Geodesy [G]

G21A

CC:Hall E Tuesday 0800h

CC:Hall E Tuesday 0800h

Recent Progress, Innovations, and Applications in Synthetic Aperture Radar II Posters

G21A-01

Surface Displacements Near The Itoiz Reservoir, Navarra, Spain

The Itoiz reservoir, located in Navarra, northern Spain, is a newly constructed gravity dam that stores the water
from the Irati and the Urrobi rivers. The dam has a total height of 121 m, a total length of 525 m and a maximum
water storage volume of 410 hm^{3}. The aim of this work is to study the surface displacement field during
the impoundment of the Itoiz water reservoir. Satellite Radar Interferometry (InSAR) has been proven to be a
useful and powerful tool in tectonic areas for surveying subtle surface deformations over several years related
to geodynamic phenomena. A DInSAR observation technique, called the Coherent Pixel Technique, (CPT), is
being applied to study the existence of deformation areas in order to obtain mean velocities and time series of
deformation. In this study the stack used is composed by 22 ENVISAT ascending images concerning to the
period 2003-2008, 15 ENVISAT descending images concerning the 2004-2007 period and 50 ERS
ascendending images concerning the 1996-2008 period. We then compare the observation results with an
analytical model. Surface water loads and the associated displacements are computed by means of the
Boussinesq solution for a vertical point load on a homogeneous elastic half-space. The surface vertical forces
can be expressed by means of F(x,y,t)=ρ gsh(x,y,t) where ρ is the water density, g is the gravity
aceleration, s is the area of the spatial sampling rate from a digital elevation model of Itoiz area and
h(x,y,t) is the water column height that depends on the spatial location at surface (x,y) and the time t
during the reservoir impoundment.

G21A-02

An Inversion of Coseismic Displacements by Means of Faults With Free 3D Geometry. Application to 2003 Bam Earthquake.

We propose a new methodology to investigate the geometry of the fault system corresponding to a coseismic deformation field given though a SAR data set. For that we apply an inversion of displacements by means of a free 3D geometrical aggregation of elemental faults in a step by step process. For each step, a suitable scale factor allows to fit the data and the regularization conditions. The growth process continues to get unit value for the scale factor. We test the new methodology using three dimensional coseismic deformation maps coming from Envisat SAR data before and after the December 26, 2003 Mw 6.6 Bam earthquake in Iran to determine a geometry of faults.

G21A-03

Classification of Full Polarimetric SAR data based on Kernel Fisher Discrimination

One of the most important applications of full polarimetric SAR data is the classification of different terrain surface types. The Cloude-Pottier decomposition method is one of the best known polarimetric SAR analysis methods. It is an eigenvalue-eigenvector-based approach for understanding the scattering mechanisms of different land cover. However, the Cloude-Pottier analysis can lead to a noisy classification due to the fact that the zone boundaries in the entropy/alpha-angle plane, which discriminate between eight classes, are arbitrarily fixed, and do not necessarily represent boundaries of observed land cover classes. In this study, a new supervised classification method based on the entropy/alpha-angle plane and the Kernel Fisher discrimination technique is proposed. The Kernel Fisher discrimination (KFD) is a state-of-the-art nonlinear machine learning method. It aims to find the optimal projection matrix such that the distance between the mean values of the projected classes is maximized while each class variance is minimized. Training samples of the eight classes in the entropy/alpha-angle plane are used to calculate a projection matrix that better discriminates between these classes. Thus, the calculated matrix is used to project the entropy and alpha-angle images. A supervised classification of the projected eight classes is performed using the projected training samples. In this study, RADARSAT-2 and ALOS full polarimetric data are used.

G21A-04

Modeling Coulomb Stress Changes Associated With the 2005 and 2008 Qeshm Island, Iran Thrust Earthquakes Through InSAR Inversions

In this research, we examine the Coulomb stress changes that occurred between two earthquakes (Mw 5.9, Nov. 7, 2005 and Mw 6.0, Sep 10, 2008) on Qeshm Island in the Iranian Zagros Mountains. Coulomb stress change has been proposed as a possible mechanism that may advance or retard the timing of ruptures of adjacent faults following an earthquake event. While magnitudes of Coulomb stress changes are often small compared to total seismic stress releases, they may accelerate the timing of rupture on faults nearing their critical failure state. Using inversions of Interferometric Synthetic Aperture Radar (InSAR), we present various potential slip distributions for each event as well as a range of fault plane solutions that are consistent with the data. From our family of models of slip and orientation for each faulting surface we calculate the Coulomb stress change on the 2008 fault surface caused by the 2005 event using the program Coulomb 3.1. Preliminary results indicate that, depending on the models used, the region where maximum slip occurred during the 2008 event can be associated with either positive or negative Coulomb stress changes due to the 2005 earthquake. This suggests that further a priori constraints are necessary to better constrain fault orientations and slip distributions, or that other post-seismic recovery mechanisms are needed to accurately model the stress evolution of the 2008 fault plane and other potential faults within the region. We then build on our models of Coulomb stress drop through finite element models of dynamic stress evolution to explore the relationship between stress fields in bedrock and local salt domes.

G21A-05

Three-Dimensional Velocity Field of the Yellowstone Deformation from Ascending and Descending ENVISAT Observations