Abstract
GPS observations since 1992 provide estimates of horizontal motion (surface velocity) at the 1 mm/yr level, or better, over much of New Zealand's South Island and between the interiors of the neighboring Australian and Pacific plates. In a few cases vertical velocities have been estimated with 1 mm/yr precision. Australia-Pacific Euler vectors from GPS are mildly inconsistent with those from longer-term (> 1 Myr) estimates, and predict 3 mm/yr faster deformation integrated across the plate boundary. We review the kinematic models that have been proposed to fit surface velocity data in central South Island. We extend the rotating, interacting, elastic block model of Wallace et al. [2006] to show that the central Alpine Fault is not presently slipping significantly between the surface and 12 km depth, consistent with the observed distribution of moderate seismicity and with occasional very large earthquakes on the Alpine Fault itself. We show that the block boundary faults in the central Southern Alps can be chosen to correlate spatially with high-conductivity zones inferred from magnetotelluric studies. We review numerical modeling studies which show that observed surface velocities can be used to constrain the width of present-day deformation in the lower crust to 30 km, and in the upper mantle to similar to 100 km. The GPS data are consistent with the doubly-vergent nature of deformation in many of these numerical models. We briefly review models that combine the long-term evolution of collision-related deformation and elastic seismic-cycle behavior; these provide insight into the relative contributions of elastic and inelastic deformation during the seismic cycle.