No abstract

Slow earthquakes are slip events on faults that have longer durations than ordinary earthquakes. These events often generate seismic waveforms characterized by dominantly lower frequency than ordinary earthquakes. Over the past two decades, research in geodesy and seismology has revealed a variety of slow earthquake phenomena, such as slow slip events (SSEs), very low-frequency earthquakes (VLFEs), low-frequency earthquakes (LFEs), and tectonic tremor (tremor). Tremor is often considered to represent a swarm of LFEs (Ide et al., 2007;Shelly et al., 2007).Along the Nankai Trough, southwest Japan, deep LFEs and tremor were first identified in the early 2000s at depths of 30-40 km in the deep extension of the megathrust seismogenic zone on the plate boundary (e.g., Katsumata & Kamaya, 2003;Obara, 2002). This type of tremor has been observed in subduction zones worldwide (Obara &

The characteristics of shallow low-frequency earthquakes (LFEs) are related to stress changes on the shallow plate boundary, which are important for understanding the megathrust earthquake cycle. The 5 September 2004 off the Kii Peninsula earthquakes (M JMA = 7.1, 7.4) occurred near the Nankai trough subduction zone, off southwest Japan. Ocean bottom seismometer observations from 22 September to 30 November 2004 detected many shallow LFEs among the ordinary aftershocks. During the observation period, the frequency of shallow LFEs steadily decayed with the exception of episodic activities, which were very sensitive to stress changes caused by tides and the 23 October 2004 Niigata earthquake (M JMA = 6.8). We also confirmed correlations between shallow LFEs and shallow very low frequency earthquakes, which suggest that these slow events represent the same slip phenomenon. These findings will contribute to clarifying the impact of LFEs on megathrust earthquakes.

The Japan Meteorological Agency (JMA) unified seismic catalog has been widely used for research and disaster prevention purposes for more than 20 years. Since the introduction in April 2016 of an improved method of automatic hypocenter determinations (PF method), the number of detected earthquakes has almost doubled due to a decrease in the completeness magnitude around the Tohoku region, where seismicity has been very active in the aftermath of the 2011 Tohoku earthquake. Automatically processed hypocenters of small events, accepted without manual modification, now make up approximately 70% of new events in the JMA unified catalog. In this paper, we show that the introduction of automated processing did not systematically bias the quality of the JMA unified catalog. Approximately 90% of automatically processed hypocenters were less than 1 km from their manually reviewed locations in inland and shallow areas. We also considered the use of automated event characterization in real-time monitoring of earthquake sequences using the example of the April 2016 Kumamoto earthquake sequence, when the PF method could have supplied the catalog with about 70,000 events in real time over the course of 2 months. We show that the PF method is capable of monitoring the migration or expansion of the hypocentral distribution and can support statistical analyses such as variations of the b-value distribution. Further improvements in automatic hypocenter determination will contribute to a better understanding of seismicity as well as rapid risk assessment, especially in cases of swarms and aftershocks.

In this article, we created a well-resolved aftershock catalog for the 2015 Gorkha earthquake in Nepal by processing 11 months of continuous data using an automatic onset and hypocenter determination procedure. Aftershocks were detected by the NAMASTE temporary seismic network that is densely distributed covering the rupture area and became fully operational about 50 days after the mainshock. The catalog was refined using a joint hypocenter determination technique and an optimal 1D velocity model with station correction factors determined simultaneously. We found around 15,000 aftershocks with the magnitude of completeness of ML 2. Our catalog shows that there are two large aftershock clusters along the north side of the Gorkha–Pokhara anticlinorium and smaller shallow aftershock clusters in the south. The patterns of aftershock distribution in the northern and southern clusters reflect the complex geometry of the Main Himalayan thrust. The aftershocks are located both on the slip surface and through the entire hanging wall. The 1D velocity structure obtained from this study is almost constant at a P-wave velocity (VP) of 6.0  km/s for a depth of 0–20 km, similar to VP of the shallow continental crust.

We have developed an empirical earthquake forecast method, Maeda's method, based on the statistical features of precursory seismic swarm activity, that is foreshocks, which sometimes appear before a mainshock, and issuing an alert of a mainshock occurrence within a certain period of time. In this study, we investigated the effectiveness of earthquake forecast of Maeda's method by applying it to seismicity under various tectonic environments of Japan such as regions characterized by interplate seismic activity, a tectonic fault line (concentrated deformation zone), and an island arc area of seismic and volcanic activity. As a result, we confirmed that Maeda's method yielded generally higher scores than a forecast model based on a stationary space‐time epidemic‐type aftershock sequence (ETAS) model. We also found that foreshocks detected along the Japan Trench were distributed along the edges of low‐velocity anomalies and among areas with background swarms related to slow slip events (SSEs). The foreshocks may have been caused by a heterogeneous stress distribution associated with the existence of a plate‐bending axis and a subducted seamount. Foreshocks off Iwate prefecture, in particular, were excited by periodic SSEs. In an inland tectonic zone and an island arc, swarm activity associated with magmatic or fluid activity related to low‐velocity anomalies tended to be followed by a mainshock. Maeda's method is a simple and efficient counting‐number‐based earthquake forecast model and may capture characteristics of foreshocks that reflect a physical phenomenon, such as a nucleation process involving precursory slip, which the stationary ETAS model is not able to represent.

S U M M A R YCurrent earthquake early warning (EEW) systems lack the ability to appropriately handle multiple concurrent earthquakes, which led to many false alarms during the 2011 Tohoku earthquake sequence in Japan. This paper uses a Bayesian probabilistic approach to handle multiple concurrent events for EEW. We implement the theory using a two-step algorithm. First, an efficient approximate Bayesian model class selection scheme is used to estimate the number of concurrent events. Then, the Rao-Blackwellized Importance Sampling method with a sequential proposal probability density function is used to estimate the earthquake parameters, that is hypocentre location, origin time, magnitude and local seismic intensity. A real data example based on 2 months data (2011 March 9-April 30) around the time of the 2011 M9 Tohoku earthquake is studied to verify the proposed algorithm. Our algorithm results in over 90 per cent reduction in the number of incorrect warnings compared to the existing EEW system operating in Japan.

The extremely active seismicity after the 2011 off the Pacific Coast of Tohoku Earthquake (Tohoku Earthquake) caused serious issues for the Earthquake Early Warning (EEW) system in Japan. Multiple concurrent earthquakes generated many false alarms due to the failure of correctly determining the source location. In this study, we propose a method to distinguish multiple concurrent events for EEW. We used the outputs of a B-delta method and principal component analysis, together with P-wave arrival times and amplitudes of triggered stations. The information of non-triggered stations is also used to constrain the locations. We constructed a likelihood function using the source parameters and solved this optimization problem using the particle filter. The appropriate choice of a likelihood function improves the accuracy of the location determination, and the probabilistic framework evaluates the uncertainty of the hypocenters. We applied this method to 72 events for which warnings were issued after the Tohoku Earthquake. The performance of EEW was greatly improved, reducing the false alarms from 22 to 0. The percentage of inaccurate warnings was also reduced from 66% to 6%.