GNGTS 2023 - Atti del 41° Convegno Nazionale

Session 3.3 ______ ___ GNGTS 2023 noise sources affect the location accuracies . (1) The first test (SYNTH-01) consists of the inversion of synthetic traveltimes computed on the DAS channels, assuming the independent location as the “true” one and adding white Gaussian noise. This exercise provides insight into the geometrical effects that influence the precision of epicentral location. In addition to this, some geometric parameters are computed to better describe the geometry of the problem in function of the location uncertainties.Then, 2) three different statistical assumptions are considered for such traveltimes, to simulate typical observations in DAS recordings, that is: i) the arrival time picking uncertainty increases with the distance from channel to the event (SYNTH-02), ii) S-wave pickings are mis-abeled as P-waves at some DAS channels (SYNTH-03) and iii) specific channel packets are associated with very-low SNRs (as a tentative to model fiber coupling inhomogeneities (SYNTH-04). Finally, these synthetic traveltimes are inverted to assess the influence of these assumptions on the ultimate outcome of the monitoring exercise: event location. Results The estimated epicentral locations obtained from the automatic traveltime picking procedure exhibit varying degrees of uncertainty. DAS arrays with greater azimuthal coverage provide more consistent locations (Fig 2a), which is an expected behavior, particularly considering the longitudinal polarization of strain/strain-rate measurements. This behavior is further supported by SYNTH-01 test (Fig 2b). Furthermore, the azimuthal gap parameter influences the epicentral uncertainty when compared to the independent location . However, exceptions to this behavior are common, indicating that geometrical factors alone do not exclusively control the uncertainties in source location with DAS. Such “biases” between observed and reference locations are indeed partially explained by arrival time uncertainties, especially the ones modeled by SYNTH-03 and SYNTH-04. As an example, Figures 2c-d show how the observed location uncertainty can be reproduced, to a certain degree, from synthetic simulations.