Joint interpretation of the size and time distributions of seismic activity around the Hengill triple junction (SW Iceland) between 1993 and 1999

A newly developed clustering algorithm is described which uses the Unified Scaling Law for earthquakes to identify correlations between events that lead to clustering. This algorithm is applied to a dataset from the Hengill triple junction in south west Iceland to separate spatially and temporally overlapping earthquake sequences. We show that the algorithm successfully identifies spatio-temporally clustered events. A search for certain patterns in the identified clusters is performed. These patterns emerged in numerical simulations of seismicity performed by Hainzl (2003) when viscous coupling was introduced in order to reproduce features of swarm earthquakes. Namely we look for a combination of increased Gutenberg-Richter b-value and decreased exponent of the waiting-time distribution α. This pattern, indicating the strongest influence of viscous coupling on the characteristics of the seismicity, could be localized beneath an extinct volcano approximately 5 km north of an area of recent crustal uplift.