Subglacial deformation of trees within overridden foreland strata, Bering Glacier, Alaska

www.elsevier.com/locate/geomorph

Geomorphology 75 (

Subglacial deformation of trees within overridden foreland strata,

Bering Glacier, Alaska

P. Jay Fleisher a,e,*, Matthew S. Lachniet b,e, Ernest H. Muller c,e, Palmer K. Bailey d,e

a Earth Sciences Department, SUNY-Oneonta, Oneonta, NY, 13820-4015, USAb Department of Geoscience University of Nevada, Las Vegas, NV 89154-4010, USAc Earth Sciences Department, Syracuse University, Syracuse, NY, 13210-2936, USA

d CRREL (retired), Anchor Point, AK, 99556-9702, USAe Prince William Sound Science Center, Cordova, AK, 99574, USA

Received 20 September 2003; received in revised form 26 January 2005; accepted 26 January 2005

Available online 20 October 2005

Abstract

The foreland stratigraphy overridden during recent Bering Glacier surges bears evidence of subglacial deformation. The pre-

existing, fine textured substrate (till and diamicton) experienced diminished strength because of saturation, thus resulting in shallow

mobilization and the formation of a new till of limited thickness. Glacial coupling with well drained sediment resulted in ploughing

that generated a diamicton that retains vestiges of outwash sorting and stratification.

The outwash sequence extending decimeters beneath the surface till contains four prominent sub-meter sand beds. Each sand

bed holds multiple small, fossil trees still rooted in underlying layers of gravel. Virtually all trees in the upper two sand beds are

deformed. Several are offset by centimeter to decimeter horizontal shears confined to thin, silt, and clay-rich zones at the base of

each sand bed. Trees that escaped shearing are warped and kinked. Deformed trees are present at depths that range from 15.76 to

5.31 m. beneath potential ice/substrate interface surfaces.

The most likely source of deforming stress in this foreland setting is related to glacial advance. The style and orientation of tree

deformation are consistent with the direction of ice movement. Therefore, the occurrence of deformed trees is attributed to stress

applied by overriding ice.

D 2005 Elsevier B.V. All rights reserved.

Keywords: Substrate deformation; Overriding glacial ice

1. Introduction

Bering Glacier, Alaska is known to have surged re-

peatedly in historic time (Post, 1972; Muller and Fle-

isher, 1995; Fleisher et al., 1995;Molnia and Post, 1995).

Each of the past two surges (1965–67 and 1993–95) were

0169-555X/$ - see front matter D 2005 Elsevier B.V. All rights reserved.

doi:10.1016/j.geomorph.2005.01.013

* Corresponding author. Earth Sciences Department, SUNY-

Oneonta, Oneonta, NY, 13820-4015, USA. Tel.: +1 607 436 3375;

fax: +1 607 436 3547.

E-mail address: [email protected] (P.J. Fleisher).

punctuated by jokulhlaups along the eastern sector ice

front. Observations during the 1993–95 surge, coupled

with stratigraphy exposed because of flood erosion of

bluffs indicate that pre-existing sediment was deformed

when overridden by ice. Although recognition of sub-

strate deformation is not new (Alley, 1989; Alley, 1993;

Benn, 1995; Boulton, 1979; Boulton and Hindmarsh,

1987; Clarke, 1987; Clarke et al., 1984, Hart et al.,

1990), seldom can the cause and effect be so clearly

and directly linked.

2006) 201–211

P.J. Fleisher et al. / Geomorphology 75 (2006) 201–211202

The deformed stratigraphic units are members of the

Weeping Peat Formation that have type sections in the

island bluffs along the eastern sector of the piedmont

lobe of the Bering Glacier (Fleisher et al., 1998) (Fig.

1). Till and diamicton at the top of the section served as

source material for a new till formed during 1993–95

surge. Young spruce trees and alder, killed by burial

within lake sand and outwash gravel while still in the

living position, were subsequently deformed at depth

when the entire stratigraphic section was overridden

by ice.

2. Methods

An archive of photographs taken at two- to five-

month intervals from four survey stations on foreland

islands documents changes in the ice front position, ice

thickness, and conditions of ground saturation and

deformation at the base of the advancing glacier during

the initial six months of the 1993–95 surge, (Fleisher et

al., 1995). Rates of advance were measured at six

ground survey stations during summer months in

1994 and 1995, until surging advance finally ceased

on the eastern sector in September 1995.

Fig. 1. Regional and eastern sector index map. Inset map illustrates the locat

Weeping Peat Island (WPI). Shown are ice front positions for the 1965–67

surge.

The stratigraphy exposed during post-surge retreat

was examined in several measured sections on Weeping

Peat Island. Buried sub-fossil trees partially exposed by

bluff erosion were completely exhumed to expose

trunks and root systems, and to establish autochthonous

occurrence. Global Positioning System coordinates

were recorded, trees and host sediment were photo-

graphed and sketched, and wood samples were

obtained for the analysis of growth rings and future

radiocarbon dating.

3. Shallow deformation of overridden substrate,

Weeping Peat Island

During the initial six months of the 1993–95 surge,

ice mounted island bluffs, advanced across the foreland

and ploughed into ice-contact, proglacial lakes at mea-

sured rates of 3 to 7 m/day (Fleisher et al., 1995). The

ground within a few meters of the ice front, consisting

of till, diamicton, and outwash, was sufficiently satu-

rated to completely lack shear strength and exhibit

physical properties similar to that of freshly poured

concrete. A few meters from the ice front, beyond

this peripheral zone of saturation, the same sediment

ions of study sites 1 (flutes and push moraine) and 2 (Waterworks) on

surge limit, retreat position in 1993, and maximum extent of 1993–95

Fig. 2. Rapid basal sliding across a saturated older till substrate. Note the exceptionally clean basal ice (9 cm knife for scale).

P.J. Fleisher et al. / Geomorphology 75 (2006) 201–211 203

remained dry, firm, and cohesive. Such conditions in-

dicate the widespread availability of infiltrating subgla-

cial water and the lack of substrate resistance to

deformation, as illustrated in Fig. 2.

Deposited on the overridden surface is a veneer of

light gray till, normally 0.5 m thick, containing rounded

pebbles held firm in a sandy matrix. Alder, freshly

crushed by overriding ice with leaves still attached,

are found flattened at the base of the till and incorpo-

rated within it. The surface of the till shows the com-

mon expression of flutes, 20–40 cm high and

decimeters in length, with conspicuous alignment par-

allel to the movement of overriding ice. This veneer of

till thickens laterally where it is continuous with a low-

relief, yet prominent, push moraine (Fig. 3).

Fig. 3. Till formed during 1993–95 surge and push moraine. Deformation till

1993–95 surge limit. Flattened alders overridden during the surge separate

The jokulhlaups of July 1994 produced an ice-

walled canyon in the glacier front that effectively

separated a large segment of the ice front on Weeping

Peat Island from further surge-related forces. Thus,

subglacial deformation on a portion of the island

ceased less than a year after it began (Fleisher et al.,

1999). The overriding ice here was relatively thin (~35

m calculated from photo station images) and active

movement was limited to approximately nine months

(Fleisher et al., 1995). Although duration of movement

was limited, the shear strength of the saturated sub-

strate was sufficiently reduced by infiltrating subglacial

water to allow the development of a deformation till.

This illustrates the susceptibility of pre-existing mate-

rial to deformation.

(0.5 m thick) is laterally continuous with the push moraine marking the

new deformation till from an older till below.

P.J. Fleisher et al. / Geomorphology 75 (2006) 201–211204

4. Weeping Peat Island stratigraphic setting

U. S. Geological Survey aerial photos show Bering

Glacier in full surge on September 2, 1966, when it

completely covered Weeping Peat Island. The till de-

posited during this surge (referred to as the bolder tillQ)lies directly beneath the new, 1993–95 till and generally

forms the upper-most stratigraphic unit beyond the

1993–95 surge limit. The bolder tillQ is gray, with an

average thickness of 1 m, and with very similar phys-

ical properties to the 1993–95 till. The older till grades

downward through a transitional contact with an under-

lying, very firm, gray, diamicton that ranges in thick-

ness from 1.5 to 2.2 m, has a sandy matrix and is

crudely sorted. The diamicton possesses vague stratifi-

cation, discontinuous lenses of clast-supported pebbles,

and distorted, wavy, coarse sand interbeds, all sugges-

tive of an origin related to deformation. A conspicu-

ously sharp contact separates the diamicton from an

underlying outwash sequence (35–40 m thick) that is

common to all bluffs on WPI. It consists of loosely held

interstratified, well-sorted, cross-bedded, coarse sand

and imbricate cobble gravel. Dispersed within the out-

wash are multiple, uniformly thick, coarse, tan sand

beds covering a thin peat horizon made conspicuous

by the occurrence of in situ, small rooted trees.

The most complete exposure of these units is in a

bluff referred to as the Waterworks site (Fig. 1). Here,

the upper 7.9 m of this 26 m high bluff consists of an

bolder tillQ that is assumed to have been deposited

during the 1965–67 surge. It is anomalously thick

compared to the sub-meter 1993–95 till and consists

of three sub-units, each of similar thickness (2.5–2.9

m). Each has uniform textural properties, yet slightly

different erosional resistance as expressed in breaks in

slope. Although thicker here than elsewhere, this till is

commonly very firm, light gray, and has a fine-to-

medium sandy matrix that holds rounded and striated

pebble clasts. Its lower contact is transitional to a 2.4 to

2.7 m thick diamicton, which is also very firm and

unsorted, yet contains decimeter zones of crudely strat-

ified gravel. A pebbly, sandy matrix holds 0.5 m thick

lenticular clusters of clast-supported, subangular and

subrounded gravel, and isolated sub-meter size

boulders. A distinct, wavy, and irregular lower contact

unconformably separates this diamicton from an under-

lying 35–40 m outwash unit that dominates most bluff

exposures on the eastern foreland (Fig. 4). The lower

contact is not exposed within the Waterworks bluff.

As is typical elsewhere as well, the outwash gravel

contains several sub-meter thick sand beds. Four sand

beds are prominent in the Waterworks bluff. At the base

of each are millimeter-scale, peat layers covered by 1 to

4 cm of silt and clay. Sand bed thickness ranges be-

tween 0.33 and 0.67 m and each bed grades upward

from fine-to-coarse sand. Although generally lacking

internal stratification, each sand bed contains faint

expressions of what appears to be deformed, remnant

trough cross bedding. Perched ground water commonly

saturates the base of the upper two sand layers causing

slow seeping drips of water from peat rootlets during

summer months, hence the name bWeeping Peat

IslandQ.

5. Nature and occurrence of deformed trees

Each of the four sand beds (A through D, from

oldest to youngest in Fig. 4) within the Waterworks

bluff is host to multiple alder (Alnus crispa, sinuata),

cottonwood (Populus balsamifera, trichocarpa), and

small spruce (Picea sitchensis) (23 growth rings or

less) that stand in situ, rooted in underlying gravel

(personal communication, Dorothy Peteet). Where not

lost to slope retreat, the trees extend upward through the

overlying blanket of sand and into the overlying gravel

bed above. Still standing upright in the living position,

the trees are held firmly within the host. A few that

extend completely through the overlying gravel are

truncated along the silt and clay zone at the base of

the next higher sand bed. All of six trees within the two

upper sand beds (C and D) show some form of defor-

mation. Most are sheared by lateral offset at the base of

the trunk, whereas others are kinked or bent within the

host sand bed. (Figs. 5–8). Shears are confined within a

2–3 cm thick layer of silt and clay that drapes the lower

trunks of all trees at the base of all sand beds.

Six sheared trees within sand bed C range in diam-

eter from 7 to 10 cm, whereas the largest deformed tree

(19 cm) is 1.93 m higher in sand bed D. All trees are

sheared where they passed through the 2–3 cm silt and

clay layer and the base of sand beds C and D, yet they

remain rooted in underlying gravel. The amount of

offset ranges from 4 cm for the largest tree to 10–14

cm for trees in sand bed C. The bark of one offset tree is

flattened onto the shear plane, as if smeared during

deformation. The direction of offset favors an azimuth

of 1508 (S 308 E). This conforms to the orientation of

surface flutes on the surface of Weeping Peat Island

(Fleisher et al., 2004).

Bent trunks are the second most common style of

deformation. Trunk curvature is greatest within the

lower portion of sand bed C, with convex curvature

also to the southeast (Fig. 8). They regain vertical

orientation as they rise through the sand bed. Also

Fig. 5. Styles of tree deformation. Three different styles of fossil-tree deformation are observed. Each represents different expressions of a common

stress condition in which forces with a strong downglacier component were exerted (arrows). Trees in the living position are rooted in gravel and

buried initially by sand, then gravel prior to being deformed. All trees remained rooted during deformation. Stippling represents basal silt and clay

within the sand. Tree #1 is offset from its root by lateral displacement that is confined to the basal silt and clay. Above the kink in tree #2, the trunk

remains upright, whereas the lower trunk is inclined. This implies differential movement of the sand across the gravel while the tree remained

rooted. Tree #3 shows a distinct curvature and tension fractures in the lower trunk where deforming forces were greatest. All trees are truncated by

offset where entering the overlying sand bed.

Fig. 4. Waterworks stratigraphic section. Deformed trees remain buried in sand beds A and B.

P.J. Fleisher et al. / Geomorphology 75 (2006) 201–211 205

Fig. 6. Trees rooted in gravel are laterally offset from the roots along a

plane of failure confined within a thin (2–4 cm) silt and clay layer

capping peat at the base of sand bed C. Direction of offset (shown by

arrows) is parallel to the direction of overriding ice movement. (a) A

cottonwood tree (7 cm diameter) rooted in gravel is offset approxi-

mately 10 cm. The displaced, lower trunk (above the arrow) is shown

in the position it occupied when excavated. The top of the trunk was

cut to obtain a wood sample. (b) A spruce tree, offset 12–14 cm,

shows lower trunk curvature and prominent horizontal tension frac-

ture. Orientation of deformation is consistent with overriding ice

movement to the south-southeast.

ig. 7. Kinked spruce tree trunk rooted in gravel. Trunk leans to the

outh-southeast below kink midway through enclosing sand bed C.

pper tree segment remains vertical through overlying gravel, where

is truncated at the base of overlying sand bed D (90 cm ice axe for

cale).

P.J. Fleisher et al. / Geomorphology 75 (2006) 201–211206

noteworthy is that the northwest, upglacier side of all

tree trunks shows evidence of abrasion and all pre-

served limbs are bent toward the southeast.

The kinked style of deformation is best displayed by

a spruce tree (12 cm diameter) within sand bed C (Fig.

7) and rooted in underlying gravel. The lower 30 cm of

the tree trunk leans to the southeast, whereas the upper

portion maintains a vertical orientation through the sand

and overlying gravel. Abrupt truncation occurs as the

tree encounters the fine-grained silt and clay layer at the

base of the overlying sand bed D. The upper-most

segment of the tree is missing, presumably lost to

bluff retreat. Adjacent to the linked tree are small, 2–

3 cm diameter alder that are offset to the southeast

approximately 10 cm from their base and root ball.

The least conspicuous style of deformation is repre-

sented by a distinct curvature in the lower 20–30 cm

trunk segment of several trees. Centimeter-scale, hori-

zontal tension fractures are distributed across the zone

of greatest flexure. This style of deformation is con-

fined to the lower portion of sand bed C, directly above

the silt and clay that accommodates lateral displacement

of adjacent trees. As with offset trees, the directional

orientation of deformation is to the south-southeast

(Fig. 8). As with the kinked tree, trunk diameter is

slightly larger (12 cm) than trees that were offset,

thus suggesting resistance to failure.

An analysis of the growth rings (Smith and Rennie,

1991; Shortle et al., 2003; Shroder, 1980; Blasing and

F

s

U

it

s

Fig. 8. Bent spruce tree trunk. Only the base of this 12 cm diameter

tree is deformed. Arching curvature and tension fractures indicate

deformation by directional stress from left-to-right, which coincides

with the direction of overriding ice (south-southeast). The tree is

rooted in gravel beneath sand bed C (90 cm ice axe for scale).

P.J. Fleisher et al. / Geomorphology 75 (2006) 201–211 207

Fritts, 1976) shows a repetitious pattern that is consis-

tent for trees from all four sand beds. The pattern

consists of thinner rings during the last 8 to 13 years

of growth, representing approximately half of the living

period. All trees, deformed and not deformed, contain

the same pattern. The thinner outer rings tend to be

thinnest on the south-southeasterly side of the trees, an

orientation that corresponds to the direction from which

the most severe weather approaches. Tree deformation

(offset, bent nor kinked) has no effect on the occurrence

or orientation of this pattern.

6. Discussion

6.1. Origin of 1993–95 deformation till

As illustrated in Fig. 2, basal ice that was thrust

forward during the surge was relatively free of debris,

thus indicating that very little new, subglacial material

was being transported. Yet, the overriding ice resulted

in the deposition of a new, sub-meter, fluted till and

push moraine. The only possible sediment source for

the new till was the overridden substrate. We suggest

that subglacial saturation by infiltrating, silt-laden melt-

water that effectively saturated the pre-existing, older

till and diamicton. Saturation led to loss of shear

strength, followed by mobilization as overburden pres-

sure exceeded substrate resistance.

Field evidence in support of this is represented by

the common occurrence of new till resting directly upon

overridden and flattened alder. Lateral continuity of this

till with the till of a push moraine indicates synchro-

nous formation. Indeed, substrate mobilization ob-

served at the ice front during the surge demonstrates

that a newly developed deformable bed was squeezed

and extruded from beneath the glacier. As the moraine

formed it was progressively overridden and in a con-

stant state of formation. Similarly, saturated, subglacial

deformation till was forced into basal ice grooves, thus,

forming flutes.

6.2. Origin of pre-1993–95 diamicton

Field evidence suggests several possible modes of

origin for the diamicton. Fine-grained texture, firmness,

and lack of sorting support deposition by ice, but zones

of crude stratification suggest substrate mobilization

that involved incorporation of till within underlying

units, including outwash. Although coarse, well-

drained outwash gravel typically maintains low pore-

water pressure and retains high sediment strength to

resist deformation by overburden pressure, fast basal

sliding (indicated in Fig. 2) is known to initiate plough-

ing (Boulton and Hindmarsh, 1987; Brown et al., 1987;

Benn and Evans, 1996; Benn and Evans, 1998), which

in turn would incorporate till within underlying out-

wash. This mechanism would account for remnants of

sorting, vestiges of stratification, and lenses of lag

boulders that are commonly found within the matrix

supported diamicton beneath the older till. Assuming

the till was deposited during the 1965–67 surge, this

mechanism would suggest that the diamicton might

have formed synchronous with the till by substrate

deformation during that surge.

An alternative to ploughing would involve the addi-

tion of a silt and clay matrix through infiltration of

highly turbid subglacial water, thus placing fines within

pore spaces of otherwise permeable outwash gravel.

The net effect would be to reduce drainage capacity

and impeding water escape, which in turn would give

rise to increased pore-water pressure, diminished

strength, and subsequent deformation through mobili-

P.J. Fleisher et al. / Geomorphology 75 (2006) 201–211208

zation. Considering the lenticular clusters of clast-sup-

ported gravel and isolated sub-meter size boulders

within the diamicton, and the unconformable lower

contact, another possibilities might include flood-relat-

ed, grain-supported, scouring debris flows known to

occur in conjunction with outburst floods (Russell et

al., 2002). The debris from such a flow would, in a

single event, be deposited on a surface locally scoured

into outwash.

The lack of soil development along the diamicton/till

contact has multiple implications. Either, 1) the diamic-

ton is the product of bed deformation beneath overrid-

ing ice that also deposited the till, 2) the time between

events was too brief for soil formation to occur, 3) the

diamicton and till are of different ages, but any soil that

may have formed on the diamicton was subsequently

effaced by the advance that deposited the till. This is

unlikely because the contact between the diamicton and

till is gradational, not erosional. None of these alter-

natives offer definitive information related to the histo-

ry of events leading to diamicton formation.

7. Deformation by overriding ice

7.1. Stratigraphic factors

The thickness of the deforming layer beneath the ice/

substrate interface represents the depth of deformation,

which is determined by the depth at which shear strength

exceeds shear stress. Water escape properties governed

by grain size will influence pore-water pressure, which

in turn controls shear strength (Clarke, 1987; Benn and

Evans, 1998; Boulton and Hindmarsh, 1987; Alley,

1989; Hart et al., 1990). Therefore, the most effective

shear stress would be confined to strata-bound zones

where grain size favors saturation, high pore-water pres-

sure, and reduced shear strength. Under these condi-

tions, deformation will be most extensive in units of

finer grain size, whereas units consisting of coarser

clasts above and below will remain undeformed.

Other factors to consider include the amount of

infiltrating water, the depths of the deformable layers

below the ice/substrate interface, and the number of

buried fossil-trees. Because subglacial water infiltration

diminishes with depth below the ice/substrate interface,

the greatest potential for substrate deformation also

diminish with depth. Therefore, deformation of trees

within the upper sand beds (C and D) would be most

pronounced. Indeed, the most conspicuous tree defor-

mation is in these beds. This supports the notion that

stratigraphic variation may control the location of de-

formation in the substrate.

Such is the case at Waterworks where it appears that

subglacial water infiltrated several meters of coarse out-

wash to saturate the thin silt and clay veneer at the base of

the upper two sand beds. Saturation caused the develop-

ment of positive pore pressure, which in turn reduced

shear strength leading to strata-bound deformation. The

most common expression of this deformation is repre-

sented by the lateral offset of trees displaced as much as

14 cm along shear planes that are confined to the silt and

clay-rich layer at the base of the two shallowest sand

beds (C and D) (Fig. 6a and b). The directional compo-

nent of all styles of deformation is very consistent and

parallels the direction of overriding ice. Trees held sta-

tionary by a firm root system were offset along shear

planes within a layer of lower competent silt and clay.

Furthermore, the trunk deformation (sheared and bent) is

concentrated within the lower portion of all sand beds.

The undeformed, still vertical trees trunks indicate a lack

of deformation within the upper portion of sand beds C

and D. It appears that lateral displacement by shearing

along the silt and clay layer shifted sand beds C and D to

the southeast as cohesive units, thus causing various

forms of tree deformation. The displacement of trees,

stacked vertically throughout the section, represents

compound lateral movement, as each sand bed and over-

lying gravel was shoved to the southeast along incom-

petent basal silt and clay.

The style of deformation represented by the kinked

tree (Fig. 7) also formed in association with shearing at

the base of the sand bed. The slightly larger diameter

tree (12 cm at the trunk base), however, offered suffi-

cient resistance to avoid failure by shearing. Instead,

differential, lateral movement at the base of the sand

shifted the entire bed and overlying units several cen-

timeters in the downglacier direction. The location of

the kink high in the sand unit suggests that while the

root system held the tree base firmly in place, the trunk

was tilted as the lower portion of the host sand was

differentially displaced more than the upper portion.

The still-vertical, upper trunk, held within the upper

sand and overlying gravel, shows no sign of deforma-

tion. Bedding within the upper sand remains undis-

turbed, which indicates differential movement was

concentrated within the basal silt and clay. The offset

of trees elsewhere along sand bed C confirms differen-

tial movement at the base of the sand. We propose that

stresses were greatest along the base of the upper two

sand units, and that while smaller diameter trees were

offset, others with slightly large trunk diameters de-

formed differently.

The flexing of tree trunks expressed in a broadly

arching curvature might be attributed to causes other

P.J. Fleisher et al. / Geomorphology 75 (2006) 201–211 209

than post-burial deformation if it were not for the

consistent orientation of the curved segments, the hor-

izontally oriented tension fractures that occupy the most

severely bent segment of the trunks, and the tree ring

pattern common to all trees. The shape and extent to

which the trunk is bent would be accommodated by the

same sense of sand bed displacement and stress distri-

bution proposed for the kinked tree. Within a sand bed

that is laterally displaced by movement along the basal

silt and clay layer, the directional stresses would be

most evident in the lower portion of a firmly rooted tree

where differential movement is the greatest. This

explains the observed orientation of bent trunks and

the distribution of associated tension fractures.

The pattern of growth rings recognized in trees from

all four sand beds, yields interesting implications re-

garding the history of deformation. All styles of tree

deformation indicate that stress was applied from the

north-northwest, thus, producing deformation to the

south-southeast. This is completely out of phase with

stresses responsible for anomalously stunted outer tree

rings. This suggests that the pattern of ring thickness

was controlled entirely by conditions that existing prior

to tree burial and before deforming forces were applied.

7.2. Depth of deformation

The common unidirectional aspect of stresses re-

quired to produce all styles of tree deformation are

attributed to overriding ice movement to the southeast.

The lack of a glacial till within the outwash sequence

above the sand beds implies that deformation at depth

may be linked to the emplacement of the overlying

diamicton and/or till units. The undulating, irregular

unconformity along the lower contact of the diamicton

represents an interval of erosion during which evidence

for an additional overriding event may have been re-

moved. This unconformity, however, is absent else-

where in the stratigraphy of Weeping Peat Island,

thus, suggesting local significance.

The directional orientation of tree deformation is

consistent with surface flutes thought to have been

produced by overriding ice during the 1966–67 surge,

as depicted on U. S. Geological Survey aerial photos

(663-28, 9-2-66, Bering and 663-23, 9-2-66, Bering).

Each style of deformation represents a response to a

common stress condition that penetrated deeply into the

substrate. The minimum depth to which deforming

stresses penetrated the Waterworks section may be esti-

mated from the distance between the overriding ice/

substrate interface and the deformed, tree-bearing strata.

The position of the ice/substrate interface depends upon

which interpretation for the origin of the diamicton and

overlying till is favored. The overlying till is anoma-

lously thick, consisting of three zones that are only

distinguished by the relative resistance to erosion. All

three may have been deposited during a single overrid-

ing event or by three separate events. Unfortunately,

field evidence leading to a confident interpretation of

which alternative is correct is lacking, thus, both are

possible. Similarly, several alternative processes may

account for the origin of the diamicton, all of which

are also related to the ice/substrate interface. Because

the depth of deformation is greatest when measured to

the base of sand beds C, it is used to calculate the

minimum depth to which deforming stress penetrated.

Several possible alternatives exist. If a single over-

riding event deposited all 7.9 m of till, as well as the

diamicton, then the minimum depth of deformation

would be measured from the top of the upper most

till to the base of sand bed C, a distance of 15.76 m (see

Fig. 4). If the anomalously thick till is the product of

multiple overriding events, however, as suggested by

differential resistance to erosion, then the depth of

penetrating stress would be less. From the top of till-

B, the depth is 12.76 m, and from till-A it is 10.26 m. If

overriding ice formed the diamicton, then the distance

is 7.86 m. If the unconformity at the base of the

diamicton is a surface of glacial erosion, the depth

would be 5.3 m. Although a definitive minimum

depth cannot be determined from existing field data,

the forces of overriding ice reached much more deeply

into the underlying strata than the thickens of a typical

deforming layer (Benn and Evans, 1998).

8. Conclusion

Clean basal ice overrode the foreland of Bering

Glacier during the 1993–95 surge, thus, very little

new, subglacial material was deposited, yet, a new

till formed at the ice/substrate interface. The overrid-

den foreland primarily consists of pre-existing till, a

till-like diamicton, and outwash. The new till devel-

oped as highly turbid, subglacial meltwater invaded

and saturated the substrate, thus, causing positive

pore-water pressure at shallow depths of less than

one meter. Fine textured sediment (till and diamicton)

experienced diminished shear strength leading to mo-

bilization and pervasive deformation, thus, generating

a newly formed deformation till. Coupling at the ice/

substrate interface above well-drained, coarser out-

wash was accompanied by ploughing and destruction

of pre-existing, primary, sedimentary structures. Poor

sorting and subtle remnants of stratification within the

P.J. Fleisher et al. / Geomorphology 75 (2006) 201–211210

resulting diamicton is interpreted to represent vestiges

of bedding in deformed outwash. Thus, the overrid-

den, near-surface deformable bed served as source

material for new till and diamicton.

Bluff exposures show that deformation at greater

depths during an earlier advances was confined to

sub-meter-thick sand beds within an outwash sequence.

Numerous small trees, rooted in gravel and buried by

sand remain preserved in the living position. All trees in

the upper two sand beds show some form of conspic-

uous deformation, such as 1) trunks offset from the root

systems along, sediment-bound horizontal shears, 2)

bent and fractured lower trunk segments, and 3) a

kinked tree trunk. The strong directional component

of deformation common to all trees is parallel to the

direction of ice movement across the foreland. Defor-

mation is attributed to shear stress related to overriding

ice. Strain was confined to silt and clay-rich layers

where shear strength was reduced because of saturation

at depths that range from 15.76 to 5.31 m beneath

potential ice/substrate interface surfaces above. Al-

though the actual minimum depth of deformation

remains speculative, it is clearly greater than a typical

deforming layer. Although deformation is linked to

overriding ice, evidence presented here cannot be

used to distinguish surging flow from a normal ad-

vance. These observations may provide a field-based,

frame of reference for investigations of deformation

beneath active glacial systems.

Acknowledgments

Funding for this investigation was from National

Geographic Society Grants to Fleisher and the SUNY-

Oneonta Foundation. Logistical assistance, cooperation

in staging operations, and access to remote field sites

were provided by the Prince William Sound Science

Center, the Chugach National Forest Office, and Wilder-

ness Helicopter, all in Cordova, Alaska. Dorothy Peteet

provided species names for trees. Eric Natel helped

collected wood samples. Moira Beach made initial edi-

torial suggestions. Input from reviews by David Barclay

and David Butler provided helpful suggestions for im-

provement. Diana Moseman applied technical skills in

figure preparation. We are grateful to all participating

colleagues of the Bering Glacier Research Group.

References

Alley, R.B., 1989. Water-pressure coupling of sliding and bed defor-

mation: II. Velocity depth profiles. Journal of Glaciology 35,

119–129.

Alley, R.B., 1993. In search of ice-stream sticky spots. Journal of

Glaciology 39, 447–454.

Benn, D.I., 1995. Fabric signature of till deformation, Breidamerkur-

jokull, Iceland. Sedimentology 41, 279–292.

Benn, D.I., Evans, D.J.A., 1996. The interpretation and classification

of subglacially deformed materials. Quaternary Science Reviews

15, 23–52.

Benn, D.I., Evans, D.J.A., 1998. Glaciers and Glaciation. Arnold,

London. 734 pp.

Blasing, T.J., Fritts, H.C., 1976. Reconstruction past climate anoma-

lies in the north Pacific and western North America from tree ring

data. Quaternary Research 6, 563–580.

Boulton, G.S., 1979. Processes of glacier erosion on different sub-

strata. Journal of Glaciology 23, 15–38.

Boulton, G.S., Hindmarsh, R.C.A., 1987. Sediment deformation be-

neath glaciers; rheology and geological consequences. Journal of

Geophysical Research 92, 9059–9082.

Brown, N.E., Hallet, B., Booth, D.B., 1987. Rapid soft bed sliding of

the Puget glacial lobe. Journal of Geophysical Research 92,

8985–8997.

Clarke, G.K.C., 1987. Subglacial till: a physical framework for its

properties and processes. Journal of Geophysical Research 92,

9023–9036.

Clarke, G.K.C., Collins, S.G., Thompson, D.E., 1984. Flow, thermal

structure, and subglacial conditions of a surge-type glacier. Cana-

dian Journal of Earth Sciences 21, 232–240.

Fleisher, P.J., Muller, E.H., Cadwell, D.H., Rosenfeld, C.L., Bailey,

P.K., Pelton, J.M., Puglisi, M.A., 1995. The surging advance of

Bering glacier, Alaska; a progress report. Journal of Glaciology

41, 207–213.

Fleisher, P.J., Cadwell, D.H., Muller, E.H., 1998. The Tsivat Basin

Conduit System persists through two surges, Bering Piedmont

Glacier, Alaska. Geological Society of America Bulletin 110,

877–887.

Fleisher, P.J., Muller, E.H., Peteet, D.M., Lachniet, M.S., 1999. Arctic

Enigma: was Alaska’s Late Pleistocene Bering Glacier really out

of step with its neighbors? Geotimes 44, 16–21.

Fleisher, P.J., Bailey, P.K., Natel, H.H., Natel, E.M., 2004. The para-

dox of overriding ice: landform continuity coupled with dynamic

change, Bering Glacier, Alaska. Abstracts with Programs - Geo-

logical Society of America 36, 172.

Hart, J.K., Hindmarsh, R.C.A., Boulton, G.S., 1990. Styles of

subglacial glaciotectonic deformation within the context of the

Anglian Ice Bed. Earth Surface Processes and Landforms 15,

227–241.

Molnia, B.F., Post, A., 1995. Holocene history of Bering Glacier,

Alaska; a prelude to the 1993–1994 surge. Physical Geography

16, 87–117.

Muller, E.H., Fleisher, P.J., 1995. Surge history and potential for

renewed retreat, Bering Glacier, Alaska. Arctic and Alpine Re-

search 27, 81–88.

Post, A., 1972. Periodic surge origin of folded medial moraines on

Bering Piedmont Glacier, Alaska. Journal of Glaciology 11,

219–226.

Russell, A.J., Tweed, F.S., Knudsen, O., Roberts, M.J., Harris, T.D.,

Marren, P.M., 2002. The geomorphic impact and sedimentary

characteristics of the July 1999 jokulhlaup on the Jokulsa a

Solheimasandi, Myrdalsjokull, southern Iceland. In: Snorasson,

A., Finnsdottir, H.P., Moss, M. (Eds.), The Extremes of the

Extremes: Extraordinary Floods, Proceedings of a symposium

held at Reykjavik, Iceland, July 2000, IAHS Publication, vol.

271, pp. 249–254.

P.J. Fleisher et al. / Geomorphology 75 (2006) 201–211 211

Shroder Jr., John F., 1980. Dendrogeomorphology: review and new

techniques of tree-ring dating. Progress in Physical Geography 4,

161–188.

Shortle, W.C., Smith, K.T., Dudzik, K.R., 2003. Tree survival and

growth following ice storm injury. USDA Forest Service Research

Paper NE 723, 1–4.

Smith, E.R., Rennie, J.C., 1991. Changes in the relationship between

annual growth and climatic variables for four hardwood species.

In: McCormick, L.H., Gottschalk, K.W. (Eds.), Proceedings, 8th

Central Hardwood Forest Conference, University Park, Pennsyl-

vania, March 4–6, 1991, USDA Forest Service General Technical

Report NE, vol. 148, pp. 224–246.