Late Readvance and Rapid Final Deglaciation of the Last Ice Sheet in the Grampian Mountains, Scotland

Edinburgh Research Explorer

Late readvance and rapid final deglaciation of the last ice sheet in the Grampian Mountains, Scotland.

Citation for published version: Hall, AM, Binnie, S, Sugden, D, Dunai, T & Wood, C 2016, 'Late readvance and rapid final deglaciation of the last ice sheet in the Grampian Mountains, Scotland.', Journal of Quaternary Science. https://doi.org/10.1002/jqs.2911

Digital Object Identifier (DOI): 10.1002/jqs.2911

Link: Link to publication record in Edinburgh Research Explorer

Document Version: Peer reviewed version

Published In: Journal of Quaternary Science

General rights Copyright for the publications made accessible via the Edinburgh Research Explorer is retained by the author(s) and / or other copyright owners and it is a condition of accessing these publications that users recognise and abide by the legal requirements associated with these rights.

Take down policy The University of Edinburgh has made every reasonable effort to ensure that Edinburgh Research Explorer content complies with UK legislation. If you believe that the public display of this file breaches copyright please contact [email protected] providing details, and we will remove access to the work immediately and investigate your claim.

Download date: 08. Oct. 2021 Journal of Quaternary Science

Rapid d eglaciation of the last ice sheet in the Grampian Mountains, Scotland.

Journal: Journal of Quaternary Science

Manuscript ID Draft

Wiley - Manuscript type: Research Article

Date Submitted by the Author: n/a

Complete List of Authors: Hall, Adrian; Stockholm University, Physical Geography Binnie, Steven; Universitat zu Köln, Institut für Geologie und Mineralogie Sugden, David; University of Edinburgh, Institute of Geography (School of Geosciences) Dunai, Tibor; Universitat zu Köln, Institut für Geologie und Mineralogie Wood, Christina

Deglaciation, British-Irish ice sheet, Strath Spey, 10Be cosmogenic Keywords: exposure ages, climate

http://mc.manuscriptcentral.com/jqs Page 1 of 27 Journal of Quaternary Science

1 2 3 4 1 TITLE 5 6 2 Rapid deglaciation of the last ice sheet in the Grampian Mountains, Scotland. 7 8 9 3 AUTHORS 10 11 4 Adrian M. Hall 1, Steven A. Binnie 2, David Sugden 3, Tibor Dunai 2, Christina Wood 4 12 13 14 15 5 ADDRESSES 16 17 6 1 Department of Physical Geography, Stockholm University, S-10691 Stockholm, Sweden. 18 2 19 7 University of Cologne, Institute of Geology and Mineralogy, Cologne, North Rhine- 20 21 8 Westphalia, Germany. 22 3 23 9 Geography, School of Geosciences, University of Edinburgh, Scotland. 24 25 10 4 Scottish Natural Heritage, Inverness, Scotland. 26 27 11 28 29 12 1 Corresponding author: [email protected] 30 31 13 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 1 59 60 http://mc.manuscriptcentral.com/jqs Journal of Quaternary Science Page 2 of 27

1 2 3 14 ABSTRACT 4 5 6 15 Towards the end of the Late Devensian glaciation, ice sourced from the Western Grampian 7 16 Mountains of Scotland flowed down Strath Spey to encroach on the northern flanks of the 8 17 Cairngorm Mountains. The maximum of this late advance and its subsequent retreat is 9 10 18 recorded by moraines, ice-marginal meltwater channels, and kame terraces that can be 11 19 traced up Strath Spey for 60 km from its down-valley limit. New cosmogenic 10 Be exposure 12 20 ages from moraines indicate deglaciation at 15.0 ± 0.6 ka. This timing matches closely 13 14 21 existing age data for deglaciation of the eastern Cairngorm plateau and for other sites along 15 22 Strath Spey. The implication is that active ice retreat from the flanks of Strath Spey occurred 16 17 23 within the ~1 ka uncertainty of the cosmogenic exposure ages. Initial ice retreat occurred 18 24 under cold conditions at the end of Greenland Stadial 2a (GS-2a) (16.9-14.7 ka). We suggest 19 25 the advance followed the collapse of the marine parts of the British-Irish Ice Sheet at ~ 16ka. 20 21 26 The combination of increased precipitation at a time of low temperatures would cause an 22 27 advance. The rapidity of deglaciation may reflect enhanced Föhn effects caused by the ice 23 28 dome in the Western Highlands. 24 25 KEY WORDS 26 29 27 28 30 Deglaciation; British-Irish ice sheet; Strath Spey; 10 Be cosmogenic exposure ages; climate. 29 30 31 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 2 59 60 http://mc.manuscriptcentral.com/jqs Page 3 of 27 Journal of Quaternary Science

1 2 3 32 INTRODUCTION 4 5 6 33 The sequence and timing of events during deglaciation of the last ice sheet in Scotland 7 34 remain poorly constrained. Strath Spey forms a major topographic corridor through the 8 35 Grampians, flanked by the Monadhliath and Cairngorm Mountains (Fig. 1). During the 9 10 36 Pleistocene, powerful flows of ice were directed along Strath Spey from Rannoch Moor, the 11 37 major snow and ice accumulation basin in the western Highlands for the Scottish ice sheet, 12 38 and towards the inner Moray Firth, one of the major outlets of the ice sheet to the North 13 14 39 Sea basin. The Cairngorm granite massif supported an independent ice centre. During the 15 40 last, Late Devensian glaciation, the carry of schist erratics shows that Strath Spey ice 16 17 41 encroached to elevations above 800 m on the northern flanks of the Cairngorms (Sugden, 18 42 1970; Hall and Phillips, 2006). This erratic limit is associated with the highest set of a 19 43 remarkable, off-lapping assemblage of ice-marginal landforms found on the mountain flanks 20 21 44 that includes lateral moraines, kame terraces and meltwater channels (Fig. 1B) (Hinxman 22 45 and Anderson, 1915; Gordon, 1993). The ice-marginal forms cover an altitudinal range of 23 46 ~500 m and can be traced for ~60 km along Strath Spey (Fig. 1A). Hence this landform 24 25 47 assemblage provides an unusual opportunity to establish the pattern and timing of the final 26 48 retreat of a major ice lobe within the last Scottish ice sheet. 27 28 49 Three conflicting models currently exist for deglaciation in Strath Spey in which the ice lobe 29 30 50 (i) decayed passively, with slow down-wasting and final downwasting leading to the 31 51 generation of huge volumes of meltwater (Sugden, 1970; Young, 1975), (ii) remained active 32 52 during a slow retreat interrupted by long stillstands when large ice-dammed lakes 33 34 53 developed at the ice margin (Hinxman and Anderson, 1915; Brazier et al., 1998; Everest and 35 54 Kubik, 2006) and (iii) remained active but retreated rapidly with only brief stillstands. These 36 55 conflicting models exist in part because of conflicting dating evidence. While cosmogenic 37 38 56 exposure ages for boulders on moraines appear to support a prolonged stillstand at ca. 39 57 14.5–14.0 ka (Everest and Kubik, 2006) ages obtained on postglacial rockslides (Ballantyne 40 58 et al., 2009) and basal radiocarbon dates (Sissons and Walker, 1974; Everest and Kubik, 41 42 59 2006) indicate that deglaciation was complete before ca. 15.5–15.0 ka (Ballantyne, 2010). 43 60 This study provides a revised interpretation of the landform assemblage associated with the 44 61 Strath Spey ice lobe in the northern Cairngorms as the product of a late advance of the last 45 10 46 62 ice sheet and of its subsequent down wastage. The study reports new cosmogenic Be 47 63 exposure ages for boulders from moraines in the northern Cairngorms that constrain the 48 49 64 time interval for deglaciation to the close of Greenland Stadial 2a (GS-2a) (16.9-14.7 ka). 50 51 65 STUDY AREA 52 53 66 The study area forms part of the northern flank of the Cairngorm Mountains and includes 54 55 67 the prominent ridge of Sròn a’ Cha-no that separates Glen More from Strath Nethy (Fig. 2). 56 68 The high ground is developed in the Caledonian Cairngorm Granite, with Moine psammitic 57 58 3 59 60 http://mc.manuscriptcentral.com/jqs Journal of Quaternary Science Page 4 of 27

1 2 3 69 schists forming the floor of the Glen More basin (Fig. 2). A remarkable assemblage of ice- 4 70 marginal features is developed along the flanks of Strath Spey (Barrow et al., 1913; Hinxman 5 6 71 and Anderson, 1915; Charlesworth, 1956), marking the final up-valley retreat of the Spey 7 72 glacier and its attendant down wasting (Figs. 2 and 3). Many ridges that project towards the 8 73 valley axis are notched by rock-cut meltwater channels (Bremner, 1934a; Gheorghiu et al., 9 10 74 2012). Re-use of many of these channels during deglaciation as routeways for ice-marginal 11 75 meltwater is indicated where the channels are in sub-parallel sets, with channel floors 12 13 76 sloping down valley, and associated with nearby ice-marginal features such as moraines and 14 77 kame terraces (Hinxman and Anderson, 1915; Greenwood et al., 2007). Many tributary 15 78 valleys entering Strath Spey from the Cairngorms and Monadhliath were blocked by ice 16 17 79 lobes that built up moraines and dammed meltwater in ice-marginal lakes and ponds 18 80 (Hinxman and Anderson, 1915; Bremner, 1934a; Brazier et al., 1998; Boston, 2013). Off- 19 81 lapping, ice-marginal landforms can be traced along hill slopes for long distances on both 20 21 82 the Cairngorm (Fig.2) and Monadhliath flanks of Strath Spey. The ice-marginal landforms 22 83 descend in altitude towards the north-east towards likely ice limits marked by moraines at 23 84 the confluence of the Avon and Livet and north-east of Grantown (Bremner, 1934b). 24 25 26 85 Within the study area, ice-marginal features record the extent of Strath Spey ice and its 27 86 subsequent unzipping from Cairngorm ice (Fig. 4). The maximum of the advance is recorded 28 87 by a trimline, with stripped granite surfaces below, and a small lateral meltwater channel at 29 30 88 832 m on Sròn a’ Cha-no (Figs. 1B and 2B). The sharpness of the morphological contrasts 31 89 above and below the trimline, the lateral continuity and the relationship to ice flow from 32 90 the west all point to the limit representing an advance along Strath Spey. In Coire Laogh Mòr 33 34 91 and southwards towards Coire na-Ciste, lateral moraine ridges and channels and an upper 35 92 limit for schist erratics occur at the slightly lower elevation of 810 m. From the col between 36 93 Sròn a’ Cha-no and Stac na h-Iolaire, sloping, schist-bearing moraines extend into Strath 37 38 94 Nethy (Fig. 3A). Lateral moraines from both Spey and Nethy ice meet near the base of the 39 95 western wall of Strath Nethy and demonstrate the contemporaneity of the two ice masses 40 96 (Fig. 4). Extensive granite boulder accumulations derived from rock slope failures also occur 41 42 97 on the valley floor in upper Strath Nethy (Ballantyne et al., 2009). The lobate form of ridges 43 98 on the surface of the boulder accumulations suggests deposition on to glacier ice, followed 44 45 99 by glacial transport a short distance down the valley (Sugden and Clapperton, 1975). 46 100 Residual Nethy ice on the valley floor also may have supported the eastern flanks of large, 47 101 steeply-sloping fan deltas that extend downslope from the col (Fig. 3A). Horizontal benches 48 49 102 cut in bedded gravel and sand on lower valley slopes (Fig. 3A) indicate that Spey ice blocked 50 103 the outlet of Strath Nethy and so ponded ephemeral lakes between the separating ice 51 104 margins. The fan deltas were fed by two main sets of meltwater channels in the col, a 52 53 105 southern set with an intake at 725 m and a deeper northern set with an intake at 650 m (Fig. 54 106 3B). The moraine M1 at 720 m marks the presence of Spey ice at a time when the lower 55 107 channels continued to pass meltwater through the col. After the ice surface dropped below 56 57 58 4 59 60 http://mc.manuscriptcentral.com/jqs Page 5 of 27 Journal of Quaternary Science

1 2 3 108 the level of the Iolaire col, meltwater was diverted northward and sequences of lateral 4 109 channels, some associated with the M2 and M3 moraines at 600 and 540 m respectively 5 6 110 (Fig. 3B), mark the down wasting of the ice margin in Glen More. 7 8 111 DATING METHODS 9 10 11 112 In order to constrain the age of the deglaciation in the study area, samples were collected 12 113 for cosmogenic exposure dating from seven large boulders, including sites on moraines M1- 13 114 M3, and two bedrock sites at elevations of 741 to 540 m O. D. on the western slopes of Sròn 14 15 115 a’ Cha-no (Fig. 1B )(Table 1). 16 17 116 10 BE SAMPLE PREPARATION AND MEASUREMENT 18 19 117 Whole rock samples were crushed then sieved and the magnetic fraction removed from the 20 21 118 250 – 710 um fraction before being prepared as AMS (Accelerator Mass Spectrometry) 22 119 targets following the standard procedures outlined in Binnie et al. (2015). The targets were 23 120 measured at CologneAMS (Dewald et al., 2013), normalized to the standards of Nishiizumi 24 10 25 121 et al. (2007). Be concentrations shown in Table 2 include the subtraction of reagent blanks 26 122 prepared alongside the samples. Analytical uncertainties are derived by summing in 27 28 123 quadrature the 1 σ uncertainty of the AMS results with an estimated 1% uncertainty in the 29 124 mass of the 9Be spike of both the samples and blanks. Exposure ages are estimated using 30 125 the CRONUS-Earth calculator version 2.2, using version 2.2.1 of the constants file and Balco 31 32 126 et al. (2008) and the Dunai (2001) ‘Du’ scaling. We used a local production rate of calibrated 33 127 from a Lateglacial Interstade age for moraines in NW Scotland (NWHLPR11.6; Ballantyne 34 128 and Stone, 2012). This production rate (4.39 ± 0.15 atoms/g/a at HLSL using Dunai, 2001, 35 36 129 scaling) is in good agreement with the estimate of Small and Fabel (2015) from Glen Roy, 37 130 based on the 10 Be ages of palaeoshorelines independently dated using a varve chronology. 38 131 Where other, previously published ages are compared with our results, they have been 39 40 132 recalculated using the local production rate calibration data set NHWLPR11.6 of Ballantyne 41 133 and Stone (2012) and scaled according to Dunai (2001). Where the recast ages are reported, 42 134 we note the erosion rate (e) used in the age derivation. Sample site coordinates and 43 44 135 elevations were recorded using handheld GNSS. Topographic shielding factors for each 45 136 sample site were measured in the field. We assumed a bulk rock density of 2.6 g/cm 3 and 46 47 137 employed the std pressure flag and 07KNSTD AMS standard flag in the CRONUS-Earth 48 138 calculator (Table 2). 49 50 139 Erosion rates. For CG12-004 the erosion rate is effectively zero because the surface of the 51 140 projecting quartz vein was glacially polished. At this site, the projection of the quartz vein 52 53 141 due to postglacial erosion indicates erosion rates on granite of 4.1 mm/ka. This is a high rate 54 142 compared to the average rate of erosion on the Cairngorm Granite of 1.6 ± 0.6 mm/ka 55 56 143 Phillips et al., (2006) derived from upstanding quartz veins (n=74). The high erosion value 57 58 5 59 60 http://mc.manuscriptcentral.com/jqs Journal of Quaternary Science Page 6 of 27

1 2 3 144 reflects weakening of the granite due to late stage hydrothermal alteration near the schist 4 145 contact. Other samples were collected from boulders with macroscopically fresh surfaces 5 6 146 and the average rate of erosion of 1.6 mm/ka is applied to those samples. 7 8 147 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 6 59 60 http://mc.manuscriptcentral.com/jqs Page 7 of 27 Journal of Quaternary Science

1 2 3 148 RESULTS 4 5 10 6 149 The results of surface exposure dating using Be are summarized in Table 2. Age 7 150 uncertainties at ± 1σ are given as both internal uncertainties, for between-sample 8 10 151 comparison, and total (external) uncertainties. The latter incorporate Be production rate 9 10 152 uncertainties, and are appropriate for comparison with exposure ages obtained in other 11 153 studies, and with radiocarbon dates. The ages presented below are those using the 12 154 production rate scaling of Dunai (2001) and total uncertainties are employed throughout. 13 14 15 155 The exposure ages exhibit significant within-site variation. Boulders at or above the M1 16 156 moraine elevation include an outlier sample, CG14-001, with a exposure age of 24.0 ± 1.3 17 157 ka, which is interpreted as reflecting nuclide inheritance. Sample CG12-004 comes from a 18 19 158 glacially-polished quartz vein surface that has experienced no postglacial erosion and 20 159 provided an exposure age of 16.6 ± 1.0 ka. This date and the overlapping age on sample 21 160 CG14-02 appear to require implausibly early deglaciation of an area close to one of the main 22 23 161 centres of the last Scottish ice sheet and so may mark an earlier phase of glacial erosion or, 24 162 alternatively, include slight nuclide inheritance. The three overlapping youngest age ranges 25 163 from the elevation of the M1 moraine comprising two boulder samples and one bedrock 26 27 164 sample together give an error weighted mean age of 14.4 ± 0.5 ka. The significant age 28 165 difference with CG12-004 and CG14-02 make it unlikely that these ages relate to the same 29 30 166 event. A single boulder on the M2 moraine provided an exposure age of 15.8 ± 0.9 ka. Two 31 167 boulders on the M3 moraine give a mean age of 14.3 ± 0.6 ka. Error weighted mean ages for 32 168 moraine ridges M1 and M3 between 720 and 540 m are the same and indicate that ice 33 34 169 down wastage by ≥200 m occurred at 15.0 ± 0.6 ka. 35 36 170 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 7 59 60 http://mc.manuscriptcentral.com/jqs Journal of Quaternary Science Page 8 of 27

1 2 3 171 DISCUSSION 4 5 6 172 The landform assemblage found in the study can be traced for ~60 km along Strath Spey 7 173 (Fig. 5). The maximum limit down-valley is probably marked by the large moraines at the 8 174 Avon-Livet confluence and in the Allt Bhreac, north-east of Grantown (Bremner, 1934b). The 9 10 175 huge outwash terraces at Bridge of Avon (Bremner, 1934b) were also probably constructed 11 176 as meltwater carried debris down the Avon from the ice margin. The M1 limit along the 12 177 eastern side of the Dorback and Abernethy basins is clearly delineated by ice marginal 13 14 178 channels, moraines and kame terraces (Hinxman and Teall, 1896; Hinxman and Anderson, 15 179 1915; Bremner, 1934a) (Fig. 2). The absence of moraines and channels east of the limit here 16 17 180 and the discharge of meltwater through a col at 700 m into the Water of Caiplich (Fig. 2) 18 181 indicate that Cairngorm ice had receded towards the Cairngorm plateau before Spey ice 19 182 advanced to its maximum limit. This is consistent with morphological evidence from 20 21 183 moraines on the flanks on the Sròn a’ Cha-no being ice-free at this stage (Fig. 4). Further 22 184 south, Strath Spey and Cairngorm ice were probably initially in contact at the line marked by 23 185 the upper limit of schist erratics on the flanks of the northern Cairngorms (Fig 2). The 24 25 186 development of the large moraines and ice marginal lakes in the Lairig Ghru and Gleann 26 187 Einich occurred after unzipping of the two ice masses. Moraine M3 marks a former ice 27 188 margin that can be traced up-valley to large moraines found at slightly higher elevations 28 29 189 west of the ski car park above Glen More (Fig. 3C) and at the mouth of the Lairig Ghru. The 30 190 M3 moraine however stands above the level of the moraines in Gleann Einich (Fig. 5). 31 191 Marginal meltwater channels on the Monadhliath flank document the further retreat of ice 32 33 192 up Strath Spey to beyond Laggan (Young, 1977; Young, 1978; Merritt et al., 2013) (Fig. 5). 34 35 193 Exposure ages from the study area can be compared to existing cosmogenic data. On the 10 36 194 Cairngorm summits. a Be exposure age for an erratic boulder at an altitude of 1156 m in 37 38 195 the eastern Cairngorms, ~14 km east of the study area, indicates deglaciation at 15.9 ± 1.0 39 196 ka (e=1.6 mm/ka; recalculated from Phillips et al. (2006). Whilst this age overlaps slightly the 40 197 ages for moraines from the study area, the age also allow final deglaciation of the 41 42 198 Cairngorm tops to have occurred up to 1-2 ka before the final ice advance down Strath Spey. 43 199 Granite boulders from the floor of upper Strath Nethy have provided recalculated 10 Be 44 200 weighted mean ages, excluding outliers, of 16.9 ± 0.8 ka (e=0; recalculated from Ballantyne 45 46 201 and Stone, 2013). The boulders accumulated by run out from major rock slope failures of 47 202 cliffs onto the surface of glacier ice in upper Strath Nethy. Sloping, schist-bearing moraines 48 49 203 extend from the M1 limit into Strath Nethy towards the margin of these boulder 50 204 accumulations. Exposure ages for the boulder accumulations are older than for the M1 51 205 moraine, suggesting that rock slope failure in Strath Nethy occurred before abandonment of 52 53 206 the M1 ice limit and supporting earlier deglaciation of the northern Cairngorm plateau. 54 10 55 207 Everest and Kubik (2006) provided Be exposure ages on granite boulders resting on the 56 208 Glen More and Gleann Einich moraines. As in the study area, a range of ages was obtained, 57 58 8 59 60 http://mc.manuscriptcentral.com/jqs Page 9 of 27 Journal of Quaternary Science

1 2 3 209 individual sample exposure ages ranging from 17.3 ± 1.4 ka to 12.9 ± 1.1 ka (e=0; 4 210 recalculated from Everest and Kubik, 2006) (Fig.4). The mean ages of 15.0 ± 0.7 ka for the 5 6 211 Glen More moraine and of the 15.3 ± 0.7 ka for the Gleann Einich moraine however are 7 212 statistically equivalent to mean ages of 15.0 ± 0.6 ka from moraines M1 and M3 in the study 8 213 area (e=0; recalculated from Everest and Kubik, 2006). Rock slope failures in the Lairig Ghru 9 10 214 occurred soon after meltwater flow had ceased through the Chalamain Gap meltwater 11 215 channel with a floor that stands below the projected elevation of M1 limit. Surface boulders 12 13 216 have provided recalculated total-uncertainty weighted mean ages of 16.0 ± 0.7 ka (e=0; 14 217 recalculated from Ballantyne and Stone, 2013), again suggesting that rock slope failure 15 218 occurred before the M1 readvance. OSL dates for lake sediments between moraines 16 17 219 marking limits of Cairngorm and Spey ice in Glen Einich are significantly older at 16.7 ± 0.5 10 18 220 and 16.7 ± 5.4 cal ka BP (Everest and Golledge, 2004). An error weighted mean Be 19 221 exposure age for boulder on a moraine at an elevation of 362 m south of Newtonmore 20 21 222 indicates deglaciation at 14.1± 0.6 ka (e=0; recalculated from Gheorghiu and Fabel, 2013). 22 223 Two radiocarbon dates obtained for basal organic sediments in a kettle hole at Loch 23 224 Etteridge (Fig. 1), 30-35 km up-valley from the study area (Sissons and Walker, 1974; Everest 24 25 225 and Golledge, 2004) yielded an uncertainty-weighted mean age of 15.6 ± 0.3 cal. ka BP. 26 226 Tephrostratigraphic evidence from the same site implies deglaciation prior to 14.0 ka and 27 227 probably 14.5 ka (Lowe et al., 2008a). A kettle hole lake in Abernethy Forest (Fig. 1), 10 km 28 29 228 north-west of the study area, has also provided a basal radiocarbon date of 14,764 ± 384 30 229 cal. a. (Vasari and Vasari, 1968). These radiocarbon dates postdate the final disappearance 31 32 230 of glacier ice on the valley floor at these sites and so add support to the results from 33 231 cosmogenic exposure dating. 34 35 232 Comparison of the dates in the study area with the detailed climatic record elaborated in 36 233 the Greenland ice cores is shown in Figure 6. Advance to the M1 limit and subsequent 37 38 234 deglaciation occurred towards the close of Greenland Stadial 2a (GS-2a) (16.9-14.7 ka) 39 235 (Björck et al., 1998). Both the cosmogenic exposure ages on moraines and the radiocarbon 40 236 ages of organic matter in kettle holes show that that deglaciation was already well 41 42 237 underway by the start of the Lateglacial (Windermere) Interstadial at 14.7 ka. Indeed, the 43 238 deglaciation occurred during a hemispheric climate that was relatively cool, if oscillating. 44 45 239 The period of deglaciation at ~15.0 ka also appears to fall entirely within the 1.0 ka 10 46 240 uncertainties of the Be cosmogenic exposure ages and so indicates that final ice retreat 47 241 was rapid. During this period, the ice margin on the flanks of Strath Spey was lowered from 48 49 242 elevations of 830 m to <360 m and retreated up-valley over a distance of more than 60 km, 50 243 requiring ice front recession at average rates of >60 m/a. 51 52 244 Rapid deglaciation requires that any stillstands of the ice margin were brief. In such a case it 53 245 is necessary to explain the large volumes of sediment found in embayments along the 54 55 246 south-eastern flank of Strath Spey. In lower Gleann Einich and in the Lairig Ghru, stream 56 247 sections in large terraces expose thick sequences of glacial lake sediments that include thin, 57 58 9 59 60 http://mc.manuscriptcentral.com/jqs Journal of Quaternary Science Page 10 of 27

1 2 3 248 horizontal beds of sand and silt (Brazier et al., 1998). The large numbers of couplets, when 4 249 interpreted as annual varves, suggest that Strath Spey ice blocked these valley exits for up 5 6 250 to 1 ka (Everest and Kubik, 2006). Varve layers apparently amounting to several centuries of 7 251 lake-bottom sediment accumulation have also been recognized in glacilacustrine sediment 8 252 sequences on the Monadhliath flank of Strath Spey (Phillips and Auton, 2013). 9 10 253 Interpretation of some of these couplets as varves however has been disputed (Brazier et 11 254 al., 1998) and poor exposure makes relationships to contemporaneous ice fronts difficult to 12 13 255 assess. To derive these large sediment volumes, it is unlikely that bedrock erosion beneath 14 256 the glaciers at the time was a significant component as hillslopes between embayments 15 257 show only small, intermittent moraine ridges. Factors favouring the build-up of thick 16 17 258 sediments include an ice surface slope that directed ice flow towards the eastern flank of 18 259 Strath Spey ice and ice flow lines that focused sediment supply towards embayments 19 260 (Brazier et al., 1998; Golledge, 2002). The thicknesses of several tens of metres of glacifluvial 20 21 261 and glacilacustrine gravel and sand found in many embayments probably requires however 22 262 the reworking or over-riding of older sediments. Stream sections in large moraines and 23 263 kame terraces in Abernethy Forest (Bremner, 1934a), below the Cairngorm Ski Car Park 24 25 264 (Sugden, 1970) (Fig. 2C) and in Gleann Einich (Brazier et al., 1998; Golledge, 2002) each 26 265 show a broadly similar, two-part stratigraphy. These sections are typically capped by several 27 266 metres of till that overlies much thicker sequences of bedded gravel, sand and diamicton. 28 29 267 Glacial deformation of the underlying sediment sequences (Golledge, 2002) and 30 268 incorporation of rounded gravel- and sand-rich debris into the upper till are recurrent 31 32 269 features, indicating that the minor ice advances to the limits marked by the moraine ridges 33 270 each took place across thick, pre-existing sediments. 34 35 271 Rapid deglaciation occurred mainly as the Strath Spey ice lobe remained active. The off- 36 272 lapping assemblages of moraines, marginal channels and kame terraces show that ice and 37 38 273 meltwater continued to flow along and towards the flanks of Strath Spey even as the ice 39 274 front retreated up-valley. In embayments such as that drained by Allt Bheadhair, east of the 40 275 Nethy basin (Fig. 2), off-lapping sets of arcuate ridges occur at elevations between 300 and 41 42 276 700 m. Average ridge spacing is ~100 m and comparison with the mean rate of ice retreat of 43 277 >60 m/a. proposed in this study allows the possibility that the ridges represent annual 44 45 278 retreat moraines. Exceptions to the general model of active retreat probably include the ice 46 279 in upper Strath Nethy that was left stagnant once the col at The Saddle (803 m) cut off the 47 280 supply of ice from Glen Avon (Hinxman and Anderson, 1915). Also, the eskers, kames and 48 49 281 kettle holes mapped on the floor of Strath Spey probably were formed in the final stages of 50 282 ice-wastage (Hinxman and Anderson, 1915; Young, 1974, 1975; Young, 1978). Extensive 51 283 dead ice held in the basins of mid Strath Spey is a feature of recent simulations of the final 52 53 284 stages of decay of the last ice sheet in Scotland (Hubbard et al., 2009). 54 55 285 The dating constraints provided here provide a test for simulations of the last Scottish ice 56 286 sheet that use a high-resolution glaciological model derived from by the NGRIP ice-core 57 58 10 59 60 http://mc.manuscriptcentral.com/jqs Page 11 of 27 Journal of Quaternary Science

1 2 3 287 record time-series and modern climate gradients (Hubbard et al., 2009). Experiment 4 288 E109b2bc suggests that, from 19.4 ka onwards, there was active, dynamic and extremely 5 6 289 rapid thinning and retreat in the present offshore area of the outer Moray Firth in response 7 290 to sustained warmer climatic conditions. Century to millennial scale oscillations reflected in 8 291 the ice core are sufficient to drive cycles of glacier advance and retreat in the maritime 9 10 292 Scottish ice sheet. A period of relative cooling after 17.4 ka stabilises a much reduced, 11 293 thinner and mainly cold-based ice sheet over much of the uplands of north-east Scotland. By 12 13 294 15.7 ka, continued warming further reduces cover to an ice cap limited to the Western 14 295 Highlands and the Cairngorms. After a final advance at 15.2 ka, ice cover is rapidly reduced, 15 296 and the simulated extent of ice in Strath Spey fits that identified in this study by 14.6 ka. It is 16 17 297 tempting to relate the field and dating evidence from our study site and its wider environs 18 298 to this latter phase of advance and rapid retreat. However, there is an important caveat. 19 299 Although the simulated timing is within error of the date of deglaciation proposed here, the 20 21 300 regional ice cover depicted in the glaciological model is much greater than indicated by our 22 301 morphological evidence for separation of Spey and Cairngorm ice by 15.0 ka. 23 24 302 Few other dating constraints exist for this period of deglaciation in the Grampian Highlands. 25 26 303 The upper Dee valley became ice-free before 14.6 ka (Huntley, 1994), requiring that a series 27 304 of ice margins identified down-valley (Fig. 1) are older features (Brown, 1993). A significant 28 305 readvance of ice in the inner Moray Firth, the Elgin Oscillation (Fig. 1), occurred before ~14.9 29 30 306 ka (Peacock et al., 1968; Merritt et al., Submitted) at a time when large parts of lower Strath 31 307 Spey were already ice-free. It is likely that the ice advance in mid-Strath Spey documented 32 308 here was broadly coeval with this event. Reconstruction of the pattern of ice retreat for the 33 34 309 entire British and Irish Ice Sheet places ice limits at 15 ka close to those proposed here (Clark 35 310 et al., 2012). Rapid deglaciation between 15 and 16 ka is also indicated by cosmogenic 36 311 exposure ages for the southern parts of the Fennoscandian (Larsen et al., 2012) and 37 38 312 Laurentide (Davis et al., 2015) ice sheets. 39 40 313 Why should there have been an advance and rapid deglaciation at a time when the ice cores 41 314 suggest a relatively cool, if oscillating, climate coinciding with the Greenland Stadial 2a? One 42 43 315 possibility acknowledged by Hubbard et al (2009) is that the lag between glacier response 44 316 and climate forcing on century to millennial time scales is that the glaciers are usually out of 45 317 phase with climate. Following a lagged advance, the glacier finds itself at the maximum of 46 47 318 an advance at a time of relative warmth and thus is vulnerable to rapid melting. An 48 319 alternative hypothesis is that there were regional factors in Scotland that dominated the 49 320 hemispheric climate signal, a point recognised by Clark et al. (2012) when commenting on 50 51 321 the contrasts between the empirical record of ice sheet decay in the UK and predictions of 52 322 models. The British-Irish Ice Sheet saw the collapse of its marine portions between 19 and 53 323 17ka and the ice sheet was confined to the mainland by 16 ka (Clark et al, 2012). The latter 54 55 324 situation would be especially important for mainland Scotland in that warmer Atlantic water 56 325 would replace ice immediately offshore to the west. Given westerly sources of precipitation, 57 58 11 59 60 http://mc.manuscriptcentral.com/jqs Journal of Quaternary Science Page 12 of 27

1 2 3 326 one would expect this new proximity to the sea to lead to a sharp increase in snowfall on 4 327 the ice dome over Rannoch Moor. Given a climate that was still cool, the result would be an 5 6 328 advance of land-based glaciers on the eastern margin of the ice sheet. But in time, overall 7 329 hemispheric warming would triumph and lead to deglaciation. Perhaps also the rapidity of 8 330 deglaciation reflects the location of the Spey ice lobe on the lee side of the ice sheet where 9 10 331 it was subject to enhanced melting by Föhn winds, a situation that is characteristic of the 11 332 comparable climate and topography of ice caps in Patagonia (Hulton and Sugden, 1995; 12 13 333 Casassa et al., 2006). 14 15 334 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 12 59 60 http://mc.manuscriptcentral.com/jqs Page 13 of 27 Journal of Quaternary Science

1 2 3 335 CONCLUSIONS 4 5 6 336 Towards the close of the Late Devensian glaciation, an ice lobe advanced down Strath Spey 7 337 to limits north of Grantown. Early unzipping from Cairngorm ice along the eastern margin of 8 338 the lobe is recorded by ice-marginal meltwater channels, kame terraces and moraines found 9 10 339 between Glen More and Strath Nethy. Correlative ice margins can be traced from ice- 11 340 marginal features for 60 km up-valley into the Lairig Ghru and Gleann Einich and further on 12 341 the Monadhliath flank of Strath Spey. 13 14 10 15 342 New cosmogenic Be exposure ages for boulder samples for moraines at 740-540 m O. D. in 16 343 the northern Cairngorms indicate deglaciation at 15.0 ± 0.6 ka. This timing matches existing 17 344 age data for deglaciation of other sites along Strath Spey. Active ice retreat from the central 18 19 345 Grampians was rapid and occurred within the ~1 ka uncertainty of the cosmogenic exposure 20 346 ages. Comparison with Greenland ice core records indicates that ice advance and the onset 21 347 of final ice retreat occurred at the close of Greenland Stadial 2a (GS-2a) (16.9-14.7 ka). 22 23 348 Perhaps the advance was a response to increased precipitation following the loss of the 24 349 marine parts of the British and Irish Ice Sheet west of the Scottish mainland. 25 26 ACKNOWLEDGEMENTS 27 350 28 29 351 Field work in the Cairngorms for AMH was supported by the Carnegie Trust for the 30 352 Universities of Scotland. Colin Ballantyne provided valuable commentary and kindly 31 10 32 353 provided the Be production rate calibration data set we used here. Jez Everest is thanked 33 354 for helping us recalculate his previously published data. 34 35 355 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 13 59 60 http://mc.manuscriptcentral.com/jqs Journal of Quaternary Science Page 14 of 27

1 2 3 356 FIGURE LIST 4 5 6 357 Figure 1. Location map of the Grampian Mountains. The maximum extent of the Strath Spey 7 358 ice lobe is shown. For comparison, the limits are shown of the Loch Lomond Stadial ice sheet 8 359 in the western Grampians, the main stillstands during ice retreat in the Dee valley (Brown, 9 10 360 1993) and the Elgin Oscillation (Peacock et al., 1968). C Cairngorm; E Elgin Oscillation; M 11 361 Monadhliath; R Rannoch Moor. Stars indicate the locations of radiocarbon-dated Lateglacial 12 362 sites. 13 14 15 363 Figure 2. NextMap model of the eastern flank of mid-Strath Spey. The spur of Sròn a’ Ch-no, 16 364 in the centre of the Box indicating area of Fig 4, separates the Glenmore basin from Strath 17 365 Nethy. The likely lateral extent of the M1-M3 ice limits shown. The dashed white line marks 18 19 366 the northern edge of the Cairngorm Granite. 20 21 367 1. Strath Spey 2. Monadhliath flanks. 3. Glean Einich. 4. Lairig Ghru. 5. Coire an t-Sneachda. 22 368 6. Glen Avon. 7. Strath Nethy. 8. Sròn a’ Cha-no. 9. Water of Caiplich. 10. Glen More. 11. 23 24 369 Kincardine Hills. 12. Allt Bheadhair. 13. Abernethy Forest. 25 26 370 Stars indicate sites with cosmogenic exposure ages. Box indicates the area of Fig. 4. 27 28 371 Figure 3. Ice-marginal landforms in the northern Cairngorms. 29 30 372 A. View west across Strath Nethy. 31 32 373 B. View north-east from the Coire na Ciste Car Park. 33 34 374 C. View west from below the Coire Cas Car Park. 35 36 37 375 a. Lateral moraine. b. Marginal meltwater channel. c. Kame terrace d. Delta fan. e. Glacial 38 376 trimline. 39 40 377 Figure 4. Geomorphological map of the area between Glen More and Strath Nethy. 41 378 Modified after Brazier et al. (1996). 42 43 44 379 1. Moraine ridge. 2. Fan or kame terrace. 3. Boulders accumulations derived from rock 45 380 slope failures. 4. Cliff. 5. Ice-contact slope. 6. Glacial drainage channel. 7. Tor. 8. 46 381 Sites with cosmogenic exposure ages. 47 48 382 Figure 5. Elevations of ice-marginal landforms along mid-Strath Spey and their relationship 49 50 383 to the M1 and M3 moraine limits. YD = Younger Dryas moraines; LE = Loch Etteridge; AF = 51 384 Abernethy Forest 52 53 385 Figure 6. Dates constraining final deglaciation of the Cairngorms compared to the GISP2 and 54 55 386 GRIP Greenland ice cores and stadial/interstadial events. The radiocarbon and most 56 57 58 14 59 60 http://mc.manuscriptcentral.com/jqs Page 15 of 27 Journal of Quaternary Science

1 2 3 387 cosmogenic dates imply that rapid deglaciation took place before the warming represented 4 388 by the Windermere interstadial. 5 6 7 389 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 15 59 60 http://mc.manuscriptcentral.com/jqs Journal of Quaternary Science Page 16 of 27

1 2 3 390 TABLE LIST 4 5 10 6 391 Table 1. Sampling sites for Be surface exposure dating 7 8 Table 1 9 Sample Altitude Grid Moraine Site description 10 ID (m) reference 11 CG12- 737 NJ 01362 M1 Unweathered, coarse-grained 12 13 001 08262 granite boulder, 1 m high, resting on 14 schist bedrock. 15 CG12- 739 NJ 01366 M1 Large coarse-grained granite block 16 002 08279 on schist 17 CG12- 740 NJ 01376 M1 Granite bedrock sample from top of 18 19 003 08248 cliff at side of meltwater channel 20 CG12- 735 NJ 01289 M1 Quartz vein projecting from 21 004 08083 hydrothermally-altered granite 22 CG14- 741 NJ 01158 M1 Flat-topped granite boulder, 4 x 4 x 2 23 001 07369 m, above general slope of ~30 24 25 degrees. Sampled 1 cm thick surface 26 chippings. 27 CG14- 723 NJ 01035 M1 Large granite boulder on smooth 28 002 07392 slope of ~20 degrees. 4 x 3 x 2 m 29 size. Sampled 3 cm thick surface 30 31 chippings. 32 CG14- 603 NJ 00781 M2 Upstanding quartz veins on 33 003 07811 psammite boulder, 2 x 2 x 2 m. 34 Quartz veins sampled are up to 4 cm 35 above more eroded schist surface. 36 37 Sample thickness ~3 cm on average. 38 CG14- 546 NJ 00513 M3 Granite boulder 1 x 1 x 1 m. Boulder 39 004 08366 rises 50 cm above ridge top. 40 Sampled 1 cm thick surface 41 chippings. 42 CG14- 540 NJ 00311 M3 Granite boulder 2 x 1.5 x 0.5 m. 43 44 005 07888 Sample is ~3cm thick. 45 46 392 All samples were from coarse-grained feldspathic granite, except CG12-004 and CG14-003, 47 393 which were from vein quartz. 48 49 394 50 51 52 53 54 55 56 57 58 16 59 60 http://mc.manuscriptcentral.com/jqs Page 17 of 27 Journal of Quaternary Science

1 2 3 395 4 5 396 Table 2. Cosmogenic 10 Be surface exposure ages 6 7 1σ age 1σ 8 Topograp Age 1σ age external 10 Be concentrat 9 Sample hic (year internal uncertai Sample Lat Long Elevati concentrat ion thickne shielding s, Du uncertai nty 10 ID (DD) (DD) on (m) ion (x10 3 uncertaint ss (cm) correctio scalin nty (years, 11 at/g) y (x10 3 n g) (years) Du 12 at/g) 13 scaling) 14 - CG12- 57.154 3.6321 6.0 1471 15 001 59 7 737 2.0 1.000 126.4 0 710 880 16 - 57.154 17 CG12- 3.6321 1447 75 18 002 1 739 5.0 1.000 121.7 6.1 0 740 900 - 19 57.154 CG12- 3.6319 5.3 1411 20 47 21 003 4 740 1.0 1.000 122.6 0 620 800 - 22 57.152 CG12- 3.6333 1665 97 23 004 1 735 4.5 1.000 142.6 7.3 0 850 1030 24 - 25 CG14- 57.146 3.6351 7.1 2398 26 001 53 9 741 2.0 0.907 186.0 0 950 1280 27 - 28 CG14- 57.146 3.6372 1644 29 002 71 3 723 3.5 0.959 132.1 5.1 0 650 870 30 - CG14- 57.150 3.6416 5.0 1578 31 003 42 0 603 1.5 0.972 117.5 0 680 890 32 - 33 CG14- 57.155 3.6462 1355 34 004 34 4 546 1.0 0.976 96.97 4.12 0 580 760 35 - 36 CG14- 57.151 3.6493 4.5 1536 37 005 01 9 540 3.5 0.974 106.7 0 660 860 38 39 397 40 41 398 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 17 59 60 http://mc.manuscriptcentral.com/jqs Journal of Quaternary Science Page 18 of 27

1 2 3 399 REFERENCES 4 5 400 Balco, G, Stone, JO, Lifton, NA, Dunai, TJ, 2008. A complete and easily accessible means of 10 26 6 401 calculating surface exposure ages or erosion rates from Be and Al measurements. 7 402 Quaternary Geochronology 3: 174-195. 8 9 403 Ballantyne, CK, 2010. Extent and deglacial chronology of the last British-Irish Ice Sheet: 10 404 implications of exposure dating using cosmogenic isotopes. Journal of Quaternary Science 11 25 12 405 : 515-534. 13 14 406 Ballantyne, CK, Schnabel, C, Xu, S, 2009. Exposure dating and reinterpretation of coarse 15 407 debris accumulations ("rock glaciers") in the Cairngorm Mountains, Scotland. Journal of 16 408 Quaternary Science 24 : 29–31. 17 18 409 Ballantyne, CK, Stone, JO, 2012. Did large ice caps persist on low ground in north-west 19 410 Scotland during the Lateglacial Interstade? Journal of Quaternary Science 27 : 297-306. 20 21 411 Barrow, G, Hinxman, LW, Craig, EHC, 1913. The geology of upper Strathspey, Gaick and the 22 412 Forest of Atholl . HMSO, Edinburgh p. 112. 23 24 413 Binnie, SA, Dunai, TJ, Voronina, E, Goral,T, Heinze, S, Dewald, A, 2015. Separation of Be and 25 414 Al for AMS using single-step column chromatography. Nuclear Instruments and Methods in 26 361 27 415 Physics Research Section B : 397–401. 28 29 416 Björck, S, Walker, MJ, Cwynar, LC, Johnsen, S, Knudsen, KL, Lowe, JJ, Wohlfarth, B, 1998. An 30 417 event stratigraphy for the Last Termination in the North Atlantic region based on the 31 418 Greenland ice-core record: a proposal by the INTIMATE group. Journal of Quaternary 32 419 Science 13 : 283-292. 33 34 420 Boston, CM, 2013. Regional and local plateau ice fluctuations during the last Glacial- 35 421 Interglacial Transition: ice-dammed lakes and local readvances in the southeast of the 36 422 Monadhliath Mounatins , In: Boston, C., Lukas, S., Merritt, J. (Eds.), The Quaternary of the 37 38 423 Monadhliath Mountains and the Great Glen: field guide. Quaternary Research Association. 39 424 Quaternary Research Association, pp. 197-204. 40 41 425 Brazier, V, Gordon, JE, Kirkbride, MP, Sugden, DE, 1996. The Late Devensian ice sheet and 42 426 glaciers in the Cairngorm Mountains , In: Glasser, N.F., Bennett, M.R. (Eds.), The Quaternary 43 427 of the Cairngorms. Quaternary Research Association, London, pp. 28-53. 44 45 428 Brazier, V, Kirkbride, M, Gordon, JE, 1998. Active ice-sheet deglaciation and ice-dammed 46 429 lakes in the northern Cairngorm Mountains, Scotland. Boreas 27 : 297-310. 47 48 430 Bremner, A, 1934a. The glaciation of Abernethy Forest. Transactions of the Edinburgh 49 431 Geological Society 13 : 1-16. 50 51 432 Bremner, A, 1934b. The glaciation of Moray and ice movements in the north of Scotland. 52 13 53 433 Transactions of the Edinburgh Geological Society : 17-56. 54 55 56 57 58 18 59 60 http://mc.manuscriptcentral.com/jqs Page 19 of 27 Journal of Quaternary Science

1 2 3 434 Brown, IM, 1993. Pattern of deglaciation of the last (Late Devensian) Scottish ice sheet: 4 435 evidence from ice-marginal deposits in the Dee valley, northeast Scotland. Journal of 5 436 Quaternary Science 8: 235-250. 6 7 437 Casassa, G. Rivera, A, Schwikowski, M, 2006. Glacier mass balance data for southern South 8 o o 9 438 America (30 -56 S) In Knight, PG, (Ed) Glacier Science and Environmental Change , Oxford. 10 439 Blackwell, 239-241. 11 12 440 Charlesworth, JK, 1956. The Lateglacial history of the Highlands and Islands of Scotland. 13 441 Transactions of the Royal Society of Edinburgh 62 : 769-929. 14 15 442 Clark, CD, Hughes, ALC, Greenwood, SL, Jordan, CJ, Sejrup, H-P, 2012. Pattern and timing of 16 443 retreat of the last British-Irish Ice Sheet. Quaternary Science Reviews 44 : 112-146. 17 18 444 Davis, PT, Bierman, PR, Corbett, LB, Finkel, RC, 2015. Cosmogenic exposure age evidence for 19 445 rapid Laurentide deglaciation of the Katahdin area, west-central Maine, USA, 16 to 15 ka. 20 446 Quaternary Science Reviews 116 : 95-105. 21 22 447 Dewald, A, Möller, O, Petkov, P, 2012. Developing the Recoil Distance Doppler-Shift 23 24 448 technique towards a versatile tool for lifetime measurements of excited nuclear states. 25 449 Progress in Particle and Nuclear Physics 67 : 786-839. 26 27 450 Dunai, T, 2001. Influence of secular variation of the geomagnetic field on production rates of 28 451 in situ produced cosmogenic nuclides. Earth and Planetary Science Letters 193 : 197-212. 29 30 452 Everest, JD, Golledge, NR, 2004. Dating deglaciation in Strathspey and the Cairngorm 31 453 Mountains , In: Lukas, S., Merritt, J.W., Mitchell, W.A. (Eds.), The Quaternary of the Central 32 454 Grampian Highlands. Quaternary Research Association, London, pp. 50-57. 33 34 455 Everest, JD, Kubik, PW, 2006. The deglaciation of eastern Scotland: cosmogenic 10 Be 35 456 evidence for a Lateglacial stillstand. Journal of Quaternary Science 21 : 95-104. 36 37 457 Gheorghiu, DM, Fabel, D, 2013. Revised surface exposure ages in the southeast part of the 38 39 458 Monadhliath Mounatins, Scotland , In: Boston, C., Lukas, S., Merritt, J. (Eds.), The Quaternary 40 459 of the Monadhliath Mountains and the Great Glen: field guide. Quaternary Research 41 460 Association. London, pp. 183-190. 42 43 461 Gheorghiu, DM, Fabel, D, Hansom, JD, Xu, S, 2012. Lateglacial surface exposure dating in the 44 462 Monadhliath Mountains, Central Highlands, Scotland. Quaternary Science Reviews 41 : 132- 45 463 146. 46 47 464 Golledge, N, 2002. Glaci-tectonic deformation of proglacial lake sediments in the Cairngorm 48 465 Mountains. Scottish Journal of Geology 38 : 126-136. 49 50 466 Gordon, JE, 1993. The Cairngorms , In: Gordon, J.E., Sutherland, D.G. (Eds.), Quaternary of 51 52 467 Scotland. Chapman and Hall, London, pp. 259-276. 53 54 468 Greenwood, SL, Clark, CD, Hughes, AL, 2007. Formalising an inversion methodology for 55 469 reconstructing ice-sheet retreat patterns from meltwater channels: application to the British 56 470 Ice Sheet. Journal of Quaternary Science 22 : 637-645. 57 58 19 59 60 http://mc.manuscriptcentral.com/jqs Journal of Quaternary Science Page 20 of 27

1 2 3 471 Hall, AM, Phillips, WM, 2006. Glacial modification of granite tors in the Cairngorms, 4 472 Scotland. Journal of Quaternary Science 21 : 811-830. 5 6 473 Hinxman, LW, Anderson, EM, 1915. The geology of mid-Strathspey and Strathdearn . Memoir 7 474 Geological Survey of Scotland . HMSO, Edinburgh. 8 9 475 Hinxman, LW, Teall, JJH, 1896. The geology of West Aberdeenshire, Banffshire, parts of Elgin 10 11 476 and Inverness. Memoir Geological Survey of Scotland . HMSO, Edinburgh. 12 477 Hubbard, A, Bradwell, T, Golledge, NR, Hall, AM, Patton, H, Sugden, DE, Cooper, R, Stoker, 13 14 478 MS, 2009. Dynamic cycles, ice streams and their impact on the extent, chronology and 15 479 deglaciation of the British–Irish ice sheet. Quaternary Science Reviews 28 : 758-776. 16 17 480 Hulton, NRJ, and Sugden, DE, 1995. Modelling mass balance on former maritime ice caps: a 18 481 Patagonian example. Annals of Glaciology , 21 , 304-310. 19 20 482 Huntley, B, 1994. Late Devensian and Holocene palaeoecology and palaeoenvironments of 21 483 the Morrone Birkwoods, Aberdeenshire, Scotland. Journal of Quaternary Science 9: 311-336. 22 23 484 Larsen, NK, Linge, H, Håkansson, L, Fabel, D, 2012. Investigating the last deglaciation of the 24 10 485 Scandinavian Ice Sheet in southwest Sweden with Be exposure dating. Journal of 25 27 26 486 Quaternary Science : 211-220. 27 28 487 Merritt, J, Connell, ER, Hall, AM, Riding, J, Submitted. Middle to Late Devensian glaciation of 29 488 north-east Scotland: implications for the north-eastern quadrant of the last British ice sheet. 30 489 Journal of Quaternary Science . 31 32 490 Merritt, JW, Auton, CA, Boston, CM, Everest, JD, Merritt, JE, 2013. An overview of the main 33 491 Late Devensian glaciation of the Central Grampian Highlands , In: Boston, C., Lukas, S., 34 492 Merritt, J. (Eds.), The Quaternary of the Monadhliath Mountains and the Great Glen: field 35 493 guide. Quaternary Research Association. London, pp. 25-40. 36 37 494 Nishiizumi, K, Imamura, M, Caffee, MW, Southon, JR, Finkel, RC, McAninch, J, 2007. Absolute 38 39 495 calibration of 10 Be AMS standards. Nuclear Instruments and Methods in Physics Research 40 496 Section B: Beam Interactions with Materials and Atoms 258 : 403-413. 41 42 497 Peacock, JD, Berridge, NG, Harris, AL, May, F, 1968. The geology of the Elgin District . Memoir 43 498 of the Geological Survey . HMSO, Edinburgh p. 155. 44 45 499 Phillips, ER, Auton, CA, 2013. Evidence for active ice retreat and ice-dammed lake formation 46 500 during deglaciation in Raitts Burn , In: Boston, C., Lukas, S., Merritt, J. (Eds.), The Quaternary 47 501 of the Monadhliath Mountains and the Great Glen: field guide. Quaternary Research 48 502 Association. London, pp. 153-163. 49 50 503 Phillips, WM, Hall, AM, Mottram, R, Fifield, K, Sugden, DE, 2006. Cosmogenic exposure ages 51 52 504 of tors and erratics on the Cairngorm plateau, Scotland: timescales for the development of a 53 505 classic landscape of selective linear glacial erosion. Geomorphology 73 : 222-245. 54 55 506 Sissons, JB, Walker, MJC, 1974. Lateglacial site in the central Grampian Highlands. Nature 56 507 249 : 822-824. 57 58 20 59 60 http://mc.manuscriptcentral.com/jqs Page 21 of 27 Journal of Quaternary Science

1 2 3 508 Small, D, Fabel, D, 2015. A Lateglacial 10 Be production rate from glacial lake shorelines in 4 509 Scotland. Journal of Quaternary Science 30 : 509-513. 5 6 510 Sugden, DE, 1970. Landforms of deglaciation in the Cairngorm Mountains, Scotland. 7 511 Transactions of the Institute of British Geographers 51 : 201-219. 8 9 512 Sugden, DE, Clapperton, CM, 1975. The deglaciation of upper Deeside and the Cairngorm 10 11 513 Mountains , In: Gemmell, A.M.D. (Ed.), Quaternary Studies in North East Scotland. Aberdeen 12 514 University., Aberdeen, pp. 30-38. 13 14 515 Vasari, Y, Vasari, A, 1968. Late- and Post-Glacial macrophytic vegetation in the lochs of 15 516 northern Scotland. Acta botanica fennicae 45 : 193-217. 16 17 517 Young, JAT, 1974. Ice wastage in Glenmore, upper Spey Valley, Inverness-shire. Scottish 18 518 Journal of Geology 10 : 147-158. 19 20 519 Young, JAT, 1975. A re-interpretation of the deglaciation of Abernethy Forest, Inverness- 21 520 shire. Scottish Journal of Geology 11 : 193-205. 22 23 521 Young, JAT, 1977. Glacial geomorphology of the Dulnain valley, Inverness-shire. Scottish 24 522 Journal of Geology 13 : 58-74. 25 26 523 Young, JAT, 1978. The landforms of upper Strathspey. Scottish Geographical Magazine 94 : 27 28 524 76-94. 29 30 525 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 21 59 60 http://mc.manuscriptcentral.com/jqs Journal of Quaternary Science Page 22 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 Fig. 1 27 318x175mm (300 x 300 DPI) 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 http://mc.manuscriptcentral.com/jqs Page 23 of 27 Journal of Quaternary Science

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Fig. 2 26 260x138mm (300 x 300 DPI) 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 http://mc.manuscriptcentral.com/jqs Journal of Quaternary Science Page 24 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 Fig. 3 196x275mm (300 x 300 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 http://mc.manuscriptcentral.com/jqs Page 25 of 27 Journal of Quaternary Science

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 184x202mm (300 x 300 DPI) 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 http://mc.manuscriptcentral.com/jqs Journal of Quaternary Science Page 26 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Fig. 5 25 160x79mm (300 x 300 DPI) 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 http://mc.manuscriptcentral.com/jqs Page 27 of 27 Journal of Quaternary Science

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Fig. 6 31 219x152mm (300 x 300 DPI) 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 http://mc.manuscriptcentral.com/jqs