Short Term Precursory Signal Detected for 2018 Sierra Negra, Galapagos Volcanic Eruption Using Seismic Ambient Noise

Abstract:

Sierra Negra volcano, located in the Galapagos Islands, erupted on June 26, 2018. This eruption followed 4.5 m of sustained intra-caldera uplift which had been detected since 2005 by a GNSS network and months of increase in the number and magnitude of seismic events. We use six months of data (April-September 2018) from 12 broad-band seismic stations deployed on the volcanic edifice and within the caldera to detect any changes in velocity associated with pre-, syn, and post-eruption processes. We perform the ambient noise Moving Window Cross Spectrum (MWCS) analysis to compute velocity variations. A two-month pre-eruption reference stack (April 1 – June 10, 2018) was compared to a 1, 5, 10, 20-day moving window to compute delay times. Sensitivity kernels were calculated and used to choose different filter bands to isolate the velocity change signal associated with geological structures and physical processes. To make sure that the comparatively small change in signal is in fact due to physical changes in the volcano, a visual inspection of the cross-correlations in the time domain was done to verify the presence of scatterers in the coda wave. Frequency-domain stability was done by computing the magnitude squared coherence function to determine which pairs to use in our analysis. A sharp decrease in seismic velocity starts on June 8 and lasts eighteen days until the day of the eruption (June 26) where it recovers its normal value. This signal is seen for a 0.3 – 1 Hz filter for all moving windows and is attributed to the physical changes in the shallow sill and overlying rocks. For station pairs crossing the NE part of the caldera, the magnitude of the velocity change reaches a minimum of -0.28% in 6 days. This crustal velocity variation seen days before the eruption constitutes a premonitory signal of the last Sierra Negra eruption, which we attribute to a change in state-of-stress. The detection of this premonitory signal has implications for eruption prediction and therefore in managing volcanic crisis in Galapagos.