GNGTS 2024 - Atti del 42° Convegno Nazionale

Session 3.1 GNGTS 2024 of orders of magnitudes higher than that of ground IP data. Beside these problems, another major issue, not always recognized, is the diference in the modelling approach. Ground IP data is usually modelled dropping the spectral informaton, e.g., turning a full secondary voltage decay into a single value of integral chargeability (e.g. Oldenburg and Li, 1994). Moreover, the efect of current waveform is ofen not modelled, resultng in inversion models in which the retrieved polarizaton magnitude strongly depends on the acquisiton setngs of the current waveform, making a quanttatve comparison between AIP and Ground IP impossible (Olsson et al., 2019). On the contrary, in this study we model the IP spectral content in both AIP and Ground IP data with the same modelling approach, in terms of data-preparaton, model space management and inversion approach. In partcular, the galvanic data have been modelled in 2D in terms of full-voltage decay (instead of the integral chargeability), taking into account the transmiter waveform and the receiver transfer functon (Fiandaca et al., 2012; Fiandaca et al., 2013; Bollino and Fiandaca, 2024). The inductve data have been modelled in 1D, and to reduce the model space and to enhance the spectral resoluton, the frequency dependence and tme constant parameters have been set to vary only horizontally (while resistvity and maximum phase change also with depth). For both the methos, the Maximum Phase Angle (MPA) re-parametrizaton of the Cole & Cole model has been used (Fiandaca et al., 2018). In the MPA Cole-Cole model, the maximum phase of the complex conductvity and the phase relaxaton tme are used instead of 0 and (of the classic Cole & Cole model). The phase of the complex conductvity can be defned in terms of both equatons 1 and 2 as: (eq. 1) The phase reaches his maximum at an angular frequency ≡1/ as: (eq. 2) Finally, the model space of the MPA Cole-Cole model can be writen as: − = { 0 , , , } The MPA parametrisaton replace the strongly-correlated parameters 0 and of the classic Cole- Cole model with the weakly-correlated parameters and (Fiandaca et al., 2018), to improve the resoluton retrieved from inversion IP data of the classical Cole-Cole model. The inversions have been performed with the inversion with EEMverter (Fiandaca et al., 2024), following a modelling scheme that uses voxel model mesh to map the solved parameters via an interpolaton of the forward mesh solutons. The decoupling of the model mesh and the forward mesh allows to work with more fexible and manageable spaces (forward and model) to perform joint inversions and tme laps inversions. In our inversion procedure, in order to increase the parametrical resoluton and the phase sensitvity in depth, we parametrized the spectral parameters ( , ) on an independent mesh respect to resistvity and phase, with diferent lateral constraints and vertcally fxed (as proposed by Viezzoli and Fiandaca in 2021). φ ( ω ) = tg −1 ( σ” ( ω ) σ ′  ( ω ) ) = − tg −1 ( ρ” ( ω ) ρ ′  ( ω ) ) φ max = tg −1 ( σ” (1/ τ φ ) σ ′  (1/ τ φ ) ) = tg −1 ( ρ” (1/ τ φ )) ρ ′  (1/ τ φ ) )