Thursday, December 13, 2018

GIS 5990 - Special Topics in GIS Archaeology - Module 7


Over the past several weeks, I have worked on a project which involved a study on detecting Scythian burial mounds, in the general area of Tuekta, Russia. The ultimate goal of the study was to generate a raster file, containing a random distribution of points which would express the possibility of presence of a Scythian burial mound, based on previously known information.

The metrics considered for a possible presence of a burial mound in the immediate area of the points were: slope, aspect and elevation. All of these layers were created, and generated in ArcMap based on DEMs downloaded from the web. The study itself was confined to a general greater, area of Tuekta, Russia, and the random points of the predictive model were merged with previously marked locations of known burial mounds, in order to have the spatial data of the known locations influence the appearance of similarly-set points.

The results of the study are represented based on an OLS model, which was created after the merging of the point files. With zero being an average value, the coloring of the points will tell us not only how compatible (or incompatible) the location of the given point is with a possible burial mound, but also express the degree of confidence in the result based on the data provided. The spatial autocorrolation test that was ran on the data showed the z-score (indication of the normal distribution of data) of 14.36 and the p-value (the likelihood that the data is not randomly distributed) at 0.0.

There are factors in the predictive model, which were not considered for this study but may be of importance to its overall accuracy. These factors may include the proximity to water, the landscaping of the landmass, proximity to other sites, and maybe even the overall space around the point. (The known burial mound locations all occur close to one another, in a very large cluster.) The model does however provide locations of potential locations based on the exact coordinates of each point. Each positive hit could benefit from a field survey, with all other factors neutral.

No comments:

Post a Comment