Technical Committee Earth Sciences

Application of a Probabilistic Moment Tensor Inversion to Data Recorded above the Northgerman Gas Fields

Even if faults are not visible on geological outcrops, source mechanisms of earthquakes occurring on those faults may distinguish between different types of faulting. We implemented a novel, probabilistic full-waveform moment tensor inversion for the application to shallow micro-earthquakes close to gas fields in North Germany. Due to the probabilistic approach, parameter trade-offs, uncertainties and ambiguities are mapped. The implemented bootstrap method implicitly accounts for modelling errors that may affect every station and phase in a unique way. Furthermore, event locations are provided by the algorithm. Such hypocentres, which are estimated simultaneously with moment tensors, are often less sensitive to uncertainties in crustal structure. We analysed events from the November 20th, 2019, Kirchlinteln series as well as the October 1st, 2018 ML 3.6 Lastrup event (00:28 UTC). We carefully investigated station quality, the usable frequency range, input data types, inversion parameters as well as velocity models and supplemented the analysis by a H/V study. The Kirchlinteln events show slightly oblique normal faulting and potentially, a complex faulting process involving several faults, fitting known local fault traces. The 00:28 UTC Lastrup event reveals thrust faulting, most likely on a single fault not fitting the regional fault trend.