11:30 AM - 11:45 AM
[SGD02-10] Detection of local displacement of GEONET stations using spatial high-pass filtering (3)
Keywords:GEONET, crustal deformation, spatial high-pass filtering, landslide, postseismic deformation
1. Introduction
By applying high-pass filtering to the spatial distribution of displacements observed at the GEONET stations of the Geospatial Information Authority of Japan (GSI), it is possible to remove large-scale crustal deformation and extract localized displacements. These include phenomena such as landslides and abnormal displacements caused by factors like GNSS signal shielding from trees (Fujiwara, 2024a). The displacements often align with the slope of the terrain where the GEONET station is situated (Fujiwara, 2024b). In this study, we present a nationwide analysis of localized displacements, employing high-pass filtering, and an example of local displacements for postseismic deformation following the 2011 off the Pacific coast of Tohoku earthquake using a larger spatial filter scale.
2. Selection of Filter Spatial Scale
The selection of an appropriate spatial scale for the high-pass filter is crucial for achieving the desired analytical objectives. As illustrated in Figure 1, the average distance between GEONET stations is approximately 15 km, with most distances being under 30 km, except for remote islands. Therefore, using a spatial filter size of 30 km effectively isolates local anomalies relative to neighboring stations while minimizing the influence of large-scale crustal deformations.
3. Nationwide Analysis
Figure 2 presents the results of high-pass filtering applied to displacement data of a 10-year period ending in November 2024. This analysis excludes coseismic displacements associated with major inland or nearshore earthquakes. These events produce large spatial variations over short distances, which cannot be fully removed by high-pass filtering.
Consistent with previous studies (Fujiwara, 2024ab), localized displacements attributable to landslides, subsidence, and other ground movements are prominent. The analysis also captures local earthquakes and volcanic activity. Landslides are observed nationwide. Despite filtering out both coseismic and short-term postseismic deformation from the 2016 Kumamoto earthquake, complex longer-term postseismic variations remain as localized anomalies. In contrast, the postseismic deformation from the similarly sized 2024 Noto earthquake does not exhibit such localized characteristics.
Volcanic activity is another notable source of displacement, particularly in eastern Hokkaido, the Izu Islands, Hakone, and Sakurajima.
4. Local Displacement in the Postseismic Deformation of the 2011 Tohoku Earthquake
In the Tohoku region, north-south subsidence along a volcanic chain in central Tohoku appears to represent postseismic deformation from the 2011 off the Pacific coast of Tohoku earthquake. Adjusting the spatial scale of the high-pass filter enables the detection of not only displacements at individual stations but also broader localized deformations. Figure 3 depicts displacement distributions for (a) the decade prior to November 2010 and (b) the period after November 2014, using a larger filter size of 100 km.
Before 2010, displacement patterns were scattered, reflecting localized phenomena such as individual instances of subsidence or landslides, with no systematic trends. After 2014, however, systematic displacements were observed across the Tohoku region. Along the volcanic chain, three notable patterns were identified in Fig. 3 (b):
N: South-southwest-oriented displacements
S: North-northeast-oriented displacements
C: Eastward displacements
These patterns are likely manifestations of postseismic deformation from the 2011 Tohoku earthquake, where fluid movements in the deep subsurface were observed at the surface.
This study demonstrates that by adjusting the spatial size of the high-pass filter, localized displacements—often obscured by large-scale crustal deformation—can be effectively isolated and analyzed.
References
Fujiwara (2024a) Detection of local displacement of GEONET stations using spatial high-pass filtering, Preprints of JpGU 2024
Fujiwara (2024b) Detection of local displacement of GEONET stations using spatial high-pass filtering (2), Proceedings of the 142nd Conference of the Geodetic Society of Japan
By applying high-pass filtering to the spatial distribution of displacements observed at the GEONET stations of the Geospatial Information Authority of Japan (GSI), it is possible to remove large-scale crustal deformation and extract localized displacements. These include phenomena such as landslides and abnormal displacements caused by factors like GNSS signal shielding from trees (Fujiwara, 2024a). The displacements often align with the slope of the terrain where the GEONET station is situated (Fujiwara, 2024b). In this study, we present a nationwide analysis of localized displacements, employing high-pass filtering, and an example of local displacements for postseismic deformation following the 2011 off the Pacific coast of Tohoku earthquake using a larger spatial filter scale.
2. Selection of Filter Spatial Scale
The selection of an appropriate spatial scale for the high-pass filter is crucial for achieving the desired analytical objectives. As illustrated in Figure 1, the average distance between GEONET stations is approximately 15 km, with most distances being under 30 km, except for remote islands. Therefore, using a spatial filter size of 30 km effectively isolates local anomalies relative to neighboring stations while minimizing the influence of large-scale crustal deformations.
3. Nationwide Analysis
Figure 2 presents the results of high-pass filtering applied to displacement data of a 10-year period ending in November 2024. This analysis excludes coseismic displacements associated with major inland or nearshore earthquakes. These events produce large spatial variations over short distances, which cannot be fully removed by high-pass filtering.
Consistent with previous studies (Fujiwara, 2024ab), localized displacements attributable to landslides, subsidence, and other ground movements are prominent. The analysis also captures local earthquakes and volcanic activity. Landslides are observed nationwide. Despite filtering out both coseismic and short-term postseismic deformation from the 2016 Kumamoto earthquake, complex longer-term postseismic variations remain as localized anomalies. In contrast, the postseismic deformation from the similarly sized 2024 Noto earthquake does not exhibit such localized characteristics.
Volcanic activity is another notable source of displacement, particularly in eastern Hokkaido, the Izu Islands, Hakone, and Sakurajima.
4. Local Displacement in the Postseismic Deformation of the 2011 Tohoku Earthquake
In the Tohoku region, north-south subsidence along a volcanic chain in central Tohoku appears to represent postseismic deformation from the 2011 off the Pacific coast of Tohoku earthquake. Adjusting the spatial scale of the high-pass filter enables the detection of not only displacements at individual stations but also broader localized deformations. Figure 3 depicts displacement distributions for (a) the decade prior to November 2010 and (b) the period after November 2014, using a larger filter size of 100 km.
Before 2010, displacement patterns were scattered, reflecting localized phenomena such as individual instances of subsidence or landslides, with no systematic trends. After 2014, however, systematic displacements were observed across the Tohoku region. Along the volcanic chain, three notable patterns were identified in Fig. 3 (b):
N: South-southwest-oriented displacements
S: North-northeast-oriented displacements
C: Eastward displacements
These patterns are likely manifestations of postseismic deformation from the 2011 Tohoku earthquake, where fluid movements in the deep subsurface were observed at the surface.
This study demonstrates that by adjusting the spatial size of the high-pass filter, localized displacements—often obscured by large-scale crustal deformation—can be effectively isolated and analyzed.
References
Fujiwara (2024a) Detection of local displacement of GEONET stations using spatial high-pass filtering, Preprints of JpGU 2024
Fujiwara (2024b) Detection of local displacement of GEONET stations using spatial high-pass filtering (2), Proceedings of the 142nd Conference of the Geodetic Society of Japan