Age and significance of aeolian sediment reworking on high plateaux in the Scottish Highlands

http://hol.sagepub.com

The Holocene

DOI: 10.1177/0959683607076448 2007; 17; 349 The Holocene

Stefan M. Morrocco, Colin K. Ballantyne, Joel Q.G. Spencer and Ruth A.J. Robinson Age and significance of aeolian sediment reworking on high plateaux in the Scottish Highlands

http://hol.sagepub.com/cgi/content/abstract/17/3/349 The online version of this article can be found at:

Published by:

http://www.sagepublications.com

can be found at:The Holocene Additional services and information for

http://hol.sagepub.com/cgi/alerts Email Alerts:

http://hol.sagepub.com/subscriptions Subscriptions:

http://www.sagepub.com/journalsReprints.navReprints:

http://www.sagepub.com/journalsPermissions.navPermissions:

http://hol.sagepub.com/cgi/content/refs/17/3/349SAGE Journals Online and HighWire Press platforms):

(this article cites 30 articles hosted on the Citations

© 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. at KANSAS STATE UNIV LIBRARIES on December 21, 2007 http://hol.sagepub.comDownloaded from

http://hol.sagepub.com/cgi/alerts

http://hol.sagepub.com/subscriptions

http://www.sagepub.com/journalsReprints.nav

http://www.sagepub.com/journalsPermissions.nav

http://hol.sagepub.com/cgi/content/refs/17/3/349


Introduction

Amongst the most striking landforms on high plateaux in theScottish Highlands are extensive deflation surfaces and relatedaeolian deposits. The latter achieve their greatest extent andthickness on lee slopes bordering Torridon Sandstone plateaux inthe NW Highlands (Peach et al., 1913; Godard, 1965; Ballantyneand Whittington, 1987; Ballantyne, 1993, 1995) but also occur onor at the margins of plateaux underlain by other lithologies,including Devonian sandstone on Orkney (Goodier and Ball,1975), granite on Shetland and in the Cairngorms (Ball andGoodier, 1974), quartzite in NW Scotland (Pye and Paine, 1983),basalts on Mull and Skye (Birse, 1980; Ballantyne 1998), ultraba-sic rocks on Rhum, Moine Schists in the NW Highlands andDalradian schists in the Grampian Highlands (Ballantyne andHarris, 1994; Figure 1). Similar deposits have been reported fromother maritime periglacial environments characterized by strong

winds, such as NW Ireland (Wilson, 1989), Iceland (Arnalds,2000) and the Faeroe Islands (Christiansen, 1998; Humlum andChristiansen, 1998).

On Scottish mountains, high-level deflation surfaces com-prise broad expanses of frost-weathered regolith from whichstrong winds have stripped all or nearly all vegetation cover andwinnowed away all exposed particles up to about 6 mm in diam-eter, leaving a featureless expanse of sterile ground armoured byboulders and carpeted by lag gravels (Figure 2). Upstandingboulders sometimes exhibit windpolishing by saltating grains(Christiansen, 2004). High-level aeolian deposits take the formof vegetation-covered sand sheets, usually less than 1 m thickbut locally achieving depths of up to 4 m (Figure 3). Thesedeposits are typically massive and poorly sorted, and some mayhave a predominantly niveo-aeolian origin (Pye and Paine,1983; Ballantyne and Whittington, 1987; Ballantyne, 1998).Granulometric analyses of 57 samples from the aeolian depositson An Teallach (Torridon Sandstone) in the NW Highlands byBallantyne and Whittington (1987, figures 9 and 10) showedthat though these are dominated by medium sand (212–600 �m;41–67% by weight), both fine sand (63–212 m) and coarse sand(600–2000 m) are well represented in all samples; silt content

The Holocene 17,3 (2007) pp. 349–360

Age and significance of aeoliansediment reworking on high plateauxin the Scottish HighlandsStefan M. Morrocco, Colin K. Ballantyne,* Joel Q.G. Spencer**and Ruth A.J. Robinson

(School of Geography and Geosciences, University of St Andrews, Fife KY16 9AL, UK)

Received 25 June 2006; revised manuscript accepted 10 October 2006

Abstract: Holocene aeolian sand sheets at the margins of high-level deflation surfaces in the ScottishHighlands commonly comprise two units: a lower weathered unit representing slow accumulation through-out much or all of the Holocene, and an upper unit of structureless sand that is inferred to represent recenterosion of aeolian deposits and aeolisols from adjacent plateaux. OSL dating of samples from above andbelow the upper–lower unit contact at three sites in NW Scotland places the onset of upper unit sedimentaccumulation within the interval AD 1550–1700. Accumulation rates calculated from OSL ages confirmrapid accumulation of upper-unit deposits. The timing of the onset of upper-unit sand accumulationexcludes expansion of sheep grazing, intrinsic instability and atmospheric pollution as triggers of plateau-surface vegetation degradation and consequent erosion, but favours climatic deterioration (increased windstress and possibly prolonged snow-lie) during the ‘Little Ice Age’ as the likely cause. This explanation doesnot, however, apply in all cases; at an OSL-dated site in the Grampian Highlands upper-unit aeolian sandaccumulation commenced around AD 1900.

Key words: Deflation surfaces, sand sheets, OSL dating, ‘Little Ice Age’, Scotland, aeolian deposition, lateHolocene.

*Author for correspondence (e-mail: [email protected])**Present address: Institut für Geologie und Paläontologie,Leopold-Franzens-Universität Innsbruck, Innrain 52, A-6020Innsbruck, Austria.

© 2007 SAGE Publications 10.1177/0959683607076448 © 2007 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.

at KANSAS STATE UNIV LIBRARIES on December 21, 2007 http://hol.sagepub.comDownloaded from


https://www.researchgate.net/publication/229664677_Suggested_amendments_to_the_world_soil_classification_to_accommodate_Scottish_aeolian_and_mountain_soils?el=1_x_8&enrichId=rgreq-c62e3e1c-3f03-4666-95e5-300f32d42dab&enrichSource=Y292ZXJQYWdlOzIzMjI3MTU5MDtBUzo5NzY1NTk1NDg3MDI4NUAxNDAwMjk0NDAyNTMz

1995). Studies by Ballantyne and Whittington (1987) of thestratigraphy and age of aeolian sand deposits on the eastern(downwind) margins of the northern plateau of An Teallach, aTorridon Sandstone mountain in NW Scotland, revealed a com-plex history of deposition. They showed that aeolian depositioncommenced in the early Holocene, and suggested that a stablevegetation-covered sand sheet up to 2.2 m thick had eventuallycovered much of the northern plateau, most of which is nowoccupied by an extensive deflation surface (Figure 2). Sectionscut through sand scarps at the margins of the plateau revealedtwo distinct stratigraphic units, sometimes separated by anunconformity. The upper unit comprises uniform, massivereddish-brown sand. By contrast, the lower unit is weathered andleached to a pale yellow colour and contains bands of iron- andmanganese-stained sand; some exposures also revealed organic-rich humic horizons with occasional black peaty layers thatyielded radiocarbon ages between 6440 �60 14C yr BP (~7.3 cal.ka BP) and 3610 �70 14C yr BP (~3.9 cal. ka BP). Ballantyneand Whittington (1987) inferred that the lower sand unit repre-sented slow accumulation of aeolian or niveo-aeolian sandthroughout much of the Holocene; the accumulation of theupper unit they attributed to relatively recent erosion and aeolianreworking of plateau-surface sand deposits. Following Goodierand Ball (1975), they suggested that widespread aeolian erosionand redeposition of plateau-surface sand deposits may haveoccurred during the ‘Little Ice Age’ of the sixteenth to eighteenthcenturies AD, a period of increased storminess across the BritishIsles (Lamb, 1977, 1979, 1984; Whittington, 1985), or alterna-tively that erosion of plateau-surface sand sheets resulted fromvegetation removal associated with the introduction of sheep tothe mountain in the late eighteenth or nineteenth centuries AD.

350 The Holocene 17 (2007)

Figure 1 The Scottish Highlands, showing the location of sam-pling sites and other sites mentioned in the text

Figure 2 Deflation surface on the northern plateau of AnTeallach, NW Scotland

(�63 �m) is generally �4% by weight, but most samples con-tained granules 2–6 mm in diameter.

The origin of plateau-surface aeolian deposits in Scotlandhas been studied at The Storr (719 m) on the Isle of Skye, wherea sheet of predominantly sandy, vegetation-covered aeoliansediment up to 2.9 m thick rests on the summit plateau above asteep cliff of basalt lavas exposed by a rockslide at 6.5 �0.5 kaBP (Ballantyne et al., 1998). Radiocarbon dating of a soilburied under the aeolian deposits suggests that the onset ofaccumulation coincided with the exposure of the cliff, and boththe thickness of the deposit and mean grain size diminish withdistance from the crest of the cliff. Ballantyne (1998) inferredfrom this evidence that the plateau-top aeolian deposits on TheStorr represent particles that were released from the cliff byweathering and blown upwards by accelerating vertical airflowson to the plateau, where they were anchored by vegetationgrowth (cf. Marsh and Marsh, 1987; Wilson, 1989; Hétu, 1992).He suggested that plateau-top sand sheets on other Scottishmountains may have accumulated in a similar manner, thoughwhere cliffs are absent grains released by weathering of exposedbedrock outcrops and exposed clasts also may form an impor-tant source of aeolian sediments.

The aeolian deposits on The Storr, however, are unusual inthat the sand deposit is intact and has not been reworked by latererosion. In contrast, most high-level sand sheets in Scotlandexhibit evidence of widespread erosion, being bounded byactively eroding scarps (Figure 3). The thickest deposits general-ly occur downwind from extensive deflation surfaces that some-times support widely spaced remnant outliers of vegetation-covered aeolian deposits (Figure 4). The latter indicate that avegetated cover of aeolian sediment was formerly much morewidespread on plateau surfaces, but has been subsequently erod-ed and redeposited on adjacent lee slopes (Ball and Goodier,1974; Goodier and Ball, 1975; Pye and Paine, 1983; Ballantyne,



https://www.researchgate.net/publication/245669835_Climate_Present_Past_and_Future-Volume_2_Climatic_History_and_the_Future?el=1_x_8&enrichId=rgreq-c62e3e1c-3f03-4666-95e5-300f32d42dab&enrichSource=Y292ZXJQYWdlOzIzMjI3MTU5MDtBUzo5NzY1NTk1NDg3MDI4NUAxNDAwMjk0NDAyNTMz

https://www.researchgate.net/publication/240541158_The_Little_Ice_Age_and_Scotland's_weather?el=1_x_8&enrichId=rgreq-c62e3e1c-3f03-4666-95e5-300f32d42dab&enrichSource=Y292ZXJQYWdlOzIzMjI3MTU5MDtBUzo5NzY1NTk1NDg3MDI4NUAxNDAwMjk0NDAyNTMz

Subsequent investigations of the stratigraphy of plateau-margin aeolian sand sheets elsewhere in Scotland have revealedthat these also often contain a lower weathered unit and anupper unit of apparently unweathered sand, the latter implying

rapid, fairly recent erosion and redeposition of plateau-surfaceaeolian deposits (Ballantyne, 1995; Morrocco, 2005). Lack ofinformation on the timing of the deposition of upper unit sands,however, has not only meant that the causes of the implied

Stefan M. Morrocco et al.: Aeolian sediment reworking on Scottish high plateaux 351

Figure 3 Aeolian sand sheets up to 4 m thick at the margins of the northern plateau of An Teallach. The vegetation-covered sand depositsoccupy the crest of a lee slope on the left of the photograph. Deflation surfaces occupy the right foreground and the plateau beyond the sandsheets

Figure 4 Remnant outliers of vegetation-covered aeolian sand deposits on the An Teallach deflation surface. Such outliers imply a formerlymuch more extensive cover of windblown sand deposits covering much or all of the area occupied by the present deflation surface



erosion of plateau-surface aeolian deposits cannot be evaluated,but also that it is not known whether reworking of plateau-surface aeolian deposits occurred synchronously across theScottish Highlands, or at different times on different mountains.

The aim of this research is to resolve these unknownsthrough a programme of optically stimulated luminescence(OSL) dating of aeolian deposits at the margins of five plateauxin the Scottish Highlands. Particular objectives are: (1) to estab-lish the age of the onset of plateau-surface erosion and associ-ated redeposition of aeolian sands at each site; (2) to determinewhether these erosion and redeposition events occurred syn-chronously across the Scottish Highlands, or at different timesat individual sites; and (3) to use the dating evidence to test arange of hypotheses for the causes of widespread stripping ofaeolian sand deposits on Scottish plateaux.

Materials and methods

Field sites were selected to represent a range of lithologies andlocations (Figure 1; Table 1). At all sites the sampling pro-gramme focused on obtaining samples for OSL dating fromimmediately above (sample suffix ‘a’) and below (sample suffix‘b’) the stratigraphic contact between the upper (unweathered)and lower (weathered) sand units to bracket the timing of theonset of deposition of the upper sand unit. At all sampled sitesthe contact between the two units is conformable. The mostintensive sampling was carried out on An Teallach, the site ofprevious investigations (Ballantyne and Whittington, 1987),where three separate exposures were sampled in this way toestablish whether plateau-surface erosion and associated rede-position had occurred synchronously across the entire northernplateau. At the remaining four sites, pairs of samples wereobtained from above and below the contact between the lowerand upper sand units at a single section (Table 1).

On An Teallach, two closely spaced pairs of samples (AT1aand AT1b, AT2a and AT2b) were obtained from sections cut inthe margins of thick plateau-margin aeolian sands (Figure 5) ata location where Ballantyne and Whittington (1987) hadobtained a radiocarbon age of 3610 �70 14C yr BP (4.1–3.7 cal.ka BP at 95% confidence using OxCal 3.1; Bronk Ramsey,

2001) from the top of an organic-rich sand layer at a depth of2.6 m, 1.6 m below the contact between the upper and lowersand units. A further pair of samples (AT3a and AT3b) wereobtained from a plateau-margin section 700 m farther NNW.On Fionn Bheinn, paired samples FB1a and FB1b were


Table 1 Sampling sites and details of samples

Site Lithology Sample Sample grid Sample Samplenumbera reference altitude depthb

An Teallach Torridon Sandstone AT1a NH 072862 725 m 163 cmAT1b NH 072862 725 m 193 cmAT2a NH 072862 725 m 64 cmAT2b NH 072862 725 m 72 cmAT3a NH 067866 720 m 113 cmAT3b NH 067866 720 m 120 cm

Fionn Bheinn Moine schist FB1a NH 168619 700 m 27 cmFB1b NH 168619 700 m 34 cm

Red Cuillin Granite RC1a NG 587235 600 m 39 cmRC1b NG 587235 600 m 46 cm

Carn nan Gobhar Moine schist CG1a NH 197350 900 m 27 cmCG1b NH 197350 900 m 34 cm

West Drumochter Dalradian schist WD1a NN 603758 970 m 40 cmWD1b NN 603758 970 m 49 cm

aSamples with suffix ‘a’ were obtained above the contact between the upper sand unit and the lower sand unit; samples with suffix ‘b’ wereobtained from the same exposure below the contact.bSample depth refers to depth below the ground surface of the midpoint of the sampling tube.

Figure 5 Section excavated in a sand scarp at the margin of thenorthern plateau of An Teallach. The contact (arrowed) betweenthe upper (unweathered) and lower (weathered) sand units occursabout halfway up the section



obtained from an exposed sand scarp (Figure 6) immediatelyeast of a deflation surface on the eastern spur of the mountain.In the Red Cuillin of Skye, samples RC1a and RC1b were col-lected from a shallow sand scarp on the northern spur of BeinnDearg Mhór, east of a zone of intact vegetation, sand sheets

and deflation surfaces. Paired samples CG1a and CG1b fromCarn nan Gobhar were obtained from a sand scarp locatedimmediately east of an extensive deflation surface. SamplesWD1a and WD1b were collected from a large sand scarp southof the A’Mharconaich plateau at West Drumochter in theGrampian Highlands.

Sample collectionA vertical face was excavated at all sampling sites and adetailed record of profile stratigraphy was logged. Where pos-sible, samples were collected from immediately above andbelow the contact between the upper and lower sand units,though at sites where the contact proved indistinct sampleswere collected slightly farther from the contact to prevent sam-pling of a mixture of upper- and lower-unit sediment. Sampleswere collected in 4 cm diameter thick-walled black plastic cylin-ders 25 cm in length, and chamfered at one end to facilitateinsertion into the exposed face. During insertion, the outer endof the cylinder was covered to prevent exposure to daylight,and following insertion a polystyrene plug was hammered intothe exposed end of the tube. The cylinder was then carefullydug out of the face and a cap was placed over the chamferedend. Sample cylinders were then sealed with strong tape andreturned to the laboratory in thick black polythene bags.

The diameter of the sampling tube and clarity of the contactlimit the precision with which the age of the contact can bedetermined. In all cases but one, the samples were collectedfrom positions 7–9 cm (vertically) apart, above and below thecontact between the upper and lower sand units; samples AT1aand AT1b were collected 30 cm (vertically) from each other, asthe contact at this site was indistinct.

The environmental dose rate of the sediment at each sam-pling site was determined in the field using a portable EG&GORTEC MicroNOMAD gamma spectrometer that was insert-ed for 30 min to determine U, Th and K concentrations fromthe measured spectrum of environmental radiation due togamma rays. The elemental concentrations were converted intoexternal beta and gamma components using dose-rate conver-sion factors from Adamiec and Aitken (1998). Total dose rate(Table 2) for each sample included a cosmic ray component(estimated following Prescott and Hutton, 1994).


Figure 6 Section excavated in a sand scarp downwind from adeflation surface on Fionn Bheinn. The contact (arrowed) betweenthe upper (unweathered) and lower (weathered) sand units occursat the top of the conspicuous dark band

Table 2 Summary of equivalent doses, total dose rate and OSL ages

Sample N Equivalent dose Total dose rate OSL age (years Age range (years AD/BC

(Gy) (Gy/ka) before AD 2004, �1�) at 95% confidence)

AT1a 24 0.45 �0.03 1.96 �0.11 231 �20 AD 1733–1813‡F

AT1b 15 1.21 �0.04 1.99 �0.11 608 �38 AD 1320–1472†

AT2a 11 0.46 �0.02 2.05 �0.11 222 �15 AD 1753–1812‡C

AT2b 51 0.94 �0.05 2.15 �0.11 435 �32 AD 1505–1633‡F

AT3a 26 0.80 �0.02 2.77 �0.11 287 �14 AD 1689–1745‡C

AT3b 25 1.03 �0.03 2.43 �0.11 424 �22 AD 1536–1624‡C

FB1a 19 0.59 �0.04 1.82 �0.11 322 �28 AD 1626–1738‡C

FB1b 15 0.89 �0.06 1.76 �0.11 507 �46 AD 1405–1589†

RC1a 21 1.05 �0.23 2.92 �0.11 358 �79 AD 1488–1804‡F

RC1b 9 5.04 �0.7 3.01 �0.11 1674 �240 150 BC–AD 810†

WD1a 23 0.27 �0.03 2.65 �0.11 101 �10 AD 1883–1923‡C

WD1b 15 3.78 �0.28 2.99 �0.11 1265 �104 AD 531–947†

N.B. Total dose rate includes a cosmic dose-rate calculated assuming constant burial depth using method described in Prescott and Hutton(1994) with an uncertainty of �10%. Error in total dose rate has been weighted with an additional �10% to allow for systematic uncertaintyin as-sampled moisture content compared with average burial moisture conditions.† Age range is based on the mean De value.‡C, ‡F Age is based on a Central (‡C) or Finite (‡F) age model of the De distribution (Galbraith et al., 1999).



https://www.researchgate.net/publication/237049156_Dose-rate_conversion_factors_Update?el=1_x_8&enrichId=rgreq-c62e3e1c-3f03-4666-95e5-300f32d42dab&enrichSource=Y292ZXJQYWdlOzIzMjI3MTU5MDtBUzo5NzY1NTk1NDg3MDI4NUAxNDAwMjk0NDAyNTMz

https://www.researchgate.net/publication/227912725_Optical_dating_of_single_and_multiple_grains_of_quartz_from_Jinmium_rock_shelter_Northern_Australia_Part_I_Experimental_Design_and_Statistical_Models?el=1_x_8&enrichId=rgreq-c62e3e1c-3f03-4666-95e5-300f32d42dab&enrichSource=Y292ZXJQYWdlOzIzMjI3MTU5MDtBUzo5NzY1NTk1NDg3MDI4NUAxNDAwMjk0NDAyNTMz

Sample preparationSamples were extracted from sampling cylinders under low inten-sity red lighting in the luminescence laboratory at the Universityof St Andrews. About 1–2 cm of sediment was removed from theends of each sample cylinder to exclude the possibility ofanalysing sediment that had been exposed to daylight duringsampling. Samples were then dried in a 50°C oven and sieved toobtain a 125–180 mm particle size fraction. Carbonates andorganic matter were removed from this fraction using 10% HCland 30% H2O2, respectively. The quartz and feldspar-rich frac-tions were isolated from heavy minerals by density separationusing sodium heteropolytungstate (LST). The quartz andfeldspar-rich fractions were then treated with 40% HF for 40 minto dissolve the feldspars and minimize the luminescence resultingfrom ionisation by alpha particles external to the quartz grains.After HF treatment the quartz fractions were washed in 10% HClto dissolve any precipitated fluorides, and finally washed in dis-tilled water and acetone before drying briefly at 50°C. PreliminaryOSL measurements indicated low luminescence intensity andtherefore in all analyses large aliquots (circles 8–9 mm in diame-ter) of quartz grains were dispensed onto 10 mm diameter stain-less steel discs coated with a layer of silicone oil.

OSL measurementsLuminescence measurements were undertaken using an auto-mated Risø TL/OSL-DA-15 system with optical stimulationprovided by blue-green light (420–550 nm) from a filtered halo-gen lamp and liquid light guide (Bøtter-Jensen, 1997). The OSLsignals were measured with an EMI 9635QA photomultipliertube and 7.5 mm of Hoya U-340 filters. Beta irradiation wasperformed with a calibrated 90Sr/90Y source (dose rate~0.1/Gys). Equivalent dose (De) estimates were determinedusing the single-aliquot regenerative-dose (SAR) protocoldeveloped by Murray and Wintle (2000). For natural andregenerative doses, pre-heat temperatures of 180–280°C for 10 swere tested and all accepted De values were measured using apre-heat of 200°C for 10 s. For test doses, the cut-heat was setto 160°C and to achieve adequate signal intensity the size ofthe test dose used ranged from ~1 to 10 Gy. The OSL was mea-sured for 100 s at a sample temperature of 125°C. For SARdata we used a net OSL integral comprising the initial 0.8 s withsubtraction of a background level averaged over the final 20 s.For some samples we modified the SAR protocol to include (1)infrared stimulation (100 s at 125°C using IR diodes) before theblue-green shine to optically remove unwanted feldspar signals(cf. Jain and Singhvi, 2001; Wallinga et al., 2002; Spencer andRobinson, 2007), and (2) an additional OSL bleach conductedfor 100 s at 280°C at the end of each SAR cycle to remove lumi-nescence due to thermal transfer (Murray and Wintle, 2003).Three standard tests of luminescence behaviour were appliedto all aliquots. The rejection criteria employed were (1) known-dose recovery outside 0.9–1.1 measured/given ratio (Murrayand Wintle, 2003), (2) recycling ratios outside 0.9–1.1 (Murrayand Wintle, 2000) and (3) thermal transfer �5% of the naturalOSL level (Murray and Wintle, 2000).

The growth curve data fitted well to a linear function and Deestimates were evaluated by interpolation with the natural OSLlevel (Figure 7). The distribution in De was displayed usingradial plots and assessed using the central and finite age mod-els (Galbraith et al., 1999). The number of aliquots acceptedfor each sample, using the rejection criteria outlined above,ranged from 9 to 51 (Table 2). The two samples from Carn nanGobhar (CG1a and CG1b) failed to recover De values from anormal SAR cycle because of low luminescence sensitivity, andhence further analysis of these two samples was impossible.The De and calculated ages for the remaining samples are sum-

marised in Table 2; Figures 7 and 8 illustrate the luminescencebehaviour for six samples.

The De distributions of the samples from the upper sand unitare typical of aeolian deposits. Samples AT2a, AT3a, AT3b,FB1a, RC1a and WD1a have low values of overdispersion(�10%) for the De distribution and are normally distributed.The De distributions for samples AT1a and AT2b are slightlyskew, with overdispersion values of 16.3% and 15.0%, respec-tively, and the finite age model (Galbraith et al., 1999) was usedto assess these data (see Figure 8B for the finite age modellingof AT2b). The age and uncertainty for the normally distributedsamples are based on the mean De value and the standard error,whereas the uncertainty for the finite-age modelled De is the rel-ative standard error (Galbraith et al., 1999). It should be notedthat sample FB1a is imprecisely measured because of lowsignal-to-noise ratios (Figure 8C).

Hypotheses of plateau-surface erosion

As outlined above, the occurrence of a unit of structureless,apparently unweathered aeolian sand overlying one ofpronounced weathered sand downwind of deflation surfaces indi-cates extensive, rapid, fairly recent wind erosion of pre-existingsand and soil from plateaux (Ballantyne and Whittington, 1987;Ballantyne and Harris, 1994). The survival of remnantvegetation-covered outliers of sandy aeolian deposits on somedeflation surfaces suggests that former episodes of reworkingof plateau-surface aeolian deposits and soils was triggered byremoval and degradation of protective vegetation cover, lead-ing to blow-outs and rapid erosion of exposed aeolian deposits(cf. Seppälä, 1995). The following hypotheses for the triggeringof plateau-surface erosion therefore focus on possible causes of


Figure 7 Examples of regenerative dose growth curves and esti-mation of equivalent dose (De) for samples AT1a and RC1a






removal or degradation of protective vegetation cover withinthe last millennium or so; the absence of evidence for weather-ing of the upper-unit sands in comparison with those of thelower sand unit rules out a longer timescale.

Climatic deteriorationClimatic deterioration, particularly associated with the ‘Little IceAge’ of the sixteenth to eighteenth centuries AD, has been sug-gested as a possible cause of plateau-surface erosion by severalauthors (Ball and Goodier, 1974; Goodier and Ball, 1975;Ballantyne and Whittington, 1987). In Scotland, this period wascharacterized by increased storminess (Lamb 1977, 1979, 1984;Whittington, 1985), lower temperatures and increased snow-lie,with perennial snowfields developing on many higher mountains(Manley, 1949, 1971; Sugden, 1971; Ballantyne and Harris,1994). Anecdotal evidence summarized by Mitchell (1999) sug-gests that some plateaux experienced several years of continuoussnowcover during this period. The resulting physiological stresson vegetation covering (and protecting) readily erodible cohe-sionless aeolian sand deposits or aeolisols may have resulted inexposure of such sediments to strong winds following eventualsnowmelt, with rapid deflation inhibiting vegetation recoloniza-tion. This hypothesis implies that the onset of plateau-surfaceerosion occurred within the approximate timescale of ‘Little IceAge’ climatic deterioration.

Summer droughtConversely, it is possible that vegetation cover overlying aeoliansand deposits and aeolisols on exposed plateaux may have beenstressed during prolonged summer drought, allowing localized

deflation, blow-out development and consequent lateral exten-sion of erosion across plateaux. Such conditions are unlikely tohave occurred during the ‘Little Ice Age’, but may have per-tained during the ‘Little Optimum’ of AD 1100–1300, whensummer temperatures were relatively high (Ballantyne andHarris, 1994), or during the twentieth century AD.

Grazing damageIntensification of grazing pressure represents a third hypothe-sis for explanation of widespread erosion of aeolian sanddeposits and aeolisols from plateau surfaces. The grasses andsedges that grow on sand sheets constitute the richest grazingon mountains such as An Teallach, and attract large numbersof sheep at present (Ballantyne and Whittington, 1987). Sheepare voracious and selective grazers, and overgrazing of plateauvegetation by sheep or deer opens up small areas of unvegetat-ed ground (Morrocco, 2005), exposing the underlying substrateto the effects of cryoturbation, particularly by needle ice (whichdetaches surface soil particles) and deflation, both of whichinhibit vegetation recolonization (King, 1971). Expansion andcoalescence of bare ground into blow-out zones may have fol-lowed, marking the transition from a stable, vegetated sand oraeolisol cover to an expanding deflation surface, with concomi-tant redeposition of wind-eroded sediment on lee slopes toform the upper sand unit. A major expansion of grazing by hillsheep on high ground in the southern Scottish Highlandsoccurred from the mid- to late-eighteenth century AD, thoughin most of the NW Highlands the advent of extensive sheepgrazing did not occur until the first half of the nineteenth cen-tury, the time of the notorious Highland clearances (Whyte and


Figure 8 Radial plots of equivalent dose (De) values for four representative samples



ResultsEquivalent dose measurements and OSL agesOSL ages were calculated by dividing the De (Gy) for each sam-ple by the environmental dose rate (mGy/yr) calculated fromthe in situ gamma dosimetry and cosmic dose rate estimate(Table 2; Figure 9). Ages expressed as years AD are given in theform of 95% confidence limits to encapsulate the uncertainty ofeach age estimate (Table 2 and below).

An important consideration for interpretation of these resultsis the vertical distance between sample midpoints and the con-tact between the upper and lower sand units. At most sites thesample midpoints were located 30–50 mm from the contact. Asthe upper sand unit is inferred to have accumulated rapidly, thetime interval between the onset of upper unit accumulationand the OSL age calculated for upper-unit samples was proba-bly brief. Conversely, accumulation of the lower-unit sandswas probably much slower, implying a greater (possibly muchgreater) time interval between lower-unit OSL ages and theonset of upper-unit sand accumulation. Thus although thepaired OSL ages bracket the onset of upper-unit sand accumu-lation, the OSL age obtained for the base of the upper sand unitis probably much closer to the timing of the initiation of upper-unit sand accumulation than the OSL age calculated for the topof the lower sand unit.

Several points emerge from these data. First, the above-contact samples (AT1a and AT2a) from the two closely spacedsampling sites on An Teallach yield statistically indistinguish-able results (AD 1733–1813 and AD 1752–1812, respectively, at95% confidence). As the contact at these two sampling sites isstratigraphically continuous across the intervening 40 m, theseresults confirm the reproducibility of the OSL ages. The greaterage obtained for AT1b (AD 1320–1472) compared with AT2b(AD 1505–1633) reflects the fact that the former was sampled ata greater depth below the contact. For the third site on AnTeallach (AT3), 700 m NNW of AT1 and AT2, the below-contact sample (AT3b) yielded an age range (AD 1536–1624)similar to that for AT2b, though the upper unit sample (AT3a)produced an age range (AD 1689–1745) slightly younger thanthose for AT1a (AD 1753–1794) and AT2a (AD 1767–1798).This may imply that plateau-surface erosion commenced a fewdecades earlier at site AT3 but, given the uncertainty intro-duced by accumulation rates, could equally imply contempo-raneity. Three inferences can be made from these results: (1)that the onset of plateau-surface erosion on An Teallachoccurred between ~ AD 1550 and ~ AD 1700; (2) that the initia-tion of erosion was simultaneous or penecontemporaneousacross the plateau; and (3), given that the upper-unit ages arecloser to the true age of the onset of upper unit accumulationthan the lower-unit ages, that plateau-surface erosion probablycommenced shortly before AD 1700.

The OSL data for Fionn Bheinn, 24 km SSE of An Teallach,suggest slightly earlier aeolian reworking of plateau-surfacesediments. Sample FB1b from near the top of the lower sandunit yielded a significantly older age range (AD 1450–1542) thansamples AT2b and AT3b on An Teallach, but this differenceprobably reflects slower accumulation of windblown sedimentbetween the sampling point and the upper–lower unit contact,as the lower sand unit on Fionn Bheinn is much thinner thanthat on An Teallach. The younger upper-unit sample (FB1a)age range of AD 1626–1738 is more telling. This falls (at 95%confidence) within the overall inferred range of AD 1550–1700for the onset of plateau-surface erosion on An Teallach, but isslightly (though not significantly) older than AT3a (AD

1689–1745). Possible differences in initial rates of upper-unitsand accumulation make it impossible to rule out synchroneity


Whyte, 1991; Richards, 2000). A second period of expansionhas also occurred over the past 50–60 years (Sydes and Miller,1988; Hudson, 1995; Scottish Executive Environment andRural Affairs Department (SEERAD), 2001). Alternatively,the onset of plateau-surface erosion may be attributable toovergrazing by sheep introduced by Norse settlers in the tenthand eleventh centuries, or possibly the expansion of red deernumbers on Scottish mountains in the late nineteenth centuryto serve the development of sporting estates.

Intrinsic instability because of sand-sheet thickeningA fourth hypothesis is that plateau-surface sand sheets are intrin-sically vulnerable to widespread erosion. As sand sheets slowlythicken, it is likely that the stable vegetation cover protecting thesediments becomes progressively distanced from the summer sea-sonal water-table. Under such conditions the likelihood ofsevere physiological stress in the protective surface vegetationmat during prolonged summer droughts increases through time,potentially triggering either the development of bare ground andblow-out zones in areas of dead or dying vegetation cover, orreplacement by species whose root architecture is less suitable formaintaining surface stability. If this hypothesis is valid, it wouldbe expected that the timing of plateau-surface erosion would beasynchronous across the Scottish Highlands and even across indi-vidual plateaux because of uneven rates of sand-sheet thickening.

Other possible triggers of plateau-surface erosionOther possible triggers of vegetation degradation and conse-quent wind erosion of plateau-surface sand deposits andaeolisols include pest or fungal infestations or atmospheric pol-lution. Pest and fungal infestations may be restricted in extent(implying asynchronous erosion in different areas) or wide-spread (implying near-synchronous erosion). A weakness ofthis hypothesis is that the freshness of upper-unit sand depositsimplies that reworking occurred only relatively recently, withno evidence for analogous reworking events earlier in theHolocene. Atmospheric pollution is only likely to have affectedupland plant communities since the Industrial Revolution (ie,since the beginning of the nineteenth century) and thus repre-sents a more plausible cause of vegetation degradation. However,evidence of physiological stress resulting from atmospheric pol-lution in present upland plant communities appears spatiallyrestricted (D.B.A. Thompson, personal communication, 2003),implying that erosion events thus triggered are likely to be spa-tially asynchronous.

Complex hypothesesSeveral of the hypotheses outlined above are temporally con-strained, whereas others are not tied to a particular time interval;some, such as climatic deterioration, are likely to be representedby near-synchronous plateau-erosion episodes in different areas,whereas others (such as intrinsic instability due to sand-sheetthickening) are inherently spatially asynchronous. Dating of theonset of plateau-surface erosion at several sites thus has thepotential to exclude several hypotheses. However, most of thehypotheses listed above are not mutually exclusive: plateau-surface erosion triggered by vegetation degradation caused byovergrazing, for example, is likely to be exacerbated by climatic-induced physiological stress (perennial or late-lying snowcover,or prolonged drought) and erosion of sediment cover by strongwinds during periods of increased storminess is likely to acceler-ate aeolian reworking of sediment irrespective of the cause ofvegetation degradation. It follows that OSL dating of the onsetof plateau-surface erosion can limit the range of possible causes,but may not allow identification of a unique cause.



with the onset of upper-unit sand accumulation on An Teallach,but it is equally feasible that that plateau-surface erosion beganhere a few decades earlier.

The OSL ages obtained for the Red Cuillin, 78 km SW of AnTeallach, are much more widely spaced: at 95% confidence, theage obtained for the below-contact sample (RC1b) falls withinthe range 150 BC–AD 810. As the lower sand unit at this site isonly ~0.2 m thick, we infer that this much older limiting agerange reflects very slow accumulation of lower-unit sand.Conversely, the above-contact sample (RC1a) yielded an agerange (AD 1488–1804) that overlaps those of all above-contactsamples from both Fionn Bheinn (AD 1626–1738) and AnTeallach (AD 1689–1813).

Collectively, all five above-contact age ranges for these threesites are consistent with the inference (from the An Teallachdates) that the onset of plateau-surface erosion and concomi-tant accumulation of upper-unit sand occurred within theinterval AD 1550–1700. It is, however, unlikely that the onset ofplateau-surface erosion was synchronous at the three sites: theupper sand unit on An Teallach is much thicker than those onFionn Bheinn and the Red Hills, implying more rapid sedimentaccumulation, yet the above-contact OSL age ranges for thelatter two sites, though overlapping that for sample AT3a, sug-gest earlier deposition.

The West Drumochter site on A’Mharconaich is 126 km SEof An Teallach in the Grampian Highlands (Figure 1) and alsoyielded a wide difference in ages bracketing the upper–lowerunit contact. The sample from near the top of the lower unit(WD1b) gave an age range of AD 531–947, and like that for thelower-unit sample from the Red Cuillin, this much older limit-ing age range (compared with the An Teallach or FionnBheinn lower-unit ages) probably reflects very low sand accu-mulation rates. The upper-unit sample (WD1a) yielded an agerange much younger than any other (AD 1883–1923). Thisresult suggests two possibilities. The first is that the onset ofupper unit sand accumulation in West Drumochter occurredwithin a similar time interval to that at the other sites (AD

1550–1700), and that the younger age reflects much slower ini-tial accumulation of upper-unit sand. The second is that theonset of upper-unit accumulation was markedly later, proba-bly around the beginning of the twentieth century. The latterinterpretation is preferred as the sample was obtained ~ 40 mmabove the base of the upper sand unit at a depth of 40 cm. Theimplied mean accumulation rate above the sampling point isthus (40 cm/sample age), or 3.3–4.9 mm/yr. If this rate isassumed for the 40 mm interval between the base of the uppersand unit and the sampling point, the extrapolated age of theupper–lower unit contact is only about 10 years younger thanthe sample age range (ie, AD 1873–1913). A very much loweraccumulation rate of 0.11–0.22 mm/yr would be required toreduce the implied age of the contact to the age range (AD

1550–1700) inferred for the other sites. The size of this dis-crepancy suggests the onset of upper-unit sand accumulationat West Drumochter was indeed two to three centuries laterthan at the other sampling sites.

In summary, the OSL age ranges imply (1) that at the sites inNW Scotland (An Teallach, Fionn Bheinn, Red Cuillin), accu-mulation of the upper sand unit began within the interval AD

1550–1700; (2) that the onset of upper-unit sand accumulationwas probably a few decades earlier on Fionn Bheinn than on AnTeallach, and possibly even earlier on the Red Cuillin, thoughsynchroneity at the three sites cannot be excluded because ofuncertainties arising from initial upper-unit sand accumulationrates; (3) that the onset of upper-unit sand accumulation at threesites at the margins of the An Teallach plateau was synchronousor near-synchronous; and (4) that the onset of upper-unit


Figure 9 Probability frequency curves of OSL ages at the datedsites. In all cases the peak on the left represents the above-contactages, and that on the right represents the below-contact ages. EachGaussian curve depicted represents a single OSL age; all Gaussiancurves enclose an identical area, with the width of the curve deter-mined by the uncertainty of the OSL age



1991), the inferred onset of plateau-surface erosion at three sitesin NW Scotland within the time interval AD 1550–1700 limitsthe viability of most of the hypotheses outlined above. Intrinsicinstability because of sand-sheet thickening and consequentincreased risk of drought-induced physiological stress in theoverlying vegetation cover appears unlikely, not only because ofthe fairly close temporal grouping of reworking events after sev-eral millennia of apparent stability, but also because reworkingaffected plateaux with only a thin cover of aeolian sand (such asthe Red Cuillin on Skye) as well as those with a thick cover oflower-unit sand (such as An Teallach). This hypothesis also failsto explain the survival of intact sand sheets up to 2.9 m thickon The Storr on Skye (Ballantyne, 1998). Post-IndustrialRevolution atmospheric pollution can also be ruled out, aswidespread erosion of plateau-surface sand sheets in NWScotland appears to have pre-dated the industrial revolution byabout two centuries. Similarly, there appears to be a temporalmismatch between the inferred timing of the onset of plateau-surface erosion at the sites in NW Scotland and the great expan-sion in sheep grazing that accompanied the Highland clear-ances, as in this area the latter occurred in the first half of thenineteenth century AD, at least a century after the inferred initi-ation of plateau-surface erosion. Overgrazing by sheep intro-duced by Norse settlers in the tenth or eleventh centuries canalso be excluded, as this occurred several centuries prior to theinferred timing of plateau-surface erosion. Pest or fungal infec-tion of plateau vegetation cover cannot be excluded as a triggerof widespread plateau-surface erosion but, as noted above, thishypothesis is difficult to reconcile with relatively recent rework-ing of plateau-surface deposits after several millennia of appar-ent stability.

The most compelling hypothesis in terms of the timing ofthe onset of plateau-surface erosion in NW Scotland is that thelatter was triggered by climatic stress during the period of cli-matic deterioration commonly referred to as the ‘Little IceAge’, even though there is no convincing evidence for regener-ation of glacier ice on Scottish mountains at this time.Historical evidence shows that the seventeenth century AD wasa period of unprecedented climatic harshness in Scotland(Lamb, 1977) with abundant documentary evidence for anincrease in the magnitude and frequency of exceptional storms(Lamb, 1979, 1984; Whittington, 1985). Whether wind stressalone was sufficient to strip the protective vegetation mat fromerodible plateau-surface sand sheets and aeolisols remainsuncertain. As suggested above, the persistence of late-lying orperennial snowcover (Lamb, 1977; Mitchell, 1999) may havebeen critical in opening the vegetation cover and thus exposingthe underlying sand deposits and soils to wind erosion, creatingblowouts that subsequently expanded and coalesced to formthe present deflation surfaces. An attractive feature of thisexplanation is that the ‘Little Ice Age’ represents the period ofcoolest and stormiest weather in Scotland for several millennia(Ballantyne and Harris, 1994), and probably since a pro-nounced cooling event at c. 8.2 cal ka BP (Klitgaard-Kristensenet al., 1998). It remains possible, however, that introduction ofsheep to high plateaux in the nineteenth century exacerbatedand accelerated the stripping of aeolian sediments andaeolisols that was already underway.

The inferred relative recency of the onset of upper-sand unitaccumulation and implied plateau surface erosion at the WestDrumochter site appears, however, irreconcilable with erosiontriggered by climatic stress associated with ‘Little Ice Age’climatic deterioration. At this site several possible explanationsremain viable, including climatic stress, overgrazing (by deer orsheep), or even localized vegetation stress resulting from diseaseor pollution. The contrast in the timing of erosion between this


accumulation in the West Drumochter Hills was much laterthan at the NW Scotland sites, and probably occurred near theend of the nineteenth century.

Mean accumulation ratesMean accumulation rates for upper-unit aeolian sediments canbe calculated by dividing sample depth by sample age range. ForAn Teallach, where plateau-margin sand scarps are up to 4 mdeep and upper-unit sand accumulations are thicker than atany other site, calculated mean accumulation rates are2.5–8.5 mm/yr. The much thinner aeolian deposits on FionnBheinn and the Red Cuillin imply lower mean accumulation rates(0.7–1.0 mm/yr and 0.8–1.9 mm/yr, respectively). Conversely, theWest Drumochter data imply a rapid mean accumulation rate of3.3–4.9 mm/yr, as here the 40 cm of sand above the samplingpoint accumulated in a much briefer time interval. These accu-mulation rates should be regarded as maximal for each location,as they reflect sampling of the thickest aeolian deposit, andbecause accumulation rate (and hence deposit thickness) tendsto decline exponentially with distance downwind from plateaumargins (Ballantyne and Whittington, 1987).

The mean accumulation rate for most lower sand units can-not be calculated, as the age of the onset of lower-unit accumu-lation is unknown. At site AT1, however, a radiocarbon age of6440 �60 14C yr BP (7460–7250 cal. yr BP at 95% confidenceusing OxCal 3.1; Bronk Ramsey, 2001) obtained for organic-rich sand at 345 cm depth by Ballantyne and Whittington (1987)provides a lower age datum for lower-unit sand accumulation.Sample AT1b from 193 cm depth near the top of the lower unityielded an OSL age of 608 �38 years. Dividing the interveningvertical distance (152 cm) by the difference between these agesyields a mean accumulation rate of 0.22–0.23 mm/yr, about 3%of the upper-sand accumulation rate (6.0–8.5 mm/yr) calcu-lated for the same site. Moreover, the biostratigraphy of theradiocarbon-dated organic sand layer at AT1 suggests that theradiocarbon age may be affected by downwash of youngerorganic carbon, and places the age of the deposit at 345 cmdepth at c. 7.9 14C ka BP (c. 8.8 cal. ka BP; Ballantyne andWhittington, 1987). If this figure is adopted, the mean accumu-lation rate of lower-unit sand reduces to roughly 0.18 mm/yr,about 2.5% of the calculated upper-sand accumulation rate.This contrast strongly supports the inference that, whereas theweathered lower sand unit represents slow accumulation of aeo-lian deposits throughout most of the Holocene, the upper sandunit represents very rapid accumulation of plateau-margin sandsbecause of rapid erosion of sand sheets and aeolisols fromplateau surfaces.

The magnitude of the lower-sand unit mean accumulation rateinferred for site AT1 on An Teallach (0.18–0.23 mm/yr) is similarto those calculated by Ballantyne (1998) from calibrated radio-carbon ages for organic horizons within the intact sand sheet onthe summit plateau of The Storr on Skye. At the latter site, sedi-ment accumulation over most of the plateau has averaged0.1–0.2 mm/yr since exposure of the adjacent cliff at c. 6.5 ka BP.The thicknesses of the lower sand units at Fionn Bheinn (0.7 m),the Red Cuillin (~ 0.3 m) and West Drumochter (0.4 m) are muchless than those on An Teallach and The Storr. If these also accu-mulated throughout most of the Holocene, even lower rates areimplied prior to the onset of upper-unit accumulation.

Discussion: timing and causes of plateau-surface erosion

Although caution is necessary in using chronological evidencealone to link upland erosion with putative causes (Ballantyne,



site and those in NW Scotland (and indeed between all of thesesites and The Storr on Skye, where vegetation-covered plateau-surface sand sheets have continued to accumulate uninterruptedsince the early Holocene) highlights the difficulty of identifyinga single factor or combination of factors responsible for destabi-lizing plateau vegetation cover, and thus triggering widespreadreworking of aeolian deposits and aeolisols, the formation ofdeflation surfaces, and rapid accumulation of reworked aeoliansediments at plateau margins.

Conclusions

(1) Extensive deflation surfaces on high plateaux in the ScottishHighlands support residual islands of vegetated aeolian sandsor aeolisols, indicating formerly much more extensive sand oraeolisol cover. In contrast, vegetation-covered sand sheets up to4 m thick occur on lee slopes at the margins of deflation surfaces.Such sand sheets often include two depositional units: a lower,weathered sand unit, representing slow accumulation of aeolianor niveo-aeolian sediments throughout much of the Holocene;and an upper unit of structureless, apparently unweathered sandthat is inferred to represent relatively rapid, relatively recent ero-sion of plateau-surface sand sheets and aeolisols, and redeposi-tion of these sediments at plateau margins.(2) The age of the contact between the lower and upper sandunit was established for four plateaux by optically stimulatedluminescence (OSL) dating of aeolian sediments below andabove the contact. For three sites on the northern plateau ofAn Teallach in NW Scotland, the results imply that the onsetof plateau-surface erosion occurred in the time interval AD

1550–1700. OSL dates for two other plateau-margin sites inNW Scotland fall within this range, but may indicate that ero-sion and associated sand redeposition commenced a few de-cades earlier than on An Teallach. A single plateau-margin sitein the West Drumochter Hills, 126 km SE of An Teallach in theGrampian Highlands, yielded a much more recent inferred age(early twentieth century) for the onset of plateau erosion andconcomitant upper sand unit accumulation.(3) Calculated mean aeolian sediment accumulation rates foraeolian deposits for An Teallach are 0.18–0.23 mm/yr for thelower sand unit and 2.5–8.5 mm/yr for the upper sand unit,confirming that the latter represents much more rapid ratesof accumulation. Both lower and upper sand units at theother sites in NW Scotland are thinner than those on AnTeallach, reflecting lower mean rates of sediment accumula-tion (0.7–1.9 mm/yr for upper-unit sands). In contrast, theupper-unit sands at the West Drumochter site have accumu-lated at an average rate of 3.3–4.9 mm/yr during the twenti-eth century.(4) Various hypotheses for the triggering of plateau-surfaceerosion after several millennia of stability are proposed, basedon the degradation of the protective vegetation mat and theexposure of underlying sand deposits and aeolisols to winderosion. The inferred timing of plateau-surface erosion at thesites in NW Scotland appears inconsistent with most hypothe-ses, including intrinsic instability, summer drought and expan-sion of sheep grazing on to high plateaux. Conversely, it coin-cides with the timing of climatic deterioration during the ‘LittleIce Age’ of the sixteenth to nineteenth centuries AD, a period ofincreased storminess, cooler temperatures and increased snow-lie on Scottish mountains. Whether increased wind stress alonewas sufficient to trigger opening of the vegetation cover isuncertain; physiological stress induced by prolonged or peren-nial snowcover may have been critical in exposing sand depositsand aeolisols to deflation.

(5) The more recent age inferred for the onset of plateau-surface erosion at the West Drumochter site, however, impliesthat not all reworking of plateau-surface sand sheets oraeolisols occurred penecontemporaneously during a period ofclimatic deterioration, and that other factors such as overgraz-ing and atmospheric pollution have probably contributed to thedestabilization and aeolian reworking of plateau-surfacedeposits.

Acknowledgements

This research was supported by the award of a PhD stu-dentship to SMM, partly funded by Scottish Natural Heritageand partly by the University of St Andrews. Additional fund-ing in support of fieldwork was made available by the BillBishop Memorial Trust, the British GeomorphologicalResearch Group, the Carnegie Trust for the Universities ofScotland and the Quaternary Research Association. The OSLanalyses were funded by the Russell Trust of the University ofSt Andrews and a Spragge Scholarship. We thank Dr SimonNelis for assistance with fieldwork, Professor John Gordon andProfessor Des Thompson for constructive comments onaspects of this research, Professor Geoff Duller, Dr HelenaRodnight and Professor Rex Galbraith for help and discussionson the central and finite age modelling, Graeme Sandeman fordrafting some of the figures, and Dr Peter Wilson and ananonymous reviewer for constructive comments on the originalsubmission.

References

Adamiec, G. and Aitken, M. 1998: Dose-rate conversion factors:update. Ancient TL 16, 37–50.Arnalds, O. 2000: The Icelandic ‘Rofabard’ soil erosion features.Earth Surface Processes and Landforms 25, 17–28.Ball, D.F. and Goodier, R. 1974: Ronas Hill, Shetland: a prelimi-nary account of its ground pattern features resulting from theaction of frost and wind. In Goodier, R., editor, The natural envi-ronment of Shetland. Natural Conservancy Council, 89–106.Ballantyne, C.K. 1991: Late Holocene erosion in upland Britain:climatic deterioration or human influence? The Holocene 1,81–85.—— 1993: An Teallach. In Gordon, J.E. and Sutherland, D.G.,editors, Quaternary of Scotland. Chapman and Hall, 110–15.—— 1995: Periglacial features on the mountains of Assynt andCoigach. In Lawson, T.J., editor, The Quaternary of Assynt andCoigach: field guide. Quaternary Research Association, 47–60.—— 1998: Aeolian deposits on a Scottish mountain summit: char-acteristics, provenance, history and significance. Earth Surface Pro-cesses and Landforms 23, 625–41.Ballantyne, C.K. and Harris, C. 1994: The periglaciation of GreatBritain. Cambridge University Press.Ballantyne, C.K. and Whittington, G.W. 1987: Niveo-aeolian sanddeposits on An Teallach, Wester Ross, Scotland. Transactions ofthe Royal Society of Edinburgh: Earth Sciences 78, 51–63.Ballantyne, C.K., Stone, J.O. and Fifield, L.K. 1998: CosmogenicCl-36 dating of postglacial landsliding at The Storr, Isle of Skye,Scotland. The Holocene 8, 347–51.Birse, E.L. 1980: Suggested amendments to the world soil classifi-cation to accommodate Scottish aeolian and mountain soils.Journal of Soil Science 31, 117–24.Bøtter-Jensen, L. 1997: Luminescence techniques: instrumentationand methods. Radiation Measurements 27, 749–68.Bronk Ramsey, C. 2001: Development of the radiocarbon programOxCal. Radiocarbon 43, 355–63.Christiansen, H.H. 1998: Highland aeolian deposits in the FaroeIslands. Fródskaparrit 46, 205–13.




—— 2004: Windpolished boulders and bedrock in the ScottishHighlands: evidence and implications of Late Devensian windactivity. Boreas 33, 82–94.Galbraith, R.F., Roberts, R.G., Laslett, G.M., Yoshida, H. and Olley,J.M. 1999: Optical dating of single and multiple grains of quartzfrom Jinmium rock shelter, Northern Australia: part 1, experimentaldesign and statistical models. Archaeometry 41, 339–64.Godard, A. 1965: Recherches de Géomorphologie en Écosse duNord-Ouest. Masson.Goodier, R. and Ball, D.F. 1975: Ward Hill, Orkney: patternedground features and their origins. In Goodier, R., editor, The natu-ral environment of Orkney. Nature Conservancy Council, 47–56.Hétu, B. 1992: Coarse cliff-top sedimentation in northernGaspésie, Québec (Canada). Earth Surface Processes andLandforms 17, 95–108.Hudson, P.J. 1995: Ecological trends and grouse management inupland Britain. In Thompson, D.B.A., Hester, A.J. and Usher, M.B.,editors, Heaths and moorland: cultural landscapes. HMSO, 282–293.Humlum, O. and Christiansen, H.H. 1998: Mountain climate andperiglacial phenomena in the Faeroe Islands. Permafrost andPeriglacial Processes, 9, 189–211.Jain, M. and Singhvi, A.K. 2001: Limits to depletion of blue-greenlight stimulated luminescence in feldspars: implications for quartzdating. Radiation Measurements 33, 883–92.King, R.B. 1971: Vegetation destruction in the sub-alpine andalpine zones of the Cairngorm Mountains. Scottish GeographicalMagazine 87, 103–15.Klitgaard-Kristensen, D., Sejrup, H.P., Haflidason, H., Johnsen, S.and Spurk, M. 1998: A regional 8200 cal yr BP cooling event innorthwest Europe, induced by final stages of the Laurentide ice-sheet deglaciation? Journal of Quaternary Science, 13, 165–69.Lamb, H.H. 1977: Climate: past, present and future. Volume 2:Climatic history and the future. Methuen.—— 1979: Climatic variation and changes in the wind and oceancirculation: the Little Ice Age in the Northeast Atlantic.Quaternary Research 11, 1–20.—— 1984: Some studies of the Little Ice Age of recent centuriesand its great storms. In Morner, N.-A. and Karlén, W., editors,Climatic changes on a yearly to millennial basis. Reidel, 291–309.Manley, G. 1949: The snowline in Britain. Geografiska Annaler 13,179–93.—— 1971: Scotland’s semi-permanent snows. Weather 26, 458–71.Marsh, W.M. and Marsh, B.D. 1987: Wind erosion and sand duneformation on high Lake Superior bluffs. Geografiska Annaler 69A,379–91.

Mitchell, I. 1999: Scotland’s mountains before the mountaineers.Luath Press.Morrocco, S.M. 2005: Terrain sensitivity on high plateaux in theScottish Highlands. Unpublished PhD thesis, University of StAndrews.Murray, A.S. and Wintle, A.G. 2000: Luminescence dating ofquartz using an improved single-aliquot regenerative-dose proto-col. Radiation Measurements 32, 57–73.—— 2003: The single aliquot regenerative dose protocol: potentialfor improvements in reliability. Radiation Measurements 37,377–81.Peach, B.N., Horne, J., Gunn, W., Clough, C.T., Hinxman, L.T.,Cadell, H.M., Greenly, E., Pocock, T.I. and Crampton, C.B. 1913:The geology of the Fannich Mountains and the country aroundUpper Loch Maree and Strath Broom. Memoirs of the GeologicalSurvey, Scotland.Prescott, J.R. and Hutton, J.Y. 1994: Cosmic ray contributions todose rates for luminescence and ESR dating: large depths andlong-time variations. Radiation Measurements 23, 497–500.Pye, K. and Paine, A.D.M. 1983: Nature and source of aeoliandeposits near the summit of Ben Arkle, Northwest Scotland.Geologie en Mijnbouw 68, 13–18.Richards, E. 2000: Highland clearances. Birlinn Press.Scottish Executive Environment and Rural Affairs Department2001: Agricultural facts and figures. SEERAD.Seppälä, M. 1995: Deflation and redeposition of sand dunes inFinnish Lapland. Quaternary Science Reviews 14, 799–809.Spencer, J.Q.G. and Robinson, R.A.J. 2007: Dating intramontanealluvial deposits using luminescence techniques: problems andpotential. Geomorphology, doi:10.1016/j.geomorph.2006.12.021.Sugden, D.E. 1971: The significance of periglacial activity on someScottish mountains. Geographical Journal 137, 388–92.Sydes, C. and Miller, G.R. 1988: Range management and natureconservation in the British uplands. In Usher, M.B. andThompson, D.B.A., editors, Ecological change in the uplands.Blackwell, 323–37.Wallinga, J., Murray, A.S. and Bøtter-Jensen, L. 2002: Measure-ment of the dose in quartz in the presence of feldspar contamina-tion. Radiation Protection Dosimetry 101, 367–70.Whittington, G.W. 1985: The Little Ice Age and Scotland’s weath-er. Scottish Geographical Magazine 101, 174–78.Whyte, I.D. and Whyte, K.A. 1991: The changing Scottish land-scape, 1500–1800. Routledge.Wilson, P. 1989: Nature, origin and age of Holocene aeolian sandon Muckish Mountain, Co. Donegal, Ireland. Boreas 18, 159–68.




Age and significance of aeolian sediment reworking on high plateaux in the Scottish Highlands

Documents