Monitoring a temporal change of seismic velocity in a volcano: Application to the 1992 eruption of Mt. Merapi (Indonesia)

Abstract

Multiplets, i.e. events with similar waveforms, are selected from shallow earthquakes recorded on Merapi volcano (Indonesia) before the eruption of February 2nd, 1992. Two multiplet families are found with their sources close to the summit. Their seismograms are analyzed using the Moving Window Cross Spectrum technique which measures the precise time delay between seismic phases in the entire seismogram. For both families of multiplets, a gradual decrease in the arrival times of coda waves is observed as a function of the date prior to the eruption: coda waves are becoming progressively faster (up to 1.2 per cent) as the time interval to the eruption shortens. This observation is interpreted as the consequence of an increase in the seismic velocity inside the volcano. The increase in velocity started in May 1991 and was observed until September 1991, 4 months before the eruption. This velocity increase may be related to an increase in pressure in the magma chamber or in the conduits and to the resulting closure of the surrounding cracks.

Locating a weak change using diffuse waves: Theoretical approach and inversion procedure

“…In seismology, the monitoring of temporal changes in the crust was initiated in the mid-80's, using repeating small earthquakes on faults [9]. Later on the method was applied to volcanoes and revealed temporal changes of velocity prior to eruptions [10]. The method was transposed to the laboratory and popularized under the terms diffuse acoustic wave spectroscopy (DAWS) [11], or codawave interferometry (CWI) [12,13].…”

“…In these approaches, changes of waveforms in the coda are interpreted in terms of travel time variations, a technique that is very sensitive [14] to detect weak changes, but gives little information concerning the location of the change. To first order, global velocity changes in the medium result in a stretching of the waveforms [10,[14][15][16] but the interpretation of a local change in terms of travel time fluctuation remains problematic. Recently DAWS has been used in damage monitoring [17,18] but a large range of other applications are possible [19].…”

Towards forecasting volcanic eruptions using seismic noise

“…Therefore, when computing a local time shift ∆τ between the reference and the current cross-correlations in a short window centered at time τ , we would expect to find ∆τ to be a linear function of τ ( (7,24). It can be seen in these measurements obtained just with one pair of receivers that the RTP changes from a negative to a nearly zero and then to a positive value 40 days, 14 days, and one day prior to the eruption, respectively.…”

“…The inter-eruptive processes also produce perturbations of the elastic properties of volcanic edifices. They can be detected as changes in the travel times of seismic waves propagating within the volcanic edifices by using coda waves from repetitive seismic sources such as multiplets (6,7,8) or small summital volcanic explosions (9, 10) or correlations of a diffuse seismic wavefield excited by long-period seismo-volcanic events (11).…”

A 3D algorithm based on the combined inversion of Rayleigh and Love waves for imaging and monitoring of shallow structures

“…In turn, such small perturbations can be detected as changes of seismic wave properties based on seismic noise (e.g. Ratdomopurbo & Poupinet 1995;Miller & Smith 1999;Grêt et al 2005;Wegler et al 2006). However, despite considerable efforts in the past, the precise monitoring of the location of subsurface changes has proven to be difficult, meaning that there is a need for novel observational methods to obtain information about the ongoing subsurface changes and their relation to volcanic processes.…”