GNGTS 2021 - Atti del 39° Convegno Nazionale

GNGTS 2021 S essione 3.1 402 The MT data set from the Larderello-Travale geothermal area Since the 90s, several MT surveys have been carried out in the Travale geothermal field to aim for industrial exploration or scientific research (Manzella et al., 2006; Santilano, 2017). The collection of the “vintage” MT data is composed of three data sets. Fig. 1 shows the data sets acquired in 1992 (black-labeled squares covering both the Larderello and Travale areas), in 2004 (55 sites in blue-labeled circles) and in 2006-07 (19 sites in red-labeled triangles). These data were studied following the latest techniques of MT data analysis to determine the geoelectrical dimensionality, directionality and phase tensor properties by means of the WALDIM software and MTpy python toolbox (Martí et al., 2009; Kirkby et al., 2019). This thorough analysis of the observed MT tensors provided an overview of the underlying resistivity distribution. It revealed a 3D behavior of the deep structures and a primary direction of N130°E for the geoelectrical strike (Pace, 2020). 3D inversion was proved to be the best approach to interpret this MT data set, while the 2D strike direction can be considered a “first-order approximation”. So far, the MT data acquired in Travale have been interpreted using both 1D PSO and 2D inversion (Santilano, 2017; Manzella et al., 2006). Therefore, both 2D PSO and 3D inversion represent new valid tools to provide novel resistivity models and further knowledge of the LTGA. PSO of MT data from the Larderello-Travale geothermal field The PSO algorithm was applied to two MT profiles located in the LTGA (Pace et al., 2019b, c). Even though the dimensionality analysis suggests 3D interpretation as the most appropriate for this data set, there are some conditions for the validity of 2D interpretation: inadequate spatial coverage (isolated MT profile), a clear strike direction (as suggested from directionality of the tensors) and the high computational and numerical complexity of 3D modeling (Ledo, 2005). The final models succeeded in imaging very complex resistivity structures and confirmed those presented in the past (Manzella et al., 2006; Santilano, 2017). PSO models offered the following advantages (Pace et al., 2019b, c): (i) the final models have not been initially biased by an external starting model derived from geology and (ii) the RMSEs associated with the final PSO models were lower than those associated to the models published in the past by using different inversion techniques. Fig. 1 - The collection of MT data sets in the Larderello-Travale area. The data set acquired in 1992 is marked with black-labeled squares. The data set acquired in 2004 (55 sites) is marked with blue-labeled circles. The data set acquired in 2006-07 (19 sites) is marked with red-labeled triangles.