Evidence for a relict glacial landscape in Quebec-Labrador

PAIAEO, ELSEVIER

Palaeogeography, Palaeoclimatology,Palaeoecology111 (1994) 217-228

Evidence for a relict glacial landscape in Quebec-Labrador Johan Kleman, Ingmar Borgstr6m, Clas Hattestrand Department of Physical Geography, Stockholm University, S-10691, Stockholm, Sweden Received 2 November 1993; revised and accepted 9 February 1994

Abstract

Two major glacial landform systems occur in Labrador. These are the radial lineation and esker swarm reflecting Late Wisconsinan decay, and the Ungava Bay lineation and esker swarm reflecting convergent northward flow. In some sectors the two landform systems give conflicting evidence regarding deglaciation pattern. We have interpreted the ice sheet dynamics in Labrador from morphological data, using a new inversion model that treats spatial patterns of deglacial meltwater landforms separately from lineation patterns. In the George River area we have found that the Ungava Bay swarm of deglacial landforms has been overprinted at a right angle by a younger regional meltwater pattern from the last deglaciation. A similar overprint also exists along the intersection line in west-central Labrador. These relations show that the previously accepted relative ages of the two landform systems (Hughes, 1964; Boulton and Clark, 1990a,b; Klassen and Thompson, 1993) have to be reversed. We interpret the Ungava Bay lineation and esker swarm to represent a 0.25 x l 0 6 km 2 pre-Late Wisconsinan relict landscape, formed during the deglaciation of an older ice sheet and later preserved in a dry-based central zone of the Labrador dome.

I. Introduction

Two major glacial landform systems (Fig. 1) occur in Quebec-Labrador. One is the radial swarm of drift lineations and eskers, that in most sectors can be traced inwards from the peripheries of the peninsula. The other one is the Ungava Bay swarm, comprising lineations and eskers that converge towards Ungava Bay. Both of these systems were previously interpreted as formed by the Late Wisconsinan Laurentide Ice Sheet, and used to reconstruct the retreat pattern of that ice sheet (Prest, 1970; Boulton et al., 1985; Dyke and Prest, 1987). Considerable uncertainty, however, still exists regarding the late-glacial Laurentide Ice sheet dynamics in central Quebec-Labrador. A key problem is that the location of some glacial lakes is in direct conflict with the retreat pattern inferred from glacial lineations and eskers. This is the case south of Ungava Bay, where traces of two 0031-0182/94/$7.00 © 1994 Elsevier Science B.V. All rights reserved SSDI 0031-0182(94)00016-2

large glacial lakes, Glacial Lake Naskaupi and Glacial Lake McLean, were described by Ives (1960). Reconstructions that do account for the existence of Glacial Lake Naskaupi involve either a tortuous ice margin outline, with a narrow "finger" projecting in a N N E direction from the main ice mass (Prest 1970), or involve a p o o r fit between postulated flow lines and observed lineation and esker pattern south of Ungava Bay (Dyke and Prest, 1987). Other ice sheet reconstructions (Boulton et al., 1985; G r a y et al., 1993) show configurations that are incompatible with the extent and location of these lakes. Ice-dammed lakes in the Arnaud, Feuilles and Mel6zes River valleys on the Ungava Peninsula (Lauriol and Gray, 1987; Vincent, 1989) required damming of ice in the east of these lakes. Yet, current reconstructions show a late-glacial Ungava Peninsula ice divide placed so far west that the damming of these lakes is unaccounted for.

218

J. Kleman et al./Palaeogeography, Palaeoclimatology, Palaeoecology 111 (1994) 217-228

Fig. 1. The eskers and till lineations in Quebec-Labrador, somewhat simplified after Prest et al. (1968). Boxes show the location of Figs. 4 and 6, based on new air photo interpretation, which give evidence that the convergent flow traces towards Ungava Bay were formed by an older ice sheet, unrelated to the Late Wisconsinan ice sheet which formed the major radial landform system. A n o t h e r key p r o b l e m concerns areas in central Q u e b e c where n o r t h w a r d - d i r e c t e d a n d southwestw a r d - d i r e c t e d flow traces intersect (i in Fig. 1). C r o s s c u t t i n g striae led H u g h e s (1964) a n d K l a s s e n a n d T h o m p s o n (1993) to c o n c l u d e t h a t flow

t o w a r d s the n o r t h p o s t d a t e d the flow t o w a r d s the southwest. H o w e v e r , H u g h e s (1964) d i d n o t e t h a t channel systems n e a r M a l a p a r t L a k e i n d i c a t e d t h a t the last ice r e m n a n t s were situated n o r t h o f the intersection zone.

J. Kleman et al./Palaeogeography, Palaeoclimatology, Palaeoecology 111 (1994) 217-228

Boulton and Clark (1990a, b) reconstructed the evolution of the Laurentide ice sheet on the basis of drift lineations mapped from Landsat images. They considered the Ungava Bay swarm to be younger than the radial swarm, which they assigned to near-maximum stages. This interpretation suggests that the eskers in the radial swarm are not deglaciation landforms, but were instead formed underneath the accumulation area of the ice sheet. We consider the evidence for nearmarginal retreat-stage formation of eskers overwhelming, and must therefore reject this specific part of the Boulton and Clark reconstruction. We have tried to resolve the Labrador ice dome dynamics, using an approach with the following components: --Some photointerpretation over the whole peninsula served to evaluate the reliability of the Glacial Map of Canada (Prest et al., 1968), which, in turn, served as the primary data for delineation of flow trace fans. --Detailed photointerpretation over two key intersection areas established relative chronologies. - - A systematic search in aerial photographs for "transparent" meltwater landform patterns, that may indicate cold-based deglaciation. Such patterns may reveal deglaciation iceflow different from ice flow directions inferred from older subglacial land forms. --Reconstruction of successive ice configurations and deglaciation events. For the reconstruction chain, going from landforms to landscapes, and resulting in reconstructed ice sheet configurations along a relative-time axis, we used a new inversion model outlined below. This inversion model was developed as a consequence of the recognition that old landform systems can be preserved in frozen-bed zones of ice sheets (Sugden and John, 1976; Dyke, 1983; Lagerb/ick, 1988; Kleman and Borgstrrm, 1990; Kleman, 1992; Dyke et al., 1992).

2. Method

The inversion model is based on the recognition that an ice sheet interacts with its substratum in three fundamentally different ways, related to wet-

219

bed conditions, dry-bed conditions and meltwater release. A frozen bed involves little or no change of preexisting basal landforms, but a thawed bed leads to continuous reshaping of the substratum. Large-scale meltwater traces such as esker and channel swarms reflect marginal retreat during decay phases. Flow traces form only under wet-based conditions. The substratum is then continuously reshaped and flow-aligned lineaments (till lineations, striae) are produced and destroyed. The amount of reshaping of older forms increases with ice velocity and time. This continuous reshaping process is terminated either by change to a frozen bed or by deglaciation (Kleman, 1990). The preservation potential of lineations is scale-dependent. Large, old, "ghost" lineations may occasionally be discernible despite later wet-based conditions at the site (Lagerb~ick, 1988; Clark, 1993), but we consider it unlikely that small-scale fluting could survive in wet-bed zones for any length of time. Swarms of lineations that lack aligned meltwater traces, such as eskers, were probably created well inside the ice margin (Kleman, 1990). Meltwater traces. The grain sizes and sedimentary structures in eskers indicate that they form by fast rhythmic sedimentation (Sugden and John, 1976), in an inward-transgressive fashion close inside retreating ice margins (Hebrand and b~nark, 1989). Beaded eskers (Norman, 1938) and varved distal esker sediments (DeGeer, 1940) demonstrate forcing by the yearly melt cycle. As the yearly cycle directly affects the ice sheet surface, but not its base, it is evident that the discharge in the conduits is largely from surficial meltwater. Surface meltwater release is typically high near the ice sheet margin and during decay phases but negligible near dome centres. On the basis of the clear link to near-marginal ablation and the evidence for continuity between esker sediments and proglacial sediments (Hebrand and .~mark, 1989) we believe that they formed a few tens of kilometres inside the ice margin at most, and therefore consider eskers to be deglaciation markers. Glacifluvial channels occur in both wet-bed deglaciation landscapes dominated by eskers (Borgstrrm, 1989), and in areas deglaciated under frozen-bed conditions (Dyke, 1993). Where chan-

J. Kleman et aL/Palaeogeography, Palaeoclimatology, Palaeoecology 111 (1994) 217-228

220

nels are the only linear meltwater traces, a coldbased deglaciation is indicated. The regional meltwater pattern indicates the approximate ice slope direction. Glacial lake traces well inside the maxi-

mum-stage ice sheet margins are part of the inward-transgressive deglaciation meltwater landform system (BorgstrOm, 1989). Glacial lake traces do not provide information of ice flow pattern in

Delineation of flow t r a c e fans Flow traces

"--~ I n t e r p r e t e d

patterns

~-~

t';'f~:l',l l l l I, I[ I/[/[,Z~/J'7 z./z;'~'' , I I / I 11

Reconstructed

'

fans

""

,

,, ,, , ,

iiiiii1

,

}

Fig. 2. The figure shows how we packed swarms of glacial landforms into the landscape-level fans shown in Fig. 8a.

Dry-bed deglaciation f a n ~

~ ~ ~

Wet-bed deglaciation~'~

~ ~ ~

"Synchronous"

fan~/~.~

fan ~

~-~ ~-~

'~

- ""

•

/ Legend

Map-plane Migrating p r e s e r vation borderline

Time-radius graphs

Wet-based

Basal ice flow

Dry-based

P r e s e r v e d flow traces

M e l t w a t e r zone

Esker

Time marker

for time-slice map