11:00 〜 13:00
[SCG44-P14] A preliminary report on slow-to-fast / fast-to-slow inflation preceding recent volcanic eruptions in Japan
キーワード：傾斜変動、時間関数、噴火
Precursory tilt changes before eruptions have been recorded at various volcanoes and have been modeled as inflation near eruptive craters. For example, an accelerating (slow-to-fast) tilt change of 450 s long showing inflation beneath the crater region immediately preceded the 2014 phreatic eruption of Ontake volcano (Maeda et al. 2017). In case of the 2015 eruption of Hakone volcano, a decelerating (fast-to-slow) tilt was observed (Honda et al. 2018). Characterization of the temporal variation of the tilt changes is a key to identify processes that control the eruptions (e.g., Genco and Ripepe, 2010; Maeda et al. 2017). The author plans to analyze precursory tilt changes worldwide with a common approach to characterize the temporal variations, as a part of the slow-to-fast project (A3 group). In this meeting, we present preliminary results from recent eruptions in Japan, as the first step toward the worldwide study.
We analyzed eruptive and non-eruptive events at Ontake, Hakone, Kusatsu, and Kirishima volcanoes. We downloaded tilt records from V-net and Japan Meteorological Agency stations, and selected time windows in which clear, summit-uplift tilt was recorded. We applied a lowpass filter of 20 s, subtracted the first value of the selected time window, and calculated the absolute value of the tilt change from E-W and N-S components. Then, we investigated the ratio of 2nd- to 1st-order integrals at each time sample; this ratio was used to investigate a differential equation that controlled the tilt change at Mt. Ontake before the 2014 eruption (Maeda et al., 2017).
Results from the 2018 phreatic eruption of Kusatsu volcano showed an accelerating (slow-to-fast) tilt change of t1.5, where t is time, similar to magmatic eruptions at Stromboli, Italy (Genco and Ripepe, 2010). In case of the 2015 phreatic eruption of Hakone volcano, the tilt change first accelerated more rapidly than t1.5, and then decelerated with close to t0.5. A non-eruptive tilt change of Ontake volcano on July 20, 2015, showed a pattern similar to Hakone. At Kirishima (Iwoyama), a station extremely close to craters showed a strange pattern on Sep. 5, 2017; the tilt record first showed SE uplift (phase 1), followed by NE uplift (phase 2), which was further followed by NW uplift (phase 3). The tilt changes in phases 1, 2, and 3 were consistent with t1.5, t, and t0.5, respectively. No eruption followed this tilt change. Another non-eruptive tilt change with the same pattern was observed on Jan. 19, 2018. Then, on Apr. 19, 2018, a phreatic eruption occurred. immediately before the eruption, a tilt change composed of the three phases (SE, NE, and NW uplifts) was recorded; however, the phase 1 was approximated by t2, and the powers of t for phases 2 and 3 of this eruptive tilt were greater than those of non-eruptive tilt on Sep. 5, 2017 and Jan. 19, 2018.
These preliminary analyses highlighted technical issues needed to be addressed. For some events, the time periods of tilt changes and uplift/downlift patterns were inconsistent among stations, or even among components of the same station as in the case of Kirishima volcano. This situation makes it difficult to select an appropriate time window for the analysis. The 2nd- to 1st-order integral ratio is sensitive to the baseline correction and thus to the selection of the time window. Introducing an algorithm to automatically select an appropriate time window, and developing a method for characterizing the temporal variations less sensitive to the baseline corrections and time window selections, are keys to apply the method to volcanoes with more frequent eruptions.
Acknowledgments: We used tilt records from V-net (https://doi.org/10.17598/nied.0006) and Japan Meteorological Agency stations downloaded from MOWLAS (https://doi.org/10.17598/nied.0009). This work was supported by JSPS Kakenhi Grant Number JP21H05203.
We analyzed eruptive and non-eruptive events at Ontake, Hakone, Kusatsu, and Kirishima volcanoes. We downloaded tilt records from V-net and Japan Meteorological Agency stations, and selected time windows in which clear, summit-uplift tilt was recorded. We applied a lowpass filter of 20 s, subtracted the first value of the selected time window, and calculated the absolute value of the tilt change from E-W and N-S components. Then, we investigated the ratio of 2nd- to 1st-order integrals at each time sample; this ratio was used to investigate a differential equation that controlled the tilt change at Mt. Ontake before the 2014 eruption (Maeda et al., 2017).
Results from the 2018 phreatic eruption of Kusatsu volcano showed an accelerating (slow-to-fast) tilt change of t1.5, where t is time, similar to magmatic eruptions at Stromboli, Italy (Genco and Ripepe, 2010). In case of the 2015 phreatic eruption of Hakone volcano, the tilt change first accelerated more rapidly than t1.5, and then decelerated with close to t0.5. A non-eruptive tilt change of Ontake volcano on July 20, 2015, showed a pattern similar to Hakone. At Kirishima (Iwoyama), a station extremely close to craters showed a strange pattern on Sep. 5, 2017; the tilt record first showed SE uplift (phase 1), followed by NE uplift (phase 2), which was further followed by NW uplift (phase 3). The tilt changes in phases 1, 2, and 3 were consistent with t1.5, t, and t0.5, respectively. No eruption followed this tilt change. Another non-eruptive tilt change with the same pattern was observed on Jan. 19, 2018. Then, on Apr. 19, 2018, a phreatic eruption occurred. immediately before the eruption, a tilt change composed of the three phases (SE, NE, and NW uplifts) was recorded; however, the phase 1 was approximated by t2, and the powers of t for phases 2 and 3 of this eruptive tilt were greater than those of non-eruptive tilt on Sep. 5, 2017 and Jan. 19, 2018.
These preliminary analyses highlighted technical issues needed to be addressed. For some events, the time periods of tilt changes and uplift/downlift patterns were inconsistent among stations, or even among components of the same station as in the case of Kirishima volcano. This situation makes it difficult to select an appropriate time window for the analysis. The 2nd- to 1st-order integral ratio is sensitive to the baseline correction and thus to the selection of the time window. Introducing an algorithm to automatically select an appropriate time window, and developing a method for characterizing the temporal variations less sensitive to the baseline corrections and time window selections, are keys to apply the method to volcanoes with more frequent eruptions.
Acknowledgments: We used tilt records from V-net (https://doi.org/10.17598/nied.0006) and Japan Meteorological Agency stations downloaded from MOWLAS (https://doi.org/10.17598/nied.0009). This work was supported by JSPS Kakenhi Grant Number JP21H05203.