Earthquake Phase Association Using a Bayesian Gaussian Mixture Model

Earthquake phase association algorithms aggregate picked seismic phases from a network of seismometers into individual earthquakes and play an important role in earthquake monitoring. Dense networks improved picking methods produce massive data sets, particularly for swarms aftershocks occurring closely time space, making challenging problem. We present new method, the Gaussian Mixture Model Association (GaMMA), that combines mixture model measurements (both amplitude), with location, origin time, magnitude estimation. treat as unsupervised clustering problem probabilistic framework, where each corresponds to cluster P S hyperbolic moveout arrival times decay amplitude distance. use multivariate distribution collection picks event, mean which is given by predicted causative event. carry out pick assignment determine parameters (i.e., magnitude) under maximum likelihood criterion using Expectation-Maximization (EM) algorithm. The GaMMA method does not require typical steps other algorithms, such grid-search or supervised training. results on both synthetic test 2019 Ridgecrest sequence show effectively associates temporally spatially dense while producing useful estimates location magnitude.