2nd Yr. B.Sc. Archaeology
St.Aidan’s College
Durham University
"The archaeological applications of GIS on the Hayton landscape" is the dissertation I plan to undertake as part of my final year of my B.Sc. Archaeology course at the University of Durham. The area in question, that of Hayton in East Yorkshire, is an area under long term study by Martin Millet, of the University of Durham and Peter Halkon. The area of Hayton, in turn, is part of a larger exploration of the River Foulness valley, in particular the Iron Age and Roman periods. The area in question is that of a 3 x 3 km square, centred around the village of Hayton.
The area comprises of fairly flat moor area, with a roman road running SE-NW and Hayton Beck, a stream which runs nearly W-E. The areas includes a Roman fort, a roman settlement and other buildings, a later Roman bath house and an Iron Age settlements complete with field systems. (See image)
I intend to conduct research to see how modern GIS techniques can be applied on a small scale area such as Hayton. This will mean taking the maximum possible amount of available data for the site and it's area, and seeing how it can be exploited and analysed using GIS. I chose this as GIS is a continually more important aspect of archaeology, and there is relatively little published reports of the applications on sites of this scale. Most published reports are either on large scale areas (60 X 60 km, for example), or on a non archaeological sites. There is still uncertainty about how useful GIS can be within the field of archaeological study. I hope not only to process and analyse the data from Hayton, so as to be able to display data in a way that would not otherwise be possible, but also to be able to comment on the effectiveness of GIS on this scale, on a site such as this.
My dissertation is to be split to have three main aims, each one dependent, although time will decide if all three can be completed.
1.) The digitising of all data from the site and its presentation.
At the moment the following data is or possibly will be available:
I plan to convert (in most cases digitise) this data so as it will be able to be used in a GIS computer package. The program most likely to be used is that of ArcInfo, available in department
Probably the first and most important part of the excavation to digitise will be the features themselves. By representing the settlements, streams, and roads as a series of lines, polygons or points a vector map can be constructed of the most important areas of study. Each piece of data will be designated as being either a road, building or whatever it may be.
The topographic data is in the form of contour maps. By digitising the contour lines, with their altitude values attached the contours can be transferred to computer which can make a 3-D (in this case a DEM - digital elevation model) image by constructing a Triangulated Irregular Network from the contours. The department has a digitiser which can be used directly with the ArcInfo GIS programme. I hope to construct a DEMof the area using ArcInfo, probably the best GIS programme available, at least within the University. It is noted however that the contour intervals of the maps are of only 5m, and the landscape of the area is that of one of gently sloping terrain, with no huge differences in gradient or height. However, I feel it important to construct one, not only for the spatial representation of sites in the area - even if only in 2-D, but also it can then be used to represent the other data listed above.
The soil data is comprehensive of the whole area. It is available in the form of a map, with each soil type represented by a different area. These can be digitised by using the digitiser in department as in the above method. The areas of soil are best done in a vector, as in this way they can be recorded as polygonal areas with lines as their borders. A resulting 2-D image can then be 'draped' over the 3-D topographic model, to form a 2.5-D image, thus representing the soil data and its relationship to the landscape.
The following is an example of a 2.5-D image. This is from a 3-D image constructed from a TIN of the holyland, with an 2-D Landsat image draped over the top to create a 2.5-D image.
There is a range of geophysical data from the excavation already available within department, and is available in a format so that it can be used with Geoplot or Insite. With these programmes the data can be processed as an image and saved as an image file, so that it can then be uploaded to the GIS computer packages. Here the geophysical data can be represented with its relationship to other features such as the settlement data, soil data and topographic. This could also be shown combined with other data, although the scale of the data would mean that a 2.5-D representation would probably be useless.
Likewise the nature of the aerial photography is unknown. Usually this type of data is scanned in rasta format, being treated more like an image, each pixel acting as cell which contains data. This then can also be hopefully integrated with other data types.
The fieldwalking data is in the form of the values of collections of artefacts found on the field surfaces. In this area the main artefact found is that of Roman pottery. However, quite how detailed the data is (whether it is overall finds, or it is subdivided into classes) is still unknown due to the fact that the data has still not be seen. This can be digitised by plotting the density of the finds with in 10m or 5m square (as has been done on certain areas already of the area). One of my aims is to construct a 2-5 D map of the terrain with the density of pot sherds found - again using ArcInfo and the method described above.
2. Construction of layer cake
In the above discussion each type of data can be considered as a layer of data in the same area. It is possible to create or combine certain layers to represent certain features. This can be done for example on a site to show the location of towns, rivers, and roads, with each field being represented by an image. An example of this is this image here:
Here, from a site in Burgundy, features such as water systems, towns and areas of archaeological interest are represented by a layer. In the Hayton area the features and settlement cover a period of time stretching from Iron Age to Later Roman times. It is my intention therefore to construct a layer cake image of Hayton showing the evolution of sites in the area over time. This will comprise of a series of images stacked upon one another, with each one showing the Hayton area and its features at a certain time.
3. Site prediction model
If time permits me, (and the data) I would like to try and construct a site prediction mode. I would use a multiple regression model based on pottery finds and their relationship with aspect, altitude, soil data and other variables. Even though all the sites within this area are known, thus there is no real need, this can be used as advantage so as to test the effectiveness of such a model on such a small area. This part of the dissertation will only be attempted if time permits, and the data is suitable.
My main concern is that of adequate time. I have never digitised anything before, and although I have used certain GIS packages on computer, including ArcView, I have not as yet really used ArcInfo, which to my understanding is slightly less user friendly. Therefore I expect to have to spend some time learning the required skills for the tasks. Although I consider myself quite adept at learning how to use computer packages, ArcInfo is still quite a complicated piece of software - but it is the best package available to me for this type of work. Another problem with time is that there is only one digitiser within department, and I will not be the only one wanting to use it. The department is only open from 9am to 5 pm, and as an undergraduate, I am not allowed to work out of hours (although this maybe open to negotiation), and from only Mon-Fri. This is means that I am unable to work at weekends and nights, which is prime time for studying, and during the weeks, no matter how keen I am, I will still have to go to lectures for other modules. As per regards the third part, my statistical knowledge is not negligble - I have an A-grade Statistics A-level, but I have never done anything like this before. I understand that ArcInfo has many options to do with statistical analysis, although I will not know the full extent of these until I become aquainted with the system. Before trying to do this, I would have to try and do some research on this area, investigating others’ methodology, and see how others have fared. It is my experience, as mentioned above, that there is relatively little published material, and even less goes into much detail of methodology used.
GIS, in theory, has limitless applications, and I could perform thousands of different tests and construct many different graphs. Therefore this dissertation, although having stated several aims, is quite open ended - depending on time and resources. Due to the fact that I am learning the technology's limitations whilst doing my dissertation it is not inconceivable that I discover a technique that I had not previously known to exist to apply to the data.
Therefore I have tried to organise my dissertation into some sort of timescale:
This I hope to be able to do before the end of the 3rd term 97/98 and with in the first few weeks of the first term 98/99.
I plan to complete this by the end of the Christmas term (which is about 10 weeks long). Obviously
the amount of time needed is still unknown , but I plan to be able to do this by then.
This will include the construction of the GIS using ArcInfo and the building of the DEM and production of the images. When this is complete, I will then be able to start building a layer cake.
This will consist of the analysis and commenting the results obtained so far. I hope to have this finished towards the end of the second term.
I will attempt this if there is enough time after the completion of the upper tasks. Enough time, that is to do the task justice, and to obtain any kind of decent results.
Finally, it should be taken into consideration that the Hayton project is still on-going, and therefore there is a good chance that if any new data was to be collected this summer, it could alter or add to any of the above mentioned aims.
Bibliography
Allen K.M.S., Green S.W., Zubrow E.B.W..(1990) Interpreting space : GIS and archaeology, London : Taylor & Francis
Halkon P. & Millet M., (1997) The Landscape Archaeology of Hayton, East Yorkshire, An Interim Report, Hull, The University of Hull
Kamermans H. & Fennema K. (ed), (1996), Interfacing the past : computer applications and quantitative methods in archaeology, Analecta praehistorica Leidensia ; 28., Leiden : University of Leiden
Madry, S. et al, GIS and remote sensing for archaeology: Burgundy France , at http://deathstar.rutgers.edu/projects/france/france.html
Murnion, S., Arc/Info tutorial page, at http://boris.qub.ac.uk/shane/arc/ARChome.html