Spatial and temporal evolution of Skaftá cauldrons floods from 2015 to 2021 through the analysis of correlograms

In September 2021, two jökulhlaups were released into the Skaftá river from western Vatnajökull icecap (Iceland). Such floods have been known since 1955. These jökulhlaups originate from two subglacial lakes under 1–3 km wide and 50–150 m deep depressions in the glacier surface, commonly referred to as ice cauldrons, formed by geothermal melting. The average time interval between jökulhlaups from each cauldron is ~2 a. The jökulhlaups travel ~40 km under a glacier that reaches maximum thickness of ~750 m and emerge in the Skaftá river at the terminus of Skaftárjökull outlet glacier. In addition to increased subglacial water flow and river discharge, these floods are responsible for seismicity and tremor onset, inside the cauldrons, along the subglacial channels, at the outlet where the subglacial channels meet the river, as well as on the river path.
We used seismic interferometry or cross-correlation of seismic noise to analyse data from 2015 to the end of 2021. We computed cross-correlation functions for 27 seismic stations and for frequencies between 0.5 and 8 Hz. To characterize these floods, we calculated the propagation velocities based on the cross-correlation functions and for each frequency band. We located seismic signatures both during the floods period and previous times by using a grid-search method based on the approach of Ballmer et al. 2013, which calculates theoretical differential times and provides probabilities of locations as the summed stack amplitudes of correlograms. These daily location grids allowed us to analyse the spatial evolution of probability of location during these floods as well as compare them with previous “non-flood” periods. Furthermore, based on these location grids, we were able to compute temporal series for isolated locations (pixels) such as ice cauldrons, a hydrological station located on the river path, or any other target, allowing us to analyse the temporal evolution of location probability during the floods as well as compare it with the last six years of data. The resolution of this temporal evolution ranges from monthly to hourly. By using six years of seismic data, we were also able to compare the 2021 floods with floods due to the same ice cauldrons, for example in August 2018 and September 2019.
This study provides an insight on how relevant seismic interferometry can be in the monitoring of such processes, with the purpose of being fully automatic in the near future.