Date of Award
6-2014
Degree Name
Master of Arts
Department
Geography
First Advisor
Dr. Charles Emerson
Second Advisor
Dr. Robert Anemone
Third Advisor
Dr. Kathleen Baker
Keywords
Paleontology, remote sensing, GEOBIA, predictive model, Wyoming
Access Setting
Masters Thesis-Open Access
Abstract
The development and testing of predictive models for identifying productive fossil localities represents a promising interdisciplinary endeavor among geographic information scientists, paleoanthropologists, and vertebrate paleontologists. This thesis analyzed high resolution (2m spatial resolution) commercial satellite imagery from the Worldview-2 satellite of five areas of the Great Divide Basin using a GEographic Object-Based Image Analysis (GEOBIA) technique, which segments the image into spectrally homogeneous, multi-pixel image objects. In addition to allowing statistical analysis of the spectral characteristics of the image objects, GEOBIA techniques also let analysts incorporate expert knowledge and contextual information to improve classification accuracy. The spectral characteristics of the image objects that represent a highly productive sandstone locality (Tim’s Confession, WMU-VP-220) were used to create a classification scheme to pinpoint areas in the Great Divide Basin that were highly likely of being fossiliferous. During the summer field season of 2013, thirty one locations across five satellite images with clusters of predicted fossil localities were surveyed. At thirteen of these locations fossils were recovered, leading to the documentation of twenty five new fossil localities.
Recommended Citation
Bommersbach, Bryan, "Predictive Modeling in the Search for Vertebrate Fossils: Geographic Object Based Image Analysis (GEOBIA) in the Eocene of Wyoming" (2014). Masters Theses. 500.
https://scholarworks.wmich.edu/masters_theses/500
Included in
Geographic Information Sciences Commons, Paleontology Commons, Physical and Environmental Geography Commons