Proceedings of the 12th International INQUA meeting on paleoseismology, active tectonic and archaeoseismology

R E S U LT S Detections and validations We conducted a comparison between the outcomes from manual and Deep Learning picker and associator (Zhu and Beroza, 2019; Zhang et al., 2019) over the entire dataset. This involved determining time diDerences between the analysts and automatic picks. The time diDerences we measured are small, the standard deviations for P and S waves are 0.20 and 0.26, respectively. One of the most important Gndings from the Grst six months of seismic monitoring in the Cochabamba region indicates the presence of an intense microseismic activity in the area. SpeciGcally, 439 earthquakes associated with magnitudes exceeding M L 0.2. The seismicity catalog inferred a completeness magnitude of M L 0.8. Joint relocation and inversion In order to estimate eventual additional biases coming from the 1D velocity model we performed the joint hypocenter, velocity model and station coeJcient determination of VELEST (Kissling et al., 1994), and the Joint Hypocenter Determination (JHD method of Pujol et al., 2000). For that purpose, we selected a set of relatively well-constrained earthquakes from the initial catalog of events (138 earthquakes with an azimuthal gap less than 180°). The Pg and Sg time corrections calculated for each station do not exceed +/- 1.5 seconds. Moment tensor inversion We performed waveform inversions incorporating stations from the temporary network (POBnet), using the FMNEAR method (Delouis, 2014). We obtained results for only two earthquakes, the M L 4.3 that occurred on May 18, 2022 (19h20 UTC) and the M L 2.5 on August 14, 2022. A visual quality control of the records was carried out in order to select the stations taken into account in the inversion. The waveform inversion results are depicted in Figure 2. The solution for the M L 4.3 presents a thrust focal mechanism characterised by a sub- horizontal nodal plane (and its subvertical conjugate), associated with motion of the upper block (hanging wall) towards the northeast (NE). Fig. 2: The ML 4.3 May 18th, 2022 earthquake FMNEAR waveform inversion. a) Map showing seismic stations (triangles) and the best focal mechanism found. Stations included in the inversion are shown in green.