Accelerating low-frequency ground motion simulation for finite fault sources using neural networks

Abstract

Summary In the context of early emergency response to moderate and large earthquake shaking, we present a simulation based low-frequency ground motion estimation workflow that expedites an existing simulation method while taking into account simplified source process information. We focus on using source information that can be expected to be available shortly after an impacting earthquake, e.g. moment-tensor and simple finite-fault parameters. We utilize physics based simulations (PBSs) which can include effects based on source orientation or finite faults, like rupture directivity. In order to keep the computational effort within feasible bounds and to apply the approach on global scale, we restrict ourselves to a low-frequency setup (standard 1D layered earth model and 2 Hz sampling frequency) for either a moment tensor or a simple kinematic finite fault model. From the simulated records we then extract ground motion parameters of interest for arbitrary locations within the area of expected impact and display the expected spatial patterns of ground motion. Although simulations are kept simple, the results from this low-frequency ground motion parameter simulation (e.g. for peak-ground-displacement) are in good agreement with observations from two well-studied earthquakes and partially more accurate than traditional, more empirical approaches (standard deviation <0.3 log10 units). However, waveform calculation and subsequent ground motion parameter extraction is computationally expensive. For a significant computational speedup in the context of rapid ground motion assessment, we directly train neural network models from large sets of source model information and their corresponding spatial ground motion distribution. We show that the trained neural networks are able to reproduce the earthquake source related effects, like directivity and focal mechanism patterns, of the ground motion in any case. Given a set of source parameters, we obtain prediction errors smaller than 0.05 log10 units (ca. 11 per cent) and a magnitude dependent increase in computational speed of more than 1,000 times compared to the initial waveform modeling. The proposed procedure enables thus to immediately compute probabilistic ground motion maps related to uncertainties in source parameters estimates, e.g. by sampling distributions based on parameter uncertainties or directly from an existing ensemble of focal parameter solutions.

Recent advances in earthquake seismology using machine learning

“…Raghucharan et al (2021) appended the simulated data from the validated seismological model to the observation data so that the trained dataset for their ML model covered a wide range of magnitude and distance, filling the data gap region. Lehmann et al (2023) prepared a synthetic database based on many PBSs considering various source mechanisms or finite faults for constructing an ML model to simulate low-frequency ground-motion parameters for arbitrary focal mechanisms or finite fault sources. ML training with simulated data alone has been done in research fields where observation data are fundamentally scarce, such as in the cases of tsunamis (Makinoshima et al 2021) and gravity waves (Licciardi et al 2022).…”