On 19 March 2013, a strong, shallow, induced seismic event struck a mining panel in the room‐and‐pillar Rudna copper mine in southeastern Poland. The event caused important damage at the mining tunnel and trapped 19 miners, who were safely rescued a few hours later. Although mining‐induced seismicity is frequent at this mine, the 19 March event was unusual because of its larger magnitude, its occurrence far from the mining stopes, and because it was accompanied by a strong hazardous rockburst. The mining inspections following the event verified the occurrence of a rockfall with tunnel floor uplift but also recognized the presence of a faulting structure at the hypocentral location. The availability of three monitoring networks (including local and regional data, short‐period and broadband seismometers, and surface and in‐mine installation) presented an optimal setup to determine rupture parameters and to compare the performance and results from different installations.
We performed waveform and spectral‐based analysis to infer source properties, with a particular interest to the determination of the rupture processes, using different moment tensor (MT) inversion techniques. Our results are surprisingly different, ranging from a dominant thrust mechanism, resolved at closest distances, to a collapse‐type rupture, resolved at regional distances. We demonstrate that a complex rupture model is needed to explain all observations and to justify these discrepancies. The final scenario indicates that the rupture nucleated as a weaker thrust mechanism along a pre‐existing weakened surface and then continued in a more energetic collapse event. The local LUMINEOS surface network has the potential to resolve both subevents but not using a standard MT decomposition. Here, we propose a new MT decomposition and an alternative MT fitting procedure that can be used to analyze the MT of collapse sources.