Reference frame access under the effects of great earthquakes: a least squares collocation approach for non-secular post-seismic evolution

The 2010, (mw 8.8) maule, chile, earthquake produced large co-seismic displacements and non-secular, post-seismic deformation, within latitudes 28 $$^{circ }$$ s–40 $$^{circ }$$ s extending from the pacific to the atlantic oceans. although these effects are easily resolvable by fitting geodetic extended trajectory models (etm) to continuous gps (cgps) time series, the co- and post-seismic deformation cannot be determined at locations without cgps (e.g., on passive geodetic benchmarks). to estimate the trajectories of passive geodetic benchmarks, we used cgps time series to fit an etm that includes the secular south american plate motion and plate boundary deformation, the co-seismic discontinuity, and the non-secular, logarithmic post-seismic transient produced by the earthquake in the posiciones geodésicas argentinas 2007 (posgar07) reference frame (rf). we then used least squares collocation (lsc) to model both the background secular inter-seismic and the non-secular post-seismic components of the etm at the locations without cgps. we tested the lsc modeled trajectories using campaign and cgps data that was not used to generate the model and found standard deviations (95 % confidence level) for position estimates for the north and east components of 3.8 and 5.5 mm, respectively, indicating that the model predicts the post-seismic deformation field very well. finally, we added the co-seismic displacement field, estimated using an elastic finite element model. the final, trajectory model allows accessing the posgar07 rf using post-maule earthquake coordinates within 5 cm for $$sim $$ 91 % of the passive test benchmarks.