The Archaeology of sand landscapes: looking for an integrated approach

David Griffiths

University of Oxford

Scottish Archaeological Internet Report 48, 2011 www.sair.org.uk

CONTENTS

List of Illustrations...... 12

1 The Archaeology of Sand Landscapes: Looking for an Integrated Approach ...... 13

2 References...... 22

11 LIST OF ILLUSTRATIONS

1. Map: , Sandwick Parish, . 15 2. Magnetometer survey on north side of Skaill Bay, 2003–06...... 17 3. Mound of Snusgar: topographic survey model (viewed as from SE)...... 18 4. Magnetometer survey near , 2003. 20

1 The archaeology of sand landscapes: LOOKING for an integrated approach

The coastal dune systems and semi-stabilised Dundrum, Co. Down (Collins 1952, 1959). The style machair of represent a vast but still largely of the time was to treat archaeological exposures in unquantified archaeological and palaeoenvironmen- sand as discrete sites and to excavate them using tal resource. They are an inherently fragile, unstable standard terrestrial techniques – there being as yet and changing environment and therefore highly little understanding of the effects on archaeologi- susceptible to erosion, providing a significant man- cal material of the mechanisms of post-depositional agement dilemma for curatorial authorities. Few change in this type of environment. The very fact such landscapes, even in the wider context of Britain that erosion and dune deflation provided the and Ireland, have yet been subjected to anything main means of identifying deposits which were like a comprehensive characterisation study, encom- subsequently investigated, means that the associ- passing prospection, sampling and investigation, ated data-capture was severely affected by in situ resulting in well-informed selective preservation or destruction, migration and turbation of deposits. mitigation strategies. Practical problems encoun- Associations of artefact and stratigraphic context tered during traditional excavation such as dune are therefore potentially suspect, and furthermore, instability and deflation, deep and unstable sand many radiocarbon determinations, particularly from overburdens and the consequent danger of breaking the earlier decades of this technique, are affected by up stabilising ground cover and accelerating growing doubts over the Marine Reservoir Effect erosion, have meant that we have remained largely and other distorting factors (Ascough et al 2004). dependent on chance exposures to determine the Consequently, through no fault of the excavators, extent of archaeological features and therefore of these interventions must generally be regarded as any intervention. However, increasing utility, ver- providing less reliable and complete a record than satility and affordability in combined prospection many contemporary interventions on ‘solid ground’. methods have begun to point towards new ways Therefore, the broad approach in the less recent forward in tackling this type of landscape. past to these types of landscape could be character- ised as: My own interest in sand landscapes was initially sparked some years ago by an aspect of my PhD • ad hoc research on Early Historic coastal trade and set- • reactive to erosion tlement (Griffiths 1992) which brought together a • responsive to in situ exposures of visible surprising number of antiquarian, and also more contexts leading to small-scale rescue or salvage recent, reports of stray finds of artefacts in coastal excavation sand dune areas in the Irish Sea Region and in • often dealing with already severely compro- Atlantic Scotland and Ireland (see also Griffiths mised deposits forthcoming 2009). These vary considerably in date, • focused on finds retrieval detail and apparent quality, but a common factor is the presence in close proximity of ceramic, lithic or By the 1970s, growing acknowledgement of the metalwork material from widely differing periods, limitations of a site-specific approach, coupled with often from the Neolithic or Bronze Age together realisation of the potential for reconstructing past with Early Historic or Medieval objects. Examples landscapes and environments, led to a broadening of include the Culbin Sands, Moray (Black 1891), and approaches encompassing more systematic survey Stevenston Sands, Ayrshire (Callander 1933). A and observation, selective excavation and pal- further extremely important and rich site of this type aeoenvironmental sampling, combined with more is Meols, Wirral, north-west England (Hume 1863; sustained and balanced radiocarbon programmes. Travis 1922; Kenna 1986; Griffiths 2001; Griffiths et A significant example of progress in tackling the al 2007). Pre-twentieth and early twentieth-century archaeology of sand landscapes is the work of Trevor reports are characterised by little or no contextual Cowie and others at Torr’s Warren, Luce Sands in information, concentrating exclusively, or almost Galloway (Cowie 1997). This landscape, character- exclusively, on describing the artefacts themselves. ised today by a vast, and now largely stable dune By the mid 20th century, chance discoveries of system, is one of the most archaeologically tanta- artefacts began to be accompanied by small-scale lising and intractable in Scotland, having produced excavation, often revealing spreads of occupa- over the decades at least 8,000 artefacts of pre- tion debris, hearths and middens, cultivation soils historic, Roman and Medieval date. Yet years of and ard-marks, and even occasionally fragments faithful and determined observation and collecting of buildings, such as at Luce Sands, Galloway by a handful of devoted local archaeologists had (Davidson 1952, Jope 1959), and Murlough Dunes, not succeeded in establishing a coherent overall

13 context for these objects, despite the occasional Sheffield University in 1988 in the southern islands discovery of individual structures and burials. The of the Western Isles. Its adequacy as a model for work of the Central Excavation Unit in 1977–79, future approaches is discussed below. An account led by Cowie, was confined to a development area and bibliography of this project is provided in these of 75 hectares within a dune landscape of over proceedings (Parker Pearson et al). 1,200 hectares. However, the sample covered by One important conclusion which flows from studies the investigation (approximately 6.25%) was large at these sites, and a range of other recent fieldwork enough to begin to create a true landscape perspec- (see elsewhere below), is that humans have actively tive. Machine-clearing of overburdens of up to 15m modified this type of landscape in the past. The in depth allowed the excavation of buried palae- build-ups of soil forming ancient land surfaces now osols which had been exposed in section by dune buried within dune sequences are not always by- deflation and blow-outs. The erosive mechanism by products of periods of ‘natural’ landscape stability which artefacts and other archaeological matter – they are evidence of the deliberate creation of had accumulated in blow-out hollows was more stasis by the spreading of pedogenic materials. In clearly documented than has been the case previ- this case, therefore, we are dealing with a relative ously in this or any similar landscape (ibid, 15). scale of past human agency – decisions to persist Buried palaeosols associated with archaeological with soil stabilisation or to abandon the process, are material were interlaminated with layers of wind- conditioned by locally specific factors and affected by blown sand; pollen analysis from dune slack area varying rates of continuing erosion and wind-driven (the results of which extended back only c 3,000 inundation by sand (the ‘Bad Year Syndrome’ affects radiocarbon years) confirmed that changes in these landscapes as much as any other, if not more vegetation had been very marked, showing alter- so). Identifying the tipping-points of agricultural nating predominance of woodland and heath since and settlement feasibility, and winning or losing the Bronze Age. The consequent interpretation was the struggle against the sand, is a critical aspect that settlement had taken place in the Neolithic of our search for explanations for the interrupted and the Bronze Age, and subsequently in the Early character of human settlement in these areas. It is to High Medieval periods, on stabilised agricultural therefore critical that we have an informed grasp soils which were separated by intervals of gradual, both of the general geomorphological background but ultimately extensive, aeolian sand inundation and also its local specifics – so building a true beginning in the mid to later second millennium multi-disciplinary landscape perspective is critical bc, and succeeded by another more sudden and here. In dune systems, past land surfaces bearing catastrophic episode in the Later to Post-Medieval archaeological and palaeoenvironmental informa- period, coinciding with the well-known effects of tion of great importance are interlaminated with climatic downturn at that stage. aeolian deposits which have very different accu- A further example of the growing ‘landscape mulation and erosional characteristics in terms of approach’, albeit still conditioned largely by visible both materials and durability. Subsequent geomor- exposure rather than prospection-based methods, is phological changes have severely compromised the the campaign of research mounted by Durham Uni- integrity of these deposits, but quantifying this is an versity at Freswick Links, Caithness, between 1979 extreme challenge to conventional reconnaissance and 1984 (Morris et al, 1995). This involved a series techniques. of sample excavations along a bay coastline badly The approach of curatorial authorities such as affected by aeolian erosion, wartime disturbance, Historic Scotland and English Heritage in more rabbit burrowing and sand quarrying, and where recent times is to recognise the need for informed (as with other sites of this type), there had been management strategies based on an effective a number of previous small archaeological exca- method of auditing the survival and loss of resource vations. The Freswick project included a detailed – and it is this need for coherent and statistically and analytical topographic survey extending 250m viable bodies of data from both exposed and eroded, north–south along the east-facing bay within an and unexposed and intact contexts, which presents overall control grid, which in itself represented an archaeologists with their current challenge. The advance on the piecemeal approaches taken pre- most pressing need now is to continue to develop viously. Column sampling, augering, small-scale a suite of prospection methods which can move excavation, cleaning and recording exposed sections, well beyond visual exposure as the starting point all provided a broad integrated stratigraphic for the archaeological response. An effective means model for the interleaved blown sand and occupa- of ‘seeing beneath the sand’ is needed not only to tion layers encountered on the links. An extensive find, locate, characterise and map a new realm of environmental programme was put in place which archaeological information currently more or less added considerably to the small stock of existing out of our reach, but also to set the state of known knowledge about Late Norse exploitation of marine archaeology (including exposed traces along coast- resources in northern Scotland. lines, records of former excavations and upstanding The most substantial and integrated approach to earthworks which have yet to be investigated) into sand-covered landscapes (and adjacent sand-free a more comprehensive data-set covering the whole areas) has been the SEARCH project initiated by landscape under study.

14 Illus 1. Map: Bay of Skaill, Sandwick Parish, Orkney

The various projects discussed below are all of direct was driven, apart from basic requirements such as interest to the wider question of how we build a more permissions for access and likely geological feasibil- representative and informative sample of archaeol- ity for magnetic survey (afforded in this case by the ogy in aeolian areas. predominance of the Old Red Sandstone, see Mykura 1976), by the aim to test and exceed A recently begun research project of my own at Birsay the scope of existing archaeological data. A short Bay and the Bay of Skaill, two extensive low-lying but intensive programme of magnetometry survey, and sandy erosive openings in the Atlantic frontage topographical survey and (in Birsay only) magnetic of the West Mainland of Orkney, has started to build susceptibility topsoil mapping was carried out in up a wider landscape perspective using integrated June–July 2003 with funding from Orkney Islands geophysical and topographical survey, around an Council. In 2004–08, with funding from Historic archaeological landscape characterised previously Scotland and with further support from Orkney by small-scale reactive excavations in the coastal Islands Council, the geophysical survey across the erosion zone. northern side of the Bay of Skaill was extended The choice of field research locations in Orkney substantially, combined with ground probing radar

15 (GPR) survey, and small-scale excavations at the centre of the low-lying bay hinterland is a group Point of Buckquoy, Birsay, and at the ‘Castle of of tumuli, at least one of which was opened by Sir Snusgar’ on the north side of the Bay of Skaill, which Joseph Banks in c 1772 (NMRS HY21 NW15). To tested the nature of the 2003 results in more depth the north of the bay is a scatter of discrete sites and detail than geophysics can provide on its own. including a broch, the ‘Knowe of Verron’ (NMRS This project is ongoing and a recent interim report HY21 NW22) and an enigmatic mound or knowe covers the details of the results of the first season known as the ‘Castle of Snusgar’ (NMRS HY21 (Griffiths 2006). NW21, NGR 2361 1960), near which was discov- Birsay Bay and the Bay of Skaill are both locations ered the tenth-century Skaill Viking silver hoard, of known significant archaeological potential – in a chance find in 1858 brought about by a local man the case of Skaill extending to part of the Orkney delving into rabbit burrows which had colonised World Heritage Site (WHS) buffer zone around the a thick layer of sand on the flanks of the mound designated site of Skara Brae on the south side (Graham-Campbell 1995). Further inland to the of the bay. Yet in both cases, the current state of east, on a higher zone of the Sand Fiold, vehicle archaeological knowledge is highly site-specific, movement associated with sand quarrying acci- based on small-scale intensive excavation strate- dentally revealed a rock-cut pit within which was a gies in the past, and the ad hoc discovery of stray cist grave (NMRS HY21 NW35, Dalland 1999). The finds and deposits. The excavation strategy of the wider context within the Sand Fiold of this unusual 1970s, whilst successful in that it produced a high- prehistoric structure has not yet been explored. quality record of the sites investigated (Ritchie 1977, Morris 1989), was essentially reactive to The Snusgar mound has been the focus of continuing rapid erosion patterns and dependent on an acute research in 2003 and 2004 (Griffiths 2006). Snusgar ‘rescue’ ethos. Moreover, the area percentage of is mentioned in NMRS, but is otherwise unsched- the archaeological landscape investigated in detail uled and falls outwith the WHS buffer zone. There by this means was tiny. The predictive element are even some cautionary suggestions that the which might be afforded by a more detailed insight mound may be a natural geomorphological feature into the archaeological potential of the landscape, (eg Morris 1985, 85) .The ‘castle’ name is probably within and away from the immediate erosion zone, connected to the visible remains of a stone building was not readily available at the time. Two to three which was noted in 1795 (Old Statistical Account of decades on from the hey-day of coastal rescue exca- Scotland 16, 458), standing on the north-west sector vation in Orkney West Mainland, greatly enhanced of the mound. Norse ‘castles’ in the form of small techniques of prospection offer the opportunity not masonry fortifications do exist in Orkney, but there only to effect new and informative data (which, by is no conclusive evidence to support the identifica- concentrating on the ‘gaps’ between known sites, tion of one here. Certainly no in situ masonry is could not only find new foci but contribute to the visible today, but a microtopographical survey (illus re-evaluation of existing evidence), but to contrib- 3) showed a flat platform on the north-west sector of ute to an updated and informed curatorial strategy the summit with a shaped and graded slope on west which seeks to audit and manage the effects of and north, about where the ‘castle’ may have stood. coastal change, rather than merely to react to Rabbit burrowing has disturbed and turbated areas them. Testing the wider applications of landscape of the summit and flanks. Other notable features prospection in Orkney was also a factor; several mapped here include a north-east spur of the mound surveys have taken place in and outwith the WHS above a depression to the east where a spring is area, but, although growing, the proportion of located, and a dry gully to the north-west dividing coverage is still lower than for other high-profile the Snusgar mound from another topographical rise, archaeological areas of Britain such as Wessex. albeit a lower and less well-defined one. For the purposes of this paper, given the scope Geophysical investigation in 2003 comprised and coverage of the project so far, the Bay of Skaill approx. seven 30 × 30m grids of magnetometer is the more appropriate case-study, being a true survey both on and off the mound flanks and aeolian landscape, whereas the Point of Buckquoy summit. The area of sandy scrub around the base at Birsay (where the survey has hitherto been con- of the mound proved unresponsive – presumably centrated), although a classic eroding landscape, because any measurable deposits are masked by lacks the same extent of wind-blown accretions. more than 1m of aeolian sand. However, the upper The central and southern hinterland of Birsay Bay part of the Snusgar mound itself proved to be a is of course another classic example of an aeolian viable prospect for magnetometry survey, producing landscape, but one which has yet to be studied in results of dense contrast, revealing it to be a highly detail as part of this particular project. The Bay complex, and apparently largely archaeological of Skaill, which is also an active erosion zone, is structure (for a detailed overview of the results, see characterised by higher ground to the north and Griffiths 2006). In 2003, some limited further recon- south with fresh water sources, which acted as naissance and magnetic scanning of the area to the past settlement attractors. The centre of the bay is north-west of the Snusgar mound, on both sides of characterised by extensive blown-sand deposits at the road, suggested that these midden-type deposits Sand Fiold, some of which is improved land. In the may be more widespread.

16 Illus 2. Magnetometer survey on north side of Skaill Bay, 2003–06 Magnetometer survey on north side of Skaill Bay, Illus 2.

17 Illus 3. Mound of Snusgar: topographic survey model (viewed as from SE)

Further geophysical survey in July–August 2004 will be reported more fully elsewhere, but prelimi- comprised resistance and magnetometry surveys nary results on Snusgar itself indicated that there carried out in conjunction with Dr Susan Ovenden, was a band of higher resistance partly surround- the recently appointed geophysical field officer at ing the base of the mound on the east and south Orkney College/UHI. The mound of Snusgar itself, sides – the reason for which is as yet unknown. A including the area of coverage of the 2003 mag- spread of lower resistance characterised the top of netometry survey, was surveyed using a Geoscan the mound, but coherent hints of structures were RM15 resistance meter. This was undertaken partly only visible in the north-west sector of the mound, in order to test the theory that the ‘castle’ mound where excavation was not planned to take place in was a shaped or even fortified structure, and also to 2004. The 2004 magnetometry survey showed that provide control for the 2003 magnetometry results. the mound or rise immediately to the north-west of Further magnetometry survey was carried out Snusgar is of considerable potential archaeological over the neighbouring mound or topographical rise interest, confirming our 2003 field observation of which is located across a narrow dry gully to the hints of midden layers and masonry rubble exposed north-west of Snusgar, and where limited augering in rabbit holes. Very much as at Snusgar itself, a had shown a further concentration of midden-type dense cluster of contrasting anomalies was mapped deposit in 2003. Two further large mounds, one yet on this neighbouring mound. The mound further to further to the west (NGR HY 2345 1970), which is the west produced some potential archaeological cut by the road, and one to the east of Snusgar (HY anomalies, but the survey response was compro- 2075 1960), were also surveyed using magnetometry. mised by high levels of later surface disturbance The westernmost mound, cut by the present Skaill– and litter caused by the fact that a large part of Quoyloo road, attracted our attention because it had its generally flat surface has been used in recent been the scene, when the road-cut was created in c memory for animal burials and also (unwittingly) 1934, of the reported discovery (alas going largely as an unofficial tourist campsite. unrecorded) of stone structures and burials of clear Accompanying the extended geophysical cam­paign archaeological interest (NMRS HY21 NW23). By in summer 2004 was a three-week excavation on contrast, the mound to the east of Snusgar, which the north-east flank of Snusgar aimed at under- is very badly scarred by rabbit burrowing, exposing standing the 2003 magnetometry results (Griffiths a large eroding section across its west side, was 2006). Firstly, beneath the sandy topsoil was selected as a control exercise as it was strongly evidence of relatively recent activity in the form suspected of being all, or largely, a sand dune (this of two c 1m-deep animal burial pits (estimated to assumption turned out to be erroneous, see below, date from within the last 150 years), and a discrete with further investigations producing substantial spread of dense red-black burnt soil with coal archaeological features). inclusions, located above small and crude bowl Work on the geophysical data is still ongoing and pits. A probable explanation for the latter is kelp

18 burning – which occurred widely in Orkney in the view, but still commemorated of course by the 19th century. Secondly, below and outwith the rela- ‘brae’ name). Another large mound some 100m to tively discrete imprint of these modern intrusions, the west is currently undergoing severe erosion the character of the upper layers of the mound (Morris 1985). Half or more of this mound has now began to emerge in the form of a series of peat-ash gone, leaving only a landward portion behind a full- layers within stone-layered features which were height semi-vertical erosive section which is visible laminated within thin bands of wind-blown sand. facing the bay, in which are exposed layers of stone Although some of the stone features as yet lack masonry and middens interleaved with deep sand clear definition, there was clear evidence of sub- accretions (perhaps, one imagines, bearing some stantial double-faced walling running east–west similarity to the appearance of Skara Brae c 150 across the area, which was congruent with the years ago). clearer traces evident in the north-west sector of Consent to carry out survey within the Skara Brae the 2003 magnetometer plot. The peat-ash layers section of the Orkney WHS buffer zone was obtained, and associated stone features contained large but targeted prospection was also deployed on a wider quantities of well-preserved animal bone and were basis in an attempt to build a more comprehensive provisionally dated by a range of stone, bone and landscape overview for the bay. Within such a large antler artefacts to the Early Middle Viking period area, a survey target zone comparable to the more (c 800–1100 ad). Further work in 2005 and 2006 coherent Buckquoy Peninsula did not immediately showed that Viking-period material constitutes the suggest itself, so work began by targeting known upper part of an archaeological ‘core’ within the locations of archaeological potential and exploring mound (Griffiths 2004–05). Midden material had their environs. A geophysical survey had been been used to stabilise successive occupation and conducted here in 1973 by A Clark and A Bartlett cultivation layers, which were regularly inundated of the Ancient Monuments Laboratory, in the field by incoming wind-blown sand. This has given us immediately to the south-west of Skara Brae. Their a significant insight into the process of mound survey, from which an unpublished stack-trace plot formation, where human intervention in building remains on file in the Sites and Monuments Record up stable and cultivable surfaces using domestic (SMR) at Orkney Archaeological Trust, was incon- refuse produces a favourable situation in the short clusive, but did show a very pronounced linear term. However by doing so, this activity exacerbates anomaly running WSW from Skara Brae. This the problem of creating an upstanding sand-trap, proved to be an igneous dyke apparently running which ultimately makes the situation worse in the beneath the southern edge of Skara Brae. Whilst long run by inflating the mound and its flanks with obviously counterproductive to magnetic reconnais- further sand accumulations. sance, the presence of this feature is not without The results from the further mound some 100m archaeological interest in such close proximity to a to the east of Snusgar produced only a very general major Neolithic site. expectation that it might contain any archaeological potential, in the form of a generally raised magnetic The southern 30 × 30m grid in the survey closest to signature but without any apparent structural Skara Brae did show some hints of stone archaeo- coherence. However, a small 10 × 5m test trench logical features, but these were partly masked by was dug here in 2005, and this, which was extended extensive ferrous metal contamination and modern in 2006 to 2008, has revealed substantial and well- drainage installations (this area is very close to preserved stone buildings which have been dated to Skaill House Farm, and the low cliff-tops and dunes the Viking/Norse period (see www.conted.ox.ac.uk/ to the west of Skara Brae are also scattered with research/birsay-skaill). These had been abandoned bits of fishing equipment, fencing and parts of old and filled up with wind-blown sand to a depth of sheds). However, the mound to the west did show 1–1.5m, which explains why the magnetometer had more promising results. A Viking burial was found registered a general spread of potential here but had here in 1888 and a cist grave in 1994 (James 1999, been unable to filter out the structural pattern later 771 ff). The grey-scale plot shows a reasonable indi- revealed by excavation from the local background. cation of the extent and shape of the southern part of the mound, and there are hints of discrete secondary Skara Brae, at the southernmost point of the structures remaining in its periphery, which would bay’s circumference, was discovered in 1850 after accord with the previous finds of burials. The 2003 marine erosion of aeolian sand deposits which had Skara Brae Magnetometer Survey has since been built up over the site in prehistory. The Neolithic superseded by a much more extensive survey ‘village’ (perhaps more accurately described as a carried out by Orkney College Geophysics Unit on ‘tell’), which was largely exposed and cleared of behalf of Historic Scotland and the WHS manage- overburden under the archaeological supervision ment programme. of V Gordon Childe in 1929–30 (Childe 1931), and is now partly reconstructed and in guardianship, In the right conditions and as far as possible free of was previously deeply buried in a combination of relatively recent ferrous contamination from fishing archaeological and aeolian sand horizons within and barbed-wire fencing, and spreads of kelp- an upstanding mound (itself now all but gone from burning detritus (which often limits its potential in

19 Illus 4. Magnetometer survey of Bay of Skaill, near Skara Brae, 2003

coastal margins), conventional fluxgate magnetom- etrates up to 6m deep – and is easy and quick to use etry can work wonders in up to 1m of blown sand. when combined with GPS. This helps us to realise This is demonstrated by an example from Harlyn a need for a faster and more extensive prospection Bay, Cornwall, which has now been ground-truthed technique which could more effectively translate in the sense that it was fully excavated for a pipeline into a three-dimensional digital terrain model. – revealing beneath the sand extant Medieval ridge Current work therefore shows the potential, and furrow overlying Iron Age round houses; here, but also the limitations of techniques of conven- magnetic susceptibility topsoil mapping might have tional magnetometer survey coupled with topsoil worked in probe form but the 100mm surface coil is magnetic susceptibility mapping and topographi- too shallow to reach the buried palaeosols (Oxford cal survey. This is very good for targeting surface Archaeotechnics, see http://dialspace.dial.pipex. exposures and settlement mounds, but where sand com/town/terrace/ld36/grad.htm). One way forward depth reaches over c 1m, its coverage becomes might be the Geonics EM31 Electrical Conductiv- more patchy and difficult to interpret, leading ity system (www.geonics.com), used to some effect in to a loss of ability to model a coherent landscape desert sites in the Middle East (such as the National sample. A sustainable methodology for modelling Museums of Scotland Saqqara Project in Egypt and ground-truthing both the visible and sub- – see www.nms.ac.uk/royal/saqqara), which pen- surface deposits is needed, which combines the

20 potential for visualising archaeological deposits • developing sustainable methodologies for both in demonstrable areas of potential and also ground-truthing and dating the resource. in the ‘gaps’ between. Moreover, this must be one • providing an enhanced and more widespread which in the case of sand areas does not in itself multi-disciplinary understanding of ongoing increase dune instability. geomorphological processes and potential for For an integrated approach, therefore, the way sudden change. forward would include a combination of the following • auditing preservation and loss of deposits. methodological objectives: The proceedings published below provide a resource • building rapid and cost-effective means of of experience, results and references which are modelling the archaeological and palaeoenvi- intended to promote the four objectives as outlined ronmental resource through survey and GIS. above.

21 2 References

Ascough, P L; Cook, G T; Dugmore, A J; Barber, J; Griffiths, D 2008 ‘Birsay-Skail landscape archae- Higney, E; Scott E M 2004 ‘Holocene Variations ology project, Orkney’, www.conted.ox.ac. in the Scottish Marine Radiocarbon Reservoir uk/research/projects/birsay-skaill. Effect’, Radiocarbon 46, 611–620. Griffiths, D forthcoming 2009 ‘The Sand Dune Sites, Black, G F 1891 ‘Report on the Archaeological Evidence for Pre-Viking trade’ in Graham- Examination of the Culbin Sands, Elginshire’, Campbell, J A; Ryan, M (eds) Ireland and Proc Soc Antiq Scot 25, 490–95. Britain in the Pre-Viking period, Proceedings Callander, J G 1933 ‘A Collection of Prehistoric of a Symposium held at the British Academy, Relics from the Stevenston Sands, Ayrshire and October 2005, British Academy/Royal Irish Other Objects in the National Museum’, Proc Academy. Soc Antiq Scot 67, 26–34. Hume, A 1863 Ancient Meols, or Some Account of the Childe, V G 1931 Skara Brae, A Pictish Village in Antiquities found on the Sea Coast of Cheshire, Orkney, Kegan Paul, London. London. Clark, A; Bartlett, A 1973 Geophysical Survey at James, H 1999 ‘Excavations of a Medieval Cemetery Skara Brae, AM Laboratory, unpub. Ms. at Skaill House, and a Cist in the Bay of Skaill, Collins, A E P 1952 ‘Excavations in the Sandhills at Sandwick, Orkney’, Proc Soc Antiq Scot 129, Dundrum, Co. Down 1950–51’, Ulster Journal of 753–77. Archaeology 15, 2–26. Jones, R; Sharpe, L (eds) 2006 ‘Going over Old Collins, A E P 1959 ‘Further Investigations in the Ground, Perspectives on Archaeological Geo- Dundrum Sandhills’, Ulster Journal of Archae- physical and Geochemical Survey in Scotland’, ology 22, 5–20. British Archaeological Reports, British Series Cowie, T G et al 1997 ‘Torrs Warren, Luce Sands, 416, Oxford. Galloway: A Report on Archaeological and Palae- Jope, E M 1966 An Archaeological Survey of County oecological Investigations Undertaken in 1977 Down, HMSO, Belfast. and 1979’, Transactions of the Dumfriesshire Jope, E M; Jope, H M; Stewart, I H; Thompson, J and Galloway Natural History and Antiquarian D A 1959 ‘A hoard of fifteenth-century coins Society 71, 11–105. from Glenluce sand dunes and their context’, Dalland, M 1999 ‘Sand Fiold: The Excavation of an Medieval Archaeology 3, 259–79. Exceptional Cist in Orkney’, Proceedings of the Kenna, R J B 1986 ‘The Flandrian Sequence of North Prehistoric Society 65, 373–413. Wirral (NW England)’, Geological Journal 21, Davidson, J M 1952 ‘Report on the Discoveries at 1–27. Glenluce Sands, Wigtownshire’, Proc Soc Antiq Morris, C D 1985 ‘Skaill, Sandwick, Orkney: Pre- Scot 86, 43–69. liminary Investigations of a Mound-site near Graham-Campbell, J A 1995 The Viking Age Gold Skara Brae’, Glasgow Archaeological Journal and Silver of Scotland, AD 850–1100. National 12, 82–92. Museum of Scotland, Edinburgh. Morris, C D 1989 ‘The Birsay Bay Project, Volume 1’, Griffiths, D W 1992 ‘Anglo-Saxon England and the Department of Archaeology Monograph Series Irish Sea Region, ad 800–1100’ Unpub. PhD thesis, 1, University of Durham. Dept of Archaeology, University of Durham. Morris, C D; Batey, C E; Rackham, J 1995 Freswick Griffiths, D 2001Great Sites: Meols, British Archae- Links, Caithness: Excavation and Survey of ology Issue 62, Dec 2001, 20–25. Norse Settlement, Historic Scotland/NABO/ Griffiths, D 2004–05 ‘Birsay-Skaill Landscape Highland Archaeology Monograph No. 1. Archaeology Project’, Discovery and Excavation Mykura, W 1976 British Regional Geology, Orkney in Scotland, New Series 5 (2004–05), 94–5. and Shetland, HMSO, Edinburgh. Griffiths, D 2006 ‘Birsay and Skaill Orkney, Old Statistical Account 1791–99 The Statistical Landscape Survey 2003–04’, in Jones, R; Account of Scotland, drawn up from the com- Sharpe, L (eds) 2006, ‘Going over Old Ground, munications of the ministers of the different Perspectives on Archaeological Geophysical parishes 15 (Orkney), Edinburgh. and Geochemical Survey in Scotland’, British Parker Pearson, M; Mulville, J; Sharples, N; Smith, H Archaeological Reports, British Series 416, 2008 ‘Archaeological remains on Uist’s machair: Oxford, 213–24. threats and potential’. SAIR 48, Society of Anti- Griffiths, D; Philpott, R; Egan, G 2007 ‘Meols: The quaries of Scotland, Edinburgh. Archaeology of the North Wirral Coast’, Oxford Ritchie, A 1977 ‘Excavation of Pictish and Viking- University School of Archaeology Monograph age farmsteads at Buckquoy, Orkney’, Proc Soc Series 68. Antiq Scot 108, 174–227.

22 Travis, W G 1922 ‘On Peaty Bands in the Wallasey Sandhills’, Proceedings of the Liverpool Geologi- cal Society 13, 207–214.

23