Seismic tomography of the Canterbury Plains and the geometry and evolution of seismicity of the Greendale fault system, New Zealand

The previously unknown Greendale fault produced the September 4, 2010 M 7.1 Darfield earthquake and later triggered the destructive February 22, 2011 M 6.3 Christchurch earthquake, as well as later M>5 aftershocks east of Christchurch. Surface rupture from the Darfield earthquake indicates up to 5 m of strike-slip motion along the main portion of the Greendale fault, while various geodetic and seismic models also indicate reverse faulting on surrounding smaller faults. We combine seismic data from a variety of sources (permanent network seismometers and strong motion instruments, temporary intermediate to broadband seismometers) to understand the geometry of these various sections of faults and the evolution of seismicity along them for the first four months of aftershocks from the Darfield earthquake. We identify 4 to 5 fault segments that were likely active in the Darfield earthquake and an additional 5 to 6 segments that were active during the study period, prior to the Christchurch earthquake. While relocating hypocenters, we also jointly invert for 3D Vp, Vs, and Vp/Vs in the Canterbury region using an extended version of the double-difference tomography code tomoDD (Zhang et al., 2009). In the area of the Greendale and associated faults, Vp, Vs, and Vp/Vs are generally reduced from the top 8 km of the average velocity model for the Canterbury region of New Zealand. from the surface to ~8 km depth, below which the resolution begins to decline. Beneath Christchurch and areas immediately to the south and west, Vp and Vs are elevated and Vp/Vs is reduced from the surface to ~8 km depth, corresponding to the location of a negative Bouguer gravity anomaly and an increase in depth to basement (Hicks, 1989). In the northwest portion of the model, Vp and Vs increase when approaching the foothills of the Southern Alps. There are no clearly defined features in the velocity model that cross or are offset by the Greendale fault and no apparent contrast in velocities across the fault, preventing us from conjecturing about the age or total offset along the fault over its lifespan based on the velocity models alone.