PRECURSORY SWARMS OF LONG-PERIOD EVENTS AT REDOUBT VOLCANO (1989-1990), ALASKA - THEIR ORIGIN AND USE AS A FORECASTING TOOL
PRECURSORY SWARMS OF LONG-PERIOD EVENTS AT REDOUBT VOLCANO (1989-1990), ALASKA - THEIR ORIGIN AND USE AS A FORECASTING TOOL
复制标题
DOI：
10.1016/0377-0273(94)90030-2
复制
发表时间：
1994-08-01
影响因子：
2.9
通讯作者：
POWER, JA
中科院分区：
文献类型：
--
作者：
CHOUET, BA;PAGE, RA;POWER, JA
在1989年12月至1990年4月的里道特火山喷发期间，阿拉斯加火山观测站根据与长周期（LP）地震事件（主要周期约为0.5 s）的偶发群发生相关的地震活动变化，发布了几次火山灰喷发的预警。12月14日的第一次喷发发生在23年的平静之后，并由23小时的LP事件群预示，这些事件随着喷发而突然结束。在接下来的几天里，经过一系列的通风口清理爆炸，圆顶于12月21日开始增长。另一个群，LP事件类似于第一个，开始于26日，并在1月2日以一次主要的火山喷发结束。接下来的两个星期里，火山继续喷发，然后停止，直到2月15日，当一次大规模喷发开始了一个长期的重复圆顶建设和圆顶破坏的阶段，持续到4月。在12月14日和1月2日的重大事件发生之前，气象部门发布了预警，但由于1月2日之后爆发序列仍在继续，群的能量下降，预测变得更加困难。在2月15日的火山爆发之前，有一个重要但不那么强烈的火山群，这是没有预测到的。这次喷发摧毁了火山建筑物上唯一的地震仪，并阻碍了预测，直到3月4日，当三个新站中的第一个安装在活动喷口3公里内时。从3月4日到4月21日序列结束，共发生了八次喷发，其中六次喷发之前都发生了可检测到的LP事件群。虽然很弱，但这些群为3月23日和4月6日爆发前发出的警告提供了基础。12月13日的初震具有以下特点：（1）持续时间短（2）在震群爆发的前18小时内，地震能量释放速度迅速加快，随后在爆发前5小时内活动减弱;（3）震级范围为-0.4至1.6级;（4）几乎相同的LP特征，主要周期接近0.5秒;（5）到处都是平行的第一次运动;（6）在火山口下1.4公里深处有一个固定的震源位置。这种长周期事件的发生表明了一种模型，涉及岩浆与地下水的相互作用，其中岩浆气体，蒸汽和水在整个群的固定点驱动一个固定的管道。这一系列事件的开始类似于连接较低的增压岩浆为主的储层和浅热液系统的减压阀的故障。Chouet最近发展的一个充满流体的振动裂纹的三维模型与地震资料相符合，并得出了LP源的下列参数：裂纹长度280-380 m，裂纹宽度140-190 m，裂纹厚度0.05-0.20 m，裂纹刚度100-200，流体声速0.8-1.3 km/s，裂纹宽度140-190 m，裂纹厚度0.05-0.20 m。岩石的纵波速度为5.1km/s，流体与岩石的密度比约为0.4，流体的体积模量与岩石的刚度比为0.03-0.07。充满流体的裂纹被脉冲压降间歇地激发，该脉冲压降的幅度在0.4至40巴的范围内变化。这种扰动似乎是一致的触发机制与阻塞流条件下的裂纹。
During the eruption of Redoubt Volcano from December 1989 through April 1990, the Alaska Volcano Observatory issued advance warnings of several tephra eruptions based on changes in seismic activity related to the occurrence of precursory swarms of long-period (LP) seismic events (dominant period of about 0.5 s). The initial eruption on December 14 occurred after 23 years of quiescence and was heralded by a 23-hour swarm of LP events that ended abruptly with the eruption. After a series of vent-clearing explosions over the next few days, dome growth began on December 21. Another swarm, with LP events similar to those of the first, began on the 26th and ended in a major tephra eruption on January 2. Eruptions continued over the next two weeks and then ceased until February 15, when a large eruption initiated a long phase of repetitive dome-building and dome-destroying episodes that continued into April. Warnings were issued before the major events on December 14 and January 2, but as the eruptive sequence continued after January 2, the energy of the swarms decreased and forecasting became more difficult. A significant but less intense swarm preceded the February 15 eruption, which was not forecast. This eruption destroyed the only seismograph on the volcanic edifice and stymied forecasting until March 4, when the first of three new stations was installed within 3 km of the active vent. From March 4 to the end of the sequence on April 21, there were eight eruptions, six of which were preceded by detectable swarms of LP events. Although weak, these swarms provided the basis for warnings issued before the eruptions on March 23 and April 6. The initial swarm on December 13 had the following features: (1) short duration (23 hours); (2) a rapidly accelerating rate of seismic energy release Over the first 18 hours of the swarm, followed by a decline of activity during the 5 hours preceding the eruption; (3) a magnitude range from -0.4 to 1.6; (4) nearly identical LP signatures with a dominant period near 0.5 s; (5) dilatational first motions everywhere; and (6) a stationary source location at a depth of 1.4 km beneath the crater. This occurrence of long-period events suggests a model involving the interaction of magma with groundwater in which magmatic gases, steam and water drive a fixed conduit at a stationary point throughout the swarm. The initiation of that sequence of events is analogous to the failure of a pressure-relief valve connecting a lower, supercharged magma-dominated reservoir to a shallow hydrothermal system. A three-dimensional model of a vibrating fluid-filled crack recently developed by Chouet is found to be compatible with the seismic data and yields the following parameters for the LP source: crack length, 280-380 m; crack width, 140-190 m; crack thickness, 0.05-0.20 m; crack stiffness, 100-200; sound speed of fluid, 0.8-1.3 km/s; compressional-wave speed of rock, 5.1 km/s; density ratio of fluid to rock, approximate to 0.4; and ratio of bulk modulus of fluid to rigidity of rock, 0.03-0.07. The fluid-filled crack is excited intermittently by an impulsive pressure drop that varies in magnitude within the range of 0.4 to 40 bar. Such disturbance appears to be consistent with a triggering mechanism associated with choked flow conditions in the crack.