Sub-surface fault slip dynamics during the 2021 Reykjanes unrest (Iceland)

The dynamics of fault slip in the upper hundreds of meters of Earth’s crust has long been an open question, as their behavior differs from classical elastic dislocation models and their observation still raises challenges. Here, we analyze centimeter-scale ground resolution aerial optical images of the surface ruptures associated with the 8 Mw ≥ 5.0 sub-surface earthquakes that stroke during the Reykjanes seismo-tectonic unrest, starting on February 24, 2021, and ending with the start of an eruption at Fagradasfjall on March 19, 2021. For four major earthquakes, we apply a sub-pixel correlation technique of pre-, syn- and post-crisis aerial and drone orthomosaics to describe the displacement field on surface blocks. We find that surface offsets reached up to 50 cm, with almost pure dextral strike-slip in a NS direction. These orientations contrast with the overall NE-SW-oriented extensional structures originating from magmatic intrusions and appear as a bookshelf faulting system conjugated to the left-lateral strike-slip plate boundary, oriented ~N070.

On hard grounds (e.g.: lava flows), shallow ruptures reached the surface, reactivating pre-existing structures and displaying an en-échelon succession of hectometric-sized fractures. We believe these ruptures are representative of medium-sized faults behavior in the last few hundred meters of the crust. On soft grounds, however, the rupture was only betrayed by meter-sized en-échelon systems, evidenced by thousands of discrete sub-metric surface fractures we were able to observe in the field and map from the orthomosaics. The sharp deformation gradient we imaged indicates that the dislocation drastically decreased above ten to a few tens of meters below the surface. In this layer, diffuse deformation takes on most of the slip deficit, mainly through inelastic processes. As a result, evidence of the February 2021 earthquake did not endure erosion for more than a few months. Except for an isolated sinkhole which allowed us to assume that one fault pre-existed, there were no markers of its presence before the earthquake. We emphasize that this issue must frequently lead to an underestimation of the seismic hazard when performed from surface traces.