Variety in Archaeo-Geophysics: The French Example

Abstract

As a result of a long history in both archaeology and geophysics, France exhibits a wide panel of practices in archaeo-geophysics, going from archaeologists using geophysics as a supplementary tool for understanding their sites to applied geophysicists using archaeological sites as better constrained features and potential ground feedback. This chapter tries to scan this variety in the practices which overlap with a variety of contexts. After a brief overview of the backgrounds which control the practices in France, we show several examples that illustrate this diversity. Firstly, we will show a set of surveys of rural areas coming from both public and private institutions. Secondly, we will present how archaeological sites in urban areas are assessed with the geophysical techniques. Thirdly, we will address what can be done in what we define as the “specific” context. In each context, we will highlight how geophysical techniques could improve itselves with the help of archaeological sites took as the place for an intensive interdisciplinary research. We conclude that archaeology can be a way to make geosciences progress by bringing together geology, soil science, geotechnics, geochemistry, and geophysics.

1 Introduction

Since the first geophysical prospecting carried out by C. Schlumberger in 1912 in Normandy, French geophysicists have acquired a wealth of experience in this field and its various applications, including archaeology (Hesse, 2000). This history combined with the variety of geoclimatic conditions in the country (oceanic or continental climatic conditions and from plain to mountain environments) have resulted in many different practices in various kinds of context. Our focus will be on the practices during the 2000–2020 period. Going from general purposes to examples, we will illustrate the diversity in the French archaeo-geophysics experience. We chose to sort the examples in three sets which have been obtained through some classification processes answering the following questions:

• What is the main objective of the archaeological study?

• Does a combination of geophysical methods in addition to other techniques allow to reach the objective?

• In which context do the studies take place?

Among the answers to each of these questions, the last one appears to us as the simplest path to outline this chapter. The distinction between rural and urban contexts is fundamental regarding the choice of geophysical methods. We added a third category which corresponds to “exotic” places and that we named the “specific” contexts. In each part we describe some specificity of each context and illustrate it with several cases, scanning some other axes of interest like innovations in geophysical methods, archaeological feedback and some state-of-the-art examples. These three parts are an attempt to overview the recent state of archaeo-geophysics in France, but firstly we will introduce the background in which it takes place.

2 French Archaeo-Geophysics Background

2.1 Short Reminder About the History of the Discipline

In term of archaeological practices, France can be distinguished amongst European countries by a very specific set of official rules. Archaeological and cultural heritage studies are indeed strictly regulated. In particular, there are two kinds of archaeology. The first, research archaeology (“archéologie programmée”), refers to all studies which take place on non-threatened sites. The second, development-led (rescue and/or preventive) archaeology (“archéologie préventive”), consists in gathering as much archaeological information and material as possible before the destruction of the site (mainly due to urban and land planning). In that latter, two phases are distinguished, the evaluation (only done by state institutions) and the excavation (open to competition). The last one occurs only if the first one appears to be sufficiently fruitful. Geophysical prospecting could intervene in both phases.

The early beginnings of archaeo-geophysics in France (during the 1960s and the 1970s) mostly concerned the research archaeology (Brézillon & Hesse, 1962; Burnez & Hesse, 1967; Tabbagh, 1971; Martinaud & Colmont, 1971). Even if the Afan (Association pour les fouilles archéologiques nationales, French association for archaeological excavation), predecessor of the Inrap (Institut national de recherches en archéologie préventive, the state organisation currently in charge of preventive archaeology), was created in 1978, the question of archaeo-geophysics in the context of rescue archaeology was not really posed until the 1990s (Dabas et al., 1994; Dabas, 1999b; Ducomet & Druelle, 1996; Marmet, 2000).

This fact did not impede the technical development effort achieved during the preceding decades. It permitted to enhance the abilities of geophysical methods to quickly study wide areas using wide mesh prospecting (Dabas, 1999a; Marmet, 2000) or towed devices (Hesse et al., 1986; Dabas et al., 1994). Those advancements made application of the geophysical survey in the preventive context more affordable.

The first decade of the twenty-first century saw a transitional period in which all the actors (institutional and private) were trying to find their place. Many misunderstandings occurred, where archaeo-geophysics was considered as a real threat by some archaeologists and authorities in the preventive context (Demoule, 2014). Around the year 2010, the situation appeared to normalise. This is well illustrated by the prospecting over a large scale planned facility, a channel linking the Seine River to the Northern part of France, the Canal Seine-Nord-Europe, CSNE (Hulin et al., 2014). The approach adopted on this huge project, in addition to a PhD work (Simon, 2012) in the Alsace region, both demonstrated the limits of geophysics but at the same time these works revealed the high consistency of the use of archaeo-geophysics during the excavation phase (Hulin & Simon, 2012; Hulin et al., 2018). These works permitted to ease the tenses about geophysics and, nowadays, archaeo-geophysics appears as a tool at the disposal of the archaeologist in both preventive and research archaeology (Hulin & Simon, 2019).

2.2 Historical Highlight: Geophysical Studies Over Large Projects

If we define a large project by as one covering more than 10 ha prospected during one campaign, the first prospecting attempts were carried out, in France, on the A77 motorway project, in 1995 by the Terra Nova company. This manual survey showed that preservation of archaeological sites could rely on geophysical information and not only on direct observations in test trenches: as a consequence, a modification of the motorway route was chosen in order to avoid an important archaeological site that was only partially excavated afterwards. For the first time also, a project was managed by a GIS (Grass and Idrisi) and prediction maps of the risk of erosion/filling of potential sites using soil maps were produced (Chazaly & Dabas, 1997; Dabas, 1999b). Another consequence was the introduction of magnetic susceptibility measurements as a proxy for detection of archaeological sites for other motorways projects in France (A89, A20 and A66) even if it was shown that soil processes were also interfering with the magnetic enhancement measured over archaeological structures (Marmet, 2000).

In 2001, the Géocarta company (a spin-off from the CNRS) began the design of motorised mapping systems for precision agriculture, and specifically electrical mapping (ARP for detail, see Sect. 3.2). Then, similar motorised systems were designed for archaeology using array of magnetic sensors (AMP in 2006) and multi-coil EM sensors (EMP). The first archaeological appraisal for large development projects began in 2006. Over the next ten years, more than ten large-scale projects were undertaken, most of which correspond to development projects (ZAC) or linear infrastructures.

In 2011, the RTE project (Electricity Transmission Network) aimed at burying a high voltage electrical cable along a 22 km route. Beside a large impact for local farmers, geotechnical and soil hazards are considered to be high (presence of voids and artefacts from First World War—trenches and UXO). Archaeology was not really considered, and, due to the small dimensions of the trench (1 m wide), no archaeological appraisal was prescribed. Nevertheless, it was decided to survey a much wider swath of 100 m for detection of archaeological sites. Beside the standard ARP and AMP methods, multi-coil electromagnetic induction (EMI) mapping (using a DualEM421 system) was carried out in order to reach larger investigation depths (6 m). Resistivity maps made it possible to detect many artefacts linked to this battle zone (trenches, bomb impacts, access roads) and many pyrotechnic elements were found. None of these elements gave rise to an archaeological excavation given the very low impact of the burial trench. The final route was defined inside the 100-meter zone, considering all the information from the geophysics. In addition, it should be emphasised that the acceptance of the project by farmers was favoured by the production of soil maps derived from resistivity to improve yields.

2.3 Environmental Background

In Europe, France is one of the countries with more than three biogeographical units over a significant part of its territory (EEA biogeographical region map 2017). These units are defined to reflect the climatic and floristic homogeneity. Pedological maps (Gis Sol, Les sols dominants de France métropolitaine, 2011; Gis Sol, 2011) are another way to illustrate the variety of contexts which can be found in France. The combination of all these facts makes the design of archaeological prospecting complex. Thus, some regional/local expertise is required to correctly define the type of geophysical method to be implemented. In addition, changes in the land-use management defines which kind of area would be investigated by archaeologists especially in preventive context.

In the last decade, according to the French central agency for sustainable development (Commissariat Général au Développement Durable, 2015), the Corine landcover index on land use exhibits changes inside and outside its major groups. This index is categorised in five major groups: artificialised surfaces, agricultural areas, forested and semi natural areas, wetlands, and waters bodies. If we gather the last three ones under the same label, it appears that these categories partially correspond to the ones that make sense for the geophysical methods implementation i.e. rural, urban and specific (this last category corresponding to surveys pushing the methods we use at their limits see Sect. 5 for more detail). We will use these as the main scheme for the following parts. The Commissariat Général au Développement Durable report (2015) mentions two key points: firstly, there is an increase of nearly 13% of the artificialised surface (350 kha) between 2000 and 2012. Secondly, one third of the artificialised surfaces which exhibit a change of land use between 2006 and 2012 were already artificialised in 2006. If archaeogeophysical surveys follow these trends, thus the ratio of surveys in preventive archaeology should be around 2/3 for rural context and 1/3 for urban context.

2.4 Trends in Archaeo-Geophysics Between 2000 and 2020

For this purpose, we built a database gathering as many surveys by as many teams as possible. More than 1100 surveys other the past 25 years were collected. Table 1 sum up the various entities which have contributed.

Table 1 Number of surveys in the database by contributors

For this study, we limited ourselves to metropolitan France surveys and to ground measurements. We chose the year 2000 as a starting point, for two main reasons:

• The selective availability for the use of the GPS signal came to its end, allowing the survey of wide areas with a more accurate positioning and faster data acquisitions.

• The digital technology became extremely cheap, increasing greatly the size of the recordings.

Both above-mentioned elements allowed a great enhancement in the quality of the geophysical dataset obtained independently of the scale of the study. In addition, the 2001 law for preventive archaeology was adopted and Inrap was created. This French specific scheme had huge implication on archaeological policy.

Considering the diversity of actors, we decided to focus on the following elements:

• year of the survey

• surface covered

• geophysical method used

• context of the study (rural, urban, specific)

• type of archaeological context (preventive or research)

The first graph (Fig. 1) shows the evolution of the number of surveys by context since 2000. Most of them were done in rural context, though it is noticeable that the urban context is present from the beginning of the period too. Specific area surveys appear mostly in the late 2000s with no explanation (maybe due to the limited number of surveys between 2000 and 2005).

Fig. 1

Number of surveys in each context over the period from 2000 to 2020 (a) in research context, (b) in preventive context

First the rural context is almost four fifths of the planned activity (Fig. 1a). In preventive archaeology (Fig. 1b), it appears that the number of surveys reached a plateau of about 30 surveys by year (with a marked decrease between 2012 and 2015 due to the combination of several factors (end of the CSNE project, end of the F.-X. Simon PhD, the “mesure 14” (Fichet de Clairfontaine, 2014), etc.) after an increase in the first year of the 2000s. In both preventive and planned archaeology, it also seems that the number of surveys in urban context is slightly increasing in proportion, mainly in preventive archaeology. It could be added that preventive surveys account for a third of the surveys done each year. Last point, the ratio urban over rural surveys in preventive archaeology seems to be around one third as expected according to the Corine landcover index.

Fig. 2a shows the number of surveys belonging to one of the three surface classes. Each category was defined according to the tertiles of the surface of our dataset (in the dataset, ‘small’ means an area under 3510 m² and include profiles, ‘intermediate’ means areas between 3510 m² and 20,000 m² and ‘large’ means areas above 20,000 m²). This graph clearly shows that in the early 2000s the ‘intermediate’ areas were the most numerous in the surveys. After that, boosted by the instrumental improvements, the ‘large’ areas saw a fast increase and certainly, some surveys which would have been scattered in ‘small’ or ‘intermediate’ if manually handled were then done as one. After 2014, the ‘small’ surveys numbers increase slightly. It is very interesting to point out that in the last years, each kind of surface class was around one third of the total amount.

Fig. 2

Repartition of the surveys: (a) by type of surface and context; (b) by geophysical method and type of surface

This overall trend can be explained by a combination of factors. In rural context, the ‘large’ areas are prominent while the ‘intermediate’ and the ‘small’ ones are more representative of the urban and specific contexts. As for the methods used (Fig. 2b), the classical methods (electrical resistivity and magnetometry) dominate. EMI proportion slightly decreases as the size of the area prospected increase. Ground penetrating radar (GPR) has a specific pattern being much less used in large areas. “Exotics” methods appear to be done over small areas which is consistent with their use for specific issues.

3 Geophysics in Rural Context

3.1 General Overview

The simplest conditions for the implementation of geophysical survey are best encountered in rural or open areas: these environments are easy to access and allow correct positioning using high precision GNSS systems. The use of towed systems for large areas and almost all geophysical methods can be implemented. In addition, these contexts are generally characterised by a lower level of ambient (mechanic and electromagnetic) noise and a limited number of metallic objects in the vicinity of geophysical sensors. For all these reasons, surveys in rural contexts still represent a significant part of the archaeo-geophysical ones undertaken in France, with 879 studies recorded between 2000 and 2020.

Despite all these advantages, some limitations should be noted. For example, the use of manure leads to soil properties modifications or entails modern metallic pieces (Dabas et al., 2021), plant sizes (vineyards, orchards), drainage and others artificial features which may impede the use of several methods (Simon et al., 2021). As soon as the soil is ploughed or harrowed, a very high degree of heterogeneity of the surface horizon can be observed, which can both prevent the use of the equipment and generate significant noise levels. France is a country with a highly intensive agricultural level and these issues are omnipresent for geophysical survey planning. Erosion issues due to agriculture should also be mentioned because it could seriously affect the archaeological remains preservation and the ability to detect them by geophysical means. As mentioned in the introduction, the variability of soil cover and climatic area on a national scale, have an impact both on the methods that could be used and the types of archaeological features we are looking for. A survey in south-eastern France on a Neolithic settlement will be different from the one of a settlement of the same period in North of France. Thus, it appears irrelevant to comment on the effectiveness of any method in France because there is no rule of the thumb about the method that could work in any given context. Therefore, and apart from specific contexts like mountain ranges and swampy areas, which will be discussed in Sect. 5, the use of several methods must be evaluated. Depending on the conditions, geophysics can be very ineffective, whatever the method. The use of towed systems makes it possible to cover large areas but also to acquire data sets with stable systems. By a combination of a GNSS system with accurate positioning, it is possible to reveal very small archaeological features when contrasts of geophysical properties are high enough. In France, these towed systems concern all geophysical methods (magnetometry, electrical resistivity, GPR and EMI) with notably the development of towed resistivity systems.

3.2 Methodology Highlight: Towed Electrical Resistivity Measurements Systems

In the middle of the 1960, a team led by Albert Hesse worked on the idea of a continuous electrical resistivity measurement for field surveying. The idea was to overcome the main limitation of the electrical resistivity measurements which requires to drive the array manually. Two major difficulties have to be overcome, the contact between the ground and the moving electrodes and a short duration measurement compatible with the motion of the array. The first efficient system was called RATEAU (Résistivimètre Auto-Tracté à Enregistrement AUtomatique, towed resistivity meter with automatic recording) and its specific electronics made the measurement of electrical resistivity possible while moving (Hesse et al., 1986; Dabas et al., 1989). Since this first version, the need to take measurements at several depths of investigation has led to a multi-electrode system, called Multi-depth Continuous Electrical Profiling (MuCEP, Dabas et al., 1994, Panissod et al., 1997). This system, first developed at the Geophysical Research Center of Garchy (later included in the UMR7619 Paris VI, Sisyphe/METIS) uses a specific geometry with electrodes arranged in a ‘V’ shaped array hence its first name of “duck flight” geometry.

The system consists of four axles, each made up with two spiked wheel electrodes. The first axle is the transmitter one. The other three axles are for measurements. According to modelling, it was shown that their distances to the injection dipole correspond more or less to the depth of investigation of each channel (Panissod et al., 1997).

In the early 2000s, the ARP system (Automatic Resistivity Profiling, Geocarta, Paris), derived from this previous development (Dabas, 2008), coupled absolute positioning information (RTK-type dGPS) with relative positioning (Doppler radar), liberating from any prior topography. This device therefore combines the advantages of continuous electrical profiling, a limited three points vertical electrical sounding (VES) and overall allows prospecting around 4 ha (up to 10 ha maximum) per day with an infra-metric spatial resolution. The distance between the profiles can be of the order of one meter at most. The measurements are sampled approximately every 10 cm along the profiles. The entire system is controlled in real time by a specific Geographic Information System (GIS).

We may point out that ploughing, or even stubble cultivation, pose the problem of moving machinery over this type of land and a change of soil density (introducing unwanted increases in resistivity). The combination of this techniques with magnetometry will be illustrated by the first example, the study of Fontaine-les-Bassets. The second example will be about the use of EMI electrical conductivity prospecting on the case of the autonomous port of Dunkerque. The third example is excavation feedback on the site of Longvic.

3.3 Combination of Magnetometer and Electrical Resistivity Survey: The Fontaine les Bassets Site

Identified in 1989 by aerial photography, the site of Fontaine les Bassets corresponds to a large Roman settlement which was not excavated at that time. In 2009, the archaeologist decided to jointly use three non-destructive prospecting methods: aerial, walkover, and geophysics in order to define the limits of this city and its internal structure. One of the objectives was to establish possible areas for archaeological excavations (Quévillon, 2012).

Like many other experiments carried out on similar sites (Le Vieil-Evreux: Dabas et al., 2005; Mandeure-Mathay: Thivet et al., 2009; Les Tours-Mirandes: Dieudonné-Glad, 2010), the electrical towed system ARP was chosen for its speed of acquisition and according to the types of structures sought. Electrical resistivity survey was carried out in three campaigns. As it is often the case, the first mission (6 ha in 2009) was carried out on the main aerial evidence showing a possible forum, streets, and a dense settlement. It allowed to validate the use of this method, all the built-up elements appearing as resistant anomalies with a strong contrast compared to the surrounding (ratio of about 1 in 7 for a background resistivity at 100 Ω.m). The missions of 2010 (14 ha) and 2012 (8 ha) made it possible to map the rest of the plots (Fig. 3a). All the anomalies appear in the first ARP channel (0 to 0.5 m) showing a superficial origin of the structures, which can also be corroborated by the fact that the cropmarks were clearly defined. Linear resistant anomalies could correspond to the ancient street network: four main axes oriented north-east/south-west as well as two north-west/south-east axes divide the urban space into irregular blocks. Within or bordering these blocks are some anomalies which can correspond to large buildings and others of more modest sizes. The density within each of the blocks is very different, showing a well-characterised central area around a quadrangular resistive anomaly which could be associated with a central square/religious complex. In addition, 200 m to the East, a semi-circular feature probably associated to a theatre is detected. The majority of anomalies were resistant, as is the usual case for features made of stone. Nevertheless, there were linear or punctual anomalies, in particular in the Southwest quarter of the plot, showing the existence of structures of a different nature, perhaps related to a small remains of ironwork material such as slag (identified during the walkover survey).

Fig. 3

Results of the prospection at Fontaine-les-Bassets. (a) 28 ha electrical resistivity ARP survey (113 to 571 Ω.m)—channel 1 (0 to 0.5 m) overlaid over orthophotograph (Bing Aerial Map), (b) 3 ha magnetometer survey (−15 to 15 nT/m) overlaying the ARP data

 Finally, in 2014, a magnetometer survey (3 ha) was decided to provide additional information on the two main buildings (square and theatre) and the area between them (Fig. 3b). As is often the case with the anomalies associated with a magnetic gradient measurement, these are of shorter spatial wavelength than in electrical resistivity prospecting and therefore allow better definition of the structural limits. But a more in-depth study of the anomalies shows a larger number of electrical anomalies, in particular for the street network, which is not very visible in the results of the magnetometer survey.

3.4 Geomorphological Study of the Dunkerque Autonomous Port

When the main objective is not to map the remains of a site but mostly reconstruct its landscape, the use of geophysical methods has been clearly proven to be efficient for decades (Castanet et al., 2015). Such a survey can be used as on its own and in the context of preventive archaeology. In that case, it allows to spatialise geomorphological features, which have been studied by the geomorphologist on a case-by-case basis and make it possible to go beyond the sampled vision of trench surveys.

This study, part of the extension of the autonomous port of Dunkerque, is a good illustration of this contribution. In this sector, an initial EMI survey was carried out in 2015 with an EM31 (Geonics Ltd) on a surface of 56 ha. It highlighted the interest of the EMI method for geomorphological applications on this coastal plain. This first campaign was followed by a survey carried out in the commune of Bourbourg as part of the same project. This second survey was carried out over two consecutive years in 2017 and 2018. During this time, approximately 180 ha were surveyed with profiles less than 10 m apart. The EM 31 was fixed on a trolley to facilitate its transport and was associated with a high precision GNSS positioning system. The results of the survey are shown in Fig. 4.

Fig. 4

Archaeological study of Bourbourg—Dunkerque autonomous port project, map of the apparent electrical conductivity (VCP mode)

The strong contrasts between the filling of the clayey features and the sandy background enabled the visualisation of ancient channels, the detection of insulated conductivity anomalies and identification of the general shape of the palaeolandscape that is now filled in. Given the very loose sampling, these geophysical datasets did not provide archaeologists with no precise evidence of human settlement. The very specific framework of this operation, as well as its annual scheduling, made it possible to plan the best time to intervene in the field.

Given the richness of the results, this type of survey was subsequently recommended by the local authorities to support the trial trenching and to provide the archaeologists with context setting maps of the remains. It also serves as a guide for geomorphological trenching, which allows the best possible targeting of the test pits and the documentation of the dynamics of sedimentary filling. The hindsight that is made possible by the multiplication of interventions in this sector has enabled the acquisition methods to evolve towards towed and multi-depth systems and to carry out cross-sections thanks to electromagnetic data inversions (Guillemoteau et al., 2019). Finally, in 2021, a total of 513 ha have been surveyed with EMI in this area to offer a global view of the palaeolandscape in correlation with archaeological features later found by trial trenching.

3.5 Excavation Feedback: The Longvic Magnetometer Survey

Excavation feedback is a great opportunity to confront geophysical data and interpretation. Due to the specific French legislative context concerning preventive archaeology, prospecting rural and open areas offer high potentiality of cross validating geophysical interpretation with archaeological feedbacks, especially during preventive excavation, even though it is currently under exploited. This cross-study of data sets enables to reflect on the undeniable difficulties encountered when geophysics is used as the only method of archaeological evaluation, but also to reconsider the interpretations. The example of the magnetometer survey on the archaeological site of Longvic illustrates this issue.

Following the discovery of a multi-phase site on a large area on the outskirts of Dijon during an archaeological field evaluation, a large-scale excavation was requested by the local authorities. Some sectors were tightly selected as a result of the trial trenching. However, other areas were designated as high potentiality but the extension and location of these evidences remained widely undefined. In order to constrain their location and to minimise unforeseen features during the excavations, a magnetometer survey was carried out over the whole set of areas to be excavated (i.e. approximately 7 ha).

Given the a priori knowledge on the nature of the soil, the land cover and the typology of the remains found in the previous step, an intervention plan was set up based on solid arguments. Magnetometry was first benchmarked over a representative test zone and, following the positive results, extended to the whole area. The map obtained made it possible to link the archaeological features observed in the various trenches to each other and to provide new clues on the potential presence of remains.

The Roman enclosure discovered during the archaeological evaluation was observed on the magnetometer survey map (Fig. 5a). Associated with this enclosure, numerous localised anomalies of medium to high amplitude could be observed and interpreted without ambiguity as archaeological pits. On the contrary, large blurry anomalies whose filling appeared to be heterogeneous were interpreted as natural features link to sedimentary features. The excavation did not validate the archaeological interpretation based on the geophysical maps. It was observed that the local anomalies corresponded to natural features (Fig. 5b), some shallow slumps in the substratum (an ancient alluvial terrace). The large and diffuse anomalies (Fig. 5c) arose from large anthropic developments with numerous storage pits that can be assessed on the geophysical data afterwards.

Fig. 5

Archaeological study of Longvic. (a) Magnetic anomalies map, (b) View of the shallow slumps in the gravel background (G. Videau, Inrap), (c) Stratigraphic section of the storage pits ensemble (G. Videau, Inrap)

This feedback, allowed by the archaeological excavation, highlights all the ambiguity between archaeological features and geophysical anomalies despite optimal conditions of surveying. As the quality of the geophysical dataset on this type of site is good, cross and a posteriori analysis is possible. However, this case study invites us to always interpret geophysical data cautiously and to advocate, as soon as possible, a comparison with the excavation which allows to account for possible and unavoidable inconsistencies.

4 Geophysics for Urban Archaeology

4.1 General Overview

Urban archaeology takes an increasingly important part in French archaeological research. The current trend in urban planning and the refocusing of development in town implies that preventive archaeological studies take place in city centre areas. Often associated with numerous constraints (accessibility, narrowness of the study areas, developed stratigraphy, backfill), urban archaeology more and more requires the support of alternative methods to strengthen their studies and among them geophysical methods play an important role. Indeed, our database includes 200 urban surveys.

For geophysics application, urban area presents a certain number of constraints. They are generally well known: presence of mechanical vibrations, electromagnetic noise, presence of infrastructures, urban furniture, pedestrians, cars obstructing the surface, presence of underground modern utilities, and most obviously, the presence of buildings and other superstructures which shatter the prospected area (e.g. Atanasova et al., 2014). These inconveniences combined to the heterogeneous state of the soil make it very challenging for the archaeogeophysicists.

Analysing the past of a city requires a wide approach based on iconographic, textual, geological and archaeological sources. Each of them could have their own timing and induce methodological biases. Geophysics, despite some constraints, has its own place in the evaluation workflow. The primary goals are mostly the determination of the density and the thickness of archaeological remains areas and/or the estimation of the stratigraphic sequence.

Results are often more complex to interpret in term of archaeological evidence than in rural areas and involves a collaborative work. In urban archaeology, more than elsewhere, the combination of all these approaches is necessary to develop reflection on a research area.

Depending on the project, several approaches on different scales can be carried out in an urban context, from the study at the city scale to recognize the ancient topography to more targeted studies at the scale of a plot allotment (Atanasova et al., 2014) or even a building (Bully et al., 2011). To deal with such issues, as in other countries, GPR is the most common method used in French urban archaeology. However, other techniques may be of interest such as the electrostatic method, also called capacitive coupled resistivity (CCR), highlighted below. This method is the counterpart of the DC electrical method and makes it possible to inject and measure current on hard ground by electrostatic poles. It is clearly complementary to the GPR method, especially for investigations over large areas or urban contexts (e.g. Dabas & Panissod, 1999). Other techniques, such as seismic methods have been used, on a more anecdotal basis, to image larger structures (e.g. in Thiesson et al., 2021 for defensive ditch).

4.2 Methodology Highlight: The Electrostatic Method with Capacitive Coupling (CCR)

The Capacitive Coupled Resistivity (CCR) or electrostatic method is certainly one particularity of current French research development. The electrical resistivity method is limited to areas where the probes or spiked wheels can be pinned in. Even if some examples of “wet” probes had been implemented successfully (Athanasiou et al., 2007), one idea to overcome this limitation is to use capacitive probes.

This development arose in the middle of the 1990s (Grard & Tabbagh, 1991; Tabbagh et al., 1993). As the geometry of the array could be similar to those used for electrical resistivity survey, it was considered at first as mean to extend the area which could be investigated using classical electrical resistivity mapping. The depth of investigation (about 10 m) is limited by the impedances of the pole (minimum operating frequency above 1 kHz) and the low induction number assumption (restricting both the maximum pole spacing and the maximum frequency). It makes this method very suitable for archaeological studies especially in the urban context (Dabas & Panissod, 1999). A significant number of studies has been achieved in monuments using this technique (e.g. Titus et al., 2001). Its combination with GPR techniques appears to be very informative as shown by the examples of Saint Germain Abbey (Sapin, 2000) or Saint-Étienne cathedral (Titus et al., 2001) both in Auxerre, the Gigny church vestibule (Bully et al., 2010) or more recently Notre-Dame de Paris (see Sect. 4.3 and Hulin et al., 2021a).

The CCR could also be adapted to evaluate the conditions of walls and identify stones (Souffaché et al., 2016) or to detect wall features hidden by surface dressing (Bully et al., 2010). One can imagine that this latter topic could be one leading development theme for this technique in future years in addition to GPR techniques (see Sect. 4.5 below).

Another development to consider is going beyond simple resistivity measurements. As the frequency can be varied and its effect is not negligible, it is necessary to take into account the polarisation phenomena to assess both the resistivity and the effective dielectric permittivity (Schamper et al., 2021; Tabbagh et al., 2021). Simultaneous measurements of several parameters could be of great interest especially in urban studies where the need for supplementary information is necessary to strengthen the interpretation.

The combination of CCR with GPR is illustrated by the first example in Notre-Dame de Paris Cathedral. As GPR is clearly the technique most adapted the urban context, the second example presents archaeological feedback on a small area in preventive archaeology. The third example shows the use of GPR indoor and belongs to the blurry limit between urban context and specific context.

4.3 Monuments Studies: The Notre-Dame de Paris Example

Geophysical survey inside buildings is an important field of application for French prospectors (Dabas et al., 2000). Requests are generally made when an excavation is difficult to carry out for obvious technical reasons. The archaeologist requires geophysics to gain insight into the study area. In France, these surveys mainly concern religious buildings, but also more atypical places such as cellars or crypts (Tabbagh et al., 2002; Bully et al., 2011). In this latter case, it is possible to get closer to the archaeological layers. “Underground” surveys represent a great opportunity for archaeologists to reach deep archaeological stratigraphic levels.

Implementation of geophysical prospection within inner areas is constrained. For instance, ground conditions such as metallic reinforcements in concrete or wooden floor with an air layer can make GPR measurement impossible.

An emblematic example is the recent study performed inside the Notre-Dame de Paris cathedral (Hulin et al., 2021a). Following the fire of 2019, the French Ministry of Culture requested the Inrap’s geophysics team to carry out a study of the cathedral’s floor. This survey was one among a wide panel of tools deployed to study the burnt cathedral. The objective of such a survey was twofold. Firstly, to take advantage of an empty cathedral to understand what was beneath the soil. Secondly, to anticipate possible restoration works affecting the near subsoil and the probable archaeological remains.

A dedicated methodology was implemented with three complementary interventions. First, a very high resolution GPR survey with a 3D-Radar step frequency multi-antenna (Fig. 6a). Then a GPR survey with a GSSI 350 MHz Hyperstacking antenna to investigate deeper anomalies. Finally, an electrostatic survey for measuring electrical resistivity and dielectric permittivity of the soil. For the electrostatic survey, the MP3 prototype (Fig. 6b; Flageul et al., 2013) from UMR METIS (Sorbonne Université, Paris) was used. The device has a V-shaped geometry and three different pole spacings (0.70/1.14/2 m).

Fig. 6

The Notre-Dame de Paris Cathedral survey: (a). 3D-Radar towed with a remote-controlled machine and results; (b) the CCR survey device and results

This geophysical survey was carried out in a particular context requiring additional safety rules. First, a very high level of pollution from the hundreds of tons of incinerated lead from the roof and the spire had to be considered for the survey implementation and for the protection of persons and equipment. In addition, a large part of the cathedral (the nave and the transept) was totally restricted due to potential falls of building materials from the roof. These particular conditions required the geophysical devices to be adapted on a remote-controlled machine which performed the survey (Fig. 6a).

The combination of GPR and electrostatic surveys was successful and enhanced the knowledge of this emblematic monument. It has allowed us to recognise a poorly known service network and has also revealed totally unknown remains with notably a large wall in the northern side-aisle. GPR gives very detailed information about the near surface whereas electrostatic gives information about electrical resistivity distribution in the ground. In such a case, electrostatic can be considered as a GPR reliability map based on the information about soil conductivity.

4.4 Geophysical Studies Over Very Small Areas

In the context of urban archaeology, studies over limited areas constitute an important part of the surveys carried out these last years. The case study of the Notre-Dame car park project in Cherbourg is a good example of this kind of survey (Paez-Rezende & Hulin, 2021). The project covers a surface of approximately 5000 m² with significant urban infrastructures and furniture. The area was already excavated in the 1970s and many archaeological features were suspected on this area as a Roman castrum, a medieval castle, a church, and its cemetery. The prior knowledge indicates a potential stratigraphy of 4 m thick. As required by French laws, an archaeological evaluation based on trial trenching was done prior to the development of the car park. In addition to the trial trenching, a GPR survey was carried out as a preliminary.

The geological context, consisting of shale and sand, led to the choice of a GSSI 350 MHz Hyperstacking combined with a robotic total station for a real time positioning. The use of such device as an accurate replacement of GNSS constitute one of the main differences with open area studies for the positioning of geophysical data. GPR results are clearly reliable and give much information at different depths (Fig. 7a). Trial trenches were implemented according to the geophysical results to test archaeological features but also areas devoid of geophysical evidence (Fig. 7b, c).

Fig. 7

Cherbourg-Notre-Dame car park: (a) GPR depth slices (b) GPR survey and orthophotograph of the excavation over the medieval castle (c) GPR survey and orthophotograph of the excavation over the cemetery

In such a context, GPR data is clearly a valuable tool, as a complement to trenches and their limitations (size, underground utilities, traffic…). Walls corresponding to the medieval castle were clearly identified by GPR in perfect accordance with trial trenching results (Fig. 7b). More discreetly, the map of the medieval cemetery was expanded through punctual anomalies. These correspond to burials covered by slabs of schist (Fig. 7c). Burials without stones were not detected by GPR and are only identified by trial trenching. These latter correspond to the majority of the burials found on this site which shows the necessity of excavations to find this kind of artefacts. The back and forth between geophysical data and archaeological observations from trial trenching, based on the use of GIS tools allows to upgrade both approaches and provides a more exhaustive view of the archaeological context. For small to very small urban areas, this kind of approach is very efficient and need to be developed. However, it requires a close interaction between the archaeologist and the geophysicist and has to be wisely used, always combined with trial trenching.

4.5 Seeing in the Wall: The Commandery of Jalès Example

Among the approaches that are still underdeveloped in archaeological context, the auscultation of the walls of buildings is an important aspect that is mostly still in its infancy stage in France. The archaeology of buildings (grounds and walls) is also subject to the French preventive archaeology law. In fact, any change on the facade of an ancient building has to be preceded by an archaeological evaluation.

Many buildings walls have been covered with cement coating and are therefore no longer accessible to archaeologists. A destructive process of staking out the coating is then carried on an ad hoc basis to assess the archaeological potential on a few test areas. Very high frequency radar antennae (>1 GHz) can be used to answer much-localised questions by imaging a filled-in opening in the wall, a particular stone arrangement and so on. As for ground surveys, positioning the geophysical data by a robotic total station can greatly improve the quality of the survey (Benech et al., 2021; Hulin et al., 2021b).

The case of the commandery of Jalès is a significant illustration of the kind of information provide by a GPR survey for the archaeological study of buildings. The commandery was founded by the Templar order during the twelfth century and has been continuously occupied since then. The study of the outer face of the walls revealed different phases of construction, especially during the twelfth, fourteenth, and eighteenth centuries. The inner faces of the same walls are mostly plastered, and it was difficult to gather more information about the inner evolution of the building without geophysics. The survey of these walls was carried out with a pulseEKKO^® Pro system (Sensors & Software) associated with a TR1000 antenna (centre frequency of 1 GHz) which allows an investigation depth of around 1 m (Fig. 8a). The walls were surveyed in both vertical and horizontal direction using a grid mesh of 5 cm.

Fig. 8

Example of GPR wall diagnostic on the commandery of Jalès (Ardèche, France) a GPR setup using a pulse EKKO PRO TR1000 with a grid mesh of 0.05 m, b Result showing traces of an arch appearing 0.25 cm-deep inside the wall

The results obtained on different inner walls of the commandery appeared to be complementary with the building study: the GPR image brought significant information about the different phases of the walls, revealing reinforcements or reconstructions not visible from outside, as well as architectural elements of the earliest phases of the building like arches, sealed windows, or heterogeneities in the use of building materials (Fig. 8b).

This experimentation showed how useful can be such survey for the building studies, even if the walls are not plastered. Such an approach provides supplementary details on the different phases of construction and/or reconstruction of the walls to the traditional building architectural studies.

5 “Tailor-Made” Prospections in Specific Contexts

5.1 General Overview

Beside of the rural and urban context, which represent most of the cases in archaeo-geophysics, a third category, accounting for 56 surveys in the database, is explained in this section. What we are calling ‘specific’ is any context (including environment and people involved) which bring us to consider a prospection pushing the limits of the methods we use.

In France, this kind of work is almost done by research teams because it requires some special designs which cannot be automated thus, they are not cost effective in time or/and in money. We gather under this cap studies taking place in challenging environments such as:

• indoor (in caves or building)

• in a humid context (coastal or wetland)

• in mountainous areas

• in forested areas.

• on stripped areas

Some of these contexts overlap with both previous ones, it is an illustration of the limitations of our classification. For example, the studies of stripped areas are undertaken in urban or rural context. Nonetheless, they need very strong interaction and high reactivity from both the archaeologist and the geophysicist which make it occurs rarely outside “specific” context.

Areas covered in this kind of studies are usually small (less than 1 ha). All geophysical methods could be used in this category on the condition that they can afford manual handling. The archaeological context of such studies goes from prehistoric sites with tenuous clues of human activities like these presented below to ancient mining facilities (e.g. Florsch et al., 2011, 2012, 2017). In these kinds of studies, the measurements of several parameters could be crucial.

The studies of salty wetlands, such as the numerous marshes along the French coast, which also present a strong archaeological potential are another good example. These environments can correspond to agricultural meadows, but also to more hostile environments such as wastelands sometimes covered with brackish water and often very vast and difficult to access.

It is necessary to define a protocol allowing to cover the whole area in spite of the difficulties related to the environment. This is most often done with a combination of several techniques. The most common set up is a wide mesh mapping used to detect areas of interest then more detailed studies on the spot chosen. Such an approach was used on the La Perroche marsh to study the physical environment of a prehistoric site (Laporte et al., 2009; Clavé-Papion et al., 2009). It has also been successfully followed for the research of ancient port installations, both on the Mediterranean (Mathé et al., 2016, 2018) and Atlantic coasts (Mathé et al., 2012, 2020).

It should be noticed that the EMI devices are a good candidate for these kinds of studies because they can measure two to more parameters simultaneously.

5.2 Methodology Highlight: (Electro)Magnetic Signal Measurements

EMI survey is mainly used to measure the electrical conductivity of soils often on large area as its acquisition time is fast and no contact between the soil and the device is required (e.g. in Sect. 2 Rural context). Nonetheless, this is only a small part of what can be assess with these methods.

Since the seminal works of Aitken, Colani, Tite and Mullins in the late 1960s (Colani & Aitken, 1966; Tite & Mullins, 1971), EMI devices are identified as sensitive to magnetic properties in addition to conductivity. In fact, Tabbagh (1974, 1986) demonstrated the conditions under which electrical conductivity and magnetic susceptibility responses can be separated.

Although devices with high quality phase detection are now available, the calibration of in-phase measurements is still challenging (Thiesson et al., 2014; De Smedt et al., 2013; Delefortrie et al., 2014, 2018) and the additional handling it requires can drastically slow down the survey. Nonetheless, after this step of calibration, the electromagnetic signal can be considered as robust and allows further processing. For example, it is possible to combine magnetic and EMI measurements to assess the remnant part of the magnetisation or to map the magnetic viscosity which allows some consideration about magnetic grain size distribution (Benech et al., 2002; Thiesson et al., 2007; Pétronille et al., 2010; Simon et al., 2012).

For approximately 20 years now, multi-spacing loop-loop systems have allowed geophysicists to work on the inversion process. 1D solution is now commonly used by the geophysical community, but this advanced processing is barely used for archaeological studies (Guérin et al., 1996; Brinon et al., 2012). These multi-spacing instruments also allow the 3D inversion of magnetic susceptibility in the frequency domain (Thiesson et al., 2017b; Guillemoteau et al., 2019). This last solution is particularly interesting for archaeological prospection as the magnetic contrast of the archaeological features is greater than electrical conductivity ones. This recent advance need now to be applied more often and associated with robust calibration procedure.

EMI prospecting, like CCR and Spectral Induced Polarisation methods, has the potential to go beyond simple measurement of electrical resistivity, and goes a step further by potentially assessing both magnetic properties (susceptibility and viscosity, Simon et al., 2015) and both electrical properties (conductivity and permittivity, Simon et al., 2019). The first step was achieved with multi-spacing, orientation or frequencies devices and has given some encouraging results (Benech et al., 2016; Thiesson et al., 2017a; Simon et al., 2019). Finally, the aim is to improve the EMI method beyond its frontiers combining measurements to propose apparent properties map which could be used for defining geophysical typology to be compared to archaeological or pedological ones (Thiesson et al., 2017b; Tabbagh et al., 2021).

The first example of this section is an illustration of the use of magnetic susceptibility measurements on a stripped surface. The second example concerns combination of geophysics and geotechnics to access the thickness of a sediment in a cave. The last example deals with coastal prospecting and its specific limitations.

5.3 Recognising and Characterising Anthropogenic Phenomena on a Stripped Surface

Geophysics can be implemented directly on a stripped surface during an excavation (e.g. David et al., 2003). This approach has been largely systematised for the first time in north of France, on the CSNE project (Hulin et al., 2012) followed by applications in Alsace (Simon, 2012; Simon et al., 2012). The objective here is different as it concerns the characterisation of soil levels and archaeological structures.

This characterisation is mainly based on magnetic parameters. It is indeed known that some human activities can modify the content and composition of iron oxides in soils and, consequently, their magnetic properties. These include heating, iron working and, to a lesser extent, organic matter (Le Borgne, 1955, 1960, 1965; Aitken, 1958; Tite & Mullins, 1971; Marmet, 2000). Thus, the measurement of parameters such as magnetic susceptibility or magnetic viscosity can provide valuable clues to these different man-made phenomena. This magnetic characterisation is then added to the observations made by the archaeologist in the field, thus allowing a better understanding of all or parts of the site.

As the measurements are taken on a stripped surface, the removal of the topsoil gives two main advantages. On the one hand, it offers the possibility to get closer to the archaeological levels, hence obtaining a stronger geophysical signal. On the other hand, it makes it possible to get rid of an important source of magnetic noise generated by the heterogeneity of the shallower horizon (ploughed layer with out of place objects for example). The measurements carried out on a stripped surface are less noisy and have a higher dynamic for the signal of archaeological origin. Finer variations are more likely to be detected especially those which have left no visible traces on the ground. These are known as magnetic ghosts (Linford, 2004; Fröhlich et al., 2005; Hulin et al., 2012; Simon et al., 2012).

One of the most relevant applications is undoubtedly the detection of iron working areas like on the site of Sauchy-Lestrée which is particularly representative. After topsoil stripping, there was no indication of the presence of an iron working area. However, the magnetic susceptibility study revealed the presence of a well-shaped anomaly on the stripped surface directly on the soil considered as natural (Fig. 9a). Clear boundaries are present and can be interpreted as wall effects which are to be related to the archaeological plan.

Fig. 9

Sauchy-Lestrée: Results of the survey over the stripped surface (a) magnetic susceptibility survey with MS2D (b) Smithing micro waste in proportion

Based on this detection, a grid was set up to take samples. Once processed, all the samples showed only a small proportion of magnetic elements per square, which can be explained by the high erosion level of the site. This corresponds to a “pollution” from the iron waste that has migrated to the depths due to bioturbation phenomena. Despite the erosion, there is a strong correlation between the micro-waste map and the geophysical map (Fig. 9b). In some places, the results diverge somewhat. Some relatively magnetic squares do not show high proportions of micro-waste. The heating hypothesis is therefore to be considered in this case. At the location where the high susceptibilities are correlated with high proportions of magnetic elements, microscopic observation has allowed the presence of typical hammerscales to be detected.

These observations give a first global picture of the organisation of the smithy which could be dated to the first half of the first century BC. Thus, the space where the metal was struck seems to be in the south-western part of the area, where one pit could have served as an anchoring point for the anvil. The mapping carried out after the treatment of the sediments also made it possible to observe an extension of the wall effects detected during the geophysical study and thus to provide a better knowledge of the forge building. The continuity of the latter can be observed towards the north-east where a second space with a concentration of magnetic elements—which was not detected during the geophysical prospection—seems to exist (Hulin et al., 2014).

This approach is now well implemented in preventive archaeology in the Inrap where it corresponds to 30% of studies carried out over during the last 5 years. The widespread use of this type of approach over the last 10 years has considerably renewed our knowledge of iron working areas, about large workshops and even more on smaller ones which have left particularly tenuous traces that are particularly difficult to observe visually.

5.4 Prospecting Prehistoric Environment Undercover

Caves and rock shelters are natural environments that people have used since prehistoric times, but also in more recent periods. Regarding the consequent number of open-air occupations, these sites are rare and are the object of reinforced conservation measures. This is particularly true in the presence of rock art. Therefore, the study of these sites by non-invasive methods is of great interest. However, caves are particularly difficult environments to explore. The small dimensions of most of these places linked to the proximity of the walls and the ceiling constitute a strong constraint that does not exist in open environments. It often results in difficulties of access and transport of equipment, but also in the impossibility of implementing the protocols of prospection routinely used in open areas.

Over the last decade, two teams led by French geophysicists have attempted to meet this challenge to study combustion paleo-structures. Among the rare archaeological structures testifying human presence in these caves, hearths are preferential targets, especially for specialists in magnetometer prospecting. However, the topography of the studied areas and the presence of archaeological remains on the ground forced the teams to adapt their protocol, mainly for the magnetometer surveys, as standing or walking on the ground with the sensors might not be possible or could damage archaeological evidence. One possible choice is to not use the “continuous” mode of the magnetometer but to prefer discrete measurements (Jrad et al., 2013). However, it is possible to fully exploit the sampling capability of current magnetometers. A specific protocol was developed to allow the sensors to be moved without walking on the archaeological soils while “continuously” recording the magnetic field and the position of the sensors in space. A system consisting of a motorised total station and a boom attached to a tripod was developed by F. Lévêque and used for the first time in the Fraux cave (Burens et al., 2014, 2019). This protocol was then implemented in several French caves and rock shelters such as Cussac, Chauvet, Castanet (unpublished studies) and Bruniquel (Jaubert et al., 2016). Another issue related to rock shelters and caves is the assessment of the volume of potentially anthropogenic sediment overlying the substrate. Estimating the volume and distribution of these deposits is an essential element for programming archaeological surveys. Relevant information can be provided by coupling dynamic penetrometer measurements with apparent electrical resistivity surveys (Martinaud et al., 1999).

The complementarity of these methods was exploited at “La Piscine Magdalenian” (Fig. 10). This deposit is located at the foot of a cliff made of Bajocian limestone with several cavities. An area of 180 m² without major topographic anomalies was prospected in pole-pole configuration for three mobile electrode spacings (0.5, 1 and 1.5 m). Except for a few modern disturbances easily identified at the soil surface, low resistivity values are attributed to a deep substratum cover by silty-clay sediments, and conversely, high values correspond to a shallow substratum. The depth of the substrate was estimated using a VES inversion software. For this, we considered the near subsoil consisting of only two layers: the overburden, with a resistivity between 60 and 75 Ω.m, and the limestone substratum, with resistivity 230 Ω.m. The result is very robust as more than 95% of the 252 points considered shows an error of less than 7%. The largest errors, of the order of 15%, are at the location of previously identified modern disturbances. Dynamic penetrometer measurements performed at five locations confirmed the satisfying quality of the estimate of the cover thickness. The depth reached by the tip of the penetrometer is in all cases close to the estimation (less than 10% error) which seems to validate the approach. Near the cliff, the rocky surface appears very close to the surface. It forms a platform with a gentle slope towards the south, limited on either side by two depressions more than 2 m deep. Based on these results, the volume of sediment cover can be estimated between 170 and 200 m³. 

Fig. 10

La Piscine, Montmorillon, France, 2010–2011. Apparent electrical resistivity map (a = 1 m, 1 m² grid). Level lines equidistant of 0.25 m indicate the estimated thickness of the cover. (Data acquired in collaboration with C. Delage and M. Druez)

5.5 Coastal Prospection

Foreshore sites, located between the extreme limit of the highest and lowest seas, are particularly difficult to study. However, most of them are in danger due to climate change, erosion, and coastal development. Tides, storms and all the phenomena causing rapid displacement of large volumes of sediment on the coast, make this intertidal space very difficult to excavate. It is therefore necessary to have alternative or complementary methods to study the archaeological sites of the foreshore. Since the end of the 1990s, geophysical prospecting methods have been implemented, on an experimental basis, on the French West Atlantic coast (Laporte et al., 2009). Trials have been multiplied in recent years, in rocky, sandy, muddy, and mixed contexts (Mathé et al., 2021).

Due to the high salinity of the environment, electrical resistivity techniques deliver small signals and GPR has been excluded. Magnetometer and Slingram-EMI techniques were chosen (Mathé et al., 2018). As the possibilities of leaving markers on the ground and the intervention slots are limited due to the tide, GNSS positioning is often preferred to locate the measurements. On the other hand, it is more difficult, if not impossible, to avoid certain specificities of intertidal zones. Micro-topography is a common source of magnetic anomalies due to the non-constancy of the ground-sensor distance; it is also the source of disturbances in the electrical conductivity signal created by “puddles” of sea water. Another important limitation to the implementation of geophysics on the foreshore is the quasi-systematic presence of metallic waste in variable quantities. When all these noise sources are under control, the survey results are of great quality. One of the most significant recent results in such an environment was acquired in 2018 on Oleron Island. The magnetometer mapping revealed the presence of three curvilinear ditches of a Neolithic enclosure on the rocky foreshore (Fig. 11). The anomalies appear clearly near the dam, then gradually disappear towards the open sea where they have probably been completely eroded.

Fig. 11

Map of total magnetic field anomalies of Ors, Le-Château-d’Oléron, France. (Data acquired in collaboration in 2018–2019 with L. Soler and G. Bruniaux)

6 Conclusion

In this chapter, we have tried to make an overview of the French situation in the field of archaeo-geophysics. We have identified some trends in the type of surveys we are bound to. First, it seems that the rural context is still dominant in terms of numbers of operations done during the last two decades in France, even if other types of contexts seem to increase yearly. Second, the amount of surface prospected which were dominated by the ‘intermediate’ scale in the beginning of the 2000s are now quite balanced between the ‘large’ (above 20,000 m²), ‘intermediate’ (between 3550 m² and 20,000 m²) and ‘small’ (below 3550 m²) categories. This reflects well the broadening of archaeological questions asked to the geophysicist. Preventive archaeology represents a third of the whole surveys and seems to be stable over the last years. Even if the Inrap’s team contribution to the database is around one third of the studies, as their activity is balanced between both preventive and research archaeology, there is no clear explanations of that trend which arise from our data.

With the case studies, we tried to show the diversity of experience that could happen in archaeo-geophysics in France. Some examples used or cited here are the first attempts to widen the panel of techniques at the disposal of archaeologist (geomorphology, pedology, micromorphology, cone penetrometer tests etc.) but these attempts are far from being standard practices nowadays. It has to be noticed that Inrap and some other preventive archaeology companies (Éveha, ArkeMine) are developing their own group of specialists in geophysics. They have begun to be places where methodological innovations and combinations could rise under a unified framework.

The geochemistry is still under development. The XRF sensors are now available for the field, but the mapping is harder than in geophysical techniques (moreover the analysis is very time consuming after field work in the case of chemical extraction on samples). This is certainly a new way to be developed to interact with the archaeologist even if it requires an important referential work.

With the very specific policy about preventive archaeology, France has the possibility to offer great feedback to compare geophysical results and trial trenching/excavation results. It is really a singular opportunity, which provides us the possibility to greatly improve our interpretation experience and methodologies though it is still underdeveloped.

Finally, the interaction with all disciplines under the scope of the SAGA group are mainly driven by the archaeological questions. It means that the archaeology training programs should involve more and more courses on geosciences (and maybe the other way round). Another point is that some of us are not specialised on specific chrono-cultural targets, but mainly in hydrological, pedological or agronomical targets. That is another point of the applied geophysics in the French context, most of geophysical studies are not performed by archaeo-geophysicists but mostly by geophysicists working with archaeologists.