Distribution and Thawing of Permafrost in the Southern Part of the

Distribution and Thawinq of Perrnafrost
lll the SorrthernPalt of the
DiscorrtinuousPerrnafi'ostZone in Manitoba
J. THIE1
ABSTRACT. This study was car.ied out to evaluate the environmental factors
which influence the distribution and collapse of perennially frozen peats in the
southern part of the discontinuous permafrost zone in Manitoba. The changes in
permafrost bodies were measuredby means of aerial photography carried out over a
period of 20 years. About 25 per cent of the once occurring permafrost is still present. Melting appears to have exceededaggradation of permafrost since about 150
years B.P. Two types of collapse were noticed: peripheral collapse around very
small permafrost bodies: and a cehtral collapse for the larger bodies. The amount
of collapsehas varied from 0 to 30 metres horizontally in a 20 year pcriod.
RESUME. Drxtrt ntion et lonte dn permageldans la pattie sud de la zone de perma'
pel tlisco,tinu au Manitoba. Ol a mene cette €tude dans le but d'6valuer les facteurs
d'environnement qui influent sur la distribution et I'affaissementdes tourbes gelees
dans la partie sud de la zone de permagel discontinu au Manitoba. On a mesur€ les
changementsdans les massesde permagel par des photographiesa€riennesrecueillies
sur une p6riode de 20 ans. Environ 25 pourcent du permagel de nallcre est encore
pr€sent.Depuisenviron 150 ansA.P., la fonte du permagela 6t€ plus important€que
son extension.On remarque deux types d'affaissement:un affaissementpdriph€rique
autour de trCs petites massesde permagel et I'afiaissementpar le centre des masses
plus grandes.Sur une p€riode de 20 ans, cet affaissementa vari€ entre 0 et 30 mCtrcs
horizontalement.
-ttcps.tu.rrpll)tntoeeb.'ktlou\uc)ttu
l'fjal()llll.Pucnpornpatrctut
)otrht o(tllpoctto?o
3a.r'?a)tttn. ?p:trbt.r?pttttrtosu 'lluttunofu l;r'I.tr nlJoBc;leHr
r)rlcHIin B.rrItHnrr olipl;Iilllolr\el1 clx';16I rltl pIlclrp{'Jo. x'rnI(' tr l)rl3p] lnorI'o r( p3.rLIx
'Ixcrn
\rl)3.IbI\
rp) HToIl lr
3lr"I('lltxlIfl
3otrbI ocrpoBHolo
ToDrhsHrrrioB 1, rot(noil
ro.Illt
xiirl(Ptt ltlrl, lIPlt. lro\{onlll
lirlJr'lrl,l\
\Iarrrrro6u.
lLrrItn,nIrn
tl)oB,rnrotrt
t'crrrtr, npoB{);lnrloiin r('l,lrrt; 20 r,r" Ilprrf' rrrrrrrcrr'rro25'i trpex;re
rrilxrtltorrx
c ) n l { \ ' T t s o r r { E r u l r rxl c p 3 . - I t ' I \ r l ) J H T o B C ) I { e C T B I C T : l O l l l l ( l T o t l q e t o B I l t ' \ l i H l l l ' T T : r l l I T L H I I c .r r o B t r ; l r r r l o \ 1 \ 'r,r r r , u t . I or r l x ' o 6 ; l . r ; l r r l n U l l l l p x ; l { q n c i l \ ( l ) r t I l ' l \ r l , \ ' H T o H l ) U
,r('T T{)lll llrirr;r. Ilt,r,Iu oruoqctttl ;tB:r rnltlr lJt3l)l lIIcHIta: I)airl)}'In('Itl'c Hl
r( prrrlrcprru;t.rfl o.rottt' rtlr._Il'r\ ri('l)t.ilHri r{r.1rJl, n l)rlP\nIoIIIK' B Itt Hrl)o J,Ifl oo. tLrnIIx
r r a r c r r n o r . l i 4 r r r : n t t r n r l ' n o c p l t 3 l ) ) r r r c t l n c a r l ; ( B i r ; I I l r | r n r o r l r l r i l r I l l n t o ; ( I n ) .I . o r r . ' I)rcL
o r 0 ; l { ) : 1 0r r r t } o B .
INTRODUCTION
Permafrost, especiallyon its sensitivesouthern fringe has to be studied on an
ecosystembasis, for the environmental and biotic factors influencing its ecosystem
are so intricately interrelatedthat only on such a basis can the dynamic aspects
of perennial frosts be understood. In an attempt to assesssome of these complex
ecological relationships, the distribution and thawing of perennially frozen peats
in the southern portion of the discontinuous permafrost zone in Manitoba were
studied, and the results are presented in this paper.
lcanada Centre for Remote Sensing, Depa ment of Energy, Mines and Resources,
C)ttawa.Ontario.
190
I ) I SI R I B L l I O N . { N D T H A W I N C O I ' P F R M A F R O S T
P e r n r r l r o s it s d c h n c c aL s c l r t h n l i r t c r i l l sb l r . i D ga t c n t p e r a t u r e
bclou 0'(l trrr
t \ \ o o r - J l r o r L ' r c a r s\ .' a r i o u s z o n e ' so f i t h i L v cb e e n i d e | t i f i e d i n C l a n a d a( B r o $ D
1 9 6 7 )( ) n t h e h a s i so f t h c d i s t r i b L r l i oann c li r c q u c n c l o f p c r c n n i a l J ifr o z c n g r o u n t l
i n r c ' l a t i o nt o a i r t c m p e r a t L r r cZs o
. l t a i a n d T a r n o c a i( 1 9 6 9 ) h a v e m a p p e d i t i n
\ I a n i t o b a o n t h c b a s i so f t h c o c c u n c n c co l p c r m a l r o s tl a n d f o r m so r r c m n a n t '
i n p c . r t .l n t h e s o u t h c r nl l f t o f i t s d i s c o n l i l u o r , rzso t e , p c r n l i i f r o s ti s g e n c n r l l r
r e ' \ t r j c t c dt o p c x t l x n d si n $ h i c h p a l s a s ,p c a t p ) a t e a u sa n c l c o l l a p s cs c a r so c c L l r
in lrLrunciirncc.
I ) a l \ a \a r c n o u n i l s o I p c i i t ] a n d a n c lm i u e r a ls o i l u i t h a p c r c n n i a l l vI r o T e nc o r .
Thcv arc generailvmuch lcss than 100 mctrcs in (liamctcr. anci varv in hciqhr
b c t $ c c n I a n d 3 m . ( S j i j r sl 9 6 l ) a n t l m a y r e a c l r5 m . P e a tp l a t e a u sa r c c l c v a t c ( l
l i a t p c a t l a n a lhsa r i n g a p e r c n n i a l l rl 'n r z c n c o r c . ' l h e v a r e u s u a l l yl o u . a n d s e l d o n ,
cxceed 120 n. in hcight. but nay covcr arcasol ovcr I km:. 1n thc south. pcrma
frost in pcat plittcausdocs not Ltsuailvreach Lrndcrlling nrincral soil. ls it doc.
i n p a l s a s( Z o l t a i l 9 7 l ) , b u t i r l b o l h c u \ c s t h c : r r c a s c o n c e r n e dm a y b e h c a r i l l
wootletl rvith blirck spnrcc itntl white birch. Pitlsaslrnd pcat platcaus are sur
roundcd by unfrozcn. \\ ter-satirftltedpeat]and. conrmonll rvilh leacJsof opcll
\\'atcr.
P a l s a sa n d p e a t p l a t e a u sa r c m o r p h o l t r - t i c lrl a r i a t i o n so f t h e s a m c p r o c e . ,
( B r o u n 1 9 6 8 ) . I c c f o r n l t t i o n i r r p c a t d u r i n g r h c \ \ i n t c r c a u s e su p l i f t i n t . T h r
pcat thus raisccl,if sullicir'n1i\ (lr\, itct\ as un inslrlttor and ma1' pt-cscr!c tht.
frozcn core throughout 1he suntnrcr. ..\ftcr rctichinq thc ntaxintunt dc'grcc ol
dcvcl{)pment pcrmittcil bl clintrtic untl locul conrlitions. thc permafrost ma\
bcgin 1o tlccav as l rcsult ol sonrc tlislurbirnccin thc balancc of thc fact{)rs ol
u a l c f t r b l e . r ' e g e t a t i o ri.n s u l i r t i n rc:o r e r a n c lc l i n r l r t c . ' l h c d c c a l 0 r d e g r a c l t r t i o ] l
s t a g eI o r p r L l s u rsi n t l p c r l p l r l c a l r \ i s n l r l k c d b v a c o l l a p s co l c i l c e s a n c l t h c
p r c s e n c co i t h e r m ( r k i l r sllc a t u l c s ( B r o \ n l 9 6 i i ) .
L i l l l c i s k n o \ \ l I a b o u t l l l e r u t . o l c l e ! f r d l t i o n o r c o l l a p s eo f p a l s a sa n c lp c a t
p l a t e a r r s( .i ) l l a p s c s a c m s1 o l . r cl n i n l c g r i r lp a r t o i t h c l i f c c l c l e o f t h c t \ \ o l a n ( l
f o r n r sr r n c Ll ) c c u r sr r n r i c rc o ( r lc l i n r r t i cc o D d i t i o n s1. h c p r c s c n c eo l p c a t p l i l t c l r . l s
clLrrinirboth thc itggrl(lati()nanti riegratlation\tl!L-\ ncar thc sorlthern limit 1)i
p c r m a i r ( r s (t Z o l l r i l 9 7 l ) : h o r r s t h u l r l c g r a c l a t i cd{ ro e \ n o t n c c c s \ a r i l vi n d i c a t e
a n a m c l i o r a t i o ni n c i i n r a t t a i l c t r r h i c h h l \ b c c n s l i l t c d b r S r e n s s o n( 1 9 7 0 )
r \ i t h r c l e r c n c el o p a l s i r ri n S c a n c l i n i r r i l .t h a s b r c n \ u g g c s t c dt h a t a r c l a t i o n s h i p
a n t Lr g - u a r i i n gp h a s c sa r t r lc l i n l a t i cc h a n g c s t: h c p r c s c n l
c \ i s 1 sb c t \ \ c ' c ld c g l a c l i n g
prcvirlcnceof clcgrldation orcr ai:g:liclalion in this iu-cl is clnrscclbl ltn ovcltll
\ r ' i l r n r i n gl r c n d i n c l i n l l r t c( J r r h n : t o tct l a / . l ! 1 6 3 ) .
\f lt IoDs
r\ stLrdl'wes ma(lc ol cc()s\'\tcn)solcr a large ar.-a (about 1300 km.:) bv
survc,vingancl analysin-ul rrrmbcr ol charactclistic clcmcnts both throucb thc
intcrprctrtion of aeriltl phololrirphs rnrl br ntcans of sclectile llclcl sanrpling.
Elcments rvhich can bc anal\scd bY these tcchniqucs are: relief. drrintcc. rt'ge
lation. landlornl iin(l parcnt mrtcrial. ancl thc importancc of environmcntal
rvhen aerial photographiccovcraqe
changcsover pcriods oI time can bc asscssc-cl
periociicalll.
has bccn repcatecl
DISTRIBUTION AND THAWING OF PERMAFROST
l9l
The resultsof four differentcoveragesof the study area by aerial photography
are available:with oblique photographsin 1926; and with vertical photographs
of 1:15,400,1:68,000and
in 1946/47,1957and 1967,with scalesrespectively
l:15,800. A preliminary photo-interpretationwas carried out to selectsites for
ground truth stationsand a representativearea for detailedaerial-photographic
analysis.After the fieldworkhad beencompleted,the relief, drainageflow, permafrost distribution, vegctation,fire history and surficial geologywere s(udiedand
mapped from the aerial photographs. A detailed interpretation of sample areas
(totalling about 225 km.2) and coveringa wide range of differentconditionswas
done on the large-scalephotographs.The objectivewas to identify and measure
changesin the size of permafrostbodies and relate theseto drainagecondition
and flow, vegetation,forest fire history, thicknessof pcat, etc. Changesin the
superficialextent of permafrostbodies were marked on the 1967 photograPhs.
Percentagecollapse of permaffost was estimatedvisually and measuredin a
horizontal line perpendicularto the permafrost edge. A scale with gradations
of 0.1 mm. built into a magnifying system was used for the taking of
measurements.
DESCRIP'IION OF THE AREA
The study arca is locatedin central Manitoba and is borderedby the latitude
of 54' N, the longitudesof 99' W and 98o W, and Lake Winnipeg(Fig. 1). The
climate is moist subhumid,and characterizedby relatively cool short summers'
cold long winters and low precipitation.The averageair temperaturein July is
Frc. l.
Pcrmafrostandsurficialdcposits.
t92
DISTRIBUTION AND THAWING OF PERMAFROST
(1) Sedgefen with open water; (2) Tamarack-sedgepatterned fens;
Ftc. 2, Vegetation ,_vpe.!.
(3) Tamarack fens; (4) Black spruce- feathermosson permafrost; (5) Black spruc€- sphagnum, no permafrost; (6) White birch on permafrost; (7) Collapse scars. mainly with sedgesdrepanocladus
and sphagnum:(8) Regeneration
of black spruceon permafrost.40yearsafter
a fire.
about I 8' C, andin Januaryabout-22' C. Avcrageannualprecipitationis about
38 cm.(Weir1960).
Topographically
the areais veryflat,varyingbetweenabout218 m. and228 m.
abovesealevel.During the Wisconsinice agesomelarge glacio-fluviallandforms
were formed in the area. Later, as the glacial Lake Agassizcovered the area,
lacustrine clays and silts were depositedand the glacio-fluvial deposits rvere
modified by wave action into sandyplateauswith strand lines. After subsidence
of the lake level,peat filled in over most of the area,forming layersup to 5 n. in
thickness.
As organic depositscover 95Vo oI the land area, the vegetationtypes are
mainly ones associatedwith the different peatlands.They are shown in Fig. 2.
patternedfens and tamarackfens are characteristicof saturated
Tamarack-sedge
minerotrophic to weakly minerotrophic wetlands,and black spruce-sphagnum
raised bogs of wet ombrotrophic conditions.On perenniallyfrozen peat, black
spruce-feathermoss
communitiesare the most common,the sites usually having
a relatively densecover of black spruce,mixed occasionnallywith white birch.
The early vegetationin collapsescarsis commonly of a Carex-Drepanocladus
type. In later stages,hummock-buildingsphagnummossesmay appear, while
tamarack,white birch and black sprucemay finally invade the drier parts of the
scar. In many parts of the area the vegetationhas been disturbedby forest fues
which were startedby lightning. In order to evaluatethe influenceof such fires
on permafrost,a map was drawn on the basisof a photographicinterpretation
and other informationsuppliedby the Forest ProtectionBranch of the Manitoba
Departmentof Mines, Resourcesand EnvironmentalManagement.
Permafrostoccursin both shallow and deep organic deposits(Fig. 1) in peat
plateausand palsas.Thicknessof permafrostwas not measured,but data from
DISIRIBUTION
AND lHAWINC
OF PERtr{AFROST
193
surrounding areasgathcrcd b) thc aulhor indicated variations of from 0.5 m. to
5 nl. Permafrostin peat has a high ice content; and the averagemoisture content
Ior the frozen peat plateau core is about 80% (Zoltai 1972). Layers of pure ice
are rare and quite thin. The pcat dcpositsvary in thicknessfrom 0.5 m. to more
than 3 m. and have a sedge, sphagnum, forest and feathermossorigin. The
relatively well-drained pcat plateaus have forest peat, generally modcratclv
dccomposcd. Betwccn pcrmafrost landforms moderately decomposed fen peat
occurs in weakly minerotrophic areas, and relatively undecomposedsphagnum
peat in morc ombrotrophic situations.
RESULTS AND DISCUSSION
Distribution an(l development ol permalrost
Permafrostdistribution is strongly rclated to vegctation,drainageand surficial
deposits.Most permafrost and collapse scars- indicators of former existence
of pcrmalrost - are found in the medium deep (1 m.-2 m.) peat deposits.These
pcats tend to be ombrotrophic. with a semi-convcx rclicf. and shorv a radial
fen-drainagepattern in between coalescentpeat plateaus and palsas (Fig. 6).
The dccper, minerotrophic peats \&ith a scmi-concaverclicf, usually found in
sedgeand tamarack fens. only occasionallycontain permafrosti here permafrost
islands are seldom greater than 400 m. to 500 m. in lcngth,50 m. to 200 m.
bcin-q common. They have a round or tear-drop shape. In the forested shallow
p c r t r t l c s st h a n I m . t h i c l ) p e r m a l r o s tm a 1 o c c u r l o c a l l y .
Thc present and past distributions of permafrost give rise to the suggestion
that formation of permafrost started in the areas where black spruce-sphagnum
communitieswere likely to exist, and confirmation of it is provided by the fact
that young (incipient) peat plateaus generally occur undel dense (more than
4 per m.r) black sprucein a fcn, or nutrient-poorfen (Zoltai 1972). Such nutrientpoor conditions are predominant in such parts as have a convex relief which in
the study area is usually causedby mineral substratathat form slight domcs or
elseescarpments.Permafrostoccurrenceis a reflectionof the presenceof buried
strandlines,glacio-fluviallandforms, bedrock escarpments,etc,
AII peat platcaus and palsasexhibit a distinct pattcrn of drainage flow. Their
smoolh forms are mor.rldcdby slow flows of ground watcr (Fig. 3), the tails of
the teardrop shapes pointing in thc downflow direction. As permafrost tends
to develop first of all in areas wherc clumps of black sprucc occur, pelmafrost
bodiesmay be expectedto have a dominant growth vector in a down0ow direction,
and ncw permafrost tcnds to develop in the tail area where the fen type of vegetation
is gradually colonized by sphagnum and black spruce. When black spruce trees
are established,seasonalfrost can pcnetrate deeper in winter, as the snow cover
near the stem and under the branchcs is then considerably thinner than in the surrounding area. The shade of the branchesin summer provides better protection
"young" permaagainstheat inflow (Zoltai and Tarnocai 1971). Only relatively
frost bodies in fens with a poorly deflned and slow groundwater flow tend to have
circular shaoes.
r. ir
"
,,
'i.
".,.
;:.',
'..
l
,4
&,
'x,
;r*
\,
DISTRIBIJ'TION AND THAWING OF PER\,IAIRoS I
195
t,to. 4- Rel(titrt ol neltins 10 sizeol pernalIost la,Idlomt: (A) \hows completecollapseol'
small palsas(l); some larger ones have undergoneperiphernlcollapse.(B) showsperipheral
collapse(2) of a small p€at plateau,all the meltinghuvingoccurredovcr a periodof 20 yellr\'
(C) showscentfal collapses(3) of a medium-sirepeat plateau.(D) sho!\'sa lLrrSepeat plateilu
in which most of the permafrostdid thaw; only a small fringe of permafrost14)remainsfacing
the fen, and lhe collapsing(5) is proceedingJromthe inside.(Scalesof picturesA'D are lhe
Rate ol nelting
ol pernwlrosl
From aerial photographs it can be seen that a considcrable amount of thawing
occurred throughout the study area over the pcriod from 1947 to 1967, and also
from 1926 to 1947. No mcasurableaggradationof pcrmafrost occurred ovcr the
aggrcgateperiod of about 40 years
Retreat by edge collapse of permafrost landforms of lcss than 100 m. in dialtctcr was in the range of 15 m. to 20 m. ovcr a pcriod of 20 years.Ncarly all small
peat platcausshowed activc collapse.Most Pcrmafrostbodics having a diamcter
of 50 m. or lcss totally disappcarcdovcr a pcriod of 20 ycars (Fig. 4), collapse
bcing gcncrallyperipheral.
In fens nedium-sizedpcat plateaus(i.c., oncs with a diamctcr of about 100 nt.
to 750 m.) melt at relatively slow ratcs. Thc proccss of collapsc starts in thc
middle - probably first in thc lorm of a thcrmokarst holc in a maturc or ov!'rmaturc body. Alter a holc is formed, a slow collapscfrom thc insidc follows. thc
rates of horizontal retreat bcing of lcss than 5 m. in 20 ycars.Rctrcat of the cdgcs
by collapsecan bc from l0 m. to 20 m. ovcr a 20-1,carpcriod. Thc anrount rrf
melting tcnds to bc high in areaswhere small fringcs of permafrost areasarc lcft.
Usually thc outer edgesstart to collapsconly aft.-r thc inncr corc has largcly. or
partially, collapsed.In cxccptionalcascsthc cxistcnccof a drainagechanncl front
the insidc melting arca to the lowcr fcn arca sccmsto havc givcn risc to outcr-cdgL'
melting.
In largc pcat platcaus (over 750 m. in diamctcr) thc collapsing proccss also
tends to start in thc ccntre corc, but closcstto thc hcadward part of thc landfotm.
Thc cdge facing large fen arcas showslcss tcndencyto collapscthan do the edgcs
facing the other permafrostbodies,though thc latter arc soparatedby small fentype drainage channels. The ratc of rctrcat by collapse of distinct cdge is of the
order of l0-20 m. pcr 20 years.Othcr parts melted at ratcs of bctwcen 1 m. and
r96
DISTRIBUTION AND
'IHAWING
OF PERMAFROST
5 m. pcr 20 ycar period.Old collapscdparts of largepeat plateausfrcquently
showcdsignsof transformation
into a raiscdbog.
Thc differcnccin typc of collapscas b.twccnsmallbodicsand largconesmay
in fcns where
bc rclatcdto agcor stagcof dcvclopmcnt.
Smallbodies,especially
pcrmafrostdevelopment
is likcly to have startcdlater, nay bc expectcdto be
oncs.
)'oungcrthanthc mcdiumandlargc-sizcd
in particularnoting
By comparingthc 1926,1946/47 and 1967photographs,
the tonesrcsultingfrom invadingvcgetation
it waspossiblcto cstimatethc ageof
scarsup
collapscscars.By applyingthis tcchniqucto thc 1946/47 photographs,
to 60 ycarsold could be mappcd.Analysisof the superficialextentof collapse
scarsshowcdthat at onc timc about 60Ea of the total land area may havc been
permafrost.Of this maximumabout25Vo is still present-i.e. about l5Vo of
of the amountof
the total land arca (Fig.5). On the basisof a measurcment
thawingin relationto the former extentof pcrmafrost,it appearsthat degradation
Ftc. 5. Meltinr: ol per,nalrost: The stereopair indicate permafrost distribution at different
linesi the dotted line shows maximum distribution at any one time; the dark areas (RIGHT)
still have permafrost(l967). irhile in the dark areas(LEFT) it meltedbetween1946and 1967.
has been outpacing aggradation of new permafrost in this area for a considerable
time.Most collapsescarswcrc foundto bc ratheryoung;theywereusuallyfilled
rvith sedges,drcpanocladusand some sphagnum.The slightly older scarscontaincd more of the hummock-formingsphagnum,dwarf birch, bog laurel, leather
Icaf, etc. On thc basisof thc agedominanceof the young and medium-agedscars
and an extrapolationof thc prcscntrate of melting,it was estimatedthat extensive
meltingstartedbetween100 and 200 yearsago, and has progresscdmost rapidly
over about 120 yearsago titl thc present- a period during which morc than half
thc total amountof meltinghastakenplace.
Forest fires havc not had a mcasurablcinflucncc on the ratc of collapseof
permafrost.For areas burned between 1900 and the present no change was
noticed in rate of melting comparcdwith that for the non-burnedareasin the
immediatevicinity(Fig. 6). Many of the matureblack sprucestandscoveringthe
DISTRIBTJTION AND THAW]NO OF PERIIAFROS'T
197
Ftc. 6. Eflect of lorest fres: TOP LEFT indicalesno significantincreasein thrwing after
forest fire-the area to the north of lhc brolen line \\rs burned abou! 40 yeafs ago. $hile
active collapseedgesare
the area to the soulb has not been bufned. IOP RICHT-some
mature black spruceon permafrost;(C) - black
rndicatedby arrows:(A) - tamarack;lB)
spruce rcgeneration;(D)-pond remaining after complete collapseof permafro\l; (E)young collapsescar overgrolvnby regetation.BO_fTOi\l \ho$s suddencollapsecausedb) r
andcolapsed
rheburned
or '*t ,' .,1;ilili:ji.r"l.SJ:i.'J,lX.,',""r.1"rhsr;
burnlcomparison
pcrmafrost are frequently of firc origin. a fact which indicatcs that. even in this
southcrn discontinuouszonc, pe'rmafrostpcrsistsfor about 150 to 160 )cars aft!'r
thc occurrcnccof fircs. lt is possiblcthat a fire dcstro]s the forest covcr without
affectingthc pcat very much. Thc dricd pcat is a good insulator which prevcnts
heat influx and thawing. Usualll'a rcgcnerationof dcnsc black sprucc or white
birch (Fig. 2) occurscarly cnough to cnsurecontinuit) of protection.Some exccptions were howevcr found: for cxamplc a 1963 ire brought in its train a collapsc
of a peat plateau about 700 m. long (Fig. 6).
r98
DISTRIBUTION AND THAWING OF PERMAFROST
Permafrost was found to within 20 m. of the shorelinc of Lakc Winnipeg, and the
rate of thawing was not measurably influenced by the modifying influence of the
iake on microclimatic conditions.
Collapse and climate
The amount of collapsing that has occurred in the past, and is still occurring, in
this area is significant. A large proportion of the original permafrost has disappc-aredand littlc, if any, has formed over the same pcriod. Of course, collapsing
did not start at 200 B.P.; it always formed a part of the development cycle of
palsas and peat platsaus. Also, permafrost may still be formed in this area, but
only on specificsitesand under suitable climatic conditions.
An approximation of the age of permafrost in this area can be obtained by
nteans of infcrencc and deduction. On permafrost landforms black spruce stands
which date from a fire are as old as 150 years. Before such stands could have
formed the pcrmafrost must have become well developed, the process taking
about one generation of trees at least, but most likely severalgenerations,as small
clumps of black spruce trees are needed to initiate perennial ice-lens formations,
which may devclop slowly into palsas or peat platcaus, as described by Zoltai
(1972). From these estimateswe can infcr that densely wooded permafrost in this
arca is at least 180-300 years old. The permafrost in the area was formed sometime after deglaciation. As Lake Agassiz covered the area for a long period of
time, permafrost could not havc developed until the lake level subsided and peat
formed. On the basis of cvidcncc which includes the radiocarbon dating of a bog
just south of the area (Klassen 1966) it is likely that peat formation startcd well
after about 4,500 B.P. Two alternative eras of origin are: the cold period between
2-500 B.P. and 1500 B.P. (Nichols 19671,Bryson and Wendland 1967; Bray
1971); or the cold pcriod aftcr 600 B.P. The first alternativc seems somewhat
unlikely, as during the warm period of about 1000 years which followed it, permafrost would havc collapsed and most collapse scars would most probably havc
become overgrown. The sccond altcrnativc is thereforc a distinct possibility: most
of the permafrost in the study arca is probably not much older than about 600
vears and not younger than about 150 years.The southcrn limit of collapsescars
as mapped by Zoltai (Zoltai and Tarnocai 1969) must have had permafrost
dating from between 600 B.P. and 120 B.P. Possibly this limit was reachcd during
thc "Littlc Icc Age" which startcd around 550 B.P. and ended about 1850 (Lamb
1963) - most likely betweenabout 400 B.P. and 200 B.P.
Thc large increasein thc rate of collapsc of pcrmafrost in the study area seems
to be closely rclated to the climatic amelioration which occurred around 1850, or
about 120 B.P. As the climate became gradually warmer thc aggradation of
permafrost probably diminished. Permafrost might even now accumulate under
certain favourable conditions (such as perhaps a succcssionof relatively snow-free
cold winters) on certain sites. Espccially with the trend of cooling of the past 20
years,new permafrost might develop again in some area.
It is inadequate to suggestthat a single meteorological parameter, such as temperature, controls permafrost development and collapse. Vegetation is the other
important factor, which is itself dependent upon a number of climatic variables.
DISTRIRUTION AND THAWING OF PERMAFROST
199
The distributionof the hummock-buildingsphagnum,which seemsto be essential
for the developmentof wooded palsasor peat plateaus,is strongly related to
changesin conditions of moisture,temperatureand evapotranspiration,which
could well mcan that within the generalpattern of climatic changes,degradation
and aggradationas phasesin peimafrost formations have altcrnatedwith each
other.
frost penZoltai and Tarnocai(1971)found that thc depthto which scasonal
of snow
etrateson woodedpalsasandpeatplatcausis closelyrelatedto thickness
layer. Under denseblack sprucestandsthe thin snow alloweddeeperpenetration
of seasonalfrost than in the opcn where the snow formed a thicker and better
insulatingcover,often preventingthc seasonalfrost lrom reachingthe permafrost.
In yearsof relativclyheavysnowfall,seasonalfrost might not reachtbe permafrost table,cvcn under a denseblack sprucestand.Therefore,in suchyears(especially whenthe hcavysnowfalls carly in wintcr) the rate of collapsemight increase
at the sametimc as the rate of accumulationdccreases.If such a situationcontinued over a numbcr of years,a considerablechangecould pcrhapsbe cxPected'
The contrarymight be the caseif yearsof tiltle snowfalloccurred.
Duringthe last 100-150yearsthe southcrnlimit of localizedanddiscontinuous
permafrosthas movcd northward,perhapsfrom the southelnmostoccurtcncesof
It thc presenttrcndof coolingof the climatcdoesnot
collapscscarson pcatlands.
continue,the slow northerlymovementwill continue;and if the samctrend is
for a prolongcdperiodof timc, this probablcmovement
andcontinues
significant,
reversed.
andcventually
to thenorthmightwellbc stopped,
SUMMARY
'l'hc
AND CONCLUSIONS
prcsentand former distributionsof permafrostin the study arca are closely
rclatcd to thc thicknessof the peat layer, types of pcatland and vcgetation' and
thc rclief of un<lcrlyingmincral substrata,as caused by buried strandlinesand
glacio-fluvialdeposits.Permafrostoccurs most abundantlyin medium-deeppeats,
and in bog-type peatlands, in thc forrn of coalcscent peat plateaus. The smaller
permafrost landforms in fcn areastend to have a teardrop shape and are aggraded
according to drainage in downflow direction. Large pcat Plateausalso havc smooth
forms, streamlinedby drainage-waterflow, but are not usuallyof a teardrop shape
At one timc 6OVoof the land portion of thc study area containedpermafrost; at
prcsentthc proportion is down to about 157o of the larld area.
Permafrost bodics (in fens) smaller than 50 m in diameter have in gcncral
completelycollapsedovcr the 20-ycar period 1947-1967.Bodies smallcr tban 100
m. in diamcter (in fcns) collapsedperiphcrally.The amount of horizontal retreat
of thc permafrostedgewas in thc range ol 15 m. to 30 m. over a 20-year pcriod'
Bodies of about 100-500 m. (in fcns) had a low rate of collapse.Collapsestartcd
gencrallyin the middlc - probably the oldestpart - in thermokarstsinks. Rates
varicd usuallybetwecn 5 m. and 20 m. over the 20-ycar interval. Distinct collapse
edges had rates of about 10-20 m. per 20 years. Some parts had melting rates ol
about 0-5 m. pcr 20 years. Larger peat plateaus manifested the same type of
collapse - one starting in the inner core The permafrost edges facing large fens
were found to collapsc later than edges facing other ncarby permafrost bodies
200
DISTRIBUTION AND THAWING OF PERMAFROST
Downstrcam cdgcs tcnd to collapsclatcr than upstrcam cdges.After collapseof
Ihc larger pcrmafrost thcrc is a noticcable trcnd towards the development of
raiscd bogs. Thesc bogs do not have permafrost,but are suitable sitcs for new
p!'rmafrost formations. Forcst fires do not sccm to influcnce lhe ratc of collapse
in thc study arca. In a fcw cases,howcvcr, total collapscof permafrost has bcen
sct in train by fircs.
Collapsc prevailcd ovcr aggradationof permafrost in the study arca for somc
thcrcfore coincided with a
is for about 100-200 ycars B.P.-and
timc-that
climatic ameliorationwhich began at about 120 years B.P. Thc author suggcsts
that most permafrost in this area may have been formed betwccn 600 B.P. and
200 B.P., and that thc southcrnlimit of coliapsescarsas mappedby Zoltai (19'l l)
may have had pcrmafrost bctween400 B.P. and 200 B.P.
REFERENCES
BR^y, J. R. 1971. Solar-climate relationships in the Post-Pl€istocene.Science, l'11:. 1242'3.
RRowN,R. J. E. 1967. Permafrostin Canada.Canacla,Ceological Surtey Map 124A. first
edition.
-.
1963 Permafrostinvestigationin northern Ontario and northeasternSaskatcheu'an. C(nada, Nationul Re.\earchCouncil, Dii.\ion ol Building Research, Techttical Papel
no.29l.
1967. Tentative climatic patterns for some late glacial
BRysoN.R. A. and w. M. WENDLAND.
cpisodesin centralNorth America Protccdin!:s,1966Cottlcrctrccott Etlti'
and post-glacial
,nutir"rrrrl Sturli"tof the Glacial Ltke AgassizRegiotl.winnipeg: Universityof Manitoba
at
l963 PermafrostinvestiSations
JoHNsroN,6. H., R. J. E. BRowNand D. N. PIcKERsclLL.
Thompson, M^nitoba. Cana(la, National ResearchCouncil, Divisiotr of Building Research,
TcchniculI'aper no. 158.
R. w. 1966. Surficialgeologyof the Waterhen- Grand Rapids area, Manitoba.
KLASSEN,
Canatla,Geological Swvey, Paper 66'36.
1r.rr. 1963. On thc nature of certain climatic epochswhich diff€r from tbe modern
L.{N1B,
( t 900-39) normal. Pr oceedings,1961Symposium on Changesot' C limate. Rome: UNESCO'
NrcHoLs, H. 1967. The post-glacial history of vegetation and climate at Ennadai Lake,
Kee*,atin, and Lynn Lake, Mafiitoba. Eis,.eitalterund Gegenwart, lgt 176-97,
sJ6Rs,H. 1961. Suface patternsin bor€al peatlands.Endeavour,20'. 217-24,
svENssoN,H. 1970. Frozen ground morphology of northeasternmostNot\'/ay. Prcceedings,
S!mposium on Ecology ol the Subarctic R?gion. Paris: UNESCO.
\\,ErR,R. R. (ed.) 1960. Econo ic Atlas oJ Mdnitoba. Winnipegi Manitoba Department of
IDdustry and Commerce. 8l PP.
7()LT^r.s. c. 1g71. Southernlinit ol penlalrcst leaturcsi peal landlormsin Manitoha and
Toronto: CeologicalAssociationof Canada(Specialpaper no. 9).
Saskatchewan.
1972, Palsasand peat plateaus in central Manitoba aod Saskatchewan.Carudian
,lournal ol Forest Researclr,2i 291-302,
zoLTAr, s. c. and c. TARNoCAI.1969. Permafrost in peat landforms ifl nofihern Matritoba'
Proieedigs, Thifteenth Annual Manitoba SoiI Science Soeiety. Winnipegi University of
Manitoba.pp.3-16.
-.
197I. Propetties of a wooded palsa in northern M^nitoba. Arctic and Alpite
ReseaI ch, 3(2)t I I 5-29.
-.
Reprintedfrom ARC'ft(l. Joufnalof the Arctic lnslituteof Norlh America.
Volume 27. Number 3, September1974.