Hello, my name is Lindsey Cochran from the University of Tennessee. I’m also with the NCPTT. I was a summer intern in 2014 for Tad Britt in Archeology and Collections. My project this summer was to create predictive modeling for the DEVA Project or Death Valley National Park. My goal was to identify and then eliminate statistical bias in these human habitational modeling patterns. So, what I was working on was expanding and refining an existing research project to look… we had the idea that humans are patterned according to their environment, so we try to exploit that through maps and statistics.
So we have four different layers based on grid based GIS systems for mapping. We have 13 environmental variables like slope, aspect, horizontal and vertical distance to water, and then we layer that on top of the states site files, which has over 2000 sites from over the last 50 or so years. Finally, we plug that into a program called MaxEnt for Maxim Entropy niche modeling, and then look at the statistical probabilities that a certain site would be located in certain environmental areas. We also created a best practices manual on how to use this software and how to take advantage of all the data sources that are available, and only used open-source software so that everything is easily and openly accessible to everyone.
A few of the other things that we’ve done so far involve the kinds of data that we’re putting into the models, like integrating climatological information, and then bringing in lots of different post-analysis statistical models. So overall, this research is significant one in that it helps to identify high probability areas for cultural resources, which is immensely useful for preservationists, archaeologists, and park and resource management professionals. Secondly, we’re hoping to identify methods that can be easily, cheaply, and efficiently recreated and reproduced, helping to integrate predictive modeling into more professional settings.
Thanks, and happy archaeology day!