GNGTS 2017 - 36° Convegno Nazionale

GNGTS 2017 S essione 3.3 729 the observed magnetic signal. At each iteration, the NRM parameters and eventually the depth and size of the object are adjusted to progressively minimize the mismatch between the model and observed anomalies along the profiles. The final result is not necessarily what we could find by direct excavation, because of the intrinsic ambiguity of potential field data. However, the availability of archaeological information can help to constrain materials and depths of the model objects, thereby allowing a realistic reconstruction of a buried settlement. Discussion. The approach presented above allows to create realistic magnetization models of archaeological sites even in the case of complex topography, granted that appropriate acquisition and processing of total field data have been performed. In ArchaeoMag , the observed and model grid anomalies are automatically assigned an orthometric height according to an input digital terrain model for the survey area. Therefore, any object in the model acquires local Cartesian coordinates depending from the burial depth specified at the time of its definition as well as from its UTM coordinates. Thus, it is possible to obtain an automatic terrain correction that accounts for the anomaly field distortion associated with topography. An example of application of ArchaoMag to a situation characterized by rugged topography is illustrated in Fig. 2A. These data were acquired in 2015 along the SE slope of the Jermë hill, southern Albania (Schettino et al. , 2017). Apparently, a segment of the western branch of the two positive anomaly stripes could be modelled by a N–S oriented rectangular prism, as illustrated in Fig. 2. This interpretation is partially supported by the E–W profile shown in Fig. 2C. However, the N–S profile (Fig. 2D) shows a northward increase of the observed anomalies, whereas the rectangular prism model predicts the opposite, according to the fact that the southern tip of this structure is closer to the surface than the northern end. As a consequence, the observed anomalies cannot be generated by an object having a flat upper surface. In Fig. 3 an alternative model is proposed, which is based on a stairway structure formed by 12 rectangular prisms. In this instance, the burial depth of each step slightly rises downslope, accounting for the increased rate of accumulation in this direction. Undoubtedly, this model provides a much better fit of the model anomalies to the observed values in N–S direction, as shown in Fig. 3C. As mentioned above, the possibility to model NRM components in addition to induced Fig. 3 - An alternative stairway model of the same anomalies considered in Fig. 2A. See text for discussion.