A framework for comparing geomechanical models of InSAR-measured surface deformation

Abstract

High-quality Interferometric Synthetic Aperture Radar (InSAR) surface deformation data for field sites around the world has become widely available over the past decade. Geomechanical models based on InSAR data occur frequently in the literature but few methods of systematically optimizing or comparing them are presented. This work discusses parameterization errors for simplified models of strike-slip, normal, thrust and reservoir-style faulting with the aim of identifying tests or characteristics that can differentiate between error types uniquely. Fault dip errors, slip errors and depth errors are modelled using a simple homogeneous elastic half-space earth model. Simple difference maps prove to be a powerful tool for identifying error types and parameter sensitivity, with gradient maps and gradient difference maps useful for distinguishing between similar cases. The fault dip proves to be more indicative of error resolving capability than the faulting regime; errors on intermediately dipping faults are very difficult to differentiate. More detailed modelling of compound errors, complex geomechanical properties and noisy data is proposed. The use of the tests as the starting point for an artificially intelligent modelling package is briefly discussed.