(PDF) Developing A Sequential Geophysical Survey Design For Norwegian Iron Age Settlements

DEVELOPING A SEQUENTIAL GEOPHYSICAL SURVEY DESIGN FOR NORWEGIAN IRON AGE SETTLEMENTS Arne Anderson STAMNES MSc. Dissertation A dissertation submitted in partial fulfilment of the requirements for the degree of Master of Science in Archaeological Prospection Division of Archaeological, Geographical and Environmental Sciences University of Bradford September 2010 Abstract There has been an increased interest in the application of geophysical prospection techniques as a non-intrusive method within Norwegian archaeology the last few years. While some projects aim towards demonstrating the usefulness of geophysical methods there is still felt that more research with clear methodological aims should be undertaken. This dissertation aims to assess the applicability of topsoil Magnetic susceptibility, Fluxgate Gradiometers, Earth Resistance Meters and Ground Penetrating Radar have for locating Norwegian Iron Age Settlements and suggest a sequential prospection design for this type of sites. By gathering high resolution field data at a site of known archaeology, it was possible to analyse and extrapolate the data collected to evaluate the effect of decreasing resolution and the geophysical response of the archaeology present. Issues concerning choice of initial site evaluation technique, sequence of methods, resolution and the value and benefit of each set of field data on its own and in combination with each other, was addressed. The analysis shows a clear benefit of a multi-method approach. The confidence level in the interpretations increases with the application of several methods and an increased resolution. The results of different methods complement each other. As their potential is associated with their geophysical response, a survey strategy involving several geophysical techniques was suggested for similar sites. Care must be taken to ensure an adequate confidence level by addressing issues of sequence, resolution and area coverage, as well as the potential benefits of each method for locating archaeological features often associated with Iron Age settlement sites. Key Words: Survey Design, Iron Age, Norway, Multi Method, Settlements i Content Abstract ...........................................................................................................................................i Content .......................................................................................................................................... ii List of illustrations ......................................................................................................................... iv List of tables ................................................................................................................................ viii Acknowledgement ........................................................................................................................ ix 1 INTRODUCTION ................................................................................................................... 10 1.1 Aims and objectives ..................................................................................................... 11 1.2 Dissertation Outline ..................................................................................................... 12 2 RESEARCH CONTEXT ........................................................................................................... 14 2.1 The application of geophysical method within Norwegian Archaeology .................... 14 2.2 Archaeological context................................................................................................. 17 2.2.1 Settlements ......................................................................................................... 17 2.2.2 Delimiting a “site” ............................................................................................... 20 2.2.3 Settlement activity and expected archaeological features: ............................... 20 2.3 The site at Gustad ........................................................................................................ 22 2.3.1 Historical and Archaeological sources ................................................................ 22 3 METHODOLOGY .................................................................................................................. 33 3.1 Methods of Geophysical Prospection .......................................................................... 33 3.1.1 Magnetic Susceptibility ....................................................................................... 33 3.1.2 Fluxgate Gradiometer ......................................................................................... 35 3.1.3 Earth Resistance .................................................................................................. 38 3.1.4 Ground Penetrating Radar .................................................................................. 41 3.2 Survey Strategies.......................................................................................................... 45 3.2.1 Sampled area ...................................................................................................... 45 3.2.2 Magnetometer scanning vs. susceptibility ......................................................... 46 3.2.3 Sampling vs. Area coverage ................................................................................ 47 3.2.4 Choice of method and sequence ........................................................................ 48 4 FIELD SURVEY RESULTS ....................................................................................................... 51 4.1 Magnetic Susceptibility Survey .................................................................................... 53 4.2 Fluxgate Gradiometer Survey ...................................................................................... 55 4.3 Earth Resistance Survey ............................................................................................... 60 4.4 Ground Penetrating Radar ........................................................................................... 65 4.5 Data comparison .......................................................................................................... 67 4.6 Archaeological interpretation of the geophysical data ............................................... 71 4.6.1 The burial mounds .............................................................................................. 74 ii 5 ANALYSIS ............................................................................................................................. 76 5.1 Applicability to Iron Age Settlement Sites ................................................................... 76 5.1.1 Magnetic Susceptibility Survey ........................................................................... 76 5.1.2 Fluxgate Gradiometer Survey ............................................................................. 79 5.1.3 Earth Resistance Survey ...................................................................................... 86 5.1.4 Ground Penetrating Radar Survey ...................................................................... 91 5.2 Sequence ...................................................................................................................... 93 5.2.1 Initial site evaluation ........................................................................................... 93 5.2.2 Sequence of methods ......................................................................................... 95 6 CONCLUSION ....................................................................................................................... 98 6.1 Further work .............................................................................................................. 100 7 Bibliography ...................................................................................................................... 102 8 Appendix ........................................................................................................................... 111 8.1 Grid Coordinates ........................................................................................................ 111 8.2 Unprocessed Earth Resistance data........................................................................... 113 8.3 Fluxgate Gradiometer – Zero Mean Traverse test ..................................................... 114 8.4 Ground Penetrating Radar – All time slices ............................................................... 115 8.5 Ground Penetrating Radar – Spatial resolution with varying RDP values. ................ 123 8.6 Ground Penetrating Radar – RDP values of different types of soil............................ 123 8.7 Known Geophysical Surveys in Norway ..................................................................... 124 8.8 Categories used for interpretation of the geophysical data ...................................... 127 8.9 Levels of data processing in Geoplot for the different geophysical images in the Data Interpretation Chapter ........................................................................................ 129 iii List of illustrations Illustration 1: Different migration period houses found in Norway. After Løken 1992 and 1999. .................................................................................................................................................... 18 Illustration 2: Map showing the location of the site compared with a predictive model for locating Iron Age Settlements (Stamnes 2008 and 2010). ......................................................... 23 Illustration 3: The original text taken from Klüwer 1823:61-62 ................................................. 24 Illustration 4: The site at Gustad by L.D.Klüwer In: (Sjöborg 1822). ........................................... 25 Illustration 5: The sketch made by Th.Petersen of the additional mounds (Petersen 1927). .... 26 Illustration 6: Map from 1965 showing the area (Sivertsen 1965:274) ...................................... 27 Illustration 7: The site of Gustad as photographed in 2007. Photo by Lars Forseth, NTFK 2007. Taken towards SSW. ................................................................................................................... 28 Illustration 8: The site of Gustad as photographed in 2007. Photo by Lars Forseth, NTFK 2007. Taken towards NW...................................................................................................................... 28 Illustration 9: Rectified Aerial Photo with interpretations by the author. Original Photo by Lars Forseth, NTFK 2007. .................................................................................................................... 29 Illustration 10: Map showing just the interpretations. Several anomalies can be isolated. ...... 29 Illustration 11: Interpretations with the map from 1822 rectified and overlaid in a GIS by the author.......................................................................................................................................... 30 Illustration 12: The interpretations of the two houses east of the "Jutulgrave"-mound. Interpretations by the author. .................................................................................................... 31 Illustration 13: Map showing the areas surveyed with the different geophysical methods. ..... 52 Illustration 14: The topsoil magnetic susceptibility results. ....................................................... 54 Illustration 15: Raw, unprocessed image of the Fluxgate Gradiometer results. ........................ 55 Illustration 16: Image after High Pass Filtering and Zero Mean Traversing (ZMT). .................... 56 Illustration 17: Image after High Pass Filtering and Zero Mean Traversing (ZMT) and interpolating times 1 in the X direction. ..................................................................................... 57 Illustration 18: Isolated anomalies of either above 3 nt or below -2 nT. ................................... 58 Illustration 19: Anomalies classified using their highest and lowest value in nT. ...................... 58 Illustration 20: Anomalies classified using their mean values in nT. .......................................... 59 Illustration 21: Initial geophysical interpretation based on the fluxgate gradiometer results alone. .......................................................................................................................................... 59 iv Illustration 22: Raw image of the Earth Resistance results for the 0.5x0.5m with a 0.5m probe spacing. This image has been edge matched and despiked. Raw, unprocessed image is presented in the appendix. ......................................................................................................... 61 Illustration 23: Filtered image of the Earth Resistance results for the 0.5x0.5m with a 0.5m probe spacing. This image has been edge matched and despiked, as well as high- and low pass filtered......................................................................................................................................... 61 Illustration 24: Geophysical interpretation of the 0.5m probe spacing Earth Resistance data.. 62 Illustration 25: Raw image of the Earth Resistance results for the 1x0.5m with a 1m probe spacing. This image has been edge matched and despiked. Raw, unprocessed image is presented in the appendix. ......................................................................................................... 62 Illustration 26: Illustration 4: Filtered image of the Earth Resistance results for the 1x0.5m with a 1m probe spacing. This image has been edge matched and despiked, as well as high- and low pass filtered................................................................................................................................. 63 Illustration 27: Geophysical interpretation of the 1m probe spacing Earth Resistance data..... 63 Illustration 28: Time slice 7 illustrated with two different palettes. .......................................... 66 Illustration 29: Time slice 18 presented with two different palettes. ........................................ 66 Illustration 30 (left): Geophysical interpretation of all time slices compiled. ............................ 67 Illustration 31: Comparison between topsoil magnetic susceptibility and the fluxgate gradiometer results. Note the good correlation between the house and an area of increased topsoil magnetic susceptibility (MS) as well as areas where stronger gradiometer anomalies coincide with the MS. ................................................................................................................. 68 Illustration 32: Visualising of the number of singular anomalies stronger than 3 nT within each survey grid................................................................................................................................... 68 Illustration 33: Simplified results from all methods. It gives a rather complicated impression of the anomalies identified, but illustrates their spatial relationship and location compared with each other. .................................................................................................................................. 69 Illustration 34: Comparison of the Earth Resistance Anomalies with the identified maximum values of the Fluxgate Gradiometer anomalies. ......................................................................... 70 Illustration 35: Comparison of GPR results with the Earth Resistance and Fluxgate Gradiometer anomalies identified. .................................................................................................................. 70 Illustration 36: Archaeological interpretation of the geophysical survey data. ......................... 72 Illustration 37: Topsoil Magnetic Susceptibility mapping with a sample point at every 5x5m. . 77 Illustration 38: Topsoil Magnetic Susceptibility mapping with a sample point at every 10x10m. .................................................................................................................................................... 78 v Illustration 39: Topsoil Magnetic Susceptibility mapping with a sample point at every 20x20m. .................................................................................................................................................... 78 Illustration 40: Processed image with a sampling resolution of 0.5m between traverses and 0.125m along the traverse .......................................................................................................... 81 Illustration 41: Processed image with a sampling resolution of 0.5m between traverses and 0.25m along the traverses .......................................................................................................... 82 Illustration 42: Processed image from a sampling resolution of 1m between traverses and 0.25m along the traverses .......................................................................................................... 82 Illustration 43a: Detailed view of interpretation of the FG data.. .............................................. 84 Illustration 43b: Raw image. 0.5m traverse and 0.125m sample interval. ................................. 84 Illustration 43c: Filtered data. 0.5m traverse and 0.125m sample............................................. 84 Illustration 43d: Filtered and Interpolated data. interpolated in x direction to 0.25m traverse interval. 0.125m sample interval. ............................................................................................... 84 Illustration 43e: Detailed view of interpretation of the FG data.. .............................................. 85 Illustration 43f: Raw image. 0.5m traverse and 0.125m sample interval. .................................. 85 Illustration 43g: Filtered Data. 0.5m traverse and 0.25m sample interval. ................................ 85 Illustration 43h: Filtered Data. 1m traverse and 0.25m sample interval.................................... 85 Illustration 44a: Raw Earth Resistance Survey results. 0.5m traverse and sampling interval. 0.5m probe spacing..................................................................................................................... 87 Illustration 44b: Raw Earth Resistance Survey results. 0.5m traverse and 1m sampling interval. 0.5m probe spacing..................................................................................................................... 88 Illustration 44c: Raw Earth Resistance Survey results. 1m traverse and 1m sampling interval. 0.5m probe spacing..................................................................................................................... 88 Illustration 45a: Raw Earth Resistance Survey results. 1m traverse interval and 0.5m sample interval. 1m probe spacing. ........................................................................................................ 89 Illustration 45b: Raw Earth Resistance Survey results. 1m traverse interval and 1m sample interval. 1m probe spacing. ........................................................................................................ 90 Illustration 46a: Comparison of 0.25m (left) and 0.5m. (right) traverse interval Slice 7. ........... 92 Illustration 46b: Comparison of 0.25m (left) and 0.5m. (right) traverse interval Slice 8............ 92 Illustration 47: Map showing what anomalies that might have been detected by scanning every 10m with a threshold of 3nT ....................................................................................................... 94 Illustration 48: Map showing the concentration of anomalies detected by scanning compared with the anomalies identified as archaeological by all methods applied at Gustad. ................. 94 Illustration 49: Map showing the grid and survey points. Measured in with a RTK GPS with an accuracy of ±2cm ...................................................................................................................... 111 vi Illustration 50: Earth Resistance, 0.5m Probe spacing. Raw, Unprocessed data...................... 113 Illustration 51: Earth Resistance, 1m Probe spacing. Raw, Unprocessed data......................... 113 Illustration 52: Fluxgate Gradiometer Data which ha only been Zero Mean Traversed. The image shows difficulties in removing striping when having strong geological anomalies affecting the mean values calculated for each traverse. .......................................................... 114 Illustration 53: GPR slice 1 - 0-2.66 ns or 0-0.09m with V=0.07m/ns ....................................... 115 Illustration 54: GPR slice 2 – 1.62-4.28 ns or 0.06-0.15m with V=0.07m/ns ............................ 115 Illustration 55: GPR slice 3 – 3.25-5.91 ns or 0.11-0.21m with V=0.07m/ns ............................ 116 Illustration 56: GPR slice 4 – 4.88-7.53 ns or 0.17-0.26m with V=0.07m/ns ............................ 116 Illustration 57: GPR slice 5 – 6.5-9.16 ns or 0.23-0.32m with V=0.07m/ns .............................. 117 Illustration 58: GPR slice 6 – 8.12-10.78 ns or 0.28-0.38m with V=0.07m/ns .......................... 117 Illustration 59: GPR slice 7 – 9.75-12.41 ns or 0.34-0.43m with V=0.07m/ns .......................... 118 Illustration 60: GPR slice 8 – 11.38-14.03 ns or 0.4-0.49m with V=0.07m/ns .......................... 118 Illustration 61: GPR slice 9 – 13-15.66 ns or 0.46-0.55m with V=0.07m/ns ............................. 119 Illustration 62: GPR slice 10 – 14.62-17.28 ns or 0.51-0.6m with V=0.07m/ns ........................ 119 Illustration 63: GPR slice 11 – 16.25-18.91 ns or 0.57-0.66m with V=0.07m/ns ...................... 120 Illustration 64: GPR slice 13 – 19.5-22.16 ns or 0.68-0.78m with V=0.07m/ns ........................ 121 Illustration 65: GPR slice 14 – 21.12-23.78 ns or 0.74-0.83m with V=0.07m/ns ...................... 121 Illustration 66: GPR slice 15 – 22.75-25.41 ns or 0.8-0.89m with V=0.07m/ns ........................ 122 Illustration 67: GPR slice 16 – 24.38-27.03 ns or 0.85-0.95m with V=0.07m/ns ...................... 122 vii List of tables Table 1: The number of Geophysical Investigations carried out in Norway sorted by year. The database is compiled by Lars Gustavsen from NIKU in collaboration with the author. Some investigations have not been dated by year and are not included. 56 surveys in total. The database should not be considered complete. .......................................................................... 15 Table 2: The Chronology for the Norwegian Iron Age (Source: Solberg 2000)........................... 17 Table 3: Different types of archaeological features that might be found at an Iron Age Settlement Site (Løken 2009, Birgisdottir 2009, Haug 1999, Myhre 1971 and 2002). ............... 21 Table 4: Finds from Gustad in the archives of The Museum of Natural History and Archaeology in Trondheim. .............................................................................................................................. 26 Table 5: Measurements of the the two houses derived from the georectified aerial photos. .. 31 Table 6: Table showing the survey parameters used at Gustad. ................................................ 52 Table 7: Area Calculations and approximate field time .............................................................. 53 Table 8: Measured length and diameter of identified burial mounds. ...................................... 74 Table 9: Coordinates and height measurement for each point in UTM 32 coordinates. ......... 112 Table 10: Calculations of spatial resolution with varying RDP values. ..................................... 123 Table 11: Typical Relative Dielectric Permittivities (RDPs) of Common Geological Materials. From Conyers 2004:47. ............................................................................................................. 123 Table 12: Known geophysical Surveys within Norwegian Archaeology including the year 2009 .................................................................................................................................................. 126 Table 13: categories used for interpreting Fluxgate Gradiometer data ................................... 127 Table 14: Categories used for interpreting Earth Resistance Data ........................................... 128 Table 15: categories used for interpreting ground Penetrating Radar Data ............................ 128 Table 16: Levels of data processing in Geoplot for the different geophysical images ............. 130 viii Acknowledgement I would like to thank my supervisor Dr. Christopher Gaffney at the Division of Archaeological, Geographical and Environmental Sciences (AGES) at the University of Bradford for his help on this project. I am grateful for all the help I have received through discussions and comments. I would also like to thank my co-supervisor Lars Stenvik at the Section for Archaeology and Cultural History at The Museum of Archaeology and Natural Sciences (NTNU Vitenskapsmuseet) at the Norwegian University of Science and Technology (NTNU) in Trondheim for comments and help while working in Norway. A big thanks also go to all the people and institutions who have helped with access to the necessary equipment: NTNU Vitenskapsmuseet with Lars Stenvik and Øyvind Ødegård for lending me the Fluxgate Gradiometer, Magnetic Susceptibility meter, Ground Penetrating Radar, Surveying equipment and office space. The AGES department at the University of Bradford with Stuart Fox for lending me the Earth Resistance equipment. Lars Forseth at Nord-Trøndelag County Council for lending me their DGPS and PDA as well as letting me use his aerial photos from Gustad. Vitec AS for surveying the grid. Kristin Foosnæs at the Section for Archaeometry at NTNU did a great job helping me out with the fieldwork. I would also like to thank the farmer at Gustad, Arne Bjørgum, for his interest in the project and for giving me access to the site. NTNU Vitenskapsmuseet and Norsk Arkeologisk Selskap have helped with funds, which made this project realisable and I owe them a huge thanks. Jamie Vandagriff is also thanked for help with proofreading. I would like to thank my family, friends and loved ones for all their help and support. I really appreciate it. ix 1 INTRODUCTION An increased interest in archaeological prospection can be noticed in Norway the last few years but it is felt that clear aims within this field of research are lacking. While some efforts have been made to learn more about the applicability and usefulness of archaeological prospection techniques (see for instance Barton et al. 2009, Trinks 2007a,b & c ,Fossum 2010 or Foosnæs 2010), any targeted research is still in early stages. Some doubt exists about the usefulness of the methods within the Norwegian Cultural Heritage Management, and there are quite a few unresolved issues in fitting a geophysical prospection scheme within the Norwegian legislative framework as the new methods still remain relatively untested. It is hard in a “developer pays” setting to argue for both the usage of new methods and techniques in addition to traditional evaluation methods such as the soil stripping technique, which has long been regarded as adequate. When that is said, it is felt that the interest in the methods and their clear advantage in investigating and learning more about sites that will not (or should not) be excavated in a non-intrusive manner, as well as their representation of a new development within field archaeology is encouraging. The Directorate for Cultural Heritage in Norway (Riksantikvaren) has indicated that they want to encourage the development of technological aids as a valuable supplement to the methods used today within registration and excavation of archaeology, with particular emphasis on non-intrusive methods (Riksantikvaren 2010:11). By this they confirm an increased interest, and also indicate that the application of non-intrusive methods will increase the next 10 years. The emphasis on a research project with clear methodological and practical aims for increasing the knowledge of the applicability of geophysical prospection methods within Norwegian 10 archaeology is therefore an important addition to Norwegian Cultural Heritage Management. This dissertation will investigate the applicability of four different methods on a site with known archaeology assumed to be related to an Iron Age Settlement site. By analysing the response of these methods, and the information provided by modeling the effect of varying resolution, an appropriate sequential survey design for similar sites can be suggested. This will contribute to the development and adaptation of geophysical prospection techniques to the archaeology present. Amongst the research question that are appropriate for this type of site is (1) which methods provide the necessary information to locate and delimit a typical Norwegian Iron Age Settlement? (2) What kind of initial prospection strategy will be most beneficial for locating Iron Age Settlements? And (3): At what resolution does a survey need to be conducted to answer archaeological questions? 1.1 Aims and objectives The aim of this project is to investigate the applicability of four different geophysical prospection techniques on a site with known archaeological components, and by in analysing the response develop a sequential geophysical survey design for Norwegian Iron Age Settlements. The fulfillment of these aims is achieved by the following objectives: - Investigate historical and archaeological information available from the Iron Age settlement site of Gustad, Levanger Kommune, Nord-Trøndelag County (Askeladden id.number 56092) for comparison with results from a geophysical survey of the site. 11 - Evaluate the applicability of the chosen methods for Iron Age Settlements, and the effect of decreased resolution by conducting a high resolution survey using the following methods: o Magnetic Susceptibility o Fluxgate Gradiometer o Electrical Resistance o Ground Penetrating Radar - Recommend a prospection strategy for future practice on similar sites. 1.2 Dissertation Outline Chapter two will present a brief history of the application of geophysical methods in Norwegian archaeology, along with the archaeological context for this dissertation. The section on geophysical methods will sum up some of the experiences we can gather from the work that has been undertaken thus far. The archaeological context will give an introduction to Iron Age Settlements in general and the associated archaeological features to be expected, as well as the archaeological and historical background of the site of Gustad, Nord-Trøndelag in Norway. This is the site where the four methods were tested. This will be followed in chapter three by a presentation of the different methods, and a preliminary assessment of their potential to locate archaeological features often associated with Iron Age Settlement sites with regard to expected geophysical contrast and the potential. Different views on survey strategies will be presented, which include choices of initial survey strategies, choice of methods, their speed and potential resolution which are all important issues when designing a 12 prospection strategy. The field results will be presented in chapter four followed by an analysis to evaluate the applicability of the methods in chapter 5. The effect of different resolutions will be modelled, and suggestions for future application will be presented. This will lead to a suggestion on the choice of methods, their resolution and sequence when planning a survey for similar sites. Conclusions and results will be summed up in chapter 6, and suggestions for further work will be presented. 13 2 RESEARCH CONTEXT 2.1 The application of geophysical method within Norwegian Archaeology The first known application of a geophysical method within Norwegian archaeology is a test-trial of a proton magnetometer at Hoset, Stjørdal North- Trøndelag in 1972 used to map a slag-mound. It took 15 years until the next identified survey, where the Cultural Historical Museum in Oslo initiated a series of tests at Borre, Vestfold County using GPR and a magnetometer. Other people and institutions involved in different work in the recent years include Tatiana Smekalova, RB Geoark, Allied Associates, Earthsound Associates in collaboration with the Norwegian University of Science and Technology (NTNU) and UV Teknik of Riksantikvarämbetet in Sweden in collaboration with Vestfold Fylkeskommune and the Norwegian Institute for Cultural Heritage Research (NIKU). The application of geophysical techniques has been given increasing attention within Archaeology in Norway the last few years. A number of investigations have been undertaken from several institutions under their own initiative, sometimes in advance of an excavation or as a part of a research project to learn more about the applicability of the methods and the results that can be acquired. Several of these investigations have been carried out by foreign companies and experts in collaboration with local archaeologists. Not all reports are easily available, but some general points can be made as a result of compiling this database and reading through the material available: - The majority of investigations involves a single-method approach, with 39 using one method, 11 using two, and the rest with 3 or more methods 14 - Magnetometer surveys dominate (40), then the use of GPR (25). Only 8 surveys include Earth Resistance and 6 include some other technique like Earth Resistance Tomography, or the measurement of Magnetic Susceptibility or Soil Conductivity. - The resolution and technical quality has been increased the last few years - Settlement sites and burial monuments have been targeted the most, but examples of ecclesiastical and fortified sites exist, as well as outfield and glacial tests. - Few surveys have been undertaken in advance of test-excavation early in the site evaluation process, but examples of geophysics being undertaken before an actual excavation exist (Barton et al. 2009, Birgisdottir 2009, Gjerpe 2005). Number of Geophysical Investigations 29 30 25 20 15 9 10 6 6 5 5 1 0 0 0 Table 1: The number of Geophysical Investigations carried out in Norway sorted by year. The database is compiled by Lars Gustavsen from NIKU in collaboration with the author. Some investigations have not been dated by year and are not included. 56 surveys in total. The database should not be considered complete. 15 The quality and technical knowledge are generally good, but a lack of definite methodological research questions, clear aims or focused research on specifically geophysical questions within the application of geophysical methods in Norwegian archaeology is eminent. The investigations undertaken have given us some important knowledge on the geophysical responses of different types of monuments, sites and geology. An analysis of the geophysical results compared directly with excavation results either from test trenching or area excavation is missing. Specific questions that are still outstanding include: - What kind of anomaly would different typical archaeological features yield on sites in Norway? - What is the effect of different geology and subsoil? - What is the reason for either the positive or negative results? - What was not found, and why? - How can the survey strategy be adapted to better yield good results? - Is the appropriate resolution used? - Should other geophysical techniques have been used instead of the ones chosen? Initiative taken by the Directorate for Cultural Heritage (Riksantikvaren) together with the county of Vestfold and UV Teknik in Sweden has led to a more targeted research on some of these methods. Similar initiative has been taken by NTNU in collaboration with Earthsound Associates from Ireland, as well as NIKU in collaboration with the Ludwig Boltzmann Institute for Archaeological Prospection and Virtual Archaeology in Vienna. The following references can be used to illustrate some of the work being 16 undertaken (Barton et. al. 2009, Fossum 2010, Binns 1994,2002a & b and 2008, Gjerpe 2005:20, Lorra 2003, Martens 2003, Smekalova and Bevans 2009a and b, Trinks et al. 2007a b & c, Persson 2006, Foosnæs 2009 and 2010). 2.2 Archaeological context The Iron Age chronology in Norway is usually denoted as follows: Period Year General terminology Pre-Roman Iron Age 500 BC – 1 AD Older Iron Age Roman Iron Age 1 AD – 400 AD Older Iron Age The Migration Period 400 AD – 560/570 AD Older Iron Age The Merovingian Period 560/570 AD – approx. 800 AD Younger Iron Age The Viking Period Approx. 800 AD – 1030 AD Younger Iron Age Table 2: The Chronology for the Norwegian Iron Age (Source: Solberg 2000). 2.2.1 Settlements Even though the Iron Age covers such a long time span, there is a tendency to find similar archaeological features when excavating settlement sites. The typical house constructions can be considered to be a variation of the same principle: a three- aisled wooden building, with pairs of roof-bearing posts along the central aisle with some of the weight of the building resting on the wall. Walls can be made up by smaller wooden post dug into the ground, turf, stone, with wattle and daub constructed walls or with wooden sillbeams either resting on the ground or dug into a ditch. A general term for this type of building is “long house”. (Løken 1999, Myhre 2002, Solberg 2000, Østmo and Hedeager 2005:184-188). A variation in length, width, and construction observed through the Iron Age. Some general traits can be mentioned, to put eventual geophysical anomalies from Gustad into a larger 17 archaeological context and give an impression of the features that is to be expected at a Norwegian Iron-Age Settlement site. In the Pre-Roman Iron Age the houses were often relatively small- usually between 10-20m in length. They often had entrances opposing each other on each side of the building, and half of the building was often used as a barn. Sometimes no trace of fireplaces are found inside the buildings (Løken 2001, Østmo and Hedeager 2005:184-188). Illustration 1: Different migration period houses found in Norway. After Løken 1992 and 1999. 18 In the end of the Pre-Roman Iron age there is a general tendency of building larger houses, 20-50m long, with a more specialised division of rooms. Rare finds of three-aisled buildings up to 90m long are known. The average width for older iron Age houses in western parts of Norway have been calculated to 6,5-7m in width. The usage of sill-beams for constructing the walls gets more common, and this form of wall construction can sometimes not be seen in the archaeological record. About this time buildings with an outer stone wall are also found. More curved outer walls are a trait that is found more often in houses dated to this period and onwards. A majority of the Iron Age settlements identified are dated to the Roman Iron Age and the Migration Period (Løken 1999, Østmo and Hedeager 2005: 184-188). Houses from the Younger Iron Age are elusive in the archaeological record, and do not appear by far as often as in the previous periods. The introduction of the notched log constructed buildings where most or all of the weight of the roof was carried by the walls resting on sill-beams leaves fewer archaeological traces behind. Typical for the farms of the Migration Period and the Viking period is that the houses gets smaller and more specialised, where barns, smithies and other activity is moved into a separate building. Three-aisled long houses are still in use, and are known to be in use at least into the 12th century. The largest three-aisled long house ever found in Scandinavia, is the 83m long and 7,5-9m wide house dated to the Viking Age, which shows that the building tradition still exists, and that large buildings still was erected with this traditional building technique (Hansson et al. 2005:109-110, Stamnes 2008:31, Solberg 2000, Göthberg 2000, Pedersen and Widgren 1998,Myhre 2002:196, Øye 2002: 277-283, Løken 1999, Østmo and Hedeager 2005:49-50, 184-189). 19 2.2.2 Delimiting a “site” What constitutes an archaeological site? When does the settlement area turn to hinterland, and where does the outfield end and turn into a landscape? These are important questions one need to ask when trying to delimit a site. An archaeological site can be defined as an area of past human activity, and these are often identified and delimited by the distribution of physical archaeological features or the scatter of objects. Geophysical and chemical methods can indicate areas of past human activity without the remnants of any physical features as such, but rather through analysis on the soil itself. Measuring the content of trace elements such as phosphate or the mapping of magnetic susceptibility can be used to indicate human activity, and in that sense extend the former definition of an archaeological site beyond any physical features. How the word “site” is used is then defined by what you are looking for, rather than a definition that maps specific features matching set criteria. 2.2.3 Settlement activity and expected archaeological features: Archaeological features and objects from both settlement excavations and burial mounds give us a more complete picture of the activities that might have been present at a traditional settlement site. The most common features are postholes, ditches, cooking pits and different pits of refuse, butchering, latrine or unknown origin. It is quite rare to find more specialized and identifiable traces relating to some form for craftsmanship, but pit houses (“grubehus/grophus”), smithies, features relating to iron- or brass-working, 20 textile production, as well as agricultural traces are known. The cooking pits are usually interpreted as pits where food was prepared. Sometimes finds like slag or random iron objects find their way into pits (Birgisdottir 2009, Haug 1999, Løken 1992 and 1999, Stamnes 2008:34). Archaeological Typical Size Properties Frequency Feature Postholes 20-100 cm The posts might have been burnt down Common in width or charred before put into the ground. Might be stone lined. Ditches 20 cm < ? Varies depending on function. Might be Uncommon stone-lined drainage ditches, stone / Common filled wall fundament, burnt sill beams etc. Fire places 40-100 cm Fireplaces for heating. Often found Common inside, or close by buildings Cooking pits 50-300cm Often filled with coal, charred stone, Common burnt sand, ash etc. Might have been reused several times. Refuse pits Varies Not easily identified. Often pits without Common any content are found. Content might have rotted away. Likely to have been backfilled at the time of settlement, and might therefore have increased magnetic susceptibility values. Pit houses 200-400 m Might be backfilled with cultural layers Rare containing refuse or activity traces from different types of craftsmanship. Could be smithies, textile workshops, storehouse etc. Stone fences/ Varies Stone fences for keeping cattle away or Regional cattle tracks marking property boundaries. variation. Rare in Trøndelag. Table 3: Different types of archaeological features that might be found at an Iron Age Settlement Site (Løken 2009, Birgisdottir 2009, Haug 1999, Myhre 1971 and 2002). 21 2.3 The site at Gustad The site investigated for this dissertation is located on the Farm Gustad Nordre, gnr 136, bnr, 6. in Nord-Trøndelag county, Norway. The site is on a relatively flat field on the north side of the main road between Skogn and Ekne. The site is registered in the Norwegian national culture-historical monument database “Askeladden” as site number 56092. The area investigated is situated on beach deposit sediments, and the soils are predominantly umbrisols and areanosols and should within ploughing depth consist of mainly of silty sand. The soil should only have a minor capacity of retaining water (Skog og Landskap 2010). The site is considered well suited for an Iron Age Settlement when compared with a predictive model constructed for Nord-Trøndelag County (Stamnes 2008 and 2010, see illustration 2). 2.3.1 Historical and Archaeological sources Gustad is first mentioned in historical sources in the first half of the 1500s concerning taxation to the church and the monastery at the Island of Tautra, and it seems like the farmland of Gustad has been divided in a privately owned part and a part owned by the church already at this time. This would indicate that the farm had been a private property in the medieval times (in Norway the medieval period is between 1030-1537 A.D.) (Vestrum 1933:836-837). 22 Illustration 2: Map showing the location of the site compared with a predictive model for locating Iron Age Settlements (Stamnes 2008 and 2010). 23 The first archaeological record is mentioned in the book by Lorentz Didrich Klüwer from 1823: Illustration 3: The original text taken from Klüwer 1823:61-62 This can be translated to modern English as follows (translation by the author): ”There is a steep cliff by the farm Gustad1, at Ekne Parish. In front of this cliff is the remains of two square buildings that are oriented North-South, lying approximately 80 alen2 (50 meters) apart. The building towards west has been 45 alen long (28,2 meters) and 17 alen (10,7 meters) wide and the other building approximately 39 alen long (24,5 meters) and 17 alen wide 1 The name Gustad is probably derived from ”Gudestad” which means the place or farm of the gods and is concidered an old placename, at least dating back to the Viking ages if not older (source?). 2 An ”alen” is an old norsk measurement of length equal to 62,75cm. 24 (10,7 meters). In between the northern ends of these buildings is a long ”ship mound”3 of 40 alens length (25,1 meter) that still is called the ”Jutul grave” 4. In this mound a battle sword is found. Between the southern ends of the houses there is a partially excavated burial mound, and around it several other smaller and larger mounds. These two building, where which only the overgrown remains are still visible, are according to old sayings said to be the remains of a cloister, but since they are oriented towards north and is so tightly surrounded by undisturbed heathen burials I rather think that this has been a fortified settlement with a sacrificial house5 at this place” (Klüwer 1823:61-62). A map of the site at Gustad by Klüwer is published Sjöborg in 1822: Illustration 4: The site at Gustad by L.D.Klüwer In: (Sjöborg 1822). In the archives of the Museum of Natural History and Archaeology in Trondheim there is a map by the famous archaeologist Th. Petersen dated 01.01.1927. He notes that a road has cut through the southern part of Klüwers map, removing some of those stone circles seen there. The remains of the “Jutul grave” were still visible. In addition to the 3 In moderen terminology a ”ship mound” is a boat shaped long mound. 4 A ”Jutul” is another word for troll used in norse mythology 5 ”Offerhuus” in the original text. 25 mounds on Klüwers map, Th.Petersen notes another 7 mounds: 4 round, 2 long mound and 1 star shaped or triangular mound (Petersen 1927): Illustration 5: The sketch made by Th.Petersen of the additional mounds (Petersen 1927). This brings the total number of known monuments from antiquarian sources up to 7 round mounds, 3 long mounds (including the “Jutul grave”), 1 star shaped mound, 7 stone rings and 2 houses - all in all 18 possible grave monuments and 2 houses. There are four objects that have been handed in to the museum from Gustad: Museum Object Date Context number T 2356 Two double layered Oval Later half of Found with buried bone Brooches of Bronze of either the 10th and charcoal in a small the type P42 or P51 century mound together with T (Jørgensen 2008) 2363 and T 2363 Sickle of Iron Younger Iron T2364. There were 4 Age larger mound close by. T 2364 Iron ring, probably part of a Younger Iron bridle Age T 13517 Axe Blade of Iron Younger Iron Found while plowing. Find Age spot marked on the map by Th.Petersen from 1927. Table 4: Finds from Gustad in the archives of The Museum of Natural History and Archaeology in Trondheim. 26 There is no mounds to be seen by 1965 (Sivertsen 1965), leaving the conclusion that they were ploughed away sometimes between 1927 and 1965. Sivertsen (1965) places the older settlement further west in between the Vestre (Western) and Nordre (Northern) Gustad. According to Sivertsen charred stone and charcoal had been found here, as well as over a wide area on the flat surface in front of the a cliff extending towards one of the Gustad Farms, which roughly the same area as indicated by Klüwer and Petersen (Sivertsen 1951 and 1965). Illustration 6: Map from 1965 showing the area (Sivertsen 1965:274) In the summer of 2007 the County Archaeologist of North Trøndelag County Council (NTFK) did a series of aerial photographs of different sites in the area, with favourable conditions for crop marks derived from archaeology to be seen (Forseth 2007a and b). A series of photos from Gustad were taken from all angles on the 12th of July 2007. The 27 area had been used to cultivate grass for cattle fodder which had been cut 1-2 weeks before the pictures were taken. Illustration 7: The site of Gustad as photographed in 2007. Photo by Lars Forseth, NTFK 2007. Taken towards SSW. Illustration 8: The site of Gustad as photographed in 2007. Photo by Lars Forseth, NTFK 2007. Taken towards NW. By rectifying this photo in the programme AirPhoto and georeferencing them in a GIS, the following interpretations can be presented: 28 Illustration 9: Rectified Aerial Photo with interpretations by the author. Original Photo by Lars Forseth, NTFK 2007. Illustration 10: Map showing just the interpretations. Several anomalies can be isolated. 29 Adding a georectified version of Klüwers map gives the following result: Illustration 11: Interpretations with the map from 1822 rectified and overlaid in a GIS by the author. If illustration 4 and 5 is also taken into consideration, there is obvious that some archaeological features that have earlier been known to be present that are not revealed in these aerial photos. On the other hand, the photos reveal features not known from any of the previous sources. Most obvious is the eastern house site, where the walls from two different houses can be seen in the aerial photo. An interpretation of these can be presented separately: 30 Illustration 12: The interpretations of the two houses east of the "Jutulgrave"-mound. Interpretations by the author. The westernmost of the houses (house 1) seen on this photo has relatively rectangular walls, and no internal features and lacks a wall at the short end towards NNW. The easternmost house (house 2) seems to have double walls, and two anomalies that might be postholes or remnants of a wall towards the northern end. Both houses seem to be aligned in the same SSE-NNW direction. Internal measurements External measurements (maximum) (maximum) House 1 23,2x6,8m 25,3x10,2m House 2 24x6,1m6 27x14m Table 5: Measurements of the the two houses derived from the georectified aerial photos. 6 Internal measurement of house 2 is taken from the inside of the two possible postholes towards the back wall in SSE. 31 Some error due to the georectification of the original oblique photo must be assumed. Worth noticing is that the walls on House 2 curve slightly. No internal postholes can be positively identified. House 1 is relatively rectangular, and is more similar to the shape and the size described by Klüwer. The fact that the walls appear on the aerial photo in a similar shade as the ditches of the burial mounds indicate that they are a cut feature rather than a stone wall, which is the opposite of what Klüwer describes. The aerial photo reveal one clear circular anomaly NE of the houses which must be one of the circular burial mounds, as well as a long-mound which with high certainty can be interpreted as the side ditches of the mentioned “Jutul grave”. In addition several pits, possible ring ditches and amourphous anomalies can be seen. These might be caused by charcoal pits, cremation burials, postholes, ditches or any other form of ancient or modern activity. The house west of the “Jutul grave” as noted by Klüwer is not to be seen on the aerial photo. The area where it was supposed to be is in a darker shade of green, and any archaeological remnant of the house might still be there, even though it can be assumed that the southern part of it might be removed or concealed by the modern road (see illustration 11). The main strength of doing a survey at this site in particular is how well documented the archaeology is, having a research history of over 200 years ranging from old antiquarian maps to aerial photos makes the site special, and well suited for the research being undertaken in this dissertation. 32 3 METHODOLOGY By gathering a high resolution datasets with a Magnetic Susceptibility meter, Fluxgate Gradiometer, Earth Resistance and Ground Penetrating Radar from a site with known archaeology, it is possible to assess their applicability and potential benefits of using several methods when evaluating and characterising Iron Age Settlement site. By analysing the collected data by it is possible to investigate the resolution required to delimit an Iron Age Settlement site and identify its archaeological components. The value of different sampling techniques can be assessed and a suggestion of an appropriate sequence can be made. These methods will be presented, along with reasons for choosing them, their potential benefits for locating and delimiting an Iron Age settlement site and the potential resolution of each method. 3.1 Methods of Geophysical Prospection 3.1.1 Magnetic Susceptibility General Principles Magnetic Susceptibility can be defined as “the ability of a material to become temporarily magnetised” (Gaffney and Gater 2003:45). This is closely linked to the distribution of magnetic minerals in the soil, where certain iron oxides are a lot more susceptible to an external magnetic field than others. Hematite (α-Fe₂O₃) is one of the most common iron oxide and has only a weak magnetic susceptibility, but can be changed into the much more susceptible iron oxides magnetite (Fe₃O₄) and maghemite 33 (γ-Fe₂O₃) through oxidation, reduction and re-oxidation processes created by heating and burning, decay of organic waste and bacterial activity. Addition of magnetic material as refuse such as broken pottery or brick fragments will also create magnetic enhancement of the topsoil (Aspinall et al. 2009:22-26, Evans and Heller 2003:231-244, Faßbinder and Stanjek 1993). Through mapping the topsoil magnetic susceptibility, zones of enhanced values can indicate anthropogenic activity. Short lived settlement sites might not enhance the values enough to create a detectable contrast, while sites settled for a longer period of time should theoretically be easier to detect (David et al. 2008: 4 and 36-37, Gaffney and Gater 2003:44-46, Aspinall et al. 2009:22-29). The Bartington MS2D-field coil is a quick and easy way of investigating the volume specific magnetic susceptibility of approximately the top 10 cm of the topsoil. While the archaeological features are often buried deeper, activity such as ploughing and bioturbation can bring materials with increased magnetic susceptibility up in the topsoil. Therefore the mapping of zones of magnetic susceptibility might be a good way of locating activity areas in the landscape. The inherent magnetic susceptibility of different soils can give an indication on the effectiveness of a magnetometer survey at a later stage (Gaffney and Gater 2003:80, Dearing 1999, Batt et al. 1995). Applicability to Iron Age Settlements At Iron Age Settlement sites in Norway there are several typical activities and archaeological features that should increase the magnetic susceptibility. Fireplaces, cooking pits, metal working, refuse in general and organic waste products both 34 domestic and agricultural, are all examples that is known from archaeological contexts (see chapter two). 3.1.2 Fluxgate Gradiometer General Principles For a positive identification of an anomaly of archaeological origin, it has to be possible to distinguish the anomaly from its background. While magnetic susceptibility is a measurement of how a material can be temporarily magnetised when introducing an external magnetic field, it is important to realise that all materials are constantly subjected to the Earths magnetic field, making it possible for all materials to behave as a magnet. If a feature consists of a material of higher or lower magnetic susceptibility than its surroundings, it can be detected by a magnetometer (Aspinall et al. 2009, Gaffney and Gater 2003: 36-42, Clark 1996: 64-98). Induced magnetisation is when materials subjected to a magnetic field create their own magnetic field as a result. This magnetic field disappear when the induced magnetic field is removed. Remanent magnetization is when materials have properties as a magnet with a local magnetic field, i.e. permanent magnetisation, without being subjected to an external magnetic field. This effect can be the result of heating, where heated material can acquire a magnetisation aligned to the Earth’s Magnetic field at the time of burning. This might not be the same alignment as the present magnetic field today (Gaffney and Gater 2003: 36-42, Aspinall et al. 2009: 21-26, Evans and Heller 2003: 13-20). 35 Variations in the magnetic properties of any materials can therefore be detected by very sensitive equipment called magnetometers which measure the strength of a magnetic field. These changes can be due to features creating a local magnetic field, similar to a bar magnet, by either induced or remanent magnetisation. The Earth’s magnetic field is about 48 000nT, while an archaeological feature can create an anomaly of maybe 10nT or less. By using two magnetometers is it possible to distinguish these tiny changes the local magnetic field create from the background. Mounting two sensors directly above each other with a fixed distance makes it possible to isolate any local changes in the vertical gradient of the magnetic field. This is called a gradiometer. A fluxgate gradiometer usually has a sensitivity down to 0.1 nT, but more sensitive magnetometers exist (Gaffney and Gater 2003: 36-44, Aspinall et al. 2009). Applicability of a gradiometer on Iron Age Settlements The method is ideal to localise industrial areas, fired brick, or features associated with burning like kilns, hearths, fireplaces or cooking pits. Stone-built features in areas of otherwise low background noise can be possible to detect, depending on the magnetic properties of the stone. It is generally not considered good for detecting structures or buildings, such as postholes from a timber building, as the magnetic contrast in the backfill or the remnants of the post is often very difficult to distinguish from its background surroundings unless the background is exceptionally quiet (Gaffney and Gater 2003:37, Gaffney et al. 2002:12-1, Neubauer et al. 2003 Former surveys also show that features such as ring-ditches from burial mounds might be unresponsive, while central cairns or stone constructions from burial mounds as 36 well as boat houses with stone walls have been positively identified (Binns 2002a &b, Lorra 2003, Martens 2003, Gjerpe 2005:20, Fossum 2010, Trinks et al. 2007a, b & c) Sample density - Fluxgate Gradiometer Surveys The English Heritage (EH) has published guidelines for archaeological (David et al. 2008), with suggestions for sample densities for different methods and field strategies. For fluxgate gradiometer surveys the EH guidelines suggest a maximum sampling interval for an area survey of 0.25m along the traverse, with a maximum spacing of 1m apart for evaluation purposes. For characterization purposes a 0.25x0.5m resolution (readings x traverse) is suggested with denser sampling for more detailed characterization. In England the standard of 1m traverse intervals is considered adequate for the majority of surveys (David et al. 2008:4-8, Gaffney and Gater 2003: 95). An increased sample interval along the traverse does not necessarily increase the overall spatial resolution unless the traverse interval is increased accordingly. A survey of 0.5x0.5m can give a higher spatial resolution than a survey of 0.25x1m. As a rule of thumb the survey density should be less than the size or depth of the archaeological feature. EH recommends a ratio of four to one between sample intervals and the traverse interval. A decreased line-spacing will be more time consuming, and increased sampling along the line creates a higher demand for spatial accuracy. The “correct” sampling density should be defined by the possible targets being searched for. The coarsest sampling resolution, most often the line spacing, can be seen as an indication of the size of the features that are possible to identify. If a clearer definition of feature size, shape and depth is needed, an adequate number of 37 data points are required for each anomaly (David et al. 2008: 20-24, Aspinall et al. 2009: 113) 3.1.3 Earth Resistance General principles This method is based upon a fairly simple principle where the ease in which electrical currents can pass through the ground can be measured systematically. The resistance to the flow of currents can be mapped, revealing areas where the current meet high- or low resistance. Where the ground is homogenous, the currents spread evenly. Any obstacles in the ground, either of archaeological origin or other, will lead to changes in the measured electrical effect (Schmidt 2009) The electrical current is carried by moving charged particles which are entirely carried by ions in the soil. These ions are created when salts crystals in the ground dissociate in the presence of soil water. The soils electrical resistivity can be described as the weakening of any current associated with the varying obstacles to the movement of the ions in the ground (Schmidt 2009:67). Resistivity can be considered as a property of the material itself that does not change with the amount or shape of the material. It is therefore more normal to use the term Resistance rather than Resistivity in area surveys, since it is strictly speaking variations due to the shape and form of any features in the subsoil that we are interested in mapping, rather than the resistivity of the material (Gaffney and Gater 2003:28). Variations of the measurements are dependent on the initial abundance of salt, where some soil types have more salts and therefore might be more responsive. More important is the 38 availability of water, which is needed to dissolve the salts into salt ions and facilitate their transportation. The soil resistivity is therefore highly dependent on the moisture content of the ground (Schmidt 2009:67). The possibility of detecting anomalies of archaeological nature is therefore dependent on their ability to retain or drain water. A ditch dug in an otherwise well drained soil will be able to contain more moisture than its surroundings, and possibly be more moist when the surroundings are less well drained. If the soil is either completely dry or totally waterlogged, no contrast can be detected. The success of a survey is therefore connected to the climate in the time before a survey. The most ideal time to survey is therefore when there is some contrast to be found, either in a dry spell after some time with rain, or after some rain in an otherwise dry period. This could be some days, weeks, or maybe even a month, and is closely connected to the ability of the soil to retain moisture (Schmidt 2009:68, Chris Gaffney pers.com.). There are many ways of arranging the probes of an earth resistance array, each with their benefits and drawbacks. Most used within archaeology is the “twin probe” arrangement, where a current and a potential electrode is mounted on a mobile frame with a set probe spacing, and two remote probes are set at a fixed point at least 30 times the probe spacing away from the mobile probes and connected to a data logger mounted on the mobile frame with a long cable. This array is less likely to respond to geological variations, and has a less complicated response than other arrays- often a single-peak rather than a set of multi-peak responses over the same feature (Gaffney and Gater 36-34). 39 Applicability of earth resistance measurements on Iron Age Settlement sites The method is well suited to detect features that cause anomalies of higher- or lower resistance than its surrounding soil depending on water content. Ditches, pits, graves or metal pipes often causes low resistance anomalies while walls, rubble, stone or made-up surfaces often cause high resistance anomalies (Gaffney and Gater 2003:26-27). Ring ditches from burial mounds, or wall-or drainage ditches of houses might be found in this way. At Gustad the method can help reveal if the walls of the houses were made out of stone or constructed in some other way. Pits or other features of interest, in addition to the possible graves known to be at the site, can be detected. Sample density - Earth Resistance surveys The EH Guidelines suggests a resolution of 1x1m for evaluation and 0.5x1m or 0.5x0.5m for characterisation purposes. They recommend Twin-Probe arrays as the preferred method of ground coverage (David et al. 2008:4,8,24-28). The wider the spacing of the probes, the deeper the current penetrates into the ground. The reading logged can then be seen as “some form of average” of the soil sampled. Due to the way the current flows, it can be said that this array is most sensitive close to the current-potential probes, and therefore samples an area around the probes. The size of the separation and the depth penetration therefore sets the limits of how high the sample density should be. Several suggestions for depth of investigation are suggested; either 1/3, ½ or 1 of the probe spacing, depending on array type. The literature suggests that the twin-probe array of 0,5m spacing can be affected by features as deep as 0.75-1m, depending on the features conductivity. 40 Sampling at 0.5x0.5m is generally considered to be a good balance between depth of investigation and spatial resolution (Schmidt 2009:3-4, Gaffney and Gater 2003: 32, 56- 60, David et al. 2008:25-28). 3.1.4 Ground Penetrating Radar General Principles This method can be used to get a three-dimensional view of the subsoil by transmitting electromagnetic waves into the ground and measuring the time it takes for the energy to be reflected back to a receiver. When the signal meet discontinuities or surfaces of contrasting properties, some of the signal is reflected back while others continue further down into the ground where they might be reflected back from other features deeper down (Gaffney and Gater 2003:47-51, Conyers 2004) The time the signal takes to travel through the ground and be reflected back can be used to estimate the depth of the buried feature, and is dependent on the dielectric constant of the material that the energy travels through. The different physical and chemical properties of the subsoil will alter the velocity of the passing waves, and both the electric conductivity and the magnetic permeability affect the propagation and reflection of these waves. As with other electric methods the water saturation will affect the way the energy is absorbed, as well as the possible contrast between interfaces or features of interest, and the surrounding soil. The dielectric constant, or the relative dielectric permittivity (RDP), can be defined as “the ability of a material to store a charge from an applied electromagnetic field and then transmit that energy” (Conyers 2004:45). Generally it can be said that the greater the RDP the 41 slower the radar energy will move through it, and the RDP can be seen as a general measurement of how well the energy will be transmitted to depth. This is therefore a measure of both the velocity of the propagating radar energy and its strength (Conyers 2004:45). Strong, or high amplitude reflections will be created where the change of velocity of the transmitted energy at boundaries are large. The greater the change in velocity of the transmitted pulse, the more likely it is that a feature can be detected (Gaffney and Gater 2003:50). The penetration is dependent on the frequency that is transmitted, where a higher frequency will give a better resolution, but less penetration because of the shorter wavelength. A low frequency with a longer wavelength will therefore penetrate deeper, but not have the same spatial resolution. Frequencies of 80 MHz – 1000 MHz (1GHz) are used for archaeological purposes, and the choice of the correct central frequency antenna is therefore important and depends on the properties of the host environment, the assumed depth of the archaeology, size and dimension of the archaeological features involved. Correct depth estimate is difficult since the velocity changes as the energy passes through materials with different RDP, but several methods exist to try and estimate this (Gaffney and Gater 2003. 48-49). A profile made of stacked reflection traces can be seen as a section through the subsoil. By surveying closely spaces transects, it is possible to arrange these parallel in a software to create a block of data that can be sliced to so called “time slices” that represent a plan view of a specific time depth down in the subsoil. By summing up the data between a selected time or depth range for every traverse, a plan view of anomalies at a particular depth can be presented. These can be viewed as a map, and 42 show how GPR can give information on both the spatial distribution of anomalies as well as the vertical properties as sections (Gaffney et al 2002.10). Ground Penetrating Radar is generally considered more time consuming than most other prospection methods, even though new multi-channel systems combined with accurate GPS measurements can cover huge amounts of ground in a day, possibly up to 4-5 hectares a day (Egil Eide. Pers. Com.). Even though the technology is venturing towards faster data capture, the post processing involved with GPR-data must be considered more time consuming and technically more complicated than for instance Earth Resistance or Gradiometer survey data (Kvamme 2003: 442-443) Applicability on Iron Age Settlement sites The method is considered good for detecting refilled pits and ditches, voids, buried roadways and paths, walls, floors etc. (Gaffney et al. 2002: 14), and has proven good potential for finding post holes of Iron Age Houses in Scandinavia (Trinks et al. 2007a). While magnetometers can indicate if a feature has a higher or lower magnetic response, it does not tell you if derives from a pit or a solid feature- or accurately at what depth the anomaly is at. Using GPR sections can aid in this interpretation, and give more accurate depth estimates. Volume calculations can be performed as a result of this- further aiding any planning of future excavations. Sample Density - Ground Penetrating Radar survey The EH Guidelines recommends a maximum traverse spacing of 0.5m and samples taken every 0.05m along the line. This method should be applied for the 43 detailed investigation of a site, gathering information about depth as individual sections and as time slices. The recommendation for the reading intervals is dependent on the centre frequency of the antenna used, where a wider spacing between traverses can be suggested for lower frequency antennas as they don’t resolve information with as high resolution as higher frequency antennas. There is a risk of spatial aliasing and the loss of necessary data to get adequate information of the features in the subsoil, if the traverse spacing is larger than the approximate footprint of the radar energy at the required depth of investigation. For a 500MHz centre frequency antenna under typical conditions any traverse spacing above 0.25m will be spatially aliased. Therefore a consideration of the gain of information compared to the time it takes to perform a survey with this traverse spacing needs to be considered. A higher sampling rate between traverses will increase the definition of features in the subsurface, and 0.5m between the traverses should be considered as a suggested spacing where spatial aliasing is not causing problems for the interpretation of the target features (David et al. 2008:4,8,28-34). The survey at Gustad was done with an antenna with a centre frequency of 400MHz, which has a wavelength (λ) of about 0.75m in air and will rarely transmit energy to greater depths than 2m. As a rule of thumb the smallest feature that can be resolved in 0.25λ, while a more pragmatic view can suggest a resolution of 0.5λ. As radar energy passes through the ground the wavelength decreases. For an antenna with a frequency of 400MHz this can indicate a spatial resolution of 4.85-9,75cm. with a typical RDP of 15. The possible resolution might change depending on the RDP in the ground. For tables presenting different spatial resolutions and typical RDP values for different type of soils, see table10 and 11 in the appendix. It is also important to 44 remember that the penetration depth decreases with a higher RDP, and the energy is not transmitted directly down vertically, but is transmitted in a “cone”-like shape with an increasing diameter of the footprint with depth, reducing the spatial resolution with depth and complicating the response (Conyers 2004:62-68, David et al. 2008:28-34). 3.2 Survey Strategies There are several points that need consideration when planning a survey of a site. This involves size of area to sample, which methods to be used- and in what sequence as well as resolution. Regarding resolution it is mostly a question of sample density, which is closely connected with the time it takes to do the actual survey and the available time/budget. The choice of method is related to the archaeology expected and a consideration of how the geology at the site might affect the results. Many of these choices will be site dependent. In this section some suggested strategies from past publications will be presented, and the EH Guidelines (David et al. 2008) is central in this respect. 3.2.1 Sampled area While a complete survey of an area should be preferred, there is always a trade-off between the time it takes to undertake a survey and the related costs. The size of areas covered by the chosen methods is often governed by the speed in which a survey can be conducted. A problem may arise when an area is too large to be fully covered with more than one technique. Some form of sampling is usually undertaken of areas larger than 2ha, depending on how archaeologically sensitive the areas are. This could be systematic, 45 random or a strategy modified depending on supporting sources like for instance past archaeological finds or aerial photos. A rapid assessment technique like scanning or magnetic susceptibility sampling might be of some help deciding where to do a more detailed survey. For large areas the guidelines state that it is desired that at least 50% should be sampled with an area survey, but it is usually expected that 10-50% of the total area that has been assessed at with more rapid methods (Gaffney et al. 2002:3, David et al. 2008:18). With the recent development in rapid data collection, it might be suggested that areas larger than 2ha should be completely covered when conducting a magnetometer survey (Chris Gaffney, Pers. Com). 3.2.2 Magnetometer scanning vs. susceptibility Scanning with a gradiometer is when an operator walks over an area with the instrument in “scan”-mode. The operator monitors the continuous readings when walking over the area and putting down some form of marker when the gradiometer is moved over an anomaly of a set value- normally 2 or 3nT. This can be performed widely spaced traverses, for instance 10-15m apart, and the anomalies marked can then give an impression on where a more detailed area survey might be conducted. This technique requires some experience, as the operator have to gain an impression of the general background value while scanning and assessing the area (Gaffney et al. 2002:3). For Magnetic Susceptibility a map of topsoil susceptibility can help delimiting a settlement or an industrial site. The topsoil susceptibility is usually measured within a grid or with a GPS with a spacing of 5-20m. As with scanning, blank control areas should be sampled. The disadvantages are that modern land use and different types of 46 geology can create spurious variations, as well as the fact that some archaeological features might be too ephemeral to create an enhanced topsoil magnetic susceptibility while still being potentially detectable with a gradiometer. The latter can be exemplified by a survey done at Charlton Villa, Wiltshire where the settlement area showed up clearly but where ring ditches from burial mounds seemed to be situated in areas without enhanced magnetic susceptibility topsoil values (Corney et al. 2004, Gaffney et al. 2002: 3, Gaffney and Gater 2003: 91-96, Batt et al. 1995). At Gustad a larger area was surveyed in high resolution with a fluxgate gradiometer, as well as magnetic susceptibility. This data is used to visualise how well different resolutions of magnetic susceptibility would yield useful results. A modelled scanning based on the gradiometer results is used to show how scanning would have indicated any of the archaeology known to be present. 3.2.3 Sampling vs. Area coverage There is always a benefit of having data from a continuous area. An area of less than 40x40m is usually considered uninformative. It is also possible to do sampling strips which should not be narrower than 20m (Gaffney et al. 2002:3, Gaffney and Gater 2003:88-101). The drawback with scanning is that no points are recorded while investigating an area, so a form of map might not be produced of these “hotspots” if they are not tied to a recorded grid or measured in with a total station or a GPS. Igneous geology and natural soil variation might cause problems, and some anomalies might not cause a response strong enough to be indicated by scanning, while still creating a detectable response on an area survey map. Apparently blank areas should therefore be included 47 in the next phase of surveying at the site, making sure that some information is documented for these areas as well; scanning should not be performed without being followed up by a gradiometer area survey (Gaffney and Gater 2003:93). Some doubts exist among practitioners on whether scanning can be trusted or not, and it has been suggested that the practice of scanning should end (Jordan 2007 and 2009). 3.2.4 Choice of method and sequence The choice of preferred method should be a trade off between the gathered data, and the ability to reveal the necessary information. The relative speed of data collection and post-processing are different from method to method. Possible archaeological features seen as geophysical anomalies by one method might not be revealed in the geophysical data from another. Gaining a good impression on the background conditions is also important, as it is easier to separate anomalies of archaeological origin from the natural ones and increases the level of confidence in the interpretations (Gaffney and Gater 2003:88-91). The EH Guidelines suggests scanning or magnetic susceptibility sampling for initial site evaluation, followed by magnetometer area survey, resistance survey and GPR in that order. This recommendation is based on the speed of acquisition and the information the methods might reveal. As a magnetometer survey offers rapid ground coverage and responds to a wide range of anomalies, this is recommended as the first technique for detailed area survey followed by slower techniques. Earth resistance surveys often complement the results from a magnetometer survey. It is considered favourable when building foundations and masonry features are expected. Areas revealed by a magnetometer survey can be targeted by selected areas of earth 48 resistance survey, especially when building remains is expected. GPR can complement the information from magnetometer and earth resistance survey, with both a higher resolution and the ability to give more accurate depth information. The sequence of preferred method should be the result of a balanced consideration for each situation, as the ground conditions, geology or potential archaeology varies from site to site (David et al. 2008). Gaffney et al. (2003) suggested a three level system for geophysical survey, where level one was for prospecting involving scanning or coarse topsoil magnetic susceptibility for locating sites or areas of interest, level two a more detailed assessment of known or suspected remains using detailed topsoil magnetic susceptibility, magnetometer survey, earth resistance and/or GPR, and with a third level for more detailed investigation. This last level was considered less common in evaluation work. The publication does not indicate in what order the specific technique should be used, but rather the level of intensity and resolution they should be utilised at the different stages of the survey. In England the use of geophysical methods is deeply integrated in the archaeological practice, but the choices made when conducting a geophysical survey might not be appropriate for the specific site and its archaeological and natural context. The choice of methods and survey density not only dictates whether or not remains might be detected, but also what kind of remains. Maybe should survey briefs rather indicate what kind of features that are expected, rather than a set “method” or resolution. A geophysical survey can then be adjusted accordingly by the contractor to be more appropriate to the specific site (Jordan 2007 and 2009). As the utilisation of geophysical techniques is not considered a standard procedure in Norwegian archaeology, an evaluation of the possibilities of the different 49 methods and the value of high resolution surveys when prospecting for Iron Age settlement sites and the response of the methods on the local type of geology is important. The survey at Gustad can be used as an example that can increase our knowledge on the applicability of geophysical methods, and point us in the right direction for future practice concerning choice of method, resolution and sequence. 50 4 FIELD SURVEY RESULTS The geophysical field survey was conducted on the 5 th to the 13th of July for the Magnetic Susceptibility, fluxgate gradiometer and GPR. The Earth Resistance survey was conducted from the 7th to the 10th if August. The magnetic susceptibility (MS) and fluxgate gradiometer (FG) data were collected in sunny and warm conditions, and the GPR were collected after a few days of rain in slightly wet conditions- which might not have been ideal. The earth resistance (ER) measurements were collected over a few days of sunny and cloudy weather, with some rain a couple of days before conducting this part of the survey. The survey was undertaken by setting up 20 x 20m grids along an East-West oriented baseline with tapes, and collecting data walking South-North along ropes with markers for every 0.25 cm, ensuring the most typical magnetic anomaly forms would be mapped. Points at 0,5m and 1m intervals were marked with a separate colour, making it easy to walk in a constant pace with the magnetometer and therefore ensuring good quality data location. All grid corners were located by a survey engineer from the company Vitec AS with an RTK GPS with an accuracy of ±2cm. A map of the grid corners with a table of coordinates is presented in the appendix. These measurements were also used for the ER Survey by accurately resetting of the grid with a Leica 500 base station at a known point and height, and a Leica Viva Rover. The topsoil mapping of the magnetic susceptibility was undertaken with a Bartington MS2D magnetic susceptibility meter and the points logged with a DGPS connected to ArcPad on a PDA via a Bluetooth connection. This should give a spatial accuracy of ±1m. The MS survey was undertaken 51 with an approximate spacing of 5m when walking along a continually updated map of the position of the surveyor on ArcPad. The following parameters were used for the different instruments: Earth Resistance Survey Details RM15-D 0,5m RM15-D 1m Electrode array Twin Probe Twin Probe Electrode separation 0.5 m 1m Grid Size 20 x 20 m 20 x 20 m Number of grids 17 17 Orientation N-S N-S Data Collection Zig-Zag Zig-Zag Traverse interval 0.5 m 1m Sampling along traverse 0.5 m 0.5 m Magnetometer Survey Bartington Ground Penetrating Radar GSSI Sir-3000 Details Grad 601 Dual Survey Details Grid Size 20 x 20 m Area surveyed 40 x 60m Number of grids 36 Number of grids 6 Orientation N-S Orientation W-E Data Collection Parallel Data collection Parallel Traverse interval 0.5 Traverse interval 0.25 Sampling along traverse 0.125 Sampling along traverse 0.02 Sensitivity 0.03 nT Centre frequency of 400 Mhz antenna Table 6: Table showing the survey parameters used at Gustad. Illustration 13: Map showing the areas surveyed with the different geophysical methods. 52 The following table indicates the approximate area surveyed with the different methods: Method Area in hectares/decares Approximate Field (“Mål”)7 Time (in days) Magnetic Susceptibility (MS) 4.75/47.5 2.5 Fluxgate Gradiometer (FG) 1.32/13.2 4 Earth Resistance (ER) 0.67/6.7 4 Ground Penetrating Radar (GPR) 0.24/2.4 2 Table 7: Area Calculations and approximate field time See the appendix, part 8.8 for categories used in the interpretation of the geophysical field data. 4.1 Magnetic Susceptibility Survey The survey was conducted with a Bartington MS2D magnetic susceptibility meter with a field coil, measuring the volume magnetic susceptibility (κ) of the topsoil. The values recorded are in dimensionless SI units. The values recorded was logged as points, and entered into the GIS software ArcGis 9.2 with 3d-analyst. By running a nearest neighbor routine on the stored points, a continuous raster was created. This was clipped with the actual extent of the survey, and visualized with a standard deviation of n=3. 7 A decare is 1/10 of a hectare, equal to 1000m². The Norwegian word “Mål” is often used to denote a decare. 53 Illustration 14: The topsoil magnetic susceptibility results. The resulting map shows a clearly delimited area with high readings of approximately 0.35 ha just north of the road. This area will later be compared with an interpretation of the results from the other methods. An increased linear signal continuing in the same direction as the main road going between the two fields surveyed, roughly from WSW towards ENE has the same orientation as a former road, which was removed in the 1970s. An area of high readings in the upper NW corner of the survey is in the same area as a now demolished summer barn. Some increased values can be seen in the NE part of the area. 54 4.2 Fluxgate Gradiometer Survey The survey was conducted with a Bartington Grad 601-2 dual gradiometer system, with a 1m sensor separation. The measurements were collected with a 0.5m traverse interval and a sample interval of 0.125 along the traverse. The same balancing point was used for the whole survey, and the data was collected by two different surveyors of the approximate same height. Illustration 15: Raw, unprocessed image of the Fluxgate Gradiometer results. The raw unprocessed data in illustration 15 reveals a low magnetic background with some broad geological trends, and a number of anomalies clearly standing out from the background and are easily interpreted from the raw image alone. Some striping due to minor differences in the balancing the two gradiometers used can be seen. The 55 broad geological trends made it difficult to get at decent data image only using the Zero Mean Traverse or Zero Mean Grid processing options in the Geoplot software (for example see appendix part 8.3, illustration 53). This is due to the geology increasing mean values along a traverse or within a grid. To process out the striping in the data, a high pass filter was applied to enhance any anomalies not caused by these geological trends. Illustration 16: Image after High Pass Filtering and Zero Mean Traversing (ZMT). 56 Illustration 17: Image after High Pass Filtering and Zero Mean Traversing (ZMT) and interpolating times 1 in the X direction. Some minor striping along grid edges are still present, and the high pass filter is also known to introduce some negative “halos” around certain anomalies (Aspinall et al. 2009: 127-128). Most of the effect of geological trends disappeared, helping to highlight smaller scale anomalies. Some horizontal trends probably caused by modern ploughing became more visible. To help classify the data all anomalies above 3 nT or monopole anomalies with a negative value of -2 nT or less were isolated causing the following anomalies can be identified. These values were chosen as these anomalie strengths clearly stand out of the general background values, and therefore are clearly anomalous. 57 Illustration 18: Isolated anomalies of either above 3 nt or below -2 nT. Further categorising can be performed by visualising the anomalies by their highest absolute values, as well as their mean value: Illustration 19: Anomalies classified using their highest and lowest value in nT. 58 Illustration 20: Anomalies classified using their mean values in nT. Illustration 21: Initial geophysical interpretation based on the fluxgate gradiometer results alone. 59 Many typical archaeological features at a settlement site, apart from ditches related to buildings and fences, are often pit-features. As a result many of the anomalies present in this data set might be caused by archaeology, but elements such as geological elements and the presence of magnetic stones, or modern disturbances might introduce anomalies of similar size, and therefore be undistinguishable from possible archaeology. Apart from the obviously elongated linear anomalies in the lower middle part of the image (“A” on illustration 21), and a number of strong positive circular anomalies with a negative halo possibly interpretable as cooking pits (“B” in illustration 21, see also illustration 34 later in this chapter), few obvious anomalies can be easily interpreted as archaeology solely based on the magnetic data alone. Systematic distribution of anomalies in lines or circles, or other shapes possibly caused by archaeological features is also hard to identify. It is interesting to note that the outline of the burial mounds that is identified on the aerial photos could not be seen in these images (see illustration 9-11), as well as only house 1 could be clearly identified but not house 2 (see illustration 12). This shows the difficulties in interpreting the magnetic data without additional information from other methods. 4.3 Earth Resistance Survey The survey was conducted with a Geoscan Research RM15-D earth resistance meter with the PA20 multiprobe array and the MPX15 multiplexer. Using this set up three readings could be collected with every movement of the array- a reading of a 1m probe spacing as well as two readings of 0.5m probe spacing adjacent to each other. 60 Illustration 22: Raw image of the Earth Resistance results for the 0.5x0.5m with a 0.5m probe spacing. This image has been edge matched and despiked. Raw, unprocessed image is presented in the appendix. Illustration 23: Filtered image of the Earth Resistance results for the 0.5x0.5m with a 0.5m probe spacing. This image has been edge matched and despiked, as well as high- and low pass filtered. 61 The following interpretation can be made: Illustration 24: Geophysical interpretation of the 0.5m probe spacing Earth Resistance data. Illustration 25: Raw image of the Earth Resistance results for the 1x0.5m with a 1m probe spacing. This image has been edge matched and despiked. Raw, unprocessed image is presented in the appendix. 62 Illustration 26: Illustration 4: Filtered image of the Earth Resistance results for the 1x0.5m with a 1m probe spacing. This image has been edge matched and despiked, as well as high- and low pass filtered. Illustration 27: Geophysical interpretation of the 1m probe spacing Earth Resistance data. 63 Most obvious are the features interpreted as archaeology- the rectangular low resistance anomalies interpreted as archaeology in the southeast (“A” on illustration 24 and 27), the circular anomaly to the northeast and the two curved low resistance anomalies central in the map (“B” on illustration 24 and 27). These must be anomalies representing houses and burial mounds respectively. Other low resistance anomalies are also present, and smaller anomalies might be pits (“C” on illustration 24 and 27), and will be compared with the magnetometer results later on to see if they display a magnetic signal indicating a backfill of increased magnetic response. A general trend going from west southwest towards east-northeast can be seen, but has not been drawn in on this map. It is likely that these are remnants of ploughing, as it is similar to the ploughing direction observed at the site and with the GPR. Some areas of higher resistance exist, that clearly are above the mean values (“D” on illustration 24 and 27). These can be hard surfaces or areas that by some reason have been more undisturbed by ploughing than others. Some random rocks might be in the subsoil, but larger boulders are expected to be rare. The general level of response can be said to be good, and the method seem to be well suited at this site. 64 4.4 Ground Penetrating Radar The GPR used at this survey was a GSSI Sir 3000 with a 400 Mhz centre frequency antenna mounted on a cart. The data was collected at 25cm traverses walking West-East, with a sample for every 2 centimeter along the traverse. The distance was measured with an odometer wheel, which was calibrated doing a sample run along the full length of a traverse (40m.). The RDP value was set at 15 for approximate depth correction, as the subsoil was assumed to be moist marine sandy silt (see Chapter three and table 11 in the appendix). Any calculated depths should be viewed as approximations, and not accurate depth indication. All data presented here are time slices made with the software GPR-slice. The data has been edited, dewoved, gained and resampled. Some mosaic correction has been performed to balance the data, as the data collected on the two days of survey reveal some differences from day to day. This is probably due to moisture content in the ground caused by rainfall, as the pre-gain set up in the field was kept the same on both days. Two georeferenced time slices are produced to illustrate the interpretation of the GPR. All interpretations were done by analysing 16 slices. These are all presented separately in the appendix. 65 Illustration 28: Time slice 7 illustrated with two different palettes. Illustration 29: Time slice 18 presented with two different palettes. 66 Illustration 30 (left): Geophysical interpretation of all time slices compiled. Any anomaly interpreted here appeared on two or more time slices, indicating a possible feature of some depth extent and appropriate RDP contrast to its surroundings. The results reveal the outline of the two houses (“A”) with some possible internal anomalies (“B”), a circular anomaly (“C”) with a central depression (“D”), a couple of anomalies with a more random distribution (E”) in addition to some straight, linear anomalies. The anomaly in the top right corner seems to be semicircular (“F”), but it is hard to say for certain without the full extent being surveyed. A low resistance anomaly was discovered at approximately the same position. 4.5 Data comparison Combining and visualising the results can illustrate how increased knowledge derived from several methods can help with the interpretation of anomalies, where anomalies identified with one method can alter the knowledge about the archaeological context and therefore the interpretation of specific anomalies. 67 Illustration 31: Comparison between topsoil magnetic susceptibility and the fluxgate gradiometer results. Note the good correlation between the house and an area of increased topsoil magnetic susceptibility (MS) as well as areas where stronger gradiometer anomalies coincide with the MS. Illustration 32: Visualising of the number of singular anomalies stronger than 3 nT within each survey grid. 68 Illustration 33: Simplified results from all methods. It gives a rather complicated impression of the anomalies identified, but illustrates their spatial relationship and location compared with each other. The topsoil MS only samples the top 10 cm, while the magnetometer used has a vertical sensor separation of 1m and therefore is influenced by anomalies deeper down. Illustrations 19 show how the MS topsoil mapping coincides with the identified settlement remains seen in the FG data, and how it can be used as an indication of the distribution of magnetic anomalies buried deeper down, as illustration 20 shows. This can therefore be interpreted as an indication of the presence of features with an increased induced magnetism, possibly settlement activity. Several of these magnetic anomalies in this area can be interpreted as archaeological based on this reasoning. Illustration 21 shows how different methods show different aspects of the site, and that only a few anomalies are present in all methods. It is interesting to note that the positions of the burial mounds are in areas of relatively low topsoil MS, while the houses are identified within an larger isolated are of increased topsoil MS values. 69 The problem with identifying subtle anomalies in the FG data as possibly archaeological can be investigated further by these illustrations: Illustration 34: Comparison of the Earth Resistance Anomalies with the identified maximum values of the Fluxgate Gradiometer anomalies. Illustration 35: Comparison of GPR results with the Earth Resistance and Fluxgate Gradiometer anomalies identified. 70 These comparisons show how subtle anomalies that are not easily interpreted when viewed on their own, taking especially anomalies present inside the house to the east (house 2). This house is not easily identified in the FG data (see illustrations 15-17 in this chapter). Quite a few of the anomalies present within this house would be very hard to interpret as potentially archaeological without the knowledge of there actually being a house there, and even then it is not certain that they are associated with any constructional features normally associated with a building. Only 6 out of 18 FG anomalies coincides with anomalies identified with the GPR, and while their peak values are above 3 nT, the mean value might only be between 0-3 nT, rendering them as quite subtle. The GPR and the ER both work well in identifying what must be ditches that we can identify as houses. Both methods also revealed the circular anomaly with a central anomaly visible on the aerial photos (see chapter two), that must be a ploughed over burial mound. It is interesting to note that the central anomaly in the burial mound interpreted as the actual burial, has a negative monopole magnetic response with a maximum value of -2.4 nT while still having a low ER. This means that the burial must have a fill of lower induced magnetism, and does not likely contain any resistant features like stone or hard rubble. 4.6 Archaeological interpretation of the geophysical data The following interpretation can be based on the geophysical results of all surveyed data. In the archaeological interpretation only the anomalies with a high degree of certainty associated with them will be included, either based on anomalies 71 being present in several methods, the nature of their geophysical response, their context and archaeological knowledge based on similar sites. It is also assumed that a lot of the anomalies not included in this interpretation might still be of archaeological nature and interest. Illustration 36: Archaeological interpretation of the geophysical survey data. It is now possible to get the exact location of the archaeology present, as well as some measurements of sizes and extent. We have two houses present which are parallel to each other, and of similar size: Internal External measurements measurements (maximum) (maximum) House 1 (west) - Aerial Photo 23.2x6.8m 25.3x10.2m House 1 (west) – 24.2x6.8m 26.3x10.6m Geophysical Survey Data House 2 (east) – Aerial Photo 24x6.1m 27x14m House 2 (east) – 25.1x6.2m 29.3x13.5m Geophysical Survey Data 72 House 1 did show up in all three detailed methods utilised. It was harder to identify on the GPR data, while still very visible in both the ER and FG results. House 2 was identified in both the ER and GPR data, and being aware of its presence made it possible to associate some anomalies with the presence of a house in the FG data as well- even though its shape and form did not show up in any way clear in the FG data. The reasons for these two houses to be so distinctly different in their magnetic response can be one or more out of several possibilities: - They were built differently. o Different building tradition might suggest a different dating. The fact that they are oriented in such a similar fashion and seem to respect the placement of each other suggest a similar dating. o Some boat houses in the area have known to have stone lined ditches (Johansen 2007, Geir Grønnesby, Pers. Com). Their low resistance suggest that the anomalies present are ditches, and not filled with stones. - They were demolished or dismantled in a different fashion. Since house 1 has such a clear magnetic signal, it might be a possibility that this house burnt down. - They had different functions. o Their orientation towards the sea and the unusually high external width compared with known settlement houses suggest a function of boat houses, even though house 2 does have an end-wall towards north- west. House 2 might have postholes centred in the middle of the building the towards south-east while possibly having post holes moved 73 closer to the walls of the building in the other end. This might indicate a building technique that can make room for a ship or a boat of up to 15- 16m in length and 4m width assuming this interpretation is correct. o The different internal magnetic response of the houses could on the other hand suggest different usage. 4.6.1 The burial mounds Two burial mounds were identified by this survey. This is quite few compared to the 18 burial monuments mentioned in the sources (see chapter two). This could be because the survey did not cover the entire area of the prehistoric burial ground, or that the monuments have been built differently in the first place. Experience mapping similar monuments would suggest that they might just as well have been built without a visible ditch. The actual burial might not be dug into the subsoil, but also be placed further up in the mound material leaving no visible trace detectable in the subsoil. The following measurements of the two positively identified mounds: Internal measurements External measurements (inside ditch) (with ditch) Round mound 13m in diameter 18.5m in diameter Boat shaped long 32x6m 32x12.4m mound Table 8: Measured length and diameter of identified burial mounds. While the size of the round mound could be considered quite ordinary, can the length of the long mound be considered quite large compared to known comparable monuments in the region. The burial pit within the round mound seem to measure roughly 2.8x2m, and is oriented NNW-SSE. 74 A smaller area of some minor magnetic noise is interpreted as an area of scattered stones, and could indicate some prehistoric activity or feature present. This could also be due to agricultural practice, especially if there is minor topographic rise present making the plough go deeper. This area coincides with some strong positive anomalies with a negative halo, as well as an area of increased topsoil magnetic susceptibility- and might therefore indicate an isolated area of cooking pits or burning. Stones from the cooking pits that have acquired a remanent magnetism by burning might have been scattered by the plough, creating this pattern. 75 5 ANALYSIS This chapter will focus on the applicability of each method discussed in chapter three for prospecting on Iron Age Settlement sites, based on the results presented in chapter four. The resolution necessary to delimit the site and identify its archaeological components will be investigated by performed by downscaling the original results. The analysis of the results will lead to discussion about an appropriate survey strategy for similar sites, involving choice of methods, appropriate sequence and resolution. 5.1 Applicability to Iron Age Settlement Sites 5.1.1 Magnetic Susceptibility Survey It can be seen in the previous chapter, as well as in illustrations 37-39, that this method proved to be very well suited to indicate an area of increased values at Gustad, coinciding very well with anomalies interpreted as houses and cooking pits. This is very encouraging results for future use for locating and delimiting settlement sites, and shows how well topsoil magnetic susceptibility mapping as an initial site evaluation might work. The location of identified burial mounds at this site was not indicated by the presence of increased topsoil magnetic susceptibility values. This is a similar situation as an investigation at Charlton Villa, Wiltshire in England (Corney et al. 1994). This is not considered surprising, as the building and demolition of a burial mound is not necessarily associated with burning unless the mound is been placed on top of the actual funeral pyre. Some degree of caution must be maintained, since short-lived past activity, varying geology, land use practice and modern disturbance might disrupt any pattern 76 from past anthropogenic activity, and render it difficult to get conclusive results. The response on marine beach sediments of sandy silt at Gustad revealed good results. Resolution of Topsoil Magnetic Susceptibility mapping Illustration 37: Topsoil Magnetic Susceptibility mapping with a sample point at every 5x5m. The central area with high and low values can be said to be well defined, and when studying the area in detail some local variation within this area can also be noticed. This might indicate some internal differences within the past activity area, and might give some spatial information on the activity within the site. 77 Illustration 38: Topsoil Magnetic Susceptibility mapping with a sample point at every 10x10m. The areas of higher and lower values are still well defined. Some local variation is lost, but the method still successfully delimits the main activity area. Illustration 39: Topsoil Magnetic Susceptibility mapping with a sample point at every 20x20m. 78 With samples taken every 20x20m only the major variations seem to be picked up, but the main activity area is still identified. Any outlying areas of potential interest may be lost, such as the location of a past summer barn in the upper north-west corner of the surveyed area. Recommendation A resolution above 10x10m is not recommended, as it is easy to miss important activity areas under a certain size. A choice should be made depending on the size of the survey area and the time available, but as the method is considered relatively quick- covering approximately 2 ha a day with the setup used at Gustad, a resolution of 5x5m might be preferable as it increases the confidence level of the survey. 5.1.2 Fluxgate Gradiometer Survey At Gustad the background values was relatively quiet, making it possible to distinguish a large number of anomalies with a peak value above 3nT (535 anomalies), and 80 monopoles of negative nT values below -2nT. Apart from the walls of house 1 and some stronger anomalies, few are easily identifiable without any shape or patterning that can be clearly associated with known archaeological features. Only when other methods positively identified anomalies could possible patterns be associated with archaeology and perceived as less ambiguous. As archaeological features associated with settlement sites often are pit-features like post holes and cooking pits of relatively small size (postholes often 0.2-1m in diameter and cooking pits up to 2-2.5m in diameter) and relatively circular, they can be quite elusive in the 79 gradiometer data. When that it said, the distribution of positive anomalies at Gustad shows generally the same pattern as the topsoil magnetic susceptibility mapping (see illustration X in the former chapter), indicating some clustering of magnetic anomalies in the area interpreted as the central settlement area with houses and cooking pits as identified by analysing all methods (chapter four). The magnetic response is also useful in comparison with other data, where for instance GPR can help distinguish whether or not the magnetic anomaly is from a pit or a solid feature, and by that be classified with a higher degree of certainty. The gradiometer data did not successfully locate the burial mounds, but some isolated features could be associated with the known mounds when located with ER and GPR. The central burial in the round mound was for instance identified as a negative monopole in the gradiometer data. The lack of response from the mounds could possibly be associated with funeral practices and the activity at the site when the features removed was backfilled with some material. At Gustad this material does not have a magnetic susceptibility contrast sufficient to be detected with the fluxgate gradiometers. This could indicate that any ditches associated with the mounds were not backfilled at the time of settlement. An area absent of anomalies might indicate that the area was not accessible at the time of settlement. For instance would no cooking pits or features be dug at the spot of occupied by a burial mound. This assumes that there were no activity at the site prior to the erection of the mound. Resolution of Fluxgate Gradiometer Surveys on Iron Age Settlement Sites It is here possible to distinguish between the delimiting of major anomalies at the site and the identification of small anomalies that can be important for the 80 identification of important archaeological components within the site. This can be seen as similar to the EH guidelines distinction between sampling for evaluation purposes and characterisation purposes (see chapter three). A series of images of different gradiometer resolution, as well as a more detailed view of a chosen area, will help analyse the effect of using different survey resolution at Gustad. Evaluation Illustration 40: Processed image with a sampling resolution of 0.5m between traverses and 0.125m along the traverse Houses and strong anomalies are easily visible. A large number of anomalies of smaller size can be seen, but a recognisable patterning is hard to distinguish. A certain ambiguity exists. 81 Illustration 41: Processed image with a sampling resolution of 0.5m between traverses and 0.25m along the traverses When projecting the survey limit at this spatial level, a clearly distinguishable change of effect due to reduced resolution along the traverse is not easily identified. Illustration 42: Processed image from a sampling resolution of 1m between traverses and 0.25m along the traverses 82 Reducing the sample resolution to 1m between the traverses still identifies the most prominent anomalies, such as house 1 and the strong anomalies with a negative halo interpreted as cooking pit in chapter four. Detailed definition of anomaly shapes disappear, and the ambiguity and uncertainty in the interpretation increases as the resolution decreases. This can be investigated further by looking at a chosen area in detail: Characterisation The raw image shown on illustration 15 is very detailed, and has been used when interpreting the data, as the high pass filtering that was needed to remove broad geological changes introduced some negative halos around some of the strong positive anomalies. Illustration 43c and 43d show how subtle anomalies can be seen that are clearly anomalous compared with the background values, but still difficult to interpret. The shape of the anomalies can be well defined at this resolution. Important properties like the orientation of negative values related to positive readings is more likely to be detected. Anomalies down to a size of 0.3x0.5m with a strength above 3nT could be identified. Illustration 43h show that the same anomalies can be said to still be easily identified, but edges that in illustration 43c seem to be quite smooth appear sharp. The extent and shape of anomalies are roughly the same, and subtle features inside house 2 can still be seen. The spatial accuracy of the survey will be decreased. Illustration 43g show how a 1m spacing between the traverses will only give a good definition of the more prominent anomalies and the broad trends, but the usefulness of this resolution to characterise subtle features is low. The spatial location of anomalies is further decreased, and important magnetic properties helping to defining the anomalies are lost. 83 Illustration 43a: Detailed view of interpretation of the FG data. Illustration 43b: Raw image. 0.5m traverse and 0.125m sample interval. Illustration 43c: Filtered data. 0.5m traverse and 0.125m sample Illustration 43d: Filtered and Interpolated data. interpolated in x interval. direction to 0.25m traverse interval. 0.125m sample interval. 84 Illustration 43e: Detailed view of interpretation of the FG data. Illustration 43f: Raw image. 0.5m traverse and 0.125m sample interval. Illustration 43g: Filtered data. 0.5m traverse and 0.25 sample interval. Illustration 43h: Filtered data. 1m traverse and 0.25 sample interval. 85 Recommendation While a resolution 1m by 0.25m will still be useful for locating strong anomalies and broad trends, it is considered too low for characterisation of the shape and the properties of the anomalies present. A resolution of 0.5m by 0.25 or 0.125 is better suited for this. While there were small, subtle differences between a sampling of 0.125 and 0.25m, these can still be important for detailed characterisation of the anomalies. If the spatial location is deemed very important, a survey with markers on ropes might be recommended for a 0.125.m resolution. Accurate surveying at this resolution also gives better result when the data are further interpolated. A 0.25 resolution might be more appropriate if walking without ropes, making edge matching easier. For usage when locating and characterising Iron Age Settlements a high resolution is considered important. In the end, the choices made have to be based on the aim of the survey. There is also a possibility to do a large coverage in low resolution detecting areas of increased response, and then increase the resolution within a more targeted area for a more detailed anomaly characterisation. 5.1.3 Earth Resistance Survey This method proved to work well to locate ditches associated with houses and burial mounds. Pits down to a size of 1.5m and ditches down to 0.5-1m width were identified with a 0.5m probe spacing with the twin probe array. There are clearly limitations to what can be detected, and much is size dependent. Not all Iron Age buildings have associated ditches, but rather built with only roof-bearing and smaller wall-bearing post holes not larger than 0.30-0.5m in diameter. Identifying features of 86 this size cannot be assumed to be possible. The effectiveness of the method is also dependent the depth, extent and nature of any archaeology is buried, which closely associated with the thickness of the topsoil. At Gustad this method gave a very good result, indicating the location of the major anomalies at the site. The relatively simple data gathering and processing is clearly an advantage. Resolution of an Earth Resistance Survey on Iron Age Settlement Sites – 0.5m probe spacing Illustration 44a: Raw Earth Resistance Survey results. 0.5m traverse and sampling interval. 0.5m probe spacing. 87 Illustration 44b: Raw Earth Resistance Survey results. 0.5m traverse interval and 1m sampling interval. 0.5m probe spacing. Illustration 44c: Raw Earth Resistance Survey results. 1m traverse interval and 1m sampling interval. 0.5m probe spacing. 88 All three resolutions give good representation of the major anomalies identified. The loss of resolution also reveals a loss of shape definition along the edges of an anomaly, but any anomalies identified with the highest resolution are not missed due to a smaller resolution. Resolution of an Earth Resistance Survey on Iron Age Settlement Sites – 1m probe spacing Illustration 45a: Raw Earth Resistance Survey results. 1m traverse interval and 0.5m sample interval. 1m probe spacing. 89 Illustration 45b: Raw Earth Resistance Survey results. 1m traverse interval and 1m sample interval. 1m probe spacing. The 1m probe spacing also shows good results for locating many of the same anomalies as the 0.5m probe spacing, but with less accuracy. Still no anomalies identified with a 0.5m sample interval is missed with a 1m sample interval. Recommendation If a more exact definition of the size and shape of a burial pit or the width of the walls are deemed necessary for the aims of the survey, then the resolution should be 0.5m traverse and interval sampling as defined by the EH. As the MP15 multiplexer easily could measure 1m sampling interval with a 0.5m traverse interval in one go for a 1m wide strip, a survey aimed for the initial evaluation of a site could well be done in this resolution without missing the presence of the ditches associated with either the houses or the burial mounds. The value of the multiplexer is that it can include a 90 measurement of the 1m probe spacing at the same time. If the topsoil is too thick to render any results with the 0.5m probe spacing, the 1m probe spacing data collected at the same time might still render some useful results. 5.1.4 Ground Penetrating Radar Survey The GPR survey did positively identify both the ditches for the round burial mound and both houses, in addition to drainage ditches, pits and possible post-holes within house 2. It proved to be a valuable addition to the other methods for evaluation Iron Age Settlement Sites. This method is considered to have the best potential to locate smaller features such as postholes, as well as having an advantage compared to other methods by its potential in giving information of depth and volume. In this way GPR could also confirm whether or not a magnetic anomaly detected with a fluxgate gradiometer could be either a dipping, dug out feature or solid. In this way the method can be used to confirm interpretations made on the results of the other methods. Resolution of GPR for locating Iron Age Settlement Sites The following comparisons can be presented, illustrating the effect of increasing the traverse interval, decreasing the resolution and increasing the speed of a survey: 91 Illustration 46a: Comparison of 0.25m (left) and 0.5m. (right) traverse interval Slice 7. Illustration 46b: Comparison of 0.25m (left) and 0.5m. (right) traverse interval Slice 8. 92 The reduced resolution will still be adequate to locate many of the anomalies present at this site. However, a clearer definition of the actual shapes of the anomalies will be lost in doing this. Recommendation As the GPR is considered to be the most appropriate method when trying to locate smaller anomalies as postholes associated with prehistoric building, it can be an important method in future investigation of similar sites. As it is considered quite normal fort houses not to have associated ditches, but rather just consist of postholes of a relatively small size, this method is considered to have a good potential for future use. Experimentation with several types of antenna frequencies and higher survey densities can potentially improve the results further. 5.2 Sequence 5.2.1 Initial site evaluation It has been shown how both the MS survey and the FG survey could be used to visualize activity areas at Gustad. By mapping the number of positive magnetic anomalies within each grid, a map was produced showing roughly the same patterning as the magnetic susceptibility (see illustrations X and X in chapter 4). If a surveyor were to scan a site, marking any anomaly above 3nt, would he then have noticed the same patterning, and resorted in focusing further investigations in the area of highest archaeological interest? This can be modeled after Gaffney and Gater 2003:97-99. 93 Illustration 47: Map showing what anomalies that might have been detected by scanning every 10m with a threshold of 3nT Illustration 48: Map showing the concentration of anomalies detected by scanning compared with the anomalies identified as archaeological by all methods applied at Gustad. 94 This modeling shows which areas that might have been investigated in more detailed following a scanning with fluxgate gradiometers. In this instance the distribution roughly resembles the Topsoil Ms survey presented in illustration X-X earlier in this chapter. Comparing the two illustrations show how the visualized does not entirely reflect the actual distribution pattern of anomalies as seen on illustration X ( anomalies per 20x20m grid) in the previous chapter, while still being useful in delimit the central activity area where the houses are identified. It also gives some indication of how large a survey area should be to reflect the actual distribution of anomalies and give a relevant picture of the background. It underlines the necessity of surveying an area of at least 40x40m, or preferably 60x60m (Gaffney et al. 2002:3, Gaffney and Gater 2003:88-101). The value of topsoil mapping is that it relies less on the experience on the operator, and all recordings are stored and easily reproduced, while keeping a record of the spatial location of the samples. These factors make the topsoil MS sampling preferable. 5.2.2 Sequence of methods It is clear that the different methods reveal different information about any anomalies present at the site, complementing each other with additional information. The gradiometer data was difficult to interpret, but the spatial distribution of the identified anomalies suggest a correlation with the identified activity area with houses and probable cooking pits as well as the topsoil MS results. This suggest that many of the anomalies may very well be related to some form of prehistoric activity. The uncertainty associated with this should be more thoroughly investigated in the future in projects by gradiometer data with actual excavation results to learn more of 95 the magnetic response of different archaeological features expected at a settlement site. The speed of data collection combined with the possibilities of a high spatial resolution is better than the other methods tested at Gustad. If the aim is to locate smaller archaeological features, a high resolution of at least 0.5x0.25m is necessary for finding anomalies of the appropriate size. The Earth Resistance with its simplicity in data collection and processing, and good results for locating ditches associated with houses and ploughed over burial mounds, make the method very appropriate at this site. At settlement sites where the building remains consist of mainly postholes and pits, this method might not have revealed any useful information. Sadly, there might not be any way of knowing this in advance. An ER survey with sampling every 1m along the traverse would have almost doubled the speed of data collection, and by that increasing the potential area surveyed, and still revealed the presence of the ditches at Gustad for evaluation purposes. The introduction of the MSP40 mobile sensor platform could improve the speed of the data gathering. The speed of the GPR data collection with a cart-mounted single frequency antenna gathering data for every 0.25m was not far from the ER survey with samples every 0.5m with a multiplexer. The possibility to better categorise any anomalies detected by FG as either dipping features or hard surfaces, is very appealing. The data processing is more time demanding and difficult compared with other methods, and this should be taken into consideration when planning and budgeting a survey. The EH guidelines suggest a survey sequence topsoil MS or scanning, followed by detailed area FG, ER and GPR in that order. Depending on the time and equipment available, using topsoil MS as the initial survey technique is preferred, followed by area 96 coverage of FG and either ER or GPR. This is because the FG has the advantage of speed and its detection of anomalies that are different in their magnetic response, therefore being more able to discriminate between the intensity of the anomaly strength, indicating features of potential burning like fireplaces or cooking pits. The GPR has an advantage of being able to identify smaller anomalies, with their inherent properties also indicated in sections as well as time slices. GPR also often responds well to the same anomalies as ER. ER cannot be expected to find anomalies as small as postholes, and might not reveal any good results in indicating buildings unless they have associated ditches. In many instances this is not the case. If it stands between a choice of doing GPR or ER, GPR should be preferred. Gathering data in the suggested resolution will be more time consuming than the ER guidelines suggestions for resolution for evaluation purposes (David et al 2008, chapter three), but an increased resolution will increase the confidence in the data gathered and be more appropriate if one hopes to locate settlements of this type. 97 6 CONCLUSION The site of Gustad has a research history spanning over 200 years, giving good archaeological information about monuments at the site. By gathering high resolution topsoil MS, ER, FG and GPR data it was possible to assess the applicability of these methods on a site like this, and suggest a survey strategy for future prospection on similar sites in a Norwegian context. At Gustad it was possible to accurately locate a range of anomalies. The geophysical response, shape, form, location and spatial distribution of anomalies have increased the knowledge about the site, and also demonstrated possibilities and limitations associated with the different methods applied. Topsoil MS mapping proved valuable for identifying the main settlement area, but did not so much indicate the location of burial mounds. FG showed a similar distribution of anomalies as the MS locating numerous anomalies possibly associated with burning and settlement activity in relatively high detail, but their nature was hard to interpret without the presence of additional information. The ER showed very clear results at this site, locating existing ditches associated with building walls and burial mounds, but has limitations in regards to resolution and speed. The GPR used showed results in higher spatial resolution- making the method very applicable for locating minor features like postholes and pits often associated with settlements of this type, while being more time consuming both in regards to data gathering and processing. As this dissertation is concerned with developing a sequential survey design for Iron Age Settlement sites, it has come clear the confidence level of the gathered data is very important. There is limited value in a survey over a large area if the resolution is too low to localise any anomalies that can be associated with key elements that make 98 up a settlement site. While certain methods might be more appropriate in an early stage of data gathering due to speed or geophysical response, a single method can only supply a certain amount of information about the site and its associated archaeology. This is related to questions of resolution and area coverage and anomaly context. The data gathered at Gustad showed how the confidence level in the interpretations and results increased as more methods were included and the resolution was increased. A necessity of an appropriate area coverage for gaining sufficient information about the general background and anomaly contrast was also demonstrated. A survey strategy for locating and delimiting similar Iron Age settlement sites should therefore be adjusted and adapted to ensure a necessary confidence level. As the location of postholes and pit-features are considered vital for the identification of a settlement sites, methods expected to yield detailed enough data for the identification of associated anomalies should be applied with the appropriate resolution. For the site of Gustad, it was clear that GPR proved very applicable because of its potential to locate minor pits like postholes, while FG showed an ability to localise anomalies of the expected size. A coarse earth resistance proved useful to identify any ditches present, while the MS showed clear results for delimiting the settlement area. As a result of the combination each method has provided context for the others, and therefore giving increased confidence in the archaeological interpretations at Gustad where the exact position, size, shape and spatial distribution of anomalies indicated at least two houses, two burial mounds and possible pits of varying kind in addition to other anomalies probably associated with past human activity on the site. 99 6.1 Further work It is clear that resolution and area coverage are important factors for increasing the confidence level of the interpretation of geophysical data. It is also clear that there are limitations to the speed of data gathering for any method of survey. Recent development in Earth Resistance and Ground Penetrating Radar technology might help improve the quality and potential of geophysical surveys in the future. The introduction of the Geoscan Research MSP40-Earth Resistance platform and GPR systems like the 3d-radar Geoscope multi channel and multi frequency system or the multichannel GPR system from Malå Geoscience can increase the acquisition speed and resolution immensely. GPS mounted gradiometer arrays might also increase data acquisition speed (Geoscan Research 2006, Strandli et al. 2007, Trinks et al. 2010, Lindford et al. 2010, Sala and Lindford 2010, Gaffney and Gater 2003, Aspinall et al. 2009). If archaeological geophysics is to be considered valuable within Norwegian Cultural Heritage Management, its potential and benefits need to be proven, and their value needs to be acknowledged by its potential users. Publication of experiences made is vital, as well as targeted research to improve the understanding of the applicability of different methods available. To properly understand the value of gathered data, it is important to acquire the geophysical data in advance any test excavation or full-scale excavation. Any excavation results can be directly compared with the geophysical data collected. The anomaly properties of typical archaeological features can then thoroughly investigated, which in return will increase the quality of future interpretations of 100 geophysical data. This is especially important within Norwegian archaeology, where the application of these techniques is still in their initial stages. 101 7 Bibliography Aspinall, A., C. F. Gaffney & A. Schmidt 2008. Magnetometry for Archaeologists. Lanham, AltaMira Press. Barton, K., L. Stenvik and B. Birgisdottir 2009. A Chieftain’s Hall or a Grave; Ground Penetrating Radar in an Archaeological Geophysics Survey to Target the Excavation of a Cropmark near Stiklestad, Nord-Trondelag,Norway. In: Gómes Martin, R., A. Rubino Bretones, S.G. Garcia, m. Fernández Pantoja & C.M. de Joeng van Coecorden (eds.). Proceedings to the 5th international workshop in Advanced Ground Penetrating Radar (IWAGPR2009),University of Granada,27th-29th May 2009, University of Granada, Granada, Spain,pages 159-164. Available online: http://www.gpr- survey.com/images/IWAGPR2009_Barton_et_al_02.pdf Last accessed: 03.05.2010 Batt, C., S. Fear and C. Heron 1995. The role of Magnetic Susceptibility as a Geophysical Survey Technique: a Site Assessment at High Cayton, North Yorkshire. Archaeological Prospection, Vol.2, 179-96. Bradford. Binns, R. 1994. Med røntgenblikk på fortiden. In Spor number 2. 1994. Trondheim, NTNU- Vitenskapsmuseet. Pages 12-15. Available online: http://www.ntnu.no/vitenskapsmuseet/les-spor-her/- /asset_publisher/MP5t/content/spor-nr-2-1994 Last accessed: 03.05.2010 Binns, R. 2002a. Gradiometer avdekker stor gravhaug på Stiklestad. In: Spor number 1, 2004. Trondheim, NTNU-Vitenskapsmuseet, page 31. Available online: http://www.ntnu.no/vitenskapsmuseet/les-spor-her/- /asset_publisher/MP5t/content/spor-nr-1-2004 Last accessed 03.05.2010 Binns, R.2002b. Grav, naust, hus. In: Spor number 1, 2004. Trondheim, NTNU- Vitenskapsmuseet, pages 32-34. Available online: http://www.ntnu.no/vitenskapsmuseet/les-spor-her/- /asset_publisher/MP5t/content/spor-nr-1-2004 Last accessed 03.05.2010 102 Binns, R.2008. Gradiometerundersøkelse ved Eikertun Prestegård (gbr. nr. 77/1) og Eikertun Sykehjem, Hokksund, Øvre Eiker k., Buskerud, Norge. Geophysical report for Buskerud County Council by RB Reoark. Available online: http://www.eiker.org/Linker/KMR/BFK- registreringer/Eikertun2.pdf Last Accessed: 03.05.2010 Birgisdottir, B. 2009. Rapport. Arkeologisk undersøkelse 2008. Stiklestad Østre 29/1. Verdal Kommune, Nord-Trøndelag. Unpublished excavation report. NTNU-Vitenskapsmuseet. Conyers, L.B. 2004. Ground Penetrating Radar for Archaeology. Plymouth, AltaMira Press. Clark A. 1996. Seeing Beneath the Soil. Prospecting Methods in Archaeology. 2nd edition. Reprinted in 2001. London, Routledge. Corney, M., C.F. Gaffney and J.A.Gater 1994. Geophysical Investigations at the Charlton Villa, Wiltshire (England). Archaeological Prospection, vol.1 -1994. Pages 121-128. David, A., N.Lindford, P.Lindford, L.Martin and A. Payne 2008. Geophysical Survey in Archaeological Field Evaluation. English Heritage. Available online: http://www.english-heritage.org.uk/upload/pdf/GeophysicsGuidelines.pdf Last accessed: 03.05.2010 Dearing J., 1999. Environmental Magnetic Susceptibility. Using the Bartington MS2 System. 2nd edition. Oxford, Bartington Instruments Ltd. Available online: http://www.bartington.com/templates/asset-relay.cfm?frmAssetFileID=96 Last accessed: 27.08.2010 Ethelberg, P., N. Hart, B. Poulsen and A. B. Sørensen. 2003. Det sønderjyske landbrugs historie: jernalder, vikingetid og middelalder.Haderslev, Haderslev museum. Evans, M.E. and Heller, F. 2003. Environmental Magnetism. Principles and Applications of Enviromagnetics. San Diego. Academic Press. Faßbinder, J.W.E and Stanjek, H. 1993. Occurence of bacterial magnetite in soils from archaeological sites. Archaeologia Polona vol.31, p. 117-128. Available online: http://www.iaepan.edu.pl/archaeologia-polona/object/463 Last accessed: 27.08.2010 103 Foosnæs, K. 2009. Jakten på det forsvunne kloster. Arkeologisk utgravninger ved Munkeby kloster 2005. In: Skevik, O. and E.Følstad 2009. Fortida – ei kraft i nåtdia. Foredrag i 2006 og 2007. Stiklestad Nasjonale Kultursenter. Foosnæs, K. 2010. Blikk gjennom bakken: geofysikk og arkeologi. Spor, nr. 1 2010. NTNU Vitenskapsmuseet. Pages 47-49. Forseth, L. 2007a. Flying to the past in Nord-Trøndelag. Aarg News 35 – 2007. Available online: http://www.univie.ac.at/aarg/php/cms/file_download/19/AARGnews35.pdf Last accessed. 20.09.2010 Pages. 18-23. Last accessed: 20.09.2010 Forseth, L. 2007b. Fly inn I forhistorien. Spor, nr.2 – 2007.Trondheim, NTNU Vitenskapsmuseet. Pages Fossum, A. 2010. Geofysikkprosjektet. Report by Vestfold County Council. Available online: http://kulturarv.aipublish.no/ai_files/prosjekter_avsluttede_geofysikkprosjektet_anitr a_634661.pdf Last accessed: 03.05.2010 Gaffney, C., J. A. Gater And S.M. Ovendsen 2002. The Use of Geophysical Techniques in Archaeological Evaluations. IFA Technical Papers 6. Reading, Institute of Field Archaeologists. Available online: http://www.archaeologists.net/modules/icontent/inPages/docs/pubs/PAPER%206%20 Geophys.pdf Last accessed: 03.05.2010 Gaffney, C. and Gater, J.A. 2003. Revealing the Buried Past: Geophysics for Archaeologists. Stroud, Tempus. Gaffney, C. and Gater, J.A.2006. Practice and method in the application of geophysical techniques in archaeology. In: Hunter, J. and I. Ralston 2006. Archaeological Resource Management in the UK: An Introduction. Stroud, Glouchestershire. Alan Sutton Publishing. Pages 292-303. Geoscan Research 2006. MSP 40 Mobile Sensor Platform. Data sheet, Issue 2, October 2006. Available online: http://www.geoscan-research.co.uk/MSP40_v2b_Data_Sheet.pdf Last accessed: 21.09.2010 104 Gjerpe, L.E. (ed.) 2005. Gravfeltet på Gulli. Varia 60, E-18 prosjektet bind 1 2005. Oslo, Kulturhistorisk museum. Available online: http://www.duo.uio.no/publ/kultmus/2005/70572/varia_60.pdf Last visited: 03.05.2010 Göthberg, H. 2000. Bebyggelse I förandring. Uppland från slutet av yngre bronsalder till tidlig medeltid. Occational papers in Archaeology (Opia) 25, Uppsala. Hansson, A., C. Olson, J.Storå, S. Welinder and Å. Zetterström 2005. Agrarkris osch ödegårdar i Jämtland. Østersund. Jamtli Förlag, Jämtland läns museum. Haug, A. 1999. Jernalderhus på Øksendalsøra. Spor nr. 1, 1999. Pages 34-37. Trondheim, NTNU Vitenskapsmuseet. Hvass, S. 1988:Jernalderens bebyggelse. In: Mortensen, P. & B.M. Rasmussen.(Eds.) 1988. Fra stamme til stat i Danmark i Jernalderens stammesamfund. Jysk arkæologisk selskab. Pages 53 –92. Johansen, H.M. 2007. Nausttufter i Midt-Norge. Utbredelse, konstruksjon og datering. Naust og samfunn I jernalderen belyst ved eksempler fra Nord-Trøndelag. Unpublished post- graduate thesis. The Norwegian University of Science and Technology. Institute for Archaeology and Religion. Jordan D. 2007. Evaluating Aggregate in North West England: the Effectiveness of Geophysical Survey. Report on English Heritage ALSF project number 3835. Available online: http://ads.ahds.ac.uk/catalogue/resources.html?nwgeophysics_eh_2007 last accessed 20.09.2010 Jordan, D. 2009. How effective is geophysical survey? A regional review. Archaeological Prospection, vol.16 – 2009. Pages 77–90. 105 Klüwer, L.D. 1823. Norske Mindesmærker aftegnede paa en reise igjennem en deel af det Nordenfieldske. Det kongelige Norske Videnskapers Selskap, Trondheim. Available online: http://books.google.com/books?id=0ZcBAAAAYAAJ&oe=UTF-8 Last accessed: 03.05.2010 Kvamme, K. 2003. Geophysical Surveys as Landscape Archaeology. American Antiquity, Vol. 68, No. 3 (Jul., 2003), pages. 435-457 Linford, N., P.Linford, L. Martin, and A.Payne 2010. Stepped frequency ground- penetrating radar survey with a multi-element array antenna: Results from field application on archaeological sites. Archaeological Prospection. Vol. 17-2010. Pages 187–198. Lorra, S. 2003. Geophysical Exploration of the sites Gulli and Rom Vestre Using Ground Penetrating Radar and Magnetics. Report on field survey and result July 2003. Report in the archives of Kulturhistorisk Museum, Oslo. Løken, 1994. En folkevandrngstids gård – Bertnem i Overhalla. Spor nr.2 – 1992. Trondheim, NTNU Vitenskapsmuseet. Pages 26-29. Løken,T. 1999. The longhouse of Western Norway from the late Neolithic to the 10th Century AD: Representatives of a common Scandinavian building tradition or a local development? In: Schjelderup, H. and Storsletten, O. (red.) 1999. Grindbygde hus i Vest-Norge. NIKU-seminar om grindbygde hus, Bryggens Museum 23-25.03.98. NIKU temahefte 030: 1-128. Available online: http://www.niku.no/archive/niku/publikasjoner/NIKU%20temahefte/30_grindbygg1.p df Last accessed: 18.09.2010 Martens, J. 2003. Gravhaugen på Rom vestre, 113/106. Tønsberg kommune, Vestfold. Rapport over forundersøkelsen 10. juni 2003. Report in the archives of Kulturhistorisk Museum, Oslo. Myhre, B. 1972. Funn, fornminner og ødegårder. Jernalderens bosetning i Høyland Fjellbygd. Stavanger museums skrifter, bind 7. Stavanger. 106 Myhre, B. 2002. Landbruk, landskap og samfunn 4000 F.KR.-800 E.KR. In: Myhre, B. and I.Øye 2002. Norges Landbrukshistorie 1. 400 f.kr.-350 e.Kr. Jorda blir levevei. Det norske samlaget, Oslo. Pages 11-213. Neubauer W., Doneus M., Eder-Hinterleitner, A., Löcker, K. 2003. The Early Neolithic monumental enclosure Weinsteig-Grossrussbach, Austria. Archaeologia Polona vol.41 (Warschau) p. 236-238. Pedersen, E.A. 1990. Rydningsrøysfelt og gravminner – spor av eldre bosetningsstruktur på Østlandet. In: Viking Bind LIII – 1990. Oslo. Pages 50-66. Pedersen, E.A. and M.Widgren 1998. Järnalder 500 f.kr. – 1000 e.kr. In: Welinder, S., E.A.Pedersen and M.Widgren 1998. Jordbrukets f¨rsta fermtusen år. 4000 f.kr. – 1000 e,kr, Det svenska jordbrukets historia. Borås, Natur och kultur/LTs förlag. Pages 237- 459. Persson, K. 2006. Rapport. Geofysisk prospektering. Avaldsnes, Karmøy. GeoFysica, April 2006. Available online: https://secure.vikingkings.com/Shared/Image.aspx?IsDoc=true&imageID=be089e58- 9d52-4499-9ef4-c7987c29467b Last accessed: 09.09.2010 Petersen, T.H. 1927. “Ekne, Gustad I Skogn Prestegjeld”. Letter in the topographical archives of The Museum of Archaeology and Natural History in Trondheim. Document number 000572. Riksantivaren 2010. Høringsutkast. Strategisk plan for forvaltningen av arkeologiske kulturminner og kulturmiljøer 2010-2020. Oslo, Riksantikvaren. Availlable online: http://www.riksantikvaren.no/filestore/STRATEGISKPLAN-HRINGSUTKAST.pdf last visited 19.09.2010 Sala, J. and N. Lindford 2010. Processing stepped frequency continuous wave GPR systems to obtain maximum value from archaeological data sets. 13th international conference on Ground Penetrating Radar (GPR) 2010. Lecce, Italy. Pages 1-6. 107 Sivertsen, M. 1951. “Buplasspor I Skogn-Området (Skogn-Frol) pr. november 1951”. Letter in the topographical archives of The Museum of Archaeology and Natural History in Trondheim. Document number 000574. Sivertsen, M. 1965. Skogns historie band III. Om Skognområdet i førhistorisk tid. Published by ”Nembda for Skogn Historie”. Levanger. Verdal Boktrykkeri. Sjöborg, N.H.1822. Samlingar för Nordens Fornälskare. Archäologiska Selskapet, Stockholm. Available online: http://books.google.com/books?id=eVAGAAAAQAAJ&oe=UTF-8 Last accessed: 03.05.2010 Skog og Landskap 2010. Jordsmonnkart. Digital maps on soil types, erosion, drift geology etc. Available online: http://www.skogoglandskap.no/kart/jordsmonnkart_og_statistikk Last accessed: 20.09.2010 Smekalova, T. and B. Bevans 2009a. A Geophysical Test at Spangereid. Report, Vest-Agder County Council. Available online: http://www.vaf.no/Spangereid_- _geofysiske_unders%C3%B8kelser_juni_2009_uYhcD.pdf.file Last accessed: 03.05.2010 Smekalova, T. and B. Bevans 2009b. A Geophysical Survey at Huseby Report, Vest-Agder County Council. Available online: http://vaf.no/Huseby_PKxWo.pdf.file Last accessed: 03.05.2010 Solberg, B. 2000. Jernalderen i Norge 500 før Kristus til 1030 etter Kristus. Cappelen Akademisk Forlag, Oslo. Stamnes, A.A. 2008. Jernalderens jordbruksbosetning i Nord-Trøndelag. Utvikling og testing av en prediktiv model. MA-thesis at the University of Science and Technology, NTNU, Trondheim, Norway. Available online: http://www.ntnu.no/eksternweb/multimedia/archive/00037/MA_Stamnes_37088a.p df Last accessed: 03.05.2010 108 Stamnes, A.A. 2010. Prediktive modeller og lokaliseringen av jernaldergården – et eksempel fra Nord-Trøndelag. Primitive tider 2010 12. årgang, p. 41-50 Strandli, C.W., I.Dragland, A. Bjerkhagen and A.A.Stamnes 2007. Archaeology and Geophysics. Exploring for Ancient Subsurface Structures Surrounding the Nidaros Cathedral in Trondheim. Unpublished Report. Statoil Research Centre, Rotvoll, Trondheim. Trinks, I., P. Karlsson, A. Eder-Hinterleitner, K. Lund, L.-I. Larsson 2007a. Borreparken. Digital report, Vestfold County Council. Available online: http://www.vfk.no/doc/kulturminnevern/geofysikk/Borre4/Borre4_hallene.pdf Last accessed: 03.05.2010 Trinks I, P. Karlsson, B. Westergaard, A. Eder-Hinterleitner, L.-I. Larsson 2007b. Professional archaeological prospection Odberg Site 1 2007. UV Teknik, Swedish National Heritage Board, Archaeological Excavations Department. Available online: http://www.vfk.no/doc/kulturminnevern/geofysikk/Odberg_Site_1/web/Odberg_Site_ 1_web_files/frame.htm Last accessed: 03.05.2010 Trinks I, P. Karlsson, B. Westergaard, A. Eder-Hinterleitner, L.-I. Larsson 2007c. Professional archaeological prospection Odberg Site 2 2007. UV Teknik, Swedish National Heritage Board, Archaeological Excavations Department. Available online: http://www.vfk.no/doc/kulturminnevern/geofysikk/Odberg_Site_2/web/Odberg_Site_ 2_files/frame.htm Last accessed: 03.05.2010 Trinks, I., B.Johansson , J..Gustafsson., J.Emilsson., J.Friborg, C.Gustafsson, J.Nissen. and A.Hinterleitner 2010. Efficient, large-scale archaeological prospection using a true three-dimensional ground-penetrating Radar Array system. Archaeological Prospection. Volume 17 – 2010. pages 175–186. Vestrum, A. 1933. Skogns historie annet bind. Levanger, Nordre Trondhjems Amtstidenes boktrykkeri. Østmo, E. And L. Hedeager 2005 (eds.). Norsk arkeologisk leksikon. Pax forlag, Oslo. 109 Øye, I. 2002. Landbruk under press 800-1350. In: Myhre, B. and I.Øye 2002. Norges Landbrukshistorie 1. 400 f.kr.-350 e.Kr. Jorda blir levevei. Det norske samlaget, Oslo. Pages 214-414. 110 8 Appendix 8.1 Grid Coordinates Illustration 49: Map showing the grid and survey points. Measured in with a RTK GPS with an accuracy of ±2cm Point number East Coordinate North Coordinate Z height 1 602092,349 7064835,407 8,908 2 602179,537 7064856,952 10,608 3 602180,075 7064876,755 8,999 4 602180,456 7064896,696 8,067 5 602180,935 7064916,540 7,013 6 602181,269 7064936,473 5,897 7 602201,708 7064956,179 6,048 8 602201,267 7064936,216 7,151 9 602200,909 7064916,204 7,733 10 602200,440 7064896,373 8,712 11 602200,066 7064876,400 9,911 12 602199,478 7064856,525 11,873 13 602219,980 7064875,836 10,822 14 602220,386 7064895,924 9,443 15 602220,868 7064915,899 8,671 111 16 602221,228 7064935,856 8,214 17 602221,612 7064955,813 7,508 18 602242,008 7064975,425 7,851 19 602241,638 7064955,474 8,201 20 602241,229 7064935,476 8,931 21 602240,820 7064915,463 9,398 22 602240,313 7064895,483 10,445 23 602239,948 7064875,557 11,938 24 602259,771 7064875,159 12,935 25 602260,352 7064895,115 11,362 26 602260,771 7064915,059 10,248 27 602261,201 7064935,100 9,710 28 602261,606 7064955,070 9,271 29 602262,370 7064974,996 7,827 30 602262,762 7064994,948 6,821 31 602282,621 7064994,430 7,168 32 602282,157 7064974,448 8,325 33 602281,739 7064954,565 9,831 34 602281,265 7064934,585 10,255 35 602280,791 7064914,599 11,093 36 602280,309 7064894,670 12,538 37 602300,277 7064894,207 13,808 38 602304,915 7064882,694 15,347 39 602300,785 7064914,131 12,210 40 602301,275 7064934,107 11,226 41 602301,691 7064954,087 10,252 42 602302,225 7064974,054 8,764 43 602302,464 7064994,045 8,161 44 602322,178 7064973,629 9,953 45 602321,703 7064953,665 10,582 46 602321,247 7064933,605 11,968 47 602320,718 7064913,717 13,329 48 602354,360 7064943,694 13,388 49 602359,875 7064934,730 14,057 Table 9: Coordinates and height measurement for each point in UTM 32 coordinates. 112 8.2 Unprocessed Earth Resistance data Illustration 50: Earth Resistance, 0.5m Probe spacing. Raw, Unprocessed data. Illustration 51: Earth Resistance, 1m Probe spacing. Raw, Unprocessed data. 113 8.3 Fluxgate Gradiometer – Zero Mean Traverse test Illustration 52: Fluxgate Gradiometer Data which ha only been Zero Mean Traversed. The image shows difficulties in removing striping when having strong geological anomalies affecting the mean values calculated for each traverse. 114 8.4 Ground Penetrating Radar – All time slices Illustration 53: GPR slice 1 - 0-2.66 ns or 0-0.09m with V=0.07m/ns Illustration 54: GPR slice 2 – 1.62-4.28 ns or 0.06-0.15m with V=0.07m/ns 115 Illustration 55: GPR slice 3 – 3.25-5.91 ns or 0.11-0.21m with V=0.07m/ns Illustration 56: GPR slice 4 – 4.88-7.53 ns or 0.17-0.26m with V=0.07m/ns 116 Illustration 57: GPR slice 5 – 6.5-9.16 ns or 0.23-0.32m with V=0.07m/ns Illustration 58: GPR slice 6 – 8.12-10.78 ns or 0.28-0.38m with V=0.07m/ns 117 Illustration 59: GPR slice 7 – 9.75-12.41 ns or 0.34-0.43m with V=0.07m/ns Illustration 60: GPR slice 8 – 11.38-14.03 ns or 0.4-0.49m with V=0.07m/ns 118 Illustration 61: GPR slice 9 – 13-15.66 ns or 0.46-0.55m with V=0.07m/ns Illustration 62: GPR slice 10 – 14.62-17.28 ns or 0.51-0.6m with V=0.07m/ns 119 Illustration 63: GPR slice 11 – 16.25-18.91 ns or 0.57-0.66m with V=0.07m/ns Illustration 16: GPR slice 12 – 17.88-20.53 ns or 0.63-0.72m with V=0.07m/ns 120 Illustration 64: GPR slice 13 – 19.5-22.16 ns or 0.68-0.78m with V=0.07m/ns Illustration 65: GPR slice 14 – 21.12-23.78 ns or 0.74-0.83m with V=0.07m/ns 121 Illustration 66: GPR slice 15 – 22.75-25.41 ns or 0.8-0.89m with V=0.07m/ns Illustration 67: GPR slice 16 – 24.38-27.03 ns or 0.85-0.95m with V=0.07m/ns 122 8.5 Ground Penetrating Radar – Spatial resolution with varying RDP values. For the 400 Mhz centre frequency antenna used in this survey, this table of calculated wavelengths of radar waves in Media of a Given RDP and frequency can be presented (1/2 to ¼ of the wavelength, see chapter three) : RDP Wavelengths at 400 MHz Spatial resolution (0,25- Frequency in Meters 0,5 λ) 1 (Air) 0.750 18,75 – 37,5 cm 3 0.433 10,83 – 21,65 cm 6 0.306 7,65 – 15,3 cm 9 0.250 6,25 – 12,5 cm 12 0.216 5,4 – 10,8 cm 15 0.194 4,85 – 9,7 cm 20 0.168 4,2 – 8,4 cm 40 0.119 4 – 5,8 cm 60 0.097 2,43 – 4,85 cm 80 0.084 2,1 – 4,2 cm Table 10: Calculations of spatial resolution with varying RDP values. 8.6 Ground Penetrating Radar – RDP values of different types of soil Material RDP Air 1 Dry Sand 3-5 Dry silt 3-30 Ice 3-4 Asphalt 3-5 Vulcanic ash/pumice 4-7 Limestone 4-8 Granite 4-6 Permafrost 4-5 Coal 4-5 Shale 5-15 Clay 5-40 Concrete 6 Saturated silt 10-40 Dry sandy coastal land 10 Average organic-rich surface soil 12 Marsh or forested land 12 Organic-rich agricultural land 15 Saturated sand 20-30 Fresh water 80 Sea water 81-88 Table 11: Typical Relative Dielectric Permittivities (RDPs) of Common Geological Materials. From Conyers 2004:47. 123 8.7 Known Geophysical Surveys in Norway Site Farm Kommun County Year Institution Performed by Company e Hoset Stjørdal Nord-Trøndelag 1973 NTNU Aalstad, I. og Åm, K. NGU Borreparken - Haug 1 Horten Vestfold 1988 KHM Pedersen, O.C. & Veslegard, G. Noteby AS Borreparken - Haug 6 Horten Vestfold 1988 KHM Pedersen, O.C. & Veslegard, G. Noteby AS Dokka Nordre Land Oppland 1988 KHM Noteby AS Borreparken - Haug 1 Horten Vestfold 1989 KHM Pedersen, O.C. & Veslegard, G. Geomap AS Borreparken - Haug 6 Horten Vestfold 1989 KHM Pedersen, O.C. & Veslegard, G. Geomap AS Borreparken - Haug 7 Horten Vestfold 1989 KHM Pedersen, O.C. & Veslegard, G. Geomap AS Rotvold Rotvold Trondheim Sør-Trøndelag 1991 NTNU Richard Binns RB Geoarch Vitenskapsmuseet Hallstein Halsan Østre Levanger Nord-Trøndelag 1993 Binns, R. RB GeoArk Gardermoen Ullensaker Akershus 1993 KHM Værem - ringforma Værem Grong Nord-Trøndelag 1993 Nord-Trøndelag Binns, R. RB GeoArk tunanlegg fylkeskommune Hadsel Hadsel Nordland 1994 Binns, R. RB GeoArk Greipstad - settlement Greipstad Kvaløya? Troms 1994 Tromsø museum Richard Binns/R.Eilertsen history Agdenes del 1 Agdenes Agdenes Sør-Trøndelag 1995 NTNU Binns, R. RB GeoArk Vitenskapsmuseet Agdenes Agdenes Sør-Trøndelag 1996 NTNU Binns, R. RB GeoArk Agdenes del 2 Vitenskapsmuseet Halvdanshaugen Stein Hole Buskerud 1997 KHM Pedersen, O.C. Geomap AS Gjevran Gjevran Steinkjer Nord-Trøndelag 1997 Egge Historielag? Binns, R. RB GeoArk Halvdanshaugen Stein Hole Buskerud 1999 KHM Pedersen, O.C. Geomap AS Kaupang Larvik Vestfold 2000 UiO Binns, R. RB GeoArk Helge Helge Steinkjer Nord-Trøndelag 2001 Egge Historielag Richard Binns RB Geoarch Olav's Wall Sarpsborg Østfold 2002 Borgarsyssel Horsley, T.J. University of Bradford museum 124 Milde Milde store Bergen Hordaland 2002 UiB Interconsult ICG ASA Greipstad Kvaløya? Troms 2002 Tromsø museum Richard Binns/R.Eilertsen RB GeoArk? Greipstad - settlement history Gulli Tønsberg Vestfold 2003 KHM Lorra, S. Allied Associates Geophysical Ltd. Stiklestad - Gravhaug Stiklestad? Verdal Nord-Trøndelag 2003 Binns, R. RB GeoArk Inderøy Grønnesby Inderøy Nord-Trøndelag 2004 Binns, R. RB GeoArk Munkeby kloster Munkeby Levanger Nord-Trøndelag 2005 NTNU Eide, E 3d-Radar AS søndre Tyin Vang Oppland 2005 KHM Smekalova, T. St. Petersburg State University Meråker - Meraker Meråker Nord-Trøndelag 2005 Meråker Binns, R. RB GeoArk haugformasjon bygdemuseum Tautra kloster Frosta Nord-Trøndelag 2006 Fortidsminneforenin Eide, E 3d-Radar AS gen Hovden Bykle Aust-Agder 2006 KHM Smekalova, T. St. Petersburg State University Bommestad Bommetad Vestfold 2006 Kulturhistorisk Smekalova, T. -Rødbøl museum UiO Stiklestad Haug Verdal Nord-Trøndelag 2007 NTNU Gibson, H., Bonsall, J. & Barton, K. SLIGO Avaldsnes Karmøy Rogaland 2007 Persson, K. Universitetet i Stockholm Nidarosdomen Trondheim Sør-Trøndslag 2007 Statoil, NiKU, NTNU, Stamnes, A., Bjerkhagen, A., Statoil, NiKU, NTNU, 3d- 3d-radar Strandli, C.W. & Dragland, I. radar Borreparken - Haug? Horten Vestfold 2007 Vestfold Trinks, I. UVTeknik fylkeskommune Gokstad Sandefjord Vestfold 2007 KHM Trinks, I. UVTeknik Benan Benum Steinkjer Nord-Trøndelag 2007 NTNU Binns, R. RB GeoArk Øvre Vitenskapsmuseet Eikertun Eikertun Øvre Eiker Buskerud 2008 Øvre Eiker Binns, R. RB GeoArk prestegård kommune Falstad Levanger Nord-Trøndelag 2008 NTNU Barton, K. & Langås, A. Earhsounds Associates 125 Stiklestad vei Stiklestad Verdal Nord-Trøndelag 2008 NTNU Barton,K. Earthsound Associates Vitenskapsmuseet Storbreen Storbreen Oppdal sør-Trødelag 2008 NTNU Barton, K & Gibson, H. Earthsound Associates Vitenskapsmuseet Borreparken Horten Vestfold 2009 NIKU 3d-Radar AS Ekeberg Oslo Oslo 2009 Byantikvaren, Oslo Barton, K. & Stamnes, A. Earthsound Associates Grefsen Oslo Oslo 2009 Byantikvaren, Oslo Barton, K. & Stamnes, A. Earthsound Associates Værne Rygge Østfold 2009 RA/NIKU Karlsson, P. UVTeknik Falstad Levanger Nord-Trøndelag 2009 NTNU Barton, K. , Foosnæs, K & Earhsounds Associates Stamnes, A. Spangereid Lindesnes Vest-Agder 2009 Vest-Agder Bevan, B. og Smekalova, T. fylkeskomune Avaldsnes Karmøy Rogaland 2009 UiO Neubauer, W. VIAS Avaldsnes Karmøy Rogaland 2009 UiO Smekalova, T. og Bevan, B. Aarhus Universitet/Geosight Storbreen Storbreen Oppdal sør-Trødelag 2009 NTNU Barton, K Earthsound associates Vitenskapsmuseet Borreparken Horten Vestfold 2007? Vestfold Trinks, I UVTeknik fylkeskommune Odberg Larvik Vestfold 2007? Vestfold Trinks, I UVTeknik fylkeskommune Domkirkeodden Hamar Hedmark 2008? Barton, K. & Bonsall, J. Earthsound associates, NTNU? Sørbø Sørbø Rogaland 2009? Arkeologisk museum Will Davies i Stavanger Tilrem Brønnøysund Nordland ? Binns, R. RB GeoArk Kvaløy Frøya Troms ? Binns, R. RB GeoArk Brønnøysund Brønnøysund Nordland ? Binns, R. RB GeoArk Table 12: Known geophysical Surveys within Norwegian Archaeology including the year 2009 126 8.8 Categories used for interpretation of the geophysical data Magnetic Archaeology Pattern, shape and form that clearly indicate an archaeological origin. Might be of any age, and certainty of classification might be backed up by other sources. These can be either positive or negative anomalies, depending on the surrounding magnetic response. ? Archaeology The strength or shape of anomaly might indicate anthropogenic origin, but might be incomplete or have some inherent uncertainty about them. Industrial Strong magnetic signal, which in their context can be assumed to be of industrial origin. This can be kilns, ovens, furnaces, metal working or similar, but the shape and nature of the anomalies can be of modern origin. Ferrous Response Strong dipolar responses interpreted as ferrous. This can be either smaller, shallow metal objects, larger strong anomalies with a distinct negative halo or linear “beaded” dipolar responses due to iron pipes. Ridge and Furrow Linear and broad anomalies typically associated with ancient agriculture. Modern activities can produce similar responses. Non-archaeological linear trend Very linear responses that might be due to modern agricultural practices or possible pipes. Might Be aligned with present day field boundaries or known from known modern agricultural practice. Pipelines connected to strong dipolar anomaly, identified by manhole covers or similar on the surface, falls into this category. Trend Linear trend of unknown origin. Might be weak, isolated or obscured. Natural Patterns of magnetic anomalies from geological or natural processes. These could be paleochannels, magnetic gravels or similar. Soft, rounded, but clear signal can be associated with this category. ? Natural Anomalies most likely to be of natural origin, but associated with some uncertainty. Areas of Increased Magnetic Where there are no visual indications or features at the Response ground surface, but the context suggest an origin that might be of archaeological interest. Areas of Magnetic Disturbance Area of increased magnetic activity, often noisy and without any possibilities of positive identification of individual anomalies. Can be the cause of modern rubble, brick, concrete or other materials. Areas of lower magnetic activity Areas that stand out with less magnetic activity than its immediate surroundings or taken the survey area into consideration as a whole. Might be the due to landscaping or difference in land use. Uncertain Origin Anomalies that stand out from the background either as negative or positive anomalies, but with no shape or context easily associated with any other category. Table 13: categories used for interpreting Fluxgate Gradiometer data 127 Resistance – can be of either high High resistance anomalies are for example hard surfaces, walls, compact ground, or well drained ditches or low resistance due to their lack of ability to retain moisture than the surrounding soil. Low resistance anomalies have often an increased ability to retain more moisture than the surrounding soil. This can be for example ditches, pits, drains, gullies or graves. Archaeology Anomalies that are probably archaeological due to their nature and context. Can be of any age. ? Archaeology The strength or shape of anomaly might indicate anthropogenic origin, but might be incomplete or have some inherent uncertainty about them. Natural Patterns of resistance anomalies from geological or natural processes. These could be paleochannels, rock outcrops etc. ? Natural Anomalies most likely to be of natural origin, but associated with some uncertainty. ? Landscaping/Topography These are responses which can be due to landscape alterations or topographical effects that can be identified. This can also be more modern paths or alterations. Vegetation Isolated or grouped narrow linear responses that can be associated with visible or known vegetation. This could also be patches of less or more drained areas, due to the evapotranspiration of plants or trees. Trend Linear trend of unknown origin. Might be weak, isolated or obscured. Uncertain Origin Anomalies that stand out from the background either as negative or positive anomalies, but with no shape or context easily associated with any other category. Table 14: Categories used for interpreting Earth Resistance Data Ground Penetrating Radar Archaeology When an anomaly has a shape, pattern or form indicating an archaeological origin. Might be supported by other evidence, and with consideration to the context. ? Archaeology Response that can be archaeologically significant, but lacks certainty due to lack of context or discrete response. Historic Anomalies showing same shape or pattern as historical evidence. ? Historic Anomalies not known from earlier historic evidence, but seem to respect features that are known from historical sources. Areas of Anomalous Response Areas of anomalous Response that is significantly different from its surrounding background. Table 15: categories used for interpreting ground Penetrating Radar Data 128 8.9 Levels of data processing in Geoplot for the different geophysical images in the Data Interpretation Chapter Earth Resistance Processing Zero Mean Zero Mean Grids Edge High Pass Low Pass Visualisation Illustration Data Grid Traverse Matched Despiked Multiplied Interpolated Filtered Filtered Range Resistance 0.5m 22 Twin Probe No No Yes X=1 Y=1 Thr= 3 Mean No No No No 105.220 Ω Resistance 0.5m X=8 Y=8, X=2 Y=2 23 Twin Probe No No Yes X=1 Y=1 Thr= 3 Mean No No Uniform Gaussian ±2 Std.Dev Resistance 1m 25 Twin Probe No No Yes X=1 Y=1 Thr= 3 Mean No No No No 100-145 Ω 26 Resistance 1m X=8 Y=8, X=2 Y=2 Twin Probe No No Yes X=1 Y=1 Thr= 3 Mean No No Uniform Gaussian ±2 Std.Dev Resistance 0,5m 44a Twin Probe No No Yes X=1 Y=1 Thr= 3 Mean No No No No 105.220 Ω Resistance 0,5m 44b Twin Probe No No Yes X=1 Y=1 Thr= 3 Mean No No No No 105.220 Ω Resistance 0,5m 44c Twin Probe No No Yes X=1 Y=1 Thr= 3 Mean No No No No 105.220 Ω Resistance 1m 45a Twin Probe No No Yes X=1 Y=1 Thr= 3 Mean No No No No 105.220 Ω Resistance 1m 45b Twin Probe No No Yes X=1 Y=1 Thr= 3 Mean No No No No 105.220 Ω Gradiometer Processing Zero Mean Zero Mean High Pass Low Pass Visualisation Illustration Data Grid Traverse Destaggered Despiked Multiplied Interpolated Filtered Filtered Range Gradiometer 15 Bartington No No No No No No No No -5 -> 5 nT Gradiometer Yes , Threshold X=10 Y=10 16 Bartington No = ±10 No No No No Gaussian No --5 -> 5 nT Gradiometer Yes , Threshold X SinX/X, X=10 Y=10 17 Bartington No = ±10 No No No times 1 Gaussian No -5 -> 5 nT 129 Gradiometer Yes , Threshold X=10 Y=10 40 Bartington No = ±10 No No No No Gaussian No -5 -> 5 nT Gradiometer Yes , Threshold X=10 Y=10 41 Bartington No = ±10 No No No No Gaussian No -5 -> 5 nT Gradiometer Yes , Threshold X=10 Y=10 42 Bartington No = ±10 No No No No Gaussian No -5 -> 5 nT Gradiometer Yes , Threshold X=10 Y=10 43c Bartington No = ±10 No No No No Gaussian No -5 -> 5 nT Gradiometer Yes , Threshold X SinX/X, X=10 Y=10 43d Bartington No = ±10 No No No times 1 Gaussian No -5 -> 5 nT Gradiometer Yes , Threshold X=10 Y=10 43g Bartington No = ±10 No No No No Gaussian No --5 -> 5 nT Gradiometer Yes , Threshold X=10 Y=10 43h Bartington No = ±10 No No No No Gaussian No --5 -> 5 nT Table 16: Levels of data processing in Geoplot for the different geophysical images 130