火山観測と噴火予知

抄録

Prediction of volcanic eruptions has been one of the major goals of scientific researches for a mankind. Short review is made from the view points of historical background and the present status. Modern scientific researches in volcanology were initiated in Japan, similar to those in seismology, in the late 19th century by the foreign scientists, who were invited by the government to help the quick modernization of Japan. Studies by Nauman and Milne are the examples. Instrumental observation of volcanic activities was initiated by Sekiya, the first Japanese professor in seismology, at the time of the cataclysmic eruption of Bandaisan in 1888. He brought a seismometer to the summit area and made a daily seismometrical observation. Omori’s studies in the early 20th century were the greatest milestone in volcanology. At the time of Torishima eruption in 1902, where all the 125 inhabitants were lost, Omori insisted firmly that we need a volcano observatory, which should conduct a continuous surveillance at the active volcanoes of Japan. After the successful prediction and well organized evacuation prior to the eruption of Usu in 1910, Omori was quite confident that 《the problem of prediction of great volcanic eruptions is, “in some cases”, not very difficult》. Omori also pointed out in 1911 that the volcanic energy is manifested essentially by pushing upwards the underground lava masses in a certain zone. The idea is again confirmed at the 1914 Sakurajima eruption through the seismic and geodetic data, and we now know that the idea is equivalent to the first attempt for the inflation-deflation model, the basic theoretical model today. Continuous monitoring of the semi-continuous eruptive activity at Asama in 1935-1961 gave a chance to Minakami for accumulating a knowledge for the eruption prediction in detail. Minakami proposed a prediction technique which involves A-type to B-type variation of volcanic earthquakes and the practical estimation of the eruption prediction probability based on B-type earthquakes. This technique coupled with the inflation modelling led us to the new era in early 1960’s for the eruption prediction study. The era, when many active volcanoes in Japan were started to be continuously monitored by Japan Meteorological Agency and hence, Japan was a leading country in volcano observations. The National Project for Prediction of Volcanic Eruptions in Japan started in 1974 for establishing a practical application of the prediction. Because of the extensive use of a new technique such as telemeters and CPU-aided observation and analysis systems, a few volcanoes are now well monitored by university observatories. The best example is the Sakurajima, where the newly installed tiltmeters and extensometers in a deep underground vault near to the crater show a consistent inflation before each summit eruption. Similar success in the prediction of volcanic eruptions was reported during the dome building eruptions at Mt. St. Helens during 1980-1982. Both are the special cases in which the eruptive activity continued quasi-continuously with similar repeated pattern. Regularly repeated earthquake swarms prior to the all seven historical eruptions of Usu give another patterns of pre-eruption process. The difference from the above cases may be only in the time scaling problems. Inflational deformation was luckily observed before 1977 eruption of Usu which was closely related to the occurrence of low-frequency earthquakes. Those regularities are believed to be caused by the highly viscous nature of dacitic magma. Even in the case of Miyakejima eruption, where basaltic magma erupted after a relatively long repose time, the minor seismic swarm was repeatedly monitored twice in 1962 and 1983, suggesting a high possibility of the successful prediction in the future if more adequate observations are available. Now we know a little better than

(View PDF for the rest of the abstract.)