Magnetic modelling and error assessment in archaeological geophysics: The case study of Urbs Salvia, central Italy

We perform an analysis of the errors that affect magnetic anomaly data in archaeological geophysics, arising from both survey time procedures and common potential fields methods of magnetic data processing. Specifically, there are errors due to: 1. positioning of total field readings, 2. the estimated diurnal drift of the Earth’s magnetic field, 3. the selected gridding algorithm, 4. the process of reduction of total field data to magnetic anomalies, 5. the application of decorrugation filters, and 6. knitting of two or more survey rectangles within the same archaeological area. Our analysis shows that in normal conditions these errors can have a magnitude up to few tens of nT and a lower limit exists to the amplitude of the anomalies that can be interpreted archaeologically. A correct error assessment is especially required when the magnetic anomalies must be interpreted quantitatively through a forward modelling procedure. We illustrate an application of these concepts to a magnetic data set acquired at the Roman settlement of Urbs Salvia (Central Italy). We show that forward modelling provides a powerful tool for the reconstruction of ancient buried settlements.