Abstract

Numerous deep short-term slow slip events (S-SSEs) and long-term slow slip events (L-SSEs) have been detected in the Nankai Trough subduction zone in southwestern Japan based on Global Navigation Satellite System (GNSS) data. Recently, slow slip events (SSEs) lasting several months, which are significantly longer than typical deep S-SSEs but similar to or shorter than L-SSEs, have been reported near the trench, particularly off the Kii Peninsula in the central Nankai Trough. To investigate whether similar SSEs occur in the shallow region of the Suruga Trough, the easternmost segment of the Nankai Trough, we analyzed changes in GNSS baseline lengths between stations on the upper and lower plates across the Suruga Trough. To identify the timing of baseline shortening, we first converted the GNSS baseline length time series into velocity time series. We then compared these velocity time series with the correlation coefficients computed between the observed data and a template function representing baseline shortening associated with shallow SSEs. As a result, a signal corresponding to a plausible SSE in late 2019, whose duration can be approximately represented by 118 days, was detected at multiple GNSS stations near Omaezaki. A Bayesian inversion of the fault slip distribution estimated a maximum slip on the order of 10 mm at depths shallower than ~ 10 km on the shallow plate boundary off Omaezaki. This result was consistent across all three assumed plate interface models. Furthermore, slip extending to the trench was also inferred. The estimated moment magnitude (Mw) was approximately 5.5–5.6. This SSE coincided with an increase in small repeating earthquake activity on the deeper side. The discovery of a months-long SSE in the shallow region of the Suruga Trough suggests that SSEs with significantly longer durations than typical deep S-SSEs may be a widespread characteristic of the shallow Nankai Trough subduction zone. Moreover, by relaxing the detection criteria and considering smaller possible SSEs, the results further imply that such events may occur at intervals of approximately 500 days.