Zoe K Mildon , Gerald Roberts, Joanna Faure Walker, Joakim Beck, Ioannis Papanikolaou, Alessandro Michetti, Shinji Toda, Francesco Iezzi, Lucy Campbell, Ken J.W. McCaffrey

Earthquakes are known to cluster in time, from historical and palaeoseismic studies, but the mechanism(s) responsible for clustering, such as evolving dynamic topography, fault interaction, and strain-storage in the crust are poorly quantified, and hence not well understood. We note that differential stress values are (1) output by calculations of fault interaction, and (2) needed as input to calculate strain-rates for viscous shear zones that drive slip on overlying active faults. However, these two separate fields of geoscience have never been linked to study earthquake clustering. Here we quantify the links between these fields, and replicate observations of earthquake clustering from a 36Cl cosmogenic study of six interacting active normal faults. We derive differential stress change values from Coulomb stress transfer calculations, and use these values in a viscous flow law for dislocation creep to calculate changes in strain-rate for shear zones, and slip-rates and earthquake recurrence on overlying active faults. Our quantification of clustering, verified with observations, reveals how brittle and viscous processes in the upper and lower crust interact, driving temporal changes in slip-rate and seismic hazard.