In order to provide a method to classify the quality of hypocenters seismic locations, we develop an empirical formula as an alternative to the definition of meaningful intervals. The idea starts by observing that quality parameters of seismic locations, coming from a probabilistic inversion, are linked to each other. For this reason, we build up an empirical formula, based on a normalized Euclidean norm, that includes simultaneously six parameters and produces the goodness of the seismic location in terms of a unique value. Aiming to verify the usefulness of our formula, we applied it to the Amatrice-Visso-Norcia seismic sequence catalog (Chiaraluce et al., 2017). The locations were performed by using the NonLinLoc code (Lomax et al., 2000) and the parameters included in the formula are: the root mean square (RMS), the azimuthal seismic gap, the number of phases, the error on the horizontal components, the error on vertical components and the distance between the maximum likelihood hypocenter and the expected one (LocDist). The last parameter was added to establish the unimodal (high-resolved location) or multimodal (poor-resolved location) character of the probability distribution. For each parameter, we determined 5 percentiles of its distribution. Then, we defined 5 quality values as function of the percentiles values, obtaining different functions for different parameters. Finally, we use a formula that includes simultaneously different quality values as function respectively of the parameters introduced above. N is the number of quality values used. The q parameter allows to classify the resulting locations in four classes (A, B, C and D). Classes (A and B) include the 52% of the relocated earthquakes, while the lowest class D only includes the 10%. The locations (in class A and B) cross-sections are similar to the ones shown in Chiaraluce et al., 2017, that were classified with standard intervals subdivision. In addition, A and B locations cross-sections allow us to observe more defined structures compared to Chiaraluce catalog.
Testing an Empirical Formula to Classify the Quality of Seismic Locations / Michele, Maddalena; Latorre, Diana; Emolo, Antonio. - (2018). (Intervento presentato al convegno 2018 Fall Meeting of the American Geophysical Union tenutosi a Washington, D.C. (USA) nel 10-14 December, 2018).
Testing an Empirical Formula to Classify the Quality of Seismic Locations
and Antonio Emolo
2018
Abstract
In order to provide a method to classify the quality of hypocenters seismic locations, we develop an empirical formula as an alternative to the definition of meaningful intervals. The idea starts by observing that quality parameters of seismic locations, coming from a probabilistic inversion, are linked to each other. For this reason, we build up an empirical formula, based on a normalized Euclidean norm, that includes simultaneously six parameters and produces the goodness of the seismic location in terms of a unique value. Aiming to verify the usefulness of our formula, we applied it to the Amatrice-Visso-Norcia seismic sequence catalog (Chiaraluce et al., 2017). The locations were performed by using the NonLinLoc code (Lomax et al., 2000) and the parameters included in the formula are: the root mean square (RMS), the azimuthal seismic gap, the number of phases, the error on the horizontal components, the error on vertical components and the distance between the maximum likelihood hypocenter and the expected one (LocDist). The last parameter was added to establish the unimodal (high-resolved location) or multimodal (poor-resolved location) character of the probability distribution. For each parameter, we determined 5 percentiles of its distribution. Then, we defined 5 quality values as function of the percentiles values, obtaining different functions for different parameters. Finally, we use a formula that includes simultaneously different quality values as function respectively of the parameters introduced above. N is the number of quality values used. The q parameter allows to classify the resulting locations in four classes (A, B, C and D). Classes (A and B) include the 52% of the relocated earthquakes, while the lowest class D only includes the 10%. The locations (in class A and B) cross-sections are similar to the ones shown in Chiaraluce et al., 2017, that were classified with standard intervals subdivision. In addition, A and B locations cross-sections allow us to observe more defined structures compared to Chiaraluce catalog.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
