NMC Scientist Claudine Gravel-Miguel Showcases Innovative Archaeological Mapping at SAA 2025
New Mexico Consortium (NMC) is pleased to highlight the work of Dr. Claudine Gravel-Miguel, who presented her latest research at the 2025 Society for American Archaeology (SAA) Annual Meeting. Her presentation, titled “A Multigenerational Workflow: Applying Deep Learning Tools on Old Maps to Detect Near-Invisible Historic Sites,” explores how artificial intelligence can be applied to historical data to uncover long-buried archaeological sites.
Dr. Gravel-Miguel’s work addresses a fundamental challenge in archaeology: many sites have become invisible due to erosion, sedimentation, or human alteration. While modern tools like satellite imagery and LiDAR help, they can miss deeply buried or modified sites. Her approach turns to a well-known Cultural Resource Management resource—19th-century General Land Office (GLO) survey maps.
These maps, though rich with historical details like homesites and roads, can be difficult to use in modern GIS systems due to their analog format. To overcome this, her team georectified maps from the Kisatchie National Forest and trained deep learning models to identify features such as roads, fields, and structures.
Using a U-Net convolutional neural network, the models performed well in detecting roads, moderately well with fields, and struggled with structures—largely due to variations in appearance and limited training data. Dr. Gravel-Miguel plans to improve model performance by standardizing map preprocessing and expanding annotations.
Ultimately, this workflow provides a powerful new tool for cultural resource managers, helping to locate and protect historic sites using legacy data enhanced by modern technology.
The NMC congratulates Dr. Gravel-Miguel on her impactful research and her role in advancing interdisciplinary approaches to archaeological discovery.
