GSA Annual Meeting in Phoenix, Arizona, USA - 2019

Paper No. 260-11
Presentation Time: 9:00 AM-6:30 PM

PENTELIC MARBLE PROVENANCE USING X-RAY FLUORESCENCE SPECTROSCOPY AND K-MEANS CLUSTERING


SHINSATO, Lara M., School of Geographical Sciences and Urban Planning, Arizona State University, Coor Hall, 5th floor, 975 S. Myrtle Ave., Tempe, AZ 85287, PIKE, Scott, Environmental Science Department and Archaeology Program, Willamette University, 900 State Street, Salem, OR 97301 and SMITH, Jeff, Environmental Science Department, Willamette University, 900 State St, Salem, OR 97301

The quarries on the south slope of Mt. Pentelikon in Attica, Greece were a primary source of white marble for ancient Greek and Roman buildings, sculptures, and monuments. The use of Pentelic marble can be traced back to the beginning of the 5th Century B.C.E. where it was first used on a large-scale for the construction of monuments atop the Athenian acropolis including the current Parthenon and its sculptural program as well as the Athena Nike Temple. The Romans used the marble for many sculptures and monuments including the facing of the porch in the Pantheon in Rome. To gain a better understanding of the trade, use, and preference of Pentelic marble in antiquity, archaeological scientists developed a stable isotope database of carbon and oxygen (δ13C and δ18O) to not only distinguish Pentelic marble from other white marble sources in the Mediterranean Basin (interquarry variation), but to also distinguish between individual quarry pits and quarry areas within the Pentelic quarry region (intraquarry variation). While previous stable isotope studies have been able to distinguish between major white marble sources, there is still a lack of significant structure within the isotope data for intraquarry site discrimination. This study aims to address this issue by combining handheld-XRF geochemical analysis with data mining techniques to differentiate amongst known ancient Pentelic quarries. By analyzing Pentelic marble geochemical data using a K-Means clustering algorithm, this study identifies clusters within the data that can be used to better classify and identify variables for future marble provenance studies.