Limitations of Strain Estimation Techniques from Continuous GPS Networks

Using an interactive strain analysis tool we have studied several common classes of model-free strain estimation techniques given geodetic measurements. Without a-priori knowledge of the deformation field, our study reveals persistent and highly-structured artifacts in the strain field at and above the spatial scale of the sampling. These artifacts may be impossible to distinguish from real features. The implemented techniques are shown to miscalculate strain in areas of non-linear (including idealized arctan) deformation. Without knowledge of the non-linear characteristics of deformation, our analysis is limited to visualizing only large-scale strain. The current GPS networks, although providing unprecedented coverage of tectonic deformation, are still too sparse to reliably depict smaller strain features. The strain analysis tool, which will be on display, demonstrates this large-amplitude and highly-organized derivative error when given a dense and regular grid of synthetic observations from a variety of systems. We provide explanation for these artifacts and comment on other results in the literature.