SEISMIC INTERFEROMETRY AND VOLCANIC TREMOR LINKED TO EXPLOSIVE REGIMES

The ability to measure in real-time information on changes of the elastic properties of the medium or on location of the seismic signal source are key parameters to forecast active ongoing processes. Cross-correlating seismic signals recorded at two stations is possible to retrieve Green’s function. Depending on which part of seismic signal is analysed, the early part or the coda, it represents position of the source originating waves or elastic characteristics of the medium through which the seismic wave propagates. Here I propose the methodology based on deconvolution technique to retrieve the Green’s function. Results have been compared with other geophysical dataset in order to infer information on volcano dynamic showing a strong correlation with seasonal variations due to atmospheric temperature. This can be explained as due to partial changes in seismic velocity induced by the ambient temperature affecting the shallower layers. The strong correlation between interferometry and ambient temperature is a strong evidence of the high sensibility and reliability of the developed technique. Once this long-term effect has been removed, I correlate interferometric changes to volcanic activity and other geophysical parameters. The interferometry changes have been then filtered in two different frequency band which seem to be the best to perform evaluations of source position and medium characteristics variations at Stromboli volcano. According also to previous studies, given the recurrence of VLP signals in the seismic record, the delay-time and noise-coda waves on both frequency bands are strongly affected by local seismicity. Changes in the time lags between the green’s functions are more reflecting changes in the delay times between the seismic events rather than changes in seismic velocity. In this case, interferometric analysis is rather reflecting changes in the position of the source. Interferometry shows in fact a very good correlation with changes in the polarization dip of the seismic VLP, revealing a strong link to changes in relative position of the seismic VLP source. In the second instance, no clear evidence of substantial medium variations in all analysed time period have been found. Volcanic seismic tremor is a persistent ground vibration due to interaction between solid earth, atmosphere and hydrosphere. It consists of both volume waves and surface waves generated by movements of fluids, either gas or magma. Due to correlation to fluid dynamics of the volcano, seismic tremor seems to be an efficient tool for understanding the dynamics of the plumbing system. I have developed a new method to analyse seismic tremor by calculating the temporal periodicity of RMS amplitude variations using a Zero-crossing technique. Analysis of frequency of tremor amplitude show sudden and clear increases close to phases of intense activity related to major gas releases. In particular the tremor switched from harmonic to pulsating-spasmodic, with large amplitude variations about 10 days before the two paroxysms which occurred on 3 of July 2019 and 28 August 2019. This behaviour probably depends on magma flow and suggest that this technique can be a valuable predictor of intense eruptions. I extended both interferometry and tremor analysis over a 4 years long dataset on 5 seismic stations. I analysed and compared results with other available geophysical datasets like seismic VLP signals and acoustic pressure of explosions.