Investigating Sedimentological Architecture Using Electrical Resistivity Tomography: A Case Study from the Archaeological Open-Air Site Shën Mitri, Southern Albania

Abstract

Large-scale and small-scale high-resolution electrical resistivity tomography (ERT) measurements were carried out to investigate the sedimentological architecture of the Shen Mitri hill and its close surroundings near the Butrint archaeological site in Albania. The ERT measurements were conducted in combination with archaeological excavations of an Upper Paleolithic/Mesolithic open-air site with the aim to better understand the early modern human migration history. The resulting inversion models are meaningful and span a wide range of subsurface resistivities. The Shen Mitri hill is mainly composed of low-resistive Pliocene clayey sediments less than 5 $$\Omega$$ m and is covered by resistive sandstone towards the hill-top reaching values more than 100 $$\Omega$$ m. The Vrina plain surrounding the hill consists of sediments of Quaternary origin with rather low resistivities. High-resistive features in the Vrina plain correspond to Pliocene units of sandstone and outcrops of evaporates with several hundred $$\Omega$$ m of resistivity. The results revealed that the neighboring hill Xarre is likely to be of older age than the hill Shen Mitri as it is composed of sediments from an older Pliocene unit with higher resistivities. The results improve the knowledge of the actual geological understanding of the hills and give detailed insight to their architecture, such as thickness and distribution of the occurring formations. Thus, not the main focus of the article, small-scale ERT results reveal zones with a high-resistive surface layer which is identified as a archaeological layer-containing stone artifacts and ceramic sherds. However, undisturbed archaeological layers were not identified yet and ERT measurements can help to clarify the taphonomic history of the Shen Mitri site. Prior to the inversion, an elaborate estimation of ERT data errors including geometrical errors and the data variation coefficient is presented. The estimated error model is meaningful and included in the two-dimensional error-weighted inversion scheme. Subsequently, the effect of surface topography on ERT data is estimated by a modeling study and, as a consequence, included in all the inversion models.