Geophysical Survey in the Archaeology of Scotland: Recent Developments and Results

Abstract

This paper reviews the current state of geophysics in Scottish archaeology, considering the scope of the surveys, the range of targets investigated and techniques deployed, as well as the practitioners and commissioners of surveys. Several issues of methodology and interpretation are illustrated through case studies taken from mainland Scotland, Orkney and the Isle of Lewis. One of these focuses on the relative frequency of poor magnetic and earth resistance responses recorded over ditch and pit features due to drift geology and soil conditions, and the efforts to explain those responses in terms of soil properties. This leads to the recommendation that archaeo-geophysics can only benefit from aligning itself on a regular basis with geoarchaeology since their respective subject areas often converge more than is usually recognised. Another recommendation is the need for fuller dissemination of the graphical output of surveys as well as access to raw data to encourage a more critical view of how interpretations of individual geophysical anomalies are made.

In Memory of Oliver O’Grady.

1 Introduction

Scotland’s heritage offers considerable scope for the application of geophysical survey: to set the scene, there is the chronological range of its sites and monuments that have been explored in this manner, from potential Mesolithic seascapes linking the Minch (Fig. 2) bordering the Isle of Harris and the Atlantic (Bicket et al., 2017; MBS—see Table 1 for abbreviations), Mesolithic shell middens on Colonsay (Finlay et al., 2019; M, R), the Heart of Neolithic Orkney (Brend et al., 2020; Table 1), Roman military presence (Hanson et al., 2019; Table 1), potential royal tombs in Dunfermline Abbey (Penman & Utsi, 2016; GPR), nineteenth century lime kilns (Bishop et al., 2017; G, MS) to WW I warships scuttled in Orkney’s Scapa Flow (Dean, 2006; MBS). This range also reflects the diversity of natural environments (lowland, upland, wetland, aeolian) that have hosted human activity, as well as the variable preservation of that activity over the course of time. Judged in this light, it may be fair to say that geophysics is now well established in most sectors of Scottish field archaeology and as such has overcome the perception, prevalent until the start of the present millennium, that its performance was uneven (Jones & Sharpe, 2006; Cuenca-Garcia et al., 2020, 6–7). This brief contribution attempts to give an overview of the current status of geophysical survey in Scotland, considering some of the routes by which geophysical surveys have been carried out and on what scale, and using case studies to highlight specific issues of methodology and interpretation.

Table 1 Case studies: sites, targets, and techniques (M- magnetometer survey, R- earth resistance, GPR- ground penetrating radar, MS-magnetic susceptibility, ALS- Airborne laser scanning or LiDAR, EMI- electromagnetic induction, MBS-multibeam sonar, P- phosphate)

An impression of the frequency with which geophysics was deployed up until the early 2000s can be gleaned from the Scottish Archaeological Geophysical Survey Database Project undertaken by the Department of Archaeology, University of Glasgow, and funded by Historic Scotland. From some 600 entries in the database Rennie (2006) reported that research-led surveys including those forming part of a broader project, such as the Traprain Law Environs Project (Hale et al., 2006; Hale & Cowley, 2009), were significantly more numerous than those relating to development proposals; prehistoric (i.e. pre-Roman) sites received the most attention. This picture seems to have altered in the intervening period to judge from the entries in Discovery & Excavation in Scotland, published annually by Archaeology Scotland. For the years 2016–2018, geophysical survey features 41, 50 and 47 times respectively as a component of fieldwork undertaken in the commercial sector, by universities and other institutions and in initiatives made by local archaeology societies and community archaeology groups. The size of the survey varies considerably, from the landscape level (up to 520 ha on the Aberdeen bypass project (see below) to a few thousand square metres in advance of targeted excavation. These numbers for 2016–2018 may be significant in themselves but should be considered in context: some 800 entries of fieldwork (including watching briefs, but excluding metal detector finds) feature, for example, in 2018. Figure 1 presents the proportions of different types of target investigated during those 3 years.

Fig. 1

Indicative relative proportions of geophysical survey targets reported in Discovery & Excavation in Scotland in 2016, 2017 and 2018

In the commercial sector, developer-funded archaeology has traditionally made some use of geophysical survey (Leslie & Banks, 2006), but recent years have seen a more critical approach to its role in guiding excavation strategy. In essence, the planning authorities in Scotland, advised by their heritage/archaeology advisor, consider all the options that non-invasive assessment can offer. The size of the development area and its location (urban, rural) are two of the many factors impacting on the suitability of deploying geophysics. For instance, the combination of large size and rural location encouraged reliance on gradiometer (hereafter magnetometer) survey in recent major road projects (for example, A9 Killiecrankie to Glen Garry section (8 ha) Headland Archaeology, 2018a; Aberdeen bypass (520 ha) https://molaheadland.com/project/aberdeen-western-peripheral-route/, Headland Archaeology, 2018b; A96 Nairn bypass (30 ha) AOC Archaeology Group, 2016). It is worth noting here the few cases of surveys, undertaken separately as developer- and research-led, that have combined to good effect, for instance at the Pictish centre at Kinneddar (Noble et al., 2019) (Fig. 2). As for the excavation strategy, the evaluation area of a proposed development site in Scotland has, since the 1990s, typically been 5–10%, compared to 2–3% elsewhere (Hey & Lacey, 2001, 43, Fig. 27). Of relevance in Hey and Lacey’s account, based on data from sites in south-east England, is the valuable assessment by Linford and David (2001) of the effectiveness of the geophysical surveys carried out at five of those (prehistoric to Roman) sites in the light of their excavation.

Fig. 2

Map of Scotland showing the locations of some of the sites mentioned in the text

Undertaking the surveys are specialised Scottish or UK-based geophysics companies, commercial archaeology units, university departments and local archaeological societies. Prominent among the many organisations commissioning such work is Historic Environment Scotland (HES), which has a diverse estate of properties in care (Sagrott, 2021); here geophysics is used as a management tool to build up a picture of the immediate environs of those properties. Indeed, they are well represented in the ecclesiastical, castle and house/palace categories shown in Fig. 1.

The planning of geophysical surveys and interpretation of their results in Scotland continues to make much use of the rich aerial photographic record, and this can now be supplemented by the availability of high-resolution ALS (Airborne Laser Scanning or LiDAR) data covering many parts of Scotland. HES, the leading body in Scotland promoting aerial survey, has recently extended its practical capability to include geophysical survey (multi-sensor magnetometer and electromagnetic sensor). This welcome development has seen the appointment of a full-time geophysics officer (Hannon, 2021).

In terms of instrumentation, one major development has been the deployment of multi-sensor gradiometers with integrated GPS capability placed on a vehicle-towed (or operator-pushed) cart, well suited for high-resolution landscape level surveys. Recent locations include Rousay (Orkney) (Beusing & Rassmann, 2018), Roman Newstead (Beusing et al., 2018a) and Roman-Iron Age Birrens/Burnswark (Beusing et al., 2018b) and the Antonine Wall (Hanson et al., forthcoming). But more generally geophysical survey has relied on the traditional trio of techniques: gradiometry, electrical resistance and ground-penetrating radar (GPR) in that order; this point is taken up in the Discussion. Methods such as electrical resistance imaging have found limited application (but see Sutherland et al., 1998 and Neighbour et al., 2001 for surveys on Shetland and in south-west Scotland respectively, and more recently Brend et al., 2020, Fig 5.29).

2 Case Studies

The case studies presented below have been selected to illustrate recent surveys. Confronting all of them is the fundamental issue of interpretating individual geophysical responses. As is demonstrated below, this has followed the traditional pragmatic approach, yielding archaeologically meaningful results without, except at Forteviot (see Fig. 2), recourse to validating those results by excavation. By contrast, there are surveys, undertaken as part of research-led and commercial projects, that have been followed by excavation; Sects. 16.2.3 and 16.3 briefly treat two such examples. However, they are in a minority and furthermore may not include a full post-excavation re-evaluation of the geophysics. In this situation in which excavation has not taken place, an awareness of the circumstances that may complicate interpretation is necessary. Notable among these is the nature of the drift geology and topsoil/subsoil conditions that may give rise to noisy data, partially masking the detection of remains of archaeological interest. It is the combination of the many natural factors likely to determine the nature of a geophysical response together with the realisation that those factors may be unique to a given location and its associated archaeological feature that has tended to undermine attempts to find helpful explanations. And to these natural factors should now be added the effects on magnetometer data of ‘green waste’ (biodegradable and organic materials) which form part of commonly used fertilisers and soil conditioners (Gerrard et al., 2015). In principle then, one way forward would be to build up a fuller characterisation of the soil that takes account of the soil’s textural and chemical attributes. The application of chemical methods to soil analysis has already played valuable roles in geoarchaeology, particularly in defining activity areas within either ancient to recent settlements or individual buildings and in determining anthropogenic soil development (see overview https://scarf.scot/thematic/scarf-science-panel-report/4-people-and-the-environment/4-2-geoarchaeology/), but it has not interacted sufficiently with the corresponding geophysical data, as discussed in 3. below. An exception, alone of its kind in Scotland, is the study by Cuenca-Garcia (2012, 2018) who monitored the combined geophysical and geochemical responses to some archaeological features present in contrasting burial environments. Part of her enquiry, which finds parallels with soil chemical responses at sites in Scotland defined by cropmarks observed in aerial photographs (Sharpe, 2004), is discussed below.

2.1 The Heart of Neolithic Orkney

In terms of scale and output, pride of place must go to the 10-year landscape study of the UNESCO World Heritage Site, The Heart of Neolithic Orkney, comprising the most well-known Neolithic settlement on Orkney, Skara Brae, and, 5 km away, a remarkable array of Late Neolithic and later monuments of the Stenness-Brodgar area on Mainland Orkney (Brend et al., 2020). Situated in that area, much of it lying on a narrow isthmus between Lochs Harray and Stenness, are upstanding stone circles and henges, the chambered tomb of Maeshowe, the settlement of Barnhouse and the Neolithic complex of the Ness of Brodgar. But the knowledge that this area is known from surface and other indications to be rich in a variety of other prehistoric sites presented a unique challenge for geophysical prospection. Covering an area of some 285 ha, the survey, conducted primarily with the magnetometer (Table 1), indeed delivered on the title of its publication: Landscape Revealed; not only are its results of major archaeological importance, but both the manner they are presented in, and the methodology employed set new standards. In brief, the combination of terrestrial survey with, on the one hand, the aerial photographic and ALS records and, on the other, historical documentation and antiquarians’ observations on past land use proved to be a powerful interpretative tool. Second, the project was able to draw on the results of marine survey on- and offshore of the two lochs just mentioned as well as sampling for microfossil and sediment analysis (including C¹⁴ dating) (Bates et al., 2016) to provide palaeo-geographic reconstruction; this approach is explored further in the Discussion. The corresponding work inland from Skara Brae in the Bay of Skaill employed conductivity survey and coring.

A selected sample of the results, from the Bay of Skaill area (Fig. 3), highlights the magnetic anomalies arising from the landscape viewed as a palimpsest: (a) prominent near-surface igneous dykes, (b) agricultural activity mainly of the rig and furrow type and (c) prehistoric and later occupation. Magnetically quiet areas in Fig. 3 are seen as former land surfaces now overlain by wind-blown sand. Of archaeological significance are first the detection of a continuation of the Skara Brae settlement and, further south, possible domestic structures; second, the inset in Fig. 3 shows the detail within the Iron Age broch of Loupandessness and, to the west, contemporary roundhouses/double houses. The post-medieval rig and furrow is prominent in this area of Iron Age occupation, yet hardly features towards the N and NE.

Fig. 3

Results of the magnetometer survey (top) and interpretation plan (bottom) of the World Heritage Area from Skara Brae to Loch of Skaill, Orkney; results at the roundhouses/‘double houses’ and Loupandessness (inset right). From Brend et al., 2020, Figs. 3.4, 3.5 and 3.25 respectively. Graphics courtesy of Nick Card

The results from the larger Stenness to Brodgar area were no less rich or informative. The marine survey concluded that by the Early Neolithic the isthmus was wider, Loch Stenness was smaller and sandstone outcrops suitable for monumental building were exposed at that time. This raised the possibility of the existence of early occupation phases on the isthmus as a whole. The agriculturally worked soils—rig and furrow—so prominent throughout the survey, deserve attention here because their magnetic signature was far from uniform. The nature of the underlying soils and deposits being brought into cultivation and the nature of the additions being made to the soil were both factors regulating the detected magnetic enhancement. As Brend et al. (2020, 102–4, Fig. 4.26–27) explain, where rig and furrow survive and weathering has been minimal, the ridge is a positive magnetic anomaly, correlating with higher earth resistance, but the reverse happens when the rig and furrow have become denuded over time. Furthermore, the authors report instances of ploughed rig and furrow passing over enhanced material belonging to earlier occupation; this manifested itself in terms of narrow positive responses due to the ploughed-out furrows ‘swapping’ to become pronounced negative responses as they pass through strong (magnetic) anomalies.

2.2 Forteviot

The corpus of geophysical, mainly magnetic responses from ditches that are part of enclosures, henges, cursus monuments and ring structures in Scotland of prehistoric date is large, as several contributions to the volume Going over Old Ground (Jones & Sharpe, 2006), more recent work in Orkney (see above) and in East Lothian (Hale & Cowley, 2009) make plain. The experience gained from those same studies also indicates that the responses from magnetometer survey over prehistoric ditches usually appear as positive anomalies, resulting from magnetic enhancement of the deposits that filled them. Efforts have been made to characterise the distribution of magnetic susceptibility across a ditch before and after excavation, notably by Kainz (2016) in Austria and now in Scotland by Cuenca-Garcia (2012, 2018).

One of Cuenca-Garcia’s programmes of survey, complementing high-resolution geophysical measurements (Table 1) with very detailed soil analysis, was located at Forteviot where some of the sites observed as cropmarks in aerial surveys were clearly identified as a complex of Neolithic-EBA enclosures and henges. Several of those prehistoric features were the targets of excavation forming a major component of Glasgow University’s long-term field project, Strathearn Environs and Royal Forteviot (SERF) (Brophy & Noble, 2020). Their presence had been confirmed by geophysical survey, which was itself undertaken at different scales, extending to one covering a 51 ha area (Poller, 2020, 44, Fig 2.27).

Working on a ring ditch, which appeared as a negative magnetic anomaly (Fig. 4 top), Cuenca-Garcia (2018) demonstrated the effect of removing the topsoil prior to excavation had on improving the quality of its resolution. The constituent ditches, which were detected more successfully by GPR, displayed lower MS values than the subsoil and topsoil and a depletion of Fe, Mn and anthropogenic trace elements; Fig. 4 (middle and bottom) shows their lateral and vertical distributions. Introducing the corresponding earth resistance measurements, higher water retention, correlating with higher organic content, in the outer ditch and the presence of a central cist burial explained their detection as resistance anomalies that were lower than the sandy surrounding soils. Thus, a picture is emerging of subtle mineralogical changes within the ditch resulting from redox reactions involving Fe and Mn oxides.

Fig. 4

Geophysical and geochemical responses at a ring-ditch enclosure at Forteviot. (a) results of the magnetometer survey before topsoil stripping (+/−10nT black/white). (b) Topsoil sampling over the enclosure before topsoil stripping (dotted yellow arrow). (c) Selected results of  (continued) the soil analyses. Total phosphate (P in mg P/kg soil), magnetic susceptibility (MS in ×10⁻⁶m³kg⁻¹), and multi-element concentrations using portable X-ray fluorescence (pXRF). The coloured bars mark the location of the outer (in green) and inner (in light orange) ditches and the cist burial (in blue). (d) results of the magnetometer survey after topsoil stripping (+/−10nT black/white). The yellow bars indicate the baulks left after the soil stripping. (e) Validation of the results of the geophysical surveys and location of the trench where the outer ditch was further explored (green double arrow). (f) Soil sampling over the enclosure after topsoil stripping (yellow arrow). (g) Selected results of the soil analyses. Total phosphate (P in mg P/kg soil), magnetic susceptibility (MS in ×10⁻⁶m³kg⁻¹), and multi-element concentrations using pXRF. The coloured bars mark the location of features as in the cist burial (in blue). (h) Exposed north-facing section (green double arrow in e) and sketch of the three backfill deposits identified in the outer ditch. (i) Selected results of the soil analyses of the samples collected from the exposed north-facing section of the outer ditch (h). (Adapted from Cuenca-Garcia 2018, Figs. 10–12)

2.3 Dalswinton Roman Fort

The results of the magnetometer survey at the Roman military complex at Dalswinton were also affected by noisy background data as well as the effects of recent agricultural activity. Interpretation by Hanson et al. (2019) proved to be an object lesson in a combination of viewing its results in the light of highly informative aerial photographs (and to a lesser extent ALS) and close cooperation between surveyors and experienced Roman specialist. That the fort had undergone a major re-organisation of its interior layout in response to changing military conditions only emphasised the need for such an approach.

While the aerial photographs (such as Fig. 5 right) provided good definition of the defences outlining the fort and its annexes, the magnetometer data revealed more detail of the fort’s interior (Fig. 5 left): the building blocks in the central sector, including two probable granaries and two courtyards (one of them the praetorium), the main gates, roads, and distinct areas of strong positive responses indicative of burnt debris reflecting the deliberate demolition at the end of the fort’s occupation. Of the two halves of the fort, the southern sector yielded better quality results since it had been less ploughed and more in pasture. Collectively, the results speak of a fort, which was of major strategic importance in south-west Scotland during the Flavian period (first century AD), accommodating probably a mixed garrison of legionaries and auxiliary cavalry. The fort’s expansion in the second phase need not have involved turning its orientation by 90°, as was originally suggested (Richmond & St Joseph, 1956, 13), but several adjustments were made to the interior. The probable legionary barracks occupying the new northerly extension (Fig. 5) may have been separated from the remainder of the fort. The via principalis (Fig. 5) continued in use, its route being altered only at the two ends to allow for the slightly changed position of the gates. The annexes, which were also expanded in the second phase, included space for animals, minor industrial activities such as metalworking but not civilian housing.

Fig. 5

Magnetometer survey (left) and aerial photograph (right) of the Roman fort and annexes at Dalswinton. 1 northern extension, 2 and 6 ovens or pits, 3 ‘parrot’s beak’ configuration of ditch end, VP (with arrows) via principalis, 4 probably praetorium courtyard building, 5 phase 2 road running parallel to the VP, 7 possible remains of bath house, 8 possible furnace, A annexe. Greyscale plotted at -10nT (white) to 10nT (black). Images: (left) Hanson et al., 2019, Fig. 8; (right) SC 165876, Crown Copyright: HES

The magnetometer and earth resistance survey at the nearby Antonine-period Roman fort and annexe at Drumlanrig (Fig. 2) (Walker et al., 2005) should be mentioned here because the opportunity was taken to investigate some of the magnetic anomalies by excavation (Wessex Archaeology, 2005). Besides the 22 m long trench (T1) across the defences whose excavation confirmed the presence of the broad rampart and the V-shaped outer ditch (Wessex Archaeology, 2005, Fig. 1), six of the other eleven trenches targeted individual magnetic anomalies in the area of the principia. Here the experience was mixed, ranging from a large rectangular negative feature that neatly related to a probable cistern or trough to instances of incomplete interpretation due in large part to the complexity and limited spatial extents of the contexts encountered.

2.4 Bothwell Castle

This castle, one of Scotland’s foremost medieval monuments, was built in the thirteenth century; it saw active involvement in the wars of Scottish Independence and as a result underwent much rebuilding only to be abandoned by the eighteenth century. As a monument in its care, HES commissioned a survey to assess some of the changes to the castle’s interior and immediate environs, as well as much of the surrounding park (Rose Geophysics Consultants, 2015). A wide array of anomalies was detected (Fig. 6: A, B), those in the park being associated less with archaeology, more with the remains of such features as palaeochannels and recent interventions comprising roads, paths, drains and fencing. Particularly around the castle’s north face, earth resistance identified some distinct structural, presumably foundation features. GPR time slices taken to a depth of 2 m were revealing; in the most shallow (Fig. 6: C) the strong responses are the result of the change from grass to paving and paths. But at 0.75–1 m (D) the time slice within the courtyard is able to show the distinctly linear high amplitude features which likely relate to drainage rather than to the amorphous shallow resistance feature. The clear message here may be well known but bears repeating: for a complex set of targets, deep and surficial, posed at a site like this castle the advantage of using complementary techniques is evident.

Fig. 6

Bothwell Castle. (a) Magnetometer survey data plotted at −6 nT (white) to 8 nT (black), (b) Earth resistance data plotted at −1SD (white) to 1SD (black), (c) 0.00–0.25 m GPR depth slice low amplitude (white) high amplitude (black), (d) 0.75–1.00 m GPR depth slice low amplitude (white) high amplitude (black). For scale, width of box a represents c. 180 m. (Graphics courtesy of Rose Geophysics and HES)

2.5 Galson, Isle of Lewis

The effects of coastal erosion on the NW coast of Lewis have prompted survey and excavation over the years (Dawson, 2015). While GPR has been generally well suited to elucidating the sequence of distinct stratigraphic layers in aeolian and coastal environments, its application to the detection of archaeological targets in such environments in Scotland has been limited (but see Parker Pearson et al. (2001, 64, 69) on survey on South Uist, also in the Outer Hebrides). Figure 7 illustrates how GPR, which has a long record of its ability to detect burials, produced some startlingly clear results at the thirteenth century chapel and associated graveyard of medieval to later date at South Galson (Rose Geophysics Consultants, 2015). High amplitude anomalies indicative of burials are evident owing to the good contrast between the sandy soil and the burial’s top and base. In many cases these burials correlate with the (surface) presence of headstones and grave markers. But to the west, the responses are less clear, probably due to the presence of older burials in this area, some of them overlapping and consisting of multiple interments. Across the graveyard as a whole, some weak or poorly defined anomalies are likely to represent natural variations in the subsoil.

Fig. 7

GPR survey at South Galson graveyard. GPR time slice at 1.0–1.25 m. (Graphic courtesy of Rose Geophysics)

3 Discussion

The presentation above has outlined something of the current status of geophysics in Scottish archaeology. In turn, the case studies have served to illustrate issues of methodology and data interpretation that continue to remain central to the development of archaeological geophysics as much in Scotland as further afield. Ahead of that discussion, some of the key points arising from those studies can be summarised.

The importance of the landscape survey conducted on Orkney lies in its spatial scale, richness of archaeological information and attention to the signatures of agricultural activity that, although very common, usually receive minimal treatment. The deliberately science-based effort to explain a particular geophysical response stands out in the case of the multi-faceted characterisation of ditches at Forteviot; subtle, localised mineralogical changes are invoked. This approach now needs to be extended elsewhere to different targets and under the same opportunities of sampling before and after topsoil removal. At Dalswinton it was the coordination of geophysical and aerial data with detailed archaeological knowledge, in this case, of Roman forts, that was crucial for optimal interpretation. Such an interpretation emerged in a different manner at Bothwell Castle where the survey served as an example of advantageous interplay between the outputs of complementary techniques. Yet, under favourable circumstances, use of a single technique—GPR at Galson—was more than sufficient.

On methodology, reference has already been made to the way that the tried and tested techniques feature in most surveys. Although developments in instrumentation of a given technique, such as vehicle-driven multi-sensor magnetometers, are making their appearance in Scotland, there has been little evidence of trialling of novel instruments or their combinations. In that light, those research-led investigations that have departed from this norm stand out, one of which deserves mention because it highlights the recognition of the role of geophysics and analysis of cores in a situation that is not uncommon in Scotland: the burial of prehistoric (and later) landscapes in peat. Just as the investigation of the palaeo-landscape around the Neolithic monuments on Orkney discussed above used a combination of land- and marine-based techniques, so the same approach was adopted to good effect at Calanish on the Isle of Lewis (Fig. 2) (Bates et al., 2019). Here, the peat covering the original ground surface of the Neolithic standing stones was explored by targeted conductivity, magnetometry and electrical tomography, while the stones’ proximity to lochs, as on Orkney, required bathymetric survey together with core analysis.

Regarding interpretation, the case studies have followed pragmatic, traditional routes: the identification of the main anomalies based on the combination of the geophysical response and archaeological criteria/experience, coupled with desk-based assessment of a range of information avenues from the aerial photographic record, historical sources to geological and soil conditions. Although its emphasis lay on detecting locations of prehistoric activity, the landscape survey on Orkney successfully took a fully diachronic approach, incorporating the recording of natural, pre-Neolithic phenomena through to agricultural and other activities of historical times. By contrast (and unsurprisingly), the priorities of the survey at Dalswinton (Hanson et al., 2019) and those at some thirty locations along the Antonine Wall (Hanson et al., forthcoming) have lain in identifying those anomalies that are known on the basis of morphology and context to be most likely related to Roman (military) presence. Such an approach only works because of the detailed and extensive knowledge of that presence derived from excavation along the Antonine Wall over the course of more than a century (Breeze & Hanson, 2020). Anomalies that did not ‘fit’ a Roman label were not ignored; indeed, they may be important in demonstrating, for example, immediate pre-Roman presence, but they were not classified with the same confidence.

This discussion can now look at a range of issues beyond the case studies. First is the experience of excavation in validating or otherwise the identification of geophysical anomalies. Acknowledging that this fund of information is limited in size, it is, however, instructive to look at two contrasting evaluations undertaken by Wessex Archaeology: Drumlanrig Roman fort, mentioned in Sect. 2.3 above, and a 30–40 m wide strip, 12.4 km long at Neart na Gaoithe (Fig. 2) (Wessex Archaeology, 2015). It is not surprising to find that the trends common to both of them arise from the criteria of the anomalies’ size, strength and spatial extent: where an anomaly is indicative of an archaeological feature, such as the main defence ditch at Drumlanrig fort or ditches of likely enclosures at Neart na Gaoithe (Wessex Archaeology, 2015, Figs. 34, 37), that is large on all three criteria the validation is usually very positive in a general sense, if not at a detailed archaeological level. At the other end of the spectrum, excavation of a smaller, weaker anomaly often reveals a decidedly mixed picture; the potential archaeological significance of such an anomaly may have been established on excavation but some uncertainty surrounds its identity owing to the contexts it was found in being complex and small in size. Alternatively, the anomaly may turn out to be geological. In this light, archaeological expectations of what geophysical survey can deliver need to be realistic, guided by many factors including, above all, the nature of the site. Although now dated, Linford and David’s (2001) assessment of the value of geophysical survey in the light of subsequent excavation of sites in south-east England, mentioned above, remains a model that could usefully have a counterpart in Scotland.

Turning now to the more common situation in which excavation has not taken place, there are some procedural and practical points to consider with interpretation in mind. Where good aerial photographs exist, they can often form a suitable and helpful point of comparison with the geophysical data, allowing, if necessary, revision of interpretation (see, for example, Cowley et al., 2019). Drawing on the experience from surveys at comparable sites and environments is another obvious route, and one that should be encouraged as access to, and dissemination of full survey reports becomes increasingly possible, notably through OASIS (the online UK-wide system for reporting archaeological investigations and linking research outputs and archives). The long-recognised issue of the importance of accessibility of raw data should receive increased attention in Scotland, as elsewhere in the UK, in the coming years. Despite good intentions, the idea of setting up a facility with known targets buried in different conditions seems to have found greater feasibility in forensic than in archaeological geophysics in the UK (e.g. https://www.keele.ac.uk/geophysics/forensicgeophysics/). On the matter of climatic effects on geophysical response there has as yet been no systematic assessment in Scotland. The results obtained by Clark (1996, Table 1) in southern England and Bonsall et al. (2014, 102, Fig. 32) in Galway, Ireland have indicated optimal resistance contrast between feature and surrounding soil occurring in spring and autumn. This should also apply widely to Scotland despite this region’s contrasting lithologies with those commonly encountered elsewhere in the UK and Ireland.

Of great relevance to the experience in Scotland are the effects of one or a combination of noisy magnetic background, localised soil conditions and lack of magnetic or other contrast between an archaeological feature and the surrounding soil, all of them impacting on the ability to detect that feature. As outlined in Sect. 2.2, the way forward at Forteviot was to take a rigorous soil science approach, and this has since diversified and developed as a result of the current European Cooperation in Science and Technology network Soil science & Archaeo-Geophysics Alliance: going beyond prospection (SAGA) (Cuenca-Garcia et al., 2019). This approach could be trialled at several large fieldwork projects in contrasting parts of Scotland, extending those already treated by Cuenca-Garcia (2012). From its outcomes, trends in the soil characterisation data should emerge allowing an informed interpretation to be made that goes beyond what is at present very site-specific to something that has a wider explanatory basis.

But such soil characterisation data has already also been used in its own right in Scotland as a prospection method, specifically in the investigation of spatial organisation and function. Following the pioneering work of Entwistle et al. (2000) on land-use activity at the 18th–19th century township of Knockaird on the Isle of Lewis (pH, LoI and P, Ca, K and Mg analysis) and the multi-element analysis of soils at known activity spots at abandoned farms by Wilson et al. (2005), this field of investigation has diversified (https://scarf.scot/thematic/scarf-science-panel-report/4-people-and-the-environment/4-2-geoarchaeology/). Yet its fuller potential remains to be realised, for a start by harnessing it, where appropriate, with geophysical survey, as was the case at the Neolithic settlement at Crossiecrown on Orkney (Jones et al., 2010). Although many of the components of soil analysis normally require more time and are more labour intensive than a geophysical survey, the objection that multi-element analysis is too costly to merit its inclusion in a field project should be revised. Analysis of soils in situ or in bulk samples by portable XRF, as used at Forteviot (see Sect. 2.2), is more cost and time effective than laboratory-based ICP-ES analysis, yet the quality of the data in terms of its interpretative ability is not usually inferior. As Save et al. (2020) have shown in their varied case studies in France based on pXRF analysis, it is the individual element trends and multi-element associations rather than absolute element concentrations that are important in interpreting geochemical signatures across a floor, a building or other area where one or more activities took place.

Matters to do with training and equipment merit attention. The general applications of the subject feature in undergraduate and many postgraduate courses at those universities in Scotland offering archaeology courses. The practical element features strongly in modules within postgraduate courses at Aberdeen University and the University of the Highlands & Islands (Orkney), both of which, together with Glasgow, have in-house equipment; some of the equipment held at NERC’s Geophysical Equipment Facility, hosted by Edinburgh University, has archaeological application. Sadly, the aerial photography with geophysical survey Masters course at Glasgow University, unique at its time, has not run during the last decade. As regards geoarchaeology, facilities and expertise are concentrated at Stirling and Aberdeen Universities. Overall, however, the opportunities for students to gain significant practical experience in Scotland are somewhat limited; the issue of further skills provision is a matter that the ScARF Archaeological Science Research Framework (www.scarf.scot/students), among other initiatives, is currently addressing.

4 Conclusions

This paper has set out some of the main roles that geophysics currently plays in Scottish archaeology. The picture is generally positive in the sense that the demand for surveys is being met, if not across all sectors. There is now a better appreciation of the techniques’ limitations and of the need to match individual techniques to the requirements of particular archaeological targets than was previously the case. This is important because the range of terrains, from the points of view of topography and drift geology, is more challenging in Scotland than, say, in many parts of England. On the debit side, for reasons of insufficient funding and lack of access to particular instrumentation (for instance in electrical resistance imaging (soundings) and magnetic susceptibility measurement) and associated expertise, there is a tendency to rely on the ‘standard’ magnetic and resistance techniques and to avoid experimentation. Avenues for the dissemination of fieldwork activity are many and varied, but they do not yet extend to opportunities on the part of practitioners and archaeologists to meet and critically discuss survey results. An example would be to review geophysical responses to individual features at prehistoric sites obtained from several surveys in the light of excavation evidence, where available, in a given region. And it goes without saying that another meeting of the kind held as long ago as 2009 between the Geophysics Group of the Chartered Institute for Archaeologists (CIFA) and the Association of Local Government Archaeological Officers (ALGAO), Scotland (O’Grady, 2009) is surely overdue. In brief, geophysics in Scotland needs to raise its profile, particularly at the present time of important developments in remote sensing more widely. Furthermore, while retaining its traditional independent role in detecting subsurface remains, this is also the moment for geophysics in Scotland to play a greater role within geoarchaeology. Such a move, although not novel, would see more geophysical surveys including some aspect of soil characterisation, in the first instance to progress from simply recording imperfect responses of the kind alluded to above to explaining them in physico-chemical terms. At another level, it would encourage greater integration of geophysical and geochemical survey, with the former taking the lead in locating areas of activity followed by the latter identifying the potential functions of those areas.