Digital Terrain Mapping in the Reconstruction

THE LAST GLACIAL MAXIMUM BRITISH-IRISH ICE SHEET:

A RECONSTRUCTION USING DIGITAL TERRAIN MAPPING

Running Head: Digital Terrain Mapping of the LGM British-Irish ice sheet

P.T Fretwell, British Antarctic Survey, High Cross, Madingley Road, Cambridge CB3 0ET, UK D.E.Smith, Oxford University Centre for the Environment, South Parks Road, Oxford OX1 3QY, UK S.Harrison, Department of Geography, University of Exeter in Cornwall, Trenough Campus, Penrhyn TR10 9EZ, UK

Corresponding author: P.T.Fretwell, British Antarctic Survey, High Cross, Madingley Road, Cambridge CB3 0ET, UK. Telephone: 44 1223 221400 E-mail: [email protected]

Sponsor: The Leverhulme Trust

Abstract The use of Digital Terrain Mapping in determining the anatomy of the Late Devensian British-Irish ice sheet at a resolution of 500m cell size is illustrated for Boulton et al.’s 1985 and 1991 models and Lambeck’s 1995 model of the ice sheet at its maximum extent as an independent ice mass. Area and volume of the ice sheet are given for each model and the spatial pattern of ice thickness shown in maps. The analyses show that if no account is taken of topography beneath the ice surface, models will seriously overestimate ice volume. It is suggested that as reconstructions of the ice sheet improve, detailed models of ice thickness at the resolution given in this paper may be of value in determining the contribution of the ice sheet to sea surface changes as well in determining the effects of ice loading on glacio-isostasy, neotectonics and possibly on paraglacial processes in areas of high relief.

Keywords: Late Devensian, British-Irish ice sheet, Digital Terrain Mapping.

1 Introduction

Most models of past Quaternary ice sheets are primarily concerned with the spatial extent of the ice. Few consider ice thickness in any detail, and most interest in the nature of these ice sheets concerns patterns of ice flow; location, age and behaviour of ice sheet margins; and implications for the palaeoclimates of these areas. Yet in recent years, interest in the thickness of Quaternary ice sheets has grown, following work in

Scandinavia (e.g. Nesje, 1990) and Scotland (e.g. Ballantyne et al., 1998). Indeed, knowledge of the vertical dimensions of the ice sheets, with the implications this has notably for ice loading and ice volume, is of particular interest in glaciology, geophysics and geomorphology. This paper describes a method of determining the anatomy of a

Quaternary ice sheet using Digital Terrain Mapping. It takes three published models of the Late Devensian British-Irish ice sheet (the BIIS) at its maximum extent as an independent ice mass: the Boulton et al. 1985 “alternative” and 1991 models and the

Lambeck 1995 model, and considers the implications for estimates of global sea surface levels, glacio-isostasy, neotectonics and paraglacial processes. Locations referred to in this paper are shown in Figure 1.

Background

Extent of the Late Devensian ice sheet in Britain and Ireland. In recent years, much empirically-based work has focussed on the likely extent and timing of the Late

Devensian ice sheet at its maximum (the LGM). Most authors now agree that the LGM of the BIIS was probably reached sometime during the period 27 000 – 20 000 radiocarbon years BP, but there is no consensus on other parameters of the ice sheet (e.g. see Hall et

2 al., 2002; Bowen et al., 2003). The view that a single maximum limit was reached, after which deglaciation occurred without noticeable interruption (e.g. Clapperton and Sugden,

1972; Paterson, 1974; Sutherland, 1984), has been replaced by views that deglaciation was interrupted by stillstands or by readvance episodes (e.g. Sissons, 1963; Sissons and

Dawson, 1981; Eyles et al., 1994; Bowen et al., 2002; McCabe and Clark, 2003; McCabe et al., 2005; Carr et al., 2006; McCabe et al., 2007), although there is no consensus on the events identified, as for example in the case of the Perth Readvance (e.g. McCabe et al., ibid.). In addition, opinion is divided on the maximum extent reached, with some authors essaying a relatively limited extent (e.g. Sutherland, ibid.; Bowen et al., ibid.) and others maintaining a more extensive limit (e.g. Hall and Bent, 1990; Hall, 1997; Hall et al.,

2002). There are also differing views on whether the BIIS at its Late Devensian maximum was confluent with the Scandinavian ice sheet (the SIS) as exemplified in the contrasting reconstructions of Sejrup et al. (1987; 2002).

Ice sheet topography and thickness. In contrast to published views on the extent of the ice sheet, there appears to be little debate on its vertical dimensions, probably because they have only been empirically examined in detail in a few areas. The most recent and detailed studies are from NW Scotland, Ireland, Wales and the Lake District, N England.

For NW Scotland, several papers have examined trimlines in local areas (e.g Ballantyne and McCarroll, 1995; Dahl et al., 1996; Ballantyne 1999; Ballantyne and Hallam, 2001;

Stone and Ballantyne, 2006) and the surface altitude of the ice sheet has been inferred across the area as a whole from this evidence (Ballantyne et al., 1998). In Ireland, Rae et al. (2004) mapped the Late Devensian limit from trimlines in the Macgillycuddy’s Reeks

3 in the SW, whilst Ballantyne et al. (2006) provide evidence for the ice sheet surface altitude from trimlines around the Wicklow Mountains. In Wales, McCarroll and

Ballantyne (2000), using trimlines, reconstructed the ice surface at about 850m above present sea level over Snowdonia. In the SW Lake District, N England, Lamb and

Ballantyne (1998) reconstructed the ice sheet surface from palaeonunataks to a maximum of 870m. Elsewhere in Britain and Ireland, reconstructions of the ice sheet surface are less detailed, for example Dury (1953) observed that watershed breaching in the NW

Highlands of Scotland indicated that as an ice sheet developed, an elongate dome was formed in that area along a NE-SW axis, migrating eastwards towards the Great Glen.

Sissons (e.g. 1967) maintained from the evidence of erratics that most of the summits in the Scottish Highlands were covered sometime in the Quaternary; suggesting that there were local ice domes and that the highest parts of the ice sheet could have reached 4000 –

5000 feet (1200 – 1500m) above present sea level. In the Southern Uplands of Scotland,

Cornish (1982) maintained that the ice surface exceeded the highest hills in Galloway (to the W), at over 840m, on the basis of erratics, while Sissons (ibid.) implied that the ice surface declined towards the E. Well defined limits spatially are recognised in the Welsh borderland and South Wales (e.g. Bowen et al., 2002), and in central and northern

England (e.g. Eyles et al., 1994), but few details are available on ice surface height in these areas.

In a theoretical study, based largely upon considerations of ice sheet dynamics, Boulton et al. (1977) sought to provide a generalised model of ice sheet limits and surface altitude at the LGM, estimating the highest ice surface altitude at over 1800m above the ice sheet

4 margin in Scotland and up to 1700m in N England, with broad domes over the

Grampians, central Ireland and the Southern Uplands, but no ice over SW Ireland.

Denton and Hughes (1981) subsequently estimated that at the LGM the BIIS had a single dome over Scotland with a maximum surface altitude of 1750m above its margin. Both the Boulton et al. (ibid.) and Denton and Hughes (ibid.) models assumed that the BIIS was confluent with the SIS at the LGM. Boulton et al. (1985) subsequently produced an

“alternative” model, with a maximum ice surface altitude above 1000m and domes over central Scotland, central Ireland and Wales, for a time when there was no connection with the SIS. Later, Boulton et al. (1991; 2003) produced another solution, with a third model, described as relating to a period when the ice sheet had separated from the SIS, but which had a smaller extent than their 1985 “alternative” model. In this third model, again no ice cover was depicted over SW Ireland, but additionally NE Scotland, parts of Caithness,

Orkney and Shetland were either partially or completely devoid of ice, while nunataks were identified in Scotland and W Ireland. Lambeck (1995) also produced a model of the ice sheet at the LGM, in which the limits were very similar to those in the Boulton et al.

(1991, 2003) model, but the maximum altitude of the ice surface reached over 1400m in

Scotland, and a small ice centre was located over SW Ireland. The models of Boulton et al. (1977; 1985; 1991; 2003), Denton and Hughes (1981) and Lambeck (1995) remain the only published models of the BIIS as a whole which provide detailed estimates of ice surface altitude. The models of Boulton et al. (1985, 1991) and Lambeck (1995) are illustrated in Figure 2.

5 Notwithstanding the many uncertainties concerning the extent of ice at the LGM, the

1981 model of Denton and Hughes has formed the basis for subsequent estimates of ice volume (e.g. Dawson, 1993), while the “alternative” model of Boulton et al. (1985) and the 1995 model of Lambeck have been used in the development of glacio-hydro-isostatic models of crustal movement and relative sea level change, firstly in the pioneering studies of Lambeck (e.g. 1991; 1993a; 1993b; 1995; 1996) and later by Shennan et al.

(2000; 2002), Lambeck and Purcell (2001) and Peltier et al. (2002) Inevitably, the use of these models involves uncertainty. In determining ice volume, the generalised nature of the Denton and Hughes (1981) model can only provide a rough estimate. In the development of glacio-hydro-isostatic models, most accounts do not explicitly refer to the role of topographic variation beneath the ice sheets modelled. Early models (Lambeck,

1991; 1993a; 1993b) would appear to have taken the ice surface altitudes of the Boulton et al. (1985) model as ice thicknesses at the LGM. Later models (Lambeck 1995; 1996;

Lambeck and Purcell, 2001) probably take account of ice thickness, although the accounts refer to both ice thickness and ice height. However, all accounts imply that there is no topography beneath the ice, as Shennan et al. (2002; 2006) and Milne et al. (2006) recognise. Milne et al. (2006) therefore, modelled ice thickness above the underlying topography (Milne et al., ibid., Figure 3, page 937). However, their resulting model of ice thickness, based on a cell size of 0.7 degrees latitude by 0.7 degrees longitude, has a resolution which is too coarse to take account of underlying topography in detail. A potentially major step forward was provided by Shennan et al. (2006), who modelled the extent and thickness of the ice sheet at different (and alternative) stages of its development and decay from circa 32000 calibrated years BP in seeking to determine the

6 effect on isostatic movement, but the detail that Shennan et al. (ibid.) provide both of extent and thickness involves the same scale as that of Milne et al. (2006).

Perspectives. There are many uncertainties arising from work on the age, extent and thickness of the BIIS at the LGM. Indeed, these uncertainties and the approach of different authors raise questions about the meaning of the term LGM itself, a topic which will not be debated here. The excellent compendium of information provided by the

BRITICE project (Clark et al., 2005) may ultimately help to resolve some of these problems, but progress will be enhanced with the development of improved numerical methods of determining the anatomy of the ice sheet.

With the advent of digital terrain mapping, it is now possible to examine in much more detail ice volumes and loading patterns from those ice sheet models for which ice surface altitudes have been determined. Such an approach will be of value in future estimates of the extent and surface altitude of ice during the Late Devensian. This paper examines the likely ice thicknesses of the three broadly comparable models of the BIIS as an independent ice sheet close to the LGM illustrated in Figure 2.

Methodology

The approach adopted in this paper was to take the models shown in Figure 2 and estimate ice volume and ice loading. A GIS (Geographical Information System) was compiled using the ARCGIS suite of programmes, (version 9 or 9.1: ARCMAP,

7 ARCSCENE, ARCCATALOGUE). The ice sheet models were projected on to the

Ordnance Survey National Grid.

The data The first step was to construct three-dimensional models of the original ice surface altitude data. The three models showing ice surface altitude were scanned and geo-rectified. The contours were digitised and then converted to a digital elevation model with a 500 metre (horizontal) resolution, to give an initial representation of the ice height models as in the original papers.

The second step was to construct a Digital Terrain Model of the present land surface of the British Isles and surrounding areas. For the land, NASA SRTM (USGS EROS Data

Center) data was downloaded and mosaiced into a 90 metre (horizontal) resolution three- dimensional grid. A comparison between the data and Ordnance Survey maps indicated that the altitude values have a vertical accuracy of around 1-2m in most areas, the only exception to this being in areas with steep slopes, where the discrepancy is greater. To reduce the error where strong topographical variations occur locally with the very detailed 90 metre cell size, the original grid was re-sampled to a 500 metre cell size by taking the mean of each 500 metre square. Offshore, ETOPO (Eastern Topographics) data, with a 1 minute grid cell size was acquired and these data were also re-sampled to a

500 metre grid. This offshore grid was then mosaiced with the SRTM data to give a seamless dataset of the land and sea topography of the area of Britain and Ireland overlain by ice at 500 metre resolution.

8 Before overlaying the original ice models upon the topography beneath, the ice edge needed to be adjusted. In the original models the ice edge had a vertical level of zero metres OD, but in overlaying the models over the topography, the vertical level of the ice edge was sometimes offshore and below the sea level of the Digital Terrain maps, or onshore and above. The models were therefore re-gridded to keep the ice edge consistent with the underlying surface. The main outputs of the analysis were digital elevation models of ice thickness with a cell size of 500 by 500 metres.

Data analysis

Extent and volume of the ice sheet. Ice thicknesses indicated by the three models shown in Figure 2 are depicted in Figure 3. All the models show considerable variation in the thickness of the ice over mountainous terrain, where in some areas thicknesses of around 1000m change to zero in a few kilometres. There are variations between the models in the broader distribution of ice thickness. Both the Boulton et al. (1985; 1991) models show a broadly central area of locally very thick ice surrounded by extensive areas of much thinner ice, whereas in the Lambeck (1995) model the area of thickest ice is distributed eccentrically around the Scottish Highlands, with only a narrow area of thinner ice to the north (the Lambeck model is assumed to be of ice surface altitude). The basic parameters of the three models are given in Table 1.

Insert Table 1 here

9 The greatest area covered by ice is depicted in the Boulton et al. (1985) model, with the other two models registering similar areas to each other. Ice thicknesses vary considerably, both in value and location. In the Boulton et al (1985) model, the main areas of ice thickness lie over the SE Grampians (where the greatest value of 946m is achieved), the inner Solway Firth and N and central Ireland. In the Boulton et al. (1991) model, the ice sheet is generally thicker, with the greatest thickness of 1192m achieved over the North Channel, and with considerable thicknesses over Loch Lomond and the inner Solway Firth. The Lambeck (1995) model identifies thickest ice along the western seaboard of Scotland, notably the Firth of Lorn, where a value of 1631m (including 231m of water depth in the cell concerned) is achieved. The ice cover over SW Ireland is probably slightly misplaced in the small scale map of the published Lambeck (1995) model, so the mean thickness indicated will be slightly less.

The ice volume statistics are derived by multiplying the area of each cell in the model by its ice thickness value, then converting into cubic kilometres. Here, the Boulton et al.

(1985) model gives by far the lowest volume, despite exhibiting the greatest extent, while the other two give surprisingly similar values, notwithstanding considerable variations in thickness between each other locally.

Sources of error. The greatest potential error concerns whether the published ice surface altitudes in the three models have taken account of glacio-isostasy, since if that is the case then comparing those altitudes with a Digital Terrain Map of the land surface unadjusted for isostasy would misinterpret ice thickness. However, none of the authors specifically

10 mention that glacio-isostasy has been taken into account, although Boulton et al. in an earlier paper (1984) did refer to the problem. In seeking to evaluate the effects of glacio- isostasy, it is noted that where the ice sheet surface modelled is intersected by nunataks, in the nunatak area at least, the vertical distance between the trimline and nearby valley floor will not have changed, even if the altitude of the trimline has. Thus, although there may be an error attributable to glacio-isostasy in all models, it may not be as large as might have been thought. As an estimate of the likely magnitude of this error, a broad estimate of the amount of glacio-isostatic depression of the land surface at the ice sheet maximum was made using the gradient of the earliest Angus-Kincardine shoreline of

Cullingford and Smith (1980), assuming an uplift pattern as demonstrated in the Main

Postglacial Shoreline of Smith et al. (2006) (see below, Figure 3A). The earliest Angus-

Kincardine shoreline is believed to have been reached not long after ice began to recede from the Wee Bankie moraine (Sissons, 1981), the maximal extent of the LGM ice sheet off eastern Scotland. This provides a deepest ice value for the Boulton et al. (1985) model of 1045m, a mean thickness of 264m and a volume of 106 187 km3 , not greatly different to the figures for this model in Table1, above.

There may additionally have been some errors in interpreting the authors’ published maps, particularly Lambeck’s (1995) map for SW Ireland, as indicated above. Further, a relatively small error will have been due to the fact that the land surface in the Digital

Terrain maps includes sediments which accumulated during or following deglaciation.

These sediments may have exceeded 50m thick in some areas, but will have been highly localised and are not considered here. Broadly, however, taking account of all possible

11 sources of error, it is maintained that the ice thickness estimates are probably close to those implied by the surface altitudes of the three models in Figure 1, perhaps with an error of up to 10%.

Implications of the thickness models

Sea surface change. The approach described here provides the basis for more accurate estimates of sea surface lowering than has been available hitherto. The Denton and

Hughes (1981) model for the LGM yielded a volume of 801,000 km3, a sea level equivalent of 1.9m (using the figures given by Denton and Hughes, ibid.), given the mass of ice implied. The volumes of the Boulton et al. (1985), Boulton et al (1991) and

Lambeck (1995) models are respectively 13.5%, 22% and 21.4% of the Denton and

Hughes (1981) model, yielding sea level equivalents of 0.26m, 0.42m and 0.41m. Since these models are for ice sheets of broadly similar extent to that of the Denton and Hughes

(1981) model (the main difference being that the Denton and Hughes model has a connection with the SIS), the differences are striking.

Glacio-isostasy. The pattern and volume of crustal loading implied by the ice thickness models here raise some interesting questions in terms of glacio-isostasy. Where glacio- hydro-isostatic modelling is concerned, the early isobase models of Lambeck (1991;

1993a, 1993b, 1995), and the generalised relative sea level graphs of Lambeck (ibid.),

Shennan et al. (2000, 2002) and Peltier et al. (2002), which were all based upon “flat bed” models, may require reconsideration.

12 Shoreline-based (“empirical”) isobase modelling is independent of ice sheet models, and comparisons may therefore be made. In Figure 4, shoreline-based isobase models for two

Holocene shorelines in Britain, based upon Gaussian Trend Surface analysis, taken from

Smith et al. (2006), are shown. These are probably the most accurate isobase models currently available for the periods concerned, and concur in identifying a centre of uplift in the SE Grampians. These models will undoubtedly improve with more data, especially from Ireland and northern England, but at present probably represent a general indication of the pattern of uplift after the mid and late Holocene. The loading patterns that the isobase models in Figure 4 imply are closest to the Boulton et al. (1985) ice sheet model, but there are important differences between these isobase models and all ice thickness models: thus the substantial ice thicknesses in SW Scotland shown in the Boulton et al models compare poorly with the isobase models; while the concentration of high ice thicknesses on the western seaboard of Scotland in the Lambeck (1995) model is at variance with the isobase models. However, further detailed comparisons are of limited value because, as Lambeck (e.g. 1995) has shown, patterns of uplift reflect water loading and unloading on the Continental Shelf, together with the effects of glaciation elsewhere, as well as the load imposed by the local ice sheet.

Neotectonics. The association of neotectonic activity with glaciation has been explored in a number of accounts (e.g. Stewart et al., 2000) and it seems possible that the considerable variations in loading identified for the upland areas of Britain and Ireland may have implications for neotectonics. Available seismic maps (e.g. as shown in British

Geological Survey Earthquake Bulletins) show a concentration of instrumentally

13 recorded earthquakes in upland areas, and for the Scottish Highlands a number of authors

(e.g. Sissons, 1972; Sissons and Cornish, 1982; Ringrose, 1989; Firth, 1992; Firth and

Stewart, 2000; Stewart et al., 2001) have cited evidence for neotectonic activity during the Late Devensian and Holocene. The correspondence between variable loading and seismic activity may not be coincidental for many types of earthquakes. However, there appear to be two types of movement recorded. The first occurred along faults which define discrete areas of crust moving with respect to one another, as in the Forth valley, where Sissons (1972) identified the dislocation of two shorelines in stratigraphical succession, yet found no evidence of differential uplift between the two, implying that uplift in the area had occurred as blocks. The second appear to be of movements along individual faults not defining discrete blocks. The question therefore arises as to the respective roles of isostasy and neotectonics in the area of the British-Irish LGM Ice

Sheet, and by implication the significance of the loading pattern shown in the ice thickness models determined in this paper.

Paraglacial processes. The ice thickness models draw attention to considerable variations in loading across glacial troughs, and where, as in the Boulton et al. (1985;

1991) models in particular, thick ice in troughs often contrasts with thin ice over adjacent interfluves, such variations are particularly marked. These contrasts may be of relevance in understanding paraglacial processes. With loading varying by over 1000m vertically in a few hundred metres horizontally across glacial troughs, pressure release may have been a factor during ice sheet decay, accelerating valley erosion and resulting in mass movement and the accumulation of substantial areas of debris at the foot of slopes. In the

14 developing literature on this subject, Wilson (2005) has drawn attention to pressure release taking place following ice sheet decay in the Lake District, for example. The distribution and extent of pressure release effects will be determined by rock structure and slope angle, but may locally have been considerable.

Conclusion

The application of Digital Terrain Mapping for the estimation of the thickness and volume of past Quaternary ice sheets has been outlined for three versions of the Late

Devensian ice sheet at its maximum as an independent ice sheet in Britain and Ireland. It has been shown that not only are there considerable variations between the ice thickneses determined from the three ice sheet models, but that all models yield substantially lower volumes and overall thicknesses than models which take no account of topography.

Where the latter models have been used in estimating sea surface change, such estimates need reconsideration. There may be implications for estimates of glacio-isostasy and neotectonics as well as the origins of paraglacial deposits.

The approach described in this paper is commended to the reader not so much for the implications which can be drawn from specific, previously-published models, but rather for its value in developing improved models for the BIIS through a better understanding of the anatomy of the modelled ice sheet. More widely, application of Digital Terrain

Mapping to models of Quaternary ice sheets globally will assist in moving towards greater detail in determining the nature and effects of their growth and decay.

Acknowledgements

15 DES acknowledges the award of a Leverhulme Emeritus Fellowship. The data used in the construction of the Gaussian trend surfaces in Figure 4 may be found on the website www.scottishsealevels.net. The referees are thanked for their helpful and constructive comments. This paper is a contribution to IGCP 495: Quaternary Land-Ocean

Interactions.

References

Ballantyne, C.K. 1999. Maximum altitude of Late Devensian glaciation on the isles of

Mull and Jura. Scottish Journal of Geology 35, 97-106.

Ballantyne, C.K. and Hallam, G.E. 2001. Maximum altitude of Late Devensian glaciation on South Uist, Outer Hebrides, Scotland. Proceedings of the Geologists’

Association 112, 155-167.

Ballantyne, C.K., McCarroll, D., Nesje, A., Dahl, S.O. and Stone, J.O. 1998. The last ice sheet in NW Scotland: reconstruction and implications. Quaternary Science Reviews 17,

1149-1184.

.Ballantyne, C.K., McCarroll, D. and Stone, J.O. 2006. Vertical dimensions and age of the Wicklow Mountains ice dome, Eastern Ireland, and implications for the extent of the last Irish ice sheet. Quaternary Science Reviews 25, 2048-2058.

Boulton, G.S., Jones, A.S., Clayton, L.M. and Kenning, M.J. 1977. A British ice sheet model and patterns of glacial erosion and deposition in Britain. In Shotton, F.W. British

Quaternary Studies: recent advances. Clarendon Press, Oxford, 231-246.

16 Boulton, G.S., Smith, G.D. and Morland, L.W. 1984. The reconstruction of former ice sheets and their mass balance characteristics using a non-linearly viscous flow model.

Journal of Glaciology 30 (105), 140-152.

Boulton, G.S., Smith, G.D., Jones, A.S. and Newsome, J. 1985. Glacial geology and glaciology of the last mid-latitude ice sheets. Journal of the Geological Society of London

142, 447-474.

Boulton, G.S.. Peacock, J.D. and Sutherland, D.G. 1991. Quaternary. In Craig, G.Y. (ed.)

The Geology of Scotland. Geological Society Special Publication, London, 503-544.

Boulton, G.S., Peacock, J.D. and Sutherland, D.G. 2003. Quaternary. In Trewin, N.H.

Geology of Scotland. Geological Society of London, London, 409-430.

Bowen, D.Q., Phillips, F.M., McCabe, A.M., Knutz, P.C. and Sykes, G.A. 2002. New data for the Last Glacial Maximum in Great Britain and Ireland. Quaternary Science

Reviews 21, 89-101.

Bowen, D.Q. and McCabe, A.M. 2003. Response to: “New data for the Last Glacial

Maximum in Great Britain and Ireland: a Scottish perspective on the paper by Bowen et al. (2002) by A.M. Hall, J.D. Peacock, E.R. Connell. Quaternary Science Reviews 22,

1555-1556.

British Geological Survey 2006. Earthquake Bulletin. http://www.earthquakes.bgs.ac.uk/

Carr, S.J., Holmes, R., van der Meer, J.J. and Rose, J. 2006. . The Last Glacial Maximum in the North Sea Basin: micromorphological evidence of extensive glaciation. Journal of

Quaternary Science 21(2), 131-153.

17 Clapperton, C.M. and Sugden, D.E. 1972. The Aberdeen and Dinnet glacial limits reconsidered. In Clapperton, C.M. (ed.) North East Scotland Geographical Essays,

Aberdeen University, Aberdeen, 5-11.

Clark, C.D., Evans, D.J.A., Khwata, A., Bradwell, T., Jordan, C.J., Marsh, S.H., Mitchell,

W. and Bateman, M. 2004. Map and GIS database of glacial landforms and features related to the last British Ice Sheet. Boreas 33, 359-375.

Clark, J., McCabe, A.M., Schnabel, C., Freeman, S., Maden, C. and Xu, S. 2006. New constraints on the deglaciation of the western margin of the British-Irish ice sheet,

Ireland, from 10 Be dating. European Geosciences Union, Geophysical Abstracts 8,

10272.

Cullingford, R.A. and Smith, D.E. 1980. Late Devensian raised shorelines in Angus and

Kincardineshire, Scotland. Boreas 9, 21-38.

Dahl, S.O., Ballantyne, C.K., McCarroll, D. and Nesje, A. 1996. Maximum altitude of

Devensian glaciation on the Isle of Skye. Scottish Journal of Geology 32, 107-115.

Dawson, A.G. 1993. Ice Age Earth. Routledge, London and New York.

Denton, G.H. and Hughes, T.J. (eds.) 1981. The last great ice sheets. Wiley, New York.

Dury, G.H. 1953. A glacial breach in the north-western Highlands. Scottish Geographical

Magazine 69, 106-117.

Earth Resources Observation and Science (EROS) Data Center. 2005. Shuttle Radar

Topography Mission (SRTM). www.http://landcover.usgs.gov

Eastern Topographics (ETOPO) Data. 2005.

18 Eyles, N., McCabe, A.M. and Bowen, D.Q. 1994. The stratigraphic and sedimentological significance of Late Devensian ice sheet surging in Holderness, Yorkshire, U.K.

Quaternary Science Reviews 13, 727-759.

Firth, C.R. 1992. Loch Ness Shorelines: evidence for isostatic depression during the Loch

Lomond (Younger Dryas) Stadial. In Fenton, C. (ed.) Neotectonics in Scotland: A Field

Guide. University of Glasgow, Glasgow, 72-75.

Firth, C.R. and Stewart, I.S. 2000. Postglacial tectonics of the Scottish glacio-isostatic uplift centre. Quaternary Science Reviews 19, 1469-1493.

Fretwell, P.T., Peterson, I.R. and Smith, D.E. 2004. The use of Gaussian trend surfaces for modelling glacio-isostatic crustal rebound. Scottish Journal of Geology 40, 175-179.

Hall, A.M. 1997. Quaternary Stratigraphy; the Terrestrial Record. In Gordon, J.E. (ed.)

Reflections on the Ice Age in Scotland. Scottish Natural Heritage, Glasgow, 59-71.

Hall, A.M. and Bent, A.J.A. 1990. The limits of the last British ice sheet in northern

Scotland and the adjacent shelf. Quaternary Newsletter 61, 2-12.

Hall, A.M., Peacock, J.D. and Connell, E.R. 2002. New data for the Last Glacial

Maximum in Great Britain and Ireland: a Scottish perspective on the paper by Bowen et al. (2002). Quaternary Science Reviews.

Lamb, A.L. and Ballantyne, C.K. 1998. Palaeonunataks and the altitude of the last ice sheet in the S.W. Lake District, England. Proceedings of the Geologists’ Association 109,

305-316.

Lambeck, K. 1991. Glacial rebound and sea-level change in the British Isles. Terra Nova

3, 379-389.

19 Lambeck, K. 1993a. Glacial rebound of the British Isles – I. Preliminary model results.

Geophysical Journal International 115, 941-959.

Lambeck, K. 1993b. Glacial rebound of the British Isles – II. A high-resolution, high- precision model. Geophysical Journal International 115, 960-990.

Lambeck, K. 1995. Late Devensian and Holocene shorelines of the British Isles from models of glacio-isostatic rebound. Journal of the Geological Society, London 152, 437-

Lambeck, K. 1996. Glaciation and sea-level change for Ireland and the Irish Sea since

Late Devensian/Midlandian time. Journal of the Geological Society, London 153, 853-

Lambeck, K. and Purcell, A.P. 2001. Sea-level change in the Irish Sea since the Last

Glacial Maxximum: constraints from isostatic modelling. Journal of Quaternary Science

16, 497-506.

McCabe, A.M. and Clark, P.U. 2003. Deglacial chronology from County Donegal,

Ireland: implications for deglaciation of the British-Irish ice sheet. Journal of the

Geological Society of London 160, 847-855.

McCabe, A.M., Clark, P.U. and Clark, J. 2005. AMS 14C dating of deglacial events in the

Irish Sea Basin and other sectors of the British-Irish ice sheet. Quaternary Science

Reviews 24, 1673-1690.

McCabe, A.M, Clark, P.U., Smith, D.E. and Dunlop, P. 2007. A revised model for the deglaciation of eastern Scotland. Journal of the Geological Society of London 164, 313-

20 McCarroll, D. and Ballantyne, C.K. 2000. The last ice sheet in Snowdonia. Journal of

Quaternary Science 15, 765-778.

Milne, G.A., Shennan, I., Youngs, B.A.R., Waugh, A.I., Teferle, F.N., Bingley, R.M.,

Bassett, S.E., Cuthbert-Brown, C. and Bradley, S.L. 2006. Modelling the glacial isostatic adjustment of the UK region. Philosophical Transactions of the Royal Society A, 364,

931-948.

Nesje, A. and Dahl, S.O. 1990. Autochthonous blockfields in southern Norway: implications for the geometry, thickness and isostatic loading of the Late Weichselian

Scandinavian ice sheet. Journal of Quaternary Science 5, 225-234.

Paterson, I.B. 1974. The supposed Perth Readvance in the Perth district. Scottish Journal of Geology, 10, 53-66.

Peltier, W.R., Shennan, I., Drummond, R and Horton, B. 2002. On the postglacial isostatic adjustment of the British Isles and the shallow viscoelastic structure of the Earth.

Geophysical Journal International 148, 443-475.

Rae, A.C., Harrison, S., Mighall, T. and Dawson, A.G. 2004. Periglacial trimlines and nunataks of the last glacial maximum: the Gap of Dunloe, southwest Ireland. Journal of

Quaternary Science 19, 87-97.

Ringrose, P.S. 1989. Recent fault movement and palaeoseismicity inwestern Scotland.

Tectonophysics 163, 305-314,

Sejrup, H.P., Aarseth, I., Ellingsen, K.L., Reither, E., Jansen, E., Lǿvlie, R., Bent, A.,

Brigham-Grette, J., Larsen, E. and Stoker, M. 1987. Quaternary stratigraphy of the

Fladen area, central North Sea: a multidisciplinary study. Journal of Quaternary Science

2, 35-58.

21 Sejrup, H.P., Larsen, E., Haflidason, H., Berstad, I. M., Hjeustuen, B.O., Jonsdottir, H.E.,

King, E., Landvik, J., Longva, O., Nygard, A., Ottesen, D., Raunholm, S., Rise, L. and

Stalsberg, K. 2003. Configuration, history and impact of the Norwegian Channel Ice

Stream. Boreas 32, 18-36.

Shennan, I., Lambeck, K., Horton, B., Innes, J., Lloyd, J., McArthur, J., Purcell, T. and

Rutherford, M. 2000. Late Devensian and Holocene records of relative sea level changes in northwest Scotland and their implications for glacio-hydro-isostatic modelling.

Quaternary Science Reviews 19, 1103-1135.

Shennan, I., Peltier, W.R., Drummond, R. and Horton, B.P. 2002. Global to local scale parameters determining relative sea-level changes and the post-glacial isostatic adjustment of Great Britain. Quaternary Science Reviews 21, 397-408.

Shennan, I., Bradley, S., Milne, G., Brooks, A., Bassett, S. and Hamilton, S. 2006.

Relative sea-level changes, glacial isostatic modelling and ice-sheet reconstructions from the British Isles since the Last Glacial Maximum. Journal of Quaternary Science 21(6),

585-599.

Sissons, J.B. 1963. The Perth readvance in central Scotland. Part I. Scottish Geographical

Magazine 79, 151-163.

Sissons, J.B. 1967. The evolution of Scotland’s scenery. Oliver and Boyd, Edinburgh.

Sissons, J.B. 1972. Dislocation and non-uniform uplift of raised shorelines in the western part of the Forth valley. Transactions of the Institute of British Geographers 55, 145-159.

Sissons, J.B. 1981. The last Scottish ice sheet: facts and speculative discussion. Boreas

10, 1-17.

22 Sissons, J.B. and Dawson, A.G. 1981. Former sea levels and ice limits in part of Wester

Ross, north-west Scotland. Proceedings of the Geologists’ Association 12, 115-124.

Sissons, J.B. and Cornish, R. 1982. Rapid localised glacio-isostatic uplift at Glen Roy,

Scotland. Nature 297, 213-214.

Smith, D.E., Fretwell, P.T., Cullingford, R.A. and Firth, C.R. 2006. Towards improved empirical isobase models of Holocene land uplift for mainland Scotland, UK.

Philosophical Transactions of the Royal Society A, 364, 949-972.

Stewart, I.S., Sauber, J. and Rose, J. 2000. Glacio-seismotectonics: ice sheets, crustal deformation and seismicity. Quaternary Science Reviews 19, 1367-1389.

Stewart, I.S., Firth, C.R., Rust, D.J., Collins, P.E.F. and Firth, J.A. 2001. Postglacial fault movement and palaeoseismicity in western Scotland: a reappraisal of the Kinloch Hourn fault, Kintail. Journal of Seismology 5 (3), 307-328.

Stone, J.O. and Balllantyne, C.K. 2006. Dimensions and deglacial chronology of the

Outer Hebrides Ice Cap, northwest Scotland: implications of cosmic ray exposure dating.

Journal of Quaternary Science 21(1), 75-84.

Sutherland, D.G. 1984. The Quaternary deposits and landforms of Scotland and the neighbouring shelves: a review. Quaternary Science Reviews 3, 157-254.

Wilson, P. 2005. Paraglacial rock slope failures in Wasdale, western Lake District,

England: morphology, styles and significance. Proceedings of the Geologists’

Association 116, 349-361.

23 List of Tables

Table 1 Ice sheet parameters Table 1 Ice sheet parameters

Ice volume Mean thickness Ice covered area Deepest ice

______Boulton et al 108 381 km3 270 m 400 679 km2 946 m (1985) Boulton et al 175 932 km3 520 m 338 181 km2 1192 m (1991) Lambeck 171 406 km3 519 m 330 442 km2 1631 m (1995)

List of Figures

Figure 1. Places mentioned in the text.

Figure 2. The ice sheet models of Boulton et al. (1985) (A), Boulton et al. (1991) (B) and

Lambeck, 1995 (C).

Figure 3. Ice thicknesses derived from the ice sheet models in Figure 1 superimposed upon a digital terrain model of the land surface beneath.

Figure 4. Isobase models for (A), Main Postglacial Shoreline, of circa 6400-7700 sidereal years BP and (B), the Blairdrummond Shoreline, of circa 4500-5800 sidereal years BP, after Smith et al. (2006).

25 26 27