Remnant echoes of the past Andreas Viberg Archaeological geophysical prospection in Sweden

Remnant echoes of the past Andreas Viberg  Archaeological geophysical prospection in Sweden

Remnant echoes of the past

Archaeological geophysical prospection in Sweden

Andreas Viberg

Doctoral Thesis in Archaeological Science 2012

Stockholm University

Abstract

Viberg, A. 2012. Remnant echoes of the past. Archaeological geophysical prospection in

Sweden. Theses and Papers in Scientific Archaeology 13.

The aim of this thesis has been to investigate the benefits, pitfalls and possibilities of using geophysical methods in archaeological projects. This is exemplified by surveys carried out at archaeological sites in different geographical and chronological contexts. The thesis also aims at investigating the cause for the under-use of the methods in Swedish archaeology by looking at previously conducted surveys. The methods used during these surveys have been Groundpenetrating radar (GPR), magnetometer, slingram and a kappameter. The surveys in the mountain tundra region of Lapland show that magnetic susceptibility surveys is a valuable aid in discovering heaps of fire-cracked stones and when combined with magnetometry, also hearths. GPR and magnetometer surveys within the Migration Period ringfort Sandbyborg provided the spatial layout of the fort and indicated, along with results from recent excavations and metal detections, many similarities with the ringfort Eketorp II. The nonmagnetic character of the sedimentary bedrock on Öland and Gotland is suitable for magnetometer surveys and the method is also highly appropriate for the detection of the remains of high-temperature crafts. GPR surveys at St. Mary’s Dominican convent in Sigtuna produced the spatial layout of the central cloister area. The investigations also show that the geology, pedology, land use and the character of commonly occurring prehistoric remains in

Sweden, in certain circumstances and in certain areas, have restricted the possibility of successfully carrying out geophysical surveys. Care must therefore be taken to choose the right instrument for the survey and to tailor the sampling density of each geophysical survey, according to the character and size of the expected archaeological remains, in order to maximize their information return. To increase the use of geophysical methods in Sweden the educational opportunities, both for surveyors and professional archaeologists, need to improve.

Keywords

: Archaeological prospection, geophysical survey, Sweden, Ground-penetrating radar, magnetometry, Slingram, magnetic susceptibility, Neolithic, Migration period, Viking

Age, Middle Ages, Öland, Gotland, Sigtuna, Lapland

.

© Andreas Viberg, Stockholm 2012

ISSN 1400-7835

ISBN 978-91-7447-549-4

Printed in Sweden by Universitetsservice US-AB, Stockholm 2012

Distributor: Archaeological Research Laboratory, Stockholm University

Cover image: The text reads: “Commemoratio omnium fratrum familiarium et

benefactorum defunctorum ordinis nostri et alibi aliorum plurimorum sanctorum

martyrum et confessorum atque sanctarum virginum”. The text is written on a wooden writing desk found at St. Mary’s Dominican Convent in Sigtuna. The gothic window frame is published with permission from The Garden Factory and the use of the font, Northumbria, by permission of David Kerkhoff. Cover image photos:

Kerstin Lidén & Siska Williams, Layout by the author.

To Helena and Elsa the lights of my life

Acknowledgements

It is almost four years and eight months since I started off as a doctoral student at the Archaeological Research Laboratory (ARL) and family, friends and colleagues have been involved in the process and as such deserve my deepest gratitude.

First I would like to thank my wife Helena and my daughter Elsa who have put up with me during the intense process of writing the thesis. I am sorry for the late nights, the weekends and vacations when I had to work to get it finished. Thank you for encouraging me to do my best, even when it was tough. This thesis is dedicated to you!

Thanks to my supervisor Kerstin Lidén who has been a solid support during these past years. Thank you for keeping me calm and motivated during times when the process of writing a thesis felt somewhat overwhelming. Thank for your inspiration and contribution to the many interesting projects we have been involved in together. We started off in the mountain tundra region in Lapland and finally ended up on Öland. It has been a fantastic time.

Next in line are my geophysical supervisors. Thanks to Immo Trinks for introducing me to the world of archaeological geophysics and for your supervision during my first year as a PhD-student. I am especially grateful to

Bo Olofsson, Rinita Dalan and Bruce Bevan who came to my aid when I needed it the most, and who have been very patient and generous with their time. This thesis would not have become a reality without your help! I would also like to thank Kjell Persson for many interesting discussions on archaeological geophysics and for stepping up and helping me with instrument and computer issues whenever needed. I would also like to thank you for your supervision during my time as an undergraduate student. It all started with a GPR survey on a cold day in Fresta, Upplands Väsby, and look where I ended up.

To my co-authors in the different articles presented in this thesis. It has been inspiring to work with you and to share your passion for the different projects and for archaeology. A special thanks to Anders Wikström who was involved in my application to the ARL and who has been providing me with suggestions to many suitable targets for geophysical prospections in Sigtuna.

It has been a privilege working with you.

I would especially like to thank Elin Fornander who has been like my

PhD mentor over the years. You were there when I arrived at the ARL and it was inspiring to see you finish your thesis last year. Thank you for

everything and don’t forget to “frysa könsögonblicket” as often as you can, just like our lecturer in university pedagogy so successfully did on many occasions.

When I started as a PhD-student I got to share room with Björn

Hjulström. I really appreciate that you took the time to answer all my rookie questions even though you were busy writing your own thesis. Remember that it is never too late to pick up the tuba again. Thanks to Anna

Linderholm, who among other things helped me collect soil samples at

Stensborg. That day was hard work for us all but ended well at the golf course restaurant. Heja Djurgården!

I have also tormented several other fellow PhD-students at the ARL to help me during my geophysical measurements. Thank you Ny Björn

Gustafsson (and Erika of course) who joined in on several surveys in miserable conditions. I especially remember the gale force winds on Öland, the GPR survey in 0.4m of snow at Fornsigtuna and that remarkably wet day on Gotland. Thank God for pear flavoured whipped cream! Thanks also to

Joakim Schultzén for helping me during the magnetometer surveys in

Hedeby and on Öland. A special thanks for taking me to Ö&B to gear up when I accidentally turned my non-magnetic rain trousers into a miniskirt while getting into your car. Thanks to my other fellow PhD-students at the

ARL: Anita Malmius, Christos Economou, Hans Ahlgren, Rachel Howcroft and Sylvia Sandelin Löfgren. Thanks also to Christina Karlsson for providing me with caffeine patches and for sharing my passion for coffee.

Snoozers are losers!

I have also had the privilege of spending a lot of time with past and present researchers and staff at the ARL. Thanks to Anders Götherström,

Birgit Arrhenius, Charlotte Hedenstierna-Jonson, Christina Schierman,

Gunilla Eriksson, Gustav Trotzig, Ioannis Panagopoulos, Laila Kitzler

Åhfeldt, Lena Holmqvist, Liselotte Bergström, Maria Wojnar-Johansson,

Margaretha Klockhoff, Mikael Lundin, Sven Isaksson and Yvonne Fors for your support that contributed to making my time as a PhD-student so interesting and rewarding. I would also like to extend a big thank you to

Ludvig Papmehl-Dufay who was very generous with his time and help during my surveys of Sandbyborg on Öland. How about surveying another ringfort?

Many others have been involved in my different projects or been helping me in one way or the other and I would like to thank Alain Tabbagh, Alois

Hinterleitner, Anders Angerbjörn, Anders Hedman, Andreas Forsgren, Anna

Kjellström, Bernth Johansson, Chris Gaffney, Christer Dahlström, Dario

Cianciarulo, Dattatray Parasnis, Dean Goodman, Elin Engström, Göran Alm,

Göran Burenhult, Henrik Alvarmo, Ivonne Dutra Leivas, Jaana Gustafsson,

Jan Storå, Joakim Kjellberg, Lars Larsson, Linda Qviström, Neil Linford,

Robin Blomdin, Svante Johansson, Tessa Evans, Timo Ibsen, Tom and

Charlotte Slettengren and Ulf Näsman. Thanks also to the participants of

Arctic Sweden, the Swedish Polar Research Secretariat, everyone

contributing to the material presented in the review article and to every researcher and staff at the Department of Archaeology and Classical Studies at Stockholm University. I would also like to thank everyone at the department who provided me with valuable comments that greatly improved the text. I hope you had “FUN” reading it.

The language in this thesis has been revised by Malcolm Hicks and I am extremely grateful for your solid work and for the incredible speed with which you corrected my texts.

This thesis would not have been possible to finish without the financial support of Albert & Maria Bergström Stiftelse, Brigit and Gad Rausings

Stiftelse för Humanistisk Forskning, Helge Ax:son Johnsons Stiftelse and

Svenska Fornminnesföreningen.

My sincere gratitude also extends to everyone in the International Society for Archaeological Prospection and especially: Adrian Butler, Alette

Kattenberg, Andrea Zeeb-Lanz, Arne Anderson Stamnes, Carmen Cuenca

Garcia, Christophe Benech, Ervan G Garrison, George Maloof, Guillaume

Hulin, Ian Moffat, James Bonsall, Jason Jeandron, Jarrod Burkes, Joep

Orbons, Johannes Frenzel, Jörg Fassbinder, Kayt Armstrong, Kelsey Lowe,

Kevin Barton, Norbert Buthmann, Piotr Szczepanik, Piotr Wroniecki, Roger

Ainslie, Steve de Vore, Tomasz Herbich, Walter Sevenants and Wesa

Perttola, who offered valuable insights into the use of archaeological prospection methods in other countries. A special thanks to Lewis Somers for interesting discussions on the early days of archaeological geophysics in

England.

To my dear friends Antje Wendt and Sven Kalmring. Thank you for your valuable comments on my thesis, for being there and encouraging me during the final stages of the writing process and for putting up with my feeble attempts of learning German.

Thanks to Mats, Maria, Anna-Carin, Nils, Ingalill, Per and Monica for your support throughout this journey. A special thanks to my mother and father, Barbro and Wincent, who has cheered for me during my years at the

University and who has been helping out in various ways so that I could spend the required time carrying out surveys and writing the thesis. Dignity, always dignity!

All that remains to be said is:

Ahoi Landratten! See you next year in Tuscany!

List of papers

I.

II.

Viberg, A., Trinks, I. & Lidén, K. 2011. A Review of the

Use of Geophysical Archaeological Prospection in Sweden.

Archaeological Prospection 18(1): 43-56.

Viberg, A., Berntsson, A. & Lidén, K. Archaeological

Prospection of a High Altitude Neolithic Site in the Arctic

Mountain Tundra Region of Northern Sweden. Manuscript submitted to Journal of Archaeological Science.

III. Viberg, A., Victor, H., Fischer, S., Lidén, K. & Andrén, A.

A Room with a View. Archaeological Geophysical

Prospection and Excavations at Sandby ringfort, Öland,

Sweden. Manuscript.

IV. Gustafsson, N-B. & Viberg, A. 2012. Tracing Hightemperature Crafts. Magnetometry on the island of Gotland,

Sweden. Archaeological Prospection, 19(3): 201-208.

V. Viberg, A. & A. Wikström, 2011. St. Mary's Dominican

Convent in Sigtuna Revisited. Geophysical and archaeological investigations. Fornvännen 106: 322-33.

Contents

1. Introduction, aim and structure of thesis ................................................11

2. Geophysical prospection methods .........................................................16

2.1. Magnetic methods ................................................................................ 16

2.1.1. The Earth’s magnetic field ............................................................ 17

2.1.2. Magnetism, basic concepts ........................................................... 19

2.1.3. Magnetic susceptibility and archaeological prospection ................. 20

2.1.4. Thermoremanent magnetism and archaeological prospection ........ 22

2.1.5. Passive instruments for magnetic prospection ............................... 22

2.2. Electromagnetic methods ...................................................................... 26

2.2.1. Ground penetrating radar, an introduction ..................................... 26

2.2.2. GPR and soil properties ................................................................ 29

2.2.3. GPR and topography .................................................................... 34

2.2.4. GPR manufacturers and suitable archaeological targets ................. 34

2.2.5. The metal detector ........................................................................ 36

2.2.6. Magnetic susceptibility and the Bartington MS2D......................... 36

2.2.7. Conductivity and susceptibility measurements using the Geonics

EM38 Slingram instrument ........................................................... 38

2.3. Geoelectric methods and other geophysical prospection methods .......... 40

2.4. Marine geophysical methods and geochemical prospection methods ...... 42

2.5. The benefits of using an integrated approach ......................................... 43

3. Development and current use of archaeological geophysical prospection methods ............................................................................. 45

3.1. A brief history of archaeological geophysics ......................................... 45

3.1.1. Geoelectric methods ..................................................................... 45

3.1.2. Magnetometry .............................................................................. 46

3.1.3. Electromagnetic methods.............................................................. 47

3.2. Use of archaeological geophysical prospection methods in other countries .............................................................................................. 48

3.2.1. Austria ......................................................................................... 48

3.2.2. Denmark ...................................................................................... 48

3.2.3. England........................................................................................ 49

3.2.4. Finland ......................................................................................... 49

3.2.5. Germany ...................................................................................... 50

3.2.6. Iceland ......................................................................................... 50

3.2.7. The Netherlands and Flanders ....................................................... 50

3.2.8. Norway ........................................................................................ 51

3.2.9. Poland .......................................................................................... 51

3.2.10. Scotland ....................................................................................... 51

3.2.11. Turkey ......................................................................................... 52

3.2.12. North and South America ............................................................. 52

3.2.13. Australia ...................................................................................... 52

3.3. Archaeological geophysics in Sweden................................................... 53

3.4. Concluding remarks ............................................................................. 60

4. Geological and pedological conditions in Sweden and their implication for archaeological geophysics ............................................. 61

4.1. Geology, pedology and land use ........................................................... 61

4.2. Implications for archaeological geophysics ........................................... 67

5. Data collection, interpretation and the character of Swedish archaeological remains ......................................................................... 69

5.1. Sampling density, data collection and georeferencing survey grids ........ 69

5.2. Interpretation ........................................................................................ 76

5.3. Swedish archaeological remains ........................................................... 77

5.3.1. Graves ......................................................................................... 77

5.3.2. Prehistoric settlement remains ...................................................... 79

5.3.3. Fortified sites ............................................................................... 80

5.3.4. Fired structures............................................................................. 81

5.3.5. Medieval buildings ....................................................................... 81

5.3.6. Gardens........................................................................................ 82

5.3.7. Detection of archaeological remains in present-day urban environments................................................................................ 82

5.3.8. Ferrous objects ............................................................................. 82

5.3.9. Other non-archaeological features ................................................. 82

6. Discussion ............................................................................................83

6.1. The Swedish situation and beyond ........................................................ 89

6.2. Brief guidelines for archaeologists ........................................................ 90

7. Conclusions and future prospects for archaeological geophysics in

Sweden ................................................................................................ 93

References............................................................................................................95

1. Introduction, aim and structure of thesis

For a very long time Sweden was in the forefront of the development and use of geophysical prospection. One of the earliest documented applications dates from the 1640s, when a simple declination compass was used to locate iron ore in Närke (Carlborg 1963:13). Since then, Swedish researchers have been involved in developing new methods and instruments, such as magnetometers (Thalén 1879; Tiberg 1884), electrical resistivity instruments

(e.g. Petersson 1907; Bergström 1913; Lundberg, 1919, 1922), electromagnetic instrumentation (Sundberg et al. 1923) and the electromagnetic instrument Slingram (Werner 1937). Metal detectors were already being used for mine detection during the Second World War and groundpenetrating radar (GPR) investigations and theoretical studies were carried out during the 1970s (Nilsson, 1973, 1976, 1978; Ulriksen 1982).

Geophysical surveys have gradually been adopted by researchers in many other countries with an interest in locating buried archaeological remains, providing important information on subsurface structures in a wide variety of archaeological environments from 1938 onwards (see, for example,

Aitken 1958; Atkinson 1953; Bevan 2000). Thus the subdiscipline of

“archaeological geophysics”, defined as “the examination of the Earth’s physical properties using non-invasive ground survey techniques to reveal buried archaeological features, sites and landscapes” (Gaffney & Gater

2003:12), was born.

Even though geophysical prospection has been successfully used in

Sweden for a long period of time and the country is the home of three geophysical equipment manufacturers, ABEM, Malå Geoscience and

Radarteam Sweden AB, this success has failed to be translated into archaeology. The methods were not used in any Swedish archaeological projects until the mid-1970s (Fridh 1982) with the exception of some early metal detector investigations in 1959 and the early 1960s (Hagberg 1961;

Nylén 1972). Despite the fact that geophysical methods are non-destructive, can pinpoint specific areas of interest within larger sites (and thus permit targeted excavations) and can provide complementary information about archaeological structures outside the areas directly affected by excavations, they have only been used sporadically in archaeological projects. This can partly be explained by the fact that many archaeologists who have used geophysical prospection have been sceptical about its benefits and reliability, as there has in many cases been a lack of correlation between the geophysical and archaeological results (see, for example, Carelli 2003;

11

Hylén 2007; Eklund 2008; Steineke 2008; Bennström & Helgesson 2010;

Lindberg 2010).

Archaeological geophysics has now become an integrated part of archaeological projects in many other countries, and its limited use in

Sweden is interesting. I have therefore, in parallel with my geophysical surveys, been engaged in examining the possible reasons for this underdevelopment.

The importance of correcting this archaeological scepticism emerges very clearly from the revised European Convention on the Protection of the

Archaeological Heritage, a convention ratified by Sweden

1

(Trotzig, 1993), which requires each signatory ‘to ensure that archaeological excavations

and prospecting are undertaken in a scientific manner’ and that ‘non-

destructive methods of investigation are applied wherever possible’ (Article

3ib).

The overall aim of this thesis is to investigate the benefits of using geophysical methods in Swedish archaeology. By surveying sites situated in different geographical contexts, from the mountain tundra region in Lapland

(Paper II) to the arable lands of Gotland (Paper IV) and Öland (Paper III), and at sites spanning a time scale from the Neolithic (Paper II) to the Middle

Ages (Paper V) (Fig. 1), I intend to gain a better understanding of the pitfalls and possibilities involved in using archaeological geophysics in different geological contexts.

The surveys have, with one exception, been carried out in the eastern part of Sweden, and this is of course a limitation, as the natural environment and archaeology of the rest of Sweden may be different. But even though the surveys have an easterly focus, it is my belief that many of the conclusions presented in this thesis are also valid for geophysical surveys in other parts of the country and other areas with similar geological, pedological and archaeological preconditions. The purposes and specific research questions connected with these different projects are presented in papers II-V.

It became evident in the process of carrying out these investigations that geophysical methods had only been used sporadically within Swedish archaeology. With the purpose of understanding the reason for this limited use, a review of prior geophysical surveys at archaeological sites in Sweden was carried out (Paper I), and an attempt was made through correspondence with several Swedish field archaeologists involved in these earlier surveys, to investigate on a limited scale their attitude towards archaeological geophysics and their level of satisfaction with the resulting data and interpretations.

1 http://conventions.coe.int/Treaty/EN/Treaties/Html/143.htm

12

Figure 1. Sites prospected by the author since 2006. 1: Tierp Church, Uppland, 2:

Alunda Church, Uppland, 3: Bälinge Church, Uppland, 4: Sigtuna, Uppland

(several sites), 5: Fornsigtuna, Uppland (several surveys), 6: RAÄ 108 in the parish of Fresta, Uppland, 7: Frescati , Uppland (several sites), 8: Karsvik,

Uppland (in collaboration with UV-Teknik, Swedish 7ational Heritage Board), 9:

Riddarholmen Church, Stockholm, 10: Stensborg in the parish of Grödinge,

Södermanland, 11: Korsnäs, in the parish of Grödinge, Södermanland, 12: Tjuls, in the parish of Eskelhem, Gotland, 13: Stånga annex, in the parish of Stånga,

Gotland, 14: Odvalds in the parish of Linde, Gotland, 15: 7ygårds, in the parish of

Eke, Gotland, 16: Sandbyborg, in the parish of Södra Sandby, Öland, 17 Hedeby

Hochburg, Germany (not shown on the map), 18: RAÄ 1372, in the parish of

Sorsele, Lapland. Coordinates in SWEREF99 TM.

13

A review of the literature further enabled a study of data collection strategies, sampling densities and the choice of instruments and methods, as these are factors which might have affected the outcomes of the surveys. In addition, the interpretations of the geophysical data were compared whenever possible with the results of subsequent excavations in order to evaluate the success of the surveys. This review was important, as I strongly believe that it is essential to evaluate thoroughly both successful and failed geophysical surveys if better decisions and strategies are to be arrived at in the future.

Related research questions include, but are not restricted to:

• What methods are especially suitable for use in Sweden?

• In what way do Sweden’s geological environments and climatic conditions affect the success of geophysical prospection?

• What are the sizes and physical properties of the archaeological structures and remains commonly found in Sweden, and what methods are best suited for their detection?

• How dense do the measurements have to be in order to identify the archaeological features and interpret them correctly?

• How does the situation for archaeological geophysics in Sweden relate to that in other countries?

Several overviews of archaeological geophysics have been written previously (e.g. Scollar et al. 1990; Clark 1996; Conyers & Goodman 1997;

Bevan 1998; Gaffney & Gater 2003; Witten 2006; Aspinall et al. 2008;

Campana & Piro 2009) and it is not my intention to provide a complete review of every aspect of the subject. As an archaeologist, my main interest is in the interpretation and implementation of geophysical data in archaeological projects, and not necessarily in the technical development of the methods. As only one doctoral thesis (Fischer 1980), one licentiate thesis

(Persson 2005a) and a short text describing archaeological geophysics and some of its benefits and disadvantages (Liljenstolpe 1998) have been written on this topic in Sweden, I felt that the time was right to take this theme up once again.

The thesis begins with a description of the theory behind the geophysical methods most commonly used in archaeology, followed by a short description of the development of archaeological prospection with an extra focus on developments in Sweden. This chapter is followed by a discussion on the impact of geological, pedological and geomorphological conditions,

14

which are vital factors for the planning and execution of any geophysical survey. Another important factor that affects the success of geophysical prospection is the character of the archaeological remains and structures. As geophysical instruments detect only contrasts between the archaeological remains and the surrounding soil matrix, one must understand the physical characteristics and properties of archaeological features if a correct choice of method and sampling density is to be made. This part of the thesis also investigates existing guidelines for data collection and suggests suitable strategies for the Swedish environment. Questions regarding sizes of survey grids and the interpretation of geophysical maps are discussed. The thesis ends with a discussion of future prospect for archaeological geophysics in

Sweden.

15

2. Geophysical prospection methods

Archaeological geophysics is defined as “the examination of the Earth’s

physical properties using non-invasive ground survey techniques to reveal

buried archaeological features, sites and landscapes” (Gaffney & Gater

2003:12), and the same authors have provided a concise list of methods for doing this (Gaffney and Gater 2003:26), labelling them as either active or passive. Active methods transmit an electromagnetic field, an electric current, or a seismic wave into the ground and measure how the soil and other materials alter its transmission, while passive methods detect voltages or fields that are already at the Earth’s surface and determine how this measurement changes within the surveyed area. GPR, Slingram, magnetic susceptibility and electrical resistance or resistivity are examples of active instruments or methods, whereas magnetometry is the most commonly used passive method (see below).

The list of publications describing geophysical prospection is extensive

(e.g. Milsom 1989; Reynolds 1997; Mussett & Khan 2000; Kearey et al.

2002), and several books focusing on geophysical methods for archaeology are available (e.g. Aitken 1974; Scollar et al. 1990; Clark 1996; Conyers &

Goodman 1997; Bevan 1998; Neubauer 2001; Gaffney & Gater 2003;

Witten 2006; Aspinall et al. 2008; Campana & Piro 2009). These provide archaeologists with information about basic techniques and about the more advanced theoretical and methodological aspects of geophysical prospection.

The summary presented here should be regarded as a basic introduction to the more common concepts of archaeological geophysics, and the reader is advised to seek more detailed information in the above publications. The methods used in the projects described in this thesis include magnetometry

(Paper III and IV), magnetic susceptibility (Paper II) and GPR (Paper III and

V).

2.1. Magnetic methods

Ever since the first magnetic surveys at archaeological sites were carried out by Aitken and Hall in 1958 the magnetic method of prospection has been increasingly widely used in archaeological projects all over the world. The fluxgate magnetometer (see below) has been referred to as the “work horse” of archaeological geophysics, for example (see Clark 1996:69; David et al.

2008) and is currently one of the most popular and most frequently used instruments for geophysical prospection at archaeological sites.

16

As with all geophysical methods, one is looking for a contrast between the buried archaeological feature and the surrounding soil. In magnetic prospection this contrast can either be related to enhancements in magnetic susceptibility (the degree to which a substance can be magnetised when subjected to a magnetic field), or to magnetic remanence. Here thermoremanent alterations are of particular interest to archaeologists because, being caused by pronounced heating and cooling of the archaeological structure, they are related to burning and represent the earth’s magnetic field at the time when this burning took place.

A basic introduction to magnetism in general and soil magnetism in particular is presented below, together with a brief introduction to magnetic susceptibility and thermoremanent magnetism and its implications for archaeological geophysics.

2.1.1. The Earth’s magnetic field

The Earth’s magnetic field is created by movement of a liquid iron alloy in the planet’s outer core (Marshak 2008:50ff); this field is essential for life as it protects us from harmful particles emitted by the sun (Reynolds

1997:131). The Earth’s magnetic field resembles that of an ordinary bar magnet, with a north and a south pole and magnetic flux lines passing through the middle of the Earth and then travelling outside it from the north magnetic pole to the south magnetic pole (see Fig. 2; Kearey & Brooks

2002:156). The strength of this magnetic field, expressed in nanotesla (nT), differs depending on where you are on Earth’s surface and varies from

c.30,000 nT at the magnetic equator to 60,000 nT at the north and south magnetic poles (Reynolds 1997:134).

Figure 2. The Earth’s magnetic field and the location of its magnetic and geographical north poles.

17

As indicated in Fig. 2, the location of magnetic north differs from that of true or geographical north. This angular difference is called the magnetic declination and is a value that varies over time as the magnetic poles are constantly moving at a slow pace (Marshak 2008:78f). The declination map for Sweden is updated regularly by the Swedish Geological Survey (SGU) and the map for 2012 is presented in Fig. 3, showing values varying between

c. 2-14 degrees. Previous locations of the Earth’s magnetic north pole are recorded by directions frozen in magnetic features such as in situ igneous rocks and fired archaeological structures such as kilns. As such, this direction can be a means for roughly dating archaeological structures fired at temperatures above the Curie point (Aitken & Weaver 1962; Sternberg

2008). The magnetic inclination, on the other hand, can be measured with a compass that is allowed to tilt freely on a horizontal axis and can be defined as the angle of dip that the compass takes at different latitudes on Earth, being parallel to the ground surface (0°) at the equator and pointing straight down (90°) at the north and south poles. The different locations of the magnetic north pole during historic and prehistoric times and the current magnetic declinations and inclinations have important implications for interpreting the results of magnetic prospection. This will be examined in greater detail below.

Figure 3. Magnetic declination values for 7orway, Sweden and Finland © Swedish

Geological Survey (SGU).

18

2.1.2. Magnetism, basic concepts

All materials show some reaction to a magnetic field, though this reaction may be weak or strong. This magnetic field can arise from electron spin and orbits at an atomic level (Thompson & Oldfield 1986:3). Some materials are permanently magnetised and will be magnetic even if they are not subjected to a magnetic field, while others are only magnetic in the presence of such a field (Reynolds 1997:122). Remanent magnetism is the name given to permanently magnetised materials, while induced magnetism describes materials that are only magnetic in the presence of a magnetic field. The magnetic property of any material is, among other things, dependent on the number of electrons present in each subshell of the atom and in which of the two possible directions the orbiting electrons are aligned when subjected to an applied magnetic field (Aspinall et al 2008:11ff). Different materials can thus be classified according to the direction of spin and the overall magnetic moment (positive or negative) created by this motion.

These materials, which in some cases are chemical elements, can be categorised as being diamagnetic, paramagnetic, ferromagnetic, antiferromagnetic and ferrimagnetic. If one were to expose these materials or chemical elements to a magnetic field, diamagnetic materials, such as quartz, with complete electron subshells, will produce an opposing weak negative magnetisation and paramagnetic materials, such as biotite, will align with the applied field and produce a weak positive magnetisation, as their atomic subshells are incomplete. Both diamagnetic and paramagnetic materials will only display this characteristic behaviour in the presence of a magnetic field.

By comparison, ferromagnetic materials such as iron and nickel are able to retain their magnetic properties without the presence of a magnetic field, as they have unpaired electrons in their atomic subshells and magnetic dipoles aligned in one direction, which produces a very strong magnetisation. Antiferromagnetic materials, on the other hand, have dipoles of equal strength that are aligned in opposite directions, resulting in a zero or almost zero net magnetisation as the magnetic moments cancel each other out, as is the case with the imperfect antiferromagnetic mineral haematite.

Magnetite is a ferrimagnetic material and has magnetic dipoles in opposite directions but of unequal strength, resulting in a positive net magnetisation

(Fig. 4 & Table 1). For a more detailed description of this classification, see

Thompson and Oldfield (1986:3ff), Reynolds (1997:122ff) or Kearey and

Brooks (2002:157f). Many materials, such as a piece of rock or a soil sample, will contain mixtures of different minerals and magnetic components, and the total magnetic susceptibility of such a material will consist of the magnetic susceptibilities of all these categories of magnetic behaviour added together (Dearing 1999:6). Magnetic susceptibility values for certain common rock types and minerals are presented in Reynolds

(1997:126ff).

19

Figure 4. Orientation of the magnetic moments for ferromagnetic, antiferromagnetic and ferrimagnetic materials.

As ferromagnetic materials rarely occur in nature, the most influential components are the ferrimagnetic iron oxides such as magnetite and maghemite. These ferrites are very common in many rock types and are responsible for the very high magnetic susceptibilities of basic igneous rocks such as basalt (Thompson & Oldfield 1986:13ff). Magnetite, and primarily maghemite, are also very important for the detection of archaeological features.

Table 1. Examples of diamagnetic, paramagnetic, ferromagnetic, antiferromagnetic and ferrimagnetic materials.

Diamagnetic Paramagnetic Ferromagnetic

Antiferromagnetic/

Imperfect

Antiferro- magnetic

Ferrimagnetic

Quartz

Feldspar

Calcite

Water

Olivine

Garnet

Biotite

Pyroxene

Iron

Cobalt

Nickel

Haematite

Chromium

Magnetite

Maghemite

2.1.3. Magnetic susceptibility and archaeological prospection

Magnetic susceptibility is the degree to which a substance can be magnetised when subjected to a magnetic field (Thompson & Oldfield

1986:25). Le Borgne has shown that the magnetic susceptibility of the soil layers within the top 20-30cm of the surface is greater than that of the underlying subsoil, and has suggested that this enhancement could be caused by soil development processes (the fermentation effect) and burning (Le

Borgne 1955, 1960). These effects have been studied by Weston (2002). The burning effect has also been studied both in the laboratory (e.g. Tite &

Mullins 1971; Mullins 1974; Graham & Scollar 1976) and during experimental firing in different parts of the world (e.g. Bellomo 1993;

Marshall 1998; Linford & Canti 2001; Weston 2002). One important finding has been that enhancement is achieved when the antiferromagnetic mineral haematite is transformed to the highly magnetic mineral magnetite during firing in a reducing environment. When the fire is put out and oxygen is

20

reintroduced, the magnetite continues to transform into the ferrimagnetic mineral maghemite. This mineral does not have as high a magnetic susceptibility as magnetite, but it is still much higher than that of haematite, thus creating a clear contrast between the fired structure and the surrounding soils. A prerequisite for this transformation is the presence of organic matter.

A relatively modest temperature of c. 150-200°C during firing is sufficient for creating a slight enhancement of magnetic susceptibility (Thompson &

Oldfield 1986:75; Linford & Canti 2001:224; Aspinall et al. 2008:24).

Forest fires or slash-and-burn agriculture would also cause such a susceptibility enhancement (Clark 1996:115) providing sufficient temperatures and burning times were achieved. Ploughing will distribute the magnetic soils of fires and burned features throughout the topsoil and cause an increase in magnetic susceptibility. This mixing can also be caused by natural factors such as bioturbation.

Enhancement of the magnetic susceptibility of the topsoil has also been shown to be caused by other mechanisms such as micron-sized soil bacteria, which contain a small amount of magnetite and found in accumulations in the unburned remains of prehistoric wooden structures in southern Germany and Austria (Fassbinder et al. 1990; Fassbinder & Stanjek 1993; Neubauer

2001). These bacteria give rise to weak magnetic anomalies that are only detectable by very sensitive magnetometers. Given the abundance of glacial till soils in Sweden (see chapter 4) it is unlikely that background levels in such soils would be low enough for such unburned features caused by the presence of magnetic bacteria to be detected, despite the use of very sensitive instruments. This mechanism, enabling the detection of unburned wood, has unfortunately been confused by one Swedish surveyor, resulting in the erroneous claim that certain bugs (beetles) supposedly containing magnetite die in situ after consuming decaying unburned wood, thus providing a magnetic contrast that is readily detectable using a magnetometer 2 (Kockum 2006). Such claims are not supported by current research, however. A concise list of the mechanisms currently known to cause enhancement of magnetic susceptibility is presented by Aspinall et al.

(2008:25).

Even though magnetic susceptibility measurements are commonly carried out using an active instrument such as the EM38 or MS2D, archaeological features that cut through the topsoil into the subsoil and are then subsequently refilled with magnetically-enhanced topsoil may also be detected with a magnetometer (Clark 1996:65; Gaffney & Gater 2003:38).

Depending on the surrounding soils, this would cause either a positive or a negative anomaly to be detected. Such prehistoric features could be pits, postholes or ditches.

2 http://www.ostran.se/NYHETER/Oeland/sensationella_fynd_vid_fornborg

21

2.1.4. Thermoremanent magnetism and archaeological prospection

Another important aspect of magnetism is the magnetic remanence of different materials, and thermoremanent magnetisation is of particular importance for archaeological prospecting. Thermoremanent magnetisation

(TRM) occurs when a magnetic material is heated above its Curie point and subsequently cooled below its blocking temperature. Above the Curie point, which is particular to given magnetic materials, the heated material is essentially paramagnetic, carrying only a slightly positive magnetic susceptibility. When the material cools, however, it regains its magnetic characteristics and the direction of the magnetisation aligns with the geomagnetic field at the time of cooling.

In prehistory, for example, this heating could be achieved by firing in hearths and kilns to temperatures exceeding the Curie temperature of the magnetic mineral present in the soil. (The Curie temperature of magnetite, for example, is 578°C and that of hematite 675°C). Upon cooling, a typical bipolar magnetic anomaly generated by thermoremanent features is created, and this is detectable by a magnetometer. Bipolar anomalies can also be caused by igneous rocks, iron artifacts or lightning (see Bevan 1995a, 2009;

Jones and Maki 2005; Maki 2005), and caution should therefore be exercised when interpreting magnetic data.

It is worth noting that ploughed out hearths lose their thermoremanent component as the hearth material slowly mixes with the surrounding soils.

This would lead to a general increase in the magnetic susceptibility of the soil (Weston 2002:208). Note that hearths will in most cases have both increased magnetic susceptibility values and a thermoremanence (Weston

2002:208).

2.1.5. Passive instruments for magnetic prospection

Magnetometers are sensitive geophysical instruments that are able to detect very small fluctuations in the Earth’s magnetic field. The unit of magnetic measurement is the tesla (T), but prehistoric remains usually only produce fluctuations in the Earth’s magnetic field of a few nanoteslas (1nT=

10

-9

T) or even picoteslas (1pT= 10

-12

T). The Earth’s magnetic field in the

Stockholm area in Sweden has an approximate strength of 50 000 nT and instruments are able to detect small deviations from this value caused by archaeological features. The detectability of prehistoric features is also dependant on factors such as sampling density and the spatial variability in the magnetic properties of the soil and rock in the area surveyed.

A single-sensor magnetometer not only measures spatial patterns caused by remains buried in the soil but also the constantly fluctuating magnetic field of the Earth (i.e. the total field). These changes, caused by solar wind or

22

solar magnetic storms (see Aspinall et al. 2008:31f) can effectively be removed from the data by using two sensors instead of one and subtracting the two values from each other. The two sensors can be mounted in different ways in relation to each other and the most common configuration for archaeological prospection being that in which one sensor is placed vertically on top of the other at a fixed distance. Both sensors are equally affected by the Earth’s magnetic field, but the lower sensor is more strongly affected by any remains buried in the soil. The Earth’s magnetic field can then be filtered out simply by subtracting the values recorded by the upper sensor from those of the lower sensor. The remainder is primarily a product of the buried structures or features at the site but is also affected by the natural geological and pedological environment. These buried structures, objects or features could be either prehistoric or modern, or due to natural soil variation. A geophysical interpretation based on the appearance of the anomaly in the data must therefore be carried out before any archaeological interpretation is attempted.

Fluctuations in the Earth’s magnetic field can also be removed by highpass filtering the collected data. Several authors have convincingly presented data favouring the use of a pair of total-field sensors in a horizontal configuration instead of arranging them vertically. This horizontal configuration essentially halves the time needed for carrying out the survey

(e.g. Tabbagh 2003; Becker 2009).

Gradiometers can use different distances between the sensors, but for archaeological purposes it is common to use a separation of 0.5–1m (Fig. 5).

The sensor separation governs the depth sensitivity, as a larger distance between the sensors results in an increased sensitivity to deep features.

Archaeological remains can normally be detected with a gradiometer when buried in the uppermost two metres of soil (Clark 1996:78f), but the nature of the buried remains and the sensitivity of the magnetometer are also factors that need to be considered (Fig. 5 & Table II), and highly magnetic materials can be detected even when buried at greater depths (David et al. 2008:16).

The depth sensitivity also increases if a total-field magnetometer is used instead of a gradiometer. Since the majority of archaeological remains in

Sweden are weakly magnetic, they would most likely not be detectable below a depth of two metres. The possibility of detecting deeply buried features is also dependent on the magnetic properties of the soil in which the feature is located and on those of the underlying bedrock.

A short distance between the magnetometer’s sensors has the benefit of being less sensitive to deeper structures or those that are more distant, such as passing cars, and should therefore be more likely to focus on the archaeological remains in the upper parts of the soil. Magnetometers are only sensitive to ferromagnetic and ferrimagnetic features, however, so that copper, gold and silver objects are not detected with the instrument.

23

Figure 5. Common gradiometers used for archaeological prospection (left)

Foerster DLG 4.032, (middle) Geoscan FM 256, (right) Bartington Grad 601-2.

Photos: 7y Björn Gustafsson & Siska Williams.

Different kinds of magnetometers such as proton magnetometers, electron spin-resonance magnetometers, alkali vapour magnetometers, SQUID

(Superconducting Quantum Interference Device) magnetometers and fluxgate magnetometers are available, all of which use different constructions and methods for measuring magnetic fields. A brief description of these instruments is presented below with special focus on the fluxgate gradiometer, as this was the instrument used for the investigations reported here.

Table 2. The sensitivity of certain magnetometers. After Aspinall et al. (2008).

Proton Electron spin resonance

0.1-0.5 nT 0.05 nT

Alkali-

Vapour

SQUID Fluxgate

0.01 nT 0.00001 nT 0.1 nT

Proton magnetometers, as used in the early days of archaeological geophysics (see, for example, Aitken 1958; Clark 1996), consist of a bottle containing a hydrogen-rich fluid with a coil of wire around it through which an electric current is passed. When the current is sent through the coil, the protons in the fluid align with the direction of the applied field, and when the current is turned off, they start to precess about the direction of the Earth’s magnetic field at the site. This precession causes a small electric current to be generated with a frequency of alternation proportional to the strength of the magnetic field. The measurement process is quite slow, and only very limited use is made of proton magnetometers in current archaeological

24

prospection. Some of these issues have been corrected in the electron spin resonance magnetometer (Overhauser) in which a second liquid is introduced which aids in the alignment of the protons and reduces the time needed to complete a measurement. Even so, their use in archaeological geophysics is limited.

Alkali vapour magnetometers use the vapour of various alkaline metals, most commonly caesium. When a ray of polarised light is sent through the alkali vapour, the translucency of the vapour is directly proportional to the strength of the magnetic field at the point of measurement.

The SQUID magnetometer is by far the most sensitive instrument available for archaeological prospection but its use is restricted by the fact that the sensors must be kept at a very low temperature in order to work, making it expensive to use for field work. The sensitivity of the measurements is very high, however, and this allows weakly magnetic prehistoric features to be measured in areas where there are few other anomalies. More detailed descriptions of the magnetometers are presented by Aspinall et al. (2008:41ff).

The fluxgate gradiometer is the most frequently used magnetometer in geophysical prospection at archaeological sites. A fluxgate magnetometer was first constructed for archaeological purposes by

Alldred (1964) and subsequently developed by

Philpot (1973). The fluxgate sensor consists of a core made up of two pieces of metal having very high magnetic permeability (Fig. 6). A copper wire is wound around them and the core is driven in and out of saturation by an electric current passing through the coil. When the core is out of saturation it is affected by the ambient magnetic field, which generates an electric current in a secondary detector coil proportional to its own magnitude (Clark 1996:69).

Three of the most frequently used gradiometers

(Foerster Ferex 4.032 DLG, Geoscan FM 256 and

Bartington Grad601) are depicted in Fig. 5. They have vertical distances between the fluxgate sensors of 0.65m, 0.5m and 1m, respectively, giving them different strengths and weaknesses

(see the discussion above).

Figure 6. Layout and functioning of a fluxgate gradiometer. After Clark

(1996).

25

2.2. Electromagnetic methods

Several electromagnetic instruments are available for archaeological prospecting, including metal detectors, ground-penetrating radar (GPR), the

Slingram and various kappameters. These are all active methods. This section provides an introduction to the physical principles of operation of some of these electromagnetic instruments, emphasising the methods used for the surveys reported in this thesis.

2.2.1. Ground penetrating radar, an introduction

RADAR, an acronym for Radio Detection And Ranging, has been used since the Second World War for the detection of airplanes (Conyers

2004a:16), and ground-penetrating radar, based on the same principles, has been known from the 1920s onwards (Stern 1929). The history, theory and application of GPR in general and GPR for archaeological use in particular have been presented by Conyers & Goodman (1997).

Figure 7. An archaeological geophysical survey using an X3M system and a

500MHz antenna manufactured by Malå Geoscience. Photo: Joakim Schultzén.

GPR (Sw. georadar or markradar) is an electromagnetic method which is similar to the echo localisation method, or sonar (Sound Navigation And

Ranging) used on boats (Fig. 7), but instead of emitting an acoustic or sound wave into the water, the GPR device emits an electromagnetic pulse into the ground from a transmitting antenna. The time in nanoseconds needed for a pulse to travel from the antenna into the soil and reflect back to a receiving antenna from a reflective structure of some kind is measured. This is called a two-way travel time (Conyers 2004a:38). Such reflective structures could be stones, bedrock, the ground-water table, archaeological features or soil boundaries. The clearest reflections are produced when the radar pulse is

26

reflected at the interface between two materials having very different electromagnetic properties (Conyers & Goodman 1997:27).

The reflected pulses recorded by the receiving antenna result in what is known as a radar trace (Fig.

8). This takes the form of an irregular, decaying sinusoid. The large amplitude reflections in the upper part of the trace are caused by the air/ground interface directly below the antenna and by the transmitted signal going directly to the receiver. At every soil layer boundary or object in the soil, some of the radar energy is reflected back towards the surface and the remainder of the wave travels deeper into the ground and continues to diminish as more and more energy is reflected back to the surface. As a consequence, reflections from objects or soil boundaries situated deep in the soil will be much less

Figure 8. Example of a radar trace. The large amplitudes noted in the upper parts of the pictures represent the radar wave’s interaction with the air/ ground interface at ground level directly below the antenna.

pronounced than those from objects or soil boundaries with equal properties that are located at shallower levels. As the radar is moved along a predetermined transect on the ground surface a set of numerous radar traces is collected, and by placing these next to each other a radar profile, or radargram, is generated (Fig. 9). The maximum positive amplitudes are depicted in black and the minimum negative amplitudes in white.

Figure 9. Example of a radar profile or radargram created by placing a number of radar traces next to each other.

A radargram is a two-dimensional vertical section containing stratigraphic information about the soil below the surveyed transect. As the radar wave is not transmitted in a narrow beam straight down into the soil but spread roughly in the shape of a cone, radar reflections from small objects such as stones will have the shape of a hyperbola (see Conyers 2004a:57; Witten

2006:229; Ernenwein & Hargrave 2009:34). This transmission cone, commonly referred to as the radar footprint (Leckebusch 2003:215; Conyers

2004a:61ff), effectively means that the GPR not only reveals objects situated

27

directly below the antenna but also objects situated in front, behind and on either side of it. The antenna begins to detect a point object when this comes into contact with the front-most part of the radiation footprint, but the radar echo is plotted as if it were directly below the antenna. This means that the object is first recorded as being deeper than it really is, as the distance between the antenna and the front part of the radiation footprint is greater than the depth of the actual object. As the antenna passes over the object, the distance, called the range, will gradually decrease, and when the antenna is directly above the object the correct depth will be recorded, provided the radar wave velocity is known (see section 2.2.2.). As the antenna continues along the survey transect the apparent depth will increase until the rearmost part of the footprint completely loses contact with the object. A hyperbolic reflection in a radargram is depicted in the centre of Fig. 10 (left).

Figure 10. A reflection hyperbola created as a result of the shape of the transmitted radar wave interacting with a buried object in the soil.

If many radar profiles are surveyed next to each other it is possible to merge these into a three-dimensional cube or rectangular cuboid (Fig. 11).

This is mainly done to increase the interpretability of the original twodimensional data. The interpretability can be increased further by subdividing the cube or rectangular cuboid into horizontal images, called depth or time slices (see Goodman & Nishimura 1993; Goodman et al. 1995;

Fig. 11). By making an animated movie of these time slices one can gain a better understanding of how structures change with depth (cf. Neubauer et

al. 2002).

Figure 11. GPR data: from densely surveyed radargrams to the creation of time or depth slices.

28

The depth penetration of a GPR is partly governed by the soil properties, but also by the antenna frequency, which is related to the wavelength of the transmitted electromagnetic pulse. A high-frequency antenna transmitting a short wavelength radio pulse will be able to resolve smaller features, but as the frequency is high the depth of penetration will be low and deeply buried features will not be detected. Low-frequency antennas have long wavelengths and will penetrate deep into the ground but will only be able to resolve larger features. For archaeological purposes this trade-off between the depth of penetration and the resolution of smaller features has resulted in a preference for antennas having frequencies in the central range of 400-500

MHz. This will provide a depth penetration of c. two metres, depending on the properties of the soil (David et al. 2008:31). When the targets are thought to be situated at shallow depths, higher-frequency antennas can be deployed.

Depending on the prerequisites and purpose of the archaeological project, antenna frequencies can range between 250 and 1000 MHz.

2.2.2. GPR and soil properties

In addition to being reduced by reflections, the electromagnetic pulse will be attenuated in conductive soil.

Highly conductive soils and soils with a high magnetic permeability will weaken the radar pulse, which in turn will affect the penetration depth of the instrument. The soil properties are consequently of vital importance to the success of any GPR survey.

One clear benefit of using GPR is the possibility for estimating the depths of targets found in the data. In order to do this, it is important to have some knowledge about the soil in which the radar wave is travelling. To obtain this information, the Relative Dielectric Permittivity (RDP) ε r

can be measured. RDP, defined as “a measure of the ability of a material to store a

charge from an applied electromagnetic field and then transmit that energy

(Conyers 2004a:45), is related to the velocity with which the radar wave travels through different soils or materials, and is generally negatively correlated with velocity (see Table 3). The characteristics of soils and other materials are of course highly diverse and the values in the table should only be regarded as rough guidelines.

If the velocity with which the radar wave travels through the soil is known, depth estimates can be produced. This can be done by one of several methods (Conyers & Lucius 1996), of which the most reliable and exact one involves hammering a reinforcement bar into the soil profile at different depths and correlating the actual depth of the bar with the GPR results, thus obtaining a body of RDP and velocity data for different depths. Note that a freshly excavated soil profile must be used if correct values are to be obtained. Another option is to use the depth of a known feature to calculate the velocity.

29

Table 3. Relative dielectric permittivity (ε r

), velocity and conductivity of common soil types and materials. Data modified from Reynolds (1997:104 and references cited therein) and Cassidy (2009a:46).

Material ε r

Velocity (m/ns) Conductivity (mS/m)

Air

Fresh water

Seawater

Seawater Ice

1

81

81

2.5-8

0.3

0.033

0.033

0.078-0.157

0

0.1-10

4000

10-100

Permafrost

Costal sand (dry)

Sand (dry)

Sand (wet)

Clay (dry)

Clay (wet)

Clayey soil (dry)

Clayey soil (wet)

Marchland

Agricultural land

1-8

10

3-6

25-30

2-20

15-40

3

10-15

12

0.106-0.3

0.095

0.12-0.17

0.055-0.06

-

0.086-0.11

0.173

-

0.086

0.1-10

0.01-1

0.0001-1

0.1-10

1-100

100-1000

0.1-100

100-1000

-

Pastoral land

Average soil

Asphalt

Concrete

15

13

16

3-5

6-30

0.077

0.083

0.075

0.134-0.173

0.055-0.112

-

-

5

-

1-100

If a soil profile is not available, one can perform either a common midpoint test (CMP) (Ulriksen 1982:13; Jol & Bristow 2003:15f) or hyperbola fitting, where a computer-generated hyperbola is remodelled to fit the echo arc in the GPR data (see Leckebusch 2000:194f; Cassidy

2009b:159; Goodman et al. 2009:485). The hyperbolas get broader as the velocity increases and narrower if it decreases (Ulriksen 1982:75f). For a

CMP test one needs two separate GPR antennas, and as most GPR systems have the transmitting and receiving antenna housed in a closed case this is not usually possible. It is advisable to make velocity measurements at different places within the survey area as soil properties can vary within one site. Furthermore, new velocity measurements should be carried out if the measurements span several days, as the soil moisture content can vary greatly with time (Conyers & Lucius 1996:38).

Attenuation of a radar wave is affected by the electrical conductivity (σ) and magnetic permeability (µ) of the soils through which it passes, and also slightly by the RDP. Electrical conductivity, defined as “the ability of a

material to pass free electric charges under the influence of an applied field

(Cassidy 2009a:54), is partly governed by the water content of the soil, but also by the salinity of the medium. As a consequence, highly conductive wet clay soils will cause a high attenuation of the radar wave, reducing the depth of penetration significantly (Leckebusch 2003:214f; Conyers 2004a:50).

30

Successful measurements have nevertheless been carried out on clay soils, which do not per se preclude the use of GPR (Weaver 2006). Certain dry soils, which have generally been accepted as suitable for GPR measurements, can also have a high conductivity and cause marked attenuation of the radar wave, and thus other factors in addition to soil moisture need to be taken into account when estimating the likely success of GPR (Conyers

2004a:50).

One option could be to perform an EM38 survey in order to measure the electrical conductivity of the soil before the GPR survey, thus estimating the likelihood of success (see Clay 2001:1f; Conyers 2004a:53; Viberg

2007:31). This is exemplified in Fig. 12, where an area of high conductivity as measured with the EM38 (see below), is directly correlated with an area of reduced depth penetration in a radargram.

Figure 12. Correlation between an area of high soil conductivity as identified by the EM38 and an area of reduced depth penetration in GPR data.

The effects of rainfall on GPR data have been studied by Bevan

(1984:203f), who showed that the same profile before and after rainfall had quite different characteristics. Many radar echoes seemed to derive from the water as it percolated down through the upper unsaturated part of the soil, and this effect was visible in the GPR data for three to four days after the rainfall, depending on the site (Bevan 1984:204). Wet soils alone are not always a problem, however, and in some areas certain kinds of structures can actually be enhanced by rain, as the soil moisture contrast between the archaeological structure and the surrounding soil could be increased (see

Bevan 1984:203f; Van Dam et al. 2002; Conyers 2004b).

This has also been confirmed during repeated seasonal measurements performed on an Iron Age house terrace (no. 6) situated on a well-drained glacial till soil at the Fornsigtuna, or Signhildsberg, site in Central Sweden

(Figs. 1 & 13). The archaeology at this site had been revealed by means of an earlier small trench in the terrace and cultural material had been shown to be situated in the upper c. 0.5 m, below which the bedrock became visible

(Damell 1986; Damell 1991:48ff).

31

Figure 13. Map of the Fornsigtuna site, showing the location of house terrace no 6.

The scale bar was missing from the original map, but the size of house terrace no. 6 is roughly 37.5m x 10m. After Damell (1991), with “no.6” added by the author.

GPR measurements were carried out in 2009-2010 using an X3M system manufactured by Malå Geoscience connected to a 500MHz antenna. Data were collected every 0.03m (inline sampling distance) in profiles located

0.25m apart (crossline sampling distance). A follow-up survey in August

2010 used a closer crossline sampling distance of 0.1m between the transects.

Measurements performed in November 2009 produced slightly clearer reflections from both shallow and deeply buried features than those made in

August 2009 and August 2010 (see Fig. 14), while the results collected in

April were somewhat clearer than those of the August measurements but not as clear as the November measurements. The depth of penetration of the

GPR seemed to be similar for all the surveys but the features situated in the upper part of the profile were more clearly defined in the spring and autumn surveys. The bedrock, which is very shallow in the middle part of the radar profile and deeper towards the ends, is clearly visible in all the profiles.

As the results were very similar, it must be concluded that future surveys of house terraces in well-drained glacial till environments could be carried

32

out in any season, although the slightly improved results obtained in

November and April perhaps speak in favour of the spring or autumn.

Winter measurements in January were also attempted, but due to several decimetres of snow covering the house terrace the survey wheel attached to the GPR equipment had severe problems and did not record the correct length of the radar profiles. This could perhaps have been solved by using another data acquisition method, such as the hip chain, but this was not available at the time of the surveys. Another problem with surveying in deep snow is that the antenna tilts as a result of the compacted snow on the previously surveyed transects. This problem could most likely be solved by simply putting the radar antenna onto a broader sledge, which spreads its weight more evenly. The results of the survey, presented as a time slice, are given in section 5.1.

Figure 14. GPR profiles collected at the seasonal measurements of house terrace no. 6 at Fornsigtuna in August 2009(top), 7ovember 2009 (second from top), April

2010 (second from bottom) and August 2010 (bottom).

The other factor affecting GPR measurements is magnetic permeability, defined as “a measure of the ability of a medium to become magnetized when

33

an electromagnetic field is imposed upon it” (Conyers 2004a:53). This factor has only been seen to affect GPR data in rare cases (e.g. Conyers 2004a:53f and references cited therein), as most soils are weakly magnetic (Van Dam

& Schlager 2000:437; Cassidy 2009a:55f).

2.2.3. GPR and topography

The topography of the survey area is of vital importance when carrying out GPR surveys (see Conyers & Goodman 1997; Conyers 2004a; Goodman

et al. 2006; Goodman et al. 2007; Leckebusch & Rychener 2007). Any tilt or roll of the GPR antenna caused by the terrain will make the radar look in a different direction. The GPR profiles will nevertheless still assume that all the collected data apply to a point directly below the antenna, and as a result, erroneous maps of the subsurface will be created. This can be remedied by topographic tilt corrections applied to the radar profiles, which can be carried out using GPR data management software such as GPR SLICE

3

.

Another important thing to remember when carrying out GPR surveys is that smaller features such as piles of grass at the surface can cause the antenna to tilt and lose its ground coupling, thereby introducing anomalies into the radar profiles (Bevan 1984:202; Conyers 2004a:70f). Larger features of this sort can be mapped with a RTK-GPS or total station to enable the correct interpretation of such anomalies.

2.2.4. GPR manufacturers and suitable archaeological targets

In the early surveys carried out in Sweden and abroad the GPR equipment was housed in a car while the antenna was pulled over the ground and the results were printed continuously on a wide strip of paper (see. Fig. 15).

Then, as personal computers became available during the 1980s, the data came to be stored on the computer’s hard disc, which enabled subsequent filtering and data processing (Leckebusch 2003:214; Conyers 2004a:5).

Examples of common GPR systems used for archaeological prospection are presented in Fig. 16. Sweden is the home of several geophysical equipment manufacturers, such as Malå Geoscience and Radarteam

Sweden.

4

The results collected using different manufacturers’ GPR systems have been evaluated in terms of quality by Goodman (2006) and Seren et al.

(2007). Before buying a GPR system, it is advisable to test different brands, and also several antennas from the same manufacturer, as these can be different and produce data of differing quality.

3

http://www.gpr-survey.com/

4 http://www.radarteam.se/

34

Figure 15. Early GPR measurements performed at the Stone Age cemetery of

Ajvide on Gotland in 1983. The GPR antenna in the top left picture is connected to the data unit housed in the car. The radargrams are continuously recorded on a roll of paper by the printer in the car. Photos: Göran Burenhult.

Figure 16. Common GPR systems and carts used for archaeological prospection by

Malå Geoscience (left), Sensors and Software (middle), Geophysical Survey

Systems Inc. (GSSI) (right). Photos: Joakim Schultzén and Siska Williams.

35

GPRs have most commonly, and most successfully, been used to detect buried remains of buildings (e.g. Leckebusch 2000; Nishimura & Goodman

2000; Leucci 2002; Linford 2002; Neubauer et al. 2002; Trinks 2005;

Persson 2005a; Berard 2008; Novo 2010; Udphuay et al. 2010; Ercul et al.

2011; Löcker et al. 2011; Paper III, IV, V), although weakly reflective archaeological structures such as graves (e.g. Goodman 1993; Conyers

2006a; Forte & Pipan 2008; Fiedler et al. 2009) and postholes (e.g. Trinks et

al. 2010b) have also been detected. Lists of suitable and unsuitable targets for GPR surveys are provided by David et al. (2008) and Ernenwein &

Hargrave (2009).

2.2.5. The metal detector

The most popular instrument for use in Swedish archaeology is the metal detector, an electromagnetic instrument using a transmitter coil to generate an alternating magnetic field. When held over the soil, the field induces currents in any buried conductive material, and these give rise to secondary magnetic fields detectable by a receiving coil (Gaffney & Gater 2003:46f:

Linford 2006:2231). The majority of these metal detectors are VLF (Very

Low Frequency) instruments and essentially function as instruments of the

Slingram type (Scollar et al. 1990:570).

Metal detectors frequently use discrimination as a means of identifying different metals. This procedure is primarily used to remove the signal produced by iron in order to focus on precious metals such as gold, silver or bronze. The discrimination procedure is, among other things, related to frequency shifts and phase shifts in the transmitted signal (Scollar et al.

1990:570). The phase shift measured for different metals is compared with values obtained during field measurements and is used as a basis for discrimination. Iron, for example, is highly magnetic and will be in-phase, as opposed to the excellent conductor, bronze, which is out of phase by 180°

(see Scollar et al. 1990:570). Care must be taken, however, as the shape of the object and its orientation and size will affect the readings (Scollar et al.

1990:570; Connor & Scott 1998:80).

Other types of metal detector, such as the Pulse Induction Meter (PIM), have been used in archaeological research but are not as common as the VLF instrument described above. The basic operation of the PIM is described by

Gaffney and Gater (2003:47).

2.2.6. Magnetic susceptibility and the Bartington MS2D

Since magnetometers are affected by both the remanence and the susceptibility of magnetic features, they are unable to determine the magnetic susceptibility of materials directly or easily. Such measurements

36

can, however, be provided by means of active instruments such as the EM38 or MS2D (Dalan 2008:3).

There are several instruments available for measuring the magnetic susceptibility of soils, as listed and discussed by Dalan (2008:3f) and David

et al. (2008:36). Two of the more common instruments are the MS2D system of Bartington Instruments and the Slingram device EM38 by Geonics

Ltd.

The Bartington instrument (see Fig. 17) consists of an induction coil with a diameter of 18.5cm that measures the volume magnetic susceptibility (κ) of the top decimetre of the soil (Dearing 1999:28f). The results are expressed in dimensionless SI units (x 10

-5

) and are calibrated to read 0.5 κ in rough soils and 0.75 κ on smooth surfaces (see the operation manual for the MS2 magnetic susceptibility system, Bartington Instruments). The data therefore needs to be corrected for roughness and then multiplied by 2 during measurements made in rough terrain to achieve a correct value for the volume magnetic susceptibility. The instrument has been described in detail by several authors, e.g. Clark (1996:102ff), Dearing (1999), Gaffney &

Gater (2003:44ff) and Dalan (2008:5).

As enhanced magnetic susceptibility has been shown to be associated with anthropogenic activity, areas with enhanced magnetic susceptibility values in soils as compared with the prevalent background values at the site, can be used as an indicator of prehistoric human activity (Gaffney & Gater

2003:44; David et al. 2008:36). Down-hole magnetic susceptibility measurements have been shown to produce detailed stratigraphic profiles, thus aiding the interpretation of archaeological sites (e.g. Dalan 2006,

2008:13ff). The method can not only be applied as a prospection tool but also during excavations (e.g. Dalan 2008:7ff). Furthermore, as evidenced by the brief theoretical discussion above, fired anthropogenic structures such as hearths or kilns would probably have elevated magnetic susceptibility values, making them readily detectable using any of the available kappameters (Dalan 2008:3).

Figure 17. Surveying in the mountain tundra region of northern Sweden using the

MS2D by Bartington Instruments. Photos: Kerstin Lidén and Andreas Viberg.

37

The magnetic susceptibility of fire-affected limestone and sandstone has been shown to be higher than that of unfired specimens of the same rock type found at the same site, which would imply that burning or heating of rocks increases their magnetic susceptibility (Gose 2000:417f).

2.2.7. Conductivity and susceptibility measurements using the

Geonics EM38 Slingram instrument

Slingram instruments, e.g. the SCM (Howell 1966), the SH3 (Parchas &

Tabbagh 1978) or the EM38 by Geonics, have been used for detecting archaeological remains since the mid-1960s and have attracted the interest of many researchers (e.g. Tite & Mullins 1970; Bevan 1983; Tabbagh 1984,

1985, 1986a, 1986b, 1994; Frohlich & Lancaster 1986; Tabbagh et al. 1988;

Benech & Marmet 1999; Simpson et al. 2009).

Slingram instruments generate a primary magnetic field in a transmitting coil and measures the secondary magnetic fields created by buried features or soil materials in the ground in a receiving coil (Fig. 18). When the instrument is held up in the air the primary field will travel unaffected from the transmitter to the receiving coil, but if the instrument is held near the ground the transmitted electromagnetic field will induce eddy currents in buried conductive features or objects. These currents will in turn generate secondary magnetic fields which can be detected by the receiving coil. The response in the receiving antenna will thus tell the operator something about the presence and properties of buried conductive features or objects (Kearey

et al. 2002; Witten 2006:151ff).

Figure 18. Principles behind electromagnetic survey instruments and their detection of buried objects. After Kearey et al. (2002:209).

The depth penetration is governed by the distance between the coils, so that a larger distance will enable a greater depth penetration. However, as increasing depth also increases the measured volume of soil, it will lower the resolution and the detectability of smaller features within the soil volume.

The distance between the coils is 1m for the EM38.

38

One survey reported in this thesis (Paper II) was carried out using the EM38 from Geonics (Fig. 19). This is a two-in-one Slingram instrument measuring both the in-phase and quadrature components of the electromagnetic field.

The in-phase component is essentially a measure of the magnetic susceptibility of a material and the quadrature component is a measure of its electrical conductivity. The instrument can be operated either in vertical mode, with a vertical orientation of the dipoles (see Fig. 19), or in horizontal mode

(lying on its side), with a horizontal orientation of the dipoles. Different depth sensitivities are achieved depending on the orientation. The greatest sensitivity in the vertical mode is at a depth of c. 0.3-0.4m when the instrument is measuring electrical

Figure 19. Surveying in the mountain tundra region of northern Sweden using the

EM38 manufactured by

Geonics Ltd. Photo: Kerstin

Lidén.

conductivity, and the greatest sensitivity in the horizontal mode is directly below the instrument, gradually diminishing with increasing depth (Fig. 20). The device can nevertheless measure down to c. 1.5m in the vertical mode and 0.75m in the horizontal mode. When it is measuring magnetic susceptibility its depth sensitivity is somewhat less, being greatest at a depth of c. 0.2m in the vertical mode below the instrument and c. 0.25m in the horizontal mode

(Fig. 20). The greatest depth penetration when measuring the magnetic susceptibility in the vertical mode is c. 0.5m (Geonics 1992:15).

Figure 20. Sensitivity of the EM38 in conductivity mode (left) and magnetic susceptibility mode (right). After Bevan (2004).

Early versions of the EM38 are also prone to drifting, as temperatures at the site change during the day (Bevan 2004:15), and it may be important to take

39

regular measurements with the instrument lifted into the air in order to check for the magnitude of the drift at particular times of day, at least when measuring susceptibility. Bevan (2004:15) made measurements in this manner before starting each survey transect in order to be able to correct for drift. The correction can be made in several ways (see Bevan 2004:15; dos

Santos et al. 2011), but failure to do so may mask important anomalies hidden in the data. An additional peculiarity of the EM38 when measuring magnetic susceptibility and operating in the vertical mode is that the data can include negative values. According to the response curve for the instrument, any magnetic layer located at c. 0.6m or below will cause the instrument to show negative values (Dalan 2008:4).

The unit for measurements of electrical conductivity is the millisiemens/m (mS/m), and in-phase measurements are expressed in parts per thousand (ppt) of the primary magnetic field. Electrical conductivity is the inverse of electrical resistivity, and consequently the two methods react to similar kinds of archaeological remains (Gaffney & Gater 2003:43), especially when comparing EM38 conductivity measurements carried out in the horizontal mode with earth resistance data (see Linford 2006:223 and references cited therein). As the EM38 does not require the inserting of probes into the ground, however, the instrument can be used in areas where ground conditions make electrical resistance or resistivity measurements impractical or impossible. This is exemplified by the electrical resistance surveys carried out by Fischer (1980) on Cyprus, where the electrodes had to be hammered some 0.4m into the soil at each measurement station to achieve sufficient contact (Fischer 1980:19). In such an environment an EM38 survey would have been less time consuming and more practical.

Archaeological features detectable by the EM38 include ditches, pits, stone constructions, brick structures, metals, etc. Several good introductions to surveying using Slingram instruments are available for interested archaeologists (see Bevan 1998; Gaffney & Gater 2003; David et al. 2008), and more technical descriptions of the method are available in McNeill

(1980a & b, 1999), Reynolds (1997) and Kearey et al. (2002).

2.3. Geoelectric methods and other geophysical prospection methods

There are several other geophysical prospection methods that have been used both frequently and infrequently for archaeological purposes. One of the more popular among these is electrical resistance or resistivity, but as this method has not been used in any of the surveys reported here, it will only be described briefly. Other less commonly used methods such as selfpotential (SP) (see Bevan 1990; Drahor 2004), induced polarisation (IP) (see

Schleifer 2002; Meyer et al. 2007), gravimetry (see Padin et al. 2012), VLF,

40

apart from the VLF metal detectors described above (see Persson 2005a;

Drahor et al. 2009) and seismic reflection and refraction (see Carson 1962;

Goulty & Hudson 1994; Andrén & Lindeberg 1997; Hildebrand et al. 2002;

Metwaly 2005; Hildebrand et al. 2007; Forte & Pipan 2008) will not be described here at all.

Earth resistance measurements are probably the most frequently used method of archaeological prospection apart from magnetometry, and as the historical review above has shown, electrical methods have been used within archaeological research since 1938 (Bevan 2000). Since this is such a popular method, it has been described extensively in the archaeogeophysical literature (see Scollar et al. 1990; Clark 1996; Gaffney & Gater 2003; Witten

2006; Campana & Piro 2009). The instruments basically measure the resistance of soils to an electrical current. As mentioned above, resistance is the inverse of conductance, and the archaeological features detectable by resistance measurements are similar to those detectable by conductivity instruments. Detectable features include ditches, graves and masonry constructions.

Electrical resistance surveying involves inserting metal electrodes into the soil and passing an alternating electric current between them. Two additional electrodes measure the electrical potential difference at every measurement point. The strength of the driving current (I), in amperes, is known and the electrical potential difference (V), in volts, is measured, whereupon the electrical resistance (R) can be calculated simply by dividing the voltage by the current, and the resulting resistance, in ohms, can be read from the instrument’s display. Resistivity in ohm/m may be calculated by multiplying the resistance by an array constant that is dependent on the geometry and spacing of the electrode configuration.

The electrodes can be arranged in different ways. If they are arranged in a straight line with the two current electrodes at the ends and the two potential electrodes between them, all spaced at equal distances, the configuration is called the Wenner array. Many alternative arrays exist, but the most popular for archaeological prospection is probably the Twin Probe array, which is described by Clark (1996:44) as a Wenner array divided in two. This was first used for archaeological prospection by Aspinall & Lynam (1970).

The distance between the probes in electrical resistance prospection is roughly equal to the depth of greatest sensitivity (see Scollar et al.

1990:321ff: Gaffney & Gater 2003:32). An electrode distance of one metre, for example, is excellent for detecting features at a depth of one metre. The

Twin Probe array with 0.5m electrode separation can measure to an approximate depth of 0.75m, however (Gaffney & Gater 2003:32). By extending the distance between the electrodes the depth of investigation can be increased, and this technique is used when carrying out vertical electrical soundings (VES) (Reynolds 1997:441). For a VES, the spacing between the electrodes is stepped up, thereby measuring to greater depths, while in a

41

related technique called resistivity tomography, or CVES, a large number of electrodes can be inserted into the soil along a profile and a switching box determines which ones are activated. The measured values, representing different soil volumes and penetration depths, are presented in a

“pseudosection”, which is a vertical representation of the soil resistance beneath the row of electrodes. For archaeological applications of this technique, see Aspinall and Crummet (1997).

The most commonly used electrical resistance instrument for archaeological prospection is the RM15 by Geoscan Research, as depicted, configured in the Twin probe array, in Fig. 21. This model was recently replaced by the RM85 (Walker 2012).

Figure 21. The Geoscan RM15. Photos: Siska Williams and Andreas Viberg.

2.4. Marine geophysical methods and geochemical prospection methods

Several marine geophysical methods, such as side scan sonar, multibeam sonar, magnetometry, etc. (see Lurton 2002; Lawrence et al. 2003) are used in archaeology, and their appropriateness for the visualisation and investigation of submerged Swedish archaeological remains have been demonstrated repeatedly (see Hjulhammar 2008, 2010). The methods are also recommended in the guidelines on marine archaeology published by the

Swedish National Heritage Board as a way of saving both money and time

5

.As this thesis explicitly deals with land-based archaeological geophysical survey methods, marine techniques will not be discussed any further, even though they are similar in their underlying physical principles to land-based methods.

5

http://www.raa.se/publicerat/9789172095250.pdf

42

The situation is similar for geochemical prospection methods (see Bethell

& Máté 1989; Crowther 1997; Heron 2001), e.g. soil phosphate analysis, which has been very popular in Sweden since the 1930s (see Arrhenius

1935). Even though soil phosphate analysis was used in one of the projects described in this thesis (Paper II), the main focus is on geophysical prospection methods.

2.5. The benefits of using an integrated approach

Since many of the geophysical methods used for archaeological prospection measure different physical parameters, an integrated approach using two or more methods within the same area can potentially provide a more complete understanding of the archaeological remains. This complementary nature of geophysical methods has been mentioned frequently in the literature (see Weymouth 1986; Clay 2001; David 2001;

Neubauer 2001; Persson 2005a; Kvamme 2006a; Kvamme & Ahler 2007;

David et al. 2008; Watters 2009). The National Park Service’s annual workshop on archaeological geophysics in the U.S.A. has on many occasions provided the possibility of surveying a site by multiple methods, and excellent European examples also exist where multiple methods have been used for the study of Roman towns (Neubauer & Eder-Hinterleitner 1997;

Buteux et al. 2000; Neubauer et al. 2002). The benefit of using more than one geophysical method is also exemplified by the survey of Sandbyborg

(Paper III), where the GPR and magnetometer provided complementary data, enabling a more complete and reliable geophysical and archaeological interpretation to be achieved. GPR provided information regarding the spatial layout of the houses within the fort and the magnetic survey provided information regarding the presence of possible hearths and pits within the houses and the fort as a whole. The magnetometer surveys on Gotland

(Paper IV) also benefited from information from earlier metal detection surveys, enabling the pinpointing of interesting buried remains of hightemperature crafts. The combined use of magnetometry and magnetic susceptibility has also been suggested for mountain tundra sites in northern

Sweden (Paper II), as this would enable the differentiation of in situ fired structures from magnetic stones.

Geophysical methods have frequently been part of wider archaeological prospection schemes, including their combination with remote sensing and aerial photography, satellite imagery, laser scanning, LIDAR surveys, geochemical prospection etc. (see Persson 2005a; Powlesland 2006, 2009;

Kvamme & Ahler 2007; Kvamme 2008; Gaffney et al. 2012). These other methods have an enormous potential of their own, but they become even more valuable as integrated parts of a wider archaeological approach to prospection.

43

The results of such multi-method surveys are commonly presented side by side and interpreted on a GIS platform along with aerial photos (see

Doneus & Neubauer 1998; Neubauer 2004), but alternatives do exist, and

Kvamme (2006a) has provided some guidelines to a more complete form of data integration in which measured values obtained by different geophysical methods can be statistically combined into something meaningful and as such strengthen the interpretation of the surveyed area.

44

3. Development and current use of archaeological geophysical prospection methods

The proton magnetometer and the computer will together, with

certainty, revolutionise field archaeology”. [Authors translation from Swedish] Gad Rausing (1971:110).

3.1. A brief history of archaeological geophysics

The Swedish archaeologist Hjalmar Stolpe used a form of primitive seismic investigation when searching for graves on the island of Birka,

Uppland, in Sweden in the 1870s (Stolpe 1888:6). He hit the ground with a cane and listened to the sound differences, as areas containing graves produced a much deeper sound. It can also be concluded from a letter written by Stolpe to Hans Hildebrandt, Director-General of the Swedish National

Heritage Board, in 1882 that Stolpe used this method in other areas as well

(see Arbman 1941:158, and references cited therein). The same method was also used by the English ethnologist and archaeologist Lieutenant-General

Augustus Pitt Rivers in his excavation in Cranborne Chase. He called it

“bosing” and used it to search for buried ditches (see Atkinson 1946:32;

Clark 1996:11; Gaffney & Gater 2003:13). This method and its limitations are also discussed by Aitken (1974:203).

3.1.1. Geoelectric methods

In November 1938 a Canadian geophysicist, Mark Malamphy, working for a company called Hans Lundberg Ltd., used the equipotential survey method (Sundberg et al. 1923:13) to search for a buried vault at the Colonial

Williamsburg historical site (Bevan 2000). Hans Lundberg, head of the company, was a Swedish geophysicist who emigrated to Canada in the

1930s (Lagerström 1968:609f) and became famous for his involvement in the discovery of Tepexpan man in Mexico in the mid-1940s (de Terra

1947:41, 1949:33ff; Fig. 22).

45

Figure 22. The Swedish geophysicist Hans Lundberg carrying out electrical measurements during the search for Tepexpan man in Mexico in 1947. Photo: ©

Juan Guzman. Picture published in Life magazine, March 31, 1947.

His interest in archaeology was shared by his younger brother Erik J.

Lundberg, a cultural historian and architect, who was involved in several

Swedish excavations: Vreta Monastery 1916-23, Nyköping 1925,

Söderköping 1925-36 and Skänninge 1928. He later became an executive at the Swedish National Heritage Board and professor at the Royal Institute of

Art in Stockholm (Meyerson 1982-1984; Berthelson 1969). The survey in

Colonial Williamsburg, which marked the actual beginning of archaeological geophysics, would, according to Bevan (2000:58), have been carried out by

Lundberg himself had he not been involved in a car accident.

The first archaeological-geophysical survey in Europe was carried out in

1946 at Dorchester-on-Thames, England, by Richard Atkinson (1953). He successfully used an electrical resistivity instrument, the Megger earth

tester, to locate ditches dug into gravel (Clark 1996:13). England was from this point on destined to take a very active part in the development and use of several new archaeological geophysical methods, but work connected with the early development and use of the methods was also carried out in

France, Italy and Germany (see Scollar 1959; Hesse 1966 and references cited therein; Carabelli 1967).

3.1.2. Magnetometry

In 1958 Martin Aitken and Edward Hall built a proton magnetometer that they used to locate buried pottery kilns at Water Newton in

Northamptonshire (Aitken 1958). Subsequent tests verified that the instrument was also able to detect pits and ditches (Aitken 1959).

46

Interest in electrical resistivity and magnetometry increased day by day, and during the following decades the methods were developed and refined to be better suited to archaeological applications. New types of magnetometer and new configurations such as the proton gradiometer (Aitken 1960), a proton magnetometer in variometer mode (Scollar 1961), the fluxgate gradiometer (Alldred 1964) and the optically pumped rubidium magnetometer (Ralph 1964; Bevan 1995b) were being used for archaeological purposes in England, Italy and Germany, for example. Later on, other types of magnetometer, the Overhauser magnetometer (Salvi

1970), the optically pumped Caesium magnetometer (Becker 1995) and the superconducting quantum interference device, or SQUID (see Chwala 2001,

2003), were used at archaeological sites. These magnetometers have improved sensitivity relative to the widely-used fluxgate magnetometer, enabling the detection of even more weakly magnetic archaeological structures in suitable environments.

3.1.3. Electromagnetic methods

Eight years after the Water Newton survey, Marc Howell (1966) developed the Soil Conductivity Meter (SCM), also known as “the banjo”, which, in spite of its name, actually measured the magnetic susceptibility of the soils at Cadbury Castle in England (cf. Tite & Mullins 1970). This was a transmitter-receiver electromagnetic instrument based on a Swedish invention, the Slingram (Werner 1937). This type of electromagnetic device was developed over the following years and led to instruments such as the

Geonics EM38 and EM31 (McNeil 1980b) and the SH3 (see Parchas &

Tabbagh 1978). Important developments, refinements and research projects were carried out by Tabbagh (1986a, 1986b), and the method was used as an archaeological prospecting tool in the USA as early as the mid-1970s (Bevan

1983).

The metal detector, an electromagnetic instrument similar to the above devices, was first used in 1945 to investigate an archaeological site in

Switzerland (von Bandi 1945), although tests with a simple version of a metal detector had been carried out in England even earlier (Passmore &

Jones 1930). These early surveys are described more thoroughly by Hesse

(2000).

The use of GPR in archaeological research began in the 1970s, and one of the earliest surveys was conducted in Chaco Canyon, New Mexico, USA, in

1974 (Vickers & Dolphin 1975; Vickers et al. 1976). This was followed by

GPR surveys in Philadelphia, USA (Bevan & Kenyon 1975), and at the pyramids of Giza, Egypt (Barakat et al. 1975).

47

3.2. Use of archaeological geophysical prospection methods in other countries

To gain a better understanding of the development of archaeological geophysics in Sweden it is necessary to form an overview of the popularity of the methods, their frequency of use, the opportunities for training and archaeologists’ attitudes towards the methods in other countries. As this information is difficult to come by and only a few publications containing such details are available (e.g. Bevan 2000, 2002; Gaffney & Gater 2003;

Herbich 2011; Visser et al. 2011; Lowe 2012) the author sent e-mail messages to all members of the International Society of Archaeological

Prospection (ISAP) 6 asking them to evaluate the discipline of archaeological geophysics in their respective countries. The answers enabled me to describe schematically the situation in a number of the countries concerned. This is admittedly an incomplete list, as information is missing for many countries, and the short descriptions should only be regarded as a brief guide to the international situation. One must also be aware that some of the information presented is anecdotal and represents a single researcher’s view. It has likewise been difficult to obtain updated information on the frequency of the use of archaeological geophysics in England, for example, and the data presented here may not be up to date in this respect.

3.2.1. Austria

According to the account by Neubauer (2001:29ff) of the development of archaeological geophysics in Austria, the first survey using electrical methods was carried out by Hampl and Fritish in 1959 (see Neubauer

2001:29 and references cited therein). The department of geophysical prospection, geodesy and photogrammetry at the University of Vienna offers training in archaeological geophysics and is currently collaborating with the recently established Ludwig Boltzmann Institute for archaeological prospection and virtual archaeology (LBI), located at the Central Institute for

Meteorology and Geodynamics in Vienna and engaged in research in the above-mentioned disciplines all over Europe together with partner organisations in Austria, Britain, Germany, Norway and Sweden

7

.

3.2.2. Denmark

The first archaeological survey using an Askania Gfz vertical intensity magnetometer was published in 1965 (Abrahamsen 1965). This was followed by several publications on archaeomagnetism. A book describing a

6 http://www.bradford.ac.uk/archsci/archprospection/menu.php?0

7

http://archpro.lbg.ac.at/

48

number of geophysical methods and containing information on many of the earliest geophysical surveys in the country was published in 1984 (Möller et

al. 1984). An estimated 4-8 surveys are currently carried out annually in

Denmark, but apart from a few lectures at Aarhus University each year, for example, there are no opportunities for specialised education in archaeological geophysics in the country (Bevan, B. pers. comm. 31.5.2012).

3.2.3. England

Gaffney & Gater (2003:23) report that c. 450 surveys were being carried out in England each year in the early 2000s. England is often considered to be the country where archaeological geophysics is most popular. This has partly been explained by the inclusion of archaeological geophysics in

Planning Policy Guideline 16, which briefly mentions geophysical surveys as a possible aspect of archaeological assessments

8

. This seems to have led to a wider awareness of the possibilities of using such methods among

English archaeologists and subsequently to their more frequent inclusion in archaeological projects (for an extended discussion of the situation in

England, see Gaffney & Gater 2003:22f). An evaluation of the use of archaeological geophysical prospection methods in northern England has been provided by Jordan (2009). Many of his conclusions are also valid for

Sweden and will be discussed below.

Archaeological geophysics is taught at many universities across the country and the M.Sc. in Archaeological Prospection-Shallow Geophysics at

Bradford University is one of the best-known university programmes in the world.

3.2.4. Finland

The first geophysical survey at an archaeological site in Finland was carried out in 1983 in a joint project between archaeologists from the

University of Turku and geophysicists from the University of Oulu (Lavento

1992 and references cited therein). A list of sites surveyed between 1983 and

1991 is presented by Lavento (1992). Since the first survey in 1983, the use of geophysical prospection in archaeology has been very limited and the methods are only used sporadically today (Pertolla, W. pers. comm.

26.5.2012). There is currently only one University in Finland that provides any training in archaeological geophysics and that is at the University of

Helsinki (Pertolla, W. pers. comm. 30.5.2012).

8 http://www.communities.gov.uk/archived/publications/planningandbuilding/ppg16

49

3.2.5. Germany

The early work by Scollar in the late 1950s marked the beginning of the discipline in Germany (Clark 1996: and references cited therein). Although accurate statistics about the use of geophysical prospection in Germany are difficult to find on account of the federal system, the country has many groups that are regularly involved in archaeological projects. Archaeological geophysics is carried out prior to 80% of all rescue excavations in the

German Palatinate region (Zeeb-Lanz, A. pers. comm. 28.5.2012), while in

Bavaria, south-eastern Germany, aerial photography has aided in the discovery of some 30,000 new archaeological sites, approximately 700 of which, c. 24 surveys annually, have been investigated using magnetic methods, electrical resistivity and GPR by the Institute of Geophysics,

University of Munich and the Bavarian State Department of Monuments and

Sites since the early 1960s (Fassbinder, J. pers. comm. 29.5.2012). For further information regarding the development of archaeological geophysics in Bavaria and Anatolia, see Fassbinder (2011).

There are currently two universities in Germany offering courses on geophysics with special focus on archaeology: the Ludwig Maximilian University in Munich and the

Christian-Albrecht University in Kiel.

3.2.6. Iceland

The use of geophysics on Iceland has been very sporadic and the first systematic surveys were carried out as late as 1999 (Horsley 2004).

Although the igneous bedrock of Iceland makes magnetic prospection more challenging, results have shown that magnetometry and resistivity combined can provide reliable evidence for archaeological structures such as long houses, pit houses, churches etc. (Horsley 2004:367ff).

3.2.7. The Netherlands and Flanders

The development of archaeological geophysics in the Netherlands has been reviewed by Kattenberg (2008), who mentions that the first two successful geophysical prospection projects in the country were carried out in 1968 and 1970 (2008:12). According to a recent report by Visser et al

(2011:27), a total of 273 surveys took place between 1996 ̶ 2010, of which

89 were electrical resistivity surveys and 38 magnetometer surveys, with between 10 and 30 surveys carried out annually since 1996 (Visser et al.

2011:28f, fig. 13). The only place which currently teaches archaeological geophysics in the Netherlands is the Institute for Geo- and Bioarchaeology in the Faculty of Earth and Life Sciences of the VU University, Amsterdam.

In Flanders the use of the methods has been quite marginal, but the frequency has increased since 2003. Research is currently conducted at the

50

University of Gent, which has been important for the development of archaeological geophysics in the area (Visser et al. 2011:35).

3.2.8. Norway

The use of archaeological geophysics is still very limited in Norway. The first survey was carried out in 1972, and recent statistics indicate that approximately 29 surveys took place between 2005 and 2009 (Anderson

Stames 2010:14f). Geophysical education is available at the Norwegian

University of Science and Technology, although there is currently no academic course that covers archaeological geophysics. Norway has recently been involved in collaboration with LBI (see above) and several surveys have already been carried out at Larvik, Vestfold

9

.

3.2.9. Poland

The first application of archaeological geophysical methods in Poland was a magnetic survey carried out on 19-21 April 1961 (

Dabrowski 1963;

Herbich 2011 and references cited therein ), and current estimates imply that

c. 10-20 surveys are carried out annually (Wroniecki, P. pers. comm.

26.5.2012). There are nowadays several active Polish groups carrying out geophysical prospection nationally and internationally. There is an introductory course available at the University of Warsaw for anyone interested in magnetic and electrical resistivity methods, and it is also possible to learn about the magnetic method at the University of Poznan

(Herbich, T. pers. comm. 30.5.2012).

3.2.10. Scotland

Geophysical measurements were first carried out in Scotland in 1972 by

Haddon-Reece and Clark

10

, and a paper dealing with magnetic measurements at the Standing Stones of Stenness, Orkney, was published by Clarke (1975).

Since then, archaeological geophysics has only been used sporadically.

Similar attitudes towards the usefulness of geophysics as noted in Sweden have also been mentioned by archaeological geophysicists in this country

(e.g. Ovenden 2010). The number of companies that do archaeological geophysics has increased, however, and an increased frequency of use can be noted as a consequence (Cuenca-Garcia, C. pers. comm. 10.7.2012).

Education in archaeological geophysics is currently available at the

University of Glasgow.

9

http://archpro.lbg.ac.at/larvik-vestfold/larvik-vestfold

10

http://sdb2.eng-h.gov.uk/Maps/britain.asp?206,31

51

3.2.11. Turkey

The first archaeological-geophysical survey in Turkey was carried out as early as in 1953 (see Drahor 2011:1 and references cited therein). The current frequency of use is increasing and the educational situation is improving thanks to the opening of the Center for Near Surface Geophysics and Archaeological Prospection at the Dokus Eylül University in 2004

(Drahor 2011:2f).

3.2.12. North and South America

The current level of use in this area is difficult to estimate, but one account provided by Bruce Bevan suggests that roughly 620 geophysical surveys were carried out between 1938 and 1999 (Bevan 2002), 6.6% of which were in Canada, 6.3% in Central and South America and the remaining 87.1% in the USA. This does not, of course, reflect the current level of use, but it still shows that an average of 223 surveys were carried out in North and South America between 1990 and 1997 (Bevan 2000:53).

Bevan also estimates that about 2% of all archaeological projects carried out in this area since 1980 made use of geophysical prospection methods (Bevan

2000:52). This figure has most likely increased, however, as current estimates for the state of Ohio alone suggests a frequency of 70-80 surveys a year (Burkes, J. pers. comm. 25.5.2012). The educational situation for archaeological geophysics in the USA is better than in many European countries, and several universities such as Minnesota State University

(Moorhead), the University of Denver, the University of Georgia and the

University of Arkansas provide courses in its use. In addition, a short course in archaeological geophysics has been taught annually by the National Park

Service (DeVore 2001).

Apart from the numbers provided by Bevan (2002), the current frequency of use of archaeological geophysics in Canada can only be conjectured upon, and it is estimated that these prospection methods are currently being used in less than 1% of the archaeological projects being carried out in the country

(Jeandron, J. pers. comm. 25.5.2012).

3.2.13. Australia

Archaeological geophysics was used in Australia for the first time in the

1970s (Connah et al. 1976), but the method was not employed very frequently after these initial trials (Lowe 2012). The frequency of use is now increasing, and several Australian universities and their departments of archaeology possess their own geophysical equipment and provide a few short courses in the subject for archaeology students (Lowe 2012).

52

3.3. Archaeological geophysics in Sweden

The historical development of archaeological geophysics in Sweden is described in Paper I and will be summarised below.

Apart from a few early metal detector surveys (Hagberg 1961; Nylén

1972), the first real geophysical prospection in Sweden with an explicit archaeological focus was carried out in 1977 at Frillesås in Halland (Fridh

1982:25f) when electrical resistivity and magnetometry surveys were performed before archaeological excavations at a site dating back to the Pre-

Roman Iron Age. The idea was to compare the results with the excavated features, but as the geophysical surveys were never properly georeferenced, no such comparison was possible. An additional early test survey using a proton magnetometer was carried out in the late 1970s/early 1980s in

Rappesta, Östergötland, but the results were invalid as plastic-coated metal measuring tapes had been used during the prospection and these had had quite a strong impact on the data (Hedman, A. pers. comm.4.11. 2008). After the initial tests, magnetometry was used sporadically in archaeological projects during the 1990s (see Johansson & Nylund 1990; Kresten &

Ambrosiani 1992; Kresten & Kresten 1994; Sträng 1995; Hedberg et al.

1996; Masters 1998; Wedmark 1999; Larsson 2000; Lorra et al. 2001;

Mercer & Schmidt 2001).

The end of the 1970s also saw the first use of GPR in Swedish archaeology, the first experiments being conducted at Ystad and Malmö in the province of Scania (Wihlborg & Romberg 1980). Successful surveys at the late Mesolithic site of Skateholm I was carried out during the spring of

1980 (Bjelm & Larsson 1980, 1984), and several other GPR surveys were subsequently carried out during the 1980s in search of graves (Burenhult

1984; Sundström 1985; Burenhult & Brandt 2002), settlement remains

(Löfroth 1987), a buried wooden causeway (Wahlström 1993) and buried walls (Carelli 2003), for example.

Very few electrical resistivity surveys have yet been carried out in

Sweden, as there are only a few examples available from the 1990s and

2000s (Sträng 1995; Hedberg et al. 1996; Larsson 2000; Dahlin 2001;

Misiewicz 2007; Sanmark & Semple 2008). Two of these (Larsson 2000;

Dahlin 2001) used the Continuous Vertical Electrical Sounding method

(CVES), producing electrical pseudo-sections (Aspinall & Crummett 1997), while in the other cases the data were collected either in single traverses or using what Gaffney and Gater (2003:34f) call an “area survey approach”, where data is collected at fixed intervals inside a survey grid.

Even though the electromagnetic Slingram instrument was invented in

Sweden as early as 1936, the first archaeological application was not until the early 1990s, when Kjell Persson of the Archaeological Research

Laboratory (ARL) at Stockholm University tested the EM31 and EM38 instruments by Geonics near the famous boat graves at Vendel, Uppland

53

(Persson & Olofsson 1995). The method later became an integrated part of the teaching provided at the ARL, and the EM38 has, along with soil phosphate analysis, been part of many undergraduate student projects and other geophysical surveys carried out there (e.g. Sanglert 1995; Kristiansson

1996; Lundberg 1997; Mattsson 1997; Stavrum 1997; Wåhlander 1997;

Englund 1999; Stålberg 2000; Vaara 2004; Persson 2005a; Sabel 2006;

Viberg 2007; Wesslén 2011). Another electromagnetic method used sporadically for archaeological prospection is VLF. This was used in a trial survey at Vendel during the 1990s, but the EM16 by Geonics, provided limited resolution, so that further surveys were cancelled (Persson 2005a).

Many laboratory measurements of the magnetic susceptibility of soil samples were carried out in Sweden starting in 1980 (Freij 1980), and the

MS2B by Bartington Instruments is regularly used at the Environmental

Archaeology Laboratory (MAL) in Umeå (see Paper 1 for further references), where field measurements are also conducted with the MS2D/F for educational purposes (Linderholm, J. pers. comm. 4.6.2012).

To date only one seismic refraction survey has been reported as a part of a

Swedish archaeological project. This was carried out on the island of Björkö in 1992 (Andrén & Lindeberg 1997) and was intended to assess the thickness of the cultural layers at the Viking Age site of Birka.

The use of dowsing as a mean of detecting buried archaeological structures has been advocated by several authors both in Sweden (e.g. Silver

1991, 2004; Andersson 1998) and internationally (e.g. Bailey et al. 1988), but many authors have been critical of the scientific evidence for the validity of the approach (e.g. Aitken 1959; van Leusen 1998; Lagerlund 2000;

Lindström 2000). As no scientific test has yet verified the reliability of the divining rod or been able to reveal its underlying principles of operation, the author strongly advises against its use as an archaeological prospecting tool.

The total number of geophysical surveys on archaeological targets between 1977 and 2008 was 280 (cf. Fig. 23). The statistics should not be considered complete, however, as it is impossible to be certain that every survey has been included in the list. Many geophysical surveys are never published properly, and in some cases they are carried out by a private contractor and might never become available to the public. There is also a risk that many unsuccessful surveys have remained unpublished, making the statistics even more uncertain. Note that the numbers presented in Fig. 23 differ from those in Paper I, as several additional surveys have been published and have become available since the publication of that paper.

This has not altered the overall trends, however, and the conclusions drawn in Paper I are still valid. Since the first Swedish archaeogeophysical investigations in 1977, the methods have been used at an increasing rate, with a clearly visible rise from the early 2000s onwards (Fig. 23). Some 137 surveys were carried out between 2005 and 2008, which corresponds to about 34 surveys a year. This number is probably rather high, since a large

54

number of surveys were carried out in 2006 and 2007, when several new archaeological prospection companies, such as the geophysical unit at the

Swedish National Heritage Board, were established in Sweden and a lot of test surveys were arranged. This establishment triggered an important renewal of the discussion regarding the use of archaeological geophysics in

Sweden, appropriate data collection routines and the visualisation and interpretation of geophysical data. Even though complete statistics for the period from 2009 to 2012 are not available, it is estimated that perhaps 10-20 surveys are now being carried out in Sweden each year. This does not include the metal detector investigations that form part of various archaeological projects, which, if included, would probably at least double the estimates for 2005-2008 (cf. Fig. 26).

When investigating the geographical distribution of these surveys, it is clear that the majority were carried out in a few Swedish provinces, namely

Uppland, Scania and Västergötland (Fig. 24), the vast majority (62) in Uppland, reflecting in part the geophysical proximity of the region to the

Swedish Geological Survey

(SGU), Uppsala University and

Figure 23. Graph showing the approximate number of archaeological geophysical prospection surveys carried out in Sweden between 1977 and 2008.

the Archaeological Research

Laboratory at Stockholm

University, who are responsible for many such surveys. Included in the Uppland statistics are surveys conducted at the Viking

Age site of Birka on the island of Björkö, which is one of the more intensely surveyed sites in Sweden.

Several surveys in Uppland have also been carried out by the prospection unit of the Swedish National Heritage Board and the archaeological prospection company GeoFysica, which is situated in the Stockholm area.

Scania, where many of the early GPR surveys were carried out, is in second place (35) and in addition to its geographical proximity to Lund Institute of

Technology, part of Lund University, has seen a lot of geophysical prospections being carried out at the Viking Age settlement and trading site

Uppåkra. The most recent prospections at the Uppåkra site, as at Birka in

Uppland, has consisted of large-scale GPR and magnetometer surveys carried out by LBI. In third place is the province of Västergötland, where 25 surveys have been carried out. Several surveys (14) have also taken place in

55

Södermanland. When analysing the geographical distribution, one must be aware that the data also reflect to some extent the development activity that has been going on in different parts of the country. The areas that have seen the least amount of prospection over the years are the provinces of northern

Sweden. This reflects not only their geographical location far away from any

Swedish institutions carrying out archaeological geophysical surveys, with the exception of the occasional surveys carried out by Luleå University of

Technology in Norrbotten, Malå Geoscience in Lapland and Radarteam

Sweden in Norrbotten, but also the relatively low property development activity, with fewer rescue excavations. It is interesting that the geographical distribution in many respects mirrors the distribution of population between the provinces (Fig. 25).

Figure 24. Graph showing the approximate number of archaeological geophysical prospection surveys carried out in Sweden between 1977 and 2008.

56

Figure 25. Population of Sweden by province. Data from Statistiska centralbyrån

(SCB) 2000 (cf. Fig. 31).

The most popular geophysical instrument is clearly the metal detector.

The relative ease with which anyone can obtain good results and the low cost of the instrument are factors that have contributed to its popularity. The problem, however, is that given the intense use of metal detecting, it is almost impossible to keep an accurate record of all the surveys ever carried out. My own list currently exceeds 400, and as the number performed on

Gotland alone has recently been estimated at c. 700 since 1972 (Fig. 26), the metal detector can be said to be indisputably the most popular geophysical instrument in Sweden. Included in the statistics for Gotland are 662 surveys, but it should be noted that many more surveys have been carried out but have been omitted from the diagram as the year of surveying is not indicated. The data on metal detecting do not include the many illegal metal detector surveys that are carried out each year, as the results of these are not published.

57

Figure 26. Graph showing the approximate number of metal detection surveys carried out on the island of Gotland between 1972 and 2010.7ote that the first yearly interval is larger than the following ones. Published with permission by 7y

Björn Gustafsson, 2012.

The use of other geophysical methods varies a lot depending on the province in question, but it is clear that, apart from metal detection, GPR is the most frequently used method for prospection (Fig. 27). The high frequency of other methods used in Uppland, and in Sweden in total, can be explained by the fact that the EM38 by Geonics has been used frequently by the Archaeological Research Laboratory during many years. The annual use made of four of the more common geophysical prospection methods in

Sweden is presented in Fig. 28.

Figure 27. Use of GPR, Magnetic Resistivity and other prospection methods such as EM surveys for archaeological surveying in the Swedish provinces of Uppland,

Scania and Västergötland. Data for Sweden as a whole presented in the lower right-hand picture.

58

The majority of the surveys performed in Sweden have used only one method or instrument, and when an integrated approach is adopted, the most popular combination is without doubt GPR and magnetometry. Other popular combinations are GPR and EM38 or earth resistance and magnetometry.

There are currently five Swedish companies or academic institutions working with archaeological geophysics: GeoFysica (GPR & Slingram),

SAGA-geofysik (Magnetometry), Modern arkeologi (GPR and Magnetometry), the prospection unit at the Swedish National Heritage Board (GPR and Magnetometry) and the Archaeological Research Laboratory (GPR,

Magnetometry and Slingram).

The educational situation for archaeological geophysics is currently very limited. The only place in Sweden where you can learn about the methods and have an opportunity to use them in archaeological projects at the bachelor’s or master’s level is the Archaeological Research Laboratory at

Stockholm University, but for several years the only instrument available for training purposes was the EM38. Several other universities and university colleges such as Uppsala University, Luleå University of Technology, the

University of Gothenburg, the Royal Institute of Technology and the department of applied geophysics at Lund University offer the possibility of studying applied geophysics, and occasionally also the possibility of working on archaeological case studies (e.g. Sträng 1995; Hedberg et al.

1996; Masters 1998; Wedmark 1999; Larsson 2000; Persson 2005; Karlberg

& Sjöstedt 2007).

Figure 28. Comparison between GPR, EM methods, magnetic methods and electrical resistance or resistivity methods in terms of usage in Sweden between

1977 and 2008.

59

3.4. Concluding remarks

When comparing Sweden to the other countries it is clear that the

Swedish situation is by no means unique. Most countries started to use geophysical prospection during the 1960s, but the continued use of the methods after these early surveys has varied greatly. In some countries such as England and Germany the methods are being used very frequently and are an integrated part of many archaeological projects. In England this has partly been explained by the inclusion of geophysical prospection in the PPG16 document, while in Germany the geological and pedological preconditions for geophysical prospecting are very suitable in many areas and this has resulted in very convincing results, contributing to the popularity of the methods. The character of the archaeological remains in these countries, including the often manifest remains dating from the Roman period, may also have been a possible factor contributing to the success of the methods.

These countries also provide education in archaeological geophysics, and this has definitely had a positive impact on the acceptance of the methods. In other countries the process has been slower. In the U.S.A., for example, acceptance and use of the methods has increased gradually over many years, while in the other Nordic countries, in Scotland and in Australia the process of adopting archaeological geophysics has been slow. This could be explained in some cases by unsuitable geology and pedology, making geophysical surveys ineffective or problematic in those countries, while in other cases the character of the archaeological remains, perhaps with a predominance of postholes, hearths and pits, may have affected the success of geophysical surveys. These factors and their implication for the likely success of archaeological geophysics in Sweden will be investigated in greater detail in the following chapters.

60

4. Geological and pedological conditions in

Sweden and their implication for archaeological geophysics

4.1. Geology, pedology and land use

The bedrock in Sweden can be divided into three principal formations.

The first is the parent rock (Sw. urberget) which is a part of the larger Baltic

Shield (Fig. 29), also including parts of Norway, Finland and Russia. This is roughly 3 billion–500 million years old (Lundqvist & Bygghammar

1994:16ff; Lundqvist 2000:29ff). This accounts for the major part of the

Swedish bedrock falls and includes many different rock types, although predominantly igneous and metamorphic rocks such as granites and gneisses. The second formation, the Scandinavian Caledonides, situated in the north-western part of the country (Fig. 29), was created between 510-400 million years ago when two tectonic plates, Baltica and Laurentia, collided

(Stephens et al. 1994:22; Lindström 2000a:209ff). As a consequence of this collision the geology of the area is highly diverse, including rock types such as quartzite, limestone, sandstone, dolomite, amphibolite, greenschist etc.

The third bedrock formation consists of sedimentary rocks such as limestone and sandstones created mainly during the Phanerozoic period, from c. 530 to

1.4 million years ago (Norling 1994:25; Lindström 2000b:230ff). This sedimentary bedrock is found on the islands of Öland and Gotland in the

Baltic Sea, but also exists in the Swedish province of Scania and in smaller areas elsewhere in Sweden (Fig. 29).

The magnetic susceptibility of Sweden’s various rock types is highly diverse, and the magnetic variations caused by the bedrock can be mapped by means of aerial measurements with a magnetometer despite the overlying soils and water (Fig. 30). The strong positive anomalies clearly visible in the northern, western and south-eastern part of the country (Fig. 30) are most often caused by the presence of the highly magnetic mineral magnetite in the bedrock (Eriksson & Henkel 1994:76ff). As the sedimentary bedrock on

Gotland has a very low magnetic susceptibility, the high readings visible on the map may be surprising. These are created by the more magnetic parent bedrock situated beneath the sedimentary rocks at depths of between 300m and 800m (Erlström et al. 2009).

61

Figure 29. Principal bedrock formations of Scandinavia: Scandinavian

Caledonides (dark green), Baltic Shield (light green, orange, pink and yellow) and the sedimentary bedrock outside the Scandinavian Caledonides (blue). © Swedish

Geological Survey (SGU).

Figure 30. Magnetic anomaly map of Sweden with the geomagnetic reference field subtracted. The data colour scale is: -464 dark blue to +1058 dark red. © Swedish

Geological Survey (SGU).

62

The quaternary deposits are dominated by glacial till, which covers roughly 75% of the country’s area (Fredén 1994:105; blue and purple in Fig.

32). This is a highly heterogeneous material that broke loose from the bedrock beneath the continental ice sheet during the various glaciations. The ice sheet would often transport this material far away from its area of origin, and consequently one can find igneous rocks deposited by the retreating ice margin on the island of Gotland even though such rocks are not native to this sedimentary environment. The melting ice has also created landforms such as eskers, moraines and drumlins which are highly visible in the landscape.

Clay and deposits of sand and gravel are other commonly occurring soils in Sweden (yellow and orange in Fig. 32). These are found predominately in the central parts of southern Sweden and along the east coast of northern

Sweden. The clay was deposited both during the melting of the ice sheet and after the glaciations due to the flushing and redeposition of sediments, and these types are distinguished as glacial and postglacial clays respectively.

Clays can be found in areas previously covered by water and are frequently used for farming because of their high fertility.

Peat soils with a high percentage of organic matter are to be found in many areas of Sweden but to an increased extent in the northern parts (brown in

Fig.32). Sweden also has large areas in the northwest and in the eastern and western parts of southern Sweden where the soil cover is very thin, thus partly or completely exposing the underlying bedrock (red in Fig. 32).

In terms of land use the Swedish landscape is dominated by forest, covering roughly 53% of the total area

(see Table 4 and Fig. 33). The figure can in fact be even higher locally, as shown by the statistics for the counties of Gävleborg, Västernorrland and

Värmland (cf. Fig. 31). The average proportion of arable land in Sweden as a whole is only 7.6%, but in Scania it reaches 47.6%. The land-use map of the Swedish counties presented in Fig.

33 reveals the dominance of marshland and open heathland in the northern parts of the country.

Figure 31. The 21 counties of

Sweden.

63

Figure 32. Quaternary deposits in Sweden.

64

Figure 33. Land use in Sweden, by county (cf. table 4).

65

Table 4. Land use in Sweden (data in percent), by county. Data from

Statistiska centralbyrån (SCB) 2005 (cf. Fig. 33).

66

4.2. Implications for archaeological geophysics

Geology, pedology and land use factors are important if one is to understand the pitfalls and possibilities entailed in carrying out geophysical measurements in Sweden. Given that most prehistoric remains found during archaeological excavations in Sweden have a small physical contrast, a knowledge of the properties of the surrounding soils is vital when predicting the likely success of geophysical prospection methods.

If the soil cover is thin, as in the mountain tundra region of north-western

Sweden, magnetic bedrock can affect magnetometry results and soil magnetic susceptibility measurements, effectively obscuring signals from the often weakly magnetic archaeological remains (Viberg et al. 2009; Paper II).

This effect can nevertheless differ from site to site, because of the geologically diverse nature of the Scandinavian Caledonides, which means that an initial assessment of the magnetic effect of the bedrock should be carried out before planning any geophysical survey. As suitable maps are provided by the Swedish Geological Survey (SGU), such an evaluation could be made reasonably quickly. Most types of sedimentary bedrock are suitable for archaeological geophysics (see Gaffney & Gater 2003:79; David

et al. 2008:15) and the limestone bedrock on Öland and Gotland provides a good example of an environment in which magnetic measurements would be appropriate and even recommendable (cf. Papers III and IV).

Because of the heterogeneous nature of glacial till, magnetic rock fragments deposited in the landscape by the melting ice sheet could have a harmful effect on geophysical data, causing spurious anomalies that can overshadow signals from the archaeological remains of interest (see Aspinall

et al. 2008:174f; Paper I). These anomalies might also be misinterpreted as archaeological features. When used for the construction of prehistoric structures such as the stone foundations of houses, however, such anomalies could be an asset and make detection much easier (cf. Paper IV).

The attenuation of radar waves in highly conductive soils has been discussed in previous chapters, and in this respect the clay soils of Sweden could be potentially problematic for GPR surveys if the archaeological features are situated at some depth. The presence of water is also important, as this reduces the penetration depth of the GPR. Areas with both clay soils and high precipitation during some parts of the year, 11 such as some parts of south western Sweden, could be potentially problematic.

Regarding land use, the most obvious difficulty for geophysical surveys is the high density of forests and forest plantations. It is difficult and timeconsuming to carry out surveys in such areas and there are often problems in covering a large enough area or carrying the survey out on a strictly regular grid. It is also important to recognise the time factor, too, as this will

11 http://www.smhi.se/polopoly_fs/1.4159!image/p117.png_gen/derivatives/fullSizeImage/p117.png

67

increase the cost of the survey. The situation is similarly complicated by the fact that tree roots are often visible in GPR data, thus limiting the possibilities for detecting archaeological remains in such areas (see Fig. 34).

A magnetometer should not be affected by roots, however (Kvamme

2001:360), and can be a practical solution if the increased time required for data collection and the negative effect of the trees on data quality are acceptable. The local topography with uneven surfaces that is common in mountainous and forested areas can also have a negative impact on the data and/or prevent the use of GPR.

Figure 34. The effect of root systems on GPR data as visible in the top and mid-part of the data from Fornsigtuna. The time slice was taken at an approximate depth of

1dm (1ns thick). Data collected with 0.1m transect spacing.

The farmlands of Sweden are very suitable for measurement purposes.

Their large open areas, at least during some seasons of the year, provide practically optimal conditions for many geophysical methods (e.g. magnetometry, cf. Paper IV). One must remember, however, that a large proportion of the country’s farmland is situated on clay soils, which may restrict the applicability of GPR in certain circumstances.

If the prehistoric remains are buried deep in the soil, detection by geophysical methods could be challenging, and consideration should be given to the information on the depth sensitivity and possible penetration depths of different geophysical instruments already provided in the previous chapters. One environment where one could expect deeply buried features is wetlands and peat bogs, but the only method that is known to produce usable results here is GPR (Clarke et al. 1999; Conyers 2004a:46). Only a few successful examples exist of GPR surveys aimed at locating archaeological remains in Swedish peat bogs (e.g. Henkel 2006:84), but since a significant portion of the country is covered by peat, GPR could be recommended for future studies.

68

5. Data collection, interpretation and the character of Swedish archaeological remains

5.1. Sampling density, data collection and georeferencing survey grids

The importance of the density of the data collected cannot be underestimated, as the sampling density governs the detectability and interpretability of archaeological features of various sizes. The inline sampling density is the density of data points along the survey transects, which is governed by the speed with which the instrument can collect data.

This is rarely a problem as, apart from electrical resistance and magnetic susceptibility instruments such as the MS2D, where any increase in inline sampling density is a factor to consider carefully, modern instruments are generally very fast. The crossline sampling density, however, i.e. the distance between the transects, is very important, as it has far-reaching implications for the speed of the survey, and thereby the costs, and also on the sizes of archaeological features detectable by the geophysical method.

The benefits of a small crossline distance and its implications for the detection of small archaeological features have been discussed by many researchers (e.g. Clark 1996:81; Neubauer 2002:140f; Gaffney & Gater

2003:70, 95; Grasmueck et al. 2003; Leckebusch 2003:216; Dennis 2004;

Kvamme 2006b:214; Aspinall et al. 2008:110f; Novo et al. 2008; Ernenwein

& Hargrave 2009:79; Jordan 2009) and it has been shown beyond reasonable doubt that an increased sampling density will enable the detection of smaller features or features buried more shallowly in the soil. This important conclusion has led the geophysical unit of the Swedish National Heritage

Board to recommend 0.25m as a standard distance between GPR profiles in order to detect many of the smaller archaeological features common in

Sweden (e.g. Trinks 2005). There are various ways of thinking about the necessary density of measurements. One rule of thumb is that in order to be able to interpret the data correctly, the sampling density must be governed by the expected size of the buried archaeological features. The archaeological feature needs to be sampled at least twice and preferably even more

(Kvamme 2003:437, 2006b:214f). This is obviously true for both the inline

69

and crossline sampling distance, but as data are as a rule collected in a much denser fashion inline than crossline, it is primarily the crossline distance that is of concern in this respect. One can expect as such to be able to detect and correctly interpret archaeological features equal to or larger than twice the size of the chosen crossline sampling distance. For archaeological purposes this means that if an archaeological feature is expected to be two metres in diameter the survey transects needs to be one metre apart at most. Similarly, for a 0.5m archaeological feature a crossline sampling distance of 0.25m is needed. Another rule of thumb states that the distance between transects should be less than the distance to features located below the instrument

(Bevan, B. pers. comm. 23.7.2012). This implies that if an object is situated at a depth of 0.5m below the instrument, a crossline sampling spacing of less than 0.5m would be desirable.

This theoretical view on sampling density obviously needs to be nuanced and weighed against the aim of the survey. In many cases 100% detection of every archaeological feature may not be necessary, and if the purpose of the survey is to delineate possible prehistoric activity areas from areas not containing such remains, a wider transect spacing may be used. The sampling density, both inline and crossline, per unit area is as such a very important measure to consider. Obviously very dense measurements have an impact on the cost of the survey, as halving the crossline distance could double the time spent in the field and increase the costs of the survey. When surveying for previously known features in restricted areas this may not present an economic limitation, and the cost of a dense survey may be considered reasonable. If the purpose is to survey larger areas where no archaeological remains are previously known, however, a different approach is needed. One way of dealing with this could be to survey a larger area with a wider sampling density. When the data have been analysed, one can carry out more detailed measurements in limited parts of this area. This approach and the whole question of how to design a geophysical survey have been discussed by Sambuelli and Strobbia (2002).

Another way of increasing the sampling density without increasing the costs is to use multiple sensors simultaneously or mounted next to each other in arrays (Fig. 35). These arrays can also be motorised for even faster data acquisition. The sampling density can then be adapted according to the desired resolution, common crossline distances, i.e. typical distances between the mounted probes in magnetic surveys, being 0.5m and 0.25m.

One of the earliest examples of a successful array investigation was the 1997 magnetic survey at Uppåkra, situated between Malmö and Lund in southern

Sweden, by Lorra et al. (2001). The survey was carried out using a Foerster fluxgate gradiometer array system consisting of five probes with very dense inline and crossline sampling densities of 0.05 x 0.2 m. The sampling densities of two of the currently available motorised GPR arrays are 0.075m x 0.075m (3D radar) and 0.08m x 0.08m (Malå MIRA system), thus

70

providing an unsurpassed data density combined with fast data acquisition for large-scale prospecting operations. The arrays are also dependent on broad, smooth expanses of terrain in order to operate well, however, and are most suitable for use in flat, open areas such as farmland. A single instrument survey as such will always be needed, especially in a country like

Sweden. A motorised system for electrical measurement is also available

(see Dabas 2009). It should be noted that the cost of buying any motorised array system is high and only a few are available for archaeological prospection in Europe. The price per unit area and the resulting data density of a motorised survey must also be compared with the corresponding figures for a single instrument survey when deciding between the two. When collecting high density data over large areas with motorised systems the time needed to interpret the collected data also increases substantially, as the resulting data files are very large.

An additional benefit of arrays mounted onto rigid frames is that noise introduced into the data during magnetometer surveys, i.e. heading errors, can be kept to a minimum, thus producing clearer pictures of the subsurface.

Heading errors are introduced into the data when the magnetometer cannot be kept in the correct upright position or when it is carried in different ways

(see Aspinall et al. 2008:120). This can be caused by strong, buffeting winds, for example (cf. Papers IV and V). One must be aware, however, that noise may also be introduced by the motorised transport and the frame itself.

Arrays, both motorised and non-motorised, have been used in Swedish archaeology for several years and have produced excellent magnetic and

GPR data (see Lorra et al. 2001; Gustafsson & Alkarp 2007; Trinks et al.

2007, 2010b; Biwall et al. 2011; Trinks & Biwall 2011b).

General recommendations regarding the sampling density of geophysical surveys performed using different methods have been published in England

(e.g. David et al. 2008). These guidelines suggest a magnetometer sampling spacing of 0.25 x 1m for evaluation purposes and a sampling spacing of

0.25m x 0.5m for a more thorough characterisation of the archaeological remains. Similarly, the suggested sampling distance for GPR surveys is

0.05m x 1m for evaluation purposes and 0.05m x 0.5m for more detailed measurements, and that for surveys with EM instruments such as the EM38 is 1m x1m for evaluation and 0.5m x 1m or 0.5m x 0.5m for characterisation

(David et al. 2008:8). These guidelines function well as general recommendations, but with regard to the Swedish situation and the nature of the Swedish geology, pedology and the sizes of the commonly occurring archaeological remains it is clear that the sampling densities need to be adjusted according to the desired resolution and archaeological questions to be answered.

71

Figure 35. Example of arrays used in archaeological prospection. Using the MIRA system manufactured by Malå Geoscience (left; photo: MALA internt,) and the 3channel gradiometer system manufactured by Foerster mounted on a non-magnetic cart (right; photo: Siska Williams).

Since the sizes of Swedish archaeological features determine the sampling density strategy, I made a small review of the lengths and widths of 984 hearths reported at archaeological sites all over the country to try to establish their mean size. The information was compiled from excavation reports were selected at random from the web pages of various Swedish National Heritage

Board excavation units (UV) (Lindman & Stibeus 2000; Streiffert 2000;

Ängeby 2001; Strömberg 2001; Andersson et al. 2003; Gustafsson et al.

2004; Grön & Sander 2005; Häringe Frisberg & Renck 2005; Karlenby et al.

2005; Seiler 2005; Björck & Appelgren 2006; Edenmo 2006; Kraft 2006;

Lindman 2006; Lindkvist & Westin 2006; Molin 2006; Westin 2006a, b;

Edlund 2007; Granlund & Karlenby 2007; Holm & Thorsberg 2007;

Stenvall 2007; Sundberg et al. 2007; Bergold 2008; Björck et al. 2008;

Dunér & Vinberg 2008; Edlund 2008; Hamilton & Östlund 2008; Hamilton

et al. 2008; Hagberg 2009; Helander 2009; Bäck et al. 2010; Eriksson &

Häringe Frisberg 2010; Stenvall 2010)

12

. For northern Sweden, where largescale excavations are rare, the sizes of hearths interpreted and compiled by

Hedman (2003) were added. The dates of the hearths ranged from the

Mesolithic up to modern times. The median sizes of the hearths (see Table

5), indicate that the sampling density when searching for hearths should be c.

0.45m, and that a crossline distance exceeding 0.5m will not succeed in mapping the median hearth and will render a substantial number of hearths difficult to interpret correctly or miss them altogether. This recommendation is also valid for EM measurements made using the EM38.

12 http://www.arkeologiuv.se/cms/arkeologiuv/publikationer/rapporter.html

72

Table 5. Mean and median lengths and widths, in metres, of hearths discovered during excavations in Sweden (see text for further details).

Average length

Average width

Median length

Median width

Hearth southern and middle Sweden

1.05

Hearth northern Sweden

Hearth entire country

1.37

1.17

0.95

0.96

0.96

0.96

1.30

1.20

0.86

0.90

0.90

Apart from the size of the archaeological feature, the sampling density for a GPR survey is dependent on the frequency of the antenna, in addition to which, in order to avoid any aliasing or interpolation of the data, the sampling distance in all directions should in optimal circumstances and when aiming at a true 3D resolution be no more than a quarter of the wavelength in the soil (see Novo et al. 2008, 2010). For a 500MHz antenna this is only c. 5cm. This resolution will probably be obtainable in future versions of motorised GPR arrays and will most certainly be feasible for detailed research surveys of smaller areas, but despite the impressively clear data and the possibility to detect very small archaeological features, it is perhaps too optimistic to think that this will become a standard survey procedure. The general recommendations that the transect spacing should not be larger than 0.5m when aiming to detect larger structures using

400MHz and 500MHz antennas (see Neubauer et al. 2002; David et al.

2008) can be considered reasonable. When aiming at smaller structures this distance should be smaller, possibly 0.25m, the spacing suggested as a standard by several authors (Leckebusch 2003; Trinks 2005). As noted above, however, this must still be governed by the questions to be answered in the research.

To exemplify the effect of different crossline sampling spacings on GPR data, a small simulated comparative study was performed (Figs. 36 and 37).

The first example is from the Dominican convent in Sigtuna (Paper V; Fig.

36), obtained using crossline distances of 0.25m, 0.5m and 1m, the results for 0.5m and 1m being simulated by removing profiles from the original data set. The time slices in all the examples below are 2ns thick and are situated at depths of 18.7ns (left-hand slices) and 25.0ns (right-hand slices).

The simulated data with 1m transect spacing are clearly substandard and fail to image many of the walls correctly, and many other features visible in the two denser examples, such as the electric cable and many of the smaller walls, are completely missing from the data or have an erroneous spatial distribution. A dramatic improvement is visible when reducing the transect spacing to 0.5m. At this point the electric cable can be interpreted correctly and the majority of the walls are visible. When the spacing is reduced further to 0.25m, the image sharpens up even more and it becomes possible to interpret some of the smaller walls, such as the one situated in the top right

73

corner of the cloister garth. The lavatorium, extending from the southern range into the cloister garth, is also easier to interpret.

Figure 36. Effects of different crossline transect spacings, as exemplified by the data set from the Sigtuna Dominican Priory (Paper V). The slices are 2ns thick and the lefthand pictures originate from a depth of 18.7ns and the right-hand pictures from a depth of 25ns. Data collected with 0.25m transect spacing (top), a simulated data set with 0.5m transect spacing generated by removing profiles from the original data set(middle), and a simulated data set with 1m transect spacing generated by removing further profiles from the original data set (bottom).The size of the survey area is 50 x 25m and references points are indicated by letters a and b. 7ote the electrical cable visible in the middle and bottom lefthand picture.

The effect of crossline spacing can also be exemplified by the GPR results from a Migration Period house terrace at Fornsigtuna, north of

Stockholm. Seasonal measurements were conducted at this site in order to evaluate the impact of soil moisture on GPR data (see section 3.2.2.). The house terrace, no. 6, had previously been excavated during the 1980s

(Damell 1991:48f; Fig. 13) and the GPR survey was carried out in August

2010 with a profile spacing of 0.1m. The time slices are 1ns thick and are situated at a depth of c. 3ns (Fig. 37). By removing profiles from the original data set, the effect of an increase in transect spacing can be simulated. The spacings presented below are 0.1m (Fig. 37: bottom picture), 0.3m (Fig. 37:

74

second picture from the bottom), 0.5m (Fig. 37: second picture from the top) and 1m (Fig. 37: top picture).

The data with 1m transect spacing is not informative enough and its interpretation would lead to many erroneous conclusions. The most dramatic improvement is visible when the crossline spacing is reduced from 1m to

0.5m, but a smaller spacing is necessary if correct assumptions are to be made regarding the size of the house as well as the correct identification and interpretation of some of the possible postholes. The tree roots in the top centre of the slices also become clearer and the smaller spacing allows these to be differentiated from any archaeological features visible in the data. The curved wall of a longhouse along with one of the gables (in the left of the picture) becomes clearer with a decrease in transect spacing.

Figure 37. Effect of different crossline transect spacings, as exemplified by the data set from Fornsigtuna. The slices are 1ns thick and originate from a depth of 3ns.

Data collected with 0.1m transect spacing (top), a simulated data set with 0.3m transect spacing generated by removing profiles from the original data set (second from top), a simulated data set with 0.5m transect spacing generated by removing further profiles from the original data set (second from bottom), and a simulated data set with 1m transect spacing generated by removing still more profiles from the original data set (bottom). The size of the survey area is 37 x 11m and references points are indicated by letters a and b.

75

The examples prove that the crossline transect spacing has an impact on the interpretability of the data. When investigating a large stone or brick structure a crossline sampling spacing of 0.5m or less is recommended in order to produce interpretable data. If the smaller walls and features of the convent building are to be mapped correctly, a spacing of 0.25m is necessary, for the smaller and more elusive remains of the longhouse 0.25m, and an even smaller transect spacing is needed if small postholes are to be mapped and interpreted correctly.

Equally important is the routine of georeferencing the data grids collected. This is crucial if subsequent excavations are to target geophysical anomalies. In order to achieve the desired high precision it is recommended that the grids that are surveyed should be tied to the national coordinate system used in Sweden, SWEREF99, by means of either a total station or an

RTK-GPS, with which an accuracy of 0.001m is achievable in favourable circumstances. These high precision instruments could also be used when establishing the survey grids. The use of low-cost handheld GPS systems, sometimes with an accuracy of several metres, or routine of georeferencing grids by means of fixed landmarks using measuring tapes, is in most cases not recommended. The quality of the data is also dependent on accurate positioning within the survey grids, in that grids that are georeferenced properly could still suffer from discrepancies caused by the surveyor failing to walk in straight lines or to maintain a constant pace. The use of guide ropes placed on the surface at regular intervals within the grid is thus recommended in order to increase the quality and precision of the data as collected.

5.2. Interpretation

After collection, the data should be filtered properly before the process of interpretation begins. There are several basic overviews of the data processing of magnetic (Aspinall et al. 2008) and GPR data (Conyers

2004a), and the art of interpreting geophysical data has been described by several authors (see e.g. Gaffney & Gater 2003; Conyers 2004a, 2006b;

Aspinall et al. 2008; David et al. 2008). The method for interpreting GPR data is presented by Poscetti et al. (2011) and should be regarded as a stateof-the-art description of the process. Different parts of the interpretation process as described in this paper could also be adopted when interpreting results obtained using other geophysical measurement techniques. As this is a fairly well covered subject it is not my purpose to present a thorough explanation of interpretation, but some crucial points need to be mentioned.

A geophysical data set can be interpreted in different ways and on different levels. At the lowest level, geophysical maps are viewed and the spatial distribution of the anomalies is described, e.g. as linear or round

76

(visual inspection level). At this level of interpretation only certain very basic information can be extracted from the data, and the data may need to be interpreted further to become meaningful to archaeologists. The next level is that of technical interpretation (parametric interpretation level), which according to Bevan (2003:3) will consider such parameters as the location of the source of the anomaly, the depth of the feature, the estimated mass of the material and an evaluation of what material is causing the anomaly. The technical interpretation could also have the purpose of singling out the anomalies that are most likely to be caused by geological features, or by shallow metal artefacts when these are unwanted. Following or accompanying such a technical interpretation, the archaeological evaluation or interpretation of the data should be added (archaeological interpretation level). This interpretation would ideally be carried out by someone with a thorough knowledge of archaeology and of the geophysical method or methods employed, possibly in cooperation with site specialists/excavators who have a comprehensive knowledge of the archaeological conditions at the site. The archaeological interpretation must depend on a thorough knowledge of earlier land use at the site as obtained by evaluating historical, geological and pedological maps and maps of previous archaeological investigations and an evaluation of the impact of agriculture. This interpretation and comparison of data from different sources is best carried out using a GIS platform (see Neubauer 2004), as the resulting georeferenced interpretation maps can provide valuable information for the planning and execution of subsequent excavations.

5.3. Swedish archaeological remains

The majority of the prehistoric remains found during excavations in

Sweden are pits, postholes and hearths, and as they are often quite small and have a weak physical contrast with the surrounding matrix, their identification by means of geophysical measurements can be problematic.

This has also been a problem in France, where, precisely because of such limitations, archaeological geophysics has recently been recommended only as a complement to more traditional archaeological preinvestigation methods such as trial trenching (Hulin & Simon 2012). A few examples are presented below of typical remains that might be encountered during archaeological excavations and their likely detectability using geophysical methods.

5.3.1. Graves

Graves are very common structures encountered during many archaeological excavations and have from the early days of geophysical prospection in Sweden been valuable discoveries (Bjelm & Larsson 1980,

77

1984). Despite occasional comments to the contrary (e.g. Alkarp & Price

2005:268), this very heterogeneous category is notoriously difficult to map

(see David et al. 2008:15), but successful surveys have been carried out in the U.S.A. primarily by using GPR, electrical conductivity and electrical resistance to map graves in both historical and modern cemeteries (see

Bevan 1991; King et al. 1993; Conyers 2006a; Fiedler et al. 2009).

Magnetometer surveys in England have in a few cases also been successful for mapping prehistoric Roman and Anglo-Saxon inhumations (see Linford

2004). As stated by Gaffney and Gater (2003:136), it is important to understand that geophysical techniques rarely detect the actual buried body, but rather graves are mostly identified on the strength of constructions associated with the burial. Historical graves are often located on the grounds of interruptions in soil layers visible in the GPR radargrams, which are in many cases evidence of grave shafts dug through these soil strata and thus provide indirect evidence of the actual burial. A few surveys have been carried out in Sweden with the explicit purpose of mapping graves, the earliest successful ones having involved the use of GPR for detecting

Mesolithic graves at the cemeteries of Skateholm I and II in Scania (Bjelm &

Larsson 1980, 1984). As the detection of graves is dependent on a contrast between the fill and the surrounding soils, it becomes clear why geophysical methods are rarely able to map such features. As the grave is dug and the body placed in it, the same soil that was removed from the hole is often used for refilling it and only a small physical contrast can be expected, unless there was a marked stratification in the natural soil. In recent years a magnetometry survey aimed at detecting graves in the Viking Age inhumation cemetery at Broby bro (RAÄ 42) in central Sweden was unsuccessful, highlighting the problems involved (Andersson 2011). A more successful survey, however, was carried out at the Franciscan Convent in

Krokek, mapping graves dating from medieval times and onwards using a

GPR (Trinks et al. 2008a). Clear evidence of graves aligned in an east-west direction is visible in the time-sliced data, acting as a valuable complement to the already existing map of burials at the site.

GPR and magnetometer surveys in Uppåkra, Scania, as also at Borre,

Norway, have been successful at identifying an overploughed grave mound

(see Trinks et al. 2010a; Biwall et al. 2011). In the Borre survey detection was possible because the stones covering the central part of the grave were magnetic. In addition, the overploughed mound was surrounded by a small ditch detectable by both the magnetometer and the GPR.

It has been shown on Öland that Early Iron Age cemeteries typically consist of circular stone settings, predominantly of limestone, while the outer kerb ring of such a grave often consists of granite or diorite stones (Hagberg

1979:14). As these rock types generally have a high magnetic susceptibility, they would most likely produce a pattern detectable by a magnetometer.

Deposited artefacts or coffin nails could also aid in the discovery of certain

78

kinds of burials when using a metal detector (David et al. 2008:16). When searching for cremation burials, which are common in Sweden, the graves could produce a magnetic response identifiable by a magnetometer if the cremation had occurred in situ. A strong response should also be produced in the GPR data if the grave contains void spaces.

Larger mounds have been surveyed by GPR on a few occasions in

Sweden and Norway

13

(e.g. Persson & Olofsson 2004) and while such surveys are feasible, it is important to correct the resulting data for tilt and elevation of the antenna in any direction if the results are to give a correct representation of the internal features of the mound (see section 2.2.3. and

Goodman et al. 2007).

5.3.2. Prehistoric settlement remains

Prehistoric settlement remains such as postholes, pits, ditches, middens, long houses, stone foundation houses etc. have been successfully detected by geophysical methods on many occasions. While some of these features are easier to detect than others, some are small and only provide a weak physical contrast with the surrounding soil matrix. One small problematic archaeological feature is the post hole. These come in a variety of sizes and in their basic form may consist merely of a post inserted into the ground.

This post may or may not be burned or lined with stone, and this means that their detectability may vary greatly. A burned post will most likely have an enhanced magnetic contrast, just as a stone-lined posthole, as the magnetic properties of the stones may enhance the possibility of detection by a magnetometer or GPR. One must also consider the possibility that the post, in most cases, may have been removed or have decayed completely, which will have negative implications for their detectability.

There are a few examples of successful detections of postholes, including the instance at Borre in Norway when the geophysical unit of the Swedish

National Heritage Board produced a convincing geophysical map with several postholes forming the typical shape of a prehistoric long house

14

. In this case the postholes were quite large and could be identified accurately with a crossline sampling spacing of 0.25m.

At the Viking Age settlement site of Birka (see above), small postholes were identified as likely parts of a Viking Age building within the Black

Earth area using the MIRA system manufactured by Malå Geoscience

(Trinks et al. 2010). The data were collected at a density of 8cm, both inline and crossline, enabling the mapping of postholes of roughly 0.16m in size.

13 http://www.vfk.no/doc/kulturminnevern/geofysikk/Borre/web/Borre_web_files/frame.htm

14 http://www.vfk.no/doc/kulturminnevern/geofysikk/Borre_oct2007/web/Borre_Oct_2007_web_files/fra me.htm

79

Postholes identified by GPR have also been confirmed at an Iron Age settlement site in the parish of Fresta in Uppland (Viberg 2007). Here roofbearing postholes, part of an Iron Age longhouse, were visible as depressions. Despite the difficult terrain and a coarse crossline sample spacing, a positive correlation could be made between the anomalies in the

GPR data and known postholes. Stone-lined postholes have also been found by magnetic methods in archaeological excavations at Skuttunge, Uppland

(Seiler & Östling 2008:61f).

At some sites entire houses have burned down and have produced distinct magnetic anomalies. One example is the Iron Age settlement site of Uppåkra in southern Sweden, where magnetic measurements indicated a burned house (Voss & Smekalova 2007) and the subsequent archaeological excavation confirmed this interpretation (Larsson, L. pers. comm. 2009).

Stone foundation houses typical of the Migration Period on the Baltic islands of Gotland and Öland have recently been shown to produce a characteristic magnetic response, as they are generally constructed of magnetic stones of igneous origin (Paper IV). At Stånga in the parish of

Lyrungs, Gotland, the data contained evidence of stone boundary walls (sw.

stensträngar) constructed of the same material which opens for the detection of overploughed prehistoric agrarian field systems in other areas. Remains from partly overploughed prehistoric agrarian field systems and clearance cairns have also been detected with a magnetometer in Västergötland

(Berglund 2007). Other stone foundation houses, such as those constructed of limestone typically found within the ringforts of Öland, have also been shown to be identifiable by GPR and to some extent with a magnetometer

(cf. Paper III).

Filled-in ditches can be detected by geophysical methods, as shown by the clear examples available from magnetic and GPR surveys at Nibble in

Uppland and Skänninge in Östergötland (Trinks et al. 2009a, b). Filled-in pits, such as cooking pits, are equally suitable targets for primarily magnetic prospection. Magnetic investigations carried out at Ryssgärdet by Gerhard

Schwartz of the Swedish Geological Survey, for example, successfully detected c. 70% of the cooking pits discovered during subsequent excavations despite a coarse crossline transect spacing of 1m (Schwartz

2005). Cooking pits have also been detected and confirmed by excavations in Uppåkra (Lorra et al. 2001:46; Larsson, L. pers. Comm. 2009) and a midden or heap of fire-cracked stones was successfully detected at RAÄ

1372 in Lapland (Paper II).

5.3.3. Fortified sites

Only a few fortified sites have been investigated geophysically, one of which is the Öland ringfort of Sandbyborg (see Paper III). This is currently the only geophysical survey that has produced clear evidence of houses

80

within a fort on Öland. Another Öland ringfort, Gråborg, has been surveyed with a magnetometer, but the results and interpretations have recently been questioned and intensively debated (Trinks & Biwall 2011a; Danielsson

2012). The argument primarily concerns the interpretation of a linear anomaly outside the fort which was originally said to indicate a moat (e.g.

Tegnér 2008), whereas Trinks and Biwall (2011) prefer to see this as alightning induced magnetic anomaly. The easiest and fastest way to resolve the issue would be to dig in a small trench over the suspected feature, or to carry out a resistivity survey.

5.3.4. Fired structures

The detection of fired structures is one of the more common applications for magnetic and electromagnetic methods internationally and in Sweden.

Successful identification of such structures has been achieved when using geophysical methods in archaeological projects in Sweden. These include hearths (Fridh 1982, 1983; Furingsten 1984, 1985; Persson 2005a; Schwartz

2005; Seiler & Östling 2008), kilns, ovens and blast furnaces for bronze, glass, brick and iron production (Fridh 1982, 1983; Furingsten 1984, 1985,

Kresten & Kresten 1994; Riisager 2003; Persson 2005a; Seiler & Östling

2008).

5.3.5. Medieval buildings

Many structures of medieval age are suitable targets for geophysical surveys. GPR have been shown to produce some outstandingly clear results in cases of medieval structures, and some of these will be referred to below.

Many of these structures are made of brick and thus produce a significant reflection in GPR data. Bricks, even though randomly oriented (Bevan

1994), can produce a strong magnetic response when surveyed with a magnetometer, but many buried medieval brick structures are situated in an urban environment, so that a magnetometer will not always be suitable because of interference from surrounding structures and passing cars.

The use of geophysical methods for mapping churches is nowadays common in Sweden, and GPR investigations of churches, monasteries and convents or parts thereof have produced clear results; see, for example, Old

Uppsala Church (Persson 2005a), Skänninge Dominican Convent (e.g.

Trinks 2005), Krokek Franciscan Convent (Trinks et al. 2008a), Ås

Cistercian Monastery (Löwenstein 2010), grave chambers in Kalmar

Cathedral (Ohlsson 2010), the porch or weapon house of Alunda Church in

Uppland (Kjellberg et al. 2010), Sigtuna Dominican convent (Paper V) and

St. Lars Church in Sigtuna (Viberg et al. manuscript). Other medieval buildings of a non-ecclesiastical nature would of course also be detectable

(see Sträng 1995; Lück et al. 2003; Trinks et al. 2009b, d).

81

5.3.6. Gardens

Historical gardens have been a popular target over the years, as many of them have been rebuilt from time to time and contain the buried remains of previous layouts, such as fountains and paths, which are traceable by geophysical methods (see Persson 2005a; Trinks 2006; Winroth et al. 2011).

5.3.7. Detection of archaeological remains in present-day urban environments

Another important factor to consider is whether the surveys are conducted in an urban or rural environment. Because of interference from passing cars, electromagnetic sources and often a complicated stratigraphy, the only geophysical instrument that seems to have been really successful in Swedish urban environments is the GPR (see Trinks et al. 2008b, 2009c, 2009e;

Trinks & Biwall 2008). Surveys of urban environments have also been carried out in Jönköping, but as the quality of the pictures in the reports available online are poor, it is difficult to evaluate the data (Petterson &

Winroth 2011a, b).

5.3.8. Ferrous objects

Ferrous objects are easy to detect using magnetometers, GPR or other electromagnetic devices. Many are of modern origin, however, and as such are unwanted, as they disturb the identification of prehistoric structures. But in some cases a ferrous response in magnetometer data may originates from prehistoric ferrous objects or artefacts (see Kresten & Kresten 1994). During interpretation of the magnetic data collected by Lorra et al. (2001) in

Uppåkra, Scania, an anomaly was interpreted as being caused by either iron slag or an accumulation of metallic or magnetic objects. Subsequent excavations proved this interpretation right when a large deposit of Iron Age swords was discovered at that point (Larsson, L. pers. Comm. 2009).

5.3.9. Other non-archaeological features

Other more modern features are also detectable by geophysical methods, the more common ones being electric cables (Trinks et al. 2009a, f; Paper

V), metal pipes (Fennö & Dyhlén Täckman 2002), refilled archaeological trenches (Trinks et al. 2009a) and drainage pipes in farm land (Sträng 1995;

Trinks et al. 2009a, f; Gustafsson & Viberg 2010). Geophysical prospection has also been used in Swedish forensic archaeology (Persson 1996).

82

6. Discussion

Despite being successfully deployed in many archaeological projects, geophysical methods have generally been used to a quite limited extent and have not enjoyed great popularity among the Swedish archaeological community. One common issue that is repeatedly mentioned in archaeological reports evaluating the success of geophysical surveys is the fact that the results are difficult to understand as they don’t seem to correlate well with the archaeology of the sites (see Eklund 2008; Steineke 2008;

Bennström & Helgesson 2010; Lindberg 2010; Andersson 2011). The examples mostly concern magnetometer surveys, but similar comments have also been expressed in connection with GPR surveys (e.g. Holback et al.

2004; Ohlsén 2006). Some archaeologists even go as far as to suggest that the methods cannot be regarded as reliable enough. A magnetometer, for example, failed to identify 77% of the hearths discovered at one site during subsequent excavations (Hylén 2007). A few other comparisons between large-area geophysical measurements and archaeological excavations have been carried out, and the results and success rates of the geophysical methods differ greatly. Geophysical measurements with GPR and a magnetometer detected c. 35% of the archaeological remains at Lövstaholm,

Uppland (Larsson 2004), but also provided clear evidence of houses that were not visible during the subsequent excavations. In a similar manner, measurements made using the EM38 at Stensborg in the parish of Grödinge,

Södermanland, seemed to indicate archaeological remains that were not found during the subsequent excavations (Viberg 2008). In rare cases the correlation between the geophysical measurements and the archaeological excavations is zero, as in a survey conducted at Skänninge (Trinks et al.

2010c). Such poor results are somewhat difficult to understand, as previous surveys in the area, in a similar geological setting and carried out by the same staff using the same instrument had correlations between 60% and

100% (Trinks 2005; Trinks et al. 2009b). The only difference seems to be that the unsuccessful survey was carried out in December as opposed to

May, June and July for the successful ones. It is unknown to what extent this influenced the results but it might be considered as a possible explanation.

It must, however, be strongly emphasised that positive examples do exist, e.g. the results presented in papers III and V and those from Ryssgärdet and

Skuttunge in Uppland, which also show an excellent correlation between the geophysical measurements and the excavation results (Schwartz 2005; Seiler

83

& Östling 2008). Note that the magnetic survey at Ryssgärdet succeeded in detecting c. 70% of the cooking pits but only 42% of the hearths. According to the prospector, this was due to the coarse crossline transect spacing (1m) combined with a complicated geology (Schwartz 2005:79). Other positive examples are also listed in chapter 5.

It can be concluded that in the majority of the more problematic surveys geological features, modern ferrous objects or magnetic stones were erroneously interpreted by the prospector as archaeological remains. It is thus not surprising that the correlation with the data from the subsequent archaeological excavations was poor (see Hylén 2007; Steineke 2007). Overoptimistic magnetometer interpretations have previously been noted in the case of a number of reports (e.g. Envall 2007b; Henriksson 2011). These surveys tend to fail to pass beyond the Visual Inspection Level discussed in chapter 5, and thus disregard the geophysical signature of the anomalies.

They also fail to translate these observations into a reasonable archaeological interpretation. It is not good praxis to leave interpretation of the geophysical data to the archaeologists, as they generally have little or no experience regarding such a procedure. As stated in the literature: “It is the

archaeological interpretation that is the common and acceptable end point

for all archaeological prospecting” (Aspinall et al. 2008:115) and as such uninterpreted data is of little or no value.

A possible problem with the attempts to correlate geophysical data and archaeological excavations is that features or objects in the plough zone might have been removed using a bulldozer or excavator before the results could be compared. If this is true, there is a risk that the cause of some of the anomalies visible in the geophysical data had been removed with the ploughed soil. This could easily be estimated by looking at the geophysical data and using the half-width rule, for example, to estimate the probable depth of some of the magnetic anomalies that are missing from the excavation record. An additional factor is the georeferencing of the geophysical grids. In many of the above examples the methods used for georeferencing are not known, and it is likely that some of the features causing the geophysical anomalies could have been situated within a buffer zone around the mapped anomaly. This buffer zone would differ in size depending on the chosen sampling density, the accuracy of the established grids and the accuracy of the subsequent georeferencing.

Regarding the geophysical measurements at Broby bro in Uppland

(Andersson 2011), another method such as GPR could probably have been more successful in detecting the inhumation burials, as magnetometers generally only produce weak responses to such features. Graves, especially those without any related stone constructions, are notoriously difficult to map, however (cf. section 5.3.1.) and a change of instrument would not necessarily yield success, but it does highlight the possible benefit of using multiple methods at a site.

84

There are also a few reports available on magnetometer prospections without any archaeological interpretations, which makes it impossible to evaluate their quality in this respect (e.g. George 2010, 2011). Many magnetic results are also presented without basic map elements such as an amplitude scale showing the magnitude of the magnetic anomalies. Such maps are almost impossible to evaluate correctly (Hylén 2007; Steineke

2008; Bennström 2009; Bennström & Helgesson 2010; George 2010; Ní

Chíobháin 2010; George 2011). Some of the reports mentioned above have such scales, but they only give the approximate magnitude of the geophysical response on a scale ranging from high to low without stating the actual nT values, and this similarly makes the results difficult to evaluate.

This could in part be due to the fact that many of the surveys are reproduced in archaeological reports, so that this information might still be present in the original survey report (e.g. Envall 2007a, b). Thus, archaeologists who summarize geophysical survey results should, if possible, publish this information together with the maps obtained from the geophysical prospection. It is also good practice, if possible, to present both the original geophysical data and the interpretation in both geophysical and archaeological reports.

If one compares the success rates in Sweden with a few international examples (e.g. Linford & David 2001; Ibsen, T. pers. comm. 2012), it can be stated that in Sweden it should be reasonable under normal circumstances to expect a correlation of at least 50% between the geophysical data and the subsequent excavations. Clearly the surveys presented above that produced a correlation of less than 30% are problematic for the general reputation of archaeological geophysics. When this is combined with high expectations from archaeologists, often caused by unrealistic promises made by the surveyor, it becomes even more problematic. Some of the surveys are still conducted with a very coarse crossline transect spacing, which might explain some of the difficulties, but even more problematic is the fact that the surveyor seems in some cases to be lacking in knowledge of how to interpret the data correctly. This is at least partly caused by the limited opportunities for education in archaeological geophysics in Sweden, so that instead of going abroad to get a proper training, it has been easier to buy an instrument and learn the procedure by trial and error. Unfortunately many poorlyexecuted surveys are the result of such a strategy and disappointed archaeologists faced with these measurements are now hesitant about using the methods again. This is not a new problem, but it does mean that all unsuccessful archaeological geophysical surveys should be re-evaluated thoroughly, if possible, in order to understand the reason behind the lack of results (cf. Conyers 2004a:8).

Not all archaeological remains, however, are likely to produce a physical response detectable by geophysical methods. One example of this has been reported by Maki (2006), in which hearths fired at high temperatures for a

85

long period of time lost the initial increase in magnetic susceptibility and inverted back from maghemite to haematite. These hearths may be readily identifiable during excavations but invisible in the geophysical data. In other cases a lack of magnetic contrast between the hearth and the surrounding soils might be caused by firing at very low temperatures so that no magnetic anomaly was created (Trinks et al. 2009a). The area in Nibble, Uppland, was nevertheless identified as a hearth during the subsequent excavation and did produce an echo in the GPR data. In other cases, the geophysical data have shown evidence of archaeological structures which were not visible in the excavation (e.g. Becker 1995; Conyers 2004a:158f; Dalan 2008:7).

Consequently, it must also be acknowledged that archaeological features may be visible in geophysical data but not to the naked eye during excavations. These features may only be detectable when using very sensitive magnetometers in magnetically quiet areas, as the increased magnetic susceptibility in one such case (Becker 1995) and was most likely caused by the presence of biogenic magnetite within the remains of micronsized magnetic bacteria living off decaying wood. In areas where the magnetic susceptibility of many soils is high, such as Sweden or northern

England, such features would most likely not be detected even with sensitive instruments. David Jordan, having evaluated the use of archaeological geophysics in northern England, came to the conclusion that the varied geology of this area, with the presence of glacial remains such as stony till, will not benefit from using more sensitive instruments but rather from increasing the sampling density according to the size and nature of the expected archaeological remains to be detected (Jordan 2009). As the major part of Sweden is covered by glacial till soils, these conclusions are also valid for the Swedish situation. Sweden has many areas, however, that are very suitable for geophysical surveys and valuable results have been produced in a wide range of geographical, geological and pedological environments.

Prior to 2005 the vast majority of geophysical areal surveys carried out in

Sweden were collected in grids very rarely exceeding 50x50m with a crossline transect spacing of between 0.5m and 5m, with 1m being the most common. The inline sampling spacing differed depending on the choice of method, but varied between 0.03m and 0.05m for GPR surveys, 0.5 and 5m for conductivity and magnetic susceptibility measurements using the EM38 by Geonics and 0.02 and 5m for magnetic surveys. Given the review of hearth sizes conducted here, such crossline sampling spacings are not sufficient for the detection of the average size of hearth present in the data set. As almost 1000 hearths were included in the data set, it is likely that this calculation can be used as a reasonable estimate for future surveys and archaeological investigations. The sampling density needs to be governed by the sizes of the archaeological features. This means that in a majority of surveys carried out before 2005 the sampling distance was too large to allow

86

correct characterisation and interpretation of the archaeology, and as a result, many of the smaller features at the sites were most likely missed or proved difficult to interpret. This is equally true for the sizes of the areas surveyed, an important consideration for the correct evaluation of noise levels in the survey area and for achieving better interpretations (see Gaffney & Gater

2003:91ff). The situation has only partially improved since 2005.

The founding of a geophysical unit at the Swedish National Heritage

Board has been important for the development of archaeological geophysics in this country and for the debate on the importance of proper routines regarding data collection, interpretation and the importance of densely sampled data, but many surveys are still performed with too coarse a sampling density and this limits the detectability of smaller archaeological remains. The amount of geophysical work carried out in larger areas has increased and this is a positive factor for the interpretability of the resulting data.

The difficulty of making correct interpretations of GPR radargrams was highlighted early on by Wihlborg & Romberg (1980) and Wahlström (1993), among others, and even though the practise of using time slices to increase the interpretability of GPR data was introduced into archaeological geophysics in the early 1990s (Goodman & Nishimura 1993; Goodman et al.

1995), it did not become commonplace in Sweden until around 2005. This was also noted by Alkarp and Price (2005:264), who strongly advised against the exclusive use of radargrams except in situations when timeslicing is impossible or inappropriate or as a complement to time slices in order to strengthen the interpretation of a data set. To my knowledge, the first application of time slicing in Swedish archaeological research was by a team from Kiel University surveying in Uppåkra in 1997 (Lorra et al. 2001), and there have been a few examples of GPR time slicing from 2003 onwards, using software such as GPR Slice and Easy 3D (e.g. Persson 2003;

Persson 2005a) and GPR process (e.g. Persson 2004, 2005b, c). It is obvious that the adoption of time-slicing have increased the reliability of interpretations of GPR data and made it easier for archaeologists to involve themselves in the interpretation process. The use of time slices to visualise

GPR data must be regarded as standard today, and this has been important for the production of the many convincing sets of GPR results published in recent years. There are, however, still many geologists and geophysicists without experience in carrying out archaeological surveys using GPR who continue to work exclusively in a 2D environment when interpreting the data they have collected.

It is also interesting to note that the magnetometer, which has been one of the most popular instruments for archaeogeophysical research abroad, was only used sporadically in Sweden before 2005. Magnetometers have probably been available at the geophysical departments of many Swedish universities, but as collaboration between archaeologists and geophysicists

87

has been sporadic, they have had a limited impact on archaeological research. The same is true for geophysical equipment in general, but as magnetometers have not been available for rental in past years, this may have been an extra factor restricting the use of the method. As the lack of geophysical instruments has been an obvious problem, it has also been difficult to educate archaeologists in the use of these methods. Students with an ambition to learn about archaeological geophysics have had to attend courses elsewhere in order to achieve the proper competence. Education in geophysical prospection methods has been available at the geophysical departments in Uppsala and Luleå, and collaborations with archaeologists has taken place in the past 10-20 years (see Sträng 1995; Masters 1998;

Wedmark 1999; Karlberg & Sjöstedt 2007), but in many cases the sampling strategies, sizes of the areas surveyed and the visualisation methods in these reports have not been adapted to the archaeological application and the authors have failed to produce any easily interpretable or clear results (e.g.

Hedberg et al. 1996; Wedmark 1999; Karlberg & Sjöstedt 2007). At present, the only section of a Swedish academic archaeological department offering the possibility of working with archaeological prospection methods such as

GPR, magnetometry and the EM38 is the Archaeological Research

Laboratory at Stockholm University, where the availability of instrumentation has recently been improved partly through collaboration with the Royal Institute of Technology in Stockholm.

The necessary level of education is also a recurring question when discussing archaeological geophysics in Sweden. As Sweden lacks any guidelines regarding the qualifications of survey personnel, one has to look at the requirements in other countries. A primary source of information regarding the use of archaeological geophysics is the English Heritage

Guidelines (David et al. 2008), where this issue is dealt with under subsection 2.8, which states that the project leader, i.e. the archaeological geophysicist, must have: “competence in basic metric survey procedure;

experience in a supervised capacity of at least 30 different site surveys, or a minimum of three full years’ supervised experience of archaeological

geophysics, and a degree in archaeology and/or an appropriate science (e.g.

MSc in Archaeological Prospection).” (David et al. 2008:5).

This topic is also discussed by Ernenwein and Hargrave (2009): “The

least effective and least responsible approach is to attempt to learn geophysics oneself, relying exclusively on the manual, while doing one’s first

“real” survey. Everyone’s initial surveys are likely to include some poor quality data, sub-optimal data processing, and possibly ill-founded interpretations. It is likely to be detrimental to a customer’s view of geophysics in general and oneself in particular to try to “learn on the job”

without the support of a competent instructor.” (Ernenwein and Hargrave

2009:98).

88

It is thus recommended that anyone planning to commission a geophysical survey should make sure that the company selected for the job can provide the necessary credentials, as this will be a guarantee that the survey and its interpretation are carried out in the best possible way. Novice users are advised to get a proper education before starting to collect data. In this regard the concept of the Learning Curve presented by Schurr (1997) could be an important starting point alongside the introduction to archaeological geophysics presented by Ernenwein and Hargrave (2009).

6.1. The Swedish situation and beyond

It can be stated as a result of e-mail correspondence with many archaeologists who have used archaeological geophysics in projects in

Sweden that the majority of them are interested in the possibilities offered by the methods and in seeing more examples of geophysical surveys carried out in different Swedish environments and geological settings before deciding if and when to re-evaluate the usefulness of the methods. Many also believe that Sweden is unsuited in its geology and pedology for geophysical prospection and this being the case, it was my hope that the present thesis, with examples of successful surveys performed using a range of methods and in different geological and pedological settings, could put forward some additional evidence to archaeologists who may be hesitant about using geophysics. From a geophysical point of view, the effort of trying to prove the method’s reliability in different areas of Sweden might seem strange, as the methods have been shown to work abroad and one would assume that a quick glance at a geological and pedological map would indicate areas that are suitable for prospection. However, as most Swedish archaeologists were relatively unexposed to the use of geophysical prospection methods before the middle of the last decade, and as several of the archaeologists’ first contacts with the methods may have been disappointing, it is reasonable to think that acceptance will take some time and that more positive examples are needed. This is similar to the experience in other countries as well, so that in the U.S.A., for example, there has been a gradual increase in the use of these methods in archaeological projects over a number of years. On a positive note, one can see that the use of geophysical prospection methods, as a valuable tool for the initial assessment of archaeological sites, is recommended in the Swedish National Heritage Board’s guidelines for contract archaeology

15

. It is to be hoped that such statements will lead to a more frequent involvement of archaeological geophysics during archaeological investigations in Sweden, as has been the case in England.

Another factor influencing the likely use of the methods is cost. The extra cost will in some cases be a limiting factor. When discussing this with some

15 http://www.raa.se/publicerat/varia2012_30.pdf

89

of the field archaeologists that I have been in contact with during my time as a PhD student, the price of the surveys has been a matter of constant concern. The change from a crossline transect spacing of 0.5m to 0.25m, for example, or from a crossline transect spacing of 1m to 0.5m, will increase both the time spent in the field and the cost of the survey. Are Swedish archaeologists ready to pay that extra money to get more accurate and easily interpretable results? I would like to think that this cost, when compared with the cost of an archaeological excavation, would turn out to be quite small, and would be bearable if the geophysical results provided were clear and accompanied by a solid interpretation that is readily understandable and usable for archaeologists.

All things considered, is the Swedish situation unique? The answer to this question is definitely no, as the small survey of the use of the methods in different countries provided in chapter 2 has shown. Many countries are in the same situation as Sweden and facing the same challenges, and Scotland and the other Nordic countries share many of the same geological and pedological characteristics that sometimes make geophysical surveying challenging. The first surveys and trials using archaeological geophysical prospection in Sweden were carried out quite late relative to the situation in other countries, and this may have had an impact on the development of the discipline here. It should be remembered, however, that although countries like Denmark and Norway started using the methods earlier than Sweden, they only use them sporadically nowadays, so that other reasons for their limited use must be sought. In this regard, the geology and pedology of the countries concerned and the character of their archaeological remains may have been two of the major causes of the failure of many surveys in the past.

When no specialised education is available for surveyors, who have to learn to use the methods on the job and without a full understanding of the underlying principles of operation or of the interpretation of geophysical data, the differentiation between anthropogenic remains and traces caused by the geology and pedology of the area concerned becomes difficult. This might explain the lack of success and ultimately the cause of a lot of the disappointment felt among archaeologists involved in such projects. The solution is to initiate specialised courses of university-level training, combined with shorter courses for archaeologists interested in learning about the basic principles of the methods and how to make the best use of them in different situations.

6.2. Brief guidelines for archaeologists

Taking the above-mentioned factors into consideration, archaeologists commissioning geophysical surveys need to make sure that:

90

• The archaeological remains are expected to give measureable contrasts with the surrounding soil matrix in terms of physical properties.

If such as contrast is not expected, a geophysical survey should not be recommended.

• The appropriate geophysical methods are chosen for detecting the presumed archaeological remains.

As shown in chapter 3, many of the methods measure different geophysical parameters, some of which may be relevant to the question in hand and some not. Often companies selling these services only use, or own, one of the many available methods and instruments, and as a consequence the archaeologist should seek quotations from different companies to make sure that the optimum method or methods will be chosen for the specific project.

Many methods provide complimentary information, and some projects could benefit from an integrated approach. An especially successful combination of methods seems to be GPR and magnetometry, but earth resistance or resistivity can provide an alternative to GPR where the terrain or other factors restrict the use of the latter. Other combinations could definitely be valid, as seen in the case of the mountain tundra survey presented in Paper

II, and care should therefore be taken to choose suitable methods depending on the archaeological, geological and pedological preconditions at each site.

• The data should be collected at a sufficient sampling density for the detection and description of the archaeological remains in question.

Let the sizes of the expected features govern the sampling density and make sure that this density allows the smallest archaeological features to be resolved to be covered by at least two consecutive transects. This also involves choosing the appropriate antenna frequency when surveying with

GPR. As GPR results are a trade-off between depth and resolution, this is an important decision to make in order to be able to reveal features of a certain size at certain depths.

• The results should include a thorough interpretation of the data and should be interpreted by someone with training in archaeology and experience in the geophysical methods used and the expected archaeological situation, or through collaboration between an archaeologist and a geophysicist.

Many reports of geophysical surveys that are available online do not contain realistic geophysical or archaeological interpretations of the data and are of

91

limited value. The goal of any geophysical survey should be the archaeological interpretation, as mentioned above, and it is only at this point that the data will become meaningful to the archaeologists involved in the project.

92

7. Conclusions and future prospects for archaeological geophysics in Sweden

Archaeological geophysics can provide answers to a range of archaeological questions. As the studies from Gotland and the mountain tundra region presented in Papers II and IV show, it can be used as a means of detecting the presence or absence of archaeological structures and features in an area. But it can also be used for detailed studies of both Iron Age and

Medieval archaeological remains, as a knowledge of the spatial layout of buried constructions is valuable for answering specific archaeological questions. It has also been shown that geophysical prospection, as at the

Dominican Convent in Sigtuna, can connect information from widely separated trenches and put these results into a meaningful context.

Furthermore, it is possible in surveys of large areas, such as those carried out by the Ludwig Boltzmann Institute at Björkö, Gamla Uppsala and Uppåkra, to connect archaeological information on an even wider scale and thereby enable more far-reaching conclusions about the archaeology of entire landscapes. Especially suitable methods for geophysical prospection in

Sweden are magnetometry and GPR, but electrical resistance or conductivity measurements could be an alternative to GPR on steep slopes or in highly conductive soils.

Geophysical surveys cannot provide the answer to all archaeological questions, however, and use of these methods might be limited in some areas by the geology, pedology and/or former land use, for example. Factors such as a lack of properly educated surveyors and the unavailability of geophysical instruments, in combination with sceptical archaeologists, may also have contributed to the low frequency of use of archaeological geophysics in Sweden. Many areas exist in Sweden, however, that are eminently suitable for geophysical surveys, and by making careful predictions as to the likely success of certain methods, based on a thorough knowledge of geological, pedological and land use factors and the character of the expected archaeological remains, the information extractable from such surveys can be increased dramatically. This includes tailoring the sample density of each geophysical survey according to the sizes of the expected archaeological remains.

Some areas in Sweden and certain kinds of archaeological remains are not suitable for the use of geophysical surveys, and an equally important task for a geophysical prospector must be to give the archaeologists responsible

93

advice on looking for other solutions where appropriate. This would make it possible to avoid many disappointing results. In some cases a combination of trial trenching and archaeological geophysics might be a sensible solution, and in other cases the use of one of these methods might suffice. The nondestructive character of a geophysical survey makes it an especially attractive possibility at sensitive sites.

In order to increase the use of geophysical methods in Sweden, professional archaeologists should be made aware of the pitfalls and possibilities of the methods, which implies an even more urgent need to address the educational situation in the country. More specialised education in archaeological geophysics, and shorter and more general courses in the subject aimed at archaeologists working with various excavations companies, county boards and universities, would be desirable. An additional and important step towards increasing the general awareness of geophysical methods and how to implement them in archaeological projects might be to formulate official Swedish guidelines, similar to those presented by English Heritage, on how to use archaeological geophysics. This would ensure that Swedish archaeologists constantly have updated recommendations to turn to. It is also suggested that a website should be constructed where companies working with archaeological geophysical surveys in Sweden can get present data from their past surveys together with their CVs, publication lists and contact information, as this would make it easier for archaeologists or contractors interested in commissioning geophysical surveys to make educated estimates of the expected quality of the results, the price of the survey, the expected data density and the quality of the subsequent interpretations. In the meantime, it is my hope that this thesis could provide a point of departure for a renewed discussion on archaeological geophysics in Sweden, a discussion which would place the remnant echoes of the past as collected by geophysical methods in a meaningful archaeological context.

94

References

Abrahamsen, N. 1965. Arkaomagnetisme og jernalderslagge. Kuml: 115-132.

Aitken, M.J. 1958. Magnetic prospecting. I – The Water Newton Survey.

Archaeometry 1: 24-29.

Aitken, M.J. 1959. Test For Correlation Between Dowsing Response And Magnetic

Disturbance. Archaeometry 2: 58-59.

Aitken, M.J. 1960. Magnetic Prospecting: The Proton Gradiometer. Archaeometry 3:

38-40.

Aitken, M.J. 1974. Physics and Archaeology. Clarendon Press. Oxford.

Aitken, M.J. & Weaver, G.H. 1962. Magnetic dating; Some archaeomagnetic measurements in Britain. Archaeometry 5: 4-18.

Alkarp, M. & Price, N. 2005. Tempel av guld eller kyrka av trä?

Markradarundersökning vid Gamla Uppsala kyrka. Fornvännen 100: 261-272.

Alldred, J.C. 1964. A Fluxgate Gradiometer for Archaeological Surveying.

Archaeometry 7: 14-19.

Anderson Stamnes, A. 2010. Developing a sequential geophysical survey design for

Norwegian Iron Age settlements. Master thesis. Division of Archaeological,

Geographical and Environmental Sciences, University of Bradford. Bradford.

Andersson, K., Annuswer, B., Karlsson, C. & Åberg, K. 2003. Särskild arkeologisk

undersökning. Två gravfält med boplatslämningar längs Mälarbanan mellan

Bro och Bålsta. RAÄ 57 i Bro socken och RAÄ 11 i Låssa socken, Uppland.

Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV

Bergslagen. Örebro.

Andersson, L. 2011.

Mellan ättebacke och kyrkogård. Arkeologisk forskningsundersökning av en gravplats från skiftet vikingatid-medeltid vid

Broby bro, RAÄ 42, i Täby socken och kommun. Undersökningarna 2007-2010.

Stockholms Läns Museum. Stockholm.

Andersson, P. 1998. Slagrutan Och Arkeologin. Slagrutan som arkeologisk metod.

Unpublished seminar paper. Institutionen för arkeologi. Lunds Universitet.

Lund.

Andrén, T. & Lindeberg, G. 1997. A Shallow Seismic Refraction Survey at Björkö,

Eastern Sweden. In Miller, U., Clarke, H

.,

Hansson, A.-M. & Johansson, B.M.

(Eds.) Environment and Vikings. Scientific Methods and Techniques.

Riksantikvarieämbetet & Statens Historiska Museer. Stockholm.

Ängeby, G. 2001. Arkeologisk undersökning. Ensamgården vid Orred – långhus

från äldre järnålder. Halland, Fjärås socken, RAÄ 499. Riksantikvarieämbetet.

Avdelningen för arkeologiska undersökningar. UV Väst. Kungsbacka.

Arbman, H. 1941. Hjalmar Stolpe som fornforskare. Fornvännen 36: 146-162.

Arrhenius, O. 1935. Markundersökning och arkeologi. Fornvännen 30: 65-76.

Aspinall, A. & Crummett, J.G. 1997. The electrical pseudo-section. Archaeological

Prospection 4: 37-47.

Aspinall, A., Gaffney, C. & Schmidt, A. 2008. Magnetometry for Archaeologists.

Altamira Press. Lanham.

95

Aspinall, A. & Lynam, J.T. 1970. An induced polarisation instrument for the detection of near surface features. Prospezioni Archaeologiche 5: 67-75.

Atkinson, R.J.C. 1946. Field Archaeology. Jarrold and sons Ltd. Norwich.

Atkinson, R.J.C. 1953. Field Archaeology. 2nd ed. Methuen. London.

Bailey, R.N., Cambridge, E. & Briggs H.D. 1988. Dowsing and church

archaeology. Intercept. Wimborne, Dorset.

Barakat, N., Dolphin, L.T., El Dessouki, T., El Hennawi, H., Moussa, A.H., Tolba,

M.F., Abdel-Wahab, S., Bollen, R.L., Johnson, D.A., Oetzel, G.N., Tanzi, J.D.,

Mokhtar, G., Hassan, A. & Buckingham, S.O. 1975. Electromagnetic sounder

experiments at the pyramids of Giza. Ain Shams University, Cairo. Stanford

Research Institute, Menlo Park, California.

Becker, H. 1995. From Nanotesla to Picotesla – a New Window for Magnetic

Prospecting in Archaeology. Archaeological Prospection 2: 217-228.

Becker, H. 2009. Caesium-magnetometry for landscape-archaeology. In Campana,

S. & Piro, S. (Eds.) Seeing the Unseen. Geophysics and Landscape

Archaeology. Taylor & Francis group. London.

Bellomo, R.V. 1993. A Methodological Approach for Identifying Archaeological

Evidence of Fire Resulting from Human Activities. Journal of Archaeological

Science 20: 525-553.

Benech C., Marmet E. 1999. Optimum Depth of Investigation and Conductivity

Response Rejection of the Different Electromagnetic Devices Measuring

Apparent Magnetic Susceptibility. Archaeological Prospection 6: 31-45.

Bennström, G. 2009. Arkeologisk förundersökning vid Leksandsbröd i Gärde by, raä

642 i Leksands socken och kommun, Dalarna. Dalarnas museum. Falun.

Bennström, G. & Helgesson, C. 2010. Arkeologiska undersökningar Sör Romme:

gravfält och kolningsanläggning, Stora Tuna 372:2 och 969, Borlänge kommun,

Dalarna. Dalarnas museum, Falun.

Berard, B.A. & Maillol, J.M. 2008. Common-and Multi-Offset Ground-Penetrating

Radar Study of a Roman Villa, Tourega, Portugal. Archaeological Prospection

15: 32-46.

Berglund, A. 2007. Väg 49 etapp Axevalla-Varnhem. Arkeologisk utredning steg II.

Västergötlands Museum. Skara.

Bergold, H. 2008. Arkeologisk undersökning. Gården Stora Mon och dess

närområde.

Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Bergslagen, Örebro.

Bergström, G. 1913. Försök med Elektrisk Malmletning. P.A. Norstedt & Söner.

Stockholm.

Berthelson, B. 1969. Några ord om Erik Lundberg. Fornvännen 64: 235-240.

Bethell, P. & Máté, I. 1989. The use of phosphate analysis in archaeology: a critique. In Henderson, J. (Ed.) Scientific Analysis in Archaeology. University committee for Archaeology. Oxford.

Bevan, B. 1983. Electromagnetics for Mapping Buried Earth Features. Journal of

Field Archaeology 10: 47-54.

Bevan, B. 1984. Environmental effects on ground-penetrating radar. Fifty-fourth

Annual International Meeting, Society of Exploration Geophysicists. Society of

Exploration Geophysicists. Atlanta.

Bevan, B. 1990. Self-potential surveys. Speleonics 4: 7-8.

Bevan, B. 1991. The search for graves. Geophysics 56: 1310-1319.

Bevan, B. 1994. The Magnetic Anomaly of a Brick Foundation. Archaeological

Prospection 1: 93-104.

Bevan, B. 1995a. Magnetic surveying and lightning. 7ear-Surface Views. October:

7-8.

96

Bevan, B. 1995b. Geophysical prospecting. American Journal of Archaeology 99:

88-90.

Bevan, B. 1998. Geophysical Exploration for Archaeology: An Introduction to

Geophysical Exploration. Special report No. 1. United States Department of the

Interior. National Park Service. Midwest Archaeological Center. Lincoln,

Nebraska.

Bevan, B. 2000. An early geophysical survey at Williamsburg, USA.

Archaeological Prospection 7: 51-58.

Bevan, B. 2002. A listing of geophysical surveys in the Americas. Unpublished report. Geosight, Weems, Virginia.

Bevan, B. 2003. What is a Geophysical Interpretation? Unpublished report.

Geosight, Weems, VA.

Bevan, B. 2004. Geophysical Exploration for Archaeology. Volume C. Detailed survey procedures. Technical Report Number 4. Geosight Weems, Virginia.

Bevan, B. 2009. The Magnetic Anomaly of a Lightning Strike. Fasttimes 14: 61-64.

Bevan B. & Kenyon J. 1975. Ground-penetrating radar for historical archaeology.

MASCA 7ewsletter 11: 2-7.

Biwall, A., Gabler, M., Hinterleitner, A., Karlsson, P., Kucera, M., Larsson, L.,

Löcker, K., Nau, E., Neubauer, W., Scherzer, D., Thorén, H., Trinks, I., Wallner

M. & Zitz T. 2011. Large-Scale archaeological prospection of the Iron and

Viking Age site Uppåkra in Sweden. First results of the LBI-ArchPro landscape archaeological case study. In Drahor, M.G. & Berge, M.A. (Eds.)

Archaeological Prospection. Extended Abstracts. 9th International Conference on Archaeological Prospection. Archaeology & Art Publications. Izmir, Turkey.

Bjelm, L. & Larsson, L. 1980. Georadar i arkeologins tjänst. META 4: 31-32.

Bjelm, L. & Larsson, L. 1984. Application of Georadar in Archaeological Research.

Meddelanden från Lunds universitets historiska museum 5: 39-46.

Björck, N., & Appelgren, K. 2006. Arkeologisk förundersökning och undersökning.

Boplats och gravar från äldre järnålder i Fyrisåns dalgång. Väg E4, sträckan

Uppsala-Mehedeby. Uppland, Gamla Uppsala socken, Fullerö 21:21 och 21:57,

RAÄ 598. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar.

UV GAL. Uppsala.

Björck, N., Lindberg, K.-F., Seiler, A. & Östling, A. 2008. Arkeologiska

undersökningar. Väg 709 – ett snitt genom Vendels förhistoria. Uppland,

Vendels socken, Fembäcke 1:6, Umberga 1:13, Brunnby 7:1, RAÄ 292, RAÄ

295:2, RAÄ 295:3, RAÄ 93. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Uppsala. Uppsala.

Burenhult, G. 1984. Projektet Arkeologiska Prospekteringsmetoder – ett vetenskapligt projekt för allmänheten. Gotländskt arkiv 1984: 23-32.

Burenhult, G. & Brandt, B. 2002. The grave-field at Ajvide. In Burenhult, G. (Ed.)

Remote sensing: applied techniques for the study of cultural resources and the localization, identification and documentation of sub-surface prehistoric remains in Swedish archaeology. Vol. 2, Archaeological investigations, remote

sensing case studies and osteo-anthropological studies. Department of

Archaeology. Stockholm University. Stockholm.

Buteux, S., Gaffney, V., White, R. & Van Leusen, M. 2000. Wroxeter hinterland project and geophysical survey at Wroxeter. Archaeological Prospection 7: 69-

80.

Bäck, M., Grenler, H., Hinterleitner, A., Nordström, A., Ramström, A., Trinks, I. &

Östlund, S. 2010. Mellan byn och staden – Birkas och Björkös bebyggelse ur

nya synvinklar. Amatörarkeologiska undersökningar år

2005-2007.

97

Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Mitt.

Stockholm.

Campana, S. & Piro, S. 2009. (Eds.) Seeing the unseen. Geophysics and landscape

archaeology. Taylor & Francis Group. London.

Carabelli, E. 1967. Ricerca sperimentale dei dispositivi più adatti alla prospezione elettrica di cavità sotteranee. Prospezioni Archaeologiche 2: 9-21.

Carelli, P. 2003. Krapperup och det feodala samhället: borgen, bygden och den

medeltida bebyggelseutvecklingen i en nordvästskånsk socken. Gyllenstiernska

Krapperupstiftelsen. Nyhamnsläge.

Carlborg, H. 1963. Om gruvkompasser, malmletning och kompassgångare. Sancte

Örjens Gille. Stockholm.

Carson, H.H. 1962. A seismic survey at Harpers Ferry. Archaeometry 5: 119-122.

Cassidy, N.J. 2009a. Electrical and magnetic properties of rocks, soils and fluids. In

Jol, H.M. (Ed.) Ground Penetrating Radar. Theory and Applications. Elsevier.

Amsterdam.

Cassidy, N.J. 2009b. Ground penetrating radar data processing, modelling and analysis. In Jol, H.M. (Ed.) Ground Penetrating Radar. Theory and

Applications. Elsevier. Amsterdam.

Chwala, A., Stolz, R., IJsselsteijn, R., Schultze, V., Ukhansky, N., Meyer, H.G. &

Schüler, T. 2001. SQUID gradiometers for archaeometry. Superconductor

Science and Technology 14: 1111.

Chwala, A., IJsselsteijn, R., May, T., Oukhanski, N., Schüler, T., Schultze, V., Stolz,

R. & Meyer, H.G. 2003. Archaeometric Prospection With High-Tc Squid

Gradiometers. IEEE Transactions on Applied Superconductivity 13: 767-770.

Clark, A. 1975. Archaeological Prospecting: A Progress Report. Journal of

Archaeological Science 2: 297-314.

Clark, A. 1996. Seeing beneath the soil. Prospecting methods in archaeology.

Routledge. Oxon.

Clarke, C.M., Utsi, E. & Utsi V. 1999. Ground penetrating radar investigations at

North Ballachulish Moss, Highland, Scotland. Archaeological Prospection 6:

107-121.

Clay, R.B. 2001. Complementary Geophysical Survey Techniques: Why two ways are always better than one. Southeastern Archaeology 20: 31-43.

Connah, G., Emmerson, P. & Stanley J. 1976. Is There a Place for the Proton

Magnetometer in Australian Field Archaeology? Mankind 10: 151-155.

Connor, M. & Scott, D.D. 1998. Metal Detector Use in Archaeology: An

Introduction. Historical Archaeology 32: 76-85.

Conyers, L. 2004a. Ground penetrating radar for Archaeology. AltaMira Press.

Walnut Creek, California.

Conyers, L. 2004b. Moisture and Soil Differences as Related to the Spatial Accuracy of GPR Amplitude Maps at Two Archaeological Test Sites. In Slob, E.,

Yarovoy, A. & Rhebergen, J. (Eds.) The Tenth International Conference on

Ground Penetrating Radar. Delft, Netherlands.

Conyers, L. 2006a. Ground-Penetrating Radar Techniques to Discover and Map

Historic Graves. Historical Archaeology 40: 64-73.

Conyers, L. 2006b. Ground-Penetrating Radar. In Johnson, J.K. (Ed.) Remote

Sensing in Archaeology. An Explicitly 7orth American Perspective. University of Alabama Press. Tuscaloosa.

Conyers, L. & Goodman, D. 1997. Ground penetrating radar. An introduction for

archaeologists. AltaMira Press. Walnut Creek, California.

Conyers L. & Lucius, J.E. 1996. Velocity Analysis in Archaeological Ground-

Penetrating Radar Studies. Archaeological Prospection 3: 25-38.

98

Crowther, J. 1997. Soil Phosphate Surveys: Critical Approaches to Sampling,

Analysis and Interpretation. Archaeological Prospection 4: 93-102.

Dabas, M. 2009. Theory and practice of the new fast electrical imaging system

ARP©. In Campana, S. & Piro, S. (Eds.) Seeing the Unseen. Geophysics and

Landscape Archaeology. Taylor & Francis Group. London.

Dabrowski, K. 1963. The application of geophysical methods to archaeological research in Poland. Archaeometry 6: 83-88.

Dahlin, T. 2001. Resistivity survey at Stora Uppåkra, Sweden. In Larsson, L. (Ed.)

Uppåkra – Centrum i analys och rapport. Lunds Universitet. Lund.

Dalan, R.A. 2006. A Geophysical Approach to Buried Site Detection using Downhole Susceptibility and Soil Magnetic Techniques. Archaeological Prospection

13: 182-206.

Dalan, R.A. 2008. A Review if the Role of Magnetic Susceptibility in

Archaeogeophysical Studies in the USA: Recent Developments and Prospects.

Archaeological Prospection 15: 1-31.

Damell, D. 1986. Fornl. nr 38, 39 m fl vid Signhildsberg, Håtuna sn, Up.,

Riksantikvarieämbetet och Statens Historiska Museer. Stockholm. Dnr.

1811/84.

Damell, D. 1991. Husgrunderna. In Andersson, B., Damell, D. & Norrman, J. (Eds.)

Fornsigtuna. En kungsgårds historia. Stiftelsen Upplands-Bro Fornforskning.

Upplands-Bro.

Danielsson, R. 2012. Svar till Trinks & Biwall om geofysik vid Gråborg.

Fornvännen 107: 124-127.

David, A. 2001. Overview — the role and practice of archaeological prospection. In

Brothwell, D.R. & Pollard, A.M. (Eds.) Handbook of Archaeological Sciences,

John Wiley. New York.

David, A., Linford, N. & Linford, P. 2008. Geophysical Survey in Archaeological

Field Evaluation. 2nd ed. English Heritage. Swindon.

Dearing, J. 1999. Environmental Magnetic Susceptibility. Using the Bartington MS2

System. Chi Publishing. England.

Dennis, T.J. 2004. Imaging geophysical data – taking the viewer into account.

Archaeological Prospection 11: 35-48.

De Terra, H. 1947. Preliminary note on the discovery of fossil man at Tepexpan in the Valley of Mexico. American Antiquity 13: 40-44.

De Terra, H. 1949. Early man in Mexico. In Linton, R. (Ed.) Tepexpan man (Viking fund publications in anthropology). Publ. for the Wenner-Gren Foundation for

Anthropological Research. University of Arizona Press. New York.

DeVore, S.L. 2001. Are We there Yet?: Geophysics in the National Park Service.

Presented at the 66th Annual Meeting of the Society for American Archaeology,

United States Department of the Interior. National Park Service. Midwest

Archaeological Center. Archaeological Assistance and Partnership Program.

New Orleans, Louisiana.

Doneus, M. & Neubauer, W. 1998. 2D Combination of Prospection Data.

Archaeological Prospection 5: 29-56.

Dos Santos, V.R.N. & Porsani, J.L. 2011. Comparing performance of instrumental drift correction by linear and quadratic adjusting in inductive electromagnetic data. Journal of Applied Geophysics 73: 1-7.

Drahor, M.G. 2004. Application of the self-potential method to archaeological prospection: some case histories. Archaeological Prospection 11: 77-105.

Drahor, M.G. 2011. Archaeological prospection in Anatolia: Past, Present and

Future. In Drahor, M.G. & Berge, M.A. (Eds.) Archaeological Prospection.

Extended abstracts. International Society for Archaeological Prospection

99

(ISAP), Center for Near Surface Geophysics and Archaeological Prospection

(CNSGAP) of Dokuz Eyül University. Archaeology and Art Publications. Izmir.

Drahor, M.G., Berge, M.A., Öztürk, C., Alpaslan, N. & Ergene, G. 2009. Integrated usage of geophysical prospection techniques in Höyük (tepe, tell)-type archaeological settlements. In Marguerie, D. & Lanos, P. (Eds.)

ArchéoSciences. Mémoire du sol, espace des hommes. Presses universitaires de

Rennes. Rennes.

Dunér, J. & Vinberg, A. 2008. Arkeologiska förundersökningar och undersökningar.

Barva – 2 000 år vid Mälarens södra strand. E20, sträckan Eskilstuna-Arphus.

Södermanland, Barva socken, Säby 4:1, RAÄ 17, RAÄ 36, RAÄ 53, RAÄ 55-57,

RAÄ 66, RAÄ 150 och RAÄ 153. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Mitt. Stockholm.

Edenmo, R. 2006.

Arkeologisk förundersökning och undersökning.

Stenåldersboplats och järnåldersgravar vid Spångastugan. Väg 53, förbifart

Malmköping. Södermanland. Lilla Malma socken, Malma-Lundby 6:2, RAÄ 69.

Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Mitt.

Stockholm.

Edlund, M. 2007. Arkeologisk undersökning. Hyddor vid Hidingebro – om

boplatslägen från stenålder, järnålder och 1700-tal. 7ärke, Hidinge socken,

Hidinge 11:2, RAÄ 113 och 114. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Bergslagen. Örebro.

Edlund, M. 2008. Arkeologisk undersökning. Treuddar och hägnader. E 18-

undersökning vid Bengtstorps gravfält. 7ärke, Täby socken, Bengtstorp 1:1,

RAÄ 34, RAÄ 35. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Bergslagen. Örebro.

Eklund, S. 2008. Undersökning med magnetometer. In Onsten-Molander, A. (Ed.)

Skälby. Bilden av byn växter fram. Societas Archaeologica Upsaliensis.

Uppsala.

Englund, T. 1999. Prospektering av Hovgårdsberget i Vendel. Inventering, elektromagnetisk kartering och fosfatkartering. In CD-uppsatser i laborativ

arkeologi 98/99, Del 1. Arkeologiska forskningslaboratoriet, Stockholms

Universitet. Stockholm.

Envall, G. 2007a. Rapport. Geofysiska mätningar Axevalla. Unpublished report.

Skandinavisk Arkeologisk Geofysik Ab. Borlänge.

Envall, G. 2007b. Rapport. Geofysiska mätningar. Ledsjö. Unpublished report.

Skandinavisk Arkeologisk Geofysik Ab. Borlänge.

Ercul, E., Wunderlich, T., Stümpel, H., Rabbel, W., Pasteka, R., Baris, S., Peksen,

E., Yazici-Cakin, O. & Niewöhner, P. 2011. Discovering a Byzantine cross-insquare church in Iznik using geophysics. In Drahor, M.G. & Berge, M.A. (Eds.)

Archaeological Prospection. Extended abstracts. International Society for

Archaeological Prospection (ISAP), Center for Near Surface Geophysics and

Archaeological Prospection (CNSGAP) of Dokuz Eyül University. Archaeology and Art Publications. Izmir.

Eriksson, L. & Henkel, H. 1994. Geofysik. In Fredén, C. (Ed.) Sveriges

7ationalatlas. Berg och jord. Sveriges Nationalatlas. Stockholm.

Eriksson, T. & Häringe Frisberg, K. 2010. Arkeologisk undersökning.

Skärvstenslager, stensättning och röjningsrösen vid Hämringe. Väg E4, sträckan Uppsala-Mehedeby. Uppland; Gamla Uppsala socken; Fullerö 17:1,

Fullerö 17:3, Fullerö 18:1 och Fullerö 18:6; RAÄ 657, RAÄ 658, RAÄ 659,

RAÄ 660, RAÄ 675, RAÄ 676, RAÄ 677. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Mitt. Stockholm.

100

Erlström, M., Persson, L., Sivhed, U., Wickström, L. 2009. Beskrivning till regional

berggrundskarta över Gotlands län. SGU. Uppsala.

Ernenwein E.G. & Hargrave, M.L. 2009. Archaeological Geophysics for DoD Field

Use: a Guide for 7ew and 7ovice Users. A Guidance Document submitted to the Environmental Security Technology Certification Program (ESTCP). U.S.

Department of Defence.

Fassbinder, J.W.E. 2011. Development of the Bavarian magnetometer system and the history of its application on archaeological sites of Anatolia. In Drahor,

M.G. & Berge, M.A. (Eds.) Archaeological Prospection. Extended abstracts.

International Society for Archaeological Prospection (ISAP), Center for Near

Surface Geophysics and Archaeological Prospection (CNSGAP) of Dokuz Eyül

University. Archaeology and Art Publications. Izmir.

Fassbinder, J.W.E., Stanjek, H. & Vali, H. 1990. Occurrence of magnetic bacteria in soil. 7ature 343: 161-163.

Fassbinder, J.W.E. & Stanjek, H. 1993. Occurrence of bacterial magnetite in soil from archaeological sites. Archaeologia Polona 31: 117-128.

Fennö, H. & Dyhlén-Täckman, I. 2002. Vetenskapsakademiens trädgård.

Stockholms stad, RAÄ 223, Arkeologisk undersökning 2002. Stockholms

Stadsmuseum. Stockholm.

Fiedler, S., Illich, B., Berger, J. & Graw, M. 2009. The effectiveness of groundpenetrating radar surveys in the location of unmarked burial sites in modern cemeteries. Journal of Applied Geophysics 68: 380-385.

Fischer, P.M. 1980. Applications of Technical Devices in Archaeology. The use of

X-rays, microscope, electrical and electro-magnetic devices and subsurface

interface radar. Paul Åströms Förlag. Göteborg.

Forte, E. & Pipan, M. 2008. Integrated seismic tomography and ground-penetrating radar (GPR) for the high-resolution study of burial mounds (tumuli). Journal of

Archaeological Science 35: 2614-2623.

Fredén, C. 1994. Jordarterna. In Fredén, C. (Ed.) Sveriges 7ationalatlas. Berg och

jord. Sveriges Nationalatlas, Stockholm.

Freij, H. 1980. Förändring av markmaterial genom mänsklig påverkan, speciellt dess magnetiska egenskaper. Arkeologiska forskningslaboratoriet, Stockholms

Universitet, Stockholm.

Fridh, B. 1982. Geofysiska metoder vid arkeologisk prospektering i Sverige.

Geologiska institutionen, Chalmers tekniska högskola och Göteborgs universitet. Göteborg.

Fridh, B. 1983. Geofysiska metoder vid arkeologisk prospektering. Arkeologiska

förundersökningar. Metodkonferens oktober 1981, Riksantikvarieämbetet och statens historiska museer. Göteborg.

Frohlich, B. & Lancaster, W.J. 1986. Electromagnetic surveying in current Middle

Eastern archaeology: Application and Evaluation. Geophysics 51: 1414-1425.

Furingsten, A. 1984. Geofysik och arkeologisk forskning i Västsverige. In

Furingsten, A., Jonsäter, M. & Weiler, E. (Eds.) Från flintverkstad till

processindustri. De första 9000 åren i Västsverige speglade av UV Västs

undersökningar 1968-1980. Riksantikvarieämbetet. Stockholm.

Furingsten, A. 1985. Archaeology and Geophysics in West-Sweden. In Edgren, T. &

Jungner, H. (Eds.) Proceedings of the third 7ordic Conference on the

application of scientific methods in archaeology. Mariehamn, Åland, Finland, 8-

11 October 1984. Suomen muinaismuistoyhdistys. Helsinki.

Gaffney, C. 2008. Detecting trends in the prediction of the buried past: A review of geophysical techniques in archaeology. Archaeometry 50: 313-336.

101

Gaffney, C., Gaffney, V., Neubauer, W., Baldwin, E., Chapman, H., Garwood, P.,

Moulden, H., Sparrow, T., Bates, R., Löcker, K., Hinterleitner, A., Trinks, I.,

Nau, E., Zitz, T., Floery, S., Verhoeven, G. & Doneus, M. 2012. The

Stonehenge Hidden Landscapes Project. Archaeological Prospection 19: 147-

155.

Gaffney, C. & Gater, J. 2003. Revealing the buried past. Geophysics for

archaeologists. Tempus. Gloucestershire.

Geonics Limited. 1992. Geonics EM38 Ground Conductivity Meter. Operating

Manual (for EM38 Models with Digital Readout). Geonics Limited.

Mississauga.

George, O. 2010. Arkeologisk kursundersökning av det Medeltida fästet Styresholm,

Raä 17 i Torsåkers socken. Delrapport 4 inom projektet hotade kulturmiljöer

längs Västernorrlands älvar. Murberget. Länsmuseet Västernorrland.

Härnösand.

George, O. 2011. Arkeologisk forskningsundersökning av stenåldersboplatsen Raä

51:1 i Ramsele sn, Ångermanland. Murberget. Länsmuseet Västernorrland.

Härnösand.

Goodman, D. 2006. GPR Survey at Ft Frederica, Georgia, National Park Service

Workshop, May 16-20, 2006. Geophysical Archaeometry Laboratory.

Woodland Hills, Los Angeles, CA.

Goodman, D., Hongo, H., Higashi, N., Inaoka, H. & Nishimura, Y. 2007. GPR surveying over burial mounds: correction for topography and the tilt of the GPR antenna. 7ear Surface Geophysics 5: 383-388.

Goodman, D. & Nishimura, Y. 1993. A ground-radar view of Japanese burial mounds. Antiquity 67: 349-354.

Goodman, D., Nishimura, Y., Hongo, H. & Higashi, N. 2006. Correcting for

Topography and the Tilt of Ground-penetrating Radar Antennae.

Archaeological Prospection 13: 157-161.

Goodman, D., Nishimura, Y. & Rogers, J.D. 1995. GPR Time Slices in

Archaeological Prospection. Archaeological Prospection 2: 85-89.

Goodman, D., Piro, S., Nishimura, Y., Schneider, K., Hongo, H., Higashi, N.,

Steinberg, J. & Damiata, B. 2009. GPR Archaeometry. In Jol, H.M. (Ed.)

Ground Penetrating Radar. Theory and Applications. Elsevier. Amsterdam.

Gose, W.A. 2000. Palaeomagnetic Studies of Burned Rocks. Journal of

Archaeological Science 27:409-421.

Goulty, N.R. & Hudson, A.L. 1994. Completion of the Seismic Refraction survey to

Locate the Vallum at Vindobala, Hadrians Wall*. Archaeometry 36: 327-335.

Graham, I. & Scollar, I. 1976. Limitations on magnetic prospection in archaeology imposed by soil properties. Archaeo-Physika 6: 1-124.

Granlund, S. & Karlenby, L. 2007. Arkeologisk undersökning. Lilla Ulvgryt. Hur

och hägnad vid Hummelbäcken. 7ärke, Vintrosa socken, Lilla Ulvgryt 1:1, RAÄ

97. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV

Bergslagen. Örebro.

Grasmueck, M., Weger, R. & Horstmeyer, H. 2003. How dense is dense enough for a `real' 3D GPR survey? SEG Technical Program Expanded Abstracts 22: 1180-

1183.

Grön, C. & Sander, B. 2005. Arkeologisk förundersökning och undersökning. Ett

hus och tre gravrester från Svista. Väg E20. Södermanland, Eskilstuna

stadsområde, Grönsta 2:3 och Grönsta 2:5, del av RAÄ 519b.

Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Mitt.

Stockholm.

102

Gustafsson, J. & Alkarp, M. 2007. Array GPR investigation of the cathedral of

Uppsala. 7ear Surface Geophysics 5: 203-207.

Gustafsson, N.B., & Viberg, A. 2010. Under Plogdjup. Arkeologiska undersökningar inom fornlämning RAÄ 109, Linde sn, Gotland 2009 & 2010.

Rapporter från Arkeologiska Forskningslaboratoriet 16, Arkeologiska

Forskningslaboratoriet, Stockholms Universitet. Stockholm.

Gustafsson, P., Hellerström, S. & Lagerås, P. 2004. Arkeologiska för- och

slutundersökningar. Förhistoriska boplatslämningar och historiska

agrarlämningar. Skåne, Ivetofta socken, Väg 116. Riksantikvarieämbetet.

Avdelningen för arkeologiska undersökningar. UV Syd. Lund.

Hagberg, U.E. 1961. Skedemosse – En första presentation. Fornvännen 56: 237-255.

Hagberg, U.E. 1979. Öland during the Iron Age and Early Middle Ages. An

Archaeological Survey. In Näsman, U. & Wegraeus, E. (Eds.) Eketorp.

Fortification and Settlement on Öland/Sweden. The setting. Royal Academy of

Letters, History and Antiquities. Stockholm.

Hagberg, L. 2009. Arkeologisk undersökning. Stjärneberg. Ett kvartsbrott i den

mesolitiska skärgården – rituell plats under yngre bronsålder och äldre järnålder. RAÄ 310-317, Kallerstad 1:27 Rystad socken, Linköpings stad och

kommun. Östergötland. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Öst. Linköping.

Hamilton, J., Heimdahl, J. & Strucke, U. 2008. Arkeologisk undersökning. Två

bebyggelsefaser med tillhörande odlingsmark i Färsna. Riksantikvarieämbetet.

Hamilton, J. & Östlund, S. 2008. Arkeologisk undersökning. Täby nya

begravningsplats, södra delen. Gård och stensträngar från äldre järnålder.

Uppland, Täby socken, Fittja 2:12, 4:12, 4:13, 4:14, 4:15, 4:17, 4:73, 4:74,

4:75, 4:76, 4:77, 4:78. RAÄ 488. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Mitt. Stockholm.

Hedberg, A., Lundström, A., Nygren, D. & Tingstorp, S. 1996. Geofysiska mätmetoder tillämpade i ett arkeologiskt sammanhang. Avdelningen för

Tillämpad Geofysik, Tekniska Högskolan Luleå. Luleå.

Hedman, S.-D. 2003. Boplatser och offerplatser: ekonomisk strategi och

boplatsmönster bland skogssamer 700-1600 AD. Institutionen för arkeologi och samiska studier. Umeå Universitet. Umeå.

Helander, C. 2009. Särskild arkeologisk undersökning. Bronsåldershärdar i

Malmslätt inför byggnation av växlarstation för fjärrvärme inom RAÄ 137.

Kärna socken, Linköpings kommun, Östergötlands län. Riksantikvarieämbetet.

Arkeologiska uppdragsverksamheten. UV Öst. Linköping.

Henkel, H. 2006. Boken om Törnskogen. Naturskyddsföreningen i Sollentuna.

Sollentuna.

Henriksson, M. 2011. Västra Vångs by. RAÄ 116 resp. 189 Hjortsberga socken,

Ronneby kommun. Magnetometerundersökning. Blekinge museum. Karlskrona.

Herbich, T. 2011. Początki metody magnetycznej w Polskich badania ch archeologicznych. Archeologia Polski LVI: 15-43.

Heron, C. 2001. Geochemical Prospecting: In Brothwell, D.R. & Pollard, A.M.

(Eds.) Handbook of Archaeological Sciences. Wiley. Chichester.

Hesse, A. 1966. Prospections géophysiques a faible profondeur. Applications a

l'Archaeologie. Dunod. Paris.

Hesse, A. 2000. Count Robert du Mesnil du Buisson (1895-1986), a French

Precursor in Geophysical Survey for Archaeology. Archaeological Prospection

7: 43-49.

Hesse, A. 2009. Short history, strategies and practical aspects of electromagnetic detection based on the description of field. In Campana, S. & Piro, S. (Eds.)

103

Seeing the Unseen. Geophysics and landscape archaeology. Taylor and Francis

Group. London.

Hildebrand, J.A., Wiggins, S.M., Henkart, P.C. & Conyers, L.B. 2002. Comparison of seismic reflection and ground-penetrating radar imaging at the controlled archaeological test site, Champaign, Illinois. Archaeological Prospection 9: 9-

21.

Hildebrand, J.A., Wiggins, S.M., Driver, J.L. & Waters, M.R. 2007. Rapid seismic reflection imaging at the Clovis period Gault site in central Texas.

Archaeological Prospection 14: 245-260.

Hjulhammar, M. 2008. 100 nya vrak. Arkeologisk analys av geofysisk kartering

inför utökad vattenverksamhet för Slussenprojektet i Stockholm, Uppland och

Södermanland, Stockholms och 7acka kommuner. Statens maritima museer.

Stockholm.

Hjulhammar, M. 2010. Tio vrak runt Beckholmen. Arkeologisk förstudie, Uppland,

Stockholms stad, Stockholms kommun. Sjöhistoriska museet. Stockholm.

Holback, T., Lindholm, P. & Runeson, H. 2004. Bjästamon. Ett kustbundet

boplatskomplex från slutet av neolitikum. Riksantikvarieämbetet, Avdelningen för arkeologiska undersökningar, UV Mitt. Stockholm.

Holm, J. & Thorsberg, K. 2007. Arkeologisk undersökning. Knutstorp – en plats

med gropar. Mesolitiska och senare lämningar. 7ärke, Hidinge socken,

Knutstorp 1:2 och Sättran 1:1, RAÄ 107. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Bergslagen. Örebro.

Horsley, T.J. 2004. The potential of geophysical prospection techniques for archaeological field evaluation in Iceland. An exploration of the methodology, environmental conditions and archaeological factors affecting the use of archaeological prospection in Iceland. Unpublished doctoral thesis. Department of Archaeological Sciences, University of Bradford. Bradford.

Howell, M. 1966. A Soil Conductivity Meter. Archaeometry 9: 20-23.

Hulin, G. & Simon, F.X. 2012. Geophysics and preventive archaeology in France: new inter-disciplinary issues. First break 30: 51-55.

Hylén, H. 2007. Ölmstad i blickfånget. Inför planerad bostadsbebyggelse inom

fastigheten Ölmstad 2:16. Ölmstad socken i Jönköpings kommun. Jönköpings

län. Jönköpings läns museum. Jönköping.

Häringe Frisberg, K. & Renck, A-M. 2005. Arkeologisk förundersökning och

undersökning. Bronsåldersboplatsen vid Svanby – ett yttre aktivitetsområde.

Väg E4, sträckan Uppsala-Mehedeby. Uppland, Tierps socken, Svanby 5:42,

RAÄ 235. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar.

UV GAL. Uppsala.

Johansson, L. & Nylund, B. 1990. Vuollerim 6000 år. Test av magnetisk gradientmätning som metod att indikera förhistoriska lämningar. Unpublished report. Sveriges Geologiska AB (SGAB).

Jol, H.M. & Bristow, C.S. 2003. GPR in sediments: advice on data collection, basic processing and interpretation, a good practice guide. In Bristow, C.S. & Jol,

H.M. (Eds.) Ground Penetrating Radar in Sediments. The Geological Society.

London.

Jones, G. & Maki, D. 2005. Lightning-induced Magnetic Anomalies on

Archaeological Sites. Archaeological Prospection 12: 191-197.

Jordan, D. 2009. How effective is geophysical survey? A regional review.

Archaeological Prospection 16: 77-90.

Karlberg, T. & Sjöstedt, D. 2007. Archaeological Exploration in Nicaragua using

Ground Penetrating Radar. A Minor Field Study. MSc Thesis in Engineering.

104

Department of Chemical Engineering and Geosciences, Luleå University of

Technology. Luleå.

Karlenby, L., Graner, G. & Johannessen, A.-C. 2005. Arkeologisk undersökning.

Hem till stenåldersbyn. En boplats fån övergången mellan senneolitikum och

äldre bronsålder vid Eriksborg och Persbo. Västmanland, Skerike socken,

Brottberga 6:13, RAÄ 645:3. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Bergslagen. Örebro.

Kattenberg, A.E. 2008. The Application of Magnetic Methods for Dutch

Archaeological Resource Management. Institute for Geo and Bioarchaeology

(Faculty of Earth and Life Sciences) of the Vrije Universiteit. Amsterdam.

Kearey, P., Brooks, M. & Hill, I. 2002. An introduction to Geophysical Exploration.

Blackwell publishing. Oxford.

King, J.A., Bevan, B. & Hurry, R.J. 1993. The reliability of geophysical surveys at historic-period cemeteries: An example from the Plains cemetery,

Mechanicsville, Maryland. Historical Archaeology 27: 4-16.

Kjellberg, J., Qviström, L. & Viberg, A. 2010. Dubbla vapenhus vid Tierp, Bälinge och Alunda kyrkor?: En rapport från ett forskningsprojekt. Uppland: 65-74.

Kockum, J. 2006. Saltus Sanctae Mariae. Bäckaskogs kloster - arkeologiska

undersökningar under ett halvt sekel. Regionmuseet Kristianstad. Kristianstad.

Kraft, A. 2006. Arkeologisk undersökning. RAÄ 109 i Äspsätter – en

järnåldersboplats i 7ärke. 7ärke, Hidinge socken, Äspsätter 1:5, RAÄ 109.

Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV

Bergslagen. Örebro.

Kresten, P. & Ambrosiani, B. 1992. Swedish vitrified forts - a reconnaissance study.

Fornvännen 87: 1-17.

Kresten, P. & Kresten, F. 1994. Faxeholm Magnetometri. Riksantikvarieämbetet,

Byrån för Arkeologiska undersökningar. Geoarkeologiskt laboratorium.

Uppsala.

Kristiansson, S. 1996. Prospektering av Kugghamn på Björkö. In CD-uppsatser i

laborativ arkeologi 95/96, del 2. Arkeologiska forskningslaboratoriet,

Stockholms Universitet. Stockholm.

Kvamme, K.L. 2001. Current Practices in Archaeogeophysics: Magnetics,

Resistivity, Conductivity, and Ground Penetrating Radar. In Goldberg, V.T.,

Holliday, V.T. & Ferring, C.R. (Eds.) Earth Sciences and Archaeology. Kluwer

Academic/Plenum Publishers. New York.

Kvamme, K.L. 2003. Geophysical Surveys as Landscape Archaeology. American

Antiquity 68: 435-457.

Kvamme, K.L. 2006a. Integrating Multidimensional Geophysical Data.

Archaeological Prospection 13: 57-72.

Kvamme, K.L. 2006b. Magnetometry: Natures Gift to Archaeology. In Johnson, J.

(Ed.) Remote Sensing in Archaeology: An Explicitly 7orth American

Perspective. University of Alabama Press. Tuscaloosa.

Kvamme, K.L. 2008. Archaeological Prospecting at the Double Ditch State Historic

Site, North Dakota, USA. Archaeological Prospection 15: 62-79.

Kvamme, K.L. & Ahler S.A. 2007. Integrated Remote Sensing and Excavation at double Ditch State Historic Site, North Dakota. American Antiquity 72: 539-

561.

Lagerlund, H. 2000. Slagrutan i teori och praktik-vatten, jordstrålning och fornminnen. In Hansson, S.-O. & Sandin, P. (Eds.) Högskolans

lågvattenmärken. Natur och Kultur. Stockholm.

Lagerström, S. 1968. Vem är det: svensk biografisk handbok, 1969. Norstedt.

Stockholm.

105

Larsson, A. 2000. Geophysics in Archaeology. Department of geology, Göteborg

University. Göteborg.

Larsson, L.-I. 2004. Jordkikare avslöjar vår förhistoria. In Fontell, Y. & Jahn, K.

(Eds.) Arkeologi E4. Årsberättelse 2003. Utgrävningar från Uppsala till Tierp.

Upplandsmuseet. Uppsala.

Lavento, M. 1992. Arkeologiassa käytettyjen sovelletun geofysiikan menetelmien alkuvaiheista Suomessa. In Huurre, M., Halinen, P., Lavento, M. & Moisanen,

J. (Eds.) Kentältä poimittua: kirjoitelmia arkeologian alalta. Museovirasto.

Helsinki.

Lawrence, M., Oxley, I. & Bates, R.C. 2003. Geophysical techniques for maritime archaeological surveys, 9th EAGE/EEGS Meeting. EAGE.

Le Borgne, E. 1955. Susceptibilité magnétique anormale du sol superficial. Annales

de Géophysique 11: 399-419.

Le Borgne, E. 1960. Influence du feu sur les propriétés magnétiques du sol et sur celles du schiste et du granite. Annales de Géophysique 16: 159-195.

Leckebusch, J. 2000. Two- and Three-dimensional Ground-penetrating Radar

Surveys Across a Medieval Choir: a Case Study in Archaeology.

Archaeological Prospection 7: 189-200.

Leckebusch, J. 2003. Ground-penetrating Radar: A Modern Three-dimensional

Prospection Method. Archaeological Prospection 10: 213-240.

Leckebusch, J. & Rychener, J. 2007. Verification and topographic correction of

GPR data in three dimensions. 7ear Surface Geophysics 5: 395-403.

Leucci, G. 2002. Ground-penetrating radar survey to map the location of buried structures under two churches. Archaeological Prospection 9: 217-228.

Liljenstolpe, P. 1998. Kan vi sluta gräva nu? Geofysiska prospekteringsmetoder

inom arkeologin: komplement eller ersättning. Uppsala Universitet. Uppsala.

Lindberg, K.-F. 2010. Arkeologisk utredning på del av STORA DICKA 9:3 i

Folkärna socken, Avesta kommun, Dalarna. Dalarnas museum. Falun.

Lindkvist, A. & Westin, Å. 2006. Arkeologisk förundersökning och undersökning.

Råberga vid Örebro-Bofors Flygplats. 7ärke, RAÄ 19, 75, 76, 79, 80, 82 och 88

i Täby socken, fastighet Råberga. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Bergslagen. Örebro.

Lindman, G. 2006. Arkeologisk undersökning. Markutnyttjande och odling på

Latorpsplatån. Arkeologisk undersökning av fossil åkermark. 7ärke, Vintrosa

socken, Västra Via 6:1, del av RAÄ 52. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Bergslagen. Örebro.

Lindman, G. & Stibéus, M. 2000. Arkeologisk undersökning. Berghem – en

arkeologisk undersökning i en vikingatida-medeltida by i Viskans dalgång.

Västergötland, Berghems sn, Berghem 16:5, fornlämning RAÄ 87.

Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Väst.

Kungsbacka.

Lindström, J. 2000. Pseudo-arkeologi vid högskolorna. Fornvännen 95: 255-258.

Lindström, M. 2000a. Fjällen. In Lindström, M., Lundqvist, J. & Lundqvist, T.

(Eds.) Sveriges geologi från urtid till nutid. Studentlitteratur. Lund.

Lindström, M. 2000b. Kambrosilurområdena. In Lindström, M., Lundqvist, J. &

Lundqvist, T. (Eds.) Sveriges geologi från urtid till nutid. Studentlitteratur.

Lund.

Linford, N.T. 2002. Ground Penetrating Radar Survey over a Suspected Roman

Amphitheatre at Richborough Castle, Kent, UK. In Kars, H. & Burke, E. (Eds.)

The 33rd International Symposium on Archaeometry. Vrije Universiteit.

Amsterdam.

106

Linford, N.T. 2004. Magnetic Ghosts: Mineral Magnetic Measurements on Roman and Anglo-Saxon Graves. Archaeological Prospection 11: 167-180.

Linford, N.T. 2006. The application of geophysical methods to archaeological prospection. Reports on progress in physics 69: 2205-2257.

Linford, N.T., & Canti, M.G. 2001. Geophysical evidence for Fires in Antiquity:

Preliminary Results from an Experimental Study Paper Given at the EGS XXIV

General Assembley in The Hague, April 1999. Archaeological Prospection 8:

211-225.

Linford, N.T. & David, A. 2001. Study of Geophysical Surveys. In Hey, G. &

Lacey, M. (Eds.) Evaluation of Archaeological Decision-making Processes and

Sampling Strategies: European Regional Development Fund Interreg IIC-

Planarch Project. Oxford Archaeological Unit. Oxford.

Linington, R.E. 1966. Test use of a gravimeter on Etruscan chamber tombs at

Cerveteri. Prospezioni Archaeologiche 1: 37-41.

Lorra, S., Kroll, S. & Thomsen, D. 2001. Geophysical Investigations at the Uppåkra

Site (Sweden). In Larsson, L. (Ed.) Uppåkra – Centrum i analys och rapport.

Lunds Universitet. Lund.

Lowe, K.M. 2012. Review of Geophysical Applications in Australian archaeology.

Australian Archaeology 74: 71-84.

Lundberg, H. 1919. Potentialmetod för elektrisk malmletning. Jernkontorets

Annaler 5: 203-215.

Lundberg, H. 1922. Practical Experience in Electrical Prospecting. P.A. Norstedt &

Söner. Stockholm.

Lundberg, L. 1997. Vägen till Vendel – analys och prospektering. In CD-uppsatser i

laborativ arkeologi 96/97, Del 2. Arkeologiska forskningslaboratoriet,

Stockholms Universitet. Stockholm.

Lundqvist, T. 2000. Sveriges prekambriska berggrund med översikter över Baltiska skölden. In Lindström, M., Lundqvist, J. & Lundqvist, T. (Eds.) Sveriges

geologi från urtid till nutid. Studentlitteratur. Lund.

Lundqvist, T. & Bygghammar, B. 1994. Urberget. In Fredén, C. (Ed.) Sveriges

7ationalatlas. Berg och jord. Sveriges Nationalatlas. Stockholm.

Lurton, X. 2002. An Introduction to Underwater Acoustics. Principles and

Applications. Springer. Heidelberg.

Löcker, K., Seren, S., Hinterleitner, A., Scherren, P., Mauthner, F. & Wallner, M.

2011. The villa rustica of Deutschkreutz. Integrated archaeological prospection of one of the largest Roman villae in Austria. In Drahor, M.G. & Berge, M.A.

(Eds.) Archaeological Prospection. Extended abstracts. International Society for Archaeological Prospection (ISAP), Center for Near Surface Geophysics and

Archaeological Prospection (CNSGAP) of Dokuz Eyül University. Archaeology and Art Publications. Izmir.s

Löfroth, B. 1987. Georadarundersökning vid Älvnäset, Vuollerim. Unpublished report. Sveriges Geologiska AB (SGAB).

Löwenstein, M. 2010. Ås kloster ringas in med radar. http://hn.se/nyheter/varberg/1.830062-as-kloster-ringas-in-med-radar. Hallands

Nyheter, Varberg.

Lück, E., Callmer, J. & Skånberg, T. 2003. The House of Bailiff of Sövestad,

Sweden – a Multi-method Geophysical Case Study. Archaeological Prospection

10: 143-151.

Maki, D. 2005. Lightning Strikes and Prehistoric Ovens: Determining the Source of

Magnetic Anomalies Using Techniques of Environmental Magnetism.

Geoarchaeology 20: 449-459.

107

Maki, D., Homburg, J.A. & Brosowske, S.D. 2006. Themally Activated

Mineralogical Transformations in Archaeological Hearths: Inversion from

Maghemite Fe2O4 Phase to Haematite Fe2O4 Form. Archaeological

Prospection 13: 207-227.

Marshak, S. 2008. Earth: portrait of a planet. Norton. New York.

Marshall, A. 1998. Visualising Burnt Areas: Patterns of Magnetic Susceptibility at

Guiting Power 1 Round Barrow (Glos., UK). Archaeological Prospection 5:

159-177.

Masters, W. 1998. A geophysical investigation of the Viking age town of Birka,

Sweden. Undergraduate thesis. Department of Geophysics, Uppsala University.

Uppsala.

Mattsson, U. 1997. Analys av den förhistoriska bebyggelsen i Burunge by, Vendel sn. Inventering, elektromagnetisk kartering och fosfatanalys. In CD-uppsatser i

laborativ arkeologi 96/97, Del 2. Arkeologiska forskningslaboratoriet,

Stockholms Universitet. Stockholm.

McNeill, J. 1980a. Electrical Conductivity of Soils and Rocks. Technical Note TN-5.

Geonics Limited. Mississauga, Ontario, Canada.

McNeill, J. 1980b. Electromagnetic Terrain Conductivity Measurement at Low

Induction 7umbers. Technical Note TN-6. Geonics Limited. Mississauga,

Ontario, Canada.

McNeill, J. & Bosnar, M. 1999. Application of 'Dipole-Dipole' Electromagnetic

Systems for Geological Depth Sounding. Technical Note TN-31. Geonics

Limited. Mississauga, Ontario, Canada.

Mercer, E. & Schmidt, A. 2001. A magnetometer survey of an Iron Age settlement site at Uppåkra, Skåne, Sweden. In Larsson, L. (Ed.) Uppåkra – Centrum i

analys och rapport. Lunds Universitet. Lund.

Metwaly, M., Green, A.G., Horstmeyer, H., Maurer, H., Abbas, A.M. & Gh.

Hassaneen, A-R. 2005. Combined Seismic Tomographic and Ultrashallow

Seismic Reflection Study of an Early Dynastic Mastaba, Saqqara, Egypt.

Archaeological Prospection 12: 245-256.

Meyer, C., Ullrich, B. & Barlieb, C.D.M. 2007. Archaeological questions and geophysical solutions: ground-penetrating radar and induced polarization investigations in Munigua, Spain. Archaeological Prospection 14: 202-212.

Meyerson, Å. 1982-1984. Erik J Lundberg. In Lager-Kromnow, B. (Ed.) Svenskt

biografiskt lexikon. Bd 24, Ljungdahl-Malmros. Svenskt biografiskt lexikon.

Stockholm.

Milsom, J. 1989. Field Geophysics. Open University Press. Milton Keynes.

Misiewicz, K. 2007. Rapport from the geophysical survey of St. Nicolaus and St.

Gertruda churches at Sigtuna, Institute of Archaeology and Ethnology, Warsaw.

Molin, F. 2006. Arkeologisk förundersökning och slutundersökning. Trädgårdstorp.

Boplatslämningar från senneolitikum och bronsålder. Arkeologiska undersökningar inför återvinningscentral samt nytt skyttecentrum. RAÄ 126-

128, Värö 3:1, Ryd 1:1 Kärna socken, Linköpings kommun Östergötland.

Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Öst.

Linköping.

Mullins, C. 1974. The magnetic properties of the soil and their application to archaeological prospecting. Archaeo-Physika 5: 143-347.

Mussett, A.E. & Kahn, A.M. 2000. Looking into the Earth: an introduction to

geological geophysics. Cambridge University Press. Cambridge.

Möller, J.-T., Jörgensen, M-S. & Höstmark, J. 1984. (Eds.) Arkaeologi og geofysiske

sporingsmetoder. Nationalmuseet. Köpenhamn.

108

Neubauer, W. 2001. Magnetische Prospektion in der Archäologie. Vlg. der

Österreichischen Akad. der Wissenschaften. Wien.

Neubauer, W. 2004. GIS in Archaeology – the Interface between Prospection and

Excavation. Archaeological Prospection 11: 159-166.

Neubauer, W. & Eder-Hinterleitner, A. 1997. Resistivity and Magnetics of the

Roman Town Carnuntum, Austria: an Example of Combined Interpretation of

Prospection Data. Archaeological Prospection 4: 179-189.

Neubauer, W., Eder-Hinterleitner, A., Seren, S. & Melichar, P. 2002. Georadar in the Roman Civil Town Carnuntum, Austria: An Approach for Archaeological

Interpretation of GPR Data. Archaeological Prospection 9: 135-156.

Ní Chíobháin, D. 2010. Jakten på det försvunna skeppet. Arkeologisk

provundersökning Tvet 1:3 och 2:8, Tegneby socken, Orust kommun. Bohusläns museum. Uddevalla.

Nilsson, B. 1973. Using Pulse-Radar for prospecting with a large penetration, a theoretical study. Boliden Metall AB, dept GP 73069. Boliden.

Nilsson, B. 1976. Geophysical prospecting by Radar, Swedish board of Technical

Development (STU). Stockholm.

Nilsson, B. 1978. Two topics in electromagnetic radiation field prospecting.

Department of Civil, Mining and Environmental Engineering, Applied

Geophysics. Luleå University of technology. Luleå.

Nishimura, Y. & Goodman D. 2000. Ground-penetrating Radar Survey at Wroxeter.

Archaeological Prospection 7: 101-105.

Norling, E. 1994. Fossilförande berggrund utanför fjällkedjan. In Fredén, C. (Ed.)

Sveriges 7ationalatlas, Berg och jord. Sveriges Nationatlas. Stockholm.

Novo, A., Grasmueck, M., Viggiano, D.A. & Lorenzo, H. 2008. 3D GPR in

Archaeology: What can be gained from dense Data Acquisition and Processing.

12th International Conference on Ground Penetrating Radar. Birmingham.

Novo, A., Lorenzo, H., Rial, F.I. & Solla, M. 2010. Three-dimensional groundpenetrating radar strategies over an indoor archaeological site: Convent of Santo

Domingo (Lugo, Spain). Archaeological Prospection 17: 213-222.

Nylén, E. 1972. Silver i Plogfåran. Smycken, mynt och andra gotländska fornfynd.

Sveriges Radio. Stockholm.

Ohlsén, M. 2006. Projekt Krokeks kloster – en förstudie. Krokeks ödekyrkogård och

klosterruin. RAÄ 28, Krokeks socken, 7orrköpings kommun, Östergötlands län.

Östergötlands länsmuseum. Linköping.

Ohlsson, A. 2010. Kalmar domkyrka. Gravkammare i kyrkans kor. Kalmar

domkyrka, Kalmar stad och kommun, Kalmar län. Riksantikvarieämbetet.

Avdelningen för arkeologiska undersökningar. UV Öst. Linköping.

Ovenden, S.M. 2010. Geophysics: A waste of time and money? Seminar on

Geophysics. ALGAO Scotland, Edinburgh.

Padín, J., Martín, A. & Anquela, A.B. 2012. Archaeological microgravimetric prospection inside don church (Valencia, Spain). Journal of Archaeological

Science 39: 547-554.

Parchas, C. & Tabbagh, A. 1978. Simultaneous measurements of electrical conductivity and magnetic susceptibility of the ground in electromagnetic prospecting. Archaeo-Physica 10: 682-691.

Passmore, A.D. & Jones, A.J. 1930. An Aid to Excavators. The Antiquaries Journal

X: 389.

Persson, K. 1996. Kort beskrivning av markradar, magnetometer och metalldetektor.

Forensisk arkeologi. Rikspolisstyrelsen. Arbetsgruppen För Forensisk

Arkeologi (AFFA). Stockholm.

109

Persson, K. 2003. Geofysisk prospektering med georadar vid S:t Nicolaus och S:t

Olofs kyrkor, Sigtuna, 3D-modellering. Unpublished report. Arkeologiska forskningslaboratoriet, Stockholms Universitet. Stockholm.

Persson, K. 2004. Geofysisk prospektering Nydala kloster. Unpublished report.

Arkeologiska forskningslaboratoriet. Stockholm.

Persson, K. 2005a.

Integrated Geophysical-Geochemical Methods For

Archaeological Prospecting. Department of Land and Water engineering. Royal

Institute of Technology. Stockholm.

Persson, K. 2005b. Geofysisk prospektering Krokek kloster. Unpublished report.

Arkeologiska forskningslaboratoriet. Stockholm.

Persson, K. 2005c. Geofysisk Prospektering Vittskövle. Unpublished report.

Arkeologiska forskningslaboratoriet. Stockholm.

Persson, K. & Olofsson, B. 1995. Kan modern teknik avslöja var båtgravsfolket bodde? Populär arkeologi 1: 12-13.

Persson, K. & Olofsson, B. 2004. Inside a mound: Applied geophysics in archaeological prospecting at the Kings' Mounds, Gamla Uppsala, Sweden.

Journal of Archaeological Science 31: 551-562.

Petersson, W. 1907. Om malmsökande medelst elektricitet. Jernkontorets Annaler 2

& 3: 153-171.

Petterson, C. & Winroth, L. 2011a. I Guvernörens grannskap. Georadar-

undersökning 2011 inom slottsområdets södra del, RAÄ 137, Jönköpings stad

och kommun. Jönköpings län. Jönköpings läns museum. Jönköping.

Petterson, C. & Winroth, L. 2011b. Under Hamnparkens gatsten. Kartering med

georadar inom förborgsområdet, Jönköpings slott RAÄ 137, Jönköpings stad

och kommun, Jönköpings län. Jönköpings läns museum. Jönköping.

Philpot, F.V. 1973. An improved gradiometer for archaeological surveys.

Prospezioni Archaeologiche 7-8: 99-105.

Piro, S. 2009. Introduction to geophysics for archaeology. In Campana, S. & Piro, S.

(Eds.) Seeing the Unseen. Geophysics and landscape archaeology. Taylor &

Francis Group. London.

Poscetti, V., Neubauer, W. & Trinks, I. 2011. Archaeological interpretation of GPR data-State-of-the-art and the road ahead. In Drahor, M.G. & Berge, M.A. (Eds.)

Archaeological Prospection. Extended abstracts. International Society for

Archaeological Prospection (ISAP), Center for Near Surface Geophysics and

Archaeological Prospection (CNSGAP) of Dokuz Eyül University. Archaeology and Art Publications. Izmir.

Powlesland, D. 2009. Why bother? Large scale geomagnetic survey and the quest for "Real Archaeology". In Campana, S. & Piro, S. (Eds.) Seeing the Unseen.

Geophysics and Landscape archaeology. Taylor & Francis Group. London.

Powlesland, D., Lyall, J., Hopkinson, G., Donoghue, D., Beck, M., Harte, A. &

Stott, D. 2006. Beneath the Sand – Remote Sensing, Archaeology, Aggregates and Sustainability: a Case Study from Heslerton, the Vale of Pickering, North

Yorkshire, UK. Archaeological Prospection 13: 291-299.

Ralph, E.K. 1964. Comparison of a Proton and a Rubidium Magnetometer for

Archaeological Prospecting. Archaeometry 7: 20-27.

Rausing, G. 1971. Arkeologien som naturvetenskap. CWK Gleerup. Lund.

Reynolds, J.M. 1997. An Introduction to Applied and Environmental Geophysics.

John Wiley & Sons Inc. Chichester.

Riisager, P., Abrahamsen, N. & Rytter, J. 2003. Research report: Magnetic

Investigations and the Age of a Medieval Kiln ar Kungahälla (south-west

Sweden). Archaeometry 45: 675-684.

110

Rundkvist, M. 2008. För en liberalisering av de svenska metallsökarreglerna.

Fornvännen 103: 118-122.

Saarenketo, T. 2009. NDT Transportation. In Jol, H.M. (Ed.) Ground Penetrating

Radar. Theory and Applications. Elsevier. Amsterdam.

Sabel, E. 2006. Arkeologisk landskapsanalys och prospektering av bebyggelselämningar och gravfält vid Alsike hage. Unpublished seminar paper.

Arkeologiska forskningslaboratoriet, Stockholms Universitet. Stockholm.

Salvi, A. 1970. Perfectionnements Apportés aux Magnetomètres a Résonance

Magnétique Nucléaire a Pompage Électronique. Revue de Physique Appliquée

5: 131-134.

Sambuelli, L. & Strobbia, C. 2002. The Buffon's needle problem and the design of a geophysical survey. Geophysical Prospecting 50: 403-409.

Sanglert, C.J. 1995. Kyrkbacken i Vendel och några av dess hemligheter.

Arkeologiska forskningslaboratoriet. In CD-uppsatser i laborativ arkeologi

94/95, Volym II. Stockholms Universitet. Stockholm.

Sanmark, A. & Semple, S. 2008. Places of Assembly: New Discoveries in Sweden and England. Fornvännen 103: 245-259.

Schleifer, N., Weller, A., Schneider, S. & Junge, A. 2002. Investigation of a Bronze

Age plankway by spectral induced polarization. Archaeological Prospection 9:

243-253.

Schurr, M.R. 1997. Using the concept of the learning curve to increase the productivity of geophysical surveys. Archaeological Prospection 4: 69-83.

Schwartz, G. 2005. Magnetometermätning av Gerhard Schwartz, SGU. In Eriksson,

T. & Östling, A. (Eds.) Ryssgärdet i Onslunda: ett fornlämningskomplex från

senneolitikum till och med 1700-talet med tyngdpunkt i bronsålder: väg E4, sträckan Uppsala-Mehedeby: Uppland, Tensta socken, Onslunda 5:1 och 3:1,

RAÄ 435: arkeologisk undersökning. Riksantikvarieämbetet, Avdelningen för arkeologiska undersökningar. UV GAL. Uppsala.

Scollar, I. 1959. Einführung in die Widerstandsmessung, eine geophysikalische

Methode zur Aufnahme von archäologischen Befunden unter der Erdoberfläche.

Bonner Jahrbücher 159: 284-313.

Scollar, I. 1961. Magnetic Prospecting in the Rhineland. Archaeometry 4: 74-75.

Scollar, I., Tabbagh, A., Hesse, A. & Herzog, I. 1990. Archaeological Prospecting

and Remote Sensing. Cambridge University Press. Cambridge.

Seiler, A. 2005. Arkeologisk undersökning. Vallby norra. En boplats från yngre

bronsålder, yngre järnålder och tidig medeltid i Tämnaråns dalgång. Väg E 4, sträckan Uppsala-Mehedeby. Uppland, Tierps socken, Fäcklinge 2:10, Vallby

1:3 och 2:4, raä 231. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV GAL. Uppsala.

Seiler, A. & Östling, A. 2008. Bönder, stormän och bronsgjutare. Senneolitikum,

bronsålder, järnålder och historisk tid vid Skuttunge kyrka Uppland, Skuttunge

socken, Skuttunge 11:1, RAÄ 317. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Uppsala. Uppsala.

Seren, S., Eder-Hinterleitner, A., Neubauer, W., Löcker, K. & Melichar, P. 2007.

Extended comparison of different GPR systems and antenna configurations at the Roman site of Carnuntum. 7ear Surface Geophysics 5: 389-394.

Silver, K. 1991. Slagrutan som arkeologisk metod. Institutionen för arkeologi,

Uppsala Universitet. Uppsala.

Silver, K. 2004. Slagrutan som arkeologiskt slagträ. In Andersson, P. & Welinder, S.

(Eds.) Mellan Tomsen och Däniken. Oenig diskussion kring alternativ

arkeologi. Bricoleur Press. Lindome.

111

Simpson, D., Van Meirvenne, M., Saey, T., Vermeersch, H., Bourgeois, J., Lehouck,

A., Cockx, L. & Vitharana, U.W.A. 2009. Evaluating the Multiple Coil

Configurations of the EM38DD and DUALEM-21S Sensors to Detect

Archaeological Anomalies. Archaeological Prospection 16: 91-102.

Stavrum, B. 1997. Förhistoriskt markutnyttjande i Borg på Björkö. En prospektering med fosfatkartering och elektromagnetisk kartering. In CD-uppsatser i laborativ

arkeologi 96/97, Del 1. Arkeologiska forskningslaboratoriet, Stockholms

Universitet. Stockholm.

Steineke, M. 2008. Utbyggnadsområdet Vintrie – område A och C. Bunkeflo socken

i Malmö stad, Skåne Län. Malmö kulturmiljö. Enheten för Arkeologi. Malmö.

Stenvall, J. 2007. Arkeologisk för- och slutundersökning. En Vråboplats i Kimstad.

Inför ombyggnad till ny pendeltågstation i Kimstad RAÄ 258 – mesolitisk och neolitisk boplats inom fastigheten Ask 5:1, invid järnvägen och länsväg 215.

Kimstads socken, 7orrköpings kommun Östergötland. Riksantikvarieämbetet.

Avdelningen för arkeologiska undersökningar. UV Öst. Linköping.

Stenvall, J. 2010. Särskild arkeologisk undersökning. En liten bit av Ektorps och

Torps förhistoria. Inför ny VA-ledning mellan Skärblacka och 7orrköping. RAÄ

19, 25 7orrköping. 7orrköpings (Borgs) socken, 7orrköpings kommun,

Östergötlands

län. Riksantikvarieämbetet. Arkeologiska Uppdragsverksamheten. UV Öst. Linköping.

Stephens, M., Beckholmen, M. & Zachrisson, E. 1994. Kaledoniderna. In Fredén, C.

(Ed.) Sveriges 7ationalatlas. Berg och jord.. Sveriges Nationalatlas. Stockholm.

Stern, W. 1929. Versuch einer elektrodynamischen Dickemessung von Gletschereis.

Beiträge zur Geophysik 23: 292-333.

Sternberg, R.S. 2008. Archaeomagnetism in Archaeometry – A semi-centennial review*. Archaeometry 50: 983-998.

Stolpe, H. 1888. Björkö i Mälaren. En vägledning för besökande. P.A. Norstedt &

Söner. Stockholm.

Streiffert, J. 2000. Arkeologisk undersökning. Förhistoriska gårdslämningar i

Onsala. Halland, Kungsbacka kn, Onsala sn, Lunden 1:121, del av fornlämning

369. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV

Väst. Kungsbacka.

Sträng, T. 1995. Geophysical Methods Applied to Two Archaeological Sites, Skåne,

Sweden. Division of applied geophysics. Luleå University of Technology.

Luleå.

Strömberg, B. 2001. Arkeologisk undersökning och förundersökning. Spår av

boplatser i Vallda socken – arkeologiska undersökningar i samband med utbyggnad av Toråsskolan. Halland, Kungsbacka kommun, Vallda socken,

Vallda

45:1. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Väst. Kungsbacka.

Stålberg, K. 2000. Hade Garnisonen en hamn? In CD-uppsatser i laborativ

arkeologi 99/00, Del 1. Arkeologiska forskningslaboratoriet. Stockholms

Universitet. Stockholm.

Sundberg, K., Lundberg, H. & Eklund, J. 1923. Electrical Prospecting in Sweden.

Sveriges Geologiska Undersökning. Stockholm.

Sundberg, K., Sillén, P. & Skyllberg, E. 2007. Arkeologisk undersökning. Gravarna

vid sundet. Södermanland, Toresunds socken, Toresunds-Sundby 1:1, del av

RAÄ 121-122 och RAÄ 299. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Mitt. Stockholm.

Sundström, J. 1985. Förundersökning vid fornl nr 181, Vivallen, Tännäs sn,

Härjedalen. Vikingatida - tidig medeltida flatmarksgravfält och kulturlager.

Sörsamiska projektet. Jämtlands läns museum. Östersund.

112

Svensson, H. & Söderberg, B. 2009. Dumpad kunskap? Om metallsökning och uppdragsarkeologins villkor. Fornvännen 104: 131-136.

Tabbagh, A. 1984. On the comparison between magnetic and electromagnetic prospection methods for magnetic features detection. Archaeometry 26: 171-

182.

Tabbagh, A. 1985. The response of a three-dimensional magnetic and conductive body in shallow depth electromagnetic prospecting. Geophysical journal of the

Royal Astronomical Society 81: 215-230.

Tabbagh, A. 1986a. Applications and advantages of the Slingram electromagnetic method for archaeological prospecting. Geophysics 51: 576-584.

Tabbagh, A. 1986b. What is the best coil orientation in the slingram electromagnetic prospecting method? Archaeometry 28: 185-196.

Tabbagh, A. 1994. Simultaneous Measurement of Electrical Conductivity and dielectric Permittivity of soils using a Slingram Electromagnetic device in medium frequency range. Archaeometry 36: 159-170.

Tabbagh, A., Bossuet, G. & Becker, H. 1988. A Comparison between Magnetic and

Electromagnetic Prospection of a Neolithic Ring Ditch in Bavaria.

Archaeometry 30: 132-144.

Tabbagh, J. 2003. Total field magnetic prospection: are vertical gradiometer measurements preferable to single sensor survey? Archaeological Prospection

10: 75-81.

Tegnér, G. 2008. Gåtan Gråborg. In Tegnér, G. (Ed.) Gråborg på Öland. Om en

borg, ett kapell och en by. Kungliga Vitterhets Historie och Antikvitets

Akademien. Stockholm.

Thalén, R. 1879. Om undersökning af jernmalmfält medelst magnetiska mätningar.

Akad. bokh. Upsala.

Thompson, R. & Oldfield, F. 1986. Environmental magnetism. Allen & Unwin.

London.

Tiberg, E. 1884. Om magnetiska mätningar vid jernmalmfält medelst den magnetiska inklinationsvågen. Jernkontorets Annaler 39: 29-66.

Tite, M.S. & Mullins, C. 1970. Electromagnetic prospecting on archaeological sites using a soil conductivity meter. Archaeometry 12: 97-104.

Tite, M.S. & Mullins, C. 1971. Enhancement of the Magnetic Susceptibility of Soils on Archaeological Sites. Archaeometry 13: 209-219.

Trinks, I. 2005. Ground Penetrating Radar surveys for Archaeological Prospection

at Skänninge 2005. Methodology, data acquisition, results, interpretation.

Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV

Teknik. Hägersten.

Trinks, I. 2006. Survey report. Ground Penetrating Radar survey for Archaeological

Prospection at Mälsåker Slott 2006. Methodology, data acquisition, results,

interpretation. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Teknik. Hägersten.

Trinks, I. & Biwall, A. 2008. Arkeologisk prospekteringsrapport. Göteborg

Skeppsbron. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Teknik. Hägersten.

Trinks, I. & Biwall, A. 2011a. Lightning-induced Remanent Magnetisation as

Plausible Explanation for a Geophysical Anomaly at Gråborg. Fornvännen 106:

350-354.

Trinks, I. & Biwall, A. 2011b. Arkeologisk prospektering av fornlämningsmiljön

Gamla Uppsala. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Teknik. Hägersten.

113

Trinks, I. & Biwall, A., Hinterleitner, A. & Larsson, L.-I. 2009e. Arkeologisk

prospekteringsundersökning. Georadarundersökning i kvarteret Magistern.

RAÄ 93, Kalmar län, stad och kommun. Riksantikvarieämbetet. Arkeologiska uppdragsverksamheten. UV Teknik. Hägersten.

Trinks, I., Fogelberg, A. & Hinterleitner, A. 2009f.

Arkeologisk prospekteringsundersökning med magnetometer på Mälby gamla bytomt. RAÄ

Tillinge 327. Uppsala län, Enköping kommun, Tillinge

socken.

Riksantikvarieämbetet. Arkeologiska uppdragsverksamheten. UV Teknik.

Hägersten.

Trinks, I., Fogelberg, A., Karlsson, P. & Hinterleitner, A. 2009b. Arkeologisk

Prospekteringsundersökning. Arkeologisk undersökning med georadar vid

Skänninge Hospital. RAÄ 2, Östergötlands län, Mjölby kommun, Skänninge

socken. Riksantikvarieämbetet. Arkeologiska Uppdragsverksamheten. UV

Teknik. Hägersten.

Trinks, I., Gansum, T. & Hinterleitner, A. 2010a. Mapping iron-age graves in

Norway using magnetic and GPR prospection. Antiquity 84.

Trinks, I., Johansson, B., Gustafsson, J., Emilsson, J., Friborg, J., Gustafsson, C.,

Nissen, J. & Hinterleitner, A. 2010b. Efficient, Large-scale Archaeological

Prospection using a True Three-dimensional Ground-penetrating Radar Array

System. Archaeological Prospection 17: 175-186.

Trinks I., Karlsson P., Biwall A., Hinterleitner A. 2009c. Mapping the urban subsoil using ground penetrating radar – challenges and potentials for archaeological prospection. In Marguerie, D. and Lanos, P. (Eds.) ArchéoSciences. Mémoire du

sol, espace des hommes. Presses Universitaires de Rennes. Paris.

Trinks, I., Karlsson, P., Biwall, A., Larsson, L.-I. & Hinterleitner, A. 2009a.

Arkeologisk prospekteringsundersökning. Arkeologisk prospektering med georadar och magnetometer vid 7ibble, Tillinge socken. Uppland, Enköping

kommun, Tillinge socken, RAÄ 335. Riksantikvarieämbetet. Arkeologiska uppdragsverksamheten. UV Teknik. Hägersten.

Trinks, I., Karlsson, P., Hedvall, R. & Hinterleitner, A. 2010c. Arkeologisk

Prospekteringsundersökning. Arkeologisk underökning med georadar invid S:t

Olofs konvent. Östergötlands län, Mjölby kommun, Skänninge stad, Allhelgona

socken. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar.

UV Teknik. Hägersten.

Trinks, I., Karlsson, P. & Hinterleitner, A. 2008a. Georadarprospektering på

Krokeks ödekyrkogård. Östergötland, 7orrköpings kommun, Krokek socken,

RAÄ 28:3, 28:5. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Teknik. Hägersten.

Trinks, I., Karlsson, P. & Hinterleitner, A. 2008b.

Arkeologisk prospekteringsundersökning. Georadarundersökning i kvarteret Lyckan,

Spinnrocken och Tulpanen i 7orrköping. RAÄ 96, Östergötlands län,

7orrköpings stad och

kommun. Riksantikvarieämbetet. Arkeologiska uppdragsverksamheten. UV Teknik. Hägersten.

Trinks, I., Karlsson, P., Stibéus, M., Ternström, C. & Hinterleitner, A. 2009d. GPR prospection at Borgholm castle, Öland, Sweden. In Marguerie, D. & Lanos, P.

(Eds.) ArchéoSciences. Mémoire du sol, espace des hommes. Presses universitaires de Rennes. Rennes.

Trinks, I., Larsson, L.-I. & Eder-Hinterleitner, A. 2007. Mapping of Sweden's first town Birka using georadar and magnetometer prospection. Študijné Zvesti

Archeologické Ústavu Sav 41: 245-246.

114

Trotzig, G. 2001. On the (Revised) European Convention on the Protection of the

Archaeological Heritage, presented in Malta 1992. The European Archaeologist

15: 1-3.

Udphuay, S., Paul, V.L., Everett, M.E. & Warden, R.B. 2010. Ground-penetrating radar imaging of twelfth century Romanesque foundations beneath the thirteenth century Gothic abbey church of Valmagne, France. Archaeological Prospection

17: 199-212.

Ulriksen, P. 1982. Application of Impulse Radar to Civil Engineering. Doctoral thesis. Department of Engineering Geology, Lund University of Technology.

Lund.

Vaara, R. 2004. Megalitgravar och flatmarksgravar på Öland under mellanneolitikum. En jämförande analys mellan Resmo och Köpingsvik, samt geofysisk prospektering av Raä 85 och Mysinge hög i Resmo sn, Öland och Raä

275 i Grödinge sn, Södermanland. Unpublished seminar paper. Arkeologiska forskningslaboratoriet, Stockholms Universitet. Stockholm.

Van Dam, R.L. & Schlager, W. 2000. Identifying causes of ground-penetrating radar reflections using time-domain reflectometry and sedimentological analyses.

Sedimentology 47: 435-449.

Van Dam, R.L., Van den Berg, E.H., Van Heteren, S., Kasse, C., Kentner, J.A.M. &

Groen, K. 2002. Influence of organic matter in soils on radar-wave reflection: sedimentological implications. Journal of Sedimentary Research 72: 341-352.

Van Leusen, M. 1998. Dowsing and archaeology. Archaeological Prospection 5:

123-138. von Bandi, H.G. 1945. Archäologische Erforschung des zukünftigen Stauseegebietes

Rossens-Broc. Jahrbuch der Schweizerischen Gesellschaft für Urgeschichte 36:

100-106.

Viberg, A. 2007. Framtidens forntid: Geofysisk och geokemisk prospektering av järnåldersgården RAÄ 108, Fresta sn, Uppland. Unpublished seminar paper.

Arkeologiska forskningslaboratoriet, Stockholms Universitet. Stockholm.

Viberg, A. 2008. Stensborg: Geofysisk och geokemisk prospektering av RAÄ

527,Grödinge socken, Södermanland. Unpublished report. Arkeologiska forskningslaboratoriet. Stockholm.

Viberg, A., Schultzén, J. & Wikström, A. Manuscript. GPR measurements and photogrammetry for the reconstruction of the spatial layout of the Church of St.

Lars, Sigtuna, Sweden.

Viberg, A., Trinks, I. & Lidén, K. 2009. Archaeological prospection in the Swedish mountain tundra region. In Marguerie, D. & Lanos, P. (Eds.) ArchéoSciences.

Mémoire du sol, espace des hommes. Presses universitaires de Rennes. Rennes.

Vickers, R.S. & Dolphin, L.T. 1975. A Communication on an Archaeological Radar

Experiment at Chaco Canyon, New Mexico. MASCA 7ewsletter 11: 6-8.

Vickers, R.S., Dolphin, L.T. & Johnson, D. 1976. Archaeological Investigations at

Chaco Canyon Using a Subsurface Radar. In Lyons, T.R. (Ed.) Remote Sensing

Experiments in Cultural Resource Studies. 7on-destructive Methods of

Archaeological Exploration, Survey, and Analysis. Reports of the Chaco Center.

National Park Service. U.S. Department of the Interior and University of New

Mexico. Albuquerque.

Visser, C.A., Gaffney, C. & Hessing, W.A.M. 2011. Het gebruik van geofysische

prospectietechnieken in de 7ederlandse archeologie. Inventarisatie, analyse en

evaluatie van uitgevoerde onderzoeken tussen 1996 en 2010. Vestigia Bv

Archeologie & Cultuurhistorie. Amersfoort.

115

Voss, O. & Smekalova, T. 2007. Magnetic survey on the site Uppåkra, Scåne,

Sveden. Geophysical prospection for archaeology. Unpublished report.

Fredriksberg.

Wahlström, K.-E. 1993. Three-dimensional presentation of Ground Penetrating

Radar radargrams. Department of Archaeology. University of Gothenburg.

Gothenburg.

Walker, R. 2012. Press Release: RM85 Resistance Meter is now available from

Geoscan Research. 7ewsletter of the International Society for Archaeological

Prospection 32: 12.

Watters, M. 2009. The complementary nature of geophysical survey methods. In

Campana, S. & Piro, S. (Eds.) Seeing the Unseen. Geophysics and Landscape

Archaeology. Taylor & Francis Group. London.

Weaver, W. 2006. Ground-penetrating Radar Mapping in Clay: Success from South

Carolina, USA. Archaeological Prospection 13: 147-150.

Wedmark, M. 1999. A small scale, high resolution magnetic survey at the

archaeological site of Birka, Sweden. Department of earth sciences. Solid Earth

Physics. Uppsala University. Uppsala.

Werner, S. 1937. Redogörelse för Sv geol undersöknings nykonstruerade elektriska

mätapparatur och med densamma utförda fältmätningar. Sveriges Geologiska

Undersökning. Stockholm.

Wesslén, E. 2011. Bebyggelsemönster och landtillväxt under järnåldern. Inventering med fosfat- och slingramsundersökning i Rällinge. Unpublished seminar paper.

Arkeologiska Forskningslaboratoriet. Stockholm University. Stockholm.

Westin, Å. 2006a. Arkeologisk förundersökning och undersökning. Medeltida odling

och förhistoriska boplatsspår vid Hummelbäcken. 7ärke, Vintrosa socken, Lilla

Ulvgryt 1:1, RAÄ 97. Riksantikvarieämbetet. Avdelningen för arkeologiska undersökningar. UV Bergslagen. Örebro.

Westin, Å. 2006b. Arkeologisk undersökning. Tidiga bönder vid Äspsätter. 7ärke,

Hidinge socken, Äspsätter 1:5, Hidinge-Lanna 4:11, RAÄ 88 och 110, Vintrosa

socken, Västra Via, 6:1, RAÄ 78., Riksantikvarieämbetet. Avdelningen för arkeolologiska undersökningar. UV Bergslagen. Örebro.

Weston, D.G. 2002. Soil and Susceptibility: Aspects of Thermally Induced

Magnetism within the Dynamic Pedological System. Archaeological

Prospection 9: 207-215.

Weymouth, J.W. 1986. Geophysical Methods of Archaeological Site Surveying.

Advances in Archaeological Method and Theory 9: 311-395.

Wihlborg, A. & Romberg, T. 1980. Georadar för att spåra fornlämningar? META 4:

33.

Winroth, L., Andreasson, A. & Petterson, C. 2011. Den dolda lustgården. Baron

Posses park på Rosenlund. Kartering med georadar hösten 2010 och våren

2011 inom den västra parken vid Rosenlunds herrgård. Jönköpings stad och

kommun, Jönköpings län. Jönköpings läns museum, Jönköping.

Witten, A.J. 2006. Handbook of Geophysics and Archaeology. Equinox. London.

Wynn, J.C. & Sherwood S.I. 1984. The Self-Potential (SP) Method: An Inexpensive

Reconnaissance and Archaeological Mapping Tool. Journal of Field

Archaeology 11: 195-204.

Wåhlander, L. 1997. Birkas stadsvall och dess förlängning fram till Borg. In CD-

uppsatser från Arkeologiska forskningslaboratoriet 97/98, Del 1. Arkeologiska forskningslaboratoriet. Stockholms Universitet. Stockholm.

116

Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement