Angelica Capozzoli

This study proposes a procedure to improve the interpretation of data from the Frequency Domain ElectroMagnetic method (FDEM), a geophysical technique with high benefit–cost ratios in archaeology. This method enables the simultaneous analysis of electrical and magnetic properties of the investigated medium, providing data as in‐phase and out‐of‐phase (quadrature) components of the electromagnetic field. Traditionally, FDEM produces individual maps for these components that are typically inspected and related to each other only visually. The proposed procedure is based on unsupervised machine learning that, by providing integrated maps of these two components, overcomes the limitations of visual inspection and streamlines the interpretative phase. The FDEM data acquired at the Torre Galli archaeological site in Calabria, Italy, were clustered using K ‐means, known for its effectiveness in partitioning densely distributed spatial data. The optimal number of clusters was determined through silhouette analysis. Once identified, the clusters were classified as either representative of the natural background or indicative of archaeological interest and mapped in real space. This analysis identified several areas with anomalous geophysical characteristics compared to the natural context, characterized by distinct ranges of EM property values, suggesting the presence of different archaeological targets, such as road, walls and iron/bronze tools, which are distributed heterogeneously in the subsoil. This classification facilitated the rapid delineation of areas with significant archaeological potential that may warrant further investigation, resulting in a reduction of both the study area size and the archaeological effort associated with excavation. A comparison of these areas with previous excavations and the results of a magnetic survey shows that the proposed procedure is promising in enhancing the efficiency and accuracy of the FDEM method for identifying areas of archaeological interest. The findings suggest a move towards automation in the interpretation process, which could improve cost‐effectiveness and time optimization in geophysical and archaeological investigations.

The 1688 Sannio–Matese earthquake, with a macroseismically derived magnitude of M_w = 7 and an epicentral intensity of I_MCS = XI, had a deep impact on Southern Italy, causing thousands of casualties, extensive damage and significant environmental effects (EEEs) in the epicentral area. Despite a comprehensive knowledge of its economic and social impacts, information regarding the earthquake’s environmental effects remains poorly studied and far from complete, hindering accurate intensity calculations by the Environmental Seismic Intensity Scale (ESI-07). This study aims to address this knowledge gap by compiling a thorough dataset of the EEEs induced by the earthquake. By consulting over one hundred historical, geological and scientific reports, we have collected and classified, using the ESI-07 scale, its primary and secondary EEEs, most of which were previously undocumented in the literature. We verified the historical sources regarding some of these effects through reconnaissance field mapping. Analysis of the obtained dataset reveals some primary effects (surface faulting) and extensive secondary effects, such as slope movements, ground cracks, hydrological anomalies, liquefaction and gas exhalation, which affected numerous towns. These findings enabled us to reassess the Sannio earthquake intensity, considering its environmental impact and comparing traditional macroseismic scales with the ESI-07. Our analysis allowed us to provide an epicentral intensity ESI of I = X, one degree lower than the published I_MCS = XI. This study highlights the importance of combining traditional scales with the ESI-07 for more accurate hazard assessments. The macroseismic revision provides valuable insights for seismic hazard evaluation and land-use planning in the Sannio–Matese region, especially considering the distribution of the secondary effects.