Search

桜島火山において発生するハーモニック微動のモーメントテンソル解析を行った。BL型地震群発後に発生する微動(HTB型)と爆発的噴火直後に発生する微動(HTE型)のモーメントテンソル成分に大きな違いはなく，等方成分は50％以上，CLVD成分は20～30％，DC成分は20％以下であった。鉛直方向のダイポール成分が大きく，鉛直方向の力が優勢な震源が推定される。震源は火口直下の浅部であり，爆発的噴火発生前に火口底直下に形成されているガス溜まりが微動の発生に関与していると考えられる。, Moment tensors of harmonic tremors observed after BL-type earthquake swarms (HTB) and immediately after explosive eruption (HTE) were estimated. HTB and HTE were excited by source of dominant vertical dipole. The source depths of HTB and HTE correspond to the location of gas pocket formed at uppermost part of the conduit before explosive eruption. It was inferred that harmonic tremor was excited by vibration of resonance of the gas pocket.

火山性地震・微動を即時自動判別するシステムのプロトタイプを作成し，口永良部島火山を対象に試行を行った。分類には，基準観測点の震動継続時間，最大振幅，スペクトルなど波形を特長付けるパラメータに加えて参照観測点との震動開始時刻の差や振幅比を用いている。8ヶ月間の試行では，台風時の擾乱と島外で発生する構造性地震の識別が出来なかった。その原因は，主として，参照観測点の選定が不適切であったことによると考えられる。台風時の擾乱と構造性地震を除けば，ほぼ適切な分類がなされた。, A seismic data processing system for automated classification of volcanic earthquakes and tremors is developed, and preliminary experiment was done using seismic data at Kuchinoerabujima volcano. The outline of the system and the results and evaluation of the preliminary experiment are described.

桜島火山において発生するハーモニック微動について、群発性B型地震から推移して発生するもの（HTB）と爆発的噴火の直後に発生するもの（HTE）に分類した。両者のスペクトルは基本周波数とその整数倍にピークを持つ特徴がある。1990年7月20日に発生したHTBは基本周波数が1.4-1.7 Hzの範囲で一定であり，9つの高次モードが見られた。2002年10月11日11時15分の爆発的噴火の3分後に始まったHTEは，微動の初期の基本周波数は0.8 Hzであったが，時間が経つにつれ高周波側にシフトし，3.7 Hzまで増加した。HTEの高次モードも基本周波数の整数倍を保ちつつ高周波側にシフトした。両者の振動軌跡はRayleigh波が卓越しており，その回転方向は地震計を設置している深度が最も深いHAR観測点でprograde，他の観測点ではretrogradeであった。火口の南もしくは西の観測点では波形にLove波が見られた。HTB，HTEともに表面波の振動パターンは類似していた。, Eruptive activity with Vulcanian type has been continued at the summit crater of Sakurajima since 1955. Harmonic tremors have been observed several hours after B-type earthquake swarms or immediately after explosive eruptions. Here, we call the first one HTB (Harmonic Tremor following B-type) and second one HTE (Harmonic Tremor after an Eruption). Particle motions and spectra of HTB and HTE were analyzed. Spectra both of HTB and HTE have peaks at fundamental frequencies and higher frequencies of multiple integers of them. Peak frequencies of HTB were kept in a certain range. In contrast, those of HTE gradually increased. Spectra of HTB, which occurred on July 20, 1990, have fundamental frequencies ranging 1.4 Hz - 1.7 Hz stably and at least 9 peaks of higher modes. While the fundamental frequencies of HTE, which occurred 3 minutes after an explosive eruption at 11h15m on October 11, 2002, gradually increased from 0.8 Hz to 3.7 Hz. The peak frequencies of higher modes of HTE were also gradually increased according to shift of the fundamental mode toward higher frequency. Particle motion analysis showed that Rayleigh waves are dominant from prograde elliptical motion at the deepest borehole station HAR and retrograde motions at the other stations. Love waves were sometimes mixed at the stations south and west of the craters. The distribution patterns of Rayleigh and Love waves of HTB are similar to those of HTE.

近年水蒸気爆発が繰り返されてきた口永良部島火山の山頂部において、2000年8月より地磁気全磁力観測を行っている。火山性地震の震源域で熱消磁が起こった場合を想定して観測点を選定し、データは、オーブコムという衛星通信を利用して研究室へ転送している。今のところ火山活動に起因すると考えられるような変動は捉えられていないが、通信に関しては100%成功しており、小型・軽量・省電力のオーブコム通信を火口近傍の観測に利用する目処がたった。, Variations of geomagnetic total intensities have been measured since August 2000 around the summit crater of Kuchierabujima volcano where phreatic eruptions were repeatedly occurred. The measurement sites were determined so as to detect the change in total intensity due to thermal demagnetization at hypocentral zone of volcanic earthquakes. The observed data are transferred to the office at Sakurajima through satellite telecommunication service called orbcomm. No volcanomagnetic effect has been observed by now. Data transfer via the orbcomm service has been perfectly working, which suggests that this concise system with a small lightweight body and low power consumption could be utilized for observations close to the active crater.

火山噴火予知計画に基づく2000年度の集中総合観測において, 口永良部島火山の地震観測を2000年11月27目から2001年3月18目まで実施した。観測された火山性地震の波形は, P波, S波の位相が明瞭なA型地震, 6-30Hzの高周波振動が卓越する高周波地震, 2-4Hzの低周波成分からなる低周波地震, 特定の周波数の波動が卓越し, コーダ部分が一定の割合で緩やかに減衰する波形をもつモノクロマティック微動の4種類に分類される。高周波地震の震源は, 新岳火口の西縁付近の深さ100-400mに集中する。その震源域は, 噴気を伴う熱異常域に対応している。そのメカニズムは東西伸張の正断層型である。これらのことから, 高周波地震は, 極めて浅い震源をもつvolcano・tectonic地震とみなすことができ, 新岳西部の地熱活動に伴う岩石のせん断破壊により発生すると考えられる。また, 高周波地震は, その約40%が, 2-4秒前に1-2個の前駆微小地震を伴い, 連発する傾向がある。モノクロマティック微動の卓越周波数は6.2Hz付近にあるが, 1.1, 3.0, 4.8, 8.Ohz, また, 10-26Hzの高周波部分にもいくつかのピークが存在する.口永良部島に設置したすべての観測点においてスペクトルのピークは共通に見られ, ピーク周波数の波動は震源において励起されていると考えられる。これらのピークのうち, 1.1, 3。0, 4.8, 6.2Hzのピークはその帯域が狭く, シャープであり, コーダ部分において緩やかに減衰する。一方, 8Hzより高周波側のピークでは帯域は広く, そのエネルギーは, 初動から数秒間に集中しており, 長いコーダ部分ではほとんど振動がみられない。8Hz以上の高周波成分は高周波地震に特徴的な周波数成分であり, 高周波地震の発生の繰返しが, 1.1, 3.0, 4, 8, 6.2Hzにピークをもつような振動を発生するresonatorを励起させたものと考えられる。, A seismic observation was conducted at Kuchierabujima volcano, south off Kyushu, during the period from November 27, 2000 to March 18, 2001. We installed 5 seismic stations with 3-component seismometers and 3 seismic stations with short-period vertical seismometer (2Hz) around Shindake crater at the summit. Most of volcanic earthquakes observed at the volcano were high-frequency events with the spectra of 6-30Hz. The high- frequency events occurred at the western part of the Shindake crater with the depth of l00-400m. The epicentral area corresponds to the geothermal area at the western part of Shindake crater. The focal mechanisms of the high-frequency events were of normal fault type with east-west extension. These facts suggest that the high-frequency events were generated by shear fracture of rock in the geothermal area. Monochromatic events have a dominant spectrum peak at 8.0 Hz and have some sub-peaks at 1.1, 3.0, 4.8, 6.2 and some peaks higher than 10Hz. The amplitudes at the peaks of 1.1, 3.0, 4.8 and 6.2 Hz were attenuated slowly with lapse time from the initial motion. The energy at the peaks higher than 8 Hz was concentrated in the initial part of the events. It is inferred that higher frequency components of the events were reflected by repetition of high-frequency events, which acted as trigger to resonance of fluid body generating seismic waves with some peak frequencies lower than 7 Hz.

1998年8月7日飛騨山脈上高地付近において群発地震が発生した。衛星通信システムの導入によって周辺観測網のデータが得られるようになった最初のイベントであり, 8カ月間に10, 000個以上の震源を求めることができた。活動は, 上高地付近の東西に帯状の地域, 穂高岳から槍ヶ岳にかけての南北の帯状の地域及び野口五郎岳付近の南北に帯状の地域に発生し, これらの活動域を移動, 再帰する現象が見られた。決められたH・2以上の地震約300個のメカニズム解は北西-南東方向に主圧力軸をもつ横ずれ型が卓越し, 群発地震の南北, 東西の並びと調和的である。また隣接する焼岳の火山活動との関連が注目されたが, 火山活動を示す現象は観測されなかった。なお, 今回の活動に関するデータを, 地元岐阜県上宝村へ随時提供して災害対策の基礎資料とした。, An earthquake swarm occurred on August 7, 1998, at the Kamikochi in the Hida Mountain range, central Honshu.The activity continued more than 8 months including a moderate earthquake of M5.4. Some temporary stations were set by use of satellite conununication system near the source area, where seismological observation is very difficult because of steep mountains. Together with the data of routinely operated stations, more than 10, 000 events have been located at the Kamitakara observatory, Kyoto University for 8 months. The activity migrated 20km from Kamikochi towards north along the Hida Mountains and returned to the beginning area again at an average speed of about 1-2km/day. The focal depths are shallow as less than 7km, concentrating at about 3-5km deep. More than 300 focal mechanisms were determined from the directions of initial motions for events with magnitude as small as 2.0, on account of the dense observation network. Focal mechanisms of most events are strike-slip type with maximum pressure axes in northwest-southeast direction, which is consistent with the global stress in this region. The damage of the earthquake was not very severe, but real time information on the seismic activity was very useful for the measures by local government offices.

薩摩硫黄島は, 島の東部にある硫黄岳の山頂火口において活発な噴気活動を続けている.京都大学, 地質調査所, 福岡管区気象台により行なわれている地震観測, GPSの繰り返し測定および火口内の地形, 噴気の調査結果から, 1995年-1998年の薩摩硫黄島の活動は次のようにまとめられる.(1)薩摩硫黄島においても桜島火山と類似した種類の火山性地震A型, B型およびC型(モノクロマティック微動)が発生している.(2)1995年から1998年のA型地震の発生頻度には, 顕著な変化はみられず, 1975年-1978年における発生頻度と比較しても大きな変化はない.また, 正断層型のメカニズムをもつA型地震が火口の西側に震源決定されたが, その震源位置は1975年-1978年における発生位置と同様である.(3)1996年6月8日に発生したM=2.9の有感地震は, それ以外のA型地震と比べて規模が大きい.硫黄岳山頂火口近傍の測量基準点は, 1995年6月から1997年4月の間に6cm北東に変位した.また, 火口の南東側周辺に割れ目が形成された.この地震の発生は火口周辺の地盤変動および火口の南東における割れ目の形成と関連があるのかもしれない.(4)B型地震は, 山頂火口直下の浅い場所において発生している.その発生頻度は, 1998年7-8月には1日に15回程度であったが, 10月以降は, 1日150回程度に増加した.B型地震は体積膨張型のメカニズムをもち, 火山ガスの膨張と関連している可能性が高い., Fumarolic activity has continued at the summit crater of Satsuma-Iwojima volcano located south off Kyushu, Japan. Sakurajima Volcanological Observatory (SVO), DPRI, Kyoto University has conducted seismic observation at a permanent station 1.5 km west of the summit crater since June 1995. Fukuoka Meteorological Observatory installed 4 stations at the volcano since September 1997. Geological Survey of Japan has repeated visual, geochemical, topographical survey at the crater. SVO and GSJ have cooperated in GPS measurement since June 1995. Volcanic activity of Satsuma-Iwojima is summarized as follows; (1)Volcanic earthquakes are classified into A-type (high frequency), B-type (low frequency) and C-type (monochromatic tremor). (2) A-type earthquakes occurred at the west of the summit crater at a depth of 1km. The focal mechanism of the A-type earthquakes is normal fault type. (3) B-type earthquakes are located at very shallow depth beneath the summit crater. B-type earthquakes are generated by expansive volumetric source, maybe expansion of gas pocket. (4) The seismicity of A-type earthquakes during 1995-1998 is the same as that in 1975-1978, except the earthquakes at 21:06 on June 8, 1996. (5) A crack with length of lOOm was formed southeast of the summit crater before October 1996. The benchmark near the crack moved eastward by 6 cm during the period from June 1995 to April 1997. The earthquakes on June 8, 1996 may be related with topographic change and ground deformation around the crater.

インドネシア・ジャワ島西部にあるGuntur火山では, 最近約150年間噴火が発生していないが, 火山性地震の活動は活発であり, 山頂付近と西部のKamojangカルデラ付近に火山性地震が分布する。Kamojangカルデラ付近の地震は, 北東から南西方向に直線状に分布し, 震源の深さは5-10kmである。発震機構は南北伸張, 東西圧縮となる横ずれ型のメカニズム解となる。一方, 山頂付近の地震は, 火山列に沿って北西-南東方向に分布し, 震源の深さは5km以下と浅い。マグニチュードも1以下と小さい。正断層または逆断層型の発震機構をもつ。1997年5月から山頂直下の深さ2-5kmにおいて火山性地震の発生頻度が増加したが, 5月29日, 10月7, 8日のように発生頻度が高いときはすべて逆断層型となる。一方, 山頂の南2kmに設置された傾斜計により, 地震活動の活発化と同期して山頂方向の地盤の隆起が観測された。山頂方向の地盤の隆起現象は, 南東山麓の水準測量によっても確認されている。これらのことから, 1997年5月以降の地震活動は, 山頂部の火山列に沿う断層において発生している。また, 火口直下で圧力の増加があり, 圧力源の直上の地盤が押し上げられ, その結果として逆断層型の地震が発生したと解釈される。1997年5月-11月の地震活動度を1991年以降の地震活動の推移の中で評価してみた。地震活動の活発化は1992年5月-7月, 1993年5月-8月にもみられる。この3期間の活動の増加に共通してみられることは, 主として山頂直下において地震が多発したこと, また, Kamojangカルデラ周辺で活動が活発化したことである。1997年5月-11月における山頂直下での地震の放出エネルギーは1992年と1993年の活動と同程度であり, 今回の活動はGuntur火山において, 1-数年の間隔で繰り返される10^15-10^16ergの規模のエネルギーの放出を伴う地震活動の活発化の1つであるといえる。, Guntur is a volcano-complex, West Java, Indonesia and is located 35km SE of Bandung. Explosive eruptions and lava flows were repeated at Guntur Crater since 17th century, however, no eruption has occurred for 150 years after the 1843 eruption. Volcanological Survey of Indonesia began to monitor the volcano by a seismometer at a station 0.8 km east of the summit crater. In 1994, 3 seismic stations were added to determine hypocenters in collaboration with DPRI, Kyoto University. Volcanic earthquakes are located in the summit area (aligned from SE to NW along the volcano line from Guntur to G. Masigit) and Kamojang area west of the summit. In Kamojang area, earthquakes are distributed from Gandapura (NE) to Darajat caldera (SW) at depth of 5-10 kim, with the strike slip mechanism, similarly to tectonic earthquakes in inland area. In contrast, volcanic earthquakes beneath the summit are shallower (

京都大学防災研究所火山活動研究センターはインドネシア火山調査所と共同して, ジャワ島西部にあるグントール火山に4観測点からなる定常地震観測網を設置し, 1994年11月から観測を続けている。また, 1995年12月から1996年2月, 1996年9月から12月にかけて臨時地震観測を行った。その結果, 次のことが明らかになった。(1)火山性地震は, 山頂の火口直下の海面下深さ0-5kmに円筒状に分布する。また, 北西部のガンダプラカルデラ, 西部のカモジャンカルデラ周辺, 南部のプツリ山直下にも分布している。特に, 臨時多点観測による精密震源決定では, カモジャンカルデラ周辺の地震は深さ10kmまでに地表に現れた断層の延長線上に直線的に分布することが明らかになった。(2)カモジャンカルデラの周辺の地震の多くは, 節線が北東-南西, 南東-北西を向く横ずれ断層のメカニズムをもつ。一方, 山頂火口周辺では正断層および逆断層となる解が得られたが, 系統的なメカニズム解は得られなかった。(3)カモジャンカルデラ周辺の地震は, この地域に特徴的な北東-南西方向の断層の右横ずれ運動によって発生したものと考えられる。一方, 山頂火口周辺の地震は, 過去の噴火によって形成された火道周辺の岩石の様々な方向への微少破壊によるものと思われる。火口周辺の地震のマグニチュードの上限が0, 7であることは, このことと調和的である。, Guntur Volcano is an andesitic volcano, located in West Java, Indonesia. A seismic network composed of 4 stations has been installed at Guntur Volcano since October, 1994, in the collaborative study between Sakurajima Volcanological Observatory, DPRI, Kyoto University and Volcanological Survey of Indonesia, DGGMR, Indonesia. Temporary seismic observations were conducted in December 1995-February 1996 and September - December 1996.Locations of hypocenters were calculated assuming a homogeneous half space of Vp=2.76 km/s. The epicenters seem to be vertically concentrated beneath Guntur summit at depths of 0-5 kim, around Kamojang caldera (west of the summit), Gandapura (northwest) and G. Putri (south). The precise determination of hypocenter by temporary observation revealed that the hypocenters are aligned along faults at depth of 5-10 km.Focal mechanisms of 23 volcanic earthquakes were determined. Mechanisms of most of the events around Kamojang Caldera are generally strike-slip fault with the fault plane direction of NE-SW and NW-SE. Considering the hypocentral distribution and surface geological structure, it is inferred that the earthquakes originate by the slip of the fault to NE and SW. Mechanisms of four events were determined. The mechanisms are normal fault and reverse type. Composite source mechanism in the area did not show consistent character. This suggests that the rocks in hypocentral zones beneath the crater are fractured more in the many directions maybe due to the previous volcanic activities.The smaller magnitude (

山頂噴火活動に伴う桜島周辺の地盤の上下変動量を測定するため, 第9回桜島火山の集中総合観測の一環として, 1996年10月に1等水準測量の再測量を実施した。前回の桜島火山の集中総合観測時の水準測量は1991年12月に行われたが, その後1995年8月と1996年3月にも桜島内の一部の路線の測量が実施されている。1995年と1996年の測量によって, 1974年以来継続して沈降していた姶良カルデラの地盤が再び隆起を始めたことが測定された。これは, 姶良カルデラ地下に想定されているマグマ溜りに, 再びマグマの貯溜が始まったことを示す現象と解釈される。, To study vertical deformations ssociated with the eruptive activity, revised leveling surveys were conducted on May, August 1995 and October 1996 at Sakurajima and a part of leveling route around the Aira caldera.Comparing relative heights of each bench marks in 1991 and 1996, ground uplifts of about 4 cm were measured at the northern part of Sakurajima. Since 1993, eruptive activity has gradually declined and the Aira caldera resumed upheaval movements. This vertical displacement suggests that magma storage begins again at the magma reservoir beneath the center of the Aira caldera.

口永良部島火山に発生する火山性地震は, P波, S波の明瞭なA型, 高周波成分の卓越するHF型, 高周波成分と低周波成分の混在するHybrid型, 低周波成分の卓越するLF型および単一の周波数にスペクトルのピークを持つMonochromatic型に分類される。1995年以前は, 1ヶ月に10個以下のHF型地震が観測されていたが, 1996年1月初めの比較的規模の大きいA型地震とLF型地震の発生以後, HF型の発生頻度が高くなっただけでなく, 低周波成分の含まれるHybrid型, LF型およびMonochromatic型が発生するようになった。1996年5月から6月に, 西側山腹に4点, 山麓に6点地震計を設置し, 臨時地震観測を行った。島内における砕石発破を利用してP波速度構造を求めた。Vp=3.3㎞/sの半無限均質構造を仮定して震源決定を行ったところ, A型地震は口永良部島北東部に, その他の種類の地震は, 新岳の火口直下海抜Om付近に集中していることがわかった。LF型およびMonochromatic型の初動は, 全観測点において引きであり, 4象限型の発震機構で説明するのは困難である。, Volcanic earthquakes at Kuchierabujima volcano are classified into 5 types; A-type: tectonic-like earthquake with clear P- and S-phases, HF-type: dominated by high frequency, Hybrid-type: mixed with high and low components, LF-type: dominated by lower frequency, and Monochromatic event. Less than 10 per month HF-type events were observed before 1995. After occurrence of A-type earthquakes and LF events in early January 1996, the number of HF events suddenly increased and Hybrid and LF with lower frequency content and Monochromatic events frequently appeared. Temporary seismic observation was conducted in May and June 1996, installing 6 seismic stations around the summit crater and a small network with 4 seismometers on the western flank. Assuming Vp = 3.3 kim/s estimated from artificial explosions, locations of the earthquakes were determined. Two A-type earthquakes were located northeast off the volcano and the other types were concentrated beneath the summit crater at depth of sea level. P-wave first motions of LF and Monochromatic event were dilatation at all the station. The mechanism of the events may not be double-couple.

Relations between temporal variations in the atmospheric predictability and quasi-stationary (QS) states, such as weather regimes, is investigated both with low- dimensional dynamical systems and with a simplified atmospheric circulation model introduced by Legras and Ghil (1985).As a measure of the predictability, we use the Lorenz index a, introduced by Lorenz (1965), which gives an ensemble average of the perturbation growth rate for a prescribed time interval. As to QS states, we apply the concept established by Mukougawa (1988).In the one-dimensional dynamical system there is the obvious relation that a increases monotonically during the QS state.However, in the multi-dimensional dynamical system there is no obvious relation between α and the QS state. Considering the trajectory only on the attractor, each QS state has its own relationship. In some chaotic solutions of the Legras and Ghil model there is the same relation as in the one-dimensional dynamical system.

On July 13th. 1989, a volcanic eruption took place east off Ito City, Izu Peninsula. Micro-tremors of unusually large amplitudes with a frequency of about 1 Hz were associated with theeruption. We carried out a temporary observation of microtremors after the eruption, by twosets of array network.Our observation determined the origins of the microtremors to be near the eruption point(Teisi knoll), which hed been located beneath the sea. However the location of the origins seemedto vary from time to time within an area of several kilometers horizontally. This shows that themagma was also activated after the eruption not only near the Teisi knoll but also in adjacentregions.Besides the major portion of the microtremors (1 Hz), some monochromatic tremors ofabout 20 Hz were also observed.

The 2nd kind volcanic micro-tremors appear correspondingly to the volcanic micro-earthquakes.Strictly speaking, a conspicuous phase (C phase) can be seen at the place about ten seconds afterthe initial motion of a volcanic micro-earthquake and the appearance of this phase is simultaneouswith that of the principal part of the 2nd kind volcanic micro-tremor. Time intervals (C-P interval)between the initial motions and the C phases of the volcanic micro-earthquakes are related to theratios of the maximum amplitudes between the initial motions and the C phases to that of the Cphases. In order to explain these phenomena, the foci of the volcanic micro-earthquakes are deter-mined. The deeper a focus is, the shorter the C-P interval becomes.

Focal mechanism of volcanic earthquakes (A type) in Sakurajima volcano are studied. Mainresults are as follows:1) At the calm stage of volcanic activity, the focal mechanism is strike slip fault type.2) At the former period of active stage of volcanic activity, the focal mechanism is obliquelyslip normal fault type.3) At the middle period of active stage of volcanic activity, the focal mechanism is obliquelyslip normal fault type or reverse fault type.4) At the last period of active stage of volcanic activity, the focal mechanism is again strikeslip fault type or reverse fault type.

The hypocenters of the explosion earthquakes of Sakurajima volcano are determined by usingthe data obtained by telemetering and precise recording system.Examining the distribution of the hypocenters, the following results are obtained.1) The majority of the epicenters of the explosion earthquakes are located within the activecrater rim.2) The depth of foci of the explosion earthquakes are located from sea level to 2 km under the sealevel. There exists no relation at present case between the focal depth of the explosion earthquakeand the magnitude of the earthquake or air shock wave, and also the stage of the volcanic activity.

Volcanic earthquakes near the active crater of Naka-dake, Volcano Aso, can be dividedinto two kinds, high frequency events and low frequency ones. High frequency events aredominated by signals with higher frequency than 10 Hz and occur in the depth range, 0-3km below sea level. Low frequency events are located in 2 km range of shallow depth directlybeneath the crater bottom, 1.2 km above sea level.The low frequency events appeared mostly as earthquake swarm. The similar eventshaving similar waveforms were recognized among the low frequency events. In a series ofseismic activity, the similar event activity occurred at first and then was accompanied withnon-similar event activity.Ten groups of the similar events were classified in 1984-1986. It is supposed that thesimilar events are associated with the movement of magmatic material in the volcanic vent.

An equipment was deviced for the purpose of measuring the kinetic energy of the volcanic micro-tremors. The data obtained by this equipment for three years of 1971 to 1973 were analyzed fromthe point of view of the correlation between the emergence of micro-tremors and the rainfall, theseismic activity or the volcanic activity. The rainfall has a great influence on the emergence ofmicro-tremors. In case the rainfall is plentiful, the kinetic energy of micro-tremors, in most cases, is increasing. The earthquakes in and near Aso Caldera have an effect on the micro-tremors, butthe earthquakes far from the Volcano Aso seem to have no relation to the micro-tremors. As regardsthe volcanic activity, the kinetic energy of micro-tremors maintains a high level when the volcanicashes are ejected together with the volcanic gases and the water vapour from the crater.

An earthquake swarm including the main event (M=4.9) occurred only several kilome-ters east to the Daisen volcano, Tottori Prefecture from the end of June 1985. This swarmstrongly drew our attention, since it could be a precursory activity of the Daisen volcanism, or could be followed by a greater earthquake which would cause a landslide. From the dis-aster prevention point of view, fortunately the number of earthquakes gradually decreased andat the present time (1986 April) this swarm almost died out and no volcanic activity has beenreported until now.We have, however, summarized this swarm here, because the main event was a compara-tively large one that ever occurred in and around the networks of our institute. The mostinteresting nature of the swarm is that we could divide it into three stages, both from timesequence and epicentral distributions. We also studied the focal mechanism solution for themain event. It shows a similar pattern to those for the 1943 Tottori Earthquake and for the1983 event in the midst of Tottori Prefecture, which indicates that the present swarm iscaused by the NWW-SEE compressional force widely predominant in the Inner zone of theSouthwest Japan. The b-value was also calculated to be 0.65 for 168 events. Moreover, weset up 4 temporary seismic stations near the epicentral region, one of which was to clarifythe relation between the earthquake occurrence and the sound that had been reported by theresidents in the epicentral region. But we could not find the relation, because no moresound was heard during the temporary observation.

The swarms of volcanic micro-earthquakes called as B-type earthquakes have been themost typical forerunner for explosive eruptions at Sakurajima Volcano since 1955. Thephysical meanings of B-type earthquake swarm was examined in relation to ground deforma-tion and eruptive activity. B-type earthquakes were categorized into two types by theirpredominant frequency in this study. The one is named as a BL-type earthquake (lowfrequency:1-3 Hz) and originates at a part shallower than 3 km beneath the summit crater.The other is a BH-type earthquake (high frequency: 5-8 Hz) and the hypocenter ranges from2 to 3.5 km. The epicenters of both type earthquakes are located in the region of the summitwith an extent smaller than the size of the crater.BH swarms were observed mainly during inflation-stage with no significant changes oferuptive activity. On the contrary, BL swarms originates during the period of rapid deflationof the summit area with the successive emission of volcanic ash and gases, and they weresometimes accompanied with the intermittent ejection of volcanic bombs similar to Strom-bolian eruption and the emission of weak infrasonic waves.These results suggest that BH-type earthquakes may reflect the intrusion of magma inthe conduit, and that BL-type earthquakes are generated closely related with the degassingof magma extruded up to the crater bottom.

In the central Kyushu, the active seismic region is restricted in the western part. The easternside area of this region is the Aso and Kuju Volcanoes.Particularly, near the Aso Caldera, several earthquake swarms have been occurred. Thefault plane solutions of earthquakes in this area were obtained by the seismic network of AVL(Aso Volcanological Laboratory, Kyoto Univ.). The solutions indicate that the faults are as-sumed the northeast-southwest normal type or strike-slip type faults. All of tension axes arehorizontal in the north-south direction. These evidences (the stress-field is characterized byremarkable north-south extension) are confirmed in the seismic active region of the central Kyushu.

Amplitude anomaly of sesimic P waves was observed in the seismic observation networkaround Mt.Naka-dake, Volcano Aso. Maximum amplitudes of P waves were relativelysmall near the active crater of Mt.Naka-dake. Result of analysis taking account of raypaths suggested that P waves were attenuated in the crater area of diameter 1-2 km andat depth 1 or 2 km under the sea level. This region contains the focal zone of volcanicearthquakes.

At the active volcano, numerous volcanic earthquakes and tremors with several kindsoccur reflecting the conditions of activity of the volcano. In order to monitor, interpret andpredict the volcanic activity, an automatic data processing system for volcanic earthquakesand tremors is indispensable.From this view point, anautomatic data processing system with the following functionshas been developed: 1) Automatic event dteection and classification of volcanic earthquakesand tremors 2) Filing and display of the results 3) Evaluation and prediction of volcanicactivity 4) Display of the results of evaluation and prediction, and generation of alarm andcontrol signals.For the purpose of obtaining the most effective threshold value of classification parameters a concept of synthetic hit rate (SHR) is introduced. This is defined as SHR=(H/J)*(H/P), where H is the number of hit of judgement, J is the number of judgement and Pis the number of phenomena being considered. Applying this concept, evaluation of judgementconsidering miss and overlooked become possible.According to a test operation in 1986, SHR of classification is 0.8-0.9 in the case ofA type, B type and explosion earthquakes, and is 0.5-0.7 in the case of other types. SHRof H day (the day judged to be high level volcanic activity and may occur som explosions)is nealy 0.4. As to the number of explosions, the number of 88% of total explosions areoccurred in H day.

Studing on the seismicity around Sakurajima volcano, following preliminary results areobtained.(1) By the projection of focus on the vertical section perpendicular to the volcanic front, clear W-B zone can be perceived. This W-B zone is characterized by the sharp change ofinclination at the depth of 70km 80km. The inclination of Shallower zone is 40-50 degreeswheras about 70 degrees in deeper zone. The thickness of deeper zone is nearly 20km.(2) The depth of seismic surface directly beneth the volcanic front is 100km. Excludingthe shallow earthquakes whose focal depth are less than 20km, no earthquake is observedin the zone between the volcanic front and the seismic surface.(3) Comparing to other island arcs, seismic activity in South Kyushu is low in the depthof 8Okm-120 km whereas high in l4Okm-l6Okm.(4) Preceding to the increasement in number of explosion at Sakurajima volcano, high seis-mic activity in the depth more than 80km was observed.

Spectral study on the explosion earthquakes of Sakurajima Volcano was carried out. Asthe results of the specral analysis based on the source mechanism of Sharp's explosion type, andthe measurement of the amplitude of the air shock wave, the following knowledges are obtained;(1) The most peak frequency of the velocity spectra of the initial motion of the explosionearthquakes are about 2 Hz, and the radii of the cavity estimated from the peak frequency areabout 0.2 Km.(2) The maximum amplitude of the air shock wave of explosion is not depend on the di-mension of the cavity of the source, but on the pressure within the cavity.

Volcanic microearthquakes called "B-type" are closely related with explosive activity.Previous researchers devided B-type earthquakes into subclasses by the predominant frequen-cy. However, hypocenters and focal mechanism of the individual subclasses in B-typeearthquakes have not been determined because of the unclear onsets.In order to detect the initial motions clearly, seismometers were installed in the boreholesaround the active crater of Sakurajima Volcano. In this paper, B-type earthquakes whosedominant frequecies are 1-3 Hz and 5-8 Hz, are named as BL-type and BH-type earthquakes, respectively. The hypocenters and distribution of the initial motions of BL-type and BH-type earthquakes were examined. The results were summarised as follows.(1) Focal depths of BL-type earthquakes were determined at the depth of less than 2 kmbelow the sea level. BH-type earthquakes had deeper hypocenters than BL-type. Themaximum depth of BH-type was 2.5 km.(2) Most of BL-type earthquakes which occurred isolatedly and BH-type earthquakeshad all compressional arrivals. BL-type earthquakes which occurred in swarms had allcompressional or all dilatational arrivals, with the ratio of 61% to 35%. The initial motionsof the remaiders were mixed with compressions and dilatations. The distributions of initialmotions did not show that focal mechanisms of BL-type and BH-type earthquakes werequadrant types.

The physical basis of the thermal model comprising the mechanism of filteration of mantlefluids has been investigated and then it has been suggested that this model can be utilized onlyin such cases that the earth's crust is characterized by sufficiently high permeability of the rocks.The filteration mechanism was offered as a possible explanation of the extremely high valuesof surface heat flow observed in central Kyushu, especially in the volcanic and geothermal area, where deep-seated faults like the Oita-Kumamoto tectonic line and differential block-movementalong them have played a leading role in the late Cainozoic geological history, implying highpermeability of the crustal rocks. On the other hand, it was noticed that this mechanism couldnot be responsible for thermal processes accompanied by extensive folding of the crust, suchas regional metamorphism at the inversion stage of geosynclinal development, since in this casethe filteration of mantle fluids is restricted by low permeability of the rocks, which rather bringsabout some advective motion of the crustal layers as manifested by folding.

This paper is concerned with some characteristics of B-type volcanic earthquake swarms inrelation to the volcanic explosions in Sakurajima volcano.1) When B-type volcanic earthquake swarm precedes volcanic explosions, seismic waveformsbecome simpler or some peculier frequencies become dominant in Fourier spectra.2) Accumulated seismic energy of B-type volcanic earthquakes in a swarm is on a log scaleapproximately propotional to that of earthquakes of volcanic explosions which followed theearthquake swarm.

Observations of volcanic earthquakes with three components of seismographs and twosmall tripartite nets were carried out in August, 1962 and in July, 1963 at Sakurajima Volcano.Examining the wave-orbits and analyzing the apparent velocities and the directions ofapproach of waves, we obtained the results as follows: (1) The nature of origin of the observedearthquake is an explosive one, since the clear S-phase, which is to be found in tectonic seismic-waves, can not be observed. (2) The origins of the earthquakes are located beneath the summitcrater (Minami-dake). (3) The seismic waves are divided in two groups, the first group isP-waves having the apparent velocity ca. 2km/sec. And the second group is surface waveshaving their phase velocity ca.1 km/sec.. (4) The approaching directions of P-waves disagreewith those of surface waves. It seems that it is not only due to the crustal structure butrelated with the depth of earthquake origin. (5) Neverthless the lack of SH-motion, the surfacewaves of Love-type are found similarly to the case of the underground nuclear explosion.

Three dinamite explosions were fired in the caldera of Aso Volcano. One wasfor construction working in the caldera and the others for quarrying at the insidemargin of the caldera-rim. Their seismic effects were observed with the routine net, spanned in Aso area, and the tripartite net at the Aso Volcanological Laboratory, forvolcanic-earthquakes. In the case of Tochinoki Shot, the observation of seismicrefraction profile along the line from the shot point to the eastern tip over the centralcones was carried out. The travel-times of seismic waves, the particle motion ofground surface and the anomalous deflection of seismic ray in the caldera are discussed.The obtained results are as follows: (1) comparing to the apparent velocity (3.2 km/sec)obtained from 10km seismic refraction profile in the caldera, the value of V5'=4. 8km/secis obtained from 8km span in the somma region; (2) orbital motion of waves causedby artificial explosion is examined and SH phase is clearly seen; (3) no anomalousdeflection of seismic waves is observed at VoIc. Lab, in the case of artificial explosionsfired within the caldera and these phenomena favour Sassa's theory that anomalousdeflection of seismic waves occurrs at the caldera-rim; (4) the value of P-wave velocity, 2. 8 km/sec, which has been conventionally used for that of upper layer in the calderais firmly examined in these analysis.

During the recent active period of the Volcano Aso, of which the main eruptionhappened on October 31, 1965, the precise seismometric net was set around the craterand the observation of a swarm of volcanic earthquakes was carried out. The fociare found in a small area near the crater and the fact suggests these earthquakescorrelating immediately with the volcanic activity. Therefore the variation of thearea occupied by the foci with the increasing or decreasing of the activity is examinedto find a clue to the problem of forecasting the volcanic eruption. On the otherhand, the source mechanism of volcanic earthquake is discussed by examining the sign(push or pull) of the initial phase of each earthquake. The main results obtainedare as follows (1) The foci are concentrated in a small region northeastwards thecrater before the main eruption and are gradually scattered in wider region andtransferred somewhat southwards the crater after the main eruption and in the laterstage. The depths are found between 500 and 1, 000 meters. (2) On the maineruption, the fragments of rocks or lava are thrown into the southwestern parts fromthe crater pit, and the other hand, the region occupied by the foci of the volcanicearthquakes is located in the northeastern part from the crater pit. This relationenables us to imagine that the region occupied by the foci could be the uppermostmagma reservoir or the store-room of volcanic energy. (3) The foci of the earth-quakes occurred immediately before the main eruption occupy the volume surroundedby 500 x 500 x 500 meters and then the strain energy which can be stored in the volumeIs amounted to 10's ergs. On the other hand, the eruption energy estimated from themass of the ejected materials and the maximum distance from the crater pit to theplace where a ejected fragment is found, is amounted to 10's ergs. The coincidenceof the both values gives us a clue to forecast the degree of eruption. (4) Theexistence of characteristic earthquake of rarefaction type, that is, all first phases aroundthe epicenter are indicated by "pull", is found and these earthquakes are frequently foundin the earlier stage or before the main eruption but scarecely in the later stage or afterthe main eruption. The fact infers that the eruption of volcano is not simpLy explosivetype.

In order to make clear the location of hypocenter of volcanic earthquakes originating in the Sa-kurajima volcano, temporal seismemetric observation was performed. Eleven observation pointswere distributed around the crater. Analyzing sha data of March and April, 1970, the followingresults are obtained.1) The large majority of hypocenter of A type volcanic earthquakes are deep. The large psrtof hypocenters of the volcanic earthquakes which are not recognized clear P and S wave phase, areshallow, but there are some one which are deep.2) The seismicity at the crater rim is heighter than that within the crater.3) There is possibillity to know the new volcanic active place from the hypocenter distribution ofthe shallow earthquakes in the volcanic region.4) Hypocenter distribution which suggeste the existance of magma resovoir, is obtained.

昭和30年10月13月から, 新しい噴火活動を始めた桜島火山は, 9年を経た今日もなおかなり活溌な活動をつずけている。この火山の活動に伴う諸現象を観測し, 噴火予知の研究を推進するために, 昭和37年に京大桜島火山観測所が設立せられ, 高感度地震計による有線遠隔多点同時観測が始められた。そして火山性地震の性質とか, 噴火前の異常現象の一部については, この論文の第1報および第2報において記述されている。第3報においては, 噴火の前2時間ほどに, おびたゾしい地震の発生する例がいくつか得られたので, これらの地震群の性質および噴火との関係について述べる。, To study the prediction of volcanic eruption, seismic routin observation with high sen-sitive seismographs has been carried out since 1962 at Sakurajima Observatory, Kyoto Univ..On the seismograms of this observation, we found sometimes the tremendous earthquakeswarm in the case of the eruption. The frequency of these earthquakes became greater andgreater with time, and at last they appered in the continuous tremores just before the eruption.Otherwise, the maximum amplitude of them did larger and larger till the eruption. After theeruption, the frequency and the maximum amplitude either became smaller and smaller astime goes on. It seems that all of these earthquakes have their origins in quite near the crater.Owing to find such a tremendous earthquake swarm, we have been able to surely predictsome eruptions, which were accompanied with the detonation and the thrown out ejects.

Since eraly houre of the 29th of May, 1968, a great many felt earthquakes have occurred.So, the precise seismometric observation using the data-recorder and the other instruments werecarried out.The results of the investigation on these earthquakes can be summarized as follows:I) Supposing that the underground structure is homogeneous and having the averragevelocity of 2 Km/sec. Or 3 Km/sec. For P wave, the epicenters of these earthquakes are estimated tobe distributed from the center to the east part of Sakurajima and to be 2-15 Km deep.2) The push pull distribution of P wave of these earthquakes does not prove any regularity.3) The coefficient of Isbimato-Lida's empirical formula, m, is 1.8 which is equivalent to oneof the deeper zone earthquakes in the volcano.4) Since the occurrence of the eatbquake swarm, the surface phenomena of the volcano havenot shown the conspicious change, and the shallow zone earthquakes near the crater have notoccurred so many too.

In 1962, new observatory of Kyoto Univ. was constructed at Harutayama beingabout 2.5 km apart westward from the top of the Volcano Sakurajima. The gejphysicalobservations are carried out in this place to study the prediction of volcanic eruption ofthis volcano, which has been active since 1955.In this first report for the seismic observation, some characters of volcanic micro-earthquakes and of eruption earthquake are described, and the origin of eruptionearthquake and the relation between the volcanic micro-earthquakes and the eruptionare discussed.

Volcanic explosions at the Minamidake crater of Sakurajima Volcano are classified intosix types, according to the mode of the occurrence of the volcanic micro-earthquakes beforeand after the explosion. The main results are as follows;1) The same type explosions are often repeated. This may be explained that the each typeexplosion represents the condition of the volcanic vent and the energy supply from the reservosr.2) The type of the large explosions are almost The D IV or the D IV' type.3) After the formation of the lava-lake in the crater, the D I type explosions occurre frequent-ly. From the arrival time difference of the sismic wave and the sonic one, the origin of the D Itype explosion is seemed to be shallower than that of the other type explosions. This fact isascertained by the another observation also, that is the observation of the angle of the incidence ofthe explosion earthquakes. Moreover the process of the occurrence of the D I type explosion isconsidered.

In the first paper, it was reported that the increasemen'. Of the number of volcanic eart-hquakes was often observed before the explosive eruption, but that to predict the explosiveeruption by only its fact was difficult.In this second paper, to obtain the better method of the prediction of explosive erupti-on, some treatments were done about the seismic data.The main results are as follows.(1) Taking where N~ the number of the volcanic earthquakes in every an hour, and N the houraverage number calculated from the total number of the volcanic earthquakes in a day, then the value of a-N becomes large in the case of the explosive eruption.(2) By the comparison of the maximum amplitudes of the volacoic earthquakes obse-rved at three stations, it is presumed that the region of the origins of the volcanic earth-quakes may become wide in some days before the explosive eruption.(3) The volcanic earthquakes with the large amplitude are often observed in somedays before the explosive eruption at every three stations.Lft9, ft~t~. ~Q altｮ )I~・.'t4), ｮ O)~1~Uft~ 5