4:30 PM - 4:45 PM
[SSS12-11] A detailed analysis of the Wakayama earthquake swarm based on hypocenter clustering
Keywords:hypocenter clustering, fluid, Wakayama earthquake swarm, spatiotemporal evolution
Introduction
Earthquake swarms are not Omori-like seismicity in a specific area without clear main shock in the first stages. Persistent earthquake swarms, such as those occurring in the western Kii Peninsula and around Ashio in Japan, are still enigmatic phenomena. Clarifying the driving mechanism of the persistent swarms is important to understand the nature of earthquake swarms.
We focused on a persistent earthquake swarm in the shallow crust near Wakayama in the western Kii Peninsula (Wakayama earthquake swarm). Previous studies suggested that fluid from the Philippine Sea Slab supplied to the geological faults in the brittle crust is a driving factor of the swarm (e.g., Maeda et al., 2018, 2021; Nakajima et al., 2023). However, the spatiotemporal characteristics of the swarm activity are not revealed in detail. In this study, we conducted a systematic and detailed analysis of the spatiotemporal characteristics in the swarm using hypocenter clustering.
Data & Methods
First, we performed hypocenter relocation of 8,157 earthquakes in the swarm area (2005–2023, M≧1.0) using the Double-Difference algorithm (Waldhauser & Ellsworth, 2000). Then we performed hypocenter clustering using DBSCAN (Ester et al., 1996), a density-based non-hierarchical clustering method. Based on the clustering result, we characterized the hypocenter distribution, estimated the b-value, and examined the earthquake recurrence interval distribution using the method of Utsu (1969) for each cluster.
Results
The hypocenter clustering classified 3,857 of the 6,754 relocated hypocenters into 28 clusters, while the remaining 2,897 hypocenters were classified as noises not belonging to any cluster. We found multiple planar hypocenter distributions with variable strike and dip by detailed visual inspection of the hypocenter distribution for each cluster.
The b-value estimated for the entire swarm activity was 1.08, whereas those in each cluster ranged from 0.52 to 1.92. We also found that the b-value for the clusters in the southwestern part of the swarm area tended to be high.
The earthquake recurrence interval distribution for the entire swarm activity showed the characteristics of persistent seismicity with an almost constant seismicity rate on the smaller side for the 1–10 days recurrence interval and a sharp decrease on the larger side. On the other hand, the results for each cluster showed various distribution shapes. For example, the distribution for the clusters with low b-values and earthquakes with magnitudes of approximately 4.0 showed a gradually decreasing seismicity rate pattern, which is characteristic of sporadic seismicity.
Discussion
Detailed inspection of the relocated hypocenter distribution revealed that multiple planar structures correspond to fault planes. Their strikes and dips varied among the hypocenter clusters, which may reflect the complex stress field in the region area of the Wakayama earthquake swarm.
Previous studies have reported a low-velocity zone beneath the swarm area (Nakajima, 2023), and an uplift of approximately 10 mm/year in the eastern part of the swarm area (Yoshida et al., 2011). Based on these observational facts, the fluid supplied from the deep has been considered a driving factor of the Wakayama earthquake swarm. However, the variety of spatiotemporal characteristics in local seismicity revealed in this study suggests that the driving mechanisms of the swarm cannot be explained by the fluid supply alone. The coexistence of clusters with constant and sporadic seismicity within a nearly 25 km square area may reflect the differences in the local mechanical and hydraulic properties of the crustal components.
This study revealed the Wakayama earthquake swarm, which has been regarded as just persistent earthquake swarm, is composed of multiple local scale seismicity with different spatiotemporal characteristics. This is an important finding to understand the driving mechanism of the persistent earthquake swarms.
Earthquake swarms are not Omori-like seismicity in a specific area without clear main shock in the first stages. Persistent earthquake swarms, such as those occurring in the western Kii Peninsula and around Ashio in Japan, are still enigmatic phenomena. Clarifying the driving mechanism of the persistent swarms is important to understand the nature of earthquake swarms.
We focused on a persistent earthquake swarm in the shallow crust near Wakayama in the western Kii Peninsula (Wakayama earthquake swarm). Previous studies suggested that fluid from the Philippine Sea Slab supplied to the geological faults in the brittle crust is a driving factor of the swarm (e.g., Maeda et al., 2018, 2021; Nakajima et al., 2023). However, the spatiotemporal characteristics of the swarm activity are not revealed in detail. In this study, we conducted a systematic and detailed analysis of the spatiotemporal characteristics in the swarm using hypocenter clustering.
Data & Methods
First, we performed hypocenter relocation of 8,157 earthquakes in the swarm area (2005–2023, M≧1.0) using the Double-Difference algorithm (Waldhauser & Ellsworth, 2000). Then we performed hypocenter clustering using DBSCAN (Ester et al., 1996), a density-based non-hierarchical clustering method. Based on the clustering result, we characterized the hypocenter distribution, estimated the b-value, and examined the earthquake recurrence interval distribution using the method of Utsu (1969) for each cluster.
Results
The hypocenter clustering classified 3,857 of the 6,754 relocated hypocenters into 28 clusters, while the remaining 2,897 hypocenters were classified as noises not belonging to any cluster. We found multiple planar hypocenter distributions with variable strike and dip by detailed visual inspection of the hypocenter distribution for each cluster.
The b-value estimated for the entire swarm activity was 1.08, whereas those in each cluster ranged from 0.52 to 1.92. We also found that the b-value for the clusters in the southwestern part of the swarm area tended to be high.
The earthquake recurrence interval distribution for the entire swarm activity showed the characteristics of persistent seismicity with an almost constant seismicity rate on the smaller side for the 1–10 days recurrence interval and a sharp decrease on the larger side. On the other hand, the results for each cluster showed various distribution shapes. For example, the distribution for the clusters with low b-values and earthquakes with magnitudes of approximately 4.0 showed a gradually decreasing seismicity rate pattern, which is characteristic of sporadic seismicity.
Discussion
Detailed inspection of the relocated hypocenter distribution revealed that multiple planar structures correspond to fault planes. Their strikes and dips varied among the hypocenter clusters, which may reflect the complex stress field in the region area of the Wakayama earthquake swarm.
Previous studies have reported a low-velocity zone beneath the swarm area (Nakajima, 2023), and an uplift of approximately 10 mm/year in the eastern part of the swarm area (Yoshida et al., 2011). Based on these observational facts, the fluid supplied from the deep has been considered a driving factor of the Wakayama earthquake swarm. However, the variety of spatiotemporal characteristics in local seismicity revealed in this study suggests that the driving mechanisms of the swarm cannot be explained by the fluid supply alone. The coexistence of clusters with constant and sporadic seismicity within a nearly 25 km square area may reflect the differences in the local mechanical and hydraulic properties of the crustal components.
This study revealed the Wakayama earthquake swarm, which has been regarded as just persistent earthquake swarm, is composed of multiple local scale seismicity with different spatiotemporal characteristics. This is an important finding to understand the driving mechanism of the persistent earthquake swarms.