Some insights into fault inversion resolution using real and synthetic geodetic data
Abstract
We examined surface deformation in the Imperial Valley of southern California, with the goal of further constraining the fault geometry and along-strike variations in slip behavior of the Imperial fault. The major difficulties facing geodetic studies in this area are 1: InSAR decorrelation due to extensive agricultural activity, and 2: the lack of GPS measurements at the spatial and temporal densities required to fully characterize behavior of the Imperial fault; therefore, we employ the Persistent Scatter InSAR (PS-InSAR) technique, that allows us to avoid the decorrelation problem and exploit numerous stable ground objects, which can provide useful deformation measurements to study the local fault system. PS-InSAR approaches estimate unwrapped deformation using spatial and temporal optimization methods on an irregular distribution of points, the robustness of which is therefore highly dependent on the spatial density of PS. To assess the validity of our PS-InSAR results, we conduct a series of synthetic tests to characterize the minimum spatial and temporal density of PS points that can potentially leads to ambiguous results. We validate the unwrapping algorithm employed by the Stanford Method of Persistent Scatterer (StaMPS) by examining sets of synthetic data exhibiting a range of spatial and temporal characteristics, which are intrinsically related to the robustness of the unwrapping results. We first attempt to unwrap sets of synthetic data with regular and irregular spacing and white noise. We examine the percentage of the successfully unwrapped PS points with respect to the level of noise and spatial and temporal densities of PS. We then add spatially correlated noise, and calculate the percentage of successfully unwrapped PS to assess the desired spatial and temporal density in order to recover the original signal. Lastly, we introduce synthetic deformation signals to the data to examine the spatial and temporal density and signal-to-noise ratio that are required to recover the signal. The second part of this study involves examining the resolution of fault behavior by a given PS distribution, which allows us to assess the confidence of our inversion. We first construct sets of faults with varying geometry and slip characteristics, we then unwrap each set of synthetic signals with different PS spatial densities and noise levels. We first search for the global minimum solution, i.e. best fitting fault geometry, using the neighborhood algorithm, and then invert for the best fitting slip distribution using weighted, regularized least-squares methods. The inversion solutions are compared to the initial model to determine the achievable resolution by a given set of PS, as well as the impact of any unwrapping errors introduced during the processing.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2011
- Bibcode:
- 2011AGUFM.G23A0843N
- Keywords:
-
- 1209 GEODESY AND GRAVITY / Tectonic deformation;
- 1241 GEODESY AND GRAVITY / Satellite geodesy: technical issues;
- 1243 GEODESY AND GRAVITY / Space geodetic surveys;
- 1294 GEODESY AND GRAVITY / Instruments and techniques