Remote Sensing in Archaeology: Monk's Mound, Cahokia

Cahokia Mounds
     The Cahokia Mounds are all that remain of a once thriving city in North America located in Collinsville, Illinois. Inhabited between 600 - 1400AD (during the late Woodland/early Mississippian periods), the site at its peak boasted a population of 10,000 to 20,000 people. It originally consisted of more than 100 earthworks and mounds but only about 80 currently remain.
     The cause for the site's eventual abandonment is unknown but it was in decline around a century before Europeans arrived. What was once the largest prehistoric Native American settlement north of Mexico is now situated amidst modern development.
     The mounds are designated a National Historic Landmark and a UNESCO World Heritage site. Monk's Mound is the largest mound in the complex. For this assignment, we performed two different types of raster classification in ArcMap. We visualized Monk's Mound after these classifications to see how they are classified.

Unsupervised Classification
Below is the map end product generated from running an unsupervised classification in ArcMap. There is a fair amount of confusion between the classes (for this assignment we used 8) with vegetation being represented by 5 of the 8 classes. Monk's Mound is classified as urban or bare earth.

A map of an unsupervised classification performed on raster data
acquired from USGS for the Cahokia site in Illinois. 

Supervised Classification
Here is the map end product of running a supervised classification for the same image. Instead of telling the program to derive the classes, samples were used to create a signature file. These signatures were in turn used to classify the raster. In this classification, I attempted to use 6 classes to separate the raster pixels. There is plenty of error in that dark features (like water and shadowed trees) are lumped together. Trees are not captured well either and it the classification makes it seem like there are less than there really are. Here, like the previous unsupervised classification, Monk's Mound is classified as an urban/road feature again.

A supervised classification performed on raster data
acquired from USGS for the Cahokia site in Illinois. 

3D Modeling of Shovel Test Data

     This week's assignment focused on learning how to manipulate archaeological data in ArcScene (a 3D environment). To prepare data for ArcScene, shovel test data was examined in ArcMap. Geological surfaces were rendered/interpolated using the IDW (inverse distance weighted tool). These surface were used to make the map you see below. The shovel test data was also imported into ArcScene and we used the Base Height tab and the Extrusion tab to project the 2D data into 3D space.

A simple map showing three interpolated surfaces derived from shovel test data. 

     Using the Fly tool (link) and Animation tool (link) I created a video that travels through the shovel tests. While not essential to understanding this data it is a neat way to visualize and interact with it. (Bonus: these types of visuals can also be used to enhance a presentation.)

     This assignment brought together several of the skills we have learned throughout the semester and helped prepare us for the final project. As a continuation of this theme, our discussion post focused on the application of 3D technology in archaeology. I mention this because you may have fun exploring some of the sites I found, CyArk, MayaArch3D, a Roman statue base, and other tech summary sources. Try to contain your excitement :).