Glacier-front measurements

(1)

Front variations of outlet glaciers from Jostedalsbreen, western Norway, during the twentieth century

STEFANWINKLER

Wink ler,S. 1996:Frontvariationsof out letglaciersfrom Jost edalsbr een,western Nor w ay, duringthetwentiethcen- tury.Nor.geo/.unders.Bull.437,33-47.

Therecord of glacie r-fro ntposit io nchanges at 15individu al out letglaciersfromJostedalsbreenduringthe twenti- ethcent uryispresent ed and interpret ed.Aspecific patt ern ofglacie r-fro ntoscillat io nsisrecognised,wit htw oread- vancesduringthefirstpartof thecent ury,foll owed byast rong ret reatinthemiddle of the cent uryturni ng intothe presentstro ngadvance.Differencesin behaviourofindivid ualout lets dependingonspecifi crespo nse timespermit theirclassifica t io ninto three groups.

Ininterpreti ngthe interactionglacier-fr ontvariations/climaticfluctuation s,nodominatingfactorcan be traced .The importanceofindi v idualmet eorologicalfactorshasbeenchanging throughoutthecent ury.Highwinter precipitati- onandlo w sum mertemperaturescausedthetwo readvancesofthe firstpartofthe twentie thcent ury.Theretreat inthe middleofthecent urywasmainlythe resultof highsummertem peratures.The presentadvanceisprimarily theresult of large winte rsnowaccumu lat io ns.Alt ho ug h mass balance data arefar mor e suitab lefor the characte ri- sat io nofthe relat io nship bet weenclimate andglaciers,therecord ofglacier-fro nt position changespresente d here alsoconta insvaluable inform ati o n.

Stefan Wink /er,Geographischeslnsti tu t,UnivetsitatWurzburg,Am Hub/and,0-97074Wurzburg,F.R.G.

Introduction

Glaciers are sensitive indicatorsof climatic fluctuations, today as in the past. Although the climatic significance is reduced by several glaciological factors as well as factors of the natural environment influencing glacier-front oscillations in addition to the influence of the climatic system, glacier-front variations are favourite topics of study.The fronts of the glaciersare easily accessible and measurement records are available for almost all glacier regions over time periods of up to 100 years and more. Therefore, climatic fluctuations are far more often viewedin relation to measured glacier-frontvariations than to mass balance measurements. Although mass balance data would be preferable for the analysis,with fewer precautions being required,there are unfortunately too few data series available.

Glacier-front oscillationsof the recent advance of the outlet glaciers of Jostedalsbreen are very remarkable as they contrastwith glaciers in other regions such asthe European Alps. At Jostedal sbreen,the first annual measurements of glacier-front position change were made around 1900 at up to 15 individual outlet glacier tongues (see Fig. 1). Unfortunately, measurements stopped around the middle of the twentieth century at a large number of glacier sites. Only at four glaciers (Brigsdalsbreen, Faberqstelsbreen, Nigardsbreen (not annually but only over a short period)and Stegholtbreen) are there continuous recordsof glacier-front measure- mentsup to today(measurement s have been carried out at Austerdal sbreen as well, but because of a break in observationsduring the 1920's,a complete record islac- king).

Although the continuous records of Brigsdalsbreen, Fabergstolsbreen and Stegholtbreen and their response to climatic fluctuations have already been studied (Nesje 1989) and as an overview of the glacier-front variations during the first half of the twentieth century has already been given by Fcegri (1948), a detailed presentation of all available recordsof the Jostedalsbreen outlets has not been published.Even though most of the glacier-front variation records only apply to the first half of the twentieth century,some important inferences on glacialdyna- mics and glacier-front oscillation chronology can be made by comparing the data of the outlet glaciersin terms of their differences in size,morphology and aspect.

This paper presents the record of glacier-front variations of all the Jostedalsbreen outlets measured during all or parts of the twentieth century,depending on the available informat ion(Winkler 1994).By using meteorological data from stations around Jostedalsbreen available for the period of glacier-frontobservations,the relationship between glacier-front oscillat ions and climaticfluctuati- onshas been analysed.In addition, the influence of different aspects of climate can be determined by referring to mass balance studies at Nigardsbreen which were begun in 1961/62(see 0strem et al. 1991).

Glacier-front measurements

Around 1900,Rekstad(1901) began his research at the outlet glaciersof Jostedalsbreen and established a series of marks for the subsequent measurement of changesin the positions of their snouts.Initially, measurements were carried out annually at only a few of the glaciers.

Unfortunately, an advance occurred during this period

(2)

34 StefanWink/er NGU-BULL4 3 1.1996

7-20

6-40 6-"0 1'. ^7-10 ^7-⁾⁰^E

10km

N DOJj

...

Byrkjtfllo

AUD Austdalsbreen AUS Austerdalsbreen BER Bergsetbreen B0D Bedalsbreen B$2lY Bevabreen BRE Brenndalsbreen BRI Brigsdalsbreen ERD Erdalsbreen FAB Fflbergst!i'l lsbreen

•

HAU Haugabreen KJE Kjenndalsbreen LOD Lodalsbreen MEKMelkevollbreen NIG Niqards breen ST E Ste gholt breen SUP Supphellebreen TUF Tuftebreen

TUN Tunsbergdalsbreen

,

Fig.1.Sketchmapof Jostedalsbreen.

(3)

NGU-BULL431,1996

and as a result neither the exact starting points nor the exact advance distances are known for a number of glaciers(see below).However,even though these measurements were not made annually at all 15 outlet glaciers, small retreat or advance distances can be interpreted as a consequence of two or three years of retreat followed by a few years of advance,result ing in a small overall positional change over the whole period.

Annual observations were made during late summer mostly by local people.The compiled resultsof the measured position changes were reported annually by Rekstad (1910) in the Bergens Museums Arbok, comple- mented by remarks on important events (dramatic changes of glacier morphology, appearance of proglacial lakes,difficulties during measurement procedure, etc.).

However, frequent field checks by Rekstad himself, as numerous photographs from the first decade of the twentieth century show (Fotoarkiv, NGU Trondheim), give us anidea about the accuracy attained in the reports of Rekstad and later of Fa=gri (1934b).The result s of the measurements are usually specified for single marks which makes it possible to distinguish between the different sections of a glacier tongue that might have under- gone appreciable changes in morphology (e.g. during periods of strong retreat or advance).Unfortunately,tho- se annual measurements stopped after 1941 at the majority of the Jostedalsbreen outlets. Apart from the five glaciers at which measurements are still taken today, the changesin glacier-front position of the other glaciers can only be determined by using ground and air photographs in addition to maps,to compare them with the recent position of the glacier tongues observed during fieldwork (Winkler 1994).

As the annual measurements of the glacier-front position change have been published (Rekstad 1901, Fa=gri 1934b, l.iestel 1963) and were thus accessible to the present author, it waspossible to reconstruct the glacier- front variations for all 15 Jostedalsbreen outlets (partly) measured duringthe twentieth century.In order to determine annual oscillations as well as the general trend of the glacier-front,two types of diagrams were construc- ted. The first shows the annual glacier-front variations (the relative positional change from one year to the next;

see Fig.8),and the second type the cumulative glacier- front variations (Le.the total change of glacier position to a fixed point set aszero representing the glacier-front position for the firstyearthe position was measured (see e.g. Fig.7 and Winkler 1994 for the complete seriesof both types of diagram ).If two or three marks existed at some glaciers and differentvalues for different sections of the glacier tongues were indicated in the annual reports, the mean was used to calculate the annual glacier-front variations. This procedure is commonly applied to glacier-frontmeasurements worldwide and,as exper i- enceduring manydecadesof glaciological research has shown, the loss of accuracyis negligable due to compen- sation by the glacier tongue itself in subsequent years (e.g.a stronger retreat on one side of a glacier-front can

StefanWink/er 35

easily be compensated in the followingyear by a stronger retreat on the other side; dramatic changes in glacier tongue morphology are rare and are described in the remarks added to every annual report).In a few cases, photographs taken at the time help to decide whether there were important changes in glacier tongue morphology.

The glacier-frontvariations of the Jostedalsbreen outlets cannot be presented here withoutnoting that mass balance data are far more suitable forcharacterisation of the interaction between climate and glaciers than data of glaciers-front variations. Among the many limiting fac- torswhich have to be controlled in the int erpret at ion of glacier-front variation data, the response time is one of the most important (Kuhn 1981,1989, Winkler 1994).In addition,glaciermorphology and relief have to be taken int o account. Mass balance data from Jostedalsbreen as a continuous record are, however, only available for Nigardsbreen. Even if these glaciological data help to increase our knowledge about the impact of individual climaticelements,thereis no more useful way to study the interaction between climate and glaciers during the whole of the twentieth century other than by using the available glacier-frontvariation data.However,in addition to their careful interpretation, consideration also needs to be given to the response time as well as to the determination of 't rends' for whole region (see e.g.

Patzelt (1985) for the Swissand AustrianAlps); correlation with meteorological parameters is possible and the results are promising (see below).In addition, the different reactions of glaciers to the same climatic fluctuations are also of great value for investigating climate-glacier interactions.

Glacier-front variations of Jostedals- breen outlets during the twentieth century

In western Norway,nearly all glaciers reached their maximum'Little Ice Age' position in the middle of the eighte- enth century (e.g. Eide 1955,Grove 1988).For the outlet glaciersof Jostedalsbreen, this 'Little Ice Age'maximum position was also the Neoglacial one, Le. the greatest advance following final deglaciation and the Erdalen Event (Nesje et al. 1991). At Jostedalsbreen, this term 'Litt le Ice Age' is far less problematic than,for example, in the European Alps where unequivocal evidence ofseve- ral major Holocene advances and Medieval glacier expan- sion episodes means that this term cannot really be used there, and consequently there is a need for alternative terms (Winkler 1994). Although there is little historical evidence for the 'Little Ice Age' maximum around Josted alsbreen (except ions being Brenndalsbreen and Nigardsbreen - see e.g.Rekstad 1901, 1904, Eide 1955), Iichenometric and morphological studies confirm the assumption that around or just a few years prior to AD

(4)

36 5refanWinkler GU-BULL431, 1996

1750, Jostedalsbreenoutl et glaciersreached their maximum positio ns(see Bickerto n & Matthews(1993)on the timing of the maximum 'Lit tle IceAge' positi ons using Iichenometry; see Winkler (1994) for some doubt s on their maximum dates for Lodals- and Steg holt breen).

Howe ver,differencesin glacier reaction time to climatic fluctuat ion s as a functi on of indivi du al response times revealedby the measuredglacier-frontvariati onsduring the twentieth century should not be neglected.Instead oftransferringNigardsbreen'sdate of maximum posit ion toallother glaciers(as wasdone by Mot tershead&Collin (1976)with theirIichenometric studi esatTunsbergdals- breen), themost probab lepatte rn of'Lit tleIceAge'maximum posit ions isthat of glaciersreaching that posit ion by differentdatesdependingontheir response timeand other glaciologicalfactors.

In the secondhalf of the eighte ent hcentu ry theglacier retreatwas very slow and commonly interruptedby read- vancesand periods of stationary glacier fronts(w inter advances!),evidence for whichisgiven bythe numerous and comp lex terminal moraine ridgescloseto the outermost end moraine (Andersen &Sollid 1971, Erikstad &

Soll id 1986, Bickerton &Matthews 1992, 1993, Winkler 1994).During thewhole of the nineteenth century there was repeated alternation between periods of glacier retreat (in part,strong retreat periods) and readvances and periods of stationary glacier-fronts,respectively ,leading to an overall retreatfar behind the outermost morai- nes.The last two decades of the nineteenth centurywere alsocharacterised by retreating glaciers,although there is evidence of minor readvances at some glaciers.

Boyabreen recorded an advance somewhere between 1880and1888 based on a comparison of historicalpho- tographs;and indications of advance at theend of the last decade of the nineteenth century have been repo r- tedfor Supphellebreen and a few other glaciers(Rekstad 1900, 1901, 1904, 0yen 1906). Just at the time when annualmeasurem ent s began,a more impo rtantreadvan- ce startedatmostoftheout let glaciers ofJostedalsbreen.

Unfortunately ,du e toti m e spans of severalyears between front position measurement sinthefirst yearsofregu-

larsurvey,theyearwhen the readvance act uallystartedis not always known; andalsothe advancedistan cecannot be calculat edat theseparticular glaciers.

Therange of known starting points ofthe first twenti- ethcenturyreadvanceis1902- 1905(Table1).Historical photograph s taken by J.Rekstad (Fotoarkiv NGU, Trondheim) give aclear visual impression of that readvance (see below),accomp anied by additional indi cati- ons of awidespreadincrease alsoinicevolu me onsmal- lerglacierson and around the Jostedalsbreen ice-capnot covered by the annualmeasurements.Even though not all glaciersparticipated in this readvance,theoverallevi- dence clearly shows that the readvancewasnotapheno- menon restricted to a few glaciers buta readvance on a regional scale.The differences in glacier-front variat ions between the Jostedalsbreen outlet scaneasily be traced back to ind ividual response times as there isa distinct pattern of smaller and steeper glaciers showingremarka- ble advance distances (e.g. Bergsetbreen, Boyabreen, Melkevollbreen)comparedto lon ger and lesssteepglaci- ers which hardly showedthe overall advanceduring that period, wit h onlysingle yearsofadvance or noadvan ce at all (e.g. Fabergstolsbreen,Lodalsbreen,Nigardsbreen, Stegholtbreen,Tunsbergdalsbreen; see also appendix).

A seriesof three photographs of Bergsetbreen takenin 1899,1903and1907gives a clearideaof the glacial dyna- mics of that readvance(Figs.2-5).As there was a frontal ret reat (112,5 m) betwee n 1899 and 1903, the photographs reveal a clear increasein icevolume of the ice- fall/upper part of the glaciertongue,whereas the low er glacier tongueissti ll veryflat, wit hou tcrevasses and showing all the typi cal features of a retreating glacier.

Betwee n1903and 1907,60.6m of glacieradvance were measured at the glacier-front and thephotograph taken in 1907 exhib its an increased ice volume of the low er tongue accompaniedbya deeply crevassedfrontal secti- on (see also Fig.6).The readvance,which was no more thana transfer of ice massesfrom the ice-cap to thelow er glaciertongue,is the only possible glacialdynamic interpretation.This expla ins why the longerglacierswith their longresponsetimes(Table2) did not participate in the

glacier 1.readva nce distance 2.readvance distance

Austerdalsbreen (1905)-09* 62.5m no data no data

Bergsetbreen (1903)-11* 145.7m 1922 - 31 127.0m

Bodalsbreen (1900) -12* 47.7 m 1922-30 62.0m

Boyabreen (1904)-11* 133.5m 1921 - 31 164.0m

Brennda lsbreen 1905 - 13 96.0 m 1922-32 73.0m

Brigsdalsbreen 1904 - 10 78.2m 1921 - 31 66.5m

Kjenndalsbreen (1907)-09* 44.0m 1922 - 32 190.5m

Melkevollbreen 1902 -11 177.0m 1922-29 208.1m

Supphellbreen (1898)-12* 95.5m 1921 - 31 91.8 m

VesleSupphellbreen 1902-11 148.0m 1922- 31 115.0m Table 1. Advance dista nces fo r thetwoadvance sin the firsthalf Fabergstolsbreen (1907)-10* 24.9m 1922-30 112.0m of the twent iet h centu ry.Theout-

Lodalsbreen no advance 1925 - 29 -5.0 m+ let glaciers are classifi ed accor-

Nigardsbreen (1903)-11* -8.9 m+ 1924-30 48.0m ding totheir respo nsetime s(* -

Stegholt breen (1903)-10* 26.3m 1922 - 28 36.5m

startof advanceunknown;+- net

Tunsbergdalsbreen 1910/11 no advance retreat during the advance peri-

od).

(5)

NGU-BULL431, 1996

,

5tefanWinkler 37

Fig.2.Bergsetbreen,photograph edby J.Reksrad 14.09.7899(Fotoarkiv NGU Trondheim

r e

^30.725]).

Fig.4.Berqsetbreen,photograph edby J.Rekstad20.08.1907(Fotoarkiv NGU Trondheim

re

^30.731]).

Table2.Responsetimesandclassification of theoutletglaciers measured (* -classification mainlybased onoscillations during the first half of thetwentiet hcent ury).

group 1:

respo nsetime2 years:

Suphellebreen responsetime3 years:

Bodalsbreen,Boyabreen,Brenndalsbreen, Brigdalsbreen,Kjenndalsbreen,Melkevollbreen response time4years:

Austerd alsbreen*,Bergsetbreen,VesleSupphellebreen group2:

responsetimec.25 years:

Fabergsto lsbreen,Lo d alsb reen,Niga rdsbreen, Stegho ltbreen

group3:

responsetimec.35years:

Tunsbergdalsbreen

Fig.3,Bergsetbreen,photographed byLkekstad17.09.7903(Fotoarkiv NGU Trondheim

re

^30.128]).

Fig.5.Bergsetbreen,photographedby S.Winkler20.08.7995.

minor readvance that culminated after a few years between 1909 and 1913.

After culmination of the firstreadvance,a short period of ret reat was registered at the glaciers up to 1921/22 (single years of advance were reg ist ered at Supphelle- breen, Stegholtbreen, Nigardsbreen and Fabergst0ls- breen).Then,asecond readvance started,again at nearly all outlets of Jostedalsbreen. As during the first readvance,there weredifferences in reaction between the glaciers covered by the annual measurements resulting from differences in individual response times. Again, aspect and local climatic factorsseem to have been of negligible influence compared to the re spon setime of theglac ie r s and their morphology.Between 1929(or 1928) and 1932 thissecond readvance ended.

There is no distinctive general difference in duration

(6)

38 Steian Winkler NGU-BUll431.1996

Fig.6.Berqsetbreen,cumulative glacier-frontvariations(data,for Figs.6- 10,takenfromRekstad,Fcegri,tiesteland NVEj .

interruption can be traced. It seems, simply, that the secondreadvancehad a slightly higher mag nit ude at the longer glaciers.Apart from the direct climaticim pact on mass balanceandglacier-front variation,thismight have beencaused bythelongerresponse timeof these glaciers which was of such a lengt h that the first climatic im pact (fi rst readvance)added to the seco ndreadvance.

Attheshorter glaciertongues,the maximum positionof the second readvance was mostly locatedafew tensof met res proximalto the ice-front positionof the firstone, because of the retreat during the period 1910- 1920 exceeding the advance distance of thefirst readvance.

Onlyat Beyabreen did the glacier-front position of 1931 reachthe position of 1911;and on some parts ofthe glacier foreland causing the formationof amult i-ridged end moraine system (FCEg ri1934a,Winkler 1994).Supphelle- breen was the only permanently observed Jostedals- breen out let reaching farth er down -valley during the second readvance and consequent ly overriding its terminal moraine built uparound 1911.Thisbehaviourmust beviewed in the light ofSupphellebreen'sspecial glacier morphology,sinceitis a regeneratedglacier.

Without exception, the mid-twentiet h century was characterised by the extensi veretreat of alloutletglaciers from Jostedalsbreen. Annual retreat distancesof more than 100 m were measured where special conditions were met (narrow glaciertongu esdescend ing through steepvalleyheads or gorges:Brenndalsbreen,Kjen ndals- breenor Melkevollbreen;proglacial lakes:Brigsdalsbreen, Beda lsbreenand Nigardsbreen).However,asthis excessi- veretreatoccurredat all theJostedalsbreenout letglaci- ers, regional or local infl uences of glacier morpholog y, proglacial lakesor relief have to be classified asfairly important secondary fact orsaugmenti ng a ret reat caused by the climatic conditions during that period.

Unfortunately, at many glaciers annual observations / measurements had ceased during this period of str ong glacierretreat.

Theretreatof the mid-twentiethcentury is veryremar- kable for its magnitude.In only a fewdecades,theglacier tong ues melt ed further backthan they had done in the preceding 180 yearsup to 1930follow ing the 'Litt le Ice

30 20

Bergsetbreen

10 1900 -400 -300 -200 -100

+50 m

advance

and advance distance between the two readvances of the first half of the twentieth century.The second read- vanceispresumed to havelasted one ortwo yearslon ger than the first one,but becausethe starti ng point ofthe firstreadvance isnot precisely delimited at mostof the Jostedalsbreenoutlets, adefinitivestate ment on its duration isnot possible. Wheth er the first orthe secondread- vanceled to agreater dista nce of glacier-front posit io n chang ebetwe enthe glaciers isuncertain.At the majority of thesmallerglaciers, the totaladvancedista nces of the first and second readvance were quite sim ilar.

A remarkable differenceoccurred at the larger glacier tongues of Fabergst01 sbreen and Niga rdsb reen with a much st ronger advance dur ing the second readvance than during thefirst one.Apartfromthis,the same major diff erences in reacti on betweenthe long er andthe shorter glacier tonguesas during the first readvance were observed.At the longer,as at the smalle rglacier tongues, the secondreadvancewas commonly interrupted forone or two years,butno distin ctivepattern of years of overall

.100 m

_.nee

Brigsdalsbreen

eetre e r

Brigsdalsb reen

Fig.7.Brigsdalsbreen,cum ula ti ve glacier-fro nt variati ons. Fig.8.Briqsdalsbreen,annualglacier-fron t variations.

(7)

NGU-BULL 431,1996

- 0 retreat -400

-BOO

·1200

-1600

-2000

-2400

Nigardsbreen

- 0 m

retreat

·200

-400

-600

StefanWinkler

Tunsbergdalsbreen

39

1900 10 20 30 40 50 60 70 80 90

Fig.9.Nigardsbreen,cumulativeglacier-fro ntvariations.

Age' maximum.Although this strong retreat was observed at every glacier, even without continuous annual measurements it becomes clear that there were differences between the longer and the shorter glacier tongues.

The strongest annual retreat of the small glaciers,marked by high retreat distances up to 100 m and more,took pla- ce during the early or late 1940's (e.g.Brenndalsbreen, Kjenndalsbreen and Brig sdalsbreen,which show ed high retreat distancesas early as 1935 due to the influence of a calving glacier-front on the proglacial Brigsdal svatnet;

see Figs.7&8).At the longer glacier tongues,the maximum retreat was recorded considerably laterduring the lat e 1960's (Fabergst0lsbreen, Stegholtbreen, Lodals- breen - prio r to observation s which stopped around 1970). Nigardsbreenwas also exposedto maximumretre- at during the 1960's,buta second maximum rate ofret reat was observed during the late1940's,probably caused by the special glacial dynamicsof the floating calving front of the glacieron Nigardsvatnet(Fig.9).The only glacier cont inuously measuredunt il 1970withoutany defi- nite retreat maximu m isTun sbergd alsbreen (Fig.10). A simple, but logical explanati o n for that observation is that the long response time in thiscaseexceeded the response times ofthe otherlargerglacier tonguesby 10

Table3. Recent advance distancesforselected outletglaciers (a- annualmeasurem ents(NVE);b- fieldobservat ionsandair photograph/topog raphical map analysis: startofadvanceunkn own).

glacier period advance

(a)

Brigdalsbreen 1955 - 95 498m

(1987 - 95) 339m

Nigardsbreen 1988 - 95 124m

(b)

Berg setb reen 1966 - 95 c.250m

Brenned alsbreen 1966 - 95 250 - 300 m

(1984-95) c.200 m

Kjenned alsbreen 1984-95 c.300m

Boda lsbreen 1984-95 150-200 m

Fig.10.Tunsberqd alsbreen,cum ulativeglacier-frontvaria tions.

years or slightly more.As the difference inresponse times between shorter and longer glacier tongues(see below) is very clearly marked by the dating of the maximum retreat in the middle of the twentieth century, the maximum retreat ofTunsbergdalsbreenis likely to have occurred later during the 1970'swhen therewere no continuous measurements.

In the 1960's,at the time when the mass balance studi- esstarted on Nigardsbreen,the retreatgradually came to an end, first on the smaller glaciers.Again in conjunction with the specific response times, the increasing ice mass of the glacier plateau of Jostedal sbreencreated stationary glacier-fronts.These were followed by advances which actually culminated in an impressive glacier advance exceedingboth readvance sof the early twentieth century;and locally nearlyat taining annual advance distances not recorded since the maximum 'Little Ice Age'advance (Figs. 11 - 14, Table 3).As the annual measurements at most of the smaller glacier tongues had ceased during the middle of the twentieth century,the present author comparedtopographic mapsand aerial photographsin order to obtain abrief ind icat ion of the magnitude of that advance on several small glacier tongues. Apart from Brigsd alsbreenwith itscontinuous record showinga con- siderable advance (210 m advance 1992 - 1995;339 m advance1987- 1995 and 498 m advance 1955-1995)and leadi ng to its classification as 'advance' (and not just

'readvance'), similar st rong advances actually occur at

Bergsetbreen , Brenndalsbreen, Bedalsbreen and Kjenndalsbreen (Table3).In addition ,othersmaller glaciers clearly increased in ice volume, asfor example at Boyabreen (t he upper, act ive glacier tongue again reaching the valley bottom and the low er,inact ive'glacier tongue'/ice-snow-debris-complex itself developing int o an active glacier-like form),Supphellebreen (lower part increasing in volume) and Melkevollbreen (advancing consider ablydown the narrowgorge ofits valley head).

At the tim ethatthe smaller glaciershad already stopped their ret reat, the larger glacierssti ll experienced a st ro ng retr eat that did not slow down until the late

(8)

40 Stefan Winkler

Fig.11.Briq sdalsbreen,photog rap hedbyS.Winkler11.08.7990

1970's.However,sincec.1975,most of the larger glacier tongues underwent a period of more or less stationary glacier-fronts(e.g.Nigardsbreen)or at least a comparable slow retreat (Fabergst0 Isbreen, Stegholtbreen). Just recently,the advancewhich has been observed for along time at the shorter glaciertongues also becamevisible at the longer glacier tongues.However,increasing ice velocity,positive mass balancesduring recent years,increa- singdissection of glaciertongues by crevasses,andincre- asingice volume inthe upperandlow er glaciertongues (e.g. at Nigardsbreen)are thought to be clearindicators of an increasing magni tude of an advance in the near

NGU-BULL431.1996

future. In 1995, almost all glacier ton gues around Jostedalsbreen werein a state ofadvance,for both the shorter and the longerglaciertongues,and fortheannu- ally measured glaciers aswell as tho se not included in annualobservat io ns.

Comparing the glacier-front variations of the outlet glaciersof Josteda lsbreen duringthe twe ntiet hcentury, significanttypesofglacier-frontoscillationbehaviourcan be diffe rentiated andaclassificationof the glaciertong - ues has become necessary. Allsmalleroutlet glacierswith short glacier tongues andsteepicefalls show verysim ilar glacier-frontvariations duringthe first half ofthetwent i- eth century.Onlyat Brigsdalsbreenisthere acon tinuous record up to the present,but thiscan betaken asarepre- sentative model oftheoscillat ionsofsmalleroutlet glaci- ersduring the last few decades.Among these glaciers there isno indicatio n of any great influenceof aspector local climaticfactors.This isremarkable because ofregis- tered climatic diff erencesbetwee n the valleyson different sidesof Jostedalsbree n (0 st rem et al. 1976,Nesje 1989,Aune 1993,Ferlan d 1993).However,asa result of the uniform accumulation area of the glacier plateau, non-parallel deviationsfrom the meansof sum mertem- perat ure and winter precipit ation recorded at different locations around Jostedalsbreen seem to be far less important than glacier size, morpholog y and result ing responsetime.

It ispossible to determine the response time of the Jostedalsbreen outlets bycorrelation analysisoftheir glacier-frontvariation s,aswell asofcombined glacier-front variati ons and meteor ological data (see below). The resultsfor the smaller glaciers give com mon response time sin therang e 2- 4 years (Table2).For most ofthe glaciers,a calculated response time of 3 years provides thebest fit tothewinterprecip itationand summer tem- perat ure records,aresultalreadyshow nby Nesje(1989) and Nesj eet al.(1995) atBrigsdalsbreen. The larger glacier tongues show much longer response times in the

Fig.12.Briqsdalsbreen, photogrophedbyS.Winkler 30.08.1995.

(9)

NGU-BULL 431, 1996

Fig.13.Kjenndalsbreen, photographed by S.Winkler 09.09.1991.

order of 21 -26 years delay compared to Supphellebreen (i.e., 23- 28 years to the climatic impulse; see Winkler 1994).Owing to smaller correlation coefficients, the exact response times are to some extent uncertain, but a

c.

25 years response time is assumed to be most realistic for the longer Jostedalsbreen outlets, as previously demonstrated by Nesje (1989) for Fflbergstolsbreen and Stegholtbreen. Among the larger glaciers, there are only small differences in response times (analogous to the smaller glaciers). Stegholtbreen and Fflbergstolsbreen show identical response times and very similar glacier- front variations(cont rast ing with quite different Iicheno- metric dates of the 'Litt le Ice Age' maximum suggested by Bickerton & Matthews 1993). In the case of Tunsbergdalsbreen a longer response time of 35 - 40 years has been recorded,due to its size as the largest outlet of Jostedalsbreen.However, even if the glacier-front variations of Lodalsbreen and Tunsbergdalsbreen are quite similar(see above),a correlation of the whole glacier-front variation record of the twentieth century leads to the classification of Lodalsbreen in the group of larger glaciers (e.g. Nigardsbreen orStegholtbreen).Although they show no important deviation from the general trend, proglaciallakes causing calvingglacier-fronts have at leastto be taken into account when the glacier-front variations of such glaciers are interpreted.The first maximum of retreat during the 1940's at Nigardsbreen not recorded at the other longer glacier tongues might be

StefanWinkler 41

Fig.14.Kjenndalsbreen,photographed by S. Winkler 31.08.1995.

the result of an influence ofthe proglaciallakes,as well as the early maximum retreat of Brigsdalsbreen and its early (and comparably strong)oscillating advance during the 1950's,1960'sand 1970's and can be traced back to the calving glacier-front.At Austerdalsbreen, the break in the record makes it difficult to give a confirmed response time.Particularly for the first readvance of the twentieth century, this glacier behaved like the shorter glacier tongues. However, considering the recent advance, Austerdalsbreen has reacted more like a longer glacier tongue with an advance just starting a few years ago (1991).Individual glacier morphology and the extraordi- nary cover of supraglacial debris (causing the formation of a dead-ice/moraine-complexin front of the active part of the glacier tongue)are assumed to be responsible for this different reaction.

The twentieth century glacier-front variations of Jostedalsbreen have not been synchronous with all other glacier variations in southern Norway. In particular,the climatically more continental mountain glaciers in Jotunheimen behaved in a different way.There was only one readvance interrupting the overall retreat during the twentieth century,around or prior to 1920, Le.at the time the Jostedalsbreen outlet glaciers underwent a phase of retreat in between the two readvances of the first half of the twentieth century (seeWinkler 1994 for more details).

Although the dating of that readvance is uncertain, due to a break in the annual observations in Jotunheimen,

(10)

42 Stetan Winkler GU-BULL 431.1996

/ stororeen

/

: Hardangerjekulef

I I I I I I II I I

~_ ~'

I I I I I I I I I I

~

I \

\\ I

•

• I 0

1980

o I •

1975 ,',..

i^l

/

./

".....

---

.--'\.^,

o I •

1970

, I I I

1965 - 6

-4

·8

·2 +2 +4 +6 +8 +12

+10

+

14,--- - - - -- -- - - -- - -- -- -- - --

m(w.e.) Nigardsbreen

!Molbreen

!

!!

! /

j

/' /

!

/

j

I

iI

"

Fig.15.Cumu lat ive net balanc es~f6qlaciersinsouth~rnNorw~y(modified after 0srrem,Dale-Selviq&Tandberg1988;data fromNVE).The'verticalprofile'is from west(top) toeast(bottom), I.e.the qlacierstha tIncreasedInvolumeare locatedin the west;and thosethat decreased, inthe east.

there isno correlation of glacier-frontoscillation s during the firstthreedecadesof the twentiet h century between Jostedalsbreen and Jotunheimen. However,the retr eat period during the middl e of the twentieth cent ury was also recordedin Jotu nh eimen,but theannualretr eatdis- tances were relatively smaller than at Jostedalsbreen.

Mass balance stu di es at Storb reen,Hellstugubreen and Grasubreen (0st rem et al. 1991) showed an overall ice massdecrease over recent decades (only the very last balance years were slightly positive in some parts of Jotunheimen;Fig.15).Incont rast to Joste dalsbreen,there isno actual advanceinJotunheimen,wit h the exception ofmore or lessstati onary glacierfronts duringthelast few years(Winkler 1994),especially in the western part, e.g. Stygged alsbreen/Hurrungane and StorbreenN is- dalen.

Meteoro log ical facto rs causin g glacier-front oscillatio ns

For anumber ofreason s,glacier-front variationsare less suitable for the characterisation and analysis of glacier- climate interaction thanmassbalance data as thenum- ber of possib le non-climatolog ical/-meteo rological factors infl uencing a 'pure climatic signal' of the glacier behaviour is larger.Inaddi ti on,theindivi d ual im pact of diffe rent meteorologi cal factors and the sign ificance of specialmeteorologicalparametersis often quite accura- telydete rmined by studying both annualmass balance and accumu latio n and ablatio n (winter and summer balances).Energy balance studies(i.e. determinati on of the contributionof singleablation factors such as solar radi ati on orconde nsation)givevaluablesupport tomass

(11)

NGU-BULL43 1,1 996

balance investigations(e.g. Moser et al. 1986).As this glaciological research is unfortunately restricted to only a few glaciertongues(at Jostedalsbreen,with few excepti- ons only at Nigardsbreen),and as there are no mass/energy balance data for Jostedalsbreen dating back more than 30-35 years,the identification ,characterisation and analysisof climatic/meteorolog icalfactorsforcing twenti - eth century glacier-frontvariation of Jostedalsbreen are limited. Althoughsome promising attemptshave been made so far (see below ),justa brief account of the results foundin the literature and of studies by the present author is given here.

One of the first attempts to calculate the mass balance variations of Jostedalsbreenwas made by Rogstad(1941, 1952) using runoff and meteorological data.He calculated higher summer temperatures (at Oppstryn) measured during his observation period than were necessary to melt winter snow accumulation recorded in the same time interval (1901 -1950;a calculated mean equilibrium summer temperature for that period 0.85°( lower than the recorded one:11.76°().In more detail,he registered an oscillation between short periods of volume increase (posit ive mass balances)and volume decrease(negat ive mass balances), corresponding to the two readvances and the intervening period of retreat during the first half of the twentieth century(bearing in mind glacier response times).Rogstad'sresults reveal that both summer temperature(below average)and winterprecip itation (above average) were favourable for a glacier volume increase and resulting readvance. However, it is not possible to decide whether one of these two forcing meteorological elements had a significantly higher influence(only for the period 1925-40do summer temperatures seem to be clearly more important).

Although Rogstad'sresults have been criticised becau-

Srefan Winkler 43

se of a less precise marking of accumulation/drainage areas for the gaugingstation and neglecting the effect of calving at Brigsdalsbreen (0 st rem & Karlen 1963), the trend demonstrated by his results has some validity (see below). Apart from Rogstad,other authors made com- ments on the forcing meteorological factors of glacier- front variations at Jostedalsbreen (e.g. Rekstad 1900, Fcegri 1934a). Unfortunately, there is no consensus in their results and conclusions (w int er precipitation and summer temperatures are alternately proposed as the most important factor).

In a recent study, Nesje (1989) concluded that both major meteorological elements,winter precipitation and summer temperature,are responsible for the two readvances of the first half of the twentieth centurybyintensi- fying theirind ividual impact (Le.winter precipitation above and summer temperatures below the average means) without the possibility of distinguishing whether one of these elements is more important than the other.A similar effect of the two major meteorological elements intensifying their impact by synchronous fluctuations, but now in an unfavourable manner for Jostedal sbreen,is shown by Nesje for the middle of the twentieth century when summer temperatures above and winter precipitation below average causeda fast retreat of the glacier tongues. Decreasing summer temperatures since the 1960'sand 1970'sin conjunction with increasing winter precipitation caused the glaciers to grow and advance again. Over the whole period studied,Nesje found a higher correlation between summer temperatures and glacier-front variations than between winter precipitationand position changes of the glacier tongues. However, the best correlation,not surprisingly, was found by using a calculated combined parameter of these two elements (see Nesje 1989 for more details) .

1900 19 10 1920 1930 19_0 1950 1960 1970 1980 1990

Oppstry n

• '500

.1000

.500

-500

-1000

-1500

t910 1930 1940 1~

Josteda l

Fig.16.Cum ulative deviation s of sum mer temp erature (T6-T9) from the meanoftheobservationperiod (12.2QC)forOppstryn(data fromDNM/).

Fig.17.Cumulativedeviations ofwin ter precip itation(N11-NS) from the mean of the observationperiod (709.9mm)forJostedal(data from DNM/).

(12)

44 SteianWinkler GU-BULL 431,1996

Table6.Correlationcoefficientsforwinter precip itation(Nl -N4)at Oppstryn and thewinter/net balanceatNigardsbreen.

Table 5.Correlationcoefficientsfor summertemperature - summer/net balance ofNigardsbreen(*-short observat ionrecord;

T=mean month lytemperature,5 = May,etc.).

period winter netbalance

1962 - 66 r =+0.99 r =+0.95

19 67 -71 r=+0.87 r=+0.87

1972 -76 r=+0.8 1 r=+0.75

1977 -81 r= +0.40 r=+0.51

1982 -86 r = +0.79 r=+0.82

1987-90 r = +0.75 r=+0.59

tongues) had a strong infl uence and caused an even more rapid ret reat;andwhen numerous regressionswere tried, calculate d negative mass balances and ice mass redu ctionwerenever of themagnitudesobservedduring that period .After 1970,winter precip itation clearlybeca- me the mostimportant meteorological element,especially during the recent years of very large winter snow accum ulation.

Howe ver,these modelling exercisesare very complex.

If,for example,correlation sover shorter periodsare calculated, there are significant differences. Taking Brigsdalsbreen as an examp le, glacier-front variations corr elat ed with winter precipitation (January - April)

-0.70 (T6 - T8) net balance

-0.89 (T5-T8) 5535 - Luster

5573 -Soqndal-Selsenq 5870-Oppstryn

-0.79 (T6 - T8) -0.82(T6-T8) -0.97(T6 - T9) -0.87 (T5-T8) -0.82 (T6 - T8) -0.23 (T6 - Tl0) -0.75(T6-T8) summer

Table4.Correlat io n coeffici ents for winte rprecipitation -winter/net balance at Nigardsbreen (N = mean of monthly precipitation, 1 = January,etc.),

stat ion 5056 5535 5545 5573 5584 5870

winterbalance

Nl 0-N5 0.71 0.55 0.70 0.77 0.8 1 0.72

N11- N5 0.74 0.71 0.74 0.83 0.81 0.8 1

N11-N4 0.74 0.70 0.74 0.83 0.81 0.79

N12-N5 0.77 0.80 0.70 0.82 0.79 0.82

N12 -N4 0.76 0.81 0.72 0.83 0.80 0.80

Nl-N4 0.73 0.89 0.81 0.85 0.8 1 0.82

net balance

Nl0 -N5 0.57 0.44 0.53 0.67 0.65 0.60

Nl1-N5 0.64 0.69 0.59 0.73 0.71 0.70

Nll-N4 0.63 0.67 0.58 0.71 0.69 0.67

N12-N5 0.66 0.67 0.67 0.70 0.67 0.68

N12· N4 0.62 0.68 0.68 0.69 0.65 0.65

Nl- N4 0.64 0.74 0.64 0.71 0.67 0.70

NlO-N9 0.59 0.55 0.54 0.71 0.67 0.62

N5-Nl0 0.14 0.31 0.19 0.30 0.26 0.06

5056 -Bergen-Fredriksberg 5545- Joste dal

5585-Brigsdal

station:

[5056]Berge n [5523JFannaraken (5529) Sognefje ll*

[5535]Luster [5543]Bj er kehauq"

[5584] Fj<erland [5870]Opp stryn

In the opinion of the author,the better correlation for summer temperaturesismostlikely to have been caused by the particularlyrapidretreat during the middle of the twentieth centurywhen summer temperature deviat ions from the average mean were much higher than comparable deviations of winter precipitation from the average mean,andwinter precip itation had less infl uence onthe mass balances(Figs. 16&17).Another factor to be taken into account is the common observation that glaciersare not assumed to be in a state of equilibriumwith the present-day climatic condit ions (Kuhn 1984).Ow ing to the delayed response of glaciers to major and lon g-term changes of the climaticenvironment,the state of the glacier (i.e.,its positionand volume)prior to an advance or retreat period alwaysshows an influence on the actual mass balance. It seems logi cal that the out let s of Jostedalsbreenatabou t 1930still contained some excess mass left over from their'Little Ice Age'advance that was not sufficiently reduced during the nineteenth century due to frequent small readvances and stillstands of the glacier-fronts. As a consequence, the glacier tong ues were larger and located farther down-valley than would be expected for the climate around 1920/ 30. Thu s,the rapid retreat during the middle of thetwentiethcentu ry could have been infl uenced by these non-eq uilibrium glacier-frontpositionsand highly negat ivemassbalances prevailing as a result of the large natural ablatio n area (w hen the glacierswere more easilyaff ected by higher summer temperatures).It should be noted that such con- siderati ons alsoapplyin other glacierregio ns such asthe Austrian Alps, where the retreat observed during the middle of the twentieth century was far more rapid than would be expectedfrom the summer temperature deviation s.

In arecent stud y,Winkler (1994) used primarily mass balance data from Nigardsbreen and other glaciers in southern Norway to characterise and analyse the im pact of separate meteorological elements and glacierparame- ters(Tables 4 - 6).Althoughthe results cannot be presented here in detail,theknown sim ilarit yofindivi dual meteorological stations around Jostedalsbreen was used to correlate meteorological parameters such as summer temperature,winter precipitationand glacier-frontvariation data (response time and changes in the system of relevant meteorological and non-meteorological factors were taken into consideration).

Although the results carry a degree of uncertainty,the present author'sresults support the conclusionsof Nesje and Rogstad that it is not possibleto diffe rent iatebetwe - en winter precipitation and summer temperatures in order to determine which meteorological factor had the most significant impact on the glaciers. However, the rapid ret reat during the middle of the twentieth century was prima rilythe resultof highersummer temperatures.

In addition, best-fit regression reconstructions of the mass balance of Nigardsbreen using meteorological parameters/stationsand mass balance data suggest that the differentstates of the glacier(larger and lower glacier

(13)

NGU-BULL431,1996

recorded at Oppstryn (empirically found to be the best fitting parameter), the correlation coefficients for 10-year periods varied between r

=

+0.8176 (p <0.05)and r

= -

0.0136 (p>0.05). This influence of individual meteorological factors showing strong variations over short inter- vals leads to the somewhat disappointing statement that it will be very difficult (or perhaps impossible) to reconstruct mass balance series. In some regions there has been progress, for example in Austria, as a result of the clear relationship between summer snow fall frequency/summertemperature and glacier mass balance as undoubtedly the most important meteorological factors(Fliri 1964,1990).However, at Jostedalsbreen with its maritime climatic environment and consequent com- plicated interactions between air temperature and precipitation, there still remains much work to do before relia- ble reconstructions can be made.

Implications and conclusions

(1) The glacier-front measurements of Jostedalsbreen have offered a unique opportunity to study the reaction of several adjacent glacier tongues during the twentieth century.The following patterns of glacier-front variations have occurred: (a)In the first half of the twentieth century, two readvances occurred at most of the smaller Jostedalsbreen outlets. Both readvances lasted 6 - 10 years and the advance distances were mostly similar.The readvances culminated around 1910/11 and 1930/31.

They were interrupted by a short period of retreat during the period 1910-1920.At the larger glacier tongues,two readvances were either only recorded as advances of one year or no advance tendencies were recorded at all. (b) After 1930, a period of rapid retreat was measured at all Jostedalsbreen outlets. At the smaller glaciers,the years of greatest retreat distancesoccurred during the 1940's for the larger glaciers during the 1960's.(c) The retreat phase ceased at the smaller glaciers in the 1960's (for some glaciers as early as the 1950's).After a period of stationary glacier-fronts an advance phase occurred which lasted until the present; and this has increased considerably in magnitude in recent years.At the larger glaciers the retreat slowed down during the 1980's. After several years of nearly stationary glacier-fronts, even the larger glaciers also started to advance at the beginning of the 1990'sand presumably they will reach the rate of advance characteristic of the smaller glaciers;this is indicated by an impressive increase in the ice mass of the lower tongues,anincreasing number of crevasses and an increasing ice velocity.

(2) There are important differences in the reaction of glacier tongues of different size throughout the twentieth century. As a result of the influence of glacier size and morphology, an d re sulting individual response time, smaller glaciers were involved in two readvances in the first half of the centu ry and underwent the most pro- nounced retreat period earlier than the larger glaciers.

StefanWink/er 45

The same'delayed' reaction of larger glaciers is evident with the recent advance. Correlation of the glacier-front variationsallows a classification of three different groups of glaciers based on their differences in response time:(a) Similar small and steep glaciers (e.g., Brigsdalsbreen, Bergsetbreen) with response times of 2 - 4 years. (b) Larger glacier tongues reacting sluggishly, with response times of about 25 years (e.g. Nigardsbreen). (c) Tunsbergdalsbreen,as the largest Jostedalsbreen outlet, can be classified as an exception with a response time of at least 35 years. However,other glaciological and c1ima- tological factors such as aspect show no important influence on the glacier-frontoscillationbehaviour, as glaciers of one group indicated above show the same patterns of front position change independent of which side of the glacier plateau they are located.

(3) Apart from some restrictions on accuracy, glacier- front variations can be correlated with meteorological data to determine the influence of various meteorological elements and parameters.The first two readvances of the twentieth century were caused by higher winter precipitation and lower summer temperatures intensifying each other, with no possibilityto decide which meteorological factor is the leading one.The fast retreat during the middle of the 20th-century was caused mainly by high summer temperatures. The actual glacier advance is clearly a result of an unusually large winter snow accumulation during recent years.

(4) In combination with glaciological data (mass and energy balances), qlacier-frontvariations can be used to reconstruct the mass balance changes of Jostedalsbreen during the twentieth century.Unfortunately, the results of the reconstructions were not successful owing to the fact that the present relationship between mass balance and meteorological parameters is not a useful basisfor regression calculations. However, it is possible that im provement s in such reconstructions can be made in the future so that the available data on glacier-front variations and climatic fluctuations can be used to a greater extent than has hithertobeen possible.

Acknowledgements

The results presented here are partof an investigationcarriedout for a doctoralthesisbythe authorat the GeographicalInst itu te,Universityof WOrzburg. The author wishes to express his thanks to J.L.Sollid, J.O.Hagen and the entire staff of Avdeling for Naturgeog rafi (Universitetet i Oslo), A.Nesje and N.Rye (Universitetet i Bergen), E.5orensen (Norges geolog iske undersokelse), E.Sonstegaard and A.R.Aa (Sogn og Fj ordane distrikthogskule), S.Kristiansen (DNMIl, J.Bogen andH.Elvehoy(NVE) and all otherpeople inNorwaywhooffe- red and gave thei r help duri ng theworkwith my thesis.R.A.5hakesby (University ofWales,Swansea)andD.Ro berts(Norgesgeologiske under- se kelse)kindly improved the Englishof this art icle.