Jul 1, 2010 - in northeastern Canada and northwest Greenland (e.g.. Koerner .... 1825. 27.90. 250.0. 252.-1-. 250.0. 251.2. 1500. 28J)5. 1800. 28.00. ~-1 9.0.
AnnalscifClaciology 25 199 7 © Int ern ati onal Gl aeiologieal Society
An assessm.ent of the regional distribution of the oxygenisotope ratio in northeastern Canada MARIO
B.
GIOVIN ETTO, I G E RALD HOLDSWORTH,2 DAVID
A.
FI SHER,3 NIG E L M. WATERS, 1 H . JAY ZWA LLy1
IEarth Sciences 356, Universi£v cifCalgmy, Calgal)l, Alberta T 2N IN4, Canada ~ Arc/ic Institute of Nor/It America, JIlL T IlthJloor, Calgal~Y, Alber/a T 2N 1.N4, Canada 3 Terra in Sciences Division, Geological Surve] cif Canada, 601 Bootlt Stree/. Otlawa, Oil /aria E1A OE8, Ca nada I Depart men/ cif GeograJJlry, linivers ily cif Calgal]. Calgal~Y, Alber/a T 2,N I.V.f-, Canada "Code 971, .ASISA/ Goddard SjJace Flight Center; Creenbelt, JlJD20771, US. A.
ABSTRACT. A compil a tion of m ea n \'alues of the oxygen-isotop e ra ti o rela ti ve to sta nda rd mea n ocean wa ter (bI8 0 , in %0 ) for 22 sites represenLa ti" e of co nditi ons in northeas tern Canad a is co mplemented with da ta on mean a nnua l surface tempera ture, latitude, 'U!'face el eva ti on, a nd mean a nnua l shortes t d ista nce to open ocea n denoted by th e 10 % sea-ice co ncentra ti on b ounda r y. Ste pwise regre. sion an a lys is is used to develop a multi\'a ri ate model suita bl e to infer th e di stribution of (j 1HO in a n a rea of complex top og raphy a nd possibl y mi xed sourcing of ad\'ec ted water "ap or. Th e best m odel is produced by a run in the backwa rd m ode at the 95% confidence level in which onl y tem pe ra ture, latiwde a nd dista nce to th e o pen occan remain in the model ( th e correl a tion coe ffi cient is 0.915, th e adjusted coeffi cient of dete rmin ation is 0.809, th e root mea n squ a re res id ua l is 1.62). Thi s model is simil a r to the best (j IBO predi cti ve model derived elsewhere for G ree nla nd , suggesting a co mmon principa l so urce of advected moisture.
INTRODUCTION E a rli er studies of th e a rea I d istribution of 180 t,O va lues relative to sta ndard mean ocea n water (SMO\ V) (bwO in %0) in n ortheastern Ca nada a nd north west Greenland (e.g. Koern er, 1979) h ave foc used on bi va ri ate relationships between (jIHO a nd indep endent va ri a bl es such as mea n a nnu al surface temperature, surface elevati on, a nd di sta nce to th e coastline. In thi s stud y wc assess th e a real di stributi on of b 180 a nd covariati on of th e indep endent va riables using stepwise regress ion a nalysis. Our compil ation of (j 18 0 va lues determin ed from urn samples coll ec ted in ice caps a nd snow fi elds in northeastern Canad a li sts 23 sites (Fi g. I a nd Ta blc 1; sit e I on th e North Ice C ap, Greenla nd, is included because it sha res the same lower tropos phere Oow w ith at least nin e other sites to th e wes t). Da ta for o ne oth er site on the D e" on Ice Cap (a t about 115 km on th e tra\'erse route, 75.60° N, 83.3° W, 500 m (Koern er a nd Russell , 1979)), and fr om two other ice caps (M eig hen Ice Cap, 79.95° N, 99.4° W, 270 m (Koerner a nd Paterson, 1974-); Ba n KS Ice Cap (fi ve sites ce ntered at about 69.75° N, 72.0 \V at ele" a tio ns betwee n 500 a nd 870 m (H ooke a nd C lausen, 1982)) we re excluded from the co mpilation because the samples were collec ted from a reas eith er of n et ablati on, or of intense summer melt a nd p ercola ti on. On so me of th e ice caps, ice shelves, a nd snow fi eld s li sted in Ta ble I there a rc ma ny more sites [or which b'!lO values have bee n dete rmined. 'Vc selected relati vely few sites from each in a n attempt to attain a regiona lly representati ve (unbiased ) d ata base.
214
BO' W
Fig. /. Location cifsi/es lis/ed in Table /. Full circles indicate sites Jar which lIZeal! anllual surface /emjJeratllre could be de /erll1inedJrom surfilce data (Ts) . Crosses indicate si/esfor which mean an.nual sll1jace temjJerature was determined excLllsivelyJrom remote£v sensed data (T, ).
G'i01'inello and others: O,-)'gell -isotojJe ratio ill Canada rc/ble 1. Database sum maT)' and l/1odeled results
Si/f
Suura
.
L oca/ioll
. \ 0.
I :2
1 '1 :l
:1 I ;)
6 7 8 9 10 11 12 13 H 15 16 17 18 113 2513 250.0 2-1-9.0 2.i 1.0 2.)6.0 2(i2.0 2W.O 252.0 2.'10.8 2-1-8.6 2.113 252. 1 253.1 2.)0.7 :231.7 251.7
257.0 2 19.9 251.0 2.i I. 2 231.2 251.2 2.iO.1 252.7 2.) 1. 1 2.18.9 :2(jI.O :232.1
I H·() 180 1770 1770 1770 1770 1500 ISOO 11;20 1620 1321 1182 1015
- 20.2.1 21.H.i 29.86 29.00 28.!l2 28.81 28J)5
1100 19 10 1670 1700 1730 Ino 1825 1800 l:l90 13.10 660 f90 1.')20 1920 1800 11 00
~ I.I O
261.0
2 k6()
2~8.8
:lO. H)
2 1B.b 250.8
:Z7.90 27.05 29.2.i 27.90 28.00 28.10 26.!l() 21. 10 2L IO
:R30 :lO.OO :2B.70 31.50
I ~OO
28.00
lOO 10 10 10 10 890
22.00 :l2.30
~5U
~5 1. 1
251.3 250.0
251. 1 252.-12513 2.1l.-t 2.12.3 25.J.!l 260.1
~-1 9.0
2.)1.0 2.;(i.O 2!i2.0 262.0 2.)2.0 230.8 2-1-8.(j
33.30 29. 10 31.30 :2100
263+ 255.2 251.1 251.1
257.0
2.12.~
252.3 252.8 251.3 2.12.1 2.15.1 2.')0.7 2.i1.7 231.7 251.7
:2.)1.7 2.')7.0
2.~ 1.(i
J(iSO
25(1.7
IHO 1515 11 70
~.i I J ~5:l. - 1
2.'-1.1 250.7 2.) 1.7 2j1.7 251.7 2.'i7.0
H~.J
126() 1 ~.J-1-
J:l3fi I:l.i 6 900
~8.8.i
27.12 2.) .')9
22.1.1 20.59 28.25 - 28.58 ~O.69
27.92 27.J:l 2.).09 32.0 I 31.21 30.71 30.67 2-1-.00
1:1. J clrrics a nd K rOl ISl' 1987 ,; 11. Il ook,' and C I""""n l'lH2; 15. Holdswo nh I97:li. t Exc luded ri'olll lill a l a na h ses. t :\0 SUllllller ace llllllll a li o ;,.
South Do m e o rth e Ba rnes Ice Ca p (site 23) prese nt cd a spccia l case. }\ firn-t cmpcra ture m eas urem ent a t a depth of 20 m ( 8 C ) was co nsidered to be higher th a n th e ac tu a l m ea n a nnu a l surface temperature d ue to melting, percolati on, a nd fre ez ing processes, a nd in a preccding stud y th e surrace tcmpcrature was estim a ted a t 15 C (H oldswo rth, (973). Th is cstim a tc was based o n m c tco rologica l o bsc r\'ati o ns a t D ewa r L a kes (68.65 N, 74.2 \\', 518 m ) using a lapsc ra te of 6.5 C km I in th e ex trap ola ti o n. Wc included thi s site a ft er a fi rst run o[ th e a na lyscs were completcd (perso na l co mmunicati o n from R. Lc B. H oo kc, Aug ust 1996) a nd did not obt a in T,. d a ta (o r it. Nc\'C rth elcss. based o n L. 1'." a nd T,. d a ta fro m thc Penny a nd D e\'o n Icc Caps (sitcs 2 a nd 7) a nd using lapse r ates deri\'Cd [rolll sta nda rd a tmospheres fo r 60 :'-J a nd 75 1\ (Co le a nd oth ers, 1965) in the interpola ti o n, th e mean surface tempera ture fo r site 23 was estim a ted a t - 17 G Th erefo re wc c ntcred a n intcrm cdi a te 1'., \'a lue 01' 257 K in the da ta basc. Va lues o r mca n a nnua l sho rtest di sta nce to the o pcn ocea n (D ) in km \\'crc meas urcd bctwee n a sitc locati o n a nd th c pos iti o n o[the 10 % sea-i ce-co ncentrati o n bo und a ry de te rmined fr om :\'imbus-7 Sca nning ~Iulti- c h a nn c l ;\ [ic ro\\'a\'C Radi om eter ( S '\I~IR ) d a ta ro r 1978- 87 (Gl ocrscn a nd o th ers. 1992). 1'\0 a ttcmpt was m adc to mcasurc d ista nce to regio na l polynya during peri od s of prec ipita ti on (c.g. Schriber a nd others, 1977). Th c d a ta base thus asse mbled has inh erent shortcomings co nce rning thc two principa l \'ari a bles, (jIHO a nd 1'." 7;.: (I) Th e co mpil cd (j IHO \'a lucs a rc reprcsc nta ti\'C of a widc \'a ri ety o r depos itio na l cll\'iro nm cnt s (e.g. Fi shc r a nd o th ers, 1983) as well as acc umul a ti o n pe ri od s, m a ny not
215
Giovinelto and others: Orygel1-isotope ratio in Canada o\'e rl a pp ing a nd r a ng ing fr om fa ll throug h spring snow sa mples to a n und e termin ed number o f summer a nd w inter firn layers. (2) Approx im a tely h a ll' o f the compil ed T" va lucs we re obta in ed indirectl y, as describ ed a boye, o r indirect ly by extrapola ti o n or il1le rpola ti on. (3) Bilinea r interp ola ti o n o f THIR d a ta d ocs not nccessarily produce a m ore rel ia bl e tempera ture va lue for a specific site du c to the to pogr aphic co mplex it y o f th e regio n, pa rtic ul a rl y wh e n the d a ta resoluti o n (ab o ut 30 x 30 km. (Co mi so, 1994)) in seven o f th e sites includes bo th la nd a r eas (snow, firn , icc a nd ex p osed roc k sur faces ) a nd sea a reas (sn ow o n sea ice, sea ice a nd water surfaces ). 11l
Th e und esira bl e a ttributes o f th e l5 0 a nd T", 1'". d a ta m enti oned a bove, as well as th e rel a ti vely sm a ll sample ava il able fo r thi s stud y (va r ying b e tween N l+ a n d N23; N de no tes th e number o f sites in a se t), req uire th a t a caution a ry note b e a pplied to o ur findin gs. Th e shortcomings presented by th e small sa mple are a ddressed , in parl, usi ng th e a c!j usted coeffi cient o f d etermin a ti on sta ti sti c, R ,,2 (co rrela ti o n co cf~ fi cie nt, R, squ a red a nd m odifi ed fo r sample size (e.g. Tabachnick a nd Fidell, 1989)). In th e (a llowing sectio ns, th e sta ti sti cs are sig nilicant a t the 99.99% co nfidence level (F sta ti sti c und er the null hypothes is sh owing a proba bility of P ::::; 0.0001 ) unless sta ted o therwi se. A co nfidence !eye! selected [or a pa rti c ul a r m odel to d e termin e which va ri a bles contribute a t th a t le\ 'eI (or be Ller ) to the ex pl a na ti o n o f \'aria ti o n, is a se pa ra te sta ti sti c fro m th e P-\'a luc atta ined by th e m odel (e.g. Dav is a nd Sa mpso ll, 1973).
N 16; R is 0.840, R o2 is 0.684, root m ea n square residu a l (rms) is 2.09) (hencefo rth a ll bi va ri a te regress io n a nal yses a re summ a ri zed in Table 2 ro r easy compa r ison ). I n Fig ure 2 th e d a ta fo r site 13 showed th e largcs t d ep a rture rrom the fitt ed lin e (::::;3 fm s). Its exclusio n, to b e fu ll y justili ed , wo uld require a d ep a rture »3 rms; howe\'c r, in a sm a ll sample a nd eve r y thing e lse being equ a l, d a ta ro r a sing le site m ay g reatl y m odi[y fin a l res ults. Therefo re, we excluded d a ta 2 rrom tha t site (N I5; R is 0.931, Ra is 0.856, J'ms is 1.24):
818 0
= - 181.305
+ 0.610 Ts .
A compa r iso n o f T" a nd 1'"., including sites fo r ,,·hi ch th ese d a ta a re 3\'a il a bl e, showed a m od erat e co rrela tion (N I4; R is 0.810, R } is 0.628, rm s is 2.31, P is 0.0004), w her eas simila r compa risons in Antarct ica and Greenl a nd rcs ult in strong correl a ti on s (R \'alues o f 0.978 a nd 0.966, r e-
·22 -24 -26
o
'i"o
·28
0
·30
-32 -34
•
2a
-36 248
250
252
256
254
258
260
·24
9
o
-26
o
Zo -28
o r sm a ll d a lase ls it is a d v isabl e to run stepwi se a na lyses b oth ro rwa rd a nd backwa rd as the res ults m ay be difTerent (e.g. D av is a nd Sa mpson, 1973) ). Neve rth eless, bi va ri a te sta ti stics ltl for c5 O = f(T,,) showed a m od e ra te co rrela ti on ('rabic 2,
-30
o
0
0
o
·32
o 248
2b
250
252
250
256
258
260
o R,,2
NHt
N J.'i N I+i
N IS N22
= f(T,) ,)iBO = f (r,) ,)IHO = f(T,) 6 1"0 = f(T,)
0.8+0 0.931 OR28 0.858
0.68+ 0.856 0.669 0.722
2.09 1.2+ 2. 13 1.95
< 0.0001 0.872
0.798 0.707
1.67 2.11
< 0.0001 < 0,0001
NI.';
1.19
• Diil l'rCIlCC oLV I in each pair OfSClS is the incl usion or exc lusio n of'si le J:~.
Ste pI,vi se a na lysis o[ se t N22 of the rOrlll c5 1H O = f (L. H, 1:t, D ) a t th e 99,9 tYo co nfidence le\'el show th a t 1:, wo uld be th e o nl y \'aria ble to enter th e m odel in th e forwa rd m ode (Ta ble 6: R is 0.858, R} is 0.722, rm s is 1.95), Th e sa m e res ult is obt a in ed in th e bac kll'a rd m ode, in which th e va ria bles a rc relllO\'Cd in the fo ll owing o rd er: H , D, L, At th e
217
Ciovinetlo and a/hers: Oxygen-isotoj)e ratio in Canada ftlbLe 6. Sllll1ma1~JI qfstepwise regression anaLyses ( N22;
COlifidel/a In'r!
>99'Y" 95%
Slel) F;B
R
Ra'2
Parlial {orrelaliolls (Rp)
7;,
FiB
I ne.
Exc.
(m ocle l)
( m ode l)
L
1'0 1"1 "
L.D.H
0.858
0.722
- 0.611 0.386
0.01-1· 0.·j.()8
0.858
T;,
0.01+ 0.+08
L,D
0.722 0.756
0.611 0,386 - 0.233
0.858
0.838 0.883
11
0.91.'> 0.91.1
0.798 0.809
1'0
7;,
FI
95°(1
1IItiel)flldelllmriables
if' 0 as the dejJendent variabLe )
F2 t
7;,.H
BD BI "
L.T;,.D
a ll
H
0.031
fins
D
( model )
0.195 0.218
171
0. 195 0.218 0.155
3.71 1.95 1.83
1.9.1
1.67 1.65
" Samc resu lls as a run inlhe IJack\\'arclmodc 99%, step B3: i.c. T;, remains inlh c moclel. H , D. a ncl L arc re mO\'('d ill lhar ordel'. 90"" confidcnce IC\'CI. , Best model. Sam e resulls as a run in the backwarclmode at the 90% con fid e nce le\'c l.
t Sam e res ulrs as a run in the forward modc al t he
95% co nfidcnce level , Ta enters the model first, H scco nd 2 while Land D do not enter (R is 0.883, Ra is 0.809, rm s is 1.65). Runs a t the 90% co nfidence level in either the forward or backward mode did not show any changes. The best model produced using th e N22 se t in the backward modc a t the 95% confidence leve l is:
818 0 = - 203.501 + 0.782T,1 + (- 0.354L)
+ 3.814E -
3D.
(5)
This model is similar to th at obta incd rrom Greenland d ata IQr +6 sites (Zwally and Giovinetto, 1997), a lso a t the 95% co nfidence level in th e bac kward mode (R is 0.987, R } is 0.973, rms is 0.53):
518 0 = - 211 .082 + 0.7001:, + 0.104L
+ (- 1.085E-3D )
(6)
where T,. Land D , ha\'e the same units and werc obtain cd on the same basis described in this study. It is remarkable that the intercept a nd temperature coefficient values in Equations (5) a nd (6) are close, altho ugh th e Greenland data are, ror th e most p a rt, represe nta ti\'C of central and so uth ern regio ns, rather th a n northwestern par t.
DISCUSSION AND CONCLUSIONS Our findings indica te that despite the relatively sma ll data18 se t available to stud y the areal di stribution of (j 0 in northeastern Canada, it is possibl e to defin e a multi va riate model at a n acceptable co nfidence level. MOl"eO\'er, the model may be used to produce a co ntoured paLLern based on mea n an nu al surface temperature, la titude a nd mean annLlal shortes t distance to open ocea n, ignoring the effec ts o[ surface elevation. Inversion of Eq ua tion (5) prod uces ratio values (6 18 0 )i, Table 1) th a t illu strate loca l differences betwee n IR observation a nd model [(c5 IH O ) - (c5 O )i], The diflcrence is largest, as exp ec ted , [or the location on Axel Heib erg Isla nd that was excluded fro m most analyses (site 13, - 5.98%0). For th e 22 sites used in th e fin a l a n a lysis, the m ea n differe nce and std d e\'. are - 0.11 ± 1.50%0, with a ra nge - 3.60- 3.05%0 (sites 16 a nd 18, I"esp ee tively, both on D evon Island ). Th e d ifferenc e is sm a llest fOI" th e loca tion on th e Penny Ice Cap (site 2, - 00.08%0). Thi s is or interest because, toge ther with the difference for th e location 011 218
th e Barn es Ic e Cap (site 23, 1.00%0), th e two sit es li e Elrthest so uth a nd away from the cluster to the no rth, indic a ting that th e model is valid for a la rge a rea. Our d ata compilation is not suita bl e to exami ne th e variati on of c5 IH O rela ti\'C to elevation (e.g. O\'er th e area of a single ice ca p ). Nevertheless, it prov idcs the basis to assess why elevation does not contribut e to pa rticul ar stepw ise models. As li sted in Table I, the (jtHO and H values show no correlation (N22; R < 0.1, P > 0.9). Exclusion or the data for ice shelves (sites 19- 22) imprO\'Cs the co rrelati on (NI8; R is 0.758, P is 0.0003). Howe\'er, th e \'a riation over a relatively sm a ll area requires using res idu a l (jtllO \'alucs produced in three steps, a t each one remO\'ing the p a rtial variation expla in ed by L, D a nd Ta , resp ectively. There is no co rrel atio n between th e (j IB O res iduals produced at the third step a nd H (NI8; R 0.2). The b est (jIRO predictive model described ror north eastern Canada (Equ ati on (5)) is similar to the best pred icti\'C model descr ibed ror Greenland (Equation (6)), the laLLer based largely on data rrom its central and so uth ern regions for which the principal so urces of advected moisture are th e l\orth Atl a ntic sector extending [ram the Labrador Sea to the Norwegia n Sea and waters to th e so uth. This suggests that the bulk or the precipitation in Canada sa mpled by the sites compil ed for thi s study may share moisture advected from that sec tor. Th e descr ipt ion of th e sector mu st be qualifi ed in that it is based on the mea n annual position of the 10 % sea-ice concentration bound ar y. If sh ortes t di stance to open ocea n is meas ured using the mea n annual position of the 50% concentration boundary, the desc r ipti on of th e sec tor co uld be sta ted as extending rrom BafIin Bay to the Greenl a nd Sea and waters to the south and cas t. J n any event, the suggestion fits th e findings ora deta il ed study on the origin of Arctic precipitation (Johnsen and others, 1989).
ACKNOWLEDGEMENTS The authors graterully acknowledge contributions or S. Fiegles a nd R . Poitras in data processing, of R. Ko erner in assessing so me of the da ta, a nd of R. L eE. H ooke (01' alerting us to include data from the Ba rn es Ice Cap. ' Vc a lso ac knowledge th e co ntributions of th e two a nonymous rev iewe rs.
Clol'inello and olhers: REFERENCES eok.. \ . E... \ . Court a lld .\ . j. ~ alltor. 196.). l\ lock l a tm o'l'hlTcs. 11I \ ·a ll ~y. S. L.. ed. flour/boo!. ~r l!fo/)/~)'JiCJ oud JPll(f fllrirolllllt' lIlI. ~C\\' l urk. ell'.. 1\ IcC;ra\\- 1-I ill. 2.1 2.22. Comiso . .). C. 19~H. Surlitce tl"mpcratulTs in thc pol~lr rq,itJlls fi'om :\imbus 7 tempcratu re humidit y illfrared radiometer. ] (;eo/JI!I'J. ReI.. 99 C::l. .'l lfll :i200. Om·is, .). C. and R..). Sampsoll. 1971 Stat;.,t;r.' allrl rlata (IIW!I'.,;.';II gl'ologl'. :\c\\' \ ork. ('tl"., .Jo iln Wile\· and Sons. Fi sher, D .. \ .1079. Compari son
or 10 -)
years oroxY~l' lI isotope and im;o lubl c
impurit\ profiles li'om the 1)t'\'O n Island and Camp CClltury ice COlTS. QJlat. R",., 11 3,299 303. Fi sher, D. .\ . and R . i\1. ~o('rner. 198!!. The crfects of \\ illd 011 blHO a nd accumu lat ion g ilT an inferred record of seasona l () a mplitude from the ,\ ga"il. Ice Cap. Ellcsm('tT Island. Canada .. filII. Glar;o/.. 10.3-1 :1 7. Fi sher, D. .\ , alld R. \d . ~ ocrncr. 1991. Si g llal and noi se in rour icc-core record s I"rol11 tlie .\ ga"is iz Ice Cap. Ellc"'l11('rc bland . Callada: dl'ta il ~ o f' lhl" la".;t milil'll11iUlll I()]' stable iso topes, melt a nd so lid conducti\'i ty.
Ilulol'flle. 4 2 . 113 120. Fisher, D.. \ .. R.:'II. ~oerner. \\'. S, B. l'aterson, I\'. Da nsgaa rcl. :\. Gunclestrup
a nd :\, Reel\. 1981 Eflect of wind sco uring on climatic record s fi'om icecore oxygen-i sotope profiles. ,\ (Itlll't', 301 !5897 1, 20.1 209. Fi sher. D.. \ ., R . :'11. Ko('rncr ancl;\. Rc... lt. 1')9.1. I-Inl oc('I t(· clilllalic record s li'()1ll ,\ "a" iz k c Cap. Ell",mcrc Island, :'\.\\'. T., C a nada. H "lorfl(f'. 5 I, 19 21. Gim·inctto. :'11. B. a nd II..J. Z\\'a ll y. 1997. , \ rcal distribut ion or the "xI'gellisotope rat io in ,\n ta rctic-a: a n assess mcnt based on ll11dti\',-uialC' model s. . filII. Glor;ol.. 25 sce paper in this 1'lIlullle . Gloer, cn. P , 11·. j. Ca mpbcJl, D.j. Cavalieri, j. C. Comi so, c.:. L. Pa r kinson ,tIld H . .). /I,'alk, 1992.. frrtir alld .-fntarrtir '1'fI ire. 1978 1981' latellite PasSir'f lIIirmll'tIl'f ObH' rl '(IliOIl.l {flld (lno(r'liJ. '\ 'a~hinglOn, DC. ~alional .\ l'ronautics and Spare .\cilllini-t ralion. , i\,\S.\ SP-JII. 1~l o l d s \"onh , G. 1973. Correspond c nce. Barnl':-i Ic e Cap and e ng-li.ll'ial debri s in g lac iers.}. (;1(1l;0/.. 12 ((j+), 1,17 liB. I-I old s\\'orth, G, 198-1. Glaciologica l rcco nnai ssance of a n ice core drilling site, Penn)' Ire Cap, Ba lTin bland.] (;Ian o/., 3011 0-1, ~ 15. Hookt'. R . l.c B, and H . B. C lause n. 1982. Wisconsin a nd Ho loccne ()IHO
°,,-Jlgen-isolojJe ra/lo In Canada
I·ariations. Bames Ice Cap. Canada. (.'fol . .1'0(, . fill, BIIII.. 93 8 . 78+ 7fl9 . .Jdl'rie,. i\1. O. a nd 1-1. R. ~rnll '(,. 19f17. Sno",I,,11 and oxyge n-isotope I'a riaIiOI1' ofT the tl orth (,O" Sl Or F:llcslll ere Island . :\.\\'.'1'.. Cana d a. ] Clor;o/.. 33 111, 19:; I')'). .JOh lhl'tI . S..J. . 1\'. O ansgaa rd and .). 1\'. C. I\·hite. I'IR'I. Th e ori~'in or . \ rnic precipitation under prese nt ;lIld glacia l cond itions, 'JflluL 4IB I+ . +32+68. ~oe rn l'r, R. ~l. 1979. Accu tllulati o n, ablation, and oXYRe n isotope I'ariati ons on the Queen Elizabeth Isla nd s ice ca ps. Canada. ] G/llr;ol .. 22 iHil. 2.1 1·1. I":'ocrttcr, R. :\1. a nd 1\'. S. B. Patcrsoll. 197·1. .\tlaksis ora COlT Ihroug h till' j\Lei g-hl' 11 Ic e Cap .. \rClj c Canada , and its palcnriimatic i ll1p li cat i{)n ~. Qjlllt. Rc, .. 4 3 ,253 263. ~ol' rner. R. J\1. and R, D. Russell. 1979. () tI'O \'ar iati olls in Sno\\' on the Dl'I'On Isla nd ice cap. :\o rthll'cst cle rrilOries. Canada. Call. ] Earth Sri .. 16 7 . 1119 I J:!7. :'IliiJl l'r. F. 1961 Ellgla c ialtel11pcrature n1l'asure nlt'nts. III :'I1,·tller. F. IIl1d 18 othm. ('Ill. ]lIrobll'"-. lhCill. frrlit' RPltrlrrh E,/)('(Iitioll IIJ.!') 1')62: /J/d;III;lIa~1 "!lOlt 1961 1%'2. :'Il ont r"a l, Que.. :'Ilceill L:ni",·r, it\. 81 89. . \ xcl Heiht'rg- I-; ianci Re'c:an.:h Repo rl :o.. J\ lcd1cr, F , B. Stalllkr and C. Schribcr. 1977. I SO l Op t'lncasuI'
