Archaeological site identification from open access multispectral imagery: Cloud computing applications in Northern Kurdistan (Iraq)

Abstract

This paper presents the results of an archaeological survey carried out in the Navkur Plain, Iraqi Kurdistan, as part of the ‘Asingeran Archaeological Project’. The survey was prepared using remote sensing products accessed via Google Earth Engine^ⓒ, a large-scale cloud computing service freely available to the scientific community that allows processing remote sensing big data. Outputs generated with a multitemporal approach are particularly successful for archaeological research, because it is possible to maximize the visibility of archaeological sites, improving their detection. Multispectral imagery from Landsat 5, Landsat 7 and Sentinel-2 collections were used and processed, testing their utility for finding unknown ancient settlements in the densely studied area of Northern Mesopotamia. Seventeen new sites were discovered in an already surveyed area of limited size (<100 km²), showing the potentialities of this method. The advantages of cloud computing for Near Eastern Archaeology and the results of the survey are also presented and discussed.