Optimal fault resolution in geodetic inversion of coseismic data

Abstract

S U M M A R YWith the continued growth in availability of differential interferometry from synthetic aperture radar and GPS data, space based geodesy has been widely applied to image the coseismic displacement field and to retrieve the static dislocation over the fault plane for almost all the significant earthquakes of the past two decades. This is performed by linear data inversion over a set of subfaults, generally characterized by a constant and predefined or manually adjusted dimensions.In this paper, we propose a new algorithm to automatically retrieve an optimized fault subdivision in the linear inversion of coseismic geodetic data. The code iteratively keeps the parameter resolution close to a predefined high value. We first discuss the rationale supporting our algorithm and, after a detailed description of its implementation, we analyse the advantages of its introduction in the data inversion.The algorithm was tested against an exhaustive range of synthetic and real data sets and fault mechanisms. Among them, we present the results for the M w 6.2, 2009 L'Aquila (Central Italy) earthquake and compare the new and previously published slip distributions showing the disappearance of misleading slip pattern and the increased resolution for shallower zones.

The 18 August 2014 Mw 6.2 Mormori, Iran, Earthquake: A Thin-Skinned Faulting in the Zagros Mountain Inferred from InSAR Measurements

“…The inversion is done in two steps: first we invert Interferometric Synthetic Aperture Radar (InSAR) observations using the Bayesian inversion (Fukuda and Johnson, 2008;Faegh-Lashgary et al, 2012) to obtain the most probable fault geometry parameters of a single uniform-slip dislocation. There is a well-known tradeoff between the amount of slip and width of the rupture when constructing a fault model using geodetic data (Atzori et al, 2009;Atzori and Antonioli, 2011), and there are also tradeoffs between the rake and amount of slip when only one InSAR line-of-sight (LOS) measurement is used. The two ascending RADARSAT-2 interferograms have LOS vectors that only differ by 7°.…”

The 2011 Mw 7.1 Van (Eastern Turkey) earthquake

“…[23] Several approaches exist to assess the resolution in slip permitted from the use of surface displacements whereby the density of data sampling and the size of subfaults in the model are varied [e.g., Fialko, 2004;Lohman and Simons, 2005b;Page et al, 2009;Atzori and Antonioli, 2011]. We address the reliability of depth extent of faulting by first examining the RMS cost for different a priori fixed choices of upper top fault depths and assessing the trade-off with other fault parameters, and second by performing checkerboard tests using known slip distributions, forward modeling the resulting displacements at the data sample locations and using this as synthetic input to the slip inversion.…”

SAR Data Analysis in Solid Earth Geophysics: From Science to Risk Management

“…In fact, the InSAR data resolving power, i.e. the maximum detail level achievable on a source, strongly depends on the location of the observed points, as shown in [60] and not on the displacement field itself. The sampling can be manually customized by defining areas with different sampling density; this also allows to have a good control on the number of observed data to handle in the inversion (Figure 9b).…”