10:00 AM - 10:15 AM
[SVC32-14] Reanalysis of tilt changes associating with the 1986 Izu Oshima flank fissure eruption: Verification of the dike intrusion process based on tilt data
Keywords:Izu Oshima , tilt, dike intrusion
1. Introduction
The Izu-Oshima Volcanic Alert Levels, which consider fissure eruptions, assume that the area where crustal deformation sources originate can be identified before an eruption. However, real-time analysis can be difficult, and rapid eruptions may leave little time for assessment. Identifying key indicators for issuing alerts based on observed data is essential. To achieve this, we must clarify how geodetic data reflect dyke intrusion leading to fissure eruptions using past cases.
For the November 1986 fissure eruption, a NW-SE dyke intrusion model (Hashimoto & Tada, 1990; Linde et al., 2016) explains seismic activity distribution (Yamaoka et al., 1988) and crustal deformation. Tilt data from three stations showed changes starting two hours before the eruption, lasting about a month. These changes likely reflect dyke extension, and we therefore examined this in this study.
2. Data
Tilt data from three stations—GJK, HAB, and OWS—were used. GJK and HAB, installed by the National Research Institute for Disaster Prevention, were digitized from Shimada et al. (1988) and Yamamoto et al. (1988). OWS, installed by the MRI, provided 30-minute data for analysis.
3. Characteristics of tilt changes
Tilt changes are described for three periods: about 2 hours before the fissure eruption (Period I), expansion of both the eruption fissure and hypocenter distribution (Period II), and expansion of only the hypocenter distribution (Period III).
Period I: At GJK, the polarity reversed from NE to E down (missing NS component), with a change of more than 50 μrad (focus①). Polarity reversal was also observed at HAB.
Period II: At GJK, the NS component reversed (focus②). OWS initially showed a large S down, shifted at 18:00, and then the W down became dominant (focus③).
Period III: At OWS, the tilt changed to SE down (focus④), and HAB shifted from SW to NE down (focus⑤).
4. Hypothesized dike Intrusion Process and Analysis
Based on dike parameters, hypocenter distribution, and propagation speed from previous studies, the following dike intrusion process was assumed, and tilt changes were calculated using the Okada model.
Common to all periods: A vertical plane (dip 90°, strike N50°W) passing near the B fissure eruption start point (139.3955°, 34.7343°).
Period I: A dike with a constant horizontal length and opening, including the B eruption start point, rises from 12 km depth to the surface (Fig. 1-I).
Period II: The two dikes—one at 12-4 km depth with a 400 cm opening, and the other at 4-0 km depth with a 150 cm opening—extend horizontally at a constant speed towards the northern end of C fissure.
Period III: The deeper (12-4 km) dike from Period II continues extending northwest direction for 10 km at 1 km/h (Fig. 1-III).
5. Results
Period I: The results calculated temporal changes of tilt at GJK (Fig. 2b) show that polarity reversal associated with dike rise (Focus①). Figures 2a and 2c show results for other points, where the polarity reversal at HAB is also qualitatively explained.
Period II: The results (dotted line, Fig. 3) show that deep-to-shallow horizontal northwesterly dike intrusion explains the polarity reversal of GJK’s NS component (Focus②). However, at OWS, the tilt bending and significant S down in the first half (Focus③) remain unexplained. Mannen (2006) suggests high plume heights during this period, indicating a high lava eruption rate and possible underground decompression. We applied a Mogi model 5 km below the B eruption start point to estimate eruption rate variation and its contraction (solid line, Fig. 3). This qualitatively explains the characteristics of OWS.
Period III: Deep-to-shallow horizontal northwesterly dike intrusion can qualitatively explain the SW down at the OWS (focus④) but not the polarity reversal of the HAB (focus⑤) (Fig. 3). A model of southeasterly dike intrusion was tested, but it resulted in a NE down, which contradicts the observations. Other factors need to be considered.
The Izu-Oshima Volcanic Alert Levels, which consider fissure eruptions, assume that the area where crustal deformation sources originate can be identified before an eruption. However, real-time analysis can be difficult, and rapid eruptions may leave little time for assessment. Identifying key indicators for issuing alerts based on observed data is essential. To achieve this, we must clarify how geodetic data reflect dyke intrusion leading to fissure eruptions using past cases.
For the November 1986 fissure eruption, a NW-SE dyke intrusion model (Hashimoto & Tada, 1990; Linde et al., 2016) explains seismic activity distribution (Yamaoka et al., 1988) and crustal deformation. Tilt data from three stations showed changes starting two hours before the eruption, lasting about a month. These changes likely reflect dyke extension, and we therefore examined this in this study.
2. Data
Tilt data from three stations—GJK, HAB, and OWS—were used. GJK and HAB, installed by the National Research Institute for Disaster Prevention, were digitized from Shimada et al. (1988) and Yamamoto et al. (1988). OWS, installed by the MRI, provided 30-minute data for analysis.
3. Characteristics of tilt changes
Tilt changes are described for three periods: about 2 hours before the fissure eruption (Period I), expansion of both the eruption fissure and hypocenter distribution (Period II), and expansion of only the hypocenter distribution (Period III).
Period I: At GJK, the polarity reversed from NE to E down (missing NS component), with a change of more than 50 μrad (focus①). Polarity reversal was also observed at HAB.
Period II: At GJK, the NS component reversed (focus②). OWS initially showed a large S down, shifted at 18:00, and then the W down became dominant (focus③).
Period III: At OWS, the tilt changed to SE down (focus④), and HAB shifted from SW to NE down (focus⑤).
4. Hypothesized dike Intrusion Process and Analysis
Based on dike parameters, hypocenter distribution, and propagation speed from previous studies, the following dike intrusion process was assumed, and tilt changes were calculated using the Okada model.
Common to all periods: A vertical plane (dip 90°, strike N50°W) passing near the B fissure eruption start point (139.3955°, 34.7343°).
Period I: A dike with a constant horizontal length and opening, including the B eruption start point, rises from 12 km depth to the surface (Fig. 1-I).
Period II: The two dikes—one at 12-4 km depth with a 400 cm opening, and the other at 4-0 km depth with a 150 cm opening—extend horizontally at a constant speed towards the northern end of C fissure.
Period III: The deeper (12-4 km) dike from Period II continues extending northwest direction for 10 km at 1 km/h (Fig. 1-III).
5. Results
Period I: The results calculated temporal changes of tilt at GJK (Fig. 2b) show that polarity reversal associated with dike rise (Focus①). Figures 2a and 2c show results for other points, where the polarity reversal at HAB is also qualitatively explained.
Period II: The results (dotted line, Fig. 3) show that deep-to-shallow horizontal northwesterly dike intrusion explains the polarity reversal of GJK’s NS component (Focus②). However, at OWS, the tilt bending and significant S down in the first half (Focus③) remain unexplained. Mannen (2006) suggests high plume heights during this period, indicating a high lava eruption rate and possible underground decompression. We applied a Mogi model 5 km below the B eruption start point to estimate eruption rate variation and its contraction (solid line, Fig. 3). This qualitatively explains the characteristics of OWS.
Period III: Deep-to-shallow horizontal northwesterly dike intrusion can qualitatively explain the SW down at the OWS (focus④) but not the polarity reversal of the HAB (focus⑤) (Fig. 3). A model of southeasterly dike intrusion was tested, but it resulted in a NE down, which contradicts the observations. Other factors need to be considered.