A comparative study on the self-similarity hypothesis of regular and slow earthquake growth

TL;DR

This study examines the self-similarity hypothesis in regular and slow earthquakes, analyzing theoretical models and seismic data to understand how source parameters and scaled energies vary across different earthquake types.

Contribution

It applies a refined self-similar rupture model to various seismic events, revealing differences in scaled energy and questioning the universality of self-similarity in slow earthquakes.

Findings

Low-frequency and very-low-frequency earthquakes share similar scaled energies.

Slow slip events show multiple modes with different scaled energies.

Seismic moments of very-low-frequency earthquakes are smaller than expected for self-similar ruptures.

Abstract

Fault ruptures of regular earthquakes typically grow in a self-similar manner, where the radiated energy is proportional to the seismic moment. Their proportionality factor, termed as scaled energy, has been conventionally described as the ratio of stress drop to stiffness. By analyzing the self-similar circular crack model by Sato and Hirasawa (1973), Matsu'ura (2024) found a correction prefactor for this theoretical representation of the scaled energy, the cubed ratio of the rupture speed to the S-wave speed. The stress drop times the cubed rupture speed is the scaling prefactor of the self-similar seismic moment, and thus, Matsu'ura's solution tells that the seismic moment rate is determined by the scaled energy in the self-similar rupture growth. We rearrange the properties of the self-similar solution from this perspective, apply this Sato-Hirasawa-Matsu'ura relation to a series of seismic events thought to be self-similar, and estimate their source parameters and scaled energies. According to our estimation using the above self-similarity model, seismologically detected low-frequency and very-low-frequency earthquakes have a common scaled energy, while geodetically detectable slow slip events indicate multiple modes with different scaled energies. Meanwhile, for very-low-frequency earthquakes and tectonic tremors, the seismic moments were significantly smaller than expected for self-similar ruptures despite their moments being proportional to cubed duration values. It keeps alive a hitherto less contemplated possibility: the proportionality of the moment to cubed duration may not be a manifestation of the self-similarity for slow earthquake rupture growth.