Journal of the North Atlantic T. Dawson 2015 Special Volume 9 83 Introduction This paper shows how threats to the coastal archaeological resource of Scotland's coast have been managed in recent years. Although using the Western Isles to provide examples, the range of threats described and the development of management strategies employed are applicable to coastlines around the globe. Coastal erosion is one of the gravest natural threats to archaeological heritage, and Scotland is particularly affected due to its large number of wellpreserved sites. Regardless of what may happen in fifty or one hundred years (with climate change predictions and fears of sea-level rise suggesting the worst), thousands of sites are at risk now. Erosion is a natural process, meaning that legislation cannot forbid damage and there is no developer who can be asked to pay for recording in advance of destruction. Thus, Historic Scotland, the Scottish Government’s agency charged with safeguarding the nation’s historic environment, has taken the lead in devising management strategies. Historic Scotland has worked with other organizations to gather data, devise plans, and undertake action at threatened coastal sites. Two of its principal partners have been the University of St Andrews and the Scottish Coastal Archaeology and the Problem of Erosion (SCAPE) Trust, a charity established in 2001 to promote action at eroding archaeological sites. This paper outlines the scale of the threat posed by coastal erosion and discusses how management strategies have developed. Using examples from the Western Isles (or Outer Hebrides), it outlines how the problem of erosion has been approached, demonstrating the effectiveness of Coastal Zone Assessment Surveys (CZAS) aimed at collecting data for heritage management. These surveys have located thousands of previously unrecorded sites and have provided recommendations for further action. The paper discusses the effectiveness of deskbased assessments (DBA) at prioritizing areas for subsequent ground survey, and analyzes the results of two DBAs that attempted to locate new sites from aerial photography. It shows how the data gathered in the CZAS and other surveys has been used to produce prioritized lists of sites where further action is required. The paper concludes by showing how local communities can play a crucial role in the management and recording of threatened archaeological sites. Archaeological Sites and the Coast Scotland has the second longest coastline in Europe (see discussion below), partly due to the large number of islands situated off the north and west coasts. The geography and topography of Scotland have encouraged coastal habitation, and the benefits of coastal settlement have included a proximity to marine resources, a transportation route by boat, and access to materials washed up or exposed on beaches. The coastal zone not only contains most of the site types that are found inland, but is also home to remains that had specific maritime-related functions. Although few studies have been undertaken to show if there is a greater density of archaeological sites at the coast edge in comparison to the hinterland, an analysis was made by researchers from the University of Sheffield during their study of the Hebridean island of Barra (Branigan 2005:63–72). Surveys between 1988 and 1999 determined that the coastal zone made up just 4 per cent of the total area of the Eroding Archaeology at the Coast: How a Global Problem is Being Managed in Scotland, with Examples from the Western Isles Tom Dawson* Abstract - Many thousands of archaeological sites are threatened by coastal erosion around the globe. The problem is particularly grave in Scotland, where a number of management strategies have been developed. Much of this work has been undertaken by Historic Scotland working in partnership with the SCAPE Trust and the University of St Andrews. This paper outlines the scale of the problem presented by coastal erosion, using recent work in the Western Isles to provide examples. It shows how action has developed over the years, with heritage managers often requiring a different approach to other coastal managers. The effectiveness of desk-based assessment and coastal survey is reviewed, and the results of two desk-based assessments in areas that were subsequently surveyed are analyzed. The paper outlines how prioritized lists of sites requiring future action have been produced and describes the important role that community groups can play in heritage management, giving examples of practical projects that include the community excavation of a site threatened with destruction. Special Volume 9:83–98 2010 Hebridean Archaeology Forum Journal of the North Atlantic *University of St Andrews, St. Katherine’s Lodge, St Andrews, KY16 9AL Scotland, UK; tcd@st-andrews.ac.uk. 2015 Journal of the North Atlantic 84 T. Dawson 2015 Special Volume 9 island, yet contained 23 per cent of all the recorded sites. Branigan (2005:68) cautiously concluded that “the coastal zone is archaeologically rich and was a preferred zone for at least some types of human activity in some periods in the past”. In some areas around the Scottish coast, access to durable building materials combined with local geomorphology has meant that sites have survived well, making the archaeological resource extremely rich. Preservation is particularly good in areas where sand has inundated structures, protecting archaeological layers from the plough, hiding stone from “robbers”, and supporting drystone walls that stand to almost full height. For example, in the Northern and Western Isles, a lack of timber in areas where there was easily accessible stone at the coast edge has resulted in the survival of entire villages buried under the sand. Recognition of the cultural value of two sites formerly buried within sand dunes has been demonstrated by the award of World Heritage status to Skara Brae, Orkney and the inclusion of Jarlshof, Shetland on the UK Tentative List for World Heritage Sites (Fig. 1; DCMS 2011). Although sites may have survived burial for millennia, they are not immune to future destruction. Many archaeological sites by the sea are vulnerable to erosion, especially if located close to the coast edge. Both Skara Brae and Jarlshof were originally revealed due to erosion and both are now defended by sea walls in an attempt to protect them from further damage (Fig. 2). Although a handful of sites have been recorded or protected, thousands of others have been neither excavated nor defended and remain vulnerable to complete destruction. Threatened site types include everything from Mesolithic shell middens to Iron Age brochs, Norse settlement sites to military structures from the two World Wars (Fig. 3). The Coast and Erosion Coastal erosion has been identified as one of the three major threats to Scotland’s archaeological heritage not covered by planning guidance (Barclay 1997:17). Although erosion is a natural process, it affects an area’s cultural heritage differently than it affects its natural heritage. Animals and birds can move and plants can recolonize cliff edges. Archaeological sites however are static, and any damage or loss to them cannot be undone. Lees (2005:19) noted at least fifteen factors which could affect the rate of erosion, including Figure 1. Iron Age structures and the Norse village at Jarlshof in Shetland, excavated after being revealed by storms in the nineteenth century. Journal of the North Atlantic T. Dawson 2015 Special Volume 9 85 Figure 2. Workmen extending the coastal defense at the Neolithic village of Skara Brae, Orkney. Continuing erosion of the coast edge has meant that the defenses need to be maintained and extended regularly. Figure 3. Part of the eroding prehistoric settlement and cemetery site at Galson, Lewis, Western Isles. Archaeological material has been reported eroding from the dune since the early twentieth century, and while much has been destroyed, new structures are still being exposed. Journal of the North Atlantic 86 T. Dawson 2015 Special Volume 9 the nature of the sediment that overlies the bedrock. Scotland’s geology is dominated by hard, ancient rocks in the west and softer, younger sedimentary rocks in the east. The igneous rock coasts are the toughest, but even sandstone coasts are relatively durable. In the Western Isles, the rock foundation is generally stable, but the soft sediments that overlie the bedrock are at risk from erosion. For example, the blown-sand landforms known as machair, which extend along much of the west coast of the Western Isles, are threatened by both sea and wind. The sand is highly mobile, and dry weather combined with strong winds can cause large amounts of sand to blow, resulting in large craters, or “blowouts”. The potential for severe erosion in these locations is especially problematic, as machair areas appear to have been a focus for settlement for millennia. For example, Simpson et al. (2003:185) noted that numerous Early Bronze Age sites had been recorded on the machair plain, leading to a suggestion that it was “eminently suitable for the establishment of early settlement”. These sites are extremely rare in other parts of Britain, and their threatened loss by erosion is a cause for concern. Another factor that affects erosion rates is exposure to wave attack exacerbated by high winds, and the entire Scottish coast is vulnerable to violent storms. Damage is proportional to the size and power of the waves, and this is increased by two factors: wind strength and fetch (the distance over which the wind can travel uninterrupted over the sea). Exposed areas of coast, not protected by offshore islands, are often most vulnerable. The west coasts of the Western Isles are fully exposed to Atlantic storms, and low-lying areas can be devastated, especially when the wind is combined with high spring tides. Managing Coastal Archaeological Sites As a first step to planning for future change at the coast, it is important to know the length of the coastline and how much of it is at risk. This apparently simple-sounding job is dependent upon the scale of map used and the method of measurement. Different surveyors using different maps have obtained widely different results. The CIA World Fact Book states that the length of the coast of the entire United Kingdom is 12,429 km long (CIA 2011), whereas the Macaulay Land Use Research Institute (MLURI) estimated that Scotland’s coast alone is 16,491 km (MLURI 1993). Archaeological surveyors have also used different map scales when reporting on lengths of coast affected by erosion, meaning that comparisons between areas can be difficult when trying to obtain a national picture of the problem. Computer mapping is helping to standardize calculations of coastal length, and as part of a review of Scotland’s coastal archaeology for Historic Scotland (Dawson 2007), GIS was used to analyze the individual segments of a 1:25,000 vector map. After cleaning the map for duplicates and removing erroneous lines, the length of each segment was added, giving a total length of 16,035 km, a figure close to the estimate given by MLURI. However, this total includes small islets and rock outcrops which, from an archaeological point of view, may never have been inhabited. Removing all rocks less than 1 km in circumference (an admittedly arbitrary figure) brought the length down to 15,415 km. The contribution of the islands to the overall length of the Scottish coast is immediately obvious in Table 1, making up three fifths of the entire coastline. Of the island groups, the Western Isles has the longest coastline, with a length of just over 3000 km Many different specialists have contributed to coastal management strategies, often during the preparation of local shoreline management plans (SMPs). In Scotland, several SMPs have been prepared (e.g., Fife, East Lothian, parts of the Inner Moray Firth, Angus, and Dumfries and Galloway). These give a range of data about coastal assets and processes and make recommendations about managing erosion. With regard to erosion, coastal managers have four main options: • advance the line—construct defenses out to sea, for example, across the mouth of a bay, • defend the line—protect the coast behind a barrier, • retreat the line—breach existing sea defenses in order to create a buffer zone, or • managed retreat—accept that erosion will occur and make plans for threatened assets; the favored option for most parts of the coast. Each of the options has economic and other implications (including for the natural and historic environment), and coastal managers must consider these Table 1. Length of the Scottish coast, divided into mainland and island groups Area Length of coast (km) Mainland 6235 Western Isles 3032 Inner Hebrides 3000 Shetland 2063 Orkney 1085 Total 15,415 Journal of the North Atlantic T. Dawson 2015 Special Volume 9 87 when recommending action. For example, effective coastal defenses can be costly both to build and maintain and can deflect problems to other areas of the coast, causing erosion elsewhere. The option recommended by coastal managers is normally driven by a perception of the “value” of what is threatened. Protection (advancing or defending the line) is more likely in areas of greater economic or political significance, e.g., population centers or key transport routes. A small number of archaeological sites have been protected in the past, (for example, St Andrews Castle in Fife, Skara Brae [Fig. 2] and the Broch of Gurness in Orkney), but the majority of sites do not have an obvious or quantifiable value for most coastal planners and are unlikely to attract resources to construct defenses. In general, it has been built-heritage managers rather than coastal specialists that have taken responsibility for managing threatened coastal archaeological sites. In Scotland, Historic Scotland has taken the lead in developing plans and gathering information. In the 1970s and 1980s, Historic Scotland’s predecessors commissioned a small number of detailed archaeological surveys along stretches of the coast. For example, parts of the Western Isles were examined in 1978 (Shepherd and Shepherd 1978) and Caithness was surveyed twice between 1980 and 1982 (Batey 1984, Mercer 1981). Heritage managers also received reports about eroding sites, and in 1983 the Central Excavation Unit (Scotland), in a follow-up to the Western Isles survey, noted that 32 sites were at immediate threat from active erosion (Barber 2003:1). As will be seen below, this figure is small compared to presentday estimates, possibly as only the larger or more noticeable sites were being counted, but the report indicates that there was a growing awareness of the problem. The authors of the St. Boniface (Orkney) excavation report noted that the site was “one of over 100 sites in Orkney where coastal erosion had been identified as a major threat to the survival and longterm integrity of the archaeological resource” (Lowe 1998:13). In the mid-1990s, Historic Scotland reported that over 350 sites were under attack by coastal erosion in Orkney, Shetland, and the Western Isles alone (Barclay and Fojut 1995:1). This growing awareness of the scale of the threat led Historic Scotland to work towards a policy for the management of archaeological sites in the coastal zone. Echoing the options available to coastal managers, Ashmore (1994:34) outlined three options for built-heritage managers with regard to threatened sites: • ignore erosion (regarded by Ashmore as an unacceptable option), • defend the coastline, or • work at sites before they are destroyed. These options were refined by Barclay and Fojut (1995:3–4) in “The Management and Conservation of the Built and Maritime Heritage in the Coastal Zone”. They listed the following options for threatened sites: • legal designation, • conscious/justified abandonment, • routine recording or extensive sampling of the site as it is destroyed, • low-cost sea defense, • small-scale excavation, • high-cost sea defense, and • large-scale excavation. Barclay and Fojut (1995) recognized that, for most sites, more information was needed before an option could be chosen, and they detailed a progression of work for developing management plans as follows: • identify coastal processes that impact upon archaeological sites in order to target highrisk areas, • conduct rapid coastal surveys to identify areas at risk, • conduct detailed follow-up surveys where necessary, • prioritize action at threatened sites, and • where possible, implement appropriate solutions at priority sites. In order to fulfil the first of these targets, Historic Scotland collaborated in an innovative partnership with Scottish Natural Heritage and the Scottish Office Agricultural, Environment, and Fisheries Department (SOAEFD) to gather information on the processes that impact upon the coast. The work examined the Scottish coastal cells, and the first report was published in 1997 (HR Wallingford 1997). Coastal cells are relatively self-contained units that are usually defined by headlands and other prominent features that block the pathways within which beach sediments move. The initial publication was followed by a series of detailed reports for each of the eleven cells identified. These reports contained maps accompanied by accounts of the geology and geomorphology of the mapped stretches of coastline, details of wave and tidal regimes, littoral processes, coastal defenses, and areas of erosion and accretion. Areas of soft sediment were highlighted, and the locations of archaeological sites within 500 m Journal of the North Atlantic 88 T. Dawson 2015 Special Volume 9 In 1996, Historic Scotland attempted to standardize the data collected during the coastal surveys with the publication of Archaeological Procedure Paper 4: Coastal Zone Assessment Survey (HS APP4; Historic Scotland 1996). HS APP4 specified that the area to be surveyed should include the coast edge, the intertidal zone, a 50-m wide landward strip for detailed survey, and the area between 50–100 m from the coast edge for a less-detailed survey. It also noted the possibility of including “an offshore area in which wrecks and other underwater artefacts should be recorded” (Historic Scotland 1996:3). Archaeological sites located were recorded on 1:25,000 maps, with additional details given in gazetteers. HS APP4 specified that descriptions were to be “merely enough to characterise the sites in terms of size; complexity; nature of their contents; and relationships.” If possible, period, archaeological potential, condition, and threat from erosion were also to be provided (Historic Scotland 1996:5). Surveyors were also required to recommend future action at the site, to be selected from one of three choices: survey (to include intrusive fieldwork), monitor, or nil. The coastal geology and geomorphology were recorded using a system developed in Wales by the Gwynedd Archaeological Trust during the Aberdaron to Great Orme survey (Smith 1995). This approach was based on British Geological Survey mapping and field verification undertaken by geomorphologists working with the archaeological team. Again, reporting was provided on 1:25,000 maps, giving much greater detail than provided in the coastal cell reports. The erosion class noted was as observed on the day of survey. There are recognized problems in classifying the stability of the coast, as both erosion and deposition can be localized and can change rapidly from one status to the other. A storm can wash away loose material such as sand, causing the coast edge to retreat temporarily until wind and tidal action replaces the lost sediment (Ramsay and Brampton 2000:10–11). However, both an accreting and an eroding coast can be classified as dynamic, and therefore warrant monitoring. As is the nature with rapid surveys, only sites visible on the day of survey were recorded. For example, thick deposits of sand obscure many wellpreserved archaeological sites, and on South Uist, Sheffield and Cardiff University surveys of the machair have located over 200 ancient settlements, many of which were identified by shell, pottery, or bone found at the entrance to rabbit burrows (Parker Pearson et al. 2011). Often, the first indication that of the coast were plotted as derived from the database of the Royal Commission on the Ancient and Historical Monuments of Scotland (RCAHMS). As each coastal cell extended for tens or hundreds of kilometers, the scale of the mapping used was generally very small. For the Western Isles (cells 8 and 9), archaeological sites were not plotted individually; instead, the coast was color-coded depending upon the density of sites per 10 km (Ramsay and Brampton 2000). For the great majority of the coast, the density of sites was under ten per 10 km (less than one site per kilometer). The notable exception was Barra, where much of the island had a density of over thirty sites per 10 km. However, the high number of sites on Barra was a reflection of the recent publication of archaeological coastal survey results (for example, Branigan and Foster 1995) and their inclusion within RCAHMS records rather than an actual difference in site density. Taken on its own, the coastal cells report for the Western Isles suggested that Barra had the best preservation of coastal sites in the Western Isles, something which has not been backed up by subsequent surveys. This example demonstrates that the management of archaeological sites can be influenced by the level of detail of preceding surveys. The reporting of archaeological sites was inadequate in the coastal cells reports. The scale of the maps was too small to allow the importance of sites and landscapes to be demonstrated in any way other than site density. This problem has persisted in the publication of most of the Scottish shoreline management plans, produced at a local rather than national level. The recording of archaeological sites has usually been undertaken in the same way as in the coastal cells report, with small-scale maps showing hundreds of sites as dots or color-coded areas. The coastal cells reports and shoreline management plans proved inadequate for archaeological management, and more detailed surveys were required. This situation led to the next stage in the progression of work suggested by Barclay and Fojut (1995), the undertaking of rapid coastal surveys. Historic Scotland initiated a program of CZAS in the mid-1990s. The CZAS were intended as rapid, though detailed, surveys that gathered data on the historic environment as well as on the geology and geomorphology of the coast and the erosion class as observed on the day of survey. The Scottish CZAS were modelled on the surveys of the Welsh coast, managed by Cadw (the Welsh Government’s historic environment service) and undertaken by the four Welsh Archaeological Trusts between 1993 and 1998 (Davidson and Jones 2002). Journal of the North Atlantic T. Dawson 2015 Special Volume 9 89 hidden by the deposition of recycled sediment. Thus, vulnerable sites may not be visible on the day of survey (Figs. 4, 5), which may have been the case during the CZAS of the west coast of South Uist (see below). In order to counteract this problem, the coast needs to be monitored or surveyed immediately such deeply buried sites are present is when they are revealed in an eroding dune face or a blowout (hollow caused by wind erosion). In such cases, the site is already vulnerable to damage. Another problem with locating sites is that, due to the dynamic nature of the coast, some sites exposed during storms may subsequently become Figure 4. Archaeological remains exposed on the beach at Kilpheder , South Uist, February 2005. Figure 5. The same beach (as in Fig. 4) at Kilpheder in May 2005 showing the archaeological remains covered in sand. Journal of the North Atlantic 90 T. Dawson 2015 Special Volume 9 after damaging storms. This is an area where local communities have made a real contribution, and in Scotland, many groups have helped to locate, monitor, and record eroding sites as part of the Shorewatch project, managed by the SCAPE Trust. Shorewatch groups are composed of local people who have been given basic training in identifying archaeological sites. Supported by SCAPE, group members are encouraged to report new discoveries, which is especially useful after damaging storms. The group members take photographs and record information about potential sites that they then submit for verification. Groups have been established around the entire Scottish coast, and they have provided much valuable information, highlighting new exposures or reporting damage to known sites. In some cases, group members have desired to undertake more detailed recording or excavation projects, and SCAPE has worked with groups to obtain funding and provide professional support to achieve these objectives (see below). Between 1996 and 2011, twenty-nine full surveys (and one stand-alone desk-based assessment) were completed (Fig. 6) and over 11,500 sites were recorded. Two of the earliest surveys completed were undertaken in the Western Isles by teams from universities, with much of the west and north of Lewis being surveyed by the University of Edinburgh (Burgess and Church 1996), and a survey of Barra completed by Sheffield University (Branigan and Grattan 1998). The length of coast surveyed, broken down into island groups and the mainland, is shown in Table 2. These data show that ≈30% of the Scottish coast has been examined, with the surveys generally targeting the areas thought to be most at risk of erosion. For this reason, a greater proportion of the coasts of the island archipelagos of Orkney, Shetland, and the Western Isles have been surveyed, and almost 44% of the entire coastline of the Western Isles has been examined to date. Reports on the first sixteen CZAS, completed between 1996 and 2000, were produced as paper reports. In general, there was limited distribution of the data, with copies sent to the Local Authority archaeologists, Historic Scotland, and RCAHMS. Since 2001, the University of St Andrews and the SCAPE Trust have worked with Historic Scotland to manage the surveys and analyze the data contained within them (Dawson 2006, 2007, 2010). Surveyors have been required to provide digital copies of reports, databases, and GIS files. In 2005, SCAPE undertook the digitization and web publication of the earlier CZAS reports, and pdf versions have been placed on the SCAPE website (www.scapetrust.org). SCAPE also produced databases from all of the survey reports and added data (including information about the coast edge) to a GIS, although much of the survey data from these earlier surveys is yet to be added to local or national Historic Environment Records. Analysis of the Results of Surveys and Desk-based Assessments In 2007, the author analyzed the data by the CZAS (Dawson 2007), including results from the Western Isles. The following discussion is based on the results of four surveys and two desk-based assessments conducted in North and South Uist between 2005 and 2007. A severe storm struck the Western Isles on 11–12 January 2005. It was associated with a deep depression (≈944 mb) that passed to the north of Scotland (Wolf 2007). Wind speeds reached hurricane force with tragic results, and the sea driven onto the land claimed the lives of five people from one South Uist family. The sea had a severe impact on the coastal zone, eroding the coast edge and damaging buildings and roads. Although some stretches of coast were relatively unchanged, there was great damage to other areas, and in extreme cases the coast edge retreated by up to 50 m (for example, parts of the Baile Sear coast, North Uist). In the immediate aftermath of the storm, numerous reports of newly exposed archaeological sites were received by Historic Scotland and the SCAPE Trust. In order to assess the effects of the storm, Historic Scotland sponsored two coastal surveys. Fieldwork was undertaken immediately after the storm, and the survey reports were completed by the end of March 2005 (Johnson et al. 2005, Moore and Wilson 2005). CFA Archaeology Ltd. worked on the west coast of North Uist from the North Causeway to the South Causeway (including the islands of Berneray, Baile Sear, and Bhalaigh)—a total length of ≈210 km. EASE Archaeology was commissioned to survey South Uist from Bagh na Creige Loisgte to Caolas Table 2. Length of coast and survey areas, ordered by percentage of coast surveyed. Length (km) Area Coast Surveyed % Western Isles 3032 1327 43.76 Orkney 1085 415 38.25 Shetland 2063 580 28.11 Mainland 6235 1683 27.00 Inner Hebridean Islands 3000 600 20.00 Total 15,415 4605 29.87 Journal of the North Atlantic T. Dawson 2015 Special Volume 9 91 Figure 6. Map showing location of all CZAS 1996–201 1. © RCAHMS. Reproduced by kind permission of ScARF. Journal of the North Atlantic 92 T. Dawson 2015 Special Volume 9 reported to be low, and rabbit burrows were not regarded as constituting a major threat (Moore and Wilson 2005). The percentage of sites carrying a recommendation for survey (the highest level of recommendation used in CZAS) was similar between the surveys, both at about 7% of the total number of sites recorded. The numbers of sites this represents is very different though, due to the larger number of sites located in North Uist (Table 5). The two surveys located a very large number of previously unrecorded sites, some of which had been revealed due to the storm. Of these, 368 sites carried a recommendation for further work, either survey or monitoring. Following the post-storm surveys of the west coast of the Uists, an examination of the east coast was undertaken in 2006. It was decided to evaluate the utility of conducting a DBA as a standalone project, as had been done in England, rather than as an integral part of the CZAS. The effectiveness of using the DBA for determining priority areas for subsequent fieldwork was tested, and the study also aimed to assess how many new sites could be identified from an analysis of aerial photographs, maps, and other sources in a remote area where few site records existed. There was also a widely held idea that there would be low site densities and a low vulnerability to erosion along much of the mountainous, steep, hard-rock-based, and sheltered east coast. A further aim of the DBA was to check these assumptions The Glasgow University Archaeological Research Division (GUARD) undertook the two DBAs (Sneddon 2006), examining a coastal strip that extended from approximately 1 km inland to 4 km offshore. The DBA report highlighted priority areas, dividing them into high, medium, and low priority. Low-priority areas, not recommended for field survey, were generally located where the coastline was deemed to be under no immediate threat from coastal processes and/or where there were very few or no previously recorded archaeological or builtheritage sites. Following completion of the DBA report, field surveys were conducted to test areas identified as high, medium, and low priority (Johnson et al. 2006, Eiriosgaih; Benbecula from Oban Uaine to Ob Saile, and the entire island of Grimsay—a total length of ≈199 km. The two surveys, although of similar lengths, obtained widely different results. Both located a large percentage of previously unrecorded sites (sites not recorded in the records of either the Western Isles Council or RCAHMS), with almost 70% of the North Uist sites and over 87% of the South Uist sites being new records (Table 3). However, the actual numbers of sites recorded are widely different, with the North Uist survey recording over four times as many sites (880 sites) as the South Uist survey (200 sites). One of the reasons for the difference is that there were a greater number of previously known sites in North Uist, partly due to the work of local antiquarian Erskine Beveridge (1911). However, discounting these sites still meant that a far greater number of sites were located in North Uist (609) compared to South Uist (175). It is possible that part of the reason for the difference may be due to some South Uist sites having been obscured by the deposition of sediment by the time the survey commenced (Figs. 4, 5). For example, two large South Uist sites (SU46 and SU19) were only revealed intermittently along the length of a coastal section, other parts of these sites were covered with sand or cobbles. There is a suspicion that many other sites may have been exposed immediately after the storm but were hidden again by the time of the survey. The dynamic nature of the South Uist coast after the storm is shown by the much larger number of the sites that were considered to be eroding: over 50% of all South Uist sites located (105 sites), as opposed to just 7% (58 sites) in North Uist (Table 4). In South Uist, wave action was the principal threat, and only minimal damage was attributed to animals, as stocking numbers were Table 4. Erosion class of sites recorded along the west coast of North and South Uist. North Uist west South Uist west Erosion class Number % Number % Eroding 58 6.59 105 52.50 Not eroding 822 93.41 92 46.00 Unknown 3 1.50 Table 3. Numbers of sites located along the west coast of North and South Uist. North Uist west South Uist west Sites Number % Number % Found by this survey 609 69.20 175 87.50 Previously known 271 30.80 25 12.50 Total 880 100.00 200 100.00 Table 5. Recommended action at sites along the west coast of North and South Uist. North Uist west South Uist west Recommendation Number % Number % Survey 54 6.14 14 7.00 Monitor 276 31.36 24 12.00 Nil, N/A, Unknown 550 62.50 162 81.00 Journal of the North Atlantic T. Dawson 2015 Special Volume 9 93 Moore and Wilson, 2007). EASE Archaeology surveyed parts of South Uist and Benbecula (171 km); CFA Archaeology examined parts of North Uist (149 km). The survey teams were asked to compare their findings with the DBA in regard to the coastal geology and geomorphology, the identification of archaeological sites, and the success in highlighting priority survey areas. Both survey teams found that the DBA was very successful at identifying geological and geomorphological features and at providing an assessment of vulnerability to erosion. The DBA was not as useful when it came to identifying archaeological sites (Table 6). Only 38 sites were noted in the DBA of Benbecula and South Uist, whereas the survey located a further 248 sites (87% of the total recorded during field survey). The North Uist DBA identified 44 sites (recorded as 70 individual sites on the ground); 660 additional sites were recorded as a result of the fieldwork (90% of the total recorded during field survey). It must be stressed that the reason for the vast difference in the numbers of sites identified through DBA and through field surveys was not due to any inadequacy on the part of the desk-based assessment. The North Uist survey report noted that there were very few examples where sites missing from the DBA would have been expected to have been visible on aerial photographs or maps. The numbers of new sites recorded led both survey teams to question the value of prioritizing survey areas on the basis of archaeological potential. The authors of the South Uist report noted that there was no appreciable difference between the archaeological potential of any of the areas surveyed, even though they had been awarded different levels of priority (Moore and Wilson 2007). Just as many sites were found in areas of low priority as in areas of high priority. The North Uist survey also noted that a large number of archaeological sites were located within an area identified as being of low priority, stating that “if this area had been excluded then some potentially important information regarding the nature and distribution of land-use and settlement in the post-medieval period would have been missed” (Johnson et al. 2006). The study showed that although a DBA could be used to highlight priority areas for survey, the criteria for prioritization should be based on the physical nature of the coast and its vulnerability to storm damage, not on the anticipated density of archaeological remains. The analysis of the DBA and survey results also demonstrated that pre-survey assumptions about site density and vulnerability along the remote and sheltered east coast were only partly correct. A comparison of Table 3 with Table 6 shows that although the length of the east coast surveyed was less than in the west (as only selected areas were surveyed), similar numbers of sites were located. The major difference between the results of the east and west coast surveys was the date of the sites, with over 90% of east coast sites being identified as post medieval or later and the rest being of unknown date. Assumptions about site vulnerability, based on the physical setting of sites, were found to be more correct. Very few east coast sites carried any recommendation for further work (Table 7), partly due to low risks from erosion. Over 96% of sites recorded during the North Uist survey carried no recommendation. In South Uist, no sites were recommended for survey of any sort, due to most of the coast being classified as stable and the majority of sites considered of limited archaeological potential. The study indicated that not all parts of the Western Isles are equally vulnerable to erosion. Sites on the sheltered east coast are not under as much threat as those along the west coast. Although this may seem obvious, the study has helped demonstrate that the selection of CZAS area to date, concentrating on the more exposed and softer coastlines, has been justified. It has also shown that stand-alone DBAs are able to identify areas most at risk and will allow future surveys to target the most vulnerable areas. Prioritizing Action The final steps in Barclay and Fojut’s (1995) progression were to prioritize action and, where possible, implement appropriate solutions. One of the aims of HS APP4 was to “encourage consistent fieldwork and reporting so that the results of future rapid surveys will be more comparable with each Table 6. Numbers of sites located during DBA and coastal surveys of the east coast of North and South Uist. North Uist east South Uist east Sites Number % Number % Found by this survey 660 90.41 248 86.71 Previously known (DBA) 70 9.59 38 13.20 Total 730 100.00 286 100.00 Table 7. Recommended action at sites along the east coast of North and South Uist. North Uist east South Uist east Recommendation Number % Number % Survey 2 0.27 0 0.00 Monitor 21 2.88 2 0.70 Nil, N/A, Unknown 707 96.85 284 99.30 Journal of the North Atlantic 94 T. Dawson 2015 Special Volume 9 other and national priorities can be decided more transparently” (Historic Scotland 1996:2). An attempt at prioritizing action based on the results of the coastal surveys has been made by this author (Dawson 2010). Prioritization was based on an analysis of the type of site recorded and the scale of the threat posed by erosion (see Dawson 2010 for detailed discussion). The threat to the sites was partly assessed on the geology, geomorphology, and erosion class of the coastline assigned during the original surveys and partly (where possible) on past rates of erosion. It can be very difficult to determine past and future rates of erosion, as its progress is not measured and orderly, with, say, one meter of land being lost every ten years. Local sedimentation and erosion patterns are complex and remain largely unpredictable. A stretch of coastline can be stable, or even accreting, over a number of years or decades until a combination of natural processes causes significant change. One useful way of judging past rates of erosion is to compare contemporary erosion surveys with previous ones. Highly accurate dGPS systems and terrestrial laser scanners allow large areas of the coast to be surveyed rapidly (e.g., Dawson, 2011a). Although it will be easier to determine erosion rates in the future, there are few detailed historical surveys to compare with modern data. Semi-permanent grid pegs have been left at some archaeological sites to allow repeat visits, for example, at Galson, Lewis, where a 1-m strip was observed to have eroded between 1997 and 2000 (Neighbour et al. 2000). Desk-based surveys can give an indication of the rate of coastal change over time, but it is known that nineteenth-century surveyors used the coast in order to rectify mapping errors (Hansom et al. 2011). The result is that coastlines were not always depicted accurately. A comparison of aerial photographs can be more accurate, but it is usually necessary to georectify the images first within a GIS to eliminate tilt distortion. A comparison of georectified images can highlight the unpredictability of erosion. As an example, a site located within the sand cliffs at A’ Cheardach Ruadh on Baile Sear, North Uist, was selected for evaluation in the 1980s after being identified as actively eroding (Barber 2003:p1). A comparison of georectified aerial photographs from 1946 and 1992 show that there was actually very little erosion in the 46 years between the photographs, with a maximum of 5 m of the site being lost. This is not to say that the evaluation was unjustified. The site’s location at the coast edge, buried in sand on an exposed, west-facing beach meant that the site was extremely vulnerable, and it would have been a high priority for full-scale excavation if it, together with up to 50 m of the coast, had not been completely lost in the storm of January 2005. The prioritization project ranked sites into one of five priority classes, with the highest-ranked sites (priority 1) requiring urgent attention. The results of the prioritization project are currently being reviewed by Historic Scotland, but the Western Isles was home to the largest number of prioritized sites, with over 80 sites being classed as either priority 1 or priority 2. Within the prioritization report, a staged set of actions was suggested for all ranked sites, with one action following on from the preceding one if necessary. In many cases, the initial action was to revisit the site to see if it still survived, as many of the sites were identified as eroding over fifteen years ago. A project is currently being planned that will ask members of local groups, such as Shorewatch group members, to assist in updating survey information and therefore to refine the priorities list. Managed by SCAPE, the Scottish Coastal Heritage at Risk Project (www.scharp.co.uk) will use mobile phone apps and an interactive, map-based website to encourage local communities to update descriptions of known sites. Participants will be asked to provide photographs and complete a simple, multiple-choice recording form to allow management information to be gathered. This approach will allow existing records to be updated and sites that have been destroyed to be removed from the prioritized list. Implement Appropriate Solutions Historic Scotland has implemented solutions at a number of sites by sponsoring surveys and excavations, and details of these can be found in the relevant site reports. In the late 1990s, a series of focal studies based on the results of coastal surveys were initiated (for example, see section III, “Coastal Archaeology and Erosion in Scotland”; Dawson 2005). Historic Scotland has sponsored several excavations in the Western Isles, including excavations prompted by aeolian erosion (for example, Dunwell et al. 1995, Sharples 2000) and coastal erosion (Armit 2006, Parker Pearson and Sharples 1999). In recent years, there have been an increasing number of community-focused excavations, enabling members of local groups to work with professional archaeologists on sites that would otherwise be lost to erosion. SCAPE has managed several of these excavations around Scotland, including one in the Western Isles. The North Uist heritage group, Journal of the North Atlantic T. Dawson 2015 Special Volume 9 95 their archaeological competence and confidence. Working to high professional standards, the group has also helped with post-excavation work, including wet-sieving and the sorting of samples. The site was open to all, either as participant or visitor. Many local people came to the site regularly to watch progress and learn about new discoveries (Fig. 7). Visits were arranged for local schoolchildren and for art students from Taigh Chearsabhagh, Lochmaddy, who used the site to help inspire their work (Fig. 8). The excavation also featured regularly in local and national press articles, further helping to increase awareness about the site locally and nationally, and helping the project to win the public vote as the Current Archaeology Rescue Dig of the Year in 2012. Conclusion Managing coastal heritage threatened by erosion is challenging in Scotland due to the length of the coast and the incredible preservation of many archaeological sites. Management strategies have been developed and revised over the years, and the lessons learned are applicable to other coastlines around the world. Desk-based assessments have been excellent at highlighting the most vulnerable areas for survey, but have been less useful at locating Access Archaeology, reported masonry exposed on the beach at Sloc Sàbhaidh, Baile Sear after the 2005 storm. The site lay 1 km south of A’ Cheardach Ruadh (above), and SCAPE worked with the group to survey and characterize the remains, revealing the eroding remains of Iron Age structures, including a wheelhouse. Erosion was rapidly destroying the site, and part of the buried wall of one structure was reduced in height by 1 m in a single year due to a change in the profile of the beach. The destruction of the site provided an opportunity for the local group and archaeologists to work together on a project that could completely excavate the site, taking down walls and looking under floor deposits. This “total excavation” allowed a large number of dating samples to be taken from above, within, and below the structure. It is unusual for such opportunities to exist at similar sites, where wellpreserved prehistoric masonry structures are often legally protected, and total excavation is prohibited. The excavation has run over several seasons as a Shorewatch project, and the group has worked with SCAPE and other archaeologists to excavate the structures (Dawson 2011b, MacDonald and McHardy 2008, Rennell and McHardy 2009, Stentoft et al. 2007). Many group members have attended regularly over the years, increasing both Figure 7. Local visitors at the excavations on the beach at Slo c Sàbhaidh, Baile Sear, North Uist. Journal of the North Atlantic 96 T. Dawson 2015 Special Volume 9 in some cases when assessing where to allocate funds. Community survey and excavation projects often receive good publicity, helping to further increase the profile of the threat to archaeological sites from erosion, both locally and nationally. Erosion is not predictable, and a single storm can completely destroy an archaeological site. Thus, time is not on the side of the heritage manager. Although erosion is a problem, it also presents an opportunity for professionals and community groups to work together. Experimental research methods can be devised and specific questions asked of sites that would otherwise be lost. There are a huge number of sites at risk, and it is up to archaeologists to take up this challenge. In Scotland, the coastal zone contains sites of all periods and all types, many of which are also found inland. Archaeologists involved in research projects should consider designing their strategies so that it is the vulnerable coastal sites that are examined. They should also involve local communities in the project as much as possible. In Scotland, information from the coastal surveys and the prioritized list of sites is starting to be used to inform excavation sites or assessing areas of high archaeological potential. Targeted Coastal Zone Assessment Surveys have located thousands of previously unrecorded sites, and the data collected has helped to produce lists of priorities for sites that need urgent work. Producing lists of priorities is not an end in itself; without further action, the sites will eventually be destroyed. In order to prioritize action, survey data needs to be analyzed in a systematic and repeatable way. This analysis has been completed in Scotland, and large numbers of sites have been classified as high priority. However, the dynamic nature of coastal erosion means that the prioritized list is not static, and erosion will destroy some sites while exposing others. In order to gain up-to-date information, heritage managers in Scotland have recognized the very important role that the general public can play in updating records and reporting new discoveries. Involving local groups in heritage management also increases awareness of the problem and the local importance of sites. The public has also played a key role in helping to determine which sites should be worked at, and the concept of public value has been adopted Figure 8. Art students from Taigh Chearsabhagh depicting archaeologists at the Sloc S àbhaidh wheelhouse on Baile Sear. Journal of the North Atlantic T. Dawson 2015 Special Volume 9 97 Dawson, T. 2006. Archaeology and coastal erosion in Scotland: The current state of knowledge and future directions. Internal report for Historic Scotland, Edinburgh, Scotland, UK. Dawson, T. 2007. A review of the Coastal Zone Assessment Surveys of Scotland, 1996–2007: Methods and collected data. Internal report for Historic Scotland, Edinburgh, Scotland, UK. Dawson, T. 2010. A system for prioritising action at archaeological sites recorded in the Coastal Zone Assessment Surveys. Internal report for Historic Scotland, Edinburgh, Scotland, UK. Dawson, T. 2011a. Fethaland Fishing Station: Shetland desk-based assessment and preservation by digital record. Unpublished report of the SCAPE Trust and the University of St Andrews, St Andrews, UK. Dawson, T. 2011b. Baile Sear Community Excavation: Sloc Sàbhaidh, Baile Sear, North Uist. Season 4 (2010) Data Structure Report. Unpublished report of the SCAPE Trust and the University of St Andrews, St Andrews, UK. Dunwell, A.J., T.G. Cowie, M.F.T. Bruce, T. Neighbour, and A.R. Rees. 1995. A Viking Age cemetery at Cnip, Uig, Isle of Lewis, Proceedings of the Society of Antiquaries of Scotland 125:719–752. Department for Culture, Media, and Sport (DCMS). 2011. The United Kingdom’s world heritage: Review of the tentative list of the United Kingdom of Great Britain and Northern Ireland. London, UK. Hansom, J.D., A.F. Rennie, A. Dunlop, and J. Drummond. 2011. A methodology to assess the causes and rates of change to Scotland’s beaches and sand dunes Phase 1. Scottish Natural Heritage Commissioned Report No. 364. Inverness, Scotland, UK. Historic Scotland. 1996. Coastal Zone Assessment Survey: Historic Scotland Archaeological Procedure Paper 4. Historic Scotland, Edinburgh, Scotland, UK. HR Wallingford. 1997 Coastal cells in Scotland. Scottish Natural Heritage Research, Survey, and Monitoring Report No. 56. Edinburgh, Scotland, UK. Johnson, M., M. Cressey, S. Badger, and P. Richardson. 2005. Coastal Zone Assessment Survey: North Uist, Report 1051. Unpublished report of CFA Archaeology, Edinburgh, Scotland, UK. Johnson, M., M. Cressey, and P. Richardson. 2006. Coastal Zone Assessment Survey: North Uist, Report 1262. Unpublished report of CFA Archaeology, Edinburgh, Scotland, UK. Lees, G. 2005. Contemporary coastal processes in Scotland. Pp. 13–21, In T. Dawson (Ed.). Coastal Archaeology and Erosion in Scotland. Historic Scotland, Edinburgh, Scotland, UK. Lowe, C. 1998. St. Boniface Church, Orkney: Coastal Erosion and Archaeological Assessment. Sutton Publishing, Stroud, UK. Mercer, R.J. 1981. Archaeological field survey in northern Scotland. Vol. II. University of Edinburgh Department of Archaeology Occasional Paper No 7, Edinburgh, Scotland, UK. strategies; hopefully, this model can be applied to many other coastlines around the globe. Literature Cited Armit, I. 2006. Anatomy of an Iron Age Roundhouse: The Cnip Wheelhouse Excavations, Lewis. Society of Antiquaries of Scotland, Edinburgh, Scotland, UK. 272 pp. Ashmore, P.J. 1994. Archaeology and the coastal zone: towards a Historic Scotland policy. Historic Scotland, Edinburgh, Scotland, UK. 35 pp. Barber, J. 2003. Bronze Age farms and Iron Age farm mounds of the Outer Hebrides SAIR 3.Available online at http://www.sair.org.uk/sair3/. Accessed 20 December 2012. Barclay, G.J. 1997. State-funded rescue archaeology in Scotland: Past, present, and future. Historic Scotland Occasional Paper No. 2. Historic Scotland, Edinburgh, Scotland, UK.38 pp. Barclay, G.J., and N. Fojut. 1995. The management and conservation of the built and maritime heritage in the coastal zone. Historic Scotland, Edinburgh, Scotland, UK. Batey, C.E. 1984 Caithness coastal survey 1980–1982: Dunnet Head to Ousdale. Durham University Department of Archaeology Occasional Paper No. 3, Durham, UK. Beveridge, E. 2001 [1911]. North Uist: Its Archæology and Topography with Notes upon the Early History of the Outer Hebrides. Birlinn, Edinburgh, Scotland, UK. 348 pp. Branigan, K. 2005. Assessment survey: Isle of Barra. Pp 63–72, In T. Dawson (Ed.). Coastal Archaeology and Erosion in Scotland. Historic Scotland, Edinburgh, Scotland, UK. Branigan, K., and P. Foster (Eds.). 1995. Barra: Archaeological Research on Ben Tangaval. Continuum, Sheffield, UK. 231 pp. Branigan, K., and J. Grattan. 1998. Coastal assessment survey: Barra and Vatersay. Unpublished Report for Historic Scotland, SEARCH, University of Sheffield, Sheffield, UK. Burgess, C., and M. Church. 1996. Coastal erosion assessment, Lewis. Unpublished Report for Historic Scotland, University of Edinburgh, Edinburgh, Scotland, UK. Central Intelligence Agency (CIA). 2011. The world factbook. Field listing: coastline. Available online at https://www.cia.gov/library/publications/the-worldfactbook/ fields/2060.html. Accessed 20 April 2011. Davidson, A., and N. Jones. 2002. The archaeological survey of the Welsh coast. Pp. 19–24, In A. Davidson (Ed.). The Coastal Archaeology of Wales. CBA Research Report 131, York, UK. Dawson, T. (Ed.). 2005 (revised). Coastal Archaeology and Erosion in Scotland. Historic Scotland, Edinburgh, Scotland, UK. 222 pp. Journal of the North Atlantic 98 T. Dawson 2015 Special Volume 9 Macaulay Land Use Research Institute (MLURI). 1993. The land cover of Scotland 1988. Aberdeen, Scotland, UK. MacDonald, K., and I. McHardy. 2008. Baile Sear community archaeology project. Sloc Sàbhaidh, Baile Sear, North Uist. Season 2 (2007) Data structure report. Unpublished report of the SCAPE Trust and the University of St Andrews, St Andrews, Scotland, UK. Moore, H., and G. Wilson. 2005. Western Isles (South) Coastal Zone Assessment Survey: Grimsay, Benbecula, and South Uist. Unpublished report of EASE Archaeology, Edinburgh, Scotland, UK. Moore, H., and G. Wilson. 2007. Report on Western Isles (South) Coastal Zone Assessment Survey: South Uist, and Benbecula (east coasts). [Unpublished report of EASE Archaeology, Edinburgh, Scotland, UK. Neighbour, T., C. Knott, M.F. Bruce, and N.W. Kerr. 2000. Excavation of two burials at Galson, Isle of Lewis, 1993 and 1996. Proceedings of the Society of Antiquaries of Scotland 130:559–584. Parker Pearson, M., and N.M. Sharples. 1999. Between Land and Sea: Excavations at Dun Vulan, South Uist. Sheffield University Press, Sheffield, UK. Parker Pearson, M., P. Marshall, J. Mulville, and H. Smith 2011. The prehistoric village at Cladh Hallan. Available online at http://www.shef.ac.uk/archaeology/ research/cladh-hallan. Accessed 23 April 2011. Ramsay, D.L., and A.H. Brampton. 2000. Coastal cells in Scotland: Cells 8 and 9. The Western Isles Scottish Natural Heritage Research, Survey, and Monitoring Report No 150. Edinburgh, Scotland, UK. Rennell, R., and I. McHardy. 2009. Baile Sear Community Archaeology Project. Sloc Sàbhaidh, Baile Sear, North Uist. Season 3 (2008) Data structure report. Unpublished report of the SCAPE Trust and the University of St Andrews, St Andrews, Scotland, UK. Sharples, N.M. 2000. The Iron Age and Norse settlement at Bornish, South Uist: An interim report on the 2000 excavations. Cardiff Studies in Archaeology Specialist Report 18. Cardiff, UK. Shepherd, I.A.G., and A.N. Shepherd. 1978. Machair sites survey: Uists and Benbecula. Internal HBM/SDD report. HBM, Edinurgh, Scotland, UK. Simpson, D.D.A., R.A. Gregory, and E.M. Murphy. 2003. Excavations at Manish Strand, Ensay, Western Isles. Proceedings of the Society of Antiquaries of Scotland 133:173–189. Smith, G. 1995. Aberdaron Bay to the Great Orme. Unpublished Report. GAT Report No. 79. Unpublished report of GAT, Bangor, Wales, UK. Sneddon, D. 2006. Coastal Zone Assessment Survey. East Coast of North Uist, Benbecula, and South Uist. Deskbased assessment. GUARD Project 2198. Unpublished report of GUARD, Glasglow, Scotland, UK. Stentoft, K., T. Dawson, and L. McKenna. 2007. Baile Sear Community Archaeology Project. Sloc Sàbhaidh, Baile Sear, North Uist. Unpublished report: The SCAPE Trust and the University of St Andrews, St Andrews, Scotland, UK. Wolf, J. 2007. Modelling of waves and set-up for the storm of 11–12 January 2005 Proudman Oceanographic Laboratory Internal Document. Liverpool, UK.