Systematic comparisons of earthquake source models determined using InSAR and seismic data

Robust earthquake source parameters (e.g., location, seismic moment, fault geometry) are essential for reliable seismic hazard assessment and the investigation of large-scale tectonics. They are routinely estimated using a variety of data and techniques, such as seismic data and, more recently, Interferometric Synthetic Aperture Radar (InSAR). Comparisons between these two datasets are frequently made although not usually in a comprehensive way. This review compares source parameters from global and regional seismic catalogues with those from a recent database of InSAR parameters, which has been expanded with 18 additional source models for this study. We show that moment magnitude (M-w) estimates agree well between the two datasets, with a trend for thrust events modelled using InSAR to have slightly larger M-w estimates. Earthquake locations determined using InSAR agree well with those reported in regional catalogues, with a median difference of 6.3 km between them, which is smaller than for global seismic catalogues. We also investigate the consistency of source parameters and source directivity by comparing ISC hypocentres with GCMT and ICMT centroid locations for earthquakes with M-w >= 6.5. In some cases the source directivity is qualitatively comparable with previous studies, especially when comparing ISC and ICMT locations. The average difference between InSAR-determined depths and those in the EHB catalogue is reduced if a layered half-space is used in the inversion of InSAR data. Overall, faulting geometry (strike, dip and rake angles) remain in good agreement with values from the GCMT catalogue, and any large discrepancies can be attributed to tradeoffs between parameters. With continued investment in satellites for radar interferometry, InSAR is a valuable technique for the estimation of earthquake source parameters. The observed trends and discrepancies between InSAR and seismically determined source parameters are the result of issues with the data, different inversion techniques and the assumed Earth structure model. (C) 2012 Elsevier B.V. All rights reserved.