SPATIO-TEMPORAL CHARACTERIZATION OF FAULTING BENEATH LONG VALLEY CALDERA

The Long Valley Caldera, situated adjacent to the Sierra Nevada in Eastern California, was created following the eruption and subsequent collapse of a supervolcano in 767 ka. Subsequently, the magmatism migrated in a southwestly direction, and most recently erupted at Mammoth Mountain between 50 and 100 ka. Since 1979, significant uplift exceeding 25 mm/yr has been observed at the resurgent dome at the centre of the caldera and significant earthquake swarms were detected in 1997, leading to concerns of increased magmatic activity. However, more recent studies have suggested that the uplift may be attributed to hydrothermal fluids released from crystallization of the historic granitic magma chamber. Further understanding of the spatio-temporal distribution of fluid migration is critical for volcanic hazard assessment.

We present a new regional high-resolution relocated seismic catalogue for the Long Valley Caldera (2000-2019), which allows us to visualize the previously unknown distribution of faults at depth, as well as the activation of these faults in time. This was achieved using cutting-edge deep-learning methods for phase detection and association, as well as a detailed 3D velocity model. Individual fault structures are identified using a density-based spatial clustering algorithm, and results suggest a pervasive network of vertically oriented structures beneath and to the south of the caldera which are periodically activated. We additionally present a corresponding catalogue of moment tensors calculated using an automated first motion polarity picker in conjunction with a probabilistic moment tensor inversion scheme. Preliminary results indicate that some events may have non-double-couple moment tensors, which are generally more common beneath Mammoth Mountain.