Article body

Introduction

The Laurentide Ice Sheet (LIS) was a fundamental influence on global climate during the Late Wisconsinan and therefore requires accurate representation in paleoenvironmental reconstructions and models. The extent, retreat pattern and gross flow geometry are relatively well constrained (Prest et al., 1968; Dyke and Prest, 1987; Dyke et al., 2003), but details of the flow dynamics are less well understood. A key component of the flow dynamics are ice streams (Clark, 1994; Andrews and Maclean, 2003) and it is essential to determine their location and timing. Once identified, their location is a valuable tool for assessing the success of numerical models which seek to reproduce changes in ice sheet configuration and the location of fast ice flow. Conversely, the location of ice streams can be used to tune models that require parameterisation of basal boundary conditions. Discovery of paleo-ice streams also provides clues to where major iceberg export events occurred, which has important implications for ocean circulation and paleoclimate (Andrews and Tedesco, 1992; Stokes et al., 2005). The aim of this synopsis is to provide an up-to-date map of paleo-ice streams in the Laurentide Ice Sheet, based on published sources where authors have suggested streaming and a concerted mapping campaign, by the authors, using satellite images and aerial photography. We do not regard the map as a definitive inventory, however, it is the most comprehensive compilation to date and should provide a useful reference, resource and stimulus, for future research.

Compilation of paleo-ice streams of the Laurentide Ice Sheet

The flow geometry and pattern of glaciation of the Laurentide Ice Sheet was reported in the Glacial Map of Canada (Prest et al., 1968). Many of the flow patterns that we now recognise as paleo-ice streams, and report in this paper, are visible on the Glacial Map of Canada, but their significance was not understood at this time. The first attempt at an overview of ice streams of the LIS was undertaken by Denton and Hughes (1981) who depicted numerous ice streams, most of which correlated with topographic troughs. Their map was somewhat speculative, but was vital in recognising the importance of ice streams. More recent overviews (Patterson, 1998; Stokes and Clark, 2001) portray between 10 and 15 ice streams of the LIS, but the last few years have seen a considerable increase in paleo-ice stream research and in the number of hypothesised ice streams in the published literature (see special issue of Boreas, vol. 32, 2003 for example).

Different approaches have been used to identify and confirm the location of paleo-ice streams (see Stokes and Clark, 2001 for a review); ranging from field-based investigations of subglacial sediments (Hicock, 1988), to large scale mapping of ice sheet flow patterns (De Angelis and Kleman, in press). A wide variety of evidence has been cited as indicative of ice streaming and recently we have seen the development of ‘diagnostic’ sedimentological (Lian et al., 2003) and geomorphological criteria for identifying ice streams (Stokes and Clark, 1999). Collectively, these criteria can be grouped into a formal ‘ice stream landsystem’ (Clark and Stokes, 2003) which provides an observational template for identifying ice stream ‘footprints’ on a former ice sheet bed.

Contemporary ice streams are defined as ‘regions in a grounded ice sheet in which the ice flows much faster than in the regions on either side’ (Paterson, 1994: p. 301). It therefore follows that to reliably identify a paleo-ice stream requires: (1) evidence for a discrete pattern and (2) some indication of fast flow within this. Geomorphological or geological evidence can define a pattern, distinct from its surroundings, and then further evidence is required to indicate fast flow velocities within this zone. We thus regard that a pattern defined by a drumlin field, for example, is insufficient to define an ice stream unless there is other evidence for fast ice flow. This could be sedimentological indications of fast flow, or a particular erratic dispersal pattern or a systematic pattern of drumlin elongation. Conversely, whilst sedimentological fieldwork at a location might suggest fast ice flow, until we can demonstrate that such flow is a spatially-discrete unit, it is in our view, premature to define an ice stream in this location. Returning to contemporary ice streams, we note that their defining characteristic is not some threshold velocity defining fast flow, but that relatively faster flow velocities are organised into discrete arteries.

Categorising paleo-ice streams of the Laurentide Ice Sheet

The paleo-ice stream map is shown in Figure 1. Strong evidence has been found for 34 ice streams, with a further 15 less certain. Further information on the ice streams in Figure 1 is provided in Table I‑IV in Appendix, including the nature of evidence on which streaming has been inferred, the key references and, where known, the dates during which streaming occurred. We categorised the ice streams according to our confidence in the evidence used to invoke streaming, and the current availability of data on their location and extent. This was a difficult task and we acknowledge that it introduces some subjectivity, but we felt it necessary given that some paleo-ice streams are well-established and robust, but for others there are only hints that fast flow existed. For example, some paleo-ice streams have excellent imprints with well defined trunks and onset zones, mega-scale glacial lineations recording fast ice flow and clear shear margin moraines (e.g. Haldane Ice Stream in Fig. 2). In contrast, paleo-ice streams have been hypothesised in the absence of supporting geomorphic or sedimentary evidence, for example those suggested by Denton and Hughes (1981) along the western margins of the ice sheet (#37‑39 in Fig. 1). To differentiate between such a range of paleo-ice streams they were categorised as: (1) Paleo-ice streams for which there is strong evidence and a well defined footprint; (2) Paleo-ice streams for which there is strong evidence and their extent can be inferred based on topography; (3) Paleo-ice streams for which evidence exists but their extent remains undefined, and (4) Ice streaming has been hypothesised but more evidence is needed to confirm this.

Figure 1

The inventory of paleo-ice streams reviewed in this paper. Ice streams are numbered and further details on each are given in Tables I‑IV. Ice streams in bold have good geomorphic evidence for their margins and extent, whilst those with dashed outlines have poorly defined margins, but it is reasonable to infer extent based on topographic setting. Bold arrows represent locations where there is clear evidence for streaming, but the extent of the ice streaming is unknown. Dashed arrows indicate locations where streaming has been hypothesised but in our view further supporting evidence is required. Ice sheet extent at 18 ka is shown by the solid black line (modified from Dyke et al., 2003). The inset map shows the ice streams on Victoria Island and Prince of Wales Island in more detail, the location of this is indicated by the box on the main map.

Inventaire des courants paléoglaciaires examinés dans cet article. Les courants glaciaires sont numérotés et des détails supplémentaires sur chacun d’eux se trouvent dans les tableaux I à IV. Les courants glaciaires en gras présentent une bonne preuve géomorphique de leurs bordures et de leur extension; ceux aux contours en pointillé présentent des bordures mal définies, mais il est raisonnable de déduire leur extension sur la base de la topographie. Les flèches en gras représentent les emplacements où il existe une preuve évidente d’écoulement mais dont l’extension n’est pas connue. Les flèches en pointillé indiquent des emplacements pour lesquels nous avons fait l’hypothèse d’un écoulement, mais, à notre point de vue, des preuves supplémentaires s’avèrent nécessaires. La ligne noire épaisse (modifiée d’après Dyke et al., 2003) détermine l’extension de l’inlandsis, il y a 18 000 ans. L’encart montre le détail des courants glaciaires sur l’Île de Victoria et l’Île du Prince-de-Galles; l’encadré sur la carte principale en indique l’emplacement.

-> See the list of figures

Many of the readily identifiable paleo-ice stream imprints are along the northern margins of the ice sheet (Fig. 1), whilst those along the southern margin tend to have a more smudged record which is often difficult to unravel. This may reflect the more dynamic nature of the southern margin, which is thought to be characterised by episodic advance and retreat (Mickelson and Colgan, 2003), leaving a geomorphic and sedimentary record which is a composite of multiple events. In concert with Jennings (2006) we contend that ice streaming played a dominant role in controlling the southern margin fluctuations, and anticipate that further paleo-ice streams will be identified here.

The distribution of paleo-ice streams provides a valuable tool for assessing the controls on fast ice flow within ice sheets. The marine troughs along the northern and eastern margins of the LIS are intuitive locations for fast flow, given topographic focusing of ice and the presence of a marine calving front. In total, ten ice streams are found in such locations, but it is interesting to note that ice streaming is by no means restricted to these situations. Fourteen ice streams have neither a calving margin nor a topographically constrained location. Ice streaming is also reported on the hard bed geology of the Canadian Shield (e.g. Dubawnt Lake number 6 in Figure 1 and Appendix), Albany Bay (#26) and Ungava Bay (#16‑17). These examples demonstrate that fast flow is not solely restricted to topographic troughs, calving bays or areas of soft deformable sediments. Other conditions must be capable of triggering and sustaining fast ice flow; possibilities include meltwater distribution and geothermal heat flux. Further research into potential controlling factors is needed and this map provides a useful sample of ice streams to test such theories.

Figure 2

A

B

Landsat satellite image of the Haldane Ice Stream (#4 on Figure 1), an extremely well defined paleo-ice stream imprint. (A) shows ice stream extent with abrupt lateral margins (dashed black) displaying a characteristic broad onset zone converging into a narrow trunk. Direction of flow is indicated by the arrow. (B) is an enlargement of the boxed area in (A) and demonstrates many of the characteristic geomorphic features used to identify paleo-ice streams. In the very north of the image there is hummocky terrain outside the ice stream margin and below this a transition to the streaming area, marked by a suite of shear margin moraines (two examples are indicated by the white arrows). The ice stream bed is much smoother and is characterised by mega-scale glacial lineations and associated streamlined bedforms. These are highly elongate (with length:width well in excess of 10:1) and show a high degree of parallel conformity.

Une empreinte très bien définie de courant paléoglaciaire : l’image satellite Landsat du courant glaciaire Haldane (#4 sur la Figure 1). (A) montre l’extension du courant glaciaire avec des bordures latérales abruptes (pointillé blanc) affichant une zone caractéristique, large à son début, convergeant en un tronc étroit. Une flèche indique la direction de l’écoulement. (B) est un agrandissement de l’aire encadrée en (A) et révèle beaucoup des traits caractéristiques géomorphiques utilisés pour identifier les courants paléoglaciaires. À l’extrême nord de cette image, un terrain accidenté délimite la bordure externe du courant glaciaire. En dessous, une suite de moraines de bordure de cisaillement souligne une transition vers l’aire d’écoulement. Le lit du courant glaciaire est beaucoup plus lisse; il est caractérisé par des lignes de coulées glaciaires à grande échelle et est associé à des formes de lit dans le sens de l’écoulement. Celles-ci sont excessivement allongées (avec un rapport de longueur/largeur supérieur à 10) et présentent un parallélisme important.

-> See the list of figures

Figure 3

Comparison of the dimensions of Laurentide paleo-ice streams with contemporary Antarctic ice streams, demonstrating the much larger variation in length and width of the Laurentide paleo-ice streams. All Laurentide paleo-ice streams whose dimensions are known are included and numbered as in Figure 1. Antarctic examples used are as follows: Lambert Glacier, Bailey Ice Stream, Recovery Glacier, Rutford Ice Stream, Blackwall Ice Stream, and the five Siple Coast ice streams. Dimensions of Antarctic ice streams were measured from the RAMP SAR imagery produced by Jezek and RAMP product team (2002).

Comparaison des dimensions des courants paléoglaciaires laurentidiens avec celles des courants glaciaires contemporains de l’Antarctique; les courants laurentidiens témoignent de plus grandes variations en longueur et en largeur. Cette figure comprend tous les courants paléoglaciaires laurentidiens aux dimensions connues et sont numérotés comme dans la Figure 1. Les exemples antarctiques sont les suivants : glacier de Lambert, coulée de Bailey, glacier de Recovery, coulée de Rutford, coulée de Blackwall et les cinq courants glaciaires de la Côte Siple. Nous avons mesuré les courants glaciaires antarctiques à partir de l’imagerie RAMP SAR élaborée par Jezek et l’équipe de produits de RAMP (2002).

-> See the list of figures

The paleo-ice stream map also provides insights into the temporal controls on ice streaming and the interaction between neighbouring ice streams. For example, on Victoria Island and Prince of Wales Island at least three generations of ice streams are revealed by overlapping imprints, each having a different size, shape and flow direction (see inset in Figure 1). Of these, the oldest and largest is the M’Clure Strait Ice Stream (#19) which extended to the shelf edge at the LGM (Last Glacial Maximum). During retreat, the M’Clintock Channel hosted a second ice stream (#10) of slightly reduced dimensions. As retreat progressed, flow ceased in the marine channel and the Transition Bay Ice Stream (#12) flowed perpendicular to earlier events, calving along the east coast of Prince of Wales Island. This example clearly demonstrates the high degree of spatial and temporal variability of Laurentide ice streams and provides a useful context with which to compare the ongoing changes taking place in contemporary ice streams in West Antarctica (Conway et al., 2002).

Finally, we note that Laurentide ice streams show a much wider range of dimensions than those currently operating in Antarctica (Fig. 3). Laurentide paleo-ice streams operating at the LGM are much larger than contemporary ice streams. During retreat, many ice streams are comparable in size to contemporary examples, but some exist below the size normally associated with streaming. This variability suggests that ice stream configurations are highly variable and that contemporary ice streams represent a relatively narrow part of the wider spectrum (Fig. 3).

Conclusions

This brief synopsis provides an up-to-date map (and associated information) on the location of known and hypothesised paleo-ice streams of the Laurentide Ice Sheet. It is compiled from previous overviews (Patterson, 1998; Stokes and Clark, 2001), a review of new published material, and synoptic mapping using satellite imagery and aerial photographs. In total, 49 hypothesised ice streams are reviewed and categorised according to the strength of evidence for streaming and knowledge of their extent. Ice streaming is found to be widespread and it is noted that several exist in areas not thought to be conducive to fast ice flow. The location of ice streams also reveals highly dynamic (perhaps unpredictable) interactions between neighbouring ice streams. The map provides a useful dataset to test theories concerning the spatial and temporal controls on streaming within ice sheets. The map should also assist in reconstructing and modelling the Laurentide Ice Sheet. Knowledge of ice stream location is valuable for testing numerical models and, alternatively, could be used as input data to parameterise areas of fast flow. In addition to discoveries of new ice streams, a crucial advance will be to improve the dating controls on ice streaming in order to gain a better understanding of the rates of change and duration of fast flow events.