Link between seismic velocity changes and the 2020 unrest in Reykjanes revealed by wavelength-dependent spatial stacking of distributed dynamic strain

The Reykjanes Peninsula, due to its active volcanism, has been subject of numerous studies. The eruption of Mount Fagradalsfjall in March 2021 followed a series of crustal inflation and subsidence cycles in 2020, as documented by GNSS stations. In this work, we monitor the crust in Reykjanes over a period of 5.5 months during the 2020 unrest with distributed dynamic strain sensing (also called DAS). At first, empirical Green’s functions are extracted through cross-correlation of ambient seismic noise (0.5 – 0.9 Hz). Subsequently, we apply coda wave interferometry to measure seismic velocity variations over time and space. We show that wavelength-dependent spatial stacking of DAS data prior to cross-correlation substantially enhances the time resolution and spatio-temporal coherency of measurements. Using this workflow, we reveal a compelling correlation between seismic velocity changes and both vertical and horizontal static ground deformation. This suggests a strong link between our measurements and geological processes. Measured velocity variations may be related to the infiltration of magmatic fluids into a shallow aquifer, although a comprehensive interpretation is impeded by the complexity of the tectonic setting, alongside increased seismic activity (>20,000 local earthquakes) and the existence of geothermal regions. Furthermore, discrepancies observed between results from causal and acausal sides of cross-correlations suggest that measurements are affected by the noise source distribution, which raises questions about the extent to which measurements are solely attributable to structural and dynamic changes of the crust. Our work not only demonstrates how the spatial sampling of DAS can be exploited to enhance seismic monitoring strategies, but also highlights conceptual limitations that need to be confronted in future investigations.