Modeling Prehistoric Sites on MSU’s Campus

As the campus archaeologists, one of our primary goals is to protect the cultural and historical heritage of the Michigan State University campus. In order to do this, we work with a number of departments to ensure that construction and other campus projects do not damage or disturb anything of historic or cultural value. It is likely that you’ve even seen us working away on the campus grounds, digging test pits, screening soil for artifacts, or surveying the ground. The relationship that we have with the campus depends on our communication with them regarding areas of archaeological sensitivity. In order to better facilitate this communication, one of my projects has been to develop a geographic information systems (GIS) map that analyzes the campus and allows us to create predictive models regarding which locations are more likely to have pre-historic materials.

When looking for areas of historic archaeological significance there are a number of archival resources that can help us determine where buildings used to be or where activity was focused. The campus has an extensive archival system and its historic past is highly documented. However, if we want to understand the history of campus before the university, colonization or writing there is a lack of information. To determine which areas were inhabited by the indigenous peoples of Michigan prior to colonization, we need to use non-archival methods.

In order to determine which areas are more likely to contain prehistoric archaeological sites, it is important to look at the variables that constrain habitation as well as those which are beneficial. Research into settlement patterns has shown that there is a number of variables which determine where people choose to live. By creating hypotheses as to what variables may have determined settlement location, we can create a model to predict where probable areas of prehistoric human activity may be located (Kohler 1988:33). Within GIS, each of these hypotheses is made into a layer of variables which can be spatially and statistically compared against the other layers. Using this program we can analyze data based on the associations between multiple variables.

In order to create this model, first it is necessary to look at what information is already known about the area and prehistoric cultures to determine which variables are important to their settlement patterns. Little is known about whether there were any indigenous groups directly in the campus area. However, there are reports of a small settlement where the football stadium currently resides, and of Chief Okemos and the Ottawa (Widder 2005). The second stage in actually creating the model is to determine what factors on the landscape will determine human activity. These have been selected based upon previous GIS studies of settlements, as well as historical knowledge of settlement patterns. For this project it was determined that settlements and human activity were most likely to occur within 300 meters of water, within a slope less than 5% steepness, near edible vegetation. Prior archaeological surveys were also included to determine which areas had already been identified as sensitive and which had been clear.

Next, we begin the modeling process in GIS, using ESRI’s ArcView. In order to do this, the various layers were downloaded from Ingham County including maps of the rivers, vegetation, and a digital elevation model. To determine proximity to water, a 100 meter and 300 meter buffer were put on the river with the former being highly sensitive and the latter moderate sensitive. Vegetation was classed into edible and non-edible sources from an 1800’s plan of the area. Slope was created through the DEM, and was divided into 0-5% rise for high sensitivity, 5-10% rise for moderate, and anything above was removed. Previous work was mapped out by hand using information from the Office of the State Archaeologist of Michigan.

The result of our analysis is a map which shows areas of high, moderate and low archaeological sensitivity. Based on the map, there are a number of locations which require further investigation and can be addressed by campus archaeology in the future.  As the map shows, surrounding the river there is a large area which is potentially highly sensitive for finding archaeological material. This is not unexpected given that we have a fresh water source (Red Cedar River) running through campus, the land is relatively flat, and most of the vegetation is edible. The two previously identified sensitive areas include the finding of two projectile points around Beaumont Tower and a reported native camp site at the southern bend of the river, both fall within the area of projected sensitivity. Future updates of this map will include the campus archaeology surveys in order to see which of these areas have been investigated and which have not.

There are some limitations to this type of modeling. We are biased towards environmental determinism (that environmental factors are the sole reason for habitation and activity location) due primarily to the lack of incorporation of cultural variables into the model (Reddy and Brewster 1999:7). The only cultural information that we can include at this moment are previous pre-historic finds. As we conduct more archaeology on campus, we can build a better interpretation of what pre-historic cultures may have lived here and what cultural constraints shaped their activity and settlement patterns. GIS predictive modeling provides archaeologists with a way to locate areas that are more likely to contain sensitive cultural material, and allows us to focus our efforts.

This map is part of an ongoing process to use GIS in order to predict archaeological site locations on campus. It is only the beginning of the process, and as we continue to add variables our resulting maps will become more accurate.

Works Cited

Dalla Boona, Luke

1994 Archaeological Predictive Modelling in Ontario’s Forests. Report for the Ontario Ministry of Natural Resources. Electronic Document. http://modelling.pictographics.com/homepage.spml

Kohler, Timothy A.

1988 Predictive Locational Modeling: History and Current Practice. In Quantifying the present andpredicting the past: Theory, method, and application of archaeological predictive modeling, edited by W.J. Judge and L. Sebastian, U.S. Government Printing Office, Washington, D.C.

Reddy, Seetha and Alice Brewster

1999 Applying GIS to Archaeological Site Prediction on Camp Pendleton, Southern California. In Pacific Coast Archaeological Society Quarterly 35(1): 7-18.

Widder, Keith R.

2005 Michigan Agricultural College. Michigan State University Press: East Lansing.