Key points• High-resolution bathymetry of the Gulf of Aqaba• Detailed map of a complex fault system including strike-slip and normal faulting• Rupture length of the main subevent of the Mw 7.3, 1995, Nuweiba earthquake and possible candidate for future earthquakes in the gulf.

The spatial distribution of large earthquakes in Slowly Deforming Continental Regions (SDCR) is poorly documented and, thus, has often been deemed to be random. Unlike in high strain regions, where seismic activity concentrates cyclically along major active faults, earthquakes in SDCR may seem to occur more erratically in space and time. This questions classical fault behavior models, posing paramount issues for seismic hazard assessment. Here, we investigate the M7, 1967, Mogod earthquake in Mongolia, a region recognized as a SDCR. Despite the absence of visible cumulative deformation at the ground surface, we found evidence for at least 3 surface rupturing earthquakes during the last 50,000 years, associated to a slip-rate of 0,06 ± 0,01 mm/yr. These results show that in SDCR, like in faster deforming regions, deformation localizes on specific structures. However, the excessive length of return time for large earthquakes along these structures makes it more difficult to recognize earthquake series, and could conversely lead to the misconception that in SDCR earthquakes would be randomly located. Thus, our result emphasizes the need for systematic appraisal of the potential seismogenic structures in SDCR in order to lower the uncertainties associated with the seismogenic sources in seismic hazard models.

<p>The north-south component of ground deformation remains difficult to derive from InSAR due to the limited sensitivity of standard InSAR observations in that direction. The new approach of burst-overlap interferometry (BOI) exploits swath overlaps of the Sentinel-1 TOPS acquisition mode to retrieve accurate north-south displacements. We applied time-series analysis to such along-track BOI observations of the roughly north-trending Dead Sea fault. Using a large number of Sentinel-1 images acquired from both ascending and descending tracks, we retrieved the horizontal displacement in the burst-overlap areas. Mis-registration errors caused by orbit errors, timing errors, or tropospheric delays are limited in burst-overlap velocities, and ionospheric delays can be reduced through spatial averaging, enhancing the surface displacement estimation. However, interferometric decorrelation is a challenge, as it degrades the co-registration performance in addition leading to fewer observations, particularly near the northern Dead Sea fault. By exploiting hundreds of images, we find a clear and consistent velocity change across different segments of the Dead Sea fault, using coherent distributed scatters optimized by integrating temporal coherence. Modeling of the ascending and descending BOI velocity results suggests that the fault-parallel velocity is in the range 4.2-5.0 mm/yr south of the Lebanese restraining bend, whereas only about half of that to the north of it. The results demonstrate the applicability of BOI time-series analysis in medium-to-low coherence regions with low deformation rates.</p>

<p>Despite multiple research efforts since the late 1950’s, many questions regarding the earthquake activity of the Dead Sea Fault (DSF) remain, in particular for its southernmost portion in the Gulf of Aqaba. This is due to its offshore location and little-known interactions with the Red Sea rift system. The emergence of the NEOM city-project in northern Saudi Arabia and the planned King Salman road crossing across the Gulf of Aqaba have made it important to find answers for these questions related to the earthquake hazard of the region. The last major earthquake in the Gulf of Aqaba occurred in 1995 along one of the main strike-slip fault segments in the gulf, bringing both extremities of the fault rupture closer to failure. Studies of the DSF have found that large events along the entire DSF cluster during relatively short active seismic periods lasting about 100-200 years, separated by longer quiescent periods of about 350-400 years. From a tectonic point of view, the time gap between 1995 and the previous major earthquake in AD1588 conforms to this scheme and suggests that the DSF might be ripe for a new earthquake sequence, with the 1995 earthquake as the starter. That said, new results from GPS and InSAR observations have pointed to possible fault creep in the southern part of the gulf, which would significantly decrease the seismic hazard in the area. To explore this possible creep and to test the clustering model, we investigate new sub-bottom profiling data acquired in December 2019 in the Gulf of Aqaba. We aim to map the extent of sand layers present in the different sub-basins of the gulf and to correlate them with seismoturbidite layers found in sediment cores collected in 2018. By looking at the geographic extent of these sand layers, we also aim to define the source of the coarse deposits, or at least, to determine whether they are related to the regular sediment influx or linked to turbidites generated by slope failures during large earthquakes. Our preliminary results indicate that the sub-bottom profiling data allow us to map sand layers up to a depth of about 8 meters. Considering a sedimentation rate in the gulf between 0.2 - 0.4 mm/year, we could be able to gain an overview of the sediment infill of the Gulf of Aqaba over the last 20 ky or more. Even if the resolution of the sub-bottom profiling data is lower than that of the sediment cores, and the assumptions made for the correlation of the sand layers, due to the scattered grid, do not help to constrain properly the source of the deposits, we can still propose a longer-term overview of the earthquake activity and discuss the temporal organization of the large events in the area.</p>