Journal ifGlaciology, Vat. 43, No. 145, 1997

Surge of Bering Glacier and Bagley Ice Field, : an up­ date to August 1995 and an interpretation of brittle­ deformation patterns

2 UTE CHRISTINA H E RZFE LD,i, 2 H E LMUT M AYER iGeomatlzemalik, Faclzbereiclz GeowissensclzaJten, Universilii/7i-ier, D-54286 Trier, Germany 21nstitute oJ Arctic and AljJine Research, Un iversity oJColorado, Bouldel; Colorado 80309, US.A.

ABSTRACT. In the summers of 1993, 1994 and 1995, video a nd Gl obal Posilioning Sys tem location data and 35 mm photographs were co ll ected in a series of sys tem ati c sur­ vey fli ghts undertaken over the Bering Glacier and Bagley Ice Field system (Al aska ) in an effort to cha racterize surge-crevasse pa tterns and surge propagation. During survey fli ghts in late August 1995, we obse rved that the 1993- 94 Bering Glacier surge was con­ tinuing a nd still ex panding, a ffecting new a reas farther up in Bagley Ice Fi eld. New cre­ \'asse field s, similar in pattern to the first surge crevasses we had observed inJune 1993 below KhitrO\' Hills a nd in other isolated a reas of centra l Bering Gl acier a nd inJuly 1994 near the head of Be ring Gl acier (near the juncti ol1 of Be ring Gl acier a nd Bagley Ice Fi eld, in both upper Bering Gl acier a nd Bagley Ice Fi eld ), were opening in eastern Bagley Ice Fiel d and in the "Stell er" side of Bagley Ice Field . T he type of crevasses seen in the new fi elds suggested that the surge was propagating into these areas. By analysis and interpretati on of the brittle-deformatio n pattern. appa rent in the crevasse p a LL ern s, so me asp ects of the past kinem ati c fr amework of the urge can be deduced. This approach may lead to a more genera l classificati on of ice-surface struclU res a nd to their linkage to ongomg processes.

INTRODUCTION lakes. These la kes are se parated from the Gulf of Alaska by a narrow stretch of coast, in one place by a sand spit onl y 200 m wide (Seal Rive r Spit). O ver-riding of the spit by the Bf'ring G lacier, Alas ka's longest glac ier, sta rted to surge ice ri'ont wo uld tra nsform Bering G lacier into a tide-water sometime bctween 13 M ay 1993, when R. Krimmel glacier and induce a rapid and dras ti c change in the emire obscrved no surge crevasses and 13- 14 June 1993, when A. coastal environment, causing rapid back- stepping erosion of Post, C. Lingle and U. C. H erzfeld noti ced fresh surge crev­ the yo un g over-ridden sediments, a nd possibl y reconnecting asses. As new surge crevasses, we defin e fi eld of fr es hly the dee p troug h under present-day Bering Glacier to its off­ opened crevasses with cha racteri sticall y sha rp and un­ shore continua ti o n (cr. Meier a nd Post (1987) fo r processes eroded edges, indicati ve of recenl extensive, substantial and im'oh-ecl ). The difference in dyna mics and mass equilibrium fas t ice deform ation. betwee n Bering a nd Stell er Glaciers is documented by Due to its size and complex it y, the Bering Glacier/Bag­ folded media l m oraines which have record ed surges during Ic y Ice Fi eld system comprises several glaciers named indep­ endently during the discovery of Alaska a nd Canada, the las t 500 years (M olnia and Post, 1995). During a surge, ineluding Bagley lee Fi eld, Columbus G lacier, Quintino the Bering- Stell er Glacier ' media l moraine is pushed to­ Sell a Glacier, J efTeri es Glacier and Stell er Glacier (Molnia ward s the Stell er Glacier side. vVith decreasing velocity of and Post, 1995; er. Fig. I). C olumbus Gl acier originales wes t Bering Glacier after the surge, the Stell er Glacier ice fl ow of the King-Logan Massifa ndjoins Quintino Sell a Glacier increases in rela ti ve strength a nd the medi al mora i ne is re­ to form Bagley Ice Fi eld , which foll ows a la rge (7.5- 12 km located towards the Bering Glacier side. The difIerence in wide) troug h in the C huga ch Mountains to the wes t. Tana the rate of cha nge in velocity (sudden increase in sp eed at Glacier drains northward from the ice fi eld, while Bering the beginning of the surge vs slow decrease of speed at the Gl acier, the largest glacier fl owing Oul of Bagley Ice Field, end ) is refl ected in the a ymmetry of the folded m orain es. drains to the so ulhwes t through a trough 13 km wide. Bcring It is not clear why Bering Glacier surges and Steller Glacier Glacier a nd Bagley Ice Field may be considered as one does not. In this contex t, the questi on arises whether Bagley glacier, the Bering- Bagley system, whi ch has a length of Ice Fi eld surges a nd how far the surge-affected ice extend. close to 200 km. During the entire surging peri od, Bering G lacier has Whereas Bering Gl acier surges qu as i-periodicall y about been in a state of retreat from its H olocene maximum pos­ every 20+ years, the neighbo uring Stell er Glacier, which is ition (Molnia a nd Post, 1995). The rapid surge ad vances are also a n a rm of Bagley Ice Field, does not surge (Post, 1972). superi mposed o n the overall retreat. Bering a nd Stell er Gl aciers form a large piedmont lobe, The progress of the recent surge has been documented in about 50 km wide, whi ch terminates in a chain of pro glacial a num ber of news notes and corres pondences. The surge

427 Downloaded from https://www.cambridge.org/core. 29 Sep 2021 at 03:36:22, subject to the Cambridge Core terms of use. JournalofGlaciology

BERING GLACIER AND BAGLEY ICE FIELD, ALASKA

Fig. 1. Flight tracks of27 July 1994, and 18 and 20 August 1995, on a topographic maj) of the Bering Glacier/Bagle)) l ee Field area, Alaska ( modified after United States Geological Survey, 1983).

reached the ice front on 23 August 1993 ( ~Io lni a, 1993), caus­ stress field ) and frozen m o\"ement (cre\"asse patterns a nd ing rapid advance of the ce ntral part of th e terminus st rain), a nd thus to reconstruct the stress fiel d from the (1500 m in 17 days) but rapid retreat of the terminus in the observed patterns, as exemplified in the secti on on "struc­ Tasali sh Arm (25 m d I) and in a neighbouring a rea, both tural geology in ice". The structural ana lysis then aid s in in th e western ice front. Terminus advance in 1994 was mon­ conclusions on th e propagation of th e surge. Geophysical itored by Fl eisher a nd others (1995). By 1994, the surface of ques tions addressed with this approach a re: how does SLll"­ Bering Glacier had become totall y crevassed and the surge face roughness relate to ice velocit y, to the fl ow fiel d, a nd to was propagating into Bagley Ice Fi eld . stress and surge propagation? How does this ultimately in­ Throughout the Bering Glacier surge, the dyna mics flu ence calving? were compl ex with fas t motion propagating into different In addition, surface classificati on by visual in spection, a reas at different times; and with the areas of greatest speed and structural interpreta ti on as demonstrated in this essay changing as the surge progressed. \ Ve attribute this to the for a few examples, provide the basis for the disc rimination probabl y complex subglacial to pography and to the enor­ of classes of crevasse types for a n automated geostatistical mous size of the Bering Glacier system, encompassing ice-surface elass ifi cation based on remote-sensing data several fl ows and catchment areas. The ad,·ance varied (Herzfeld, 1993, in press; H erzfeld and others, 1996a, b). regionall y, i.e. from one glac ier region to another. In the 1994 and 1995 field ex periments, we coll ected vi­ In contras t to the situati on in the well-prepared study of deo data a nd Global Positioning System (GPS) data fr om the Vari egated Glacier surge (Kamb and others, 1985; R ay­ small aircraft. G PS data are used for geolocation of the vi­ mond, 1987), only few geophys ical data a re a,·ail abl e for the deo da ta a nd of photog raphs, and all ow compari son of vi­ Bering Glacier surge. H ydrographic a nd glaciologic da ta of deo data with satellite Syntheti c Aperture R adar (SAR) the glacier at a quiescent stage are generall y lacking, with a data from ERS-l (European Spaee Agency). Video data few exceptions (cf. Fleisher and others, 1995; Molnia a nd are collected to study surface structures a t high resolu tion Post, 1995; and references therein). Subglacial topography and to do so automaticall y, a nd are preferable to photo­ influences the locati on of crevasse fields but unfortunately graphs for digital evaluation. not many so undings of Bering Glacier are m·ail able (per­ Geostatistical surface cl assificati on provides a robust sonal communicati on from D. Traba nt). In part, the size of technique for the analysis of complex movement. Interfero­ the Bering Gl acier-Bagley Ice Field system has prohibited metry is not applicable to the surge, because the ice move­ comprehensive ground-based geophysical study. This fact, ment leads to a lack of correlati on between SAR images which impedes "classical" geophys ical im·esti gations, consti­ (Fatl and a nd Li ngle, 1994). From the perspective of provid­ tutes both a chall enge and an opponunit y for investigations ing ground truth for the a nalysis of SAR data, the Bering based on remote-sensing data. Glacier surge has presented a unique opportunity to sample The basic idea of our analyses is to exploit the relation­ and characteri ze the surface of catas trophicall y accelerat­ ships between movement in time (surge propagatio n and ing ice a nd rapid changes in surface roughness.

428 Downloaded from https://www.cambridge.org/core. 29 Sep 2021 at 03:36:22, subject to the Cambridge Core terms of use. Herz;feld and Jlayer: Surge qfB erillg Glacier and Bagley lee Field, Alaska

On 27 August 199+, a j 6kulhlaup (catastrophic water outburs t) sta rted; this even t was generally considered to mark the end of the surge (Molni a, 1994·; Fleisher and others, 1995). During a week of exceptiona ll y clear weather in the ra iny 1995 ummer season, wc undertook two ni ghts, during 'which photography a nd video data were co-regis­ tered with GPS data, to investigate the current (18 and 20 August 1995) state and extent of the surge. The visual analysis in this essay is documented by the sha rper 35 mm photographs, taken on the same nights as the videos and also G PS registered. \ Ve now use these obser vati ons to ad­ dress the foll owing ques tions:

(I) Does Bagley Ice Field surge?

(2) Where in the Bagley Ice Field- Bering Glacier- Stell er Fig. 3. .Yew surge cre/'asses cross -m l old lopographica l£y in ­ Glacier system is the boundary between surging and duced crevasses qfJ SPill' 6290 on 8640 ridge al 60 ~ 25'. 'V, non-surging ice? 141"40' r 1 ~' 18 Augllsl1995.

(3) Does a fl ood outburst genera ll y mark the end of a surge? at the begin ni ng o f the surge (cr. fi gs 2, 3 and 4 in Lingle and others, 1993) a nd is typical of a surge front propagating in to (+) How do surges propagate? ice pre\'iously not affec ted by the surge. Fres h en-echelon These questi ons relate to m ore general ques ti ons formu­ surge crevasses as existing in ce ntral Bering Glacier in 1993 lated by R aym ond (1987): a re depicted in Fi g ure 4.

(a ) Wha t is the mechanism orrast moti on during a surge?

(b) What initiates and terminates the fas t m oti on?

The Va riegated Glacier study (Kamb a nd others, 1985) prO\' ided importa I1l data needed to answer these ques ti ons but the surge mechanisms a re still not fully understood. Ber­ ing Glacier prO\'ides an example of an obser ved surge in a much la rger glacier in a elimatic setting simila r to that of Variegated Glacier.

ADVANCE OF SURGE FRONT

By the summer of 1995, the surge had continued to propag­ ate fa rther into Bagley Ice Field. New surge crevasse fi elds Fig. 4. Fresh surge crevasseJield il1 cenlral Bering Glacier on (Figs 2 a nd 3) occurred as far east as 1410 10' \ V. In the cre­ 24 J une 1993. Compare 10 Figure 2 and Figures 9 an d 10. vasse field opposite cast Ba rk ley Ridge (near 60 0 30' ~ , 1+1 50' W ) in Baglcy Ice Field, cre\'asses opened in a direc­ Figure 3 shows the surge cl'C \'asse fi eld located fa rther ti on different fr om that of pre-ex isting ones, creating se racs cast (upstream in Bagley Ice Ficld/Colum bus G lacier) in August a t a bo ut N, at the foot of the with squa re, pentagonal or hexagonal tops (Fi g. 2). This re­ 1995, 60°25' 141 040' \ \' sembles the situati on in centra l Bering Glacier inJune 1993 rock spur that conta ins point 6290 on thc USGS I: 250 000 topographic m ap "Bering Glacier, A laska" and lead s off the ridge labeled "8640" on the m ap (we call this o ne "86

STAGES OF SURGE EXEMPLIFIED BY CREVASSE PATTERNS

In our analysis, we use the following terminology: (a ) beg­ inning su rge stage: earl y summer 1993, few crevasse fields; (b) mature surge stage: after surge reached terminus in 1993 and summer 1994; (c) late-surge tage: after onset of the water outburst during August 1994, continuing through August 1995. Bagley Ice Field in 1995 was reminiscent of Bering Glacier in 1993 a nd of the Bering Glacier/Bagley Ice Field area in 1994 - a smooth ice stream into which the urge wasj ust propagating, as indicated by disc rete surge-crevasse fields several tens of kilometres apart. Our observations in the three seasons revealed that the surge propagated up­ Fig. 6. Survey area, central Bering Clacie1: The high water glacier and down-glacier from a starting point in the lower level in the crevasses is rypical oJ the late stages of a surge; central Bering Glacier (below Khitrov Hills; cr. Lingle a nd Augusl1995. others, 1993, fi g. 5), and that crevasse patterns followed t yp­ ical successions, indicative of the cha nge in the stress a nd velocity field s, as the surge eontinued. The movement is lit­ and parts of Bagley Ice Fi eld was higher in 1995 than in el-all y frozen in ice: a record of the movement is contained in 1994. This indicates that there are several distinct water sys­ the crevasse patterns and the patterns are characteristic of tems in the glacier. In o ur flight of 18 August 1995, we the surge activities. This suggests that it is possible to char­ noticed upwelling in several places in the a rea of the jokul­ acteri ze the succession of surge stages and related crevasse hlaup. According to the local pilots, who frequently fl ew patterns, and to use them in a classification of the ice sur­ over the glacier and observed its actiyity, water had contin­ face. Such a classification may be applied to reconstruct the ued to drain from this a rea after August 1994 but at a ra te stress and velocity fi eld s of past surge stages. lower tha n the initia l rate. Observation and a na lysis ofstrain patterns as evident in the crevasse patterns facilitates th e deduction of some as­ pec ts of the kinematic framework in which these crevasses formed. The ice surface in Figure 6 shows the signature of seve ral surge-deformatio n phases: shear patterns and exten­ sional patterns, overprinting an undul ating ice surface. At least three processes can be di stinguished: I. Shear, indic­ ated by offsets along two main crevasse directions at oblique angle; 2. Extension, indicated by den ely spaced crevasses; and 3. Extension, indicated by widely spaced large crevasses filled with water and fallen seracs. Several strain fi elds adja­ cent to each other can be distinguished in the crevasse pat­ terns of Fi gure 7 taken in the survey area in central Bering Glacier. In the foreground of the photograph, there a re: l. ~~.....- ---- Block structures; 2. Cross-cutting crevasses; and 3. Wavy crevasses a nd water-filled gaps form ed by extension. In the Fig. 5. Later stage ifsurge in Bagley Ice Field, with old and background, we see boundari es delineating different strain new crevasses and wale1; near]uniper Ridge; 18 August 1995. patterns. Fig ure 8 shows the location where the surge started in Figure 5 shows an area that was first affected by the surge in the summer of 1994 (located in Bagley Ice Field nearJuniper Ridge, in the background). The previous year's crevasses have taken on a wavy shape a nd are, in part, fi !led with clear blue water. Newer (1995) crevasses trend across the glacier at about 45°to the 1994 crevasses. New crevasses intersect the snow cover, whereas patches of melting snow adhere to the ridges between old crevasses. High water levels in a glacier are typical of the mature stage of a su rge. They indicate that the crevasses are not, or no longe r, connected to the bottom. In the lower part of the fast-flowing glacier, water conduits are apparently shut off regionally. No turbid water is ex pelled from the bed to the surface nor is the clear surface meltwater effectively drained. Even more so than in 1994, in the summer of 1995 large areas of central Bering Glacier appeared to be flooded Fig. 7. Survey area, central Bering Clacie1: Old wary crevasses by water (Fi g. 6). D espite the occurrence of the large out­ resulting from superposition of several diformatiol1 phases. burst in late August 1994, th e water level in Bering Glacier High water level; August 1995. 430 Downloaded from https://www.cambridge.org/core. 29 Sep 2021 at 03:36:22, subject to the Cambridge Core terms of use. Her

Fig. 8. Area where the sll1gefirst started ill 1993, below Kh itrov Fig. 10. Fresh en -echelon crevasses in eastern Bagle) Ice Field Hills, as it ajJJJeared ill Augusl1995. (near JeJ!eries Gap) indicating the contempor(1)1 jJmjJag ­ ation Vthe slllge into this area; 18 August 1995.

June 1993, below Khitrov Hills (cf. fi gs 5 and 6 in Lingle cries Glacier gap (141° 45' W ), gives an excell ent example of and others, 1993). It is impossibl e to di stinguish individual en-echelon structures. These are the urst manifes tations of a crevasse lines or surge-deformati on phases. IVIost se racs cha nge in stress-field orientati on. The older direction is in­ have fallen o\'er, leaving isla nds of standing or leaning se r­ dicated by crevasses that trend in the view direc ti on. In the acs, arranged in an approxi m ately square pattern, with a foregroulld, they a rc covered by snow. The yo unger direc­ maze of sera cs in the foreground. ti on is at about 30 to the older direc ti o n (crevasses depicted by thin sharp lines ). Only rciati ve ages can be determined by this method. Some measure of age difference can be STRUCTURAL GEOLOGY IN ICE estimated fr om the differences with respect to snow cover a nd eroded or degraded edges but this is hard to qua ntify. Crevasse patterns on the glacier surface reveal stra in states. The stress of the second generati on is broken and defl ected Strain state is defin ed as a particul ar state of deformati on at the old lines of weakness, forming the en-echelon pa t­ res ulting from the process of deform ati on. Furthermore, terns. This is exactl y what is see n in the oceani c crust when crevasses reveal strain histori es by their cross-cutting and transform faults rc-ori ent them elves pa ra ll el to the new O\'e rprinting relati onships. In a kinemati c framework, these spreading directi on when a change of plate moti on occurs strain patterns allow the reconstructi on of stress fields and a long m id-ocean ic ridges (e.g. ~I ac d o n a ld , 1982), their changes. \ Ve also observed fl'esh cre\'assing on the "St ell er" side of Fi gure 9 shows old (1993 ol-ja nd 1994) crevasses with a Bagley Ice Fi eld (Fig. 11), Figure 11 is, in principle, an typi cal pattern of two approx i m ately orthogona l directi ons example of the a me situati on as Fig ure 10, showing a com­ (two-directi o na l patte rn ), indicative of a later stage of the posite strain res ulting fr om two deformati on events that sa me process that created the strain pattern in Figure 2. occurred at different times but with a different relative spa­ The two-directi onal pattern dominated the su rface of ci ng of the crevasse systems, which intersect at a 40° a ng le. ce ntral Bering Glacier inJuly and August 1994. This indi­ The stress secms to have been propagated along an old crev­ cates th at the ice mass had been a fTen ed, in the course of asse line for a di sta nce a lmost equal to the spacing of the old the surge up to that time, by a t least two di stinct deforma­ crevasses. A new cre\'asse is then created, connecting two ti onal e\'e nts charac teri zed by different stress-field ori enta­ adj acent old ones; this process leads to fairly regul ar zig­ ti ons. zag lines of fr ac ture. The feathery en-echelon pattern of the Figure 10, taken in eastern Bagley Ice Field near th e J e f1~ pre\'ious example is rarel y seen in this image.

Fig. 9. Old (1994) two -directional crevasses on B ering Fig. 11. Fresh sll1ge crevassing superimposed on older crevasses Glacier; August 1995. on tlte Steller Glacier side V Bagley Ice Field; 20 August 1995. 431 Downloaded from https://www.cambridge.org/core. 29 Sep 2021 at 03:36:22, subject to the Cambridge Core terms of use. J ournaloJGlaciology

Our preliminary analyses indicate that certain crevasse BOUNDARIES OF SURGE-AFFECTED ICE IN patterns are cha racteristic of certain surge phases a nd are BAGLEY ICE FIELD preserved in the ice. Contiguous a reas of specific patterns lend themselves to (automated ) surface classification (Herz­ Bering Glacier is pan of the Bering Glacier-Bagley Ice fc ld, in press ), thus permitting the reconstruction of the Field system regarding glacial dynamics. Although it has area I extent of surge phases a nd the kinemati c evolution of long been known that Bering Glacier surges quas i-period­ the surge. ically and that Steller Glacier does not (Post, 1972), it was Deform ati on patterns of ice surfaces may be used to not known until the 1993- 95 surge whether Bagley Ice Field study forces a nd processes in glaciers or ice sheets analogous is affected by surging, a nd if so, which pa rts of it. In 1994, the to those acti ve in the Earth's lithosphere, and have been an surge had propagated just into Bagley Ice Field at the out­ flow of Bering Glacier and some di sta nce eastward (below obj ecti ve of structural- geological a nalysis in a few instances , to about 142° W ). (e.g. Allen and others, 1960; M eier, 1960; Hambrey and Observations during our survey fli ght on 18 August 1995 Milnes, 1977). In their paper entitled "Structural geology of revealed that the easternmost field of surge-related cre­ an Alpine glacier (Griesgletscher, Valais, Switzerl and )", vasses was located near 141 °10' W, as m entioned above. The H ambrey a nd Milnes (1977) produced a structural map of occurrence of new crevasses in Bagley Ice Field in the late Griesgletscher with pole diagram s and analyses of strain surge stage may have res ulted from an upflow transmission (foli ation, crevasse traces in glacier ice, en-echelon zones, of stress reli ef and kinemati c effects from the surge, ex­ boudins and m ylonite zo nes ). Those authors fo cused on duc­ plained by th e following mechani sm: the mass transfer from tile deform ati on in the deeper level s of an Alpine glacier ex­ the anomalous ice build-up (as explai ned in the model of posed in the ablati on zone. The proccsses involved Fowler (1987)) typical of the onse t of the surge causes a def­ correspond to those occ urring at a deep crustallevel during icit of mass in neighbouring areas, which in turn induces orogenetic deformation. further mass transfer and velocity increase up-glacier from In this paper we arc concerned with the brittle deforma­ the starting point of the surge. The defi cit of mass leads to a ti on apparent in surfacc crevasse patterns of a surging decrease in upwa rd or back-holding force, which induces glacier. The surge-type crevasse patterns of Bering Glacier tensile stress in flow direction res ulting in new crevassing. correspond to extensional faulting in upper erustal le\'els. By a continuation of this mechanism, the surge "propagates The elongated blocks separated by parallel crevasses and up-glacier". 'Ve attribute the fresh crevasse fi eld s n ear the en- echelon crevasses are reminiscent of faults in yo ung 141 °10' 'V to this effect, which is part of the surge. In this oceanic crust, such as those in the abyssal hill terrain on sense, th e field of surge crevasses fa rthest east marks the the fl anks of the slow-spreading Mid-Atlantic Ridge eastern limit of the extent of th e surge (at time of obser­ (Tucholke a nd Lin, 1994; H erzfeld and Higginson, 1996) vation). and in their corresponding subaerial surfaces as they occur To es tablish the western limit of the surge-affec ted ice, in Iceland, for instance. A good example is the Alma nnagj a we investigated western Bagley Ice Field and Steller Glacier, and its surroundings in Iceland. Brittle deformation pat­ a n area that is seldom visited, on 20 August 1995. The terns observed on Bering Gl aci er during th e surge not only boundary between the eastern "Bering" side and the western resemble brittle crust deform ation features visuall y, but are "Steller" side ofBagley Ice Field is the juncti on with Bering also characterized by the same types of parameters in auto­ Glacier. The new generation of crevasses depicted in Figure mated geostati stical surface cl assificati on, as used for an 11 (described in the previous section) is attributed to the example fr om the western flank of the Mid-Atla ntic Ridge same effect of up-glacier-pro pagated stress reli ef as the cre­ near 46° N (H erzfeld and Higginson, 1996). These obser­ vasse fi elds near 141° 10' \ V. Because Figure 11 was ta ken vati ons and compariso n with the work on Griesgletscher about 15 km west of the wes tern corner of th e Bagley Ice (Hambrey a nd Milnes, 1977) suggest a velocity a nalogy Field- Bering Glacier j uncti on, the western bounda ry of between different tectonic environments a nd types of surge-affected ice was located in the "Stell er" side of Bagley glacier dynamics: Ice Field in August 1995.

(1) lVlctamorphic deformations - ductile deformation at lowe r- crust temperature a nd high lithostatic pres­ OBSERVATIONS ON STELLER GLACIER sure- basal parts of an Alpine glacier (Griesgletsc her). Steller Gl acier itself did not show surge-related crevasses. As (2) Deformation of newly genera ted oceani c crust of Atlan­ a n exception, we noted minor fres h crevassing in two areas ti c (slow-spreading) type - upper-crust brittle defor­ in Stellcr Gl acier on 20 August 1995. Some fr es h crevasses in mation (mainly ex tensional) - disintegration of a large Steller Glacier close to the folded medial moraine between glacier during a surge (Bering Glacier). Bering and Stellcr Glaciers might have resulted from the push of the surging Bering Glacier transmitted across the Can this a nalogy be extended fur ther? For instance, fas t­ medial moraine. During a surge, the entire sys tem Bagley spreading-typ e oceanic-crust deformation may corres pond Ice Fi eld- Bering Glacier- Stell er Glacier is oul of equil­ to the dynamic deform ati on in fast-flowing ice streams and ibrium and not all of this is compensated for by the (rather glaciers. slow) di slocation of the folded moraine, as described by Post If the analogies proposed h ere can be confirmed, they (1 972). In another possible interpretation, this lower cre­ can be exploited, on the one hand, to learn more about crus­ vasse field may be attributed to the retreat ofTasalish Arm tal deformati on, since ice deformati on operates on shorter (interpretati on by A. Post (personal communicati on, J a n­ time-scales. The longe r record of dynamic deformati on in uar y 1997)). crustal rock , on the oth er hand, will provide us with new Minor fres h crevassing obse rved in upper Stell er Glaci er ways of understa nding glacier a nd ice-sheet dyna mics. a nd in the aforementioned location next to the Bering 432 Downloaded from https://www.cambridge.org/core. 29 Sep 2021 at 03:36:22, subject to the Cambridge Core terms of use. Her;Jeld and lHayer: Swge of Bering Glacier and Bagley Ire Field, Alaska

Glacier-Stell er Glacier moraine may a lso be the res ult of a theory, a branch of a lgebra ic topology. The catas trophe­ "pulse" of Stell er Glacier at the time. This interpretation classification th eo rem, proved in the mid-1960s by R. Thom derives from the obser vati ons of A. Post (personal commun­ (1972; cr. Poston and Stewart, 1978, p. 99 122), states th at a ny ication, J anu ary 1997) that Stell er Glacier "pulses" m ore catastrophe is desc ribed by onc of seven types of catastrophe frequentl y than Bering Gl acier surges, with the lates t manifolds. There are catastrophe manifolds that encompass "pulse" in 1993- 94 being most active in the Stell er Glacier a continuous transition from onc se t of parameters charac­ valley. In summary, there are three poss ible reasons for teri stic of a system that has only continuous properties to a fr es h crevassing in Stell er Glacier: (1) inducti on by the se t of p a rameters characteristi c of a sys tem th at switches Bering Glacier surge, (2) pulsing of Stell er Glacier and (3) states. In this sense, we hypothesize th at surge-t ype glaciers retreat of1asalish Arm. may be considered intermediate in th ei r dynami cs between "norma l" (non- surging ) glaciers and fast-fl owing ice streams (er. Fowler, 1987; R aymond, 1987). \Ve . howed that CONCLUSIONS the dyna mic deformation during surges is akin to brittle de­ formation in slow-spreading oceani c ridges. D e\·eloping this Observations of Bering Gl ac ier and Bagle y Icc Fi eld with analogy betwee n dynamic ice sys tems a nd dynami c crust video and GPS data coll ected fr om small aircraft a nd aeri al sys tems further, we suggest that defo rma ti on in continu­ photography aided in answering the qu estions intiall y ously fast-moving ice streams may be a nalogous to brittle posed: (I) Large pa rts of Bagley Ice Fi eld are affected by deforma ti on in fas t-spreading ocean-ridge environments. the surge. (2) The eastern boundary of surge-affected ice is located at least as fa r east as 141 °10' \V (near l\Ilount Hux­ ACKNOWLEDGEMENTS Ie y), the wes tern boundary about 15 km wes t of the Bering Glacier- Bagley Ice Field corn er (or possibl y farther west). Ma ny tha nks to cveryone who provided help and unforget­ (3) The surge propagated from a sta rting point in lower table memories, in particul a r to S. and G. RanneyofFishing central Bering Glacier both up-glacier and down-glacier. and Flying, COI·dm·a, Al aska, R. Carlson of Era Helicop­ (4) The up-glacier propagation of fr esh surge crevass ing tcrs, Anchorage, Alaska, to B. ~[o lni a for the opportunity conti nued for at least 12 months after a maj or water o utburst to participate in the USGS fi eld camp 1994·, to A. Post a nd which started on 27 August 1994 a nd continued with a D. Traba nt (or \·a lu abl e disc uss ions (all U. S. G eologica l Sur­ decreased di sc harge through Au gust 1995. \"Cy ), toJ. Roush (Uni vC'rsity of Alaska, Fairbanks) and K . The ice di splacement is probabl y initiated at the locatio n Lohuis (Albion College) for field assista nce in 199+, to K. of la rgest build-up of a nomalous ice foll owing the model of Steffen (Universit y of C olol-ado, Boulder) for use of a GPS Fowl er (1987); aft er this ice has been transported to the receiver in 1995, to H . D enksc herz (Univcrsitat Tri er) 10 r receiving area, the upstream press ure component induced drafting the map of Fi g ul-c I, to R. Leb. Hookc (Uni\"C rsity by the anomalous ice loading in the central part dccays, of Minnesota, Minneapoli s), D. R. ~fac Ay e a l (Uni\·e rsity of which releases waves of materi al for transport down-glacier Chi cago), C. Ray mond (U ni\"C rsit y of Washington, Seattle), from a reas farther upstream. This process continues a ft er and J. H ollin (Universit y of Colorado, Bo ulder) for helpful the main surge activity in th e mature stage of the surge. It co mments on the manusC'ripl. The proj ect has in part been is interes ting to note that in th e late surge stage the surge supportcd through NASA Office of Polar Programs grant fr ont is still advancing in both cast a nd wes t directions in NAG'V-3790, U. S. DOE grant DE-FG03-93 ER61689 Bagley Ice Fi eld, while at the same time onl y local surge (Principa l ImTs tigato r ~f. F. f\I eier) and Deutsche activity is obse n -e d in the centra l, heavil y creyassed p a rts Forschungsgemeinsc ha ft through a Heise nberg Fell owship tha t were affec ted by the surge during the mature stage. to UCH. Ana logies with brittle-deformati o n patterns from struc­ tura l geology are employed in crevasse-pattern class ific­ ati o n and automa ted In a geostati stical surface REFERENCES cl assificati on method (H erzfeld , in press ) in an effort to reconstruct stress fields during surge propagation and to ,\Ilc n. c. R ., \\' 13. Kamb, ;"1. I '~ l\kier a nd R. P. Sha rp. 1960. St ruCl ure o /" Ihe 10w("I" Blue Gl ac ier, Washing ton.,7 C(ol .. 68(6), 601 625. lea rn about the relationships between surface roughness, Hlt land, D. R . a nd C. S. Lingle. 199+. The surface "clocity lidd on BagJc y Icc­ \·el ocit y, stlTSS and st ra in during a surge. Not onl y may tech­ fidel. Alaska, before and du ring the 1993 9+ surge of He ring Glacier, from niques and insights from structura l geology be applied, ERS-I SAR iIll rrfcromclry. I Abstrac l.] EOS. 75 (H ), Supplement. 62. a na logies of deforma ti on types a rc obvious. Sl owl y cha n­ Fleisher, 1'..J. and 601hm. 1995. Correspondence. The s urging aeh-a nce o f Bering G lacier. ,\Ias ka. U.S. A .: a progress reporl. ] Clacial., 41(1 37), ging a lpine glaciers exhibit ductil e deform ati on patterns 207 213. si m i la r to orogeneti c metamorphic processes. Crevassi ng Fowler. A. C. 1987. A theory ofglacicr surges.] Geo/lh),5. Re.\., 92(13 91, 91 11 9120. of Bc ring Gl acier, during the surge, is reminisce nt of crusta l Hambrey, l\1..J. a nd A. G. ;"[ilnes. 1977. St ruc tural geology of an Alpinc glacier deformati on in slow-spreading oceanic ridges rather than in G ri csgJc-t schcr, \ ',tl ais, Switzerl a nd ). Er/agae (:ea/. I-Ielv., 70 (3 ), 667 68 ct. Herzfeld, U. G 1991 :\ method for scafl oor clas< ificatio n using direCt iona l fa st- spreading ones, a nd a simila r cl ass ifi ca ti on scheme ,·a riogra ms, dCl11 0mtraled fo r cla ta from the western Il a nk of the i\lid­ applies. Ex amples of applicati ons in a regime of brittle de­ ,\tla lllic ridge . .Ilolh. Ceol., 25(7), 90! 92+. formation (on the western fl ank of the Mid-Atlantic Ridge) Herzfeld, U. G In press. G CllS latistical imerpolati o n a mi d ass iflcat ion o f ha\ "C been gi\ "C n by H erzfeld and Higginso n (1996). remolc-sensing data from ice s llrl ~lces. 1nl.] Remote Sell . ~illg. Hcrzfcl d. C. G and C. i\. Higginson. 199(1. ,\ ulO m a lecl geostat isti cal sea- Glaciologists have considered surge glaciers a different fl oor classifica ti on princ ipl es, pa rameters, feCl lul'(' \ 'CClOI'';';, a lld spec ies from non-surging glaciers, because surge glaciers discrimina tion criteria. CO)))/)))t. (;eolci., 22 1,35 52. switch from onc dyna mical state into another. In the sense Hrrzfcld. U. G. H. !llaye r. C. 1\. Higginso n ancl ;"1. l\l a tassa. 1996a. (;co­ of a catastrophe model, this is not necessarily th e case. The stati sti cal approaches to inte rpolation and cl assificati o n o f" remotc-sen­ sing d ata li·om ice su rf'ICCS. III I;illlrlh CirCllIII/),J/ar s.)'III/Jolilllll 011 Remole study of sys tems where continuolls input leads to di seOlllin­ Sensillg if Ihe Polar Ellvironlllellts. 29 , I/Iril I ,I/,!1' 1996. lel'llg~)'. f) (,))lIIark. uous eYe illS (catastrophes ) is the obj ecti ve of catastrophc Noord ll'ij k, European Space Agency, 59 63. (ESA SP-391.).

433 Downloaded from https://www.cambridge.org/core. 29 Sep 2021 at 03:36:22, subject to the Cambridge Core terms of use. J ournal qfGlaciology

Herzfcld, U. c. , O. Zahncr, H. i\Iayer, C. A. Higginson a nd M . Stauber. i\Iolnia, B. 10 1993. M aj or glacier surge continues. EOS, 74 (45),521,524. 199610. I mage analys is by geos tati stical and neura l-ne twork mcthods i\lolnia, B. 1994. Modern surge of glac ier comes to an cnd. EOS, 75(47), 549. ~ applicati ons in glaciology. In Fourth CircumjJolar Sym.posium 011 Remote Molnia, B. 1O. and A. Post. 1995. Holocenc history of Bcring G lacier, Al aska: Sensil1g qf the Polar El111irol1mel1ts. 29 April - I ,I fay 1996. Lyngb)', Denmark. a prel udc to the 1993- 1994 su rge. Phys. Geogr., 16(2\ 87-11 7. Noord wijk, Eu ropean Space Agency, 87- 9\. (ESA SP-39\.). Pos t, A. 1972. Periodic surge ori gin of folded med ial moraines on Bering Kamb. B. alld 7 others. 1985. Glacier surge mec hanism: 1982 1983 surge of pied mont glacier, Al as ka. J GlaciaL. , 11 (62),219- 226. Variegated Glacier, Al aska. Science, 227 (-1686), 469-479. Poston, T. and 1. N. Stewart. 1978. Catastrophe theory alld its applications. Lon­ Lingle, C. S., A. Post, U. C. Herzfel d, B. F. ~Io lni a, R. M . Krimmcl and ].]. don, Pitma n Publishing Limited. Roush. 1993. Corres pondence. Bering Gl ac ier surge a nd iceberg-cal ving R ay mond, C. F. 1987. H uw clo glaciers surge? A re\·iew. J Geopl!Js. Res. , mecha nism at Vitus Lake, Al aska, U.s.A. J GlaciaL., 39(133), 722- 727. 92 (B9),9121 9134. ~l acdo n a l d, K. C. 1982. i\lid-occan ridges: fi ne scale tectonic, \"olca nic and Thom. R. 1972. Slabilite strueturelle et lIl07plwgbtese. New York , Benj amin. hydrotherma l processes within the pl ate bounda ry zone. Anllu. Rev. Tucholke, B. E. a nd]. Lin. 1994. A geological model for the structure of Earth Planet. Sei., 10, 155 190. ridge segments in slow spreading ocean crust. J GeojJh.. Ys. Res., 99(B6), ;\Ieier. M . F. 1960. i\ fode of now ofSaskatchell"an G l ac i e l ~ A lberta, Ca nada. 11,937- 11.958. u.s. Geo!. SlIrD. Pro[ PajJ. 35 \. Uni ted States Geological Surl"ey. 1983. Bering Glacier. Alaska. Fa irba nks, AK, ~" e i e r. M . F. a nd A. Pos t. 1987. Fast tidewater glaciers. J GeopI9's. Res., U. S. Geological Sun·ey. (Topographic map. sca lc I : 250,000.) 92 (B9), 9051- 9058.

MS received 29 March 1996 and accepted in revisedJorm 29 M arch 1997

434 Downloaded from https://www.cambridge.org/core. 29 Sep 2021 at 03:36:22, subject to the Cambridge Core terms of use.