The goal of analyzing the spatial relationship between Roman marching camps and the features of the areas they are placed in would be hard to attain without the use of a Geographic Information System (GIS). Using GIS allows analysts to combine separate layers of geographic data and analyze the relationships between them.
The GIS used for this project was Esri Software’s ArcGIS — more specifically, ArcGIS Desktop 10.4.1 for the analysis and heavy lifting, and ArcGIS Online for easy display via the Web. ArcGIS is known for its complexity almost as much as its power, but on the advice of more experienced GIS users, ArcGIS was chosen to ensure that the tools would have all of the needed features.
Connecting the various pieces of information needed to study the positioning of the camps required pulling information from a variety of data sets. This data breaks into two main pieces — the position data of Roman remains in Scotland, including the camps themselves and Roman roads, and the data concerning the elevation, water features, and land boundaries of Scotland.
Perhaps the most important data set — the locations of temporary Roman camps — was provided by Canmore, an online database of important archaeological sites and buildings throughout Scotland. I cross referenced the sites of camps with the camps listed in Roman Camps In Scotland, a book written by Rebecca Jones that includes a gazetteer of Roman temporary camps throughout Scotland1.
The camp locations in Canmore are broken down into “temporary camps” and “possible temporary camps”, and in Jones they are broken down further into “known”, “probable”, and “possible” camps. I removed the “possible” or “probable” camps in both cases– even if they were listed differently between the sources — as well as any camps that appeared in one but not the other. This may have removed some camps that have been been verified in more recent academic works, but I felt that this was better than admitting camps I was unsure about into the data, and individually determining the mood of scholars for every conflict between the two sources would have taken an amount of time beyond the scope of this project.
The shape files (a default file type used by ArcGIS to represent data) for the Roman road network came from Harvard University’s Digital Atlas of Roman and Medieval Civilizations (DARMC). DARMC provides a great deal of other data sets relating to the Roman civilization, and the maps they have created and made available on their website are certainly worth looking at.
As for the data concerning Scotland itself, I owe a great deal of thanks to the British Ordnance Survey (OS) for providing much of the data I needed. OS Boundary-Line provided shape files for various administrative districts in Britain — I sifted through these and eventually decided on a map of the entire island divided by European electoral regions as my base map to build off of. This particular shape file was great for this application because it both provided enough geographic context for the island of Great Britain and did not clutter the map with unnecessary low-level information (as Scotland is an undivided electoral region). Another OS product — OS Open Rivers — provided shape files for Scotland’s waterways.
I initially thought that the final piece — a digital elevation model — would come also come from the OS in the form of OS Terrain 50. However, while this product is highly detailed, the shape files it provides are broken down by National Grid Reference squares — the individual squares would have needed to be pieced together, and I was advised against doing this. Instead, I used the digital elevation model (DEM) created by NASA’s Shuttle Radar Topography Mission (SRTM). This DEM was better for my purposes, as there were only two files that had to be loaded into ArcGIS as opposed to dozens in OS Terrain 50.
One issue with these models of the various aspects of Scotland is that this data is modern, and as a result there are some things that may have changed in the years since the Roman camps were placed. Although the physical features of an area change slowly in ways that are sometimes nearly imperceptible, they do change, and as a result it is expected that there will be some difference between the current distance between a camp and a land feature and the actual distance at the time the camp was placed. However, to my knowledge there is no comprehensive elevation and water feature data from the time of the Roman empire, and I decided to make do with our modern equivalents.
- Jones, Rebecca H., Roman Camps In Scotland (Edinburgh: Society of Antiquaries of Scotland, 2011, Print).