Seasonal variation of crustal seismicity in San-in district, southwest Japan

It has been reported that earth surface's process such as tidal loading, ice and snow load, and heavy rainfall load, modulates seismicity in the crust (e.g., Heki, 2003; Amos et al., 2014). Elucidation of the periodicity of seismicity provides us a new insight into seismicity model and earthquake physics.

San-in district, southwest Japan, is an active seismicity zone characterized as high shear strain rate (Nishimura and Takada, 2017). Moreover, Ogata (1983) has shown seasonal variation of seismicity in Inner zone of southwest Japan. We here investigate seasonal variation of the crustal seismicity in San-in district.

We used the JMA catalog (constructed by Japan Meteorological Agency) from 1975 to 2017 (magnitude M >= 3.0). We applied space-time Epidemic Type Aftershock Sequence (ETAS) model (e.g., Ogata, 1998; Zhuang et al., 2004) to the catalog to calculate the possibility how likely each event is to be categorized as an aftershock or independent event. Then, we produced 10,000 stochastic copies of the declustered catalog, following a method proposed by Zhuang et al. (2002).

Using each stochastic-declustered catalog, we made a time-series of a background seismicity rate for the studied period. Then, we performed Fourier transform to each catalog, calculated the amplitude of sinusoidal and cosine functions, and evaluated the significance of seasonal variation of background seismicity rate. As a result, we found out a significant half-year seasonal variation of seismicity during the whole studied period; a background seismicity rate slightly increases in spring and autumn. Interestingly, past large earthquakes in the studied region (from 1850 to 2017, magnitude M >= 6.2) show a similar seasonal variation to smaller earthquakes.

The semiannual variation revealed in the present study would likely be explained by a combination of following two factors; 1) a change in stress loading rate induced by the thaw in spring, 2) a fault-weakening by increase in pore fluid pressure associated with heavy rainfall in autumn.