Trans-regional Agricultural Deintensification: An AI-Assisted Survey of Agricultural Infrastructure in the South-Central Andes

Summary

This is an abstract from the "SAA 2023: Individual Abstracts" session, at the 88th annual meeting of the Society for American Archaeology.

Since late prehispanic times, peoples throughout the central Andean highlands have created highly productive anthropogenic agricultural landscapes on a monumental scale through terracing. Yet a large proportion of these terrace systems fell into disrepair and abandonment through the Spanish colonial period, even in the face of food shortages. The mechanisms and patterns of such agricultural deintensification remain poorly understood, in large measure because we lack a trans-regional view of the extent and distribution of terrace abandonment. Field-based pedestrian survey cannot capture trans-regional distributions of agricultural deintensification, while manual digitization of terraced areas via imagery survey is slow and labor intensive. This research develops and deploys convolutional neural network-based image segmentation of high-resolution satellite imagery in the South-Central Andes to map the distribution of active and abandoned terracing. This AI-assisted image segmentation renders a continuous distributional view of terrace infrastructure in an area in excess of 100,000 km2. Given the innovative approach to large-scale imagery survey, we evaluate the data for quality and potential biases before using multivariate spatial statistics to identify factors driving patterns of agricultural deintensification.

This Resource is Part of the Following Collections

• SAA 2023: Individual Abstracts   •
• Society for American Archaeology 88th Annual Meeting, Portland, OR (2023)

Keywords

Spatial Coverage

Record Identifiers

Abstract Id(s): 37381.0