Finding a Relatively Flat Archaeological Site with Minimal Ceramics: A Case Study from Iraqi Kurdistan

ABSTRACT

We present a new technique to detect flat archaeological sites with minimal ceramics using an unmanned aerial vehicle that maps surface stone concentrations. Methods deployed include point pattern analysis of stone concentrations and a machine-learning technique using unsupervised classification of visible stone signature qualities, which are used in simple linear regressions to compare with geophysical and ceramic surface survey results of a site in Iraqi Kurdistan. There is a stronger fit (r²= 0.77) between surface stone concentrations and architecture identified by geophysical measurement, while surveyed ceramics show a weaker fit to defined architecture (r²= 0.31). Surface stone concentrations are potentially a better proxy than ceramics for determining the presence of past settlement in regions where stone was commonly used, sites are relatively flat, and ceramics are found in low concentrations. The methods advanced here can be scaled to wider areas, particularly in mountainous regions, where surface stone features are present.

KEYWORDS:

• Remote sensing
• site detection
• geophysics
• stone concentration
• pottery survey
• image classification
• unmanned aerial vehicles (UAVs)
• machine-learning

Acknowledgments

This paper originated from research undertaken within the Alexander von Humboldt Foundation and the Peshdar Plain Project directed by Karen Radner (LMU Munich, Alexander von Humboldt Professorship of the Ancient History of the Near and Middle East). A first draft was written while Mark Altaweel held a fellowship at the Center for Advanced Studies (CAS) at LMU Munich as part of the Research Focus group “Settlements between Diversity and Homogeneity” led by Radner. We thank both the Alexander von Humboldt Professorship and the CAS for providing generous support and assistance for this research. We thank the Sulaymaniyah Directorate of Antiquities and Heritage for permission and assistance in making this work possible.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Notes on Contributors

Mark Altaweel (Ph.D. 2004, University of Chicago) is a Reader in Near Eastern Archaeology at the Institute of Archaeology at University College London. He is interested in developing computational and information science methods to enhance fieldwork methods and archaeological theory.

Andrea Squitieri (Ph.D. 2015, University College London) is a postdoctoral researcher at the Ludwig-Maximilans University of Munich (LMU) and field director of the Peshdar Plain Project. His research interests include stone tool technology, archaeology of empires, and statistical methods applied to archaeology.

The authors contributed equally to this work.

ORCID

Mark Altaweel http://orcid.org/0000-0003-1807-3737