Fi g. 2). fie re gi o n wit h t he hi g hest c orrel ati o n ( r i n t he r a n ge of − 0. 3 t o − 0. 6) w as Fi g ure 2. C orrel ati o n b et we e n a n n u all y b ur n e d f orest are a i n N ort h er n S we d e n ( n ort h of 6 0° N) a n d m o nt hl y distri b uti o n of S S Ts i n t h e n ort h er n N ort h Atl a ntic pri or t o a n d d uri n g t h e fire s e as o n. fie are a wit h t he hi g her c orrel ati o n ( 4 0 – 5 0° N, 5 0 – 4 0° W) is i n dic ate d b y a s q u are o n all ma ps. C oe fficie nts si g ni fic a nt at p < 0. 1 a re mar ke d b y c ol o ur. C ol o ur sc ale refers t o t he v al ues of c orrel ati o n c oe fficie nt. S ee Fi g. SI 1 C f or t he ma ps of si g ni fic a nce le vels.
o bs er ve d wit hi n 5 0 – 4 0° W a n d 4 0 – 5 0° N, w hic h is cl os e t o t he Gr a n d B a n ks, a n are a m ar ke d b y a fr o nt b et wee n t he L a bra d or C urre nt a n d t he N A C. fie c orrel ati o n was hi g hest f or A pril- Ma y, w hic h is t he peri o d of t he ye ar f ol-l o wi n g t he m a xi m u m se a ice s pre a d i n t he n ort h west N ort h Atol-l a ntic a n d i m me di ateol-l y prece di n g t he fire se as o n. It persiste d o ver t he J u ne t hr o u g h A u g ust, s o me w hat di mi nis hi n g z o nall y b y e xte n di n g al o n g t he e aster n c o ast of N ort h A meric a. fie S S T vari a bilit y i n A pril- Ma y wit hi n t he re gi o n o ff t he Gra n d B a n ks acc o u nte d f or ne arl y 4 0 % of t he v ari a bilit y i n b ur ne d are as ( Fi g. SI 1 A). M ore o ver, a m o vi n g c orrel ati o n a nal ysis i n dic ate d t hat t he p atter n w as te m p or all y st a ble ( Fi g. SI 1 B). Alt h o u g h a nal yses of telec o n necti o n bet wee n N ort her n Atl a ntic S S Ts a n d fire
acti vit y s h o we d m o der ate le vels of si g ni fic a nce at t he sc ale of t he s electe d ge o gr a p hic al are a ( pfiel d < 0. 0 5 – 0. 2 0),
t his is a res ult of a li mite d ge o gr a p hic al e xte nt of t he li kel y “ce ntre of acti o n” ne ar t he Gr a n d B a n ks ( a n are a s ur-r o u n de d b y 4 0 – 5 0° N a n d 5 0 – 4 0° W) a n d t he v ast aur-re a i ncl u de d i n t hes e a n al ys es ( 2 0 – 8 0° N a n d 8 0 W – 3 0° E). C o nsiste ntl y l o w p v al ues i n t h at “ce nter of acti o n” ( p < 0. 0 0 5, Fi g . SI 1 C) s u p p ort t he n oti o n of telec o n necti o n bet wee n N ort h S c a n di na vi a n fire acti vit y a n d S S Ts.
S u per p ose d e p oc h a nal ysis b ase d o n t he fi ve L F Y i n t he m o der n fire rec or d ( 1 9 5 9, 1 9 6 6, 1 9 6 9, 1 9 9 7 a n d 2 0 1 4) a p plie d t o t he S S T fiel d p oi nte d t o t he s a me i m p ort a nt re gi o n of n ort her n N ort h Atl a ntic. H o we ver, t he ge o gra p h -ic al a n d te m p or al e xte nt of si g ni fic a nt S S T a n o m alies w as m ore ge o gr a p h-ic all y li mite d t h a n f or t he c orrel ati o n a nal ysis o n t he e ntire rec or d, a n d it w as te m p or all y restricte d t o A pril-J u ne ( Fi g. SI 2).
fiere was a str o n g ne gati ve rel ati o ns hi p bet wee n se a ice c o ver e xte nt a n d S S Ts i n t he n ort h west N ort h Atl a ntic (r ~ 0. 4 – 0. 6, Fi g. 3 A), es p eci all y d uri n g t he wi nter a n d s pri n g. fieref ore, we c alc ul ate d t he c orrel ati o n b et wee n Fe br u ar y- A pril s e a i c e c o n c e ntr ati o ns i n t h e Gr a n d B a n ks are as a n d t h e b ur n e d are as: i n t h e 4 8 – 5 2° N a n d 5 3 – 4 8° W re gi o n, a str o n g p ositi ve c orrel ati o n li n ke d t he t w o v ari a bles (r ~ 0. 5 – 0. 7, Fi g. 3 A).
fie ne g ati ve rel ati o ns hi p b et we e n S S T a n d a n n u all y b ur ne d are as s u g geste d t h at t here mi g ht b e a ge ner al rel ati o ns hi p bet wee n S S Ts i n p arts of t he N ort h Atl a ntic a n d t he s u m mer ari dit y of N ort her n Sc a n di na vi a. To test t his hy p ot hesis fr o m m o der n i nstr u me nt al d at a, we c orrel ate d t he self-c ali br ate d Pal mer Dr o u g ht S e verit y I n de x ( P D SI) i n J ul y- A u g ust o ver N ort her n S we de n i n t he 6 0 – 6 4° N a n d 4 – 1 4° E re gi o n wit h Atl a ntic S S T f or A pril- Ma y ( Fi g. 3 B). fie res ults s u g geste d t h at a w ar mer e xtr atr o pic al N ort h Atl a ntic w as ass o ci ate d wit h hi g her P D SI, w hic h c orres p o n ds t o m ore h u mi d a n d less fire- pr o ne c o n diti o ns i n n ort her n S we de n. B ec a use t he c h oice of t he ge o gr a p hic re gi o n t o a nal yze a ffects t he res ults, we als o r a n t he a nal ysis usi n g S S Ts a ver a ge d o ver t he are a li mite d b y 4 0 t o 5 0° N a n d 5 0 t o 4 0° W a n d c orrel ate d t he m wit h P D SI at e ac h gri d p oi nt o ver t he w h ole S c a n di n a vi a n Fi g ure 3. T elec o n necti o n bet wee n t he n ort her n N ort h Atl a ntic a n d fire we at her i n Sc a n di n avi a. (A ) C orrelati o n bet wee n a n n ually b ur ne d area i n n ort her n S we de n a n d sea ice c o nce ntrati o n ( N SI D C dataset) ar o u n d Ne wf o u n dla n d f or Fe br uar y t hr o u g h A pril f or t he peri o d 1 9 9 6 – 2 0 1 4. (B ) C orrelati o n bet wee n avera ge A pril- May S S T of t he n ort her n N ort h Atla ntic f or t he peri o d 1 8 7 0 – 2 0 1 0 a n d self-cali brate d P D SI f or t he area of n ort her n S we de n (a re gi o n li mite d by 6 0 – 6 4 ° N a n d 4 – 1 4 ° E) f or J uly, A u g ust. fie area li mite d by 4 0 – 5 0° N a n d 5 0 – 4 0° W is i n dicate d by a s q uare. N ote t hat l o wer P D SI val ues mea n i ncrease d dr o u g ht c o n diti o ns. (C ) C orrelati o n bet wee n self-cali brate d avera ge J uly, A u g ust P D SI f or t he area of n ort her n Fe n n osca n dia a n d avera ge A pril, May S S T f or area li mite d by 4 0 t o 5 0° N a n d 5 0 t o 4 0° W. I n all gra p hs c oe fficie nts si g ni fica nt at p < 0. 1 a re mar ke d by c ol o ur. C ol o ur scale refers t o t he val ues of t he c orrelati o n c oe fficie nts. fie fiel d si g ni fica nce a n d t he fracti o n of t he ma p wit h si g ni fica nt c orrelati o ns ( p < 0. 0 5), a re s h o w n f or eac h ma p. See Fi g. SI 1 D f or t he ma ps of si g ni fica nce le vels.
w w w. n at ur e. c o m/ s ci e ntifi cr e p ort s/
Pe ni ns ul a ( Fi g. 3 C). fie o bs er ve d p atter n s h o we d a p ositi ve rel ati o ns hi p b et wee n A pril- Ma y S S T a n d s u m mer h u mi dit y f or p arts of t he re gi o n, wit h t he str o n gest telec o n necti o ns o bser ve d i n s o ut he aster n S we de n, n ort her n S we de n a n d n ort her n Fi nl a n d. fie preci pit ati o n d o mi n ate d t his rel ati o ns hi p ( Fi g. SI 3). D uri n g t he 1 9 0 0s, t he a n n u all y b ur ne d are as i n n ort her n S we de n were c orrel ate d wit h preci pit ati o n r at her t ha n wit h te m per at ure ( Fi g. SI 3 A vs SI 3 C).
D uri n g t he fire s e as o n, hi g h press ure s yste ms d o mi n ate d o ver S c a n di n a vi a a n d t he Britis h Isles, acc or di n g t o t he S E A a n al ysis ( Fi g. 4, Fi g. SI 4). D uri n g t he wi nter m o nt hs, ye ars wit h a n a n o m al o usl y l ar ge are a b ur ne d were ass o ci ate d wit h ne g ati ve a n o m alies i n 5 0 0 h Pa press ure hei g hts, ce ntre d ar o u n d 5 5° N acr oss t he Atl a ntic. L o w press ure s yste ms l ater de vel o p e d o ver t he D a vis Str ait, s o ut her n Gre e nl a n d a n d n ort her n Q ue b e c d uri n g t he s pri n g ( Fi g. 4). C orrel ati o n a nal yses bet wee n 5 0 0 h Pa press ure hei g hts a n d a n n u all y b ur ne d are as, o per ati n g o n t he m o der n rec or d, i ncl u di n g b ot h L F Y a n d n o n L F Ys, re ve ale d ge ner all y si mil ar b ut a st atistic all y less pr o -n o u -nce d p atter -n ( Fi g. SI 4).
R el ati o ns hi p b et w e e n fir e a cti vit y a n d t e m p er at ur e v ari a bilit y o v er 1 4 0 0 – 1 9 9 0 A D.
A n n u all y b ur ne d are as rec o nstr ucte d at t he Bj ur h ol m a n d Ti ve de n sites re ve ale d te m p or al s y nc hr o nicit y acc or di n g t o t he re gi me s hi fi a nal ysis ( Fi g. 5), des pite di ffere nces i n t he le vels of a ver a ge fire acti vit y a n d site are as ( Ta ble SI 1). Fire pr o ne i nter v als i ncl u de d t he 1 6 0 0s at b ot h sites, a n d t he 1 8 0 0s f or t he Bj ur h ol m site o nl y. L o wer a ver a ge le vels of fire acti vit y were o bser ve d d uri n g t he 1 4 0 0 – 1 6 0 0 p eri o d ( 1 5 0 0 – 1 6 0 0 i n t he s h orter Ti ve de n rec or d) a n d d uri n g t he 1 7 0 0s. I ncre ase d fire acti vit y occ urre d d uri n g c ol der peri o ds, as s u g geste d b y t he re gi me s hi fi detecti o n al go-rit h m a p plie d i n de pe n de ntl y t o t he tree-ri n g b ase d te m perat ure rec o nstr ucti o n ( Fi g. 5)3 6. B et wee n 1 3 0 0 a n d 1 8 8 0,
L F Ys occ urre d si g ni fic a ntl y m ore fre q ue ntl y d uri n g t he c o oler t h a n w ar mer peri o ds, acc or di n g t o t he b o otstr a p a nal ysis. fie e m piric al di ffere nce bet wee n t he n u m ber of L F Ys i n c ol d a n d w ar m peri o ds ( 7) w as l ar ger t h a n t he 9 9. 9 % q u a ntile ( 5) of t he b o otstra p pe d ( n = 1 0 0 0) di stri b uti o n of t hat di ffere nce u n der t he ass u m pti o n of r a n d o m L F Y occ urre nce.
Fir e a cti vit y a n d e n vir o n m e nt al v ari a bilit y at H ol o c e n e s c al e.
N ort h er n S w e dis h fir e a cti v-it y si nc e 8 0 0 0 c al ye ars B P as re c o nstr ucte d fr o m s e di me nt c h arc o al4 1 te n d e d t o b e i n vers el y ass o ci ate d wit h
Fi g ure 4. S u p er p os e d e p oc h a n al ysis of m o nt hl y 5 0 0 m b press ure fi el ds ( N C E R/ N C A R d at as et) f or t h e fi ve l ar g est fire ye ars i n n ort h er n S we d e n o ver p eri o d fr o m D ec e m b er t o A u g ust of t h e c urre nt fire s e as o n f or 1 9 4 8 – 1 9 7 5 a n d 1 9 9 6 – 2 0 1 4. Are as wit h de vi ati o ns si g ni fic a nt at p < 0. 1 0 a re mar ke d wit h c ol o ur . S ee Fi g. SI 4 C f or t he ma ps of si g ni fic a nce le vels.
c hir o n o mi d- b ase d J ul y te m perat ure esti mates3 7 ( Fi g. 6), acc or di n g t o t he re gi me s hi fi a nal ysis. I ncre ase d c harc o al
fl u xes occ urre d bet wee n 8 0 0 0 a n d 7 0 0 0 c al ye ars B P, 6 0 0 0 – 5 5 0 0 c al ye ars B P, a n d si nce 3 0 0 0 c al ye ars B P. fie t w o l ater i nter v als were ass oci ate d wit h c ol der c o n diti o ns, e xce pt f or t he e arliest peri o d w he n hi g her fire acti vit y c oi n-ci de d wit h a w ar mer cli mate. H o we ver, i ncre ase d fire acti vit y br o a dl y c oi nn-ci de d wit h l o wer S S T i n t he N or we gi a n
S e a as i nferre d fr o m al ke n o ne- b ase d rec o nstr ucti o n of c ore M D 9 5 2 0 1 14 9 i n all t hree i nter v als.
S e a ice c o ver rec o nstr ucti o n of t he L a br a d or S e a fr o m c ore H U 8 4- 0 3 0- 0 2 1 re ve ale d t hree peri o ds of i ncre ase d ice c o ver si nce 8 0 0 0 B P, acc or di n g t o re gi me s hi fi detecti o n a nal ysis ( Fi g. 6). fiese were 8 0 0 0 – 7 0 0 0 B P, 6 0 0 0 – 5 0 0 0 B P, a n d si nce 4 0 0 0 B P u ntil t he prese nt ti me. I ncre ases i n L a br a d or ice c o ver br o a dl y c oi nci de d wit h i ncre ases i n fire acti vit y rec o nstr ucte d fr o m l a ke se di me nts i n N ort her n S we de n. Alt h o u g h we o bser ve d a li mite d s y nc hr o nic -it y bet wee n t he L a br a d or ice c hr o n ol o g y a n d t he c harc o al- b ase d rec o nstr ucti o n of Sc a n di na vi a n fire acti v-it y o ver 4 0 0 0 – 3 0 0 0 B P, we n ote d t hat d uri n g t he l atest peri o d wit h i ncre ase d fire acti vit y (si nce 3 0 0 0 c al B P), t he m aj orit y of d at a p oi nts re prese nti n g ice rec o nstr ucti o ns i n dic ate d a b o ve a ver a ge ice c o nce ntr ati o ns.
Rec or ds of I R D, rel ate d t o s e a ice s pre a di n g i n t he n ort h wester n (c ore K N 1 5 8- 4 G G C 2 2) a n d e aster n N ort h Atla ntic (c ore V M 2 9- 1 9 1), t o get her wit h se a ice c o ver rec o nstr ucti o n fr o m t he L a bra d or se a (c ore H U 8 4- 0 3 0- 0 2 1), i n dic ate d i ncre as es i n wi nter s e a ice c o nce ntr ati o n a n d its are al e xte nt ar o u n d 8 0 0 0 – 7 0 0 0 a n d 6 0 0 0 – 5 2 0 0 c al ye ars B P. B ot h i nter v als c oi nci de d wit h i ncre ase d fire acti vit y i n N ort her n S we de n ( Fi g. 6), alt h o u g h t he re gi me s hi fi a nal ysis di d n ot s h o w t he I R D i n de x i n t he peri o d si nce 3 0 0 0 B P as si g ni fic a ntl y di ffere nt fr o m t h at bet wee n ar o u n d 3 0 0 0 – 5 0 0 0 B P.
Dis c ussi o n
fie c orrel ati o n b et wee n t he 2 0 t h ce nt ur y fire acti vit y i n n ort her n S c a n di n a vi a a n d S S T a n d s e a ice c o ver i n t he
n ort h west N ort h Atl a ntic p oi nte d t o a telec o n necti o n bet wee n oce a n c o n diti o ns a n d t he fire c o n d ucti ve we at her of N ort her n Sc a n di na vi a. fie oce a nic re gi o n wit h t he hi g hest c orrel ati o n is l oc ate d o ff t he Gr a n d B a n ks, w hic h is a critic al are a i n t he N ort h Atl a ntic circ ul ati o n d ue t o t he c ol d L a br a d or C urre nt, fl o wi n g s o ut h w ar d a n d m a ki n g
a fr o nt wit h t he w ar m N ort h Atl a ntic C urre nt ( N A C) fl o wi n g n ort he ast w ar d6 1, 6 2. He nce, i ncre ase d se a ice s pre a d
a n d fres h w ater e x p ort vi a t he L a br a d or C urre nt i nt o t he N A C6 1 ma y i n d uce re d ucti o n of t he Atl a ntic Meri di o n al
O vert ur ni n g Circ ul ati o n ( A M O C). S u peri m p ose d e p oc h a nal yses o n t he n ort her n S we dis h L F Ys d uri n g t he 2 0 t h
ce nt ur y, s u g geste d a cl ose rel ati o ns hi p of re gi o n al fire acti vit y wit h wi nter se a ice c o ver a n d de vel o p me nt of l o w press ure s yste ms at t he wi nters pri n g tr a nsiti o n ( Fe br u ar y Marc h) i n t he s u b p ol ar N ort h Atl a ntic, f urt her pr o -m oti n g s o ut h w ar d s e a ice tr a ns p ort fr o -m t he L a br a d or S e a. A str o n g c orrel ati o n b et we e n t he a n n u all y b ur ne d are a i n N ort her n S we de n a n d t he se a ice c o nce ntr ati o n o ff Ne wf o u n dl a n d d uri n g t he wi nters pri n g m o nt hs s u p p orte d t he hy p ot hesis of a telec o n necti o n bet wee n t he oce a n o gr a p hy of t he n ort h west N ort h Atl a ntic a n d h y dr o -cli mat ol o gic al c o n diti o ns i n Sc a n di na vi a. A nal yses of hist oric al hy dr o gra p hic d at a, rec o nstr ucti o n a n d m o delli n g st u dies t o get her s u g gest t hat se a ice a n d fres h w ater e x p ort t hr o u g h t he L a br a d or C urre nt is o ne of t he i m p ort a nt f act ors c o ntr olli n g t he pr o perties of t he N A C at a n n u al6 1, dec a d al6 3, 6 4, ce nte n ni al6 5, 6 6, a n d mille n ni al ti me sc ales5 0.
fie i ncre as e d fl o w of c ol d a n d l o w s ali ne Arctic w ater i nt o t he wester n N ort h Atl a ntic we a ke ns t he N A C a n d
le a ds t o a s o ut h w ar d s hi fi of t he wester n st or m tr ac ks6 7 f osteri n g t he de vel o p me nt of hi g h press ure s yste ms o ver
n ort her n E ur o p e ( Fi g. 4) wit h d o mi n a nt dr y Arctic air m ass es i n s u m mer. This is c o nsiste nt wit h t he ne g ati ve c orrel ati o n bet wee n s u m mer ari dit y o ver m ost of t he S c a n di n a vi a n Pe ni ns ul a a n d t he a ver a ge A pril- Ma y S S T i n
Fi g ure 5. D y n a mics of rec o nstr ucte d fire acti vit y a n d M a y – A u g ust te m p er at ure (r u n ni n g 1 1- ye ar m e a n)3 6
i n n ort h er n Sc a n di n a vi a si nc e 1 3 0 0 A D. Thic k h oriz o nt al li nes re prese nt peri o ds of si mil ar me a n c o n diti o ns
i de nti fie d t hr o u g h re gi me s hi Th detecti o n met h o d b ase d o n se q ue nti al t-tests5 8 o n ori gi nal ( u ns m o ot he d) d at a
p oi nts. Arr o ws i n dic ate l ar ge fire ye ars est a blis he d t hr o u g h a c o nti n ge nc y a nal ysis o n t he net w or k of sites wit h
rec o nstr ucte d fire hist ories ( Fi g. 1)3 9. Fire acti vit y at sites Bj ur h ol m a n d Ti ve de n is prese nte d as c u m ul ati ve
a n n u al are a b ur ne d o ver a m o vi n g 1 0 0 ye ar fr a me wit h a 1 0 ye ar s hi Th. Yell o w b ars i n dic ate peri o ds wit h i ncre ase d fire acti vit y.
w w w. n at ur e. c o m/ s ci e ntifi cr e p ort s/
Fi g ure 6. H ol oc e n e-l o n g d y n a mics of fire acti vit y a n d e n vir o n m e nt al pr o xi es. C harc o al- b ase d pr o x y of fire
acti vit y4 1 a n d c hir o n o mi d- b ase d te m per at ure rec o nstr ucti o n3 7 re prese nte d c o n diti o ns o ver n ort her n S we de n.
S e a-s urf ace c o n diti o ns are ill ustr ate d fr o m S S T rec o nstr ucti o n o n t he Vøri n g Pl ate a u, N or we gi a n S e a c ore
M D 9 5 2 0 1 1 4 9, I R D occ urre nces i n t he n ort h west (c ore K N 1 5 8- 4 G G C 2 2) a n d n ort he ast (c ore V M 2 9- 1 9 1)
N ort h Atl a ntic 5 0, a n d se a ice c o ver esti mates fr o m t he L a br a d or S e a c ore H U 8 4- 0 3 0- 0 2 15 1, 5 2. S ee Ta ble SI 1 f or
e x act site l oc ati o ns. C harc o al rec o nstr ucti o n is a n a ver a ge of t hree l a ke se di me nt c hr o n ol o gies a n d c hir o n o mi d rec o nstr ucti o n is a n a ver a ge of t w o rec o nstr ucti o ns. fiic k h oriz o nt al li nes re prese nt peri o ds of si mil ar me a n
c o n diti o ns i de nti fie d t hr o u g h re gi me s hi fi detecti o n met h o d b ase d o n se q ue nti al t-tests5 8. 9 5 % c o n fi de nce
t he n ort h west N ort h Atl a ntic ( Fi g. 2 A). fiis is als o c o nsiste nt wit h t he res ults of m o delli n g e x peri me nts si m ul
at-i n g t he res p o ns e of t he fres h w ater dat-isc h ar ges o n l ar ge-sc ale wat-i n d fiel ds o ver t he n ort her n N ort h Atl a ntat-ic6 5. fie
critic al r ole of N ort h Atl a ntic S S T a n d N A C o n t he ari dit y of N ort her n E ur o p e h as b e e n f urt her de m o nstr ate d
fr o m o x y ge n is ot o p e d at a i n l ate H ol oce ne l ac ustri ne rec or d of S we dis h L a pl a n d: δ 1 8O
si de pleti o n e ve nts rel ate d
t o s hi fis i n preci pit ati o n s o urce fr o m Atl a ntic δ 1 8O
si - ric h air masses t o dr yer Arctic air masses were c o nc o mit a nt
wit h t he I R D ma xi ma i n t he n ort her n N ort h Atl a ntic 6 8.
C ha n ges i n press ure fiel ds o ver t he N ort h Atl a ntic a p p are ntl y a m plif y t he e ffect of S S T o n fire acti vit y i n n ort h -er n E ur o p e. C orrel ati o n a n al ysis o n S S T fiel ds s u g geste d a str o n g rel ati o ns hi p b et wee n a v-er a ge S S T c o n diti o ns a n d fire acti vit y ( Fi g. 2) w hile s h o wi n g a m uc h we a ker si g n al wit h res p e ct t o L F Ys i n S E A a n al ys es ( Fi g. SI 2). fie p atter n was re verse d i n t he a nal yses of 5 0 0 h P a press ure hei g hts, w hic h re ve ale d a m uc h str o n ger p atter n d ur -i n g L F Ys ( S E A a nal ys-is, F-i g. 4) t ha n c orrel at-i o n a nal yses o n t he c o m plete d at aset ( F-i g. SI 4). We s pec ul ate t hat t he rel ati ve c o ntri b uti o n of S S T d y n a mics vs. c h a n gi n g press ure p atter ns o ver t he N ort her n Atl a ntic o n fire acti vit y i n N ort her n E ur o pe v ar y al o n g a gr a die nt i n t he re gi o nal fire haz ar d, wit h press ure a n o malies bei n g i ncre asi n gl y i m p ort a nt d uri n g ye ars of ge ner all y c ol der S S T i n t he N ort h Atl a ntic. Al o n g wit h i nter n al c ycles a n d fee d b ac ks
wit hi n t he s u b p ol ar g yre 6 6, 6 9, e xter n al f orci n gs h a ve b e e n s u g geste d t o a ffe ct t he fl o w of c ol d Arcti c w ater t o
t he n ort h west N ort h Atl a ntic5 0, 7 0, 7 1. B as e d o n det aile d p ale o ce a n o gr a p hic re c o nstr ucti o ns, li n k a ges h a ve b e e n
pr o p ose d bet wee n t he decli ne i n s ol ar irr a di a nce a n d t he bl oc ki n g of hi g h press ure s yste ms t hat m o di fie d st or m
tr ac ks a n d we at her o ver Wester n E ur o pe7 2.
A n n u all y b ur n e d are as d e n dr o c hr o n ol o gi c all y re c o nstr u cte d f or t w o are as i n n ort h er n S we d e n re ve al e d str o n g s y nc hr o nicit y at ce nte n ni al ti me s c ales. We o bs er ve d i ncre as e d fire acti vit y d uri n g t he 1 6 0 0s a n d m ost of t he 1 8 0 0s, w hi c h were m ar ke d b y s urf a c e air te m p er at ures b el o w a ver a ge as i n d e p e n d e ntl y re c o nstr u cte d
fr o m tree-ri n g d at a ( Fi g. 5)3 6. fiis i ncl u des t he c ol dest i nter v al of t he Little Ice A ge ( LI A) t h at w as rec or de d i n
Sc a n di na vi a d uri n g 1 6 0 0s. C ol d i nter v als e x hi bite d a hi g her fre q ue nc y of L F Ys as i de nti fie d t hr o u g h s y nc hr o nic -it y a n al ysis of t he N ort h S we d e n- wi d e net w or k of fire hist or y s-ites. We, t heref ore, pr o p os e t h at c o ol cli m ate peri o ds i n Sc a n di na vi a were ass oci ate d wit h l o w at m os p heric h u mi dit y a n d i ncre ase d fre q ue nc y of hi g h press ure s yste ms d o mi n ate d b y c ol d a n d dr y air m ass es. S uc h a h y p ot hesis is s u p p orte d b y s u ns hi ne re c o nstr ucti o n i n Fe n n os c a n di a i n dic ati n g ge ner all y s u n n y s u m mers d uri n g t he LI A b ut cl o u d y c o n diti o ns d uri n g t he Me die v al
War m A n o m al y 7 3. fie ne g ati ve c orrel ati o n b et we e n s urf ace air te m p er at ure a n d fire acti vit y re p orte d f or t he
N ort her n E ur o p e m a y n ot b e u ni q ue. A n i ncre as e d fire fre q ue nc y d uri n g t he LI A h as b ee n re p orte d fr o m de n -dr o c hr o n ol o gic al rec o nstr ucti o n i n t he b ore al f orest of Wester n Q ue b ec a n d rel ate d t o a n i ncre as e i n s u m mer ari dit y7 4.
C ol der p eri o ds i n n ort her n S c a n di n a vi a mi g ht b e m ore pr o ne t o cli m atic e xtre mes as c o m p are d t o w ar mer
p eri o ds, pr o b a bl y as a res ult of m ore u nst a bl e circ ul ati o n m o d es a n d a n i n cre as e i n m eri di o n al fl o w7 5. We
o bs er ve d a n ass o ci ati o n b et we e n c o ol er cli m ate a n d t he l e vels of fire acti vit y w hil e c o nsi d ere d at t he d e c a d al t hr o u g h mille n ni a s c ales ( Fi gs 5 a n d 6). We s p ec ul ate t h at t he fre q ue nc y of fire- pr o ne p eri o ds li kel y i ncre as e d u n der c o ol c o n diti o ns d ue t o b ot h i ncre ase d we at her i nst a bilit y a n d t he fre q ue nc y of e xtre me we at her e pis o des
d uri n g t he ge ner all y c o oler p eri o ds. A st u d y of preci pit ati o n v ari a bilit y i n S we de n3 3 h as i n dic ate d a n i ncre as e d
v ari a bilit y at s u b- dec a d al ti me sc ales d uri n g t he LI A. F urt her m ore, l o w δ 1 8 O
si v al ues fr o m l a kes (s u g gesti n g a n
Arctic pr o ve n a nce of preci pit ati o n) h a ve b e e n n ote d f or p eri o ds c orres p o n di n g t o l o w te m p er at ure as i nferre d
fr o m tree-ri n g rec or ds3 5, 7 6. I n a d diti o n, si m ul ati o n e x p eri me nts s u g geste d i ncre as e d i nte nsit y of st or ms d uri n g
LI A7 7. O n t he a n n u al sc ales, h o we ver, te m perat ure d uri n g t he fire se as o n di d n ot a p pe ar t o c orrel ate wit h a n n u all y
b ur ne d are as i n N ort her n S c a n di na vi a ( Fi g. SI 3).
C h arc o al i n fl u x i n l a ke s e di me nts i n N ort her n S we de n 4 1, a pr o x y f or re gi o n al fire acti vit y, w as s y nc hr o nize d
wit h i n de p e n de ntl y re c o nstr ucte d re gi o n al te m p er at ure, S S T d y n a mics of t he N or we gi a n a n d Atl a ntic s e a ice f or a n i nter v al e nc o m p assi n g m ost of t he H ol o ce ne ( Fi g. 6). C h arc o al fl u x s h o we d a n i n vers e rel ati o ns hi p wit h
s u m mer air te m per at ures rec o nstr ucte d fr o m di at o m a n d c hir o n o mi d asse m bl a ges i n l a ke se di me nts3 7 a n d wit h
S S Ts of t he N or we gi a n S e a. fie d at a s u g gest hi g her re gi o n al fire acti vit y d uri n g p h as es of N A C we a ke ni n g a n d
s u bse q ue nt re gi o nal c o oli n g. I R D fr o m t he n ort her n N ort h Atl a ntic5 0 i n dic ate d p arti al c orres p o n de nce bet wee n
p ale o fire rec or ds4 1 a n d se a ice c o ver e xte nt, w hic h w as ne g ati vel y c orrel ate d t o S S T. I n p artic ul ar, a n i ncre ase i n
c harc o al fl u x c oi nci de d wit h a n i ncre ase d a m o u nt of se a ice c o ver i n t he L a br a d or S e a5 2 a n d I R D i n t he n ort her n
N ort h Atl a nti c at 7 0 0 0 – 8 0 0 0 a n d 5 0 0 0 – 6 0 0 0 c al. ye ars B P, w here as t he 7 0 0 0 – 6 0 0 0 c al ye ars B P i nter v al c or -res p o n ds t o a de cli ne i n c h arc o al fl u x a n d mi ni m u m s e a ice c o ver e xte nt ( Fi g. 6). A fier 5 0 0 0 c al ye ars B P, t he al g orit h m f aile d t o detect p eri o ds wit h si g ni fic a ntl y di ffere nt le vels of I R D. fie f act ors b e hi n d t he l oss of s y n c hr o nicit y b et we e n I R D a n d c h arc o al fl u x a Ther 5 0 0 0 c al. ye ars B P re m ai n u ncle ar. H o we ver, we n ote t h at ge o -gra p hic all y cl oser esti mates of cli mate vari a bilit y s uc h as A u g ust S A T rec o nstr ucti o n i n n ort her n Sc a n di na vi a a n d S S T rec or ds fr o m t he N or we gi a n S e a, c o nsiste ntl y e x hi bite d a cl ose c orres p o n de nce wit h fire acti vit y as i nferre d fr o m l ac ustri ne c h arc o al fl u x. Direct c o m p aris o n b et wee n t he l ac ustri ne is ot o p e rec or d a n d t he N ort h Atl a ntic
I R D als o s h o ws a g o o d c orres p o n de nce b ut m ostl y si nce a b o ut 2 0 0 0 ye ars3 5.
The pr o xi mit y of c ol d a n d dr y Arctic air masses ma kes t he n ort her n S c a n di na vi a n f orests se nsiti ve t o cli m ate v ari a bilit y. B ec a use t he s o ut h w ar d e x p a nsi o n of Arctic air masses is l ar gel y c o ntr olle d b y t he N ort h Atl a ntic S S Ts t hr o u g h t he s hi Th of westerl y st or m tr ac ks, o ne s h o ul d e x p ect str o n ger telec o n necti o ns b et we e n o ce a n a n d fire acti vit y i n N ort her n S c a n di n a vi a t h a n i n m ore s o ut herl y l o c ate d p arts of t he E ur o p e a n s u b-c o nti ne nt. I n de e d, n u meric al e x p eri me nts h a ve re ve ale d a m pli fie d di ffere nces i n preci pit ati o n o ver N ort her n E ur o p e fr o m str o n g
t o we a k A M O C7 8. This is c o nsiste nt wit h t he li mite d s y nc hr o nicit y of H ol o c e ne fire acti vit y i n n ort her n4 1 vs.
s o ut her n Sc a n di na vi a7 9, a n d wit h t he di ffere nt p atter ns of a n n u al fire acti vit y e vi de nce d fr o m b ot h de n dr oc hr o
n-ol o gic al rec o nstr ucti o ns2 8 a n d 2 0t h ce nt ur y o bser v ati o nal d at a1 3.
We pr o p ose t h at c ol d c o n diti o ns o ver t he n ort her n N ort h Atl a ntic, ass oci ate d wit h l o w S S Ts, re direct s o ut h -w ar d t he pre ci pit ati o n ass o ci ate d -wit h t he -westerli es a n d res ult i n i ncre as e d re gi o n al fire a cti vit y. O ur st u d y d e m o nstr ate d a pre vi o usl y u n d o c u m e nte d tel e c o n n e cti o n b et we e n t h e d y n a mi cs of Atl a nti c O c e a n a n d t h e
w w w. n at ur e. c o m/ s ci e ntifi cr e p ort s/
re gi o nal fire acti vit y i n n ort her n E ur o pe. fie c o nsiste nc y of res ults o bt ai ne d fr o m m o der n o bser v ati o ns, de n dr o -c hr o n ol o gi-c al, p ale oli m n ol o gi-c al a n d p ale o-ce a n o gr a p hi-c al re-c o nstr u-cti o ns s u g gests t hat tele-c o n ne-cti o n bet wee n N ort h Atl a ntic S S T a n d fire acti vit y o per ate d o n m ulti ple te m p or al sc ales d uri n g t he H ol oce ne, w hic h is rele v a nt f or u n derst a n di n g f ut ure fire acti vit y i n N ort her n E ur o p e. T w o fe at ures of t he re gi o n al cli m ate li kel y di vert fire acti vit y fr o m t he a p atter n of i ncre asi n g wit h te m p er at ure. fies e are: ( a) t he d o mi n ati n g e ffect of preci pit ati o n o n fire acti vit y ( Fi g. SI 3 C), a n d ( b) t he dec o u pli n g bet wee n S A T a n d preci pit ati o n ( Fi g. SI 3 E) t hat res ults i n fire acti vit y m ostl y res p o n di n g t o preci pit ati o n a n d at m os p heric h u mi dit y ( Fi g. SI 3 D).
fie li n k bet wee n S S T a n d fire acti vit y i n N ort her n S c a n di na vi a e xe m pli fies a f u n d a me nt al c o ntr ol of oce a nic circ ul ati o n o n n at ur al dist ur b a nce re gi mes i n terrestri al ec os yste ms, as alre a d y de m o nstr ate d i n ot her re gi o ns of t he gl o b e. Peri o ds wit h l ar ge wil d fires i n b ore al C a n a d a a n d Al as k a h a ve b e e n li n ke d t o t he Paci fic D e c a d al
Oscill ati o n ( P D O), El Ni n o S o ut her n Oscill ati o n ( E N S O) 2 6, 8 0, 8 1 a n d gl o b al S S T d uri n g t he n ort her n he mis p here
wi nter 2 4. I n p artic ul ar, t he w ar mer p h as es of P D O h a ve b e e n s h o w n t o res ult i n i ncre as e d fire acti vit y o ver t he
N ort h A meric a n b ore al f orest. I n Pat a g o ni a, ye ars of i ncre ase d fire haz ar d ha ve bee n s h o w n t o rel ate t o l o wer S S T
( L a Ni ña e ve nts) le a di n g t o he a vil y re d uce d preci pit ati o n d uri n g wi nter a n d s pri n g peri o ds8 2. I n E ur o pe, n o pre
vi-o us st u dies ha ve e xa mi ne d t he li n k bet wee n Atl a ntic S S T a n d fire acti vit y, alt h vi-o u g h t he pre dict a bilit y vi-of E ur vi-o pe a n
we at her fr o m oce a n d y na mics has bee n d oc u me nte d e arlier 8 3 – 8 5.
R ef er e n c es
1. B o n d- L a m bert y, B., Pec k h a m, S. D., A hl, D. E. & G o wer, S. T. Fire as t he d o mi n a nt dri ver of ce ntr al C a n a di a n b ore al f orest c ar b o n b al a nce. N at ure ( L o n d o n) 4 5 0, 8 9 – 9 2, d oi: 1 0. 1 0 3 8/ n at ure 0 6 2 7 2 ( 2 0 0 7).
2. B o w m a n, D. M. et al. F orest fire m a n a ge me nt, cli m ate c h a n ge, a n d t he ris k of c at astr o p hic c ar b o n l oss es. Fr o nt. Ec ol. E nvir o n. 1 1, 6 6 – 6 8, d oi: 1 0. 1 8 9 0/ 1 3. W B. 0 0 5 ( 2 0 1 3).
3. A mir o, B. D. et al. Fire we at her i n d e x s yste m c o m p o ne nts f or l ar ge fires i n t he C a n a di a n b ore al f orest. I nt. J. Wil dl a n d Fire 1 3, 3 9 1 – 4 0 0, d oi: 1 0. 1 0 7 1/ W F 0 3 0 6 6 ( 2 0 0 4).
4. d e Gr o ot, W. J. et al. A c o m p aris o n of C a na di a n a n d R ussi a n b ore al f orest fire re gi mes. Forest Ec ol. M a n ag. 2 9 4, 2 3 – 3 4, d oi: 1 0. 1 0 1 6/j. f orec o. 2 0 1 2. 0 7. 0 3 3 ( 2 0 1 3).
5. v a n d er Werf, G. et al. Gl o b al fire e missi o ns a n d t he c o ntri b uti o n of d ef orest ati o n, s a v a n n a, f orest, a gri c ult ur al, a n d p e at fires ( 1 9 9 7 – 2 0 0 9). At m os. C he m. P hys. 1 0, 1 1 7 0 7 – 1 1 7 3 5, d oi: 1 0. 5 1 9 4/ ac p- 1 0- 1 1 7 0 7- 2 0 1 0 ( 2 0 1 0).
6. M cr ae, D. et al. Vari a bilit y of fire be ha vi or, fire e ffects, a n d e missi o ns i n Sc otc h pi ne f orests of C e ntr al Si beri a. Mitig. a d a pt. str ategies gl o b. c h a ng. 1 1, 4 5 – 7 4, d oi: 1 0. 1 0 0 7/s 1 1 0 2 7- 0 0 6- 1 0 0 8- 4 ( 2 0 0 6).
7. B er ger o n, Y., G a ut hier, S., Fl a n ni g a n, M. & Kaf k a, V. Fire re gi mes at t he tr a nsiti o n bet wee n mi xe d w o o d a n d c o nifer o us b ore al f orest i n N ort h wester n Q ue bec. Ec ol og y 8 5, 1 9 1 6 – 1 9 3 2, d oi: 1 0. 1 8 9 0/ 0 2- 0 7 1 6 ( 2 0 0 4).
8. S t oc ks, B. J. & Ly n ha m, T. J. Fire i n ec osyste ms of b ore al E ur asi a (e ds. G ol d a m mer, J. G. & F ur y ae v, V. V.). 4 8 1 – 4 9 4 ( Kl u wer Ac a de mic P u blis hers, D or drec ht, 1 9 9 6).
9. Fl a n ni g a n, M. D. & Wott o n, B. M. F orest fires: be h a vi or a n d ec ol o gic al e ffects . ( e ds. J o h ns o n, E. A. & Mi y a nis hi, K.), 3 1 5 – 3 7 3 ( Ac a de mic Press, S a n Fr a ncisc o, C A, 2 0 0 1) d oi: 1 0. 1 0 1 6/ B 9 7 8- 0 1 2 3 8 6 6 6 0- 8/ 5 0 0 1 2- X ( 2 0 0 1).
1 0. P y ne, S. J. Worl d Fire: fie C ult ure of Fire o n E art h ( U ni versit y of Was hi n gt o n Press, 1 9 9 7).
1 1. B r a ds h a w, R. H., Li n d bl a d h, M. & Ha n n o n, G. E. fie r ole of fire i n s o ut her n S c a n di n a vi a n f orests d uri n g t he l ate H ol oce ne. I nt. J. Wil dl a n d Fire 1 9, 1 0 4 0 – 1 0 4 9, d oi: 1 0. 1 0 7 1/ W F 0 9 1 0 8 ( 2 0 1 0).
1 2. N i kl ass o n, M. & Gr a nstr ö m, A. N u m bers a n d sizes of fires: L o n g-ter m s p ati all y e x plicit fire hist or y i n a S we dis h b ore al l a n dsc a p e. Ec ol og y 8 1, 1 4 8 4 – 1 4 9 9, d oi: 1 0. 2 3 0 7/ 1 7 7 3 0 1 ( 2 0 0 0).
1 3. Dr o b ys he v, I., Ni kl ass o n, M. & Li n der h ol m, H. F orest fire acti vit y i n S we de n: Cli matic c o ntr ols a n d ge o gra p hic al p atter ns i n t he 2 0t h
ce nt ur y. Agr. F orest Mete or ol. 1 5 4- 1 5 5, 1 7 4 – 1 8 6, d oi: 1 0. 1 0 1 6/j. a grf or met. 2 0 1 1. 1 1. 0 0 2 ( 2 0 1 2).
1 4. B ir ks, C. J. A. & K oc, N. A hi g h-res ol uti o n di at o m rec or d of l ate- Q u ater nar y se a-s urface te m perat ures a n d oce a n o gra p hic c o n diti o ns fr o m t he e aster n N or we gi a n S e a. B ore as 3 1, 3 2 3 – 3 4 4, d oi: 1 0. 1 1 1 1/j. 1 5 0 2- 3 8 8 5. 2 0 0 2.t b 0 1 0 7 7. x. ( 2 0 0 2).
1 5. T al a n dier, C. et al. I m pr o ve me nts of si m ul ate d Wester n N ort h Atl a ntic c urre nt s yste m a n d i m p acts o n t he A M O C. Oce a n M o del. 7 6, 1 – 1 9, d oi: 1 0. 1 0 1 6/j. oce m o d. 2 0 1 3. 1 2. 0 0 7 ( 2 0 1 4).
1 6. M ars hall, J., J o h ns o n, H. & G o o d ma n, J. A st u d y of t he i nteracti o n of t he N ort h Atl a ntic oscill ati o n wit h oce a n circ ul ati o n. J. Cli m ate 1 4, 1 3 9 9 – 1 4 2 1, d oi: 1 0. 1 1 7 5/ 1 5 2 0- 0 4 4 2( 2 0 0 1) 0 1 4< 1 3 9 9 % 3 A A S O TI O > 2. 0. C O % 3 B 2 ( 2 0 0 1).
1 7. T r o uet, V., S c o urse, J. D. & R ai ble, C. C. N ort h Atl a ntic st or mi ness a n d Atl a ntic Meri di o nal O vert ur ni n g Circ ul ati o n d uri n g t he l ast Mill e n ni u m: R e c o n cili n g c o ntr a di ct or y pr o x y re c or ds of N A O v ari a bilit y. Gl o b. Pl a net. C h a ng. 8 4- 8 5, 4 8 – 5 5, d oi: 1 0. 1 0 1 6/j. gl o pl ac ha. 2 0 1 1. 1 0. 0 0 3 ( 2 0 1 2).
1 8. P e arce, C. et al. Oce a n le a d at t he ter mi nati o n of t he Yo u n ger Dr y as c ol d s pell. N at. C o m m u n. 4, d oi: 1 0. 1 0 3 8/ nc o m ms 2 6 8 6 ( 2 0 1 3). 1 9. S utt o n, R. T. & H o ds o n, D. L. R. Atl a ntic Oce a n f orci n g of N ort h A meric a n a n d E ur o pe a n s u m mer cli m ate. Scie nce 3 0 9, 1 1 5 – 1 1 8,
d oi: 1 0. 1 1 2 6/scie nce. 1 1 0 9 4 9 6 ( 2 0 0 5).
2 0. H urrell, J. W. & Va n L o o n, H. D e c a d al v ari ati o ns i n cli m ate ass o ci ate d wit h t he n ort h Atl a ntic os cill ati o n. Cli m atic C h a nge 3 6, 3 0 1 – 3 2 6, d oi: 1 0. 1 0 2 3/ A: 1 0 0 5 3 1 4 3 1 5 2 7 0 ( 1 9 9 7).
2 1. R o gers, J. C. N ort h Atl a ntic st or m tr ac k v ari a bilit y a n d its ass o ci ati o n t o t he n ort h Atl a ntic os cill ati o n a n d cli m ate v ari a bilit y of n ort her n E ur o pe. J. Cli m ate 1 0, 1 6 3 5 – 1 6 4 7, d oi: 1 0. 1 1 7 5/ 1 5 2 0- 0 4 4 2( 1 9 9 7) 0 1 0< 1 6 3 5: N A S T V A > 2. 0. C O; 2 ( 1 9 9 7).
2 2. V e ble n, T. T., Kitz ber ger, T. & D o n ne ga n, J. Cli matic a n d h u ma n i n fl ue nces o n fire re gi mes i n p o n der os a pi ne f orests i n t he C ol ora d o Fr o nt R a n ge. Ec ol. A p pl. 1 0, 1 1 7 8 – 1 1 9 5, d oi: 1 0. 1 8 9 0/ 1 0 5 1- 0 7 6 1( 2 0 0 0) 0 1 0[ 1 1 7 8: C A HI O F] 2. 0. C O; 2 ( 2 0 0 0).
2 3. G e d al of, Z., Pet ers o n, D. L. & M a nt u a, N. J. At m os p h eri c, cli m ati c, a n d e c ol o gi c al c o ntr ols o n e xtre m e wil dfire ye ars i n t h e n ort h wester n U nite d St ates. Ec ol. A p pl. 1 5, 1 5 4 – 1 7 4, d oi: 1 0. 1 8 9 0/ 0 3- 5 1 1 6 ( 2 0 0 5).
2 4. S ki n ner, W. R., S h a b b ar, A., Fl a n ni g a n, M. D. & L o g a n, K. L ar ge f orest fires i n C a n a d a a n d t he rel ati o ns hi p t o gl o b al s e a s urf ace te m per at ures. J. Ge o p hys. Res.- At m os. 1 1 1, d oi: 1 0. 1 0 2 9/ 2 0 0 5J D 0 0 6 7 3 8 ( 2 0 0 6).
2 5. Y oc o m, L. L. et al. El Ni n o- S o ut her n Oscill ati o n e ffect o n a fire re gi me i n n ort he aster n Me xic o h as c h a n ge d o ver ti me. Ec ol og y 9 1, 1 6 6 0 – 1 6 7 1, d oi: 1 0. 1 8 9 0/ 0 9- 0 8 4 5. 1 ( 2 0 1 0).
2 6. S ha b b ar, A., S ki n ner, W. & Fl a n ni g a n, M. D. Pre dicti o n of se as o nal f orest fire se verit y i n C a na d a fr o m l ar ge-sc ale cli mate p atter ns. J. A p pl. Mete or ol. Cli m . 5 0, 7 8 5 – 7 9 9, d oi: 1 0. 1 1 7 5/ 2 0 1 0J A M C 2 5 4 7. 1 ( 2 0 1 1).
2 7. M ar g olis, E. Q. & S wet n a m, T. W. Hist oric al fire-cli m ate rel ati o ns hi ps of u p p er ele v ati o n fire re gi mes i n t he s o ut h- wester n U nite d St ates. I nt. J. Wil dl a n d Fire 2 2, 5 8 8 – 5 9 8, d oi: 1 0. 1 0 7 1/ W F 1 2 0 6 4 ( 2 0 1 3).
2 8. Dr o b ys he v, I. et al. M ulti-ce nt ur y rec o nstr ucti o n of fire acti vit y i n N ort her n E ur o pe a n b ore al f orest s u g gests di ffere nces i n re gi o nal fire re gi mes a n d t heir se nsiti vit y t o cli m ate. J. Ec ol. 1 0 2, 7 3 8 – 7 4 8, d oi: 1 0. 1 1 1 1/ 1 3 6 5- 2 7 4 5. 1 2 2 3 5 ( 2 0 1 4).
2 9. R a a b, B. & Ve di n, H. Kli mat, sj ö ar oc h v atte n dr a g. S veri ges Nati o nal Atl as. S N A F örl ag, St ock h ol m ( 1 9 9 5).
3 0. A hti, T., Hä met- A hti, L. & Jal as, J. Ve get ati o n z o nes a n d t heir secti o ns i n n ort h wester n E ur o pe. A n n. B ot. Fe n n. 5, 1 6 9 – 2 1 1 ( 1 9 6 8). 3 1. Dr o b ys he v, I. et al. R e c o nstr ucti o n of a re gi o n al dr o u g ht i n de x i n s o ut her n S we de n si nce A D 1 7 5 0. H ol oce ne 2 1, 6 6 7 – 6 7 9, d oi:
