The most significant earthquake sequence in North Iceland for almost 50 years

The Tjörnes Fracture Zone (TFZ) in North Iceland links the offshore Kolbeinsey Ridge to the onshore Northern Volcanic Zone. About 98% of the seismicity in the TFZ concentrates along the Húsavík-Flatey Fault (HFF), a ~120-km-long right-lateral strike-slip fault, and the Grímsey Oblique Rift (GOR), an oblique rift with associated volcanism. Many earthquake swarms have occurred in the TFZ during the past decades. A notable swarm began on June 19, 2020, on the HFF and lasted until October 2020. It included three earthquakes of magnitude 5.4, 5.7 and 6.0, making it the most energetic earthquake sequence in North Iceland since the 1976 M6.2 Kópasker earthquake. The seismicity began on the western HFF and propagated to the northwest and southeast along the HFF. The first large event (M5.4) occurred on June 20 at 15:05h and was followed by the M5.7 earthquake four hours later at 19:26h. While the earthquake activity on the western HFF increased significantly, these events and their aftershocks also show a distinct N-S lineation, roughly perpendicular to the HFF, indicating that a conjugate fault may have been activated. The third large event and the largest of the sequence (M6.0) struck on the following day at 19:07h. It occurred at the eastern shoulder of the N-S oriented Eyjafjarðaráll Rift (ER), which is the southern continuation of the Kolbeinsey Ridge. This M6.0 event is the largest normal faulting earthquake recorded in Iceland, and thus constraining the parameters of this sequence is crucial for seismic hazard assessment in the region. However, the offshore location of these events makes this task challenging. Here, we attempt to overcome this challenge by combining waveform data from broadband stations up to 4,000 km away from the earthquakes and co-seismic GNSS offsets from 20 continuous and 10 campaign stations in North Iceland to estimate the earthquake source parameters, utilizing Bayesian inference. An initial double couple source was estimated from waveform data only as a prior for the joint geodetic and waveform data inversion. The M5.4 and M5.7 events correspond to strike-slip faulting on a WNW-ESE right-lateral fault or a N-S left-lateral fault, while the M6.0 event indicates normal faulting on a N-S fault. Our preliminary results show that incorporating GNSS data yields shallower depths than with waveform data alone (7.9 km for the first event, 5.1 km for the second event, and 10.2 km for the third event). However, previous results from moment tensor inversions of local broadband data indicate shallower depths for the event in the ER than for the strike-slip earthquakes, emphasizing the difficulty in constraining the depths of these offshore events. For their magnitudes, the first event was estimated as M5.3, with and without geodetic data, and the latter two events were found slightly larger in the joint inversion (M5.8 and M6.1) than in the seismic data solution, indicating limited aseismic moment release. These findings highlight the importance of combining diverse datasets to improve our understanding of complex tectonic settings and to constrain the proportion of seismic/aseismic moment release in earthquake sequences.