I'm writing here to get your thoughts on the methodology of a simple GIS project that I hope to have form a chapter in my (history) doctoral dissertation. Most of what I know about historical GIS is coming from a range of articles in the Journal of Archaeological Science and is there applied to pre-literate or pre-historical cultures, but I'm attempting to bring some of their methodology into the study of the near east in the Middle Ages. So I'm looking at a contested border zone, a sort of no-man's land. I built myself a cost surface from a DEM, hydrological features, and some ground cover data. In my area of study I have about a dozen points indicating settlements or fortresses. So I ran a series of least-cost paths, from every point to every other point.
We have a historical text pertaining to this time and place which discusses the sort of low-intensity warfare that was conducted. The idea behind generating the LCPs was to provide an alternative to our poor knowledge of the roads in the area, as well as to get a sense of where else people might be moving when the roads were being watched. Ambush and tracking of raiding parties was a big part of warfare in this region. So what I'm doing is taking the polylines generated by the LCPs, turning them into rasters, and then performing viewsheds along the entire length. Rather than identifying what can be seen, I'm interested in what the viewer on these paths cannot see, ie: potential locations where one might hide ambushing forces, and how those defending might have been able to move around the landscape without being detected or intercepted.
So I have a couple of questions that those with more GIS knowledge might be able to help me with.
1. Does anyone know of any similar, published work? As a GIS novice, all my methodologies are borrowed from others.
2. I have been using the line rasters to create the viewsheds, but am I correct in believing that the viewshed tool is taking vertices along the line as the points in which it makes the viewshed from? For example, in this image the green line is the original cost path, the purple is the raster line made from the cost path, and the circle un-seeable zone is probably there because it's only taking viewsheds from the two vertices rather than any points along the line?
3. If the answer to number 2 is yes, I should probably manually create a series of points along the line and then use Modelbuilder to run many viewsheds automatically?
4. Are there better tools or methods to go about doing this project?
Thank you for your time.