@article{NamikiRivaltaWoithetal.2018,
  author    = {Namiki, Atsuko and Rivalta, Eleonora and Woith, Heiko and Willey, Timothy and Parolai, Stefano and Walter, Thomas R.},
  title     = {Volcanic activities triggered or inhibited by resonance of volcanic edifices to large earthquakes},
  series = {Geology},
  volume    = {47},
  journal   = {Geology},
  number    = {1},
  publisher = {American Institute of Physics},
  address   = {Boulder},
  issn      = {0091-7613},
  doi       = {10.1130/G45323.1},
  pages     = {67 -- 70},
  year      = {2018},
  abstract  = {The existence of a causal link between large earthquakes and volcanic unrest is widely accepted. Recent observations have also revealed counterintuitive negative responses of volcanoes to large earthquakes, including decreased gas emissions and subsidence in volcanic areas. In order to explore the mechanisms that could simultaneously explain both the positive and negative responses of volcanic activity to earthquakes, we here focus on the role played by topography. In the laboratory, we shook a volcanic edifice analogue, made of gel, previously injected with a buoyant fluid. We find that shaking triggers rapid migration of the buoyant fluid upward, downward, or laterally, depending on the fluid's buoyancy and storage depth; bubbly fluids stored at shallow depth ascend, while low-buoyancy fluids descend or migrate laterally. The migration of fluids induced by shaking is two orders of magnitude faster than without shaking. Downward or lateral fluid migration may decrease volcanic gas emissions and cause subsidence as a negative response, while upward migration is consistent both with an increase in volcanic activity and immediate unrest (deformation and seismicity) after large earthquakes. The fluid migration is more efficient when the oscillation frequency is close to the resonance frequency of the edifice. The resonance frequency for a 30-km-wide volcanic mountain range, such as those where subsidence was observed, is ∼0.07 Hz. Only large earthquakes are able to cause oscillation at such low frequencies.},
  language  = {en}
}
@article{BindiCottonSpallarossaetal.2018,
  author    = {Bindi, Dino and Cotton, Fabrice and Spallarossa, Daniele and Picozzi, Matteo and Rivalta, Eleonora},
  title     = {Temporal variability of ground shaking and stress drop in Central Italy},
  series = {Bulletin of the Seismological Society of America},
  volume    = {108},
  journal   = {Bulletin of the Seismological Society of America},
  number    = {4},
  publisher = {Seismological Society of America},
  address   = {Albany},
  issn      = {0037-1106},
  doi       = {10.1785/0120180078},
  pages     = {1853 -- 1863},
  year      = {2018},
  abstract  = {Ground-motion prediction equations (GMPEs) are calibrated to predict the intensity of ground shaking at any given location, based on earthquake magnitude, source-to-site distance, local soil amplifications, and other parameters. GMPEs are generally assumed to be independent of time; however, evidence is increasing that large earthquakes modify the shallow soil conditions and those of the fault zone for months or years. These changes may affect the intensity of shaking and result in time-dependent effects that can potentially be resolved by analyzing between-event residuals (residuals between observed and predicted ground motion for individual earthquakes averaged over all stations). Here, we analyze a data set of about 65,000 recordings for about 1400 earthquakes in the moment magnitude range 2.5-6.5 that occurred in central Italy from 2008 to 2017 to capture the temporal variability of the ground shaking at high frequency. We first compute between-event residuals for each earthquake in the Fourier domain with respect to a GMPE developed ad hoc for the analyzed data set. The between-events show large changes after the occurrence of mainshocks such as the 2009 Mw 6.3 L'Aquila, the 2016 Mw 6.2 Amatrice, and Mw 6.5 Norcia earthquakes. Within the time span of a few months after the mainshocks, the between-event contribution to the ground shaking varies by a factor 7. In particular, we find a large drop in the between-events in the aftermath of the L'Aquila earthquake, followed by a slow positive trend that leads to a recovery interrupted by a new drop at the beginning of 2014. We also quantify the frequency-dependent correlation between the Brune stress drop Δσ and the between-events. We find that the temporal changes of Δσ resemble those of the between-event residuals; in particular, during the period when the between-events show the positive trend, the average logarithm of Δσ increases with an annual rate of 0.19 (i.e., the amplification factor for Δσ is 1.56 per year). Breakpoint analysis located a change in the linear trend coefficients of Δσ versus time in February 2014, although no large earthquakes occurred at that time. Finally, the temporal variability of Δσ mirrors the relative seismic-velocity variations observed in previous studies for the same area and period, suggesting that both crack healing along the main fault system and healing of microcracks distributed at shallow depths throughout the surrounding region might be necessary to explain the wider observations of postearthquake recovery.},
  language  = {en}
}