Seismic Amplitude Ratio Analysis of the 2014–2015 Bárarbunga‐Holuhraun Dike Propagation and Eruption

Magma is transported in brittle rock through dikes and sills. This movement may be accompanied by the release of seismic energy that can be tracked from the Earth's surface. Locating dikes and deciphering their dynamics is therefore of prime importance in understanding and potentially forecasting volcanic eruptions. The Seismic Amplitude Ratio Analysis (SARA) method aims to track melt propagation using the amplitudes recorded across a seismic network without picking the arrival times of individual earthquake phases. This study validates this methodology by comparing SARA locations (filtered between 2 and 16 Hz) with the earthquake locations (same frequency band) recorded during the 2014–2015 Bár arbunga‐Holuhraun dike intrusion and eruption in Iceland. Integrating both approaches also provides the opportunity to investigate the spatiotemporal characteristics of magma migration during the dike intrusion and ensuing eruption. During the intrusion SARA locations correspond remarkably well to the locations of earthquakes. Several exceptions are, however, observed. (1) A low‐frequency signal was possibly associated with a subglacial eruption on 23 August. (2) A systematic retreat of the seismicity was also observed to the back of each active segment during stalled phases and was associated with a larger spatial extent of the seismic energy source. This behavior may be controlled by the dike's shape and/or by dike inflation. (3) During the eruption SARA locations consistently focused at the eruptive site. (4) Tremor‐rich signal close to ice cauldrons occurred on 3 September. This study demonstrates the power of the SARA methodology, provided robust site amplification; Quality Factors and seismic velocities are available. Locating earthquakes usually implies picking phase arrivals ( P and S waves). Another technique called Seismic Amplitude Ratio Analysis (SARA) was recently introduced to locate them only by using the amplitude recorded at different pairs of seismic stations. However, this technique was never proven to be true. This study shows that the earthquake locations derived by SARA compares remarkably well with the locations of 30,000 seismic events triggered when magma migrated in the Icelandic crust prior to the 2014–2015 Holuhraun eruption. But the results also provide new insight into the magma dynamics that led to the largest eruption of the last two centuries in Europe. We show that ground vibration was continuously triggered during the 2 week perio