dil

ioUniC, D

Keywords:Till

eam-cen, ted, and the contacts between the units are either sharp or transitional. The naturerlying sediments, ductile deformation structures, largely undeformed clayey clasts,

and an

Sedimentary Geology 293 (2013) 3044

Contents lists available at SciVerse ScienceDirect

Sedimentar

e lsutes to abrupt climatic shifts (Heinrich, 1988; Oppenheimer, 1998).However, despite their profound impact on the Earth system both atpresent and in the past, ice streams are still poorly understood(Bennett, 2003; Winsborrow et al., 2010).

Ice streams are arteries of fast ice ow within ice sheets in which theow velocity is orders of magnitude greater than in the surrounding ice(Echelmeyer and Harrison, 1990; Echelmeyer et al., 1991; Whillans andvan der Veen, 1993; Holland et al., 2008). Ice-stream parameters suchas lateral migration and changes of width (Jacobel et al., 1996;Bindschadler andVorenberger, 1998), changes inowvelocity (including

British Isles (Boulton and Hagdorn, 2006) and North America(Clayton et al., 1985, 1989; Mickelson and Colgan, 2004). Topogra-phy, bed lithology, geothermal heat ux, and meltwater under andin front of advancing ice sheets played an important role in the for-mation of ice streams (Winsborrow et al., 2010). Most studies pointto elevated water pressure in subglacial sediments and along theicebed interface as triggers of fast ice ow (Boulton and Jones,1979; Clayton et al., 1985; Brown et al., 1987; Clayton et al., 1989;Hicock and Fuller, 1995; Colgan and Mickelson, 1997; Lian et al.,2003; Mickelson and Colgan, 2004; Kehew et al., 2005; Boulton andstoppage) (Retzlaff and Bentley, 1993) or chanare temporarily and spatially variable (Conway

Pleistocene ice streams that terminated onalogues (Stokes and Clark, 2001; Bennett, 2003;tendedhundreds of kilometres beyond themain

Corresponding author. Tel.: +48 566112590.E-mail address: w.narloch@umk.pl (W. Narloch).

0037-0738/$ see front matter 2013 Elsevier B.V. Allhttp://dx.doi.org/10.1016/j.sedgeo.2013.05.001d Antarctica lose their90%) (Bennett, 2003).al sea level and contrib-

Thomason and Iverson, 2009). Highly lobate, palaeo-ice sheet marginsare known from northern Europe (Kleman and Borgstrm, 1996;Punkari, 1997; Houmark-Nielsen, 2004; Johansson et al., 2011), themass mainly by ice stream drainage (c. 70The behaviour of ice streams inuences globSubglacial processesIce streamScandinavian Ice SheetWeichselian glaciationPermafrost

1. Introduction

Modern-day ice sheets of Greenlated upper surfaces of pebbles in the till indicate both bed deformation and enhanced basal sliding under highsubglacial water pressure conditions. It is suggested that the till is a hybrid deposit generated by some combina-tion of lodgement, deformation and ploughing punctuated by periods of basal decoupling. The depth of deforma-tion at any point in time was thinner (up to several decimetres) than the maximum till thickness (c. 2.5 m). Theice sliding velocity estimations indicate velocities of less than 100 to over 2000 m yr1, which suggests anunstable and highly dynamic ice lobe, consistent with spatial variability of till characteristics. Sand wedges inthe deposits beneath the till and the nature of the till/bed interface indicate permafrost under the advancingice sheet. We suggest that under the increasing ice thickness, a layer of thawed, water-saturated sedimentformedon top of the still-frozen ground due to inefcient drainage, and contributed to ice streaming by promotingpervasive deformation and basal sliding.

2013 Elsevier B.V. All rights reserved.

formed lobes (e.g., Clayton et al., 1989; Colgan and Mickelson, 1997;Stokes and Clark, 2001; Mickelson and Colgan, 2004; Jennings, 2006;Editor: J. Knight tectonic lamination, thin horizontal stringers of sorted sediments, ploughingmarks, boulder pavements, and stri-Available online 15 May 2013scopically massive or beddeof the contacts with the undeSedimentological record of subglacial conof the Vistula Ice Stream (north-central Po

Wodzimierz Narloch a,, Wojciech Wysota a, Jan A. Pa Department of Geology and Hydrogeology, Faculty of Earth Sciences, Nicolaus Copernicusb Department of Geoscience, Aarhus University, Hegh-Guldbergs Gade 2, DK-8000 Aarhus

a b s t r a c ta r t i c l e i n f o

Article history:Received 28 January 2013Received in revised form 21 April 2013Accepted 5 May 2013

Deposits of the Vistula Ice Stred at four eld sites in northunits with specic structural

j ourna l homepage: www.ges in ice-ow directionset al., 2002).land have no modern an-Jennings, 2006). They ex-body of the ice sheet and

rights reserved.tions and ice sheet dynamicsand) during the Last Glaciation

trowski b

versity, Lwowska 1, 87-100 Toru, Polandenmark

draining the Scandinavian Ice Sheet during the Last Glaciationwere investigat-tral Poland using micro- and macroscale features. The study reveals several tillxtural and lithological characteristics. The individual till units are either macro-

y Geology

ev ie r .com/ locate /sedgeoHagdorn, 2006; Jennings, 2006; Thomason and Iverson, 2009).Under warm-based ice that was generating meltwater (Paterson,1994; Clarke, 2005; Hooke, 2005), bed deformation varied in timeand space with enhanced basal sliding on top of a pressurisedwater layer (Piotrowski et al., 2004), which is reected in the prop-erties of tills formed under variable basal conditions (Piotrowskiand Tulaczyk, 1999; Evans et al., 2006; Piotrowski et al., 2006;Menzies et al., 2012; Phillips et al., 2013).

Studies of modern subglacial environments are typically conductedunder warm-based glaciers (e.g., Alley et al., 1986; Iverson et al.,1995; Hart et al., 2009) and the importance of permafrost, especiallyin the early stages of ice overriding, has possibly been underestimated(Clarke, 2005; Evans et al., 2006; Waller et al., 2009). Accordingly, de-formation models involving unfrozen bed conditions were typically ap-plied to the reconstructions of past continental ice sheets under whichsuch conditions did not necessarily prevail (Boulton et al., 2001b;Piotrowski et al., 2004). Some researchers have addressed the interac-tions between subglacial permafrost and ice sheets (e.g., Hughes,1992; Cutler et al., 2000) and in most cases such studies were limitedto narrow ice-marginal zones (Piotrowski, 2006; Waller and Murton,2006) and specic geographical regions such as Siberia (Astakhov etal., 1996), British Isles (Waller et al., 2009) or Greenland (Waller andMurton, 2006).

This paper follows upon the pilot study by Narloch et al. (2012)that investigated the strain signature in till deposited by the VistulaIce Stream in Poland. The aim of the present study is to further con-strain subglacial processes under this palaeo-ice stream with focuson the origin of the till and ice-movement dynamics using an ap-proach involving till sedimentology, structural geology, petrographyand micromorphology from four eld sites located within the outer-most 130 km of the ice sheet.

2. Study area

ice lobe during the Pozna (=Frankfurt) phase and reached its max-imum 60 km to the south-east of the Leszno phase limit about18.5 ka BP (Fig. 1) (Wysota et al., 2009; Wysota and Molewski,2011) and this marks the terminus of the Vistula Ice Stream. Thisice stream is a major land-based palaeo-ice stream of the southernScandinavian Ice Sheet (named ice stream B3 by Stokes and Clark,2001) fed by the Baltic Ice Stream further to the north (Punkari,1997; Boulton et al., 2001b; Marks, 2002; Wysota, 2002; Marks,2005; Wysota, 2007; Wysota et al., 2009).

3. Methods

We investigated the Pozna phase till together with the upper partof the underlying deposits at four eld sites (Fig. 1B). The Mielnicaand Obrki sites are located on either ank of the Vistula ice lobe,Nieszawa is in the centre and Kozowo furthest up-ice. In most cases,the Pozna phase till is underlain by sand, similar to sites elsewherein the ice stream area (e.g., Wysota, 2002; Molewski, 2007; Wysota,2007; Wysota et al., 2009).

At each eld site, sedimentary characteristics, bed contacts, faciesgeometry, texture and lithology of the deposits were recorded, and anup to 3.5-m-high vertical prolewas selected for detailed investigations(Fig. 2). The sedimentary successions in all sites have been subdividedinto units of local range (within each site): K1K4 (Kozowo site),M1M4 (Mielnica site), N1N2 (Nieszawa site) and O1O5 (Obrki

of

31W. Narloch et al. / Sedimentary Geology 293 (2013) 3044Fig. 1. Location of the study area. A The Vistula ice lobe area and the main trunkThe area covered by theVistula ice lobe area is located in north-centralPoland (Fig. 1). The southern boundary of the lobe is the maximumice-sheet extent during Pozna (=Frankfurt) phase of the Weichselianglaciation (Wysota et al., 2009; Wysota and Molewski, 2011).

The stratigraphy of the Weichselian glaciation includes at leasttwo till complexes deposited in the Middle and Late Weichselian.The two youngest tills in the Vistula ice lobe area are of LateWeichselian age (e.g., Marks, 2002; Wysota, 2002; Marks, 2005;Molewski, 2007; Wysota et al., 2009; Wysota and Molewski, 2011;Marks, 2012). The lower till belongs to the Leszno (=Brandenburg)phase, which reached its maximum extent about 20.5 ka BP (Fig. 1).After retreating from the area, the ice sheet re-advanced as the Vistulanorth-central Poland. B Study area and the location of eld sites.

Vistula Ice Stream in relation to the Late Weichselian maximum ice sheet extent insite) based on their spatial, structural and textural characteristics.Lithofacies codes of Eyles et al. (1983) were used for logging the de-posits. In each prole, till segments 20 cm high and 30 cm wide sepa-rated by approximately 10 cm gaps were designated for till-fabricmeasurements and sampling for grain-size and lithologic analysis. Ineach segment, undisturbed oriented till samples were taken forthin-section production to study sediment micromorphology.

Till fabrics were determined in each segment by measuring theorientation of 35 elongated pebbles with a-axis lengths between 1and 7 cm and a/b axes ratios of 1.5. The results are presented ascontours on an equal-area Schmidt projection, lower hemisphere.Eigenvectors (V1) and eigenvalues (S1) are calculated according toMark (1973).

udyEyd, m

32 W. Narloch et al. / Sedimentary Geology 293 (2013) 3044Grain-size distribution (56 samples from tills and underlying sedi-ments) in the b2 mm fraction was examined using sieve and laseranalysis and is presented as weight percentage distributions. Lithologywas determined on 28 samples whereby at least 300 pebbles in therange of 510 mmwere identied in each sample according to methodestablished in Polish Geological Institute (Rzechowski, 1971). Grainswere subdivided into two groups: far-travelled rocks of northern prov-enance (Scandinavia and the Baltic depression, comprising crystallinerock, Palaeozoic limestone, dolomite, sandstone and shale; quartziticsandstone, quartzite and quartz) and rocks of local provenance (fromthe southern Baltic depression and northern Poland, comprising Meso-zoic limestone, marl, sandstone, int and chert and PalaeogeneNeo-gene mudstone).

For all proles, till-fabric, grain-size and lithologic analyses carriedout in each segment have been grouped together for each till unit tovisualise their major characteristics (Fig. 3).

Samples for till micromorphology were collected in 80 60

Fig. 2. Vertical logs spanning the Pozna phase till and the underlying sediments at the stal. (2009), and at Obrki from Wysota and Molewski (2011). Lithofacies codes based onmatrix-supported, stratied; Dml diamicton, matrix-supported, laminated; Sm sandeformed; Fd silt, deformed.40 mm standard aluminium Kubina boxes (Carr, 2004). All sampleswere oriented vertically and parallel to the ice-ow direction as inferredfrom till-fabricmeasurements. Analyses of thin sections (4 6 cm, 28 intotal) were carried out using a petroscope and a standard petrographicmicroscope under plain and polarised light with a magnication up to40 times (Piotrowski et al., 2006; Larsen et al., 2007). For the descriptionof micromorphology, we use the terminology of S-matrix microstruc-tures and plasmic fabric types of van der Meer (1993), Menzies (2000)andMenzies et al. (2006)with special focus on the set ofmicrostructuresoccurring in massive tills (Larsen et al., 2007).

Ice sheet sliding velocity, U, was calculated from the dimensions ofploughing clasts as outlined by Iverson and Hooyer (2004):

U RLRS

12

1RL

1RS

1

where is sliding constant 3.43 1010 m3 s1 for normal (clean) iceand 1.47 1010 m3 s1 for basal (debris-rich) ice, and RS and RL areradii of smallest and largest clasts that ploughed, respectively. Smallervalues of RS imply higher sliding velocities because small clasts promotemovement by regelation whereas larger clasts inhibit it (Iverson andHooyer, 2004). At each site, the diameter (averaged longest andshortest principal axes) of each clast associated with distinct ploughingmark was measured, in the vertical face of the section, to determinesize-range of clasts that ploughed. At Mielnica and Obrki, ploughedskeleton grains (treated as the smallest clasts that ploughed) visible inthin sections were also measured to infer the ice sliding velocity.

4. Results

4.1. Kozowo site

The Kozowo site is located on the wiecie Plateau (Fig. 1B). In thesand pit, a Pleistocene sedimentary succession up to 30 m thick is ex-posed. A c. 3.5 m thick prole comprising four sedimentary units wasinvestigated in the upper part of the section (Fig. 2).

4.1.1. Unit K1The unit K1 (up to 5 m thick) rests on varved clays and consists of

stratied sand and sandy silt lithofacies. It has one grain-size mode in

sites. OSL dates at Kozowo from Piotrowski et al. (in prep.), at Mielnica fromWysota etles et al. (1983) are: Dmm diamicton, matrix-supported, massive; Dms diamicton,assive; Sh sand, horizontally stratied; Sr sand, ripple cross-stratied; Sd sand,the ne-sand fraction where sand (58%) dominates over silt (42%)(Fig. 3). Sedimentary structures in coarser grained beds includeundisturbed low-angle planar cross-stratication, ripple and climbingripple cross stratication, while the more silty lithofacies exhibit loadcasts (Figs. 2, 4A). Thickness of individual beds range from 10 to250 cm and sharp, erosional contacts between them are common.Within the undeformed sand beds of unit K1 occurs a structure with asharp boundary which creates a partly-hanged wall (Fig. 5). This struc-ture is internally very chaotic consisting of irregular clasts (Fig. 5) andundisturbed chunks of laminated sand similar to the surroundingmate-rial. The unit is cut by high-angle faults (083/65S) which continuethrough all the overlying units (Figs. 2, 4A).

4.1.2. Unit K2Unit K2 is a 25-cm-thick, dark brown, clayey, massive till (Figs. 2,

4A). It consists of very ne material (50% clay, 45% silt, 5% sand) withmodes in the ne-silt and clay fractions (Fig. 3). Its lower contact issharp and conformable (Fig. 4A) with till draping the surface ofclimbing ripples without truncating or deforming them. Unit K2 con-tains vertical and horizontal cracks and faults that continue from unitK1 upward into K3 (Figs. 2, 4A).

Thin sections exhibit frequent grain stacks and grain lineations(Fig. 6). Also observed are clasts and circular structures with andwithout core stones. There are shear zones in which the orientationof plasmic fabric and a-axes of skeleton grains are very well dened

Fig. 3. Grain-size distribution, petrographic composition and till fabrics in sediment units at the eld sites (Figs. 1, 2).

33W.N

arlochet

al./Sedim

entaryGeology

293(2013)

3044

Fig. 4. Details of sediment sections studied at the eld sites. Kozowo (ice movement from left to right): A Sharp contact between till K2 and unit K1 with load casts structures, high-anglefaults (black arrows) and a clastic dyke (white arrow); B bedded till K4 with boulder that ploughed and high-angle faults (black arrows). Mielnica (ice movement from left to right and to-wards the observer): CD Sharp contact between undeformed M1 and laminated M2 units, and the transition from M2 to M3. Note both sharp and transitional contacts between laminaswithin M3 (D) and a ploughing clast (black arrow). Nieszawa (ice movement from left to right): E N1 and N2 tills with lodged boulders (black arrows) in the lower parts of bothtills; F deformed sand wedges and sub-horizontal shears (white arrows) within clayey till N1. Obrki: G sharp, largely undeformed contact between clayey till O2 and O1 sandbelow (icemovement from left to right). Note theploughing furrowlledwith till; H sharp contact between bedded till O3 andmassive till O4 (icemovement from right to left obliquelyto the observer). Note lodged boulders in both tills. Black arrows show concentrations of clasts that ploughed within sand stringers of unit O3.

34 W. Narloch et al. / Sedimentary Geology 293 (2013) 3044

35W. Narloch et al. / Sedimentary Geology 293 (2013) 3044(Fig. 7A). The zone of banded plasma also contains pressure shadowsand crushed grains (Fig. 7A). Skel-lattisepic plasmic fabrics are rare,while multisepic plasmic fabric types are common.

4.1.3. Unit K3Within a 6-cm-thick zone of mixing, K2 grades upward into the till

of unit K3, which is a dark-brown, sandy and loosely packed diamictonup to 175 cm thick (Fig. 2). The grain-size distribution shows peaks inthe ne-sand and ne-silt fractions (Fig. 3), and sand (51%) dominatesover silt (31%) and clay (18%). The till appears macroscopically massiveand homogenous, but elongated lenses of sorted sediments are occa-sionally found in its lower part. Pebbles up to 7 cm in diameter carrystriations on their upper surfaces. Till fabric is strong (S1 = 0.82) anddips up-glacier, whereby most clasts have NESW a-axis orientation(Fig. 3). The orientation of striations on the pebbles corresponds tothis till fabric trend. Lithology of K3 till ismainly of northern provenance(Fig. 3) with limestone (45%) and crystalline rocks (30%) dominating.Palaeozoic shales occur sporadically (1%). The content of local rocks is

Fig. 5. Field site Kozowo. Example of thermal erosion within sediment unit K1 markedby a broken line. Sediment facies symbols as in Fig. 2.10%.Similar to the underlying till, thin sections obtained from the K3

till show a high content of grain stacks and grain lineations (Fig. 6).The unit exhibits multiple circular structures with core stones. Thefrequency of crushed grains, pressure shadows and necking structuresis moderate. Till pellets are very rare and occur in the upper part of tillonly. The plasmic fabric is skel-lattisepic, and in the lower part of thetill also masepic.

4.1.4. Unit K4Till unit K3 is covered by a structurally different, up to 50-cm-thick

till referred to as unit K4 (Fig. 2). Its matrix (52% sand, 36% silt, 11%clay) is texturally similar to K3 with two modes in ne-sand andne-silt fractions (Fig. 3). The matrix is loosely packed and darkbrown in colour. The boundary between K3 and K4 is marked by anappearance of sand stringers (Fig. 4B) with thickness of 550 mm andlength of b250 cm. In general, the bottom contacts of these sandstringers are sharpwhile the upper contacts are transitional and includea few-mm-thick zone of mixing. Up to 1-cm-large folds and faults arepresent in the sand stringers in the vicinity of ploughed clasts(Fig. 4B). The diameter of clasts that ploughed range between 10 and90 mm. The depths of ploughing marks below these clasts are up to c.1.5 times the clast diameter. Striations were observed on the upperattened surfaces of some clasts. Calculation of sliding velocities

rise at the icebed interface to the vicinity of ice oatation point, leading

to ice decoupling from the bed and meltwater ow in thin layersresulting in sediment sorting (cf., Alley, 1989; Piotrowski and Tulaczyk,1999; Piotrowski et al., 2006; Lesemann et al., 2010). As pore-water pres-sure dropped, basal re-coupling triggered till deposition anddeformation(cf., Weertman, 1968; Iverson et al., 1995; Piotrowski and Tulaczyk,1999) recorded by transitional, smeared upper contacts of the sandstringers with the overlying till. Shearing and strain-related particle ad-vection is indicated by the occurrence of skel-lattisepic plasmic fabrictypes (Jim, 1990). The presence of largely undisturbed sand stringerssuggest that the deformation was thin-skinned, affecting a sedimentdepth of no more than few centimetres beneath the ice sole (see alsoLarsen et al., 2004; Piotrowski et al., 2006). Clast ploughingwas common,facilitated by low sediment strength caused by high pore-water pressure(Tulaczyk, 1999). Layers of massive till between the sand stringers weremainly formed by lodgement and deformation (cf., Dreimanis, 1989;Evans et al., 2006). The system of high-angle faults cross-cutting allunits probably resulted from icemass load changes on sediments duringusing sliding constants for clean and debris-rich ice yields 160 m yr1

and 69 m yr1, respectively, according to the ice sliding velocityestimation of Iverson and Hooyer (2004). A network of high-anglefaults (Fig. 4B) similar to faults found in the underlying units is presentin this unit.

Till fabrics are strong (S1 = 0.81) with dips in both up-glacier anddown-glacier directions and the clasts have a preferred NESW a-axisorientation (Fig. 3). The orientation of striations on the upper surfacesof clasts corresponds to the till fabric trend.

K4 till is characterised by a large content of far-travelled northernlimestones (45%) and crystalline rocks (35%) (Fig. 3). There are veryfew dolomites (4%), and shales are entirely absent. The content oflocal rocks (7%) is lower than in the K3 till.

Grain lineations and grain stacks dominate the micromorphologyof till K4 (Fig. 6). Frequency of crushed grains, pressure shadows,necking structures and circular structures with core stones is moder-ate (as in K3) and till pellets are present (Fig. 6).

4.1.5. InterpretationUnit K1 represents a glaciolacustrine environment as indicated by

the varved clays (Narloch, 2006), whereas sandy part of this unit isinterpreted as fan-delta deposits (Narloch, 2011). The structureshown in Fig. 5 was probably formed in two stages. First, irregular inci-sion was formed by uvial erosion on the surface of the fan delta. Sec-ond, the hanging wall of incision partly collapsed that gave materialwhich inlled but not destroyed part of the incision. Sedimentary char-acteristics of this structure suggest uvial erosion of frozen ground,probably permafrost.

The ne-grained unit K2 is probably a ow till (Dreimanis, 1989;Phillips, 2006). The high content of multisepic plasmic fabric typessuggests the presence of multi-directional strains during till deposi-tion (Menzies, 2000; Menzies et al., 2006). The clayey diamicton K2draped an older accumulation surface consisting of climbing ripplesof unit K1. The lack of deformation below unit K2 suggests that unitK1 may have been in a frozen state.

Subsequently, the ice sheet overrode the Kozowo site and depos-ited subglacial till K3. Fresh, far-travelled material released from theice sole was mixed with local material in the transitional zone be-tween units K2 and K3 (cf., Weertman, 1968; Boulton and Jones,1979; Boulton and Hindmarsh, 1987; Piotrowski et al., 2006). Thetill is a hybrid product of lodgement and deformation processes,where deformation and mixing of far-travelled material with local li-thologies dominated in the lower part of the unit K3.

The origin of the bedded till K4 is envisaged as some combination ofsediment deformation and meltwater washing in basal cavities. Thepresence of sand stringers suggests recurrent events of water pressuredeglaciation.

Fig. 6. S-matrix elements observed in 28 vertical micrographs from the study sites. Ice movement from left to right (Kozowo, Mielnica, and Nieszawa) and from right to left(Obrki). View eld of each micrograph is 17 23 mm. Location in Fig. 2.

36 W. Narloch et al. / Sedimentary Geology 293 (2013) 3044

37W. Narloch et al. / Sedimentary Geology 293 (2013) 30444.2. Mielnica site

The Mielnica site is a sand pit previously studied by Molewski(2007) and Wysota et al. (2009) and is located in the eastern part ofWielkopolska Lowland (Fig. 1B). Exposed is a succession of Pleistocenedeposits up to 15 m thick. A 3.5 m thick prole consisting of four sedi-mentary units was investigated in the upper part of the section (Fig. 2).

4.2.1. Unit M1UnitM1 is up to 3 m thick and consists of stratied sand and silt. The

uppermost 70 cm are made up of ripple cross-laminated sand (Figs. 2,4C). It has a grain-size mode in the ne-sand fraction, and sand (65%)dominates over silt (35%) (Fig. 3). The uppermost c. 5 cm shows de-formed and exhibit minute folds with southerly vergence within setsof ripple cross laminations.

4.2.2. Unit M2M1 is covered by sorted sediment (unit M2) where silt (76%) domi-

nates over sand (24%) with modal values in the ne-silt fraction(Fig. 3). The thickness of the unit ranges from 5 to 20 cm (Figs. 2, 4C,D). Its base is sharp and irregular whereas the top is sharp and planar(Figs. 2, 4D). Unit M2 is laminated with individual laminae up to 3 mmthick and transitional contacts between them. Most common are siltylaminae but sandy ones also occur. In the lower part of unit M2 are iso-lated, intact clasts of unit M1 in which original ripple cross-laminationis preserved. In the upper part of M2 there are dark brown, oval and de-formed clasts of clay.

Fig. 7. Examples of vertical micrographs in polarised light (left) and their interpretations (rieton grains and occurrence of fractured grains with asymmetric pressure shadow within zo4.2.3. Unit M3The unit is a sandy, bedded sediment which is light brown in colour

and up to 65 cm thick (Figs. 2, 4C, D). The grain-size distribution is bi-modal withmodes in the ne-sand and ne-silt fractions (Fig. 3). On av-erage, sand (62%) dominates over silt (28%) and clay (10%). The contactbetween unitsM2 andM3 is horizontal and is sharp to graded; in the lat-ter case there is a few-cm-thick transition zone between the sorted sed-iments below and the till above, characterised by mixing of the twodeposits whereby the content of diamicton increases upwards. Individu-al beds in M3 are 270 mm thick and consist of silt, sand and diamictonwith clayey clasts up to 5 mm in size dispersed throughout the unit. Thecontacts between the individual beds are variably sharp to transitional,in the latter case with a mm-thick zone of mixing. Lateral interngeringof layers made up of different grain sizes is also observed. Ploughingmarks are common in the unit (Fig. 4D). The diameter of clasts associatedwith ploughing marks range between 10 and 45 mm. Till deformationdepth below the clasts corresponds to c. 1.52 times the clast diameter.The sliding velocities calculated using the formula of Iverson andHooyer (2004) for clean and debris-rich ice are 2258 m yr1 and968 m yr1, respectively.

The till fabric of M3 is strongly clustered around the NS direction(S1 = 0.82) and dips up-glacier (Fig. 3). Lithology reveals mainlyfar-travelled components of northern provenance (Fig. 3) with lime-stone (49%) and crystalline rocks (33%) dominating. The content oflocal rocks is relatively small (7%).

Till micromorphology shows abundant grain stacks and grain linea-tions (Figs. 6, 8A). Necking structures, crushed grains and circular

ght) from till units K1 and M3. Ice movement from left to right. Note alignment of skel-ne of banded plasma (A). Location in Fig. 2.

38 W. Narloch et al. / Sedimentary Geology 293 (2013) 3044structures with and without core stones are also present (Figs. 7B, 8A).Other microstructures are associated with domains. Till pellets withinthe domains are rounded and intact (Fig. 8A). The frequency ofskel-lattisepic plasmic fabric is moderate (Fig. 7B) while masepic andbimasepic plasmic fabrics are rare. Also observed were micro-ploughingmarks with and without skeleton grain (c. 2 mm in diameter).

4.2.4. Unit M4Within a 10-cm-thick zone of decreasing-upward stratication,

unit M3 grades into unit M4 which is a c. 2 m thick till (Fig. 2). Thetill is a brown diamicton with sand (61%) dominating over silt(25%) and clay (14%). The grain-size distribution shows two peaksin the ne-sand and ne-silt fractions (Fig. 3). Within the otherwisemassive till M4, up to 5-cm-thick horizontal layers of sand and clayoccur with vertical spacing of c. 30 cm. Small, mm-sized clasts ofclay are found throughout unit M4. The content of these clasts ishighest in the lower part of the till and gradually decreases upwards.Occasionally, laminae consisting of brecciated clay clasts are found.

Till fabric measurements reveal a moderate (S1 = 0.66) NS cluster-ing with a secondmode oriented EW(Fig. 3). M4 till is characterised bya large content of northern limestones (49%) and crystalline rocks (29%)(Fig. 3). Dolomites are few (4%) while Palaeozoic shales also occur (1%)and the content of local rocks is relatively high (10%).

Fig. 8. Examples of vertical micrographs in plane light (left) and their interpretations (right)high content of till pellets and different characteristics of domain contacts with the surrounmicroploughing structure behind the till pellet with skeleton grain core. Location in Fig. 2.Thin sections show high content of grain stacks and grain lineations(Fig. 6). Till matrix exhibits circular structures with and without corestones. The content of crushed grains is smaller than in the underlyingM3 till. Pressure shadows and necking structures also occur. Till pelletsare much more common in the lower than in the upper part of the till(Fig. 6). The plasmic fabric is skel-lattisepic and multisepic.

4.2.5. InterpretationUnit M1 has been earlier interpreted by Molewski (2007) as

glaciolacustrine sediment deposited on a prograding fan delta(Brodzikowski, 1993), which is conrmed by our observations. As theglacier overrode unit M1, a thin deformation zone developed at the topof the glaciolacustrine sand just beneath the sharp, irregular contactwith M2.

Bed deformation and sediment advection and mixing by intensiveshearing were the major mechanisms for the formation of unit M2. In-tact clasts of unit M1 with preserved ripple-cross lamination foundwithin unit M2 point at incorporation of frozen sediment. Transitionalcontacts between the laminae and their lateral interngering suggesttectonic processes (Benn and Evans, 1996; Boulton et al., 2001a) andwe interpret unit M2 as a glacitectonite (Evans et al., 2006) formedwithin a 20-cm-thick zone of deformation.

from till units M3 and O4. Ice movement from right to left. Note that image (B) showsding till matrix: sharp (top contact) and transitional (bottom contact). There is also a

39W. Narloch et al. / Sedimentary Geology 293 (2013) 3044We also interpret the bedding in till M3 to be caused by deformation.Shearing of sand and silt mobilised from the bed resulted in tectoniclamination (cf., Benn and Evans, 1996; Boulton et al., 2001a). This is alsosuggested by the transitional contacts between sediment layers indicativeof strain-induced grain diffusion (Weertman, 1968), and by the inter-ngering of laminae composed of different grain sizes. Microstructuressuch as skel-lattisepic plasmic fabrics, crushed grains and domains pointto high shear-stress concentration (Hooke and Iverson, 1995; Hiemstraand van der Meer, 1997). Layers of massive till were formed by a combi-nation of lodgement (Dreimanis, 1989) and deformation (Evans et al.,2006) although the deformation zone did not exceed a few cm inthickness, which corresponds to observations made in similar palaeo-environments elsewhere (e.g., Piotrowski and Tulaczyk, 1999; Larsen etal., 2004; Piotrowski et al., 2006). Strain and advective grain transportare indicated by diffusive contacts between laminae composed of differ-ent grain sizes, whereby far-travelled lithologies were mixed with localmaterial. However, sharp contacts between sand stringers and thediamicton suggest deposition during basal decoupling and ice slidingover a thinwater layer, which togetherwith the evidence of deformation,indicate variable subglacial conditions and shifting ice-movement mech-anisms during sediment accretion at this site in accord with models ofBoulton et al. (2001a), Knight (2002), Jrgensen and Piotrowski (2003),van der Meer et al. (2003) and Piotrowski et al. (2004). Under highwater pressure conditions, ploughing processes were common (see alsoTulaczyk, 1999).

Till M4 originated from a combination of lodgement, deformationand washing by meltwater. The sorted sediment layers intercalatingthe till were waterlain during phases of basal decoupling. Verticalspacing between the individual layers constrains the maximum de-formation depth to c. 30 cm. Laminae of clayey clasts and areas oftill matrix with micro-clast concentrations formed during stages ofstrong basal coupling as a result of fracturing and redeposition of ma-terial from the sorted sediment layers when subglacial water pressuredecreased.

4.3. Nieszawa site

The Nieszawa site is located in the north-eastern part of KujawyPlateau (Fig. 1B). A sand pit exposes 15 m of Pleistocene deposits pre-viously studied by Molewski (2007). Here, we investigated in detailthe uppermost c. 2.7 m of the section spanning two till units (Fig. 2).

4.3.1. Unit N1The lower till constituting unit N1 is massive, dark grey-brown

diamicton c. 70 cm thick (Figs. 2, 4E). The till has a ne-grained matrix(29% sand, 60% silt, 13% clay) with modes in the ne-sand and ne-siltfractions (Fig. 3). It is dissected by numerous vertical and horizontalcracks. A distinct layer of lodged boulders is found in the lower part ofthe unit (Fig. 4E), whereas wedge-shaped sand bodies are common inits upper part (Fig. 4F). The wedges are 3040 cm high and up to10 cm wide at the top and consist of laminated sand with laminationmimicking the outlines of the wedges. The spaces between individualsand bodies range from 50 to 80 cm. All wedges are deformed towardsthe east and are cut by subhorizontal faults (Fig. 4F).

Orientation of clasts measured in this unit by Molewski (2007)shows a NS ice-ow direction. Lithology is dominated by limestone(55%) and crystalline rocks (32%) with other rock types being muchless abundant (Fig. 3). The amount of far-travelled dolomites and shalesis small (1%), whereas the proportion of local rocks is high (9%).

4.3.2. Unit N2A massive brown till of unit N2 is c. 2 m thick (Figs. 2, 4E, F). The

grain-size distribution is bimodal with modes in the ne-sand andne-silt fractions (Fig. 3). On average, sand (65%) prevails over silt(23%) and clay (12%). The contact with the underlying N1 till is sharp

to gradual and is highlighted by attenuated sand bodies along theboundary (Fig. 4F). Located 25 cm above the base of N2 is a horizontalpavement of lodged clasts and boulders extending laterally for over40 m. Their diameter range from 10 to 115 mm which, based on thesliding constants for clean and debris-rich ice according to Iverson andHooyer (2004), yields sliding velocities of 138 m yr1 and 59 m yr1,respectively. Till beneath the stone pavement is deformed and exhibitsdetached and attenuated chunks of sand from the sand wedges. Defor-mations associated with the lodged boulders reach the base of unit N2.The upper surfaces of the boulders are often striated.

The till fabric is strong (S1 = 0.76) and oriented WNWESE with adistinct up-glacier dip (Fig. 3). Wherever visible, the orientation of stri-ae on the upper surfaces of stones corresponds to the till fabric trend.

N3 till is characterised by a high content of northern limestones(42%) and crystalline rocks (40%) (Fig. 3). The amount of dolomitesis small (4%), while Palaeozoic shales are absent and the content oflocal rocks is 6%.

Thin sections obtained from this till show a predominance of grainstacks and grain lineations (Fig. 6). Pressure shadows and neckingstructures occur. The unit exhibits also circular structures with andwithout core stones. The content of crushed grains is relatively high,whereas till pellets are very rare (Fig. 6). Content of plasmic fabric ismoderate and represented by skel-lattisepic and multisepic fabrics.

4.3.3. InterpretationUnit N1, according to Molewski (2007), is basal till and was formed

by a combination of lodgement, deformation and ploughing processes.In this area permafrost was established under periglacial conditions

as indicated by numerous sandwedges formed by strong soil contractionunder low temperatures, as postulated earlier by Molewski (2007),Wysota et al. (2009) and Wysota and Molewski (2011). The sandwedges likely formed a polygonal network initiated by primary and sec-ondary cracks (French, 2007).When during the Pozna phase the glacieroverrode the frozen soil, the icemovementwas slow and accommodatedprimarily by internal deformation (e.g., Hooke, 2005). Bed deformationfollowed down-ice stretching and fracturing the sandwedges, indicatingboth ductile and brittle strain. Permafrost was subjected to gradual deg-radation under the thickening ice (cf., French andHarry, 1990) andwhenthe icebed interface thawed, debris release from the ice base com-menced causing deposition of the unit N2. The transitional contact be-tween units N1 and N2 suggests advective mixing of far-travelled andlocal material under warm ice-base conditions (see also Boulton andHindmarsh, 1987; Larsen et al., 2004; Baroni and Fasano, 2006; Evanset al., 2006; Piotrowski et al., 2006).

The underlying ne-grained N1 till impeded drainage of subgla-cial meltwater into the bed. Elevated water pressure decreasedbasal coupling, which in turn reduced the thickness of the deforminglayer. Ploughing of clasts projecting from the basal ice into the bedthrough a layer of meltwater was common, and accumulation ofclasts released from the ice sole resulted in the formation of a boulderpavement (cf., Tulaczyk, 1999; Jrgensen and Piotrowski, 2003). Amas-sive till (unit N2) accreted above the boulder pavement from somecombination of lodgement and deformation whereby the depth of de-formation was up to several tens of centimetres and migrated upwardsthrough the prole as the till thickness increased, as predicted by themodel of Larsen et al. (2004).

4.4. Obrki site

This sand pit is located in the north-eastern part of the Dobrzy Pla-teau (Fig. 1B). Earlier investigations conducted here focused on the sed-imentology (Narloch et al., 2012) and stratigraphy (Wysota andMolewski, 2011) of the c. 12-m thick succession of Pleistocene deposits.The topmost 3.5 mof the prolewasdescribed in detail byNarloch et al.(2012) and below we follow their subdivision of sedimentary units

(Fig. 2).

40 W. Narloch et al. / Sedimentary Geology 293 (2013) 30444.4.1. Unit O1The unit consists of well-sorted, ne-grained sand that is up to

2 m thick. The sand is undeformed and exhibits a range of sedimenta-ry structures. Two horizontal pavements of wind-polished pebblesoccur in this unit (Fig. 2).

4.4.2. Unit O2Unit O2 is an up to 60-cm thick till (Fig. 2). Its basal contact is

sharp and occasionally marked by ploughing marks (Fig. 4G). TheO2 till has a dark brown, generally massive matrix with few smudgesof clay. It is ne-grained with two modes in the ne-sand and ne-siltranges (Fig. 3). Pebbles and boulders with wind-polished surfacesoccur in the till. Some stones are so heavily weathered that they easilydisintegrate when handled.

There is a moderate fabric (S1 = 0.66) dipping both up-glacierand down-glacier (Fig. 3), and the a-axis orientations of elongatedclasts show a NWSE maximum. The lithology reveals predominantlynorthern provenances (90%) with relatively high content of localrocks (10%) (Fig. 3). Grain lineations and grain stacks dominate inthis till (Fig. 6). Crushed grains, necking structures and circular struc-tures with and without core stones were also observed.

4.4.3. Unit O3The contact between tills O2 and O3 is generally transitional within

a 10-cm-thick zone of mixing. Till O3 is light brown in colour and has asandy matrix with two modes in the ne-sand and ne-silt ranges(Fig. 3). Horizontal sand and clay stringers occur, and they are disruptedby ploughing marks behind clasts in some places (Fig. 4H). The diame-ter of measured clasts ranges between 6 and 70 mm. The sliding veloc-ities calculated using the formula of Iverson and Hooyer (2004) forclean and debris-rich ice are 392 m yr1 and 168 m yr1, respectively.

Till fabric is relatively weak throughout unit O3 (S1 = 0.57) andshows a NS ow direction (Fig. 3). The till is characterised by ahigh content of northern-provenance rocks (93%) (Fig. 3). Theamount of local rocks (7%) is less than in the unit O2 below. Thin sec-tions reveal that grain lineations and grain stacks are more commonthan over other microstructures (Fig. 6). Till pellets are abundant inthe middle part of O3 (Fig. 6), while some domains appear in itsupper part. Necking structures, crushed grains and circular structuresalso occur. The content of multisepic plasmic fabric is moderate.

4.4.4. Unit O4O3 till is covered by the structurally different, up to 70-cm-thick

O4 till (Fig. 2). This till is characterised by multiple layers and laminaeof diverse texture ranging from sand to gravel. In some laminae, clay-ey clasts several mm in diameter occur. Unit O4 contains lodged claststhat ploughed with sizes between 4 and 85 mm (Fig. 4H).

Till fabric orientation shows a NESW clustering and moderatestrength (S1 = 0.66) (Fig. 3). Similar to the underlying till, unit O4has a large content of northern provenance rocks (93%) and relativelysmall proportion of local rocks (7%) (Fig. 3).

Thin sections obtained from this till show a predominance of grainstacks and grain lineations (Figs. 6, 8B). The content of crushed grainsis high. The unit exhibits domains and circular structureswith andwith-out core stones. The high content of plasmic fabric is represented bymultisepic types. We also observed microploughing marks with skele-ton grains (c. 2 mm in diameter) (Fig. 8B). Sliding velocities estimatedfor micro- and macroscale clasts that ploughed for clean anddebris-rich ice are 1581 m yr1 and 678 m yr1, respectively.

4.4.5. Unit O5The unit consists of macroscopically massive, dark brown, sandy, up

to 20-cm-thick till (Fig. 2). Its grain-size distribution shows two modesin the ne-sand and ne-silt ranges (Fig. 3). The till fabric is strong

(S1 = 0.81) clustering in the NESW direction and dipping mainlyup-glacier (Fig. 3). As with units O3 and O4, northern-provenancerocks dominate (93%) and the content of local rocks is around 7% (Fig. 3).

4.4.6. InterpretationUnit O1 is interpreted as uvial sand deposited under periglacial con-

ditions (Narloch et al., 2012). The overlying till consists of four units, eachexhibiting different textural and structural characteristics except for thelithology which is almost uniform in the whole prole. Narloch et al.(2012) postulated that the till succession was generated by a combina-tion of various processes at the icebed interface related to stressesexerted on the bed by active ice. Bed deformation interrupted by basaldecoupling played an important role during till formation, which sug-gests complex and temporarily variable subglacial conditions caused byuctuating water pressures.

5. Discussion

5.1. Sedimentological record of fast ice ow

Sedimentary evidence of fast ice ow found within the Vistula icelobe area includes thin till sheets, deformation tills, bedded tills, stri-ated boulder pavements, ploughing marks, and severe basal erosion.These features often occur across large areas of palaeo-ice lobes andice streams and characterise tills of high structural and textural con-sistency generated by dynamic processes related to high-water pres-sures at the icebed interface and in the underlying soft sediments(e.g., Brown et al., 1987; Patterson, 1997; Jrgensen and Piotrowski,2003; Lian et al., 2003; Jennings, 2006).

Ploughing marks behind clasts are common in the Pozna phase tilland their presence suggests high pore-water pressures (Tulaczyk, 1999;Jrgensen and Piotrowski, 2003). Ploughing process occurs when thedeposit below the ice sole is weak and the ice is sliding over the bed(Brown et al., 1987). Tulaczyk (1999) suggested that deformation be-neath ploughing clasts may reach depths 4.5 times the size of the clastdiameter. Ploughing marks frequently occur in the bedded tills the Vis-tula Ice Stream area in Kozowo, Mielnica and Obrki. Data show thatthe ploughing clasts have generated deformation up to twice the clastdiameter only, which suggests a signicant downward-increase instrength of the deposits affected by ploughing. This is consistent withHooke (2005) who showed that, beneath the ice sole, normal effectivepressure strengthening the deposit typically increases with depth at agreater rate than the shear stress.

Although there are no modern analogues for the past land-terminating ice streams as the Vistula Ice Stream, some broad compari-sonswith present ice streams regarding theowvelocities are instructive.Ice ow velocities of modern marine-terminating ice streams of the SipleCoast in Antarctica are 100800 m yr1, which are orders of magnitudegreater than in the ice outside the ice streams (5 m yr1) (Shabtaie etal., 1987; Whillans and van der Veen, 1993; Bennett, 2003). Velocity ofthe topographically-conned Jakobshavn Isbr in Greenland increasedfrom 4000 to 12,300 m yr1 during the last three decades (Echelmeyerand Harrison, 1990; Echelmeyer et al., 1991; Holland et al., 2008).

Iverson and Hooyer (2004) proposed method of calculation ofsliding velocity based on clast-size parameters obtained from claststhat ploughed into a sediment substrate. This approach is a promisingtool in investigations of sliding velocities of past ice sheets but thismethod is not conrmed by measurements on contemporary slidingglaciers. Fluidity parameters for ice show higher velocities for normal(clean) ice and smaller velocities for basal (debris-rich) ice. Iversonand Hooyer (2004) point at the poorly known effects of solutes andof friction between debris in ice and ploughing clasts which maylead to overestimation of sliding velocities. On the other hand, ourstudy shows that sliding velocity may be underestimated because ofthe difculties in recognition of smallest clasts that ploughed, espe-cially in case of microploughing structures. This method is instructive

but need to be veried by eld observations.

41W. Narloch et al. / Sedimentary Geology 293 (2013) 3044The full range of sliding velocities estimated for eld data in ourstudy is 592258 m yr1 according to the sliding velocity estimationof Iverson and Hooyer (2004), which broadly corresponds to the mod-ern values quoted above but also shows a substantial spread from siteto site and even within individual beds of a single till unit (e.g., atObrki). Wysota et al. (2009) and Wysota and Molewski (2011),based on OSL ages, concluded that the Vistula ice lobe advanced rapidlyduring the Pozna phase with an average velocity of 400 m yr1, con-sistent with our results. The lowest velocities were obtained from gen-erally massive tills and the highest from bedded tills. This suggests thatthe highest velocities were generated by basal decoupling and en-hanced sliding on a layer of pressurised water (as evidenced by layersof sorted sediments) whereas bed deformation yielded lower ice owvelocities. However, it should be noted that mapping of ploughingmarks is, especially in the case of mm-sized clasts, much easier in bed-ded than in apparentlymassive tills, whichmay be a source of some op-erator bias.

The very wide range of sliding velocities within the lobe area sug-gests frequent changes in ice ow dynamic in time and space that arenot, however, recorded in the grain-size distribution and pebble li-thology of the tills, both being fairly uniform. This indicates thatchanges in ice-movement velocities were not accompanied bymajor shifts in ice-movement directions or the magnitude of subgla-cial erosion. This situation can be expected when the pressure of sub-glacial meltwater uctuates around the ice-oatation pressure,causing frequent shifts in till strength and in consequence ice-owvelocities and movement mechanisms (enhanced sliding vs. bed de-formation), but little impact on the textural composition of the till.The occurrence of microploughing structures suggests a low strengthof the till matrix most likely caused by high pore-water pressure.

A repeated basal coupling and decoupling facilitating phases of fastice ow is indicated by the stringers of sorted sediments giving sometill areas at our eld sites a distinctly bedded appearance. This interpre-tation is consistent with numerous other studies on bedded tills in gen-eral (e.g., Brown et al., 1987; Piotrowski and Tulaczyk, 1999; Piotrowskiet al., 2006) and bedded tills found in areas of palaeo-ice streams (Cofaigh and Evans, 2001; Cofaigh et al., 2010), all documenting hy-draulic lifting of ice from its bed when the pressure of basal meltwaterreached the ice-oatation threshold.Water owing in shallow subglacialcavities would lead to material sorting and deposition of sand and silt inthin layers of large lateral extent. Since water cannot transmit shearstress, appearance of subglacial water sheets causes a zero-friction con-dition best promoting fast movement of ice. We note that bedded tillsat Kozowo, Mielnica and Obrki are found at different depths in the in-vestigated proles. Assuming an across-site synchronicity of till accretionthis indicates that areas of basal decoupling and thus fastest ice owswitched temporarily and spatially under the ice stream similarly towhat is predicted by various versions of the mosaic model of subglacialsystems (see also Boulton et al., 2001a; Knight, 2002; van der Meer etal., 2003; Piotrowski et al., 2004;Menzies, 2012; Trommelen et al., 2012).

Indirectly, basal decoupling is also indicated by the occurrence of tilllayers enriched in clasts consisting of ne-grained materials, typicallyclay. We interpret these clasts as remnants of material originally depos-ited in stagnant water trapped in subglacial cavities (cf., Munro-Stasiuk,2000). Fuller and Murray (2002) postulated that such conditions mayexist prior to glacier surges. Under the highly dynamic Vistula Ice Stream,switches between basal coupling and decoupling would have destroyedthe clay layers creating rip-up clasts, later to be re-deposited within theotherwise massive till matrix of the fast moving ice.

The occurrence of both brittle (grain lineation) and ductile (circularstructures) micro-deformation with necking structures, crushed grainsand different plasmic fabric types in the till proles indicates deforma-tion under uctuating pore-water pressure conditions (van der Meeret al., 2003; Menzies et al., 2006; Piotrowski et al., 2006; Larsen et al.,2007; Narloch et al., 2012). Co-existence of these microstructures

show that tills in all sites were subjected to different styles ofdeformation and are palimpsest structures, which togetherwithmacro-scale features indicate that tills in the Vistula ice lobe area are hybridtills (Piotrowski et al., 2006) or subglacial traction tills (sensu Evans etal., 2006).

5.2. The inuence of bed conditions and permafrost on the dynamics ofthe Vistula Ice Stream

Among numerous factors controlling ice-sheet stability, basal cou-pling and the strength of the bed are of importance. These processesare controlled by the volume and pressure of subglacial meltwater,which in turn depends on the drainage capacity of the bed. Lowwater pressure, typically associated with coarse-grained subglacialdeposits of high hydraulic transmissivity promotes strong basal cou-pling and high sediment strength (Boulton and Jones, 1979; Iversonet al., 1995; Clarke, 2005; Hooke, 2005; Piotrowski et al., 2006). Con-versely, high water pressure occurring in areas of ne-grained de-posits decreases basal coupling and sediment strength which maydestabilise the ice sheet. Boulton and Jones (1979) suggested thatne-grained subglacial areas with elevated water pressure favouredthe formation of elongated ice lobes that advanced much beyondthe general limit of the Laurentide Ice Sheet such as the prominentJames Bay Lobe and De Moines Lobe (Patterson, 1997; Mickelsonand Colgan, 2004; Jennings, 2006). Within heterogeneous subglacialdeposits, pore-water pressures would vary depending on grain sizeand thus a non-systematic distribution of strain rates in the bed cangenerate complicated systems of deformation patterns (e.g., Evanset al., 2006). In the study area, deposits underlying the ice streamtill are either sands or older tills with contrasting hydraulic conduc-tivities, which could have inuenced local ice-ow dynamics.

Although grain-size differences in the subglacial sediments probablyaffected the iceowdynamics of the Vistula Ice Stream,we consider sub-glacial permafrost as an equally important factor. Similar to grain-sizecomposition, the role of permafrost would also control drainage capacityof the bed. Hydraulic conductivity of frozen granular materials is up toabout six orders ofmagnitude lower than of the samematerials in unfro-zen state (Williams and Smith, 1991) making silt, sand, gravel anddiamictons practically impermeable. The maximum thickness of perma-frost in the Central European Lowland during the Last Glaciation couldhave been about 300 m (Szewczyk and Nawrocki, 2011). Delisle (1998,2003, 2007) suggested that the best conditions for permafrost develop-ment occurred around 3018 ka with a peak at about 21 ka, which cor-responds to the time between Leszno and Pozna ice advances. For thistime period, Delisle (2007) estimated permafrost thickness of up to170 m in NE Germany. Concluding from the above calculations, thick-ness of permafrost reached at least tens of metres in the Vistula icelobe area. It seems that 2000 years of cold condition between the Leszno(most of the Vistula ice lobe area was not covered by ice during thisphase, see Fig. 1) and Pozna phases is enough for permafrost aggrada-tion: Taylor and Wang (2008) estimated that 150-m-thick permafrostdeveloped in 800 years in the Canadian Arctic. This time span wouldalso be enough for development of sand wedges, based on OSL ages, asreported by Sokoowski (2007) and Wysota et al. (2009) in the Vistulaice lobe area.

Our data indicate that permafrost existed both in front and below theadvancing Vistula ice lobe. This is shownby the truncated sandwedges atNieszawa, and uvial erosion phenomena in sub-till sediments atKozowo. Permafrost features were also documented elsewhere in theVistula lobe area (Molewski, 2007; Sokoowski, 2007; Wysota et al.,2009; Wysota and Molewski, 2011) and in the neighbouringWielkopolska Lowland (Ewertowski, 2009). Brodzikowski (1987) point-ed to steep walls of subglacial channels and rafts of intact beddedsediments found within Weichselian age deposits as further evidence ofpermafrost.

Cutler et al. (2000) proposed amodel of permafrost growth in front

of an advancing glacier and its subsequent decay under the ice. For

42 W. Narloch et al. / Sedimentary Geology 293 (2013) 3044conditions broadly resembling our study area, they concluded that thetime required for a total thawing of frozen bed would be at least a fewthousand years. That is why during ice advances of shorter durationonly top and bottom fringes of permafrost layer degraded (French andHarry, 1990). The decaywas caused by lowering of the pressuremeltingpoint under the thickening glacier, frictional heat generated at the icebed interface, isolation of the bed from cold atmosphere, and geother-mal heat trapped under the ice. Under the Vistula ice lobe, permafrostprobably degraded gradually under the advancing ice sheet. In theearly stages of ice overriding a mosaic of thawed and frozen patchesformed, with successively larger areas of the bed becoming unfrozenas ice advance persisted (Fig. 9), as predicted by the model of Hughes(1992).

The effect of subglacial permafrost on ice ow dynamics changedwith time. As long as the icebed interface was frozen, the ice sheetwas stable and no streaming occurred. As thawing of the permafrostcommenced, meltwater started lubricating the bed. Porewater pressureincreased rapidly in the thin layer of thawed sediment immediatelybelow the ice sole because drainage into the bed was impeded by thefrozen sediments below. At this stage, probably the entire layer ofthawed sediment constituted a deforming bed thatwas thin but accom-modating very high strains. If the effect of subglacial permafrost on icestreaming was signicant, through blocking evacuation of basal melt-water into the bed, it follows that ice-ow velocities were highest rela-tively early during ice advance. In the course of time the gradual decayof permafrost caused an increase in hydraulic transmissivity of the bed

Fig. 9. Hypothetical distribution of unfrozen (black) and frozen (white) areas (upper panel)Stream overriding permafrost. The upper panel is based on the theoretical model of Hughewhich could have slowed down the ice by increasing basal coupling andstabilising the deforming layer.

6. Conclusions

The investigated sites provide a sedimentological record of fast iceow of the Vistula Ice Stream, one of the major land-terminatingpalaeo-ice streams in the southern portion of the Scandinavian IceSheet during the Last Glaciation. Our observations suggest that:

1. Hydrological conditions at the icebed interface and in the sedi-ments below controlled the ice-movement mechanism and themode of till formation. Under high water pressure conditions, beddeformation occurred contributing to high ice-ow velocities andsediment advection. In places where water pressure reached theice oatation point basal de-coupling occurred leading to enhancedsliding over a thin water layer.

2. Tills generated by the ice stream are hybrids bearing evidence oflodgement, deformation and ploughing. Based on the measuredsizes of clasts that ploughed, the estimated sliding velocity usingthe formula of Iverson and Hooyer (2004) was in the range ofless than 100 to over 2000 m yr1.

3. Bed deformation and basal sliding occurred in a thin layer ofthawed bed between the ice base and permafrost in the substra-tum. Progressive ice advance and ice thickening led to gradual deg-radation of the subglacial permafrost. Ice movement mechanisms,

and deformed and undeformed areas of unfrozen bed (lower panel) beneath Vistula Ices (1992).

Clarke, G.K.C., 2005. Subglacial processes. Annual Review of Earth and Planetary Sci-

43W. Narloch et al. / Sedimentary Geology 293 (2013) 3044ences 33, 247276.Clayton, L., Teller, J.T., Attig, J.W., 1985. Surging of the southwestern part of the

Laurentide Ice Sheet. Boreas 14, 235241.Clayton, L., Mickelson, D.M., Attig, J.W., 1989. Evidence against pervasively deformed

bed material beneath rapidly moving lobes of the southern Laurentide ice sheet.Sedimentary Geology 62, 203208.

Colgan, P.M., Mickelson, D.M., 1997. Genesis of streamlined landforms and ow historyof the Green Bay lobe, Wisconsin, U.S.A. Sedimentary Geology 111, 725.

Conway, H., Catania, G., Raymond, C.F., Gades, A.M., Scambos, T.A., Engelhardt, H., 2002.Switch of ow direction in an Antarctic ice stream. Nature 419, 465467.

Cutler, P.M., MacAyeal, D.R., Mickelson, D.M., Parizek, B.R., Colgan, P.M., 2000. A numer-ical investigation of ice lobepermafrost interaction around the southernLaurentide ice sheet. Journal of Glaciology 46, 311325.

Delisle, G., 1998. Numerical simulation of permafrost growth and decay. Journal ofQuaternary Science 13, 325333.

Delisle, G., 2003. Permafrost in north-central Europe during the Weichselian: howdeep? In: Phillips, M., Springman, S.M., Arenson, L.U. (Eds.), Permafrost:modes of sediment transport and deposition, and bed deformationprocesses were transient in time and space.

4. We consider subglacial permafrost as one major factor contribut-ing to ice streaming by impeding evacuation of meltwater fromthe ice sole into the bed and thus promoting sediment deforma-tion and basal sliding. In the course of ice advance and permafrostdegradation this effect decreased as more water could drain intothe bed, which suggests slowing down of the ice stream due tostronger basal coupling.

5. This study emphasises the role of subglacial conditions in ice move-ment dynamics and delivers a set of features thatmay help to identifyfast ice ow in areas of less well dened geomorphological signature.

Acknowledgements

The following research grants are acknowledged: EU and PolishGovernment programme ZPORR scholarship grant toWN, Polish Min-istry of Science and Higher Education grant no. NN306 316835 toWW, FNU grant no. 09-062326 to JAP, and UMK grant no. 363-G toWW and WN. We thank Pertti Sarala, an anonymous reviewer andjournal Editor Jasper Knight for constructive comments that helpedto improve this paper.

References

Alley, R.B., 1989. Water pressure coupling of sliding and bed deformation: II. Velocitydepth proles. Journal of Glaciology 35, 119129.

Alley, R.B., Blankenship, D.D., Bentley, C.R., Rooney, S.T., 1986. Deformation of till be-neath ice stream B, West Antarctica. Nature 322, 5759.

Astakhov, V.I., Kaplyanskaya, F.A., Tarnogradskiy, V.D., 1996. Pleistocene permafrost ofWestSiberia as a deformable glacier bed. Permafrost and Periglacial Processes 7, 165191.

Baroni, C., Fasano, F., 2006. Micromorphological evidence of warm-based glacier depo-sition from the Ricker Hills Tillite (Victoria Land, Antarctica). Quaternary ScienceReviews 25, 976992.

Benn, D.I., Evans, D.J.A., 1996. The interpretation and classication of subglacially-deformed materials. Quaternary Science Reviews 15, 2352.

Bennett, M.R., 2003. Ice streams as the arteries of an ice-sheet: their mechanics, stabil-ity and signicance. Earth-Science Reviews 61, 309339.

Bindschadler, R.A., Vorenberger, P., 1998. Changes in the West Antarctic Ice Sheet since1963 from declassied satellite photography. Science 279, 689692.

Boulton, G.S., Hagdorn,M., 2006. Glaciology of theBritish Isles Ice Sheet during the last gla-cial cycle: form, ow, streams and lobes. Quaternary Science Reviews 25, 33593390.

Boulton, G.S., Hindmarsh, R.C.A., 1987. Sediment deformation beneath glaciers: rheol-ogy and geological consequences. Journal of Geophysical Research 92, 90599082.

Boulton, G.S., Jones, A.S., 1979. Stability of temperate ice sheets resting on beds of de-formable sediment. Journal of Glaciology 24, 2943.

Boulton, G.S., Dobbie, K.E., Zatsepin, S., 2001. Sediment deformation beneath glaciers andits coupling to the subglacial hydraulic system. Quaternary International 86, 328.

Boulton, G.S., Dongelmans, P., Punkari, M., Broadgate, M., 2001. Palaeoglaciology of anice sheet through a glacial cycle: the European ice sheet through the Weichselian.Quaternary Science Reviews 20, 591625.

Brodzikowski, K., 1987. rodowiskowe podstawy analizy i interpretacji glacitektonizmuEuropy rodkowej. Acta Universitatis Wratislaviensis 934, 1331.

Brodzikowski, K., 1993. Sedymentacja glacilimniczna, cz. I, Procesy de pozycyjne orazcharakterystyka litofacjalna. Acta Geograa Lodziensia 62.

Brown, N.E., Hallet, B., Booth, D.B., 1987. Rapid soft bed sliding of the Puget glacial lobe.Journal of Geophysical Research 92, 89858997.

Carr, S.J., 2004. Micro-scale features and structures. In: Evans, D.J.A., Benn, D.I. (Eds.), APractical Guide to the Study of Glacial Sediments. Arnold, London, pp. 115144.Proceedings of the 8th International Conference on Permafrost, 2125 July 2003.Taylor & Francis, Zurich, Switzerland, pp. 187191.

Delisle, G., 2007. Estimating episodic permafrost development in northern Germanyduring the Pleistocene. In: Hambrey, M.J., Christoffersen, P., Glasser, N.F.,Hubbard, B. (Eds.), Glacial Sedimentary Processes and Products. Blackwell Publish-ing Ltd, Oxford, pp. 109119.

Dreimanis, A., 1989. Tills, their genetic terminology and classication. In: Goldthwait,R.P., Matsh, C.L. (Eds.), Genetic Classication of Glacigenic Deposits. A.A. Balkema,Rotterdam, pp. 1784.

Echelmeyer, K.A., Harrison, W.D., 1990. Jakobshavns Isbr, West Greenland: seasonalvariations in velocity or a lack thereof. Journal of Glaciology 36, 8288.

Echelmeyer, K.A., Clarke, T.S., Harrison, W.D., 1991. Surcial glaciology of JakobshavnsIsbrae,West Greenland. Part I. Surfacemorphology. Journal of Glaciology 37, 368382.

Evans, D.J.A., Phillips, E.R., Hiemstra, J.F., Auton, C.A., 2006. Subglacial till: formation,sedimentary characteristics and classication. Earth-Science Reviews 78, 115176.

Ewertowski, M., 2009. Ice-wedge pseudomorphs and frost-cracking structures inWeichselian sediments, central-west Poland. Permafrost and Periglacial Processes20, 316330.

Eyles, N., Eyles, C.H., Miall, A.D., 1983. Lithofacies types and vertical prole models; analternative approach to the description and environmental interpretation of glacialdiamict and diamictite sequences. Sedimentology 30, 393410.

French, H.M., 2007. The Periglacial Environment, Third Edition. John Wiley and Sons,Chichester.

French, H.M., Harry, D.G., 1990. Observations on buried glacier ice andmassive segregatedice, western Arctic Coast, Canada. Permafrost and Periglacial Processes 1, 3143.

Fuller, S., Murray, T., 2002. Sedimentological investigations in the foreeld of an Icelan-dic surge-type glacier: implications for surge mechanism. Quaternary Science Re-views 21, 15031520.

Hart, J.K., Rose, K.C., Martinez, K., Ong, R., 2009. Subglacial clast behaviour and its impli-cation for till fabric development: new results derived from wireless subglacialprobe experiments. Quaternary Science Reviews 28, 597607.

Heinrich, H., 1988. Origin and consequences of cyclic ice rafting in the Northeast Atlan-tic Ocean during the past 130,000 years. Quaternary Research 29, 142152.

Hicock, S.R., Fuller, E.A., 1995. Lobal interactions, rheologic superposition, and implica-tions for a Pleistocene ice stream on the continental shelf of British Columbia. Geo-morphology 14, 167184.

Hiemstra, J.F., van der Meer, J.J.M., 1997. Pore-water controlled grain fracturing as indi-cator for subglacial shearing in tills. Journal of Glaciology 43, 446454.

Holland, D.M., Thomas, R.T., Young, B., Ribergaard, M.H., Lyberth, B., 2008. Acceleration ofJakobshavn Isbr triggered by warm subsurface ocean waters. Nature Geoscience 1,659664.

Hooke, R. LeB, 2005. Principles of GlacierMechanics. Cambridge University Press, Cambridge.Hooke, R. LeB, Iverson, N.R., 1995. Grain-size distribution in deforming subglacial tills:

role of grain fracture. Geology 23, 5760.Houmark-Nielsen, M., 2004. The Pleistocene of Denmark: a review of stratigraphy and

glaciation history. In: Ehlers, J., Gibbard, P.L. (Eds.), Quaternary Glaciations Ex-tent and Chronology. Elsevier B.V., pp. 3546.

Hughes, T., 1992. Abrupt climatic change related to unstable ice-sheet dynamics: to-ward a new paradigm. Global and Planetary Change 97, 203234.

Iverson, N.R., Hooyer, T.S., 2004. Estimating the sliding velocity of a Pleistocene icesheet from plowing structures in the geologic record. Journal of Geophysical Re-search Earth Surface 119, 111.

Iverson, N.R., Hanson, B., Hooke, R. LeB, Jansson, P., 1995. Flow mechanism of glacierson soft beds. Science 267, 8081.

Jacobel, R.W., Scambos, T.A., Raymond, C.F., Gades, A.M., 1996. Changes in congurationof ice stream ow from the West Antarctic Ice Sheet. Journal of Geophysical Re-search 101, 54995504.

Jennings, C.E., 2006. Terrestrial ice streams a view from a lobe. Geomorphology 75,100124.

Jim, C.Y., 1990. Stress, shear deformation and micromorphological clay orientation: asynthesis of various concepts. Catena 17, 431447.

Johansson, P., Lunkka, J.P., Sarala, P., 2011. The glaciation of Finland. In: Ehlers, J., Gibbard,P.L., Hughes, P.D. (Eds.), Quaternary Glaciations Extent and Chronology A closerlook. : Developments in Quaternary Sciences, 15. Elsevier B.V., pp. 105116.

Jrgensen, F., Piotrowski, J.A., 2003. Signature of the Baltic Ice Stream on Funen Island,Denmark during the Weichselian glaciation. Boreas 32, 242255.

Kehew, A.E., Beukema, S.P., Bird, B.C., Kozlowski, A.L., 2005. Fast ow of the Lake Mich-igan Lobe: evidence from sedimentlandform assemblages in southwestern Mich-igan, USA. Quaternary Science Reviews 24, 23352353.

Kleman, J., Borgstrm, I., 1996. Reconstruction on palaeo-ice sheets: the use of geomor-phological data. Earth Surface Processes and Landforms 21, 893909.

Knight, J., 2002. Glacial sedimentary evidence supporting stickslip basal ice ow. Qua-ternary Science Reviews 21, 975983.

Larsen, N.K., Piotrowski, J.K., Kronborg, C., 2004. Amultiproxy study of basal till: a time trans-gressive accretion and deformation hypothesis. Journal of Quaternary Science 19, 921.

Larsen,N.K., Piotrowski, J.A.,Menzies, J., 2007.Microstructural evidence of low-strain, timetransgressive subglacial deformation. Journal of Quaternary Science 22, 593608.

Lesemann, J.-E., Alsop, G.I., Piotrowski, J.A., 2010. Incremental subglacial meltwater sedimentdeposition and deformation associated with repeated ice-bed decoupling: a case studyfrom the Island of Funen, Denmark. Quaternary Science Reviews 2324, 32123229.

Lian, O.B., Hicock, S.R., Dreimanis, A., 2003. Laurentide and Cordilleran fast ice ow:some sedimentological evidence from Wisconsinian subglacial till and its sub-strate. Boreas 32, 102113.

Mark, D.M., 1973. Analysis of axial orientation data, including till fabrics. Geological So-ciety of America Bulletin 84, 13691374.

Marks, L., 2002. Last Glacial Maximum in Poland. Quaternary Science Reviews 21, 103110.

Marks, L., 2005. Pleistocene glacial limits in the territory of Poland. Przegld Geologiczny53, 988993.

Marks, L., 2012. Timing of the Late Vistulian (Weichselian) glacial phases in Poland.Quaternary Science Reviews 44, 8188.

Menzies, J., 2000. Micromorphological analyses of microfabrics and microstructures in-dicative of deformation processes in glacial sediments. In: Maltman, A.J., Hubbard,B., Hambrey, M.J. (Eds.), Deformation of Glacial Materials: London, Geological Soci-ety Special Publication, 176, pp. 245257.

Menzies, J., 2012. Strain pathways, till internal architecture and microstructures perspectives on a general kinematic model a blueprint for till development.Quaternary Science Reviews 50, 105124.

Menzies, J., van der Meer, J.J.M., Rose, J., 2006. Till a glacial tectomict, a microscopicexamination of a till's internal architecture. Geomorphology 75, 172200.

Menzies, J., Gao, C., Kodors, C., 2012. Microstructural analyses of a Middle Pliocene till fromthe James Bay Lowlands, Canada evidence of potential fast ice streaming. Proceed-ings of the Geologists' Association. http://dx.doi.org/10.1016/j.pgeola.2012.07.002.

Mickelson, D.M., Colgan, P.M., 2004. The southern Laurentide Ice Sheet. In: Gillespie,A.R., Porter, S.C., Atwater, B.F. (Eds.), The Quaternary Period in the United States.Elsevier, pp. 116.

Molewski, P., 2007. Neotektoniczne i glacidynamiczne uwarunkowania wyksztaceniaplejstocenu Wysoczyzny Kujawskiej. Wydawnictwo Naukowe UMK, Toru.

Munro-Stasiuk, M.J., 2000. Rhythmic till sedimentation: evidence for repeated hydrau-lic lifting of a stagnant ice mass. Journal of Sedimentary Research 70, 94106.

Narloch, W., 2006. Stratygraa i rodowiska sedymentacji zlodowacenia wisy wstanowisku Kozowo 2 koo wiecia nad Wis. Nicolaus Copernicus University,Toru (M.Sc Thesis).

Punkari, M., 1997. Glacial and glaciofuvial deposits in the interlobate areas of theScandinavian Ice Sheet. Quaternary Science Reviews 16, 741753.

Retzlaff, R., Bentley, C.R., 1993. Timing of stagnation of Ice Stream C, West Antarctica,from short-pulse radar studies of buried surface crevasses. Journal of Glaciology39, 553561.

Rzechowski, J., 1971. Granulometryczno-petrograczne waciwoci glin zwaowych wdorzeczu rodkowej Widawki. Biuletyn Instytutu Geologicznego 254, 111155.

Shabtaie, S., Whillans, I.M., Bentley, C.R., 1987. The morphology of Ice Streams A, B andC, West Antarctica, and their environments. Journal of Geophysical Research 92,88658883.

Sokoowski, R.J., 2007. Stratygraa i sedymentologia osadw plejstocenu w kamienioomieWapienno, NE Wielkopolska. Nicolaus Copernicus University, Toru (Ph.D. Thesis).

Stokes, C.R., Clark, C.D., 2001. Palaeo-ice streams. Quaternary Science Reviews 20,14371457.

Szewczyk, J., Nawrocki, J., 2011. Deep-seated relict permafrost in northeastern Poland.Boreas 40, 385388.

Taylor, A.E., Wang, K., 2008. Geothermal inversion of Canadian Arctic ground temper-atures and effect of permafrost aggradation at emergent shorelines. Geochemistry,Geophysics, Geosystems 9, 125.

Thomason, J.F., Iverson, N.R., 2009. Deformation of the Batestown till of the Lake Mich-igan lobe, Laurentide ice sheet. Journal of Glaciology 55, 131146.

Trommelen, M.S., Ross, M., Campbell, J.E., 2012. Inherited clast dispersal patterns: implica-tions for palaeoglaciology. Boreas. http://dx.doi.org/10.1111/j.1502-3885.2012.00308.

Tulaczyk, S., 1999. Ice sliding over weak, ne-grained tills: dependence of icetillinteractions on till granulometry. In: Mickelson, D.M., Attig, J.W. (Eds.), GlacialProcesses: Past and Present: Geological Society of America, Special Paper, vol.337, pp. 159177.

van der Meer, J.J.M., 1993. Microscopic evidence of subglacial deformation. QuaternaryScience Reviews 12, 553587.

van der Meer, J.J.M., Menzies, J., Rose, J., 2003. Subglacial till: the deforming glacier bed.

44 W. Narloch et al. / Sedimentary Geology 293 (2013) 3044podczas stadiau gwnego ostatniego zlodowacenia. Nicolaus Copernicus Univer-sity, Toru (Ph.D. Thesis).

Narloch, W., Piotrowski, J.A., Wysota, W., Larsen, N.K., Menzies, J., 2012. The signatureof strain magnitude in tills associated with the Vistula Ice Stream of the Scandina-vian Ice Sheet, central Poland. Quaternary Science Reviews 57, 105120.

Cofaigh, C., Evans, D.J.A., 2001. Deforming bed conditions associated with a major icestream of the last British ice sheet. Geology 29, 795798.

Cofaigh, C., Evans, D.J.A., Smith, I.R., 2010. Large-scale reorganization and sedimenta-tion of terrestrial ice streams during late Wisconsinian Laurentide Ice Sheet degla-ciation. Geological Society of America Bulletin 122, 743756.

Oppenheimer, M., 1998. Global warming and the stability of the West Antarctic IceSheet. Nature 393, 325332.

Paterson, W.S.B., 1994. The Physics of Glaciers, Third Edition. Pergamon, Oxford.Patterson, C.J., 1997. Southern Laurentide ice lobes were created by ice streams: Des

Moines Lobe in Minnesota, USA. Sedimentary Geology 111, 249261.Phillips, E., 2006. Micromorphology of a debris ow deposit: evidence of basal shear-

ing, hydrofracturing, liquefaction and rotational deformation during emplacement.Quaternary Science Reviews 25, 720738.

Phillips, E., Everest, J., Reeves, H., 2013. Micromorphological evidence for subglacialmultiphase sedimentation and deformation during overpressurized uid ow as-sociated with hydrofracturing. Boreas 42, 395427.

Piotrowski, J.A., 2006. Groundwater under ice sheets and glaciers. In: Knight, P.G. (Ed.),Glacier Science and Environmental Change. Blackwell, London, pp. 5060.

Piotrowski, J.A., Tulaczyk, S., 1999. Subglacial conditions under the last ice sheet innorthwest Germany: icebed separation and enhanced basal sliding? QuaternaryScience Reviews 18, 737751.

Piotrowski, J.A., Larsen, N.K., Junge, F.W., 2004. Reections on soft subglacial beds as amosaic of deforming and stable spots. Quaternary Science Reviews 18, 737751.

Piotrowski, J.A., Larsen, N.K., Menzies, J., Wysota, W., 2006. Formation and subglacial tillunder transient bed conditions: deposition, deformation, and basal decouplingunder Weichselian ice sheet lobe, central Poland. Sedimentology 53, 83106.Quaternary Science Reviews 22, 16591685.Waller, R., Murton, J., 2006. The behavior of glaciers on frozen beds: modern and Pleis-

tocene examples. In: Knight, P.G. (Ed.), Glacier Science and Environmental Change.Blackwell Publishing Ltd., pp. 326328.

Waller, R., Murton, J., Whiteman, C., 2009. Geological evidence for subglacial deformation ofPleistocene permafrost. Proceedings of the Geologists' Association: , 120, pp. 155162.

Weertman, J., 1968. Diffusion law for the dispersion of hard particles in an ice matrixthat undergoes simple shear deformation. Journal of Glaciology 7, 161165.

Whillans, I.M., van der Veen, C.J., 1993. New and improved determinations of velocityof Ice Streams B and C, West Antarctica. Journal of Glaciology 39, 483490.

Williams, P.J., Smith, M.W., 1991. The Frozen Earth. Fundamentals of Geocryology.Cambridge University Press, Cambridge.

Winsborrow, M.C.M., Clark, C.D., Stokes, C.R., 2010. What controls the location of icestreams? Earth-Science Reviews 103, 4559.

Wysota, W., 2002. Stratygraa i rodowiska sedymentacji zlodowacenia wisy wpoudniowej czci dolnego Powila. Wydawnictwo Naukowe UMK, Toru.

Wysota, W., 2007. Successive subglacial depositional processes as interpreted frombasal tills in the Lower Vistula valley (N Poland). Sedimentary Geology 193, 2131.

Wysota, W., Molewski, P., 2011. Chronologia i zasigi nasuni ldolodu na obszarze lobuWisy podczas stadiau gwnego ostatniego zlodowacenia. Przegld Geologiczny 59,214225.

Wysota, W., Molewski, P., Sokoowski, R.J., 2009. Record of the Vistula ice lobe advancesin the LateWeichselian glacial sequence in north-central Poland. Quaternary Interna-tional 207, 2641.Narloch, W., 2011. Procesy depozycji i deformacji glin subglacjalnych w lobie Wisy

Sedimentological record of subglacial conditions and ice sheet dynamics of the Vistula Ice Stream (north-central Poland) du...1. Introduction2. Study area3. Methods4. Results4.1. Kozowo site4.1.1. Unit K14.1.2. Unit K24.1.3. Unit K34.1.4. Unit K44.1.5. Interpretation

4.2. Mielnica site4.2.1. Unit M14.2.2. Unit M24.2.3. Unit M34.2.4. Unit M44.2.5. Interpretation

4.3. Nieszawa site4.3.1. Unit N14.3.2. Unit N24.3.3. Interpretation

4.4. Obrki site4.4.1. Unit O14.4.2. Unit O24.4.3. Unit O34.4.4. Unit O44.4.5. Unit O54.4.6. Interpretation

5. Discussion5.1. Sedimentological record of fast ice flow5.2. The influence of bed conditions and permafrost on the dynamics of the Vistula Ice Stream

6. ConclusionsAcknowledgementsReferences