ALBERT

Description: One of the contemporary trend in seismology is process huge data sets automatically with use of Neural Network (NN) formalism. We present seismogram onsets interpretation obtained both by Convolution NN as well as Recurrent NN approach. We investigated data from Acoustic Emission loading experiment with Westerly Granite. Such data appeared to be suitable for testing of NN approach as they are more homogeneous then data originated from natural earthquakes, but simultaneously they are complex enough not to be of trivial interpretation. We designed NN architecture, learned in and compare the results with biased interpretation. We were searching not only for onsets on individual seismograms but we try to identified the whole events. In addition to automatic onsets identification we (also automatically) determined event location and seismic moment tensor. Comparison with biased data proved that these automatically obtained values can be successfully used as preliminary estimation at least. Problems of multiple events identification are discussed as well. The method has a potential to be applicable on natural earthquake seismograms.

Description: The focal mechanism is a parameter of an earthquake that typically demands observations surrounding the focus. If such observations are not available, only limited information can be retrieved regarding the process of rupturing (e.g., geometry, and the orientation of the fault plane and slip), not its shear versus non‐shear nature. In extreme cases of monitoring that incorporate a single station, a mechanism can only be estimated if data from additional seismic phases are available. However, such is not the case for weak seismic events, in which a station often only records direct P and S waves. Under such conditions, information is severely limited, and standard synthetics‐to‐data‐matching does not allow even the simplest and most constrained seismological source model, namely a double‐couple focal mechanism, to be determined. During data review, we detected an internal ambiguity for solutions of the inverse task using single‐station geometry. Nevertheless, we found that at least some important features of a focal mechanism may be retrieved. Such scenarios occur for particular configurations of focal mechanism orientation and source‐station direction, as follows: (1) If a station lies in a direction within the quadrant of compressions, the T axis can be retrieved with reasonable error. However, the P axis remains ambiguous. (2) Conversely, if a station lies in a direction within the quadrant of dilatations, reasonable resolution of the P axis is possible, while the T axis remains ambiguous. (3) If the data are noisy, apart from extreme cases of contamination, the possibility of estimating the direction of the T/P axis remains intact. The closer source‐station direction is to the center of the quadrant of compressions or dilatations, the better the resolution of the T/P axis. Our method was applied to selected weak earthquakes recorded by the three‐component borehole seismic station MDBI, located near the Dead Sea in Israel.

Description: We retrieved source mechanisms in the moment-tensor description (MT) of an intense earthquake-swarm activity on the Reykjanes Peninsula (SW Iceland) in 2020 to 2021 preceding the Fagradalsfjall volcanic eruption in March 2021. We inverted displacement amplitudes of the direct P- and S-waves from a local network REYKJANET complemented the network of the University of Cambridge and by the near stations of the regional SIL network; a total of about 30 stations covering well the Fagradalsfjall are used. For a sufficient resolution of the source mechanism an accurate enough response of the medium the Green’s function (GF) is demanded. For this purpose we used a 1-D velocity model for the Reykjanes Peninsula by Vogfjörd et al. (2002). For the GF construction, we applied the ray theory. Substantial effort has been devoted to facilitate a batch processing of the events, first of all it means to allow the ANRAY to work in a semi-automatic regime, which is not its inherent mode of work. As a result, we have a comprehensive tool for retrieval of MTs on the Reykjanes Peninsula. We demonstrate its usefulness by treating a set of several tens of events. Source mechanisms of some of them are compared to those obtained by ISOLA, a software for processing of regional earthquakes, which however applies different data and different approach. Our set of MTs retrieved provides a deeper insight into the mechanism of faulting due to dyke propagation in the region in question from the beginning of 2020 to March of 2021.