Seismicity and 3-D body-wave velocity models across the Hengill geothermal area, SW Iceland

Abstract

We image shallow crustal structures and analyze seismicity patterns in the Hengill high-enthalpy geothermal area in SW Iceland, exploiting a temporary densification of the seismic network 2018 to 2020. Using a subset of 6,300 high-quality manually picked P- and S-phases, we compute a minimum 1-D model for the region. Our results suggest that the most consistent and accurate hypocenter locations are derived from a joint inversion of P and S arrival times for the Hengill area. We demonstrate that this minimum 1-D model in combination with SeisComP detection and location algorithms can be used to produce fully-automated yet high-quality earthquake catalogs. Our analysis established that both the induced and natural seismicity in the Hengill area occurs in several distinct, spatially constrained clusters. In production and injection areas, the depth of the clusters is at about 2 km, near the bottom of the production and injection wells. These are most likely triggered by the injection and induced by the production, respectively. Outside of these clusters, the seismicity is generally deeper, with the depth of the deepest seismicity indicating the brittle-ductile transition zone. This zone is encountered at about 4 km near the center of the Hengill volcanic area and deepens with increasing distance from its volcanic center, to about 7 km in the southernmost region. A spatial analysis of b-values shows slightly increased values in areas with numerous injection wells and slightly decreased values in production areas. Three-dimensional crustal imaging of VP, VS, VP/VS shows a SE-NW trending fast velocity that extends, at 1–3 km depth between the extinct Grensdalur volcanic center and the presently active Hengill volcanic center. The fastest velocities are found in the NW corner of the Grensdalur volcanic center coinciding with a gravity high and probably reflecting dense solidified magmatic intrusion(s). This trend coincides with traces of geothermal surface manifestations, a shallow lying low resistivity anomaly and an aero-magnetic low. All these anomalies are caused by high temperature at some point in the geological history of the area and are most likely due to migration of the crustal accretion and volcanic activity between the two volcanic centers. Below-average VP/VS ratios at similar depth, coincide with the main production field. We suggest that this anomaly is caused by the extensive fluid extraction, which lowers the pore-pressure in the field and consequently increases the steam dominated zone, leading to lower Vp/Vs ratios. Most of the earthquakes are within the Vp/Vs low and at the boundary of the high and low Vp/Vs anomalies, which might indicate a region of good permeability.