GNGTS 2016 - Atti del 35° Convegno Nazionale

506 GNGTS 2016 S essione 3.1 Multiscale analysis. The multiscale analysis consists in two main steps: • ������ ������������ �� ��� ������� �� ������� �������� �� ����� ��� ��������� �������������� upward continuation of the dataset at several altitude by using the classical level-to-level upward continuation formula: (1) • ������ �� ��� ���������� ��� ���� ������� �� �������� ��� ������ ������ employ of the Multiridge and DEXP methods to localize the source depth. Multiscale methods will allow the analysis of the whole crust at different scales, and so they may characterize sources at different depths in the crust, without using band-pass or similar filtering. This follows from being multiscale algorithms entirely built with physically based tools, such as upward continuation and differentiation. The source depth and shape are extracted by the simultaneous analysis of the field at multiple scales. The Multiridge method (Fedi et al. , 2009) is based on the search of the ‘extreme points’ of the field and its horizontal and vertical derivatives at several altitudes along a profile. For extreme points we intend the maxima of the field and of its horizontal and vertical derivatives at several altitudes. By joining these extreme points along the scales, we form the so-called ‘ridges’ in the harmonic region. The ridges may be prolonged down to the source region, using the geometric approach (Fedi et al. , 2009), as long as they intersect one each other at the field singular points, that is at the depth position of the sources generating the anomaly field. Important features of this method are: a) its high stability and low computation time; b) it does not need a priori information of source: it is independent of the directions of the magnetic induced field. Moreover, Euler deconvolution may be applied along such ridges in order to estimate the homogeneity degree and, consequently, retrieve information about the kind of source (Florio and Fedi, 2006; Fedi et al. , 2009; Milano et al. , 2016). The DEXPmethod allows an estimate of the excess dipole moment intensity for the magnetic field, readily found from the value of the scaled potential field at the extreme points (Fedi, 2007). The final result will then be a sort of image of the source distribution, which may be tuned at several resolutions. It enjoys a great stability versus high-order differentiations of the field. This stability is a physical property based on the regular behavior of potential fields versus the altitude z and it is shared also by the other multiscale methods. Finally, it may be switched to different resolutions to provide a comprehensive description of complex structures, as in the case of the basement reconstruction. The magnetic data were upward continued up to 20 km altitude and the multiscale analysis was performed on a profile (AA’) crossing the Adriatic magnetic anomaly (Fig. 2). In particular, we chose the profile coinciding with profile ‘A’ in Mancinelli et al. (2015) in order to compare the multiridge results with the magnetic model. In Fig. 2 we show the multiridge section of the 3 rd derivative order allowing the interference between the effects of different sources to be reduced. As expected, the multiridge section shows a high density of ridges converging at singular points representing the depth position of the sources of the magnetic anomalies. Then, a geometric approach was used prolonging the ridges down to the source region obtaining depth values ranging between 5 and 14 km. We considered exclusively the high altitude portion of the ridges in order to retrieve the deeper (basement) source position. In Fig. 3A we show a comparison between the multiridge section and the magnetic field model of Mancinelli et al. (2015). The depth values retrieved are fairly well correlated to the top of the magnetic basement where the singular points distribution shows a progressive uplift of the deep magnetic boundary toward the NE direction. In particular, the intersection points seem to be related to local change of depth of the basement, probably coincident to the horst and graben structures formed during the extensional movements and crustal thinning of the Middle Triassic. Moreover, the nature of such magnetic sources was evaluated by employing the ‘scaling function’ method along the ridges of the Multiridge section to evaluate the homogeneity degree