“InSAR Time Series and ALOS Data, Focusing on Ground Deformation in Cascadia”

Larger imageLauren pointing out the Cascadia region on a map of North America.

The Cascadia subduction zone is a convergent plate boundary between the oceanic Juan de Fuca plate and the continental North American plate, just off the northwestern coast of the United States. Recent GPS and seismic observations show evidence of slow or silent earthquakes in Cascadia. These slow earthquakes occur over several week periods with total surface displacements similar to those generated by a M~6 earthquake. Not much is known about the driving forces behind slow earthquakes, but they may play a role in reducing the energy released by large earthquakes.

Larger imageLauren working on InSAR processing.

We created interferograms using synthetic aperture radar (SAR) data acquired in the Cascadia region and further analyzed the data through time series inversions. We explored three methods for time series analysis of the Cascadia region: the Stanford Method for Persistent Scatterers (StaMPS), the Small BAseline Subset (SBAS) algorithm, and stacking. Since ground motion is not necessarily linear with time, single interferograms only provide partial information about the deformation behavior. Time series analysis allows us to take advantage of the existence of a number of interferograms and invert them to create a model of the process over the entire time period. This gives us a better idea of the time-dependent history of the deformation, as opposed to just the beginning and end result.

Larger imageInterferogram from 8 July 2007 to 23 August 2007

We also worked with Advanced Land Observation Satellite (ALOS) SAR data. ALOS is a Japanese satellite, launched in 2006, that uses L-band radar. L-band radar performs better than the more typical C-band radar wavelengths in vegetated areas like Cascadia. Working with ALOS data, we have produced interferograms using Repeat Orbit Interferometry PACkage (ROI_PAC), and made a few observations based on the radar images that have been acquired so far. It is worth noting that the digital elevation model (DEM) used in processing the interferograms was missing a few islands and land formations, and we also noted interesting patch-like patterns in the farming regions around Cascadia. Similarly, when all acquired ALOS scenes for a specific area were plotted in respect to baseline and time, we observed that the baselines appeared almost linearly aligned across time, which may be due to steering of the satellite to acquire large-baseline interferograms for vegetation studies. The processing scripts used to plot satellite baselines, mask data, and generate interferograms stacks were our own, and were written in Perl and Matlab.

View field notes from other SURE Interns:   Select an intern... Alex Wilk Alexandra Millar Christina Velasquez Hannah Waterhouse Joshua Soble Melissa Nunley Nathan Williams Patrick Dooling Trevor Thomas Tyanna Schlom