Detecting historic tar kilns and tar production sites using high-resolution, aerial LiDAR-derived digital elevation models: Introducing the Tar Kiln Feature Detection workflow (TKFD) using open-access R and FIJI software. (February 2022)

• Record Type: Journal Article
• Authors: Snitker, Grant; Moser, Jason D.; Southerlin, Bobby; Stewart, Christina
• Is Part Of: Journal of archaeological science. Volume 41(2022)
• Publication Date: 2022-02
• Subjects: Tar kilns -- Naval stores industry -- Historical archaeology -- LiDAR -- Feature detection -- Ecological restoration -- Longleaf pine ecosystems; Archaeology -- Periodicals; Archaeology -- Research -- Periodicals; 930.1
• DOI: 10.1016/j.jasrep.2022.103340
• Languages: English
• ISSNs: 2352-409X

Abstract

Highlights: Historical reference conditions have both social and ecological drivers ·The TKFD workflow is an open-access, scripted, and replicable process Over 2, 700 tar kilns are identified within the Francis Marion National Forest Dataset informs the distribution of historic longleaf pines and disturbance regimes Abstract: Significant declines in longleaf pine ecosystems during the last 150 years have motivated research and conservation programs focused on their restoration to perceived historical reference conditions. However, the ecological impacts of the historical naval stores industry on forest structure, fire regimes, and fuel conditions are not currently considered in reference conditions constructed from 18th, 19th, and 20th century sources. We present the Tar Kiln Feature Detection workflow (TKFD), an open-access, scripted, and replicable process developed in R and FIJI to identify archaeological tar kilns within high-resolution digital elevation models derived from aerial LiDAR datasets. The workflow is developed and validated on the entirety of the Francis Marion National Forest in coastal South Carolina. The TKFD has identified and measured over 2, 700 tar kilns within our 420, 000-acre study area and validation studies demonstrate a balanced identification accuracy of 90.6%. This is the most comprehensive dataset of tar production sites in North America and has implications for understanding the historical distribution of longleaf pine stands, anthropogenic impacts on fire and fuels, and the nature of these unique archaeological sites.