今週のうなぎセミナーについてお知らせいたします。

Here is information of the Unagi-seminar(May, 7th).

************** Seminar on Seismology IV A, C /地震学ゼミナールIV A, C (Unagi Seminar) **************

科目：地震学ゼミナールIV A, C / Seminar on Seismology IV A, C（修士・博士）
日時：2026年 5月 7日 (木) 13:30~
場所：京都大学 防災研究所 本館E-232D

Date and Time：2026-05-07(Thursday), 13:30~
Place：Uji Campus Main Building E232D

---------- ---------- ---------- ----------

Speaker 1（発表者）:　西沢　貴志（Takashi Nishizawa）
Title（題目）:
Detection of Possible Slow Slip Events Using Seismic & Geodetic Approaches in Costa Rica

Abstract（要旨）:
Slow slip events (SSEs), spontaneous transient aseismic slip lasting from days to years, are common in subduction zones (e.g., Obara & Kato, 2016). Stress perturbations associated with SSEs can temporarily enhance background seismicity, inducing earthquake swarms and disturbing the recurrence intervals of repeating earthquakes (Uchida et al., 2019). Such seismic activity provides useful indirect indicators of transient aseismic slip. In contrast, GNSS displacement time series record crustal deformation associated with aseismic fault slip more directly, offering independent information on SSEs.
In this study, we attempt to detect SSEs by integrating analyses of earthquake swarms, repeating earthquakes, and GNSS time series offshore Costa Rica in the Central American subduction zone. Although repeating earthquakes have been reported in many subduction zones worldwide (Uchida & Bürgmann, 2019), their detailed spatial distribution remains unclear in several regions, including offshore Costa Rica. Furthermore, various types of fault slip events, ranging from slow to fast, have been observed in this area (e.g., Dixon et al., 2014). Therefore, we focus on offshore Costa Rica as the study area.
The analysis period spans from 2015 to 2020. Stable seismic waveform records from a broadband seismic network are available throughout this period. An earthquake catalog was constructed using deep-learning-based phase picking with PhaseNet (Zhu & Beroza, 2019) and event association with GaMMA (Zhu et al., 2022), and hypocenters were determined using hypomh (Hirata & Matsu’ura, 1987). Earthquake swarm activity was identified by applying a spatiotemporal ETAS model (Zhuang et al., 2002) and evaluating temporal variations in the difference between the observed and predicted cumulative numbers of earthquakes (Nishikawa & Ide, 2018). Repeating earthquakes were detected using cross-correlation analysis of filtered waveforms, and event pairs with correlation coefficients greater than 0.95 at two or more stations were regarded as repeating earthquakes (Uchida et al., 2019). Based on the occurrence dates of earthquake swarms and repeating earthquakes, temporally proximal events were grouped and compiled into a unified catalog.
Transient crustal deformation potentially associated with SSEs was extracted from GNSS displacement time series. As preprocessing, coseismic offsets due to earthquakes were removed based on the ANSS earthquake catalog, and common-mode errors were corrected. Linear models with and without a step were fitted to the displacement time series in the plate convergence direction using a 90-day moving window. Transient deformation signals were detected based on differences in the Akaike Information Criterion between the two models (ΔAIC, Nishimura et al., 2013). Based on the ΔAIC time series at all stations, temporally proximal events were grouped to construct an SSE catalog.
We examined whether peaks in ΔAIC occur around the dates of earthquake swarms and repeating earthquakes. The results show that ΔAIC peaks are significantly more likely to occur near these dates than expected from randomly assigned catalog dates. We also examined whether earthquake swarms and repeating earthquakes occur around the dates of SSEs. Although such seismic events were often observed in association with SSEs, no strong temporal correlation was found between the timing of these events and ΔAIC peaks.
Finally, based on these catalogs, the slip amounts and locations of SSEs were estimated from the detected deformation signals assuming a circular crack model with constant stress drop (Eshelby, 1957; Madariaga & Ruiz, 2016). Transient deformation events exhibit good spatiotemporal correspondence with seismic activities detected in the eastern and western parts of the peninsula. In particular, several earthquake swarms and repeating earthquakes were found to be distributed adjacent in space and time to SSEs. These results suggest the occurrence of transient aseismic slip offshore the Nicoya Peninsula.

* * * * * * * * * * * * * *

Speaker 2（発表者）:　Jan Sher Khan
Title（題目）:
Greedy Automatic Signal Decomposition (GrAtSiD) for Ocean Bottom Pressure Data at the Hikurangi Subduction zone New Zealand

Abstract（要旨）:
Slow Slip Events (SSEs) are an important class of transient tectonic phenomena occurring along subduction zones, such as the Hikurangi margin in New Zealand, without generating strong seismic shaking. Ocean Bottom Pressure (OBP) data can play a key role in detecting these events.
In this study, we first apply the Baytap08 tidal analysis method to isolate and remove tidal components from the OBP time series. A small slow slip event is identified in the processed data. The same non-tidal data are further analyzed using a modified version of the Greedy Automatic Signal Decomposition (GrAtSiD) algorithm, originally developed for GNSS time series.
This approach decomposes the residual pressure signal into secular, seasonal, and transient components through a greedy basis selection framework combined with L2-regularized inversion.