Uncovering a hidden neolithic landscape. Locating neolithic monumental sites through remote sensing, geophysics, and archaeology.
The role of central gathering places in Neolithic societies has been stressed in much of recent archaeology. In this project, we wish to add to this discussion by analysing a particular region, Falbygden in western Sweden. The area has unique advantages for such a study, due to good preservation, highly structured landscape, and well researched Neolithic archaeology.
The central places are sites enclosed by ditch systems and palisades, thought to have played a major role in the reproduction of Neolithic social systems. Such sites are known from South Scandinavia and much of central and western Europe but have so far not been identified in western Sweden.
We wish to locate Neolithic central sites in the area and analyse their relation to other elements of the landscape such as settlements and passage graves, as well as communication routes and natural features. This will add significantly to our understanding of Neolithic landscape and society.
To do this, we wish to employ a combination of remote sensing, geophysical prospection, GIS modelling and archaeological fieldwork. Other types of monuments, such as long barrows and palisade enclosures may also be found. Promising sites identified by geophysical prospection, remote sensing and GIS modelling will be evaluated in the field by archaeological test excavations.
Final report
Purpose and development of the project
The purpose of this project was to contribute to the understanding of Neolithic landscape structure in Falbygden by locating enclosures, so-called central places, and viewing these in relation to other natural and human-made features. In this, we wished to employ a combination of remote sensing, geophysical prospection, GIS modelling and archaeological fieldwork.
The role of enclosures, also called Sarup sites or causewayed camps, in Neolithic societies has been stressed in much of recent archaeology. These are sites enclosed by ditch systems and palisades, thought to have played a major role in the reproduction of Neolithic social systems. Enclosures are known from South Scandinavia and much of central and Western Europe but not yet in western Sweden.
Finding Neolithic enclosures through field surveys is not feasible. Instead, we proposed to use a combination of methods, including drone, aerial and satellite images, geophysical prospection, and predictive modelling in order to find promising sites. The size of these sites (3-20 ha) and the substantial ditch systems makes them suitable for such methods.These would then have to be substantiated in the field by archaeological investigations.
A second approach used predictive GIS modelling. This method departs from properties of known locations to predict new locations. Such studies were performed for south Scandinavia and northern Germany with good results. Although the landscape in Falbygden is different, some of the variables pointed out in these studies may be applicable, such as locations on prominent positions, proximity to streams or wetlands etc. Many of these aspects are known or can be modelled with the high resolution data now available. The predicted locations were then targeted for prospection.
The project developed largely as outlined in the project description, although we decided to put more effort into geophysical and drone prospection and less into interpretation of satellite images than was originally planned. In the final phase, we also got access to a Lidar drone, which was used for detailed terrain mapping of a few sites.
Implementation
The project can be divided into two main phases. In the first phase, prospection aimed at covering a large and as far as possible contiguous area, including different landscape niches and with differing relations to known neolithic monuments. The aim was to cover all different parts of the landscape, without consideration to previous knowledge about the location of neolithic enclosures. The chosen test area was an area from the bottom of the Åsle valley, through the village of Karleby with its monumental passage graves, and westwards over the limestone plateau towards Falköping town.
In a second phase, predictive modelling of probable enclosure locations was made, based on our knowledge of Danish and north German enclosure sites from the TRB period. The predicted areas were then targeted for geophysical and drone survey.
Acquisition of satellite images was limited to data available at low or no cost, through resources such as EOS and Rymdstyrelsen. We also downloaded Swedish modern and historical airphotos, DEM data, Lidar data, and soil maps from Lantmäteriverket and SGU.
A dedicated multispectral drone (DJI P4 multispectral) was purchased at the start of the project, and was used extensively for surveying during the growth and maturing season of crops. In all, 5 sq km were surveyed by drone, but some fields were surveyed repeatedly in various growing conditions. Later, limited areas were prospected by Lidar with drones from the DAI and from an in-house drone in Gothenburg.
Magnetometry surveys were performed by 1-2 teams yearly from the Deutsche Archäologische Institut, Frankfurt. These surveys were limited by accessibility and were mainly performed after harvest, in September-November. A total of 8 sq km were surveyed, mostly in a continuous area in central Falbygden, but also in some other areas chosen for topographic reasons.
GPR surveys were made in limited areas, based on indications from magnetometric and/or drone prospection.
Image data was analysed in dedicated software such as Erdas Imagine, ArcGis Pro, and Saga Gis. ArgGis Pro and Saga Gis were used for terrain analysis of DEM data and for predictive modelling of probable enclosure locations, using variables such as topographic position, distance to megalithic tombs, distance to wetland and open water, distance to historic roads, etc.
A selection of five possible sites in Karleby were tested archaeologically by machine trenches and archaeological excavation. These could all be discarded, cf below. A further possible site in Friggeråker remains to be tested in 2024.
Results and discussion
We have carried out large-scale geophysical and remote sensing prospection of strategic parts of the Falbygden landscape, covering 8 sq km and different landscape niches. The most convincing indications of neolithic enclosures or other neolithic monuments were tested by GPR survey and archaeological excavation. In spite of this large-scale prospection, none of the suspected enclosures could be substantiated. In most cases, the tested indications could instead be explained as late glacial ice-lake shoreline formations or by glaciofluvial tectonic processes. In one case, a cropmark resembling a long barrow or long dolmen turned out to be historic in origin.
The main result of this project is therefore negative. We can at present not confirm the presence of neolithic enclosures in Falbygden, and in view of the large area that has been surveyed, we believe that they actually do not exist.
This conclusion must be qualified, however. First of all, the interpretation of magnetometric data turned out to be challenging due to the noisy character of the data. This problem arises from the presence of magnetic stone material in the moraines in the area, emitting strong magnetic signals that tend to mask the finer nuances related to archaeological features. Various filter techniques were tested to deal with this problem, but this remains a largely unsolved issue. It is therefore theoretically possible that archaeological features could be found in our data by applying better analysis methods.
In spite of these difficulties, we were able to distinguish large numbers of non-natural features in our data, such as old field boundaries, roads, drainage systems etc. Most of these were late historical features. Certain natural features were also clearly visible, such as water courses or bedrock fissures. This demonstrates the feasibility of the prospection methods in this landscape, although certain features may have been missed or overlooked.
Secondly, in the final phase of the project we identified a previously unrecognised type of enclosures in the area, consisting of small hilltops enclosed by earthen ramps and possibly ditches. These were visible in Lidar and DEM data, partly also in airphotos and drone imagery. These sites do not match the south Scandinavian Neolithic enclosures, but they also are not typical Iron Age defensive sites. Four such sites have so far been recognised. No archaeological investigations have yet been made at these sites, and their date and function remains unknown.
New research questions
Three main questions emerge from these results, one methodical, one empirical and one interpretative. The methodical question concerns how to develop tools to deal with and filter out the strong noise in Scandinavian geophysical data, arising from moraine containing partly magnetic rocks. The empirical question regards the dating and function of the newly discovered group of walled enclosures on small hilltops, none of which has been excavated yet. On the interpretative level, the apparent lack of enclosures of Sarup type leads to questions about the organisation of the Neolithic societies on Falbygden and why they differ from south Scandinavian TRB societies.
Dissemination and collaboration
We have presented and discussed our project at seminars in Gothenburg, Stockholm and Oslo universities, and at the Megalithic studies group conference in Santiago de Compostela 2023. Through cooperation with Heather Reese, Department of Earth Sciences, GU, we were given access to software (Erdas Imagine) for image analysis, and to high resolution lidar data for parts of our area. The geological interpretations of our excavation data have been discussed in the field with Torbjörn Persson, Falköping. In cooperation with Stephen Davis, University of Dublin, Tony Axelsson participated in georadar surveys in the area of Newgrange, Ireland, in the context of a similar large-scale survey project as the one in Falbygden. A common paper is forthcoming, comparing the experiences from the two projects (Davis et al in prep). Project details will be reported in Axelsson & Sjögren (in prep).
The purpose of this project was to contribute to the understanding of Neolithic landscape structure in Falbygden by locating enclosures, so-called central places, and viewing these in relation to other natural and human-made features. In this, we wished to employ a combination of remote sensing, geophysical prospection, GIS modelling and archaeological fieldwork.
The role of enclosures, also called Sarup sites or causewayed camps, in Neolithic societies has been stressed in much of recent archaeology. These are sites enclosed by ditch systems and palisades, thought to have played a major role in the reproduction of Neolithic social systems. Enclosures are known from South Scandinavia and much of central and Western Europe but not yet in western Sweden.
Finding Neolithic enclosures through field surveys is not feasible. Instead, we proposed to use a combination of methods, including drone, aerial and satellite images, geophysical prospection, and predictive modelling in order to find promising sites. The size of these sites (3-20 ha) and the substantial ditch systems makes them suitable for such methods.These would then have to be substantiated in the field by archaeological investigations.
A second approach used predictive GIS modelling. This method departs from properties of known locations to predict new locations. Such studies were performed for south Scandinavia and northern Germany with good results. Although the landscape in Falbygden is different, some of the variables pointed out in these studies may be applicable, such as locations on prominent positions, proximity to streams or wetlands etc. Many of these aspects are known or can be modelled with the high resolution data now available. The predicted locations were then targeted for prospection.
The project developed largely as outlined in the project description, although we decided to put more effort into geophysical and drone prospection and less into interpretation of satellite images than was originally planned. In the final phase, we also got access to a Lidar drone, which was used for detailed terrain mapping of a few sites.
Implementation
The project can be divided into two main phases. In the first phase, prospection aimed at covering a large and as far as possible contiguous area, including different landscape niches and with differing relations to known neolithic monuments. The aim was to cover all different parts of the landscape, without consideration to previous knowledge about the location of neolithic enclosures. The chosen test area was an area from the bottom of the Åsle valley, through the village of Karleby with its monumental passage graves, and westwards over the limestone plateau towards Falköping town.
In a second phase, predictive modelling of probable enclosure locations was made, based on our knowledge of Danish and north German enclosure sites from the TRB period. The predicted areas were then targeted for geophysical and drone survey.
Acquisition of satellite images was limited to data available at low or no cost, through resources such as EOS and Rymdstyrelsen. We also downloaded Swedish modern and historical airphotos, DEM data, Lidar data, and soil maps from Lantmäteriverket and SGU.
A dedicated multispectral drone (DJI P4 multispectral) was purchased at the start of the project, and was used extensively for surveying during the growth and maturing season of crops. In all, 5 sq km were surveyed by drone, but some fields were surveyed repeatedly in various growing conditions. Later, limited areas were prospected by Lidar with drones from the DAI and from an in-house drone in Gothenburg.
Magnetometry surveys were performed by 1-2 teams yearly from the Deutsche Archäologische Institut, Frankfurt. These surveys were limited by accessibility and were mainly performed after harvest, in September-November. A total of 8 sq km were surveyed, mostly in a continuous area in central Falbygden, but also in some other areas chosen for topographic reasons.
GPR surveys were made in limited areas, based on indications from magnetometric and/or drone prospection.
Image data was analysed in dedicated software such as Erdas Imagine, ArcGis Pro, and Saga Gis. ArgGis Pro and Saga Gis were used for terrain analysis of DEM data and for predictive modelling of probable enclosure locations, using variables such as topographic position, distance to megalithic tombs, distance to wetland and open water, distance to historic roads, etc.
A selection of five possible sites in Karleby were tested archaeologically by machine trenches and archaeological excavation. These could all be discarded, cf below. A further possible site in Friggeråker remains to be tested in 2024.
Results and discussion
We have carried out large-scale geophysical and remote sensing prospection of strategic parts of the Falbygden landscape, covering 8 sq km and different landscape niches. The most convincing indications of neolithic enclosures or other neolithic monuments were tested by GPR survey and archaeological excavation. In spite of this large-scale prospection, none of the suspected enclosures could be substantiated. In most cases, the tested indications could instead be explained as late glacial ice-lake shoreline formations or by glaciofluvial tectonic processes. In one case, a cropmark resembling a long barrow or long dolmen turned out to be historic in origin.
The main result of this project is therefore negative. We can at present not confirm the presence of neolithic enclosures in Falbygden, and in view of the large area that has been surveyed, we believe that they actually do not exist.
This conclusion must be qualified, however. First of all, the interpretation of magnetometric data turned out to be challenging due to the noisy character of the data. This problem arises from the presence of magnetic stone material in the moraines in the area, emitting strong magnetic signals that tend to mask the finer nuances related to archaeological features. Various filter techniques were tested to deal with this problem, but this remains a largely unsolved issue. It is therefore theoretically possible that archaeological features could be found in our data by applying better analysis methods.
In spite of these difficulties, we were able to distinguish large numbers of non-natural features in our data, such as old field boundaries, roads, drainage systems etc. Most of these were late historical features. Certain natural features were also clearly visible, such as water courses or bedrock fissures. This demonstrates the feasibility of the prospection methods in this landscape, although certain features may have been missed or overlooked.
Secondly, in the final phase of the project we identified a previously unrecognised type of enclosures in the area, consisting of small hilltops enclosed by earthen ramps and possibly ditches. These were visible in Lidar and DEM data, partly also in airphotos and drone imagery. These sites do not match the south Scandinavian Neolithic enclosures, but they also are not typical Iron Age defensive sites. Four such sites have so far been recognised. No archaeological investigations have yet been made at these sites, and their date and function remains unknown.
New research questions
Three main questions emerge from these results, one methodical, one empirical and one interpretative. The methodical question concerns how to develop tools to deal with and filter out the strong noise in Scandinavian geophysical data, arising from moraine containing partly magnetic rocks. The empirical question regards the dating and function of the newly discovered group of walled enclosures on small hilltops, none of which has been excavated yet. On the interpretative level, the apparent lack of enclosures of Sarup type leads to questions about the organisation of the Neolithic societies on Falbygden and why they differ from south Scandinavian TRB societies.
Dissemination and collaboration
We have presented and discussed our project at seminars in Gothenburg, Stockholm and Oslo universities, and at the Megalithic studies group conference in Santiago de Compostela 2023. Through cooperation with Heather Reese, Department of Earth Sciences, GU, we were given access to software (Erdas Imagine) for image analysis, and to high resolution lidar data for parts of our area. The geological interpretations of our excavation data have been discussed in the field with Torbjörn Persson, Falköping. In cooperation with Stephen Davis, University of Dublin, Tony Axelsson participated in georadar surveys in the area of Newgrange, Ireland, in the context of a similar large-scale survey project as the one in Falbygden. A common paper is forthcoming, comparing the experiences from the two projects (Davis et al in prep). Project details will be reported in Axelsson & Sjögren (in prep).