COSMO-SkyMed Synthetic Aperture Radar Interferometry Observations of Volcano and Earthquake Processes
Fielding, Eric; Lundgren, Paul R.
Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, UNITED STATES

Volcano and earthquake processes occur over a significant range of spatial and temporal scales that require a similar broad spectrum of observations to better understand their source parameters and to improve hazard response and mitigation. The COSMO-SkyMed (CSK) SAR data have high spatial resolution and potentially short revisit times (1, 3, 4, 8, 9, 16 days) from the constellation of four CSK satellites.
Kilauea volcano is the most active volcano at the Hawaii hot spot, with continuous eruption of lava from the east rift zone (ERZ) since 1983. Kilauea has been the target of several synthetic aperture radar (SAR) satellites, as well as the NASA-JPL UAVSAR airborne SAR, each system suitable for differential InSAR measurements of surface deformation changes. Since late July 2010 CSK data from ascending and descending tracks have been acquired at Kilauea, with each track acquired at repeat intervals usually 8 days prior to the March 2011 and at shorter time sampling since. These data present the highest temporal sampling of any existing system (35, 46, and 11 day repeat interval for Envisat, ALOS PALSAR, and TerraSAR-X, respectively) and the possibility to produce higher resolution InSAR time serie to more tightly resolve short-lived transient deformation such as volcanic eruptions and inflation/deflation pulses. On March 5, 2011, the Kamoamoa fissure eruption began along the ERZ. It followed several months of pronounced inflation beneath Kilauea's summit and was the first dike intrusion into the ERZ since June 2007. The eruption began in the late afternoon of March 5, 2011 (Hawaii Standard Time; UTC-10:00 hrs) with rapid deflation beginning at Pu`u `O`o on the ERZ and about 30 minutes later at the summit. Magma from both locations fed the intrusion and an eruption that included lava fountaining along a set of eruptive fissures ~2 km in length between Napau and Pu`u `O`o. The eruption continued, jumping between fissure segments, until all activity terminated the night of March 9.
A rich InSAR data set exists for the Kamoamoa eruption from the COSMO-SkyMed (CSK), TerraSAR-X (TSX), ALOS PALSAR, and UAVSAR sensors. CSK data acquired on March 7 and processed that same day provided the earliest, quasi-real-time SAR data for this event. By March 10, when the eruption had ended, we had three CSK acquisitions and one ALOS scene acquired and processed. In all we have a number of satellite data (UTC dates) during and immediately after the eruption: ALOS March 6, 9, 11; CSK March 7, 10, 11; TSX March 11; from a mixture of ascending and descending tracks. UAVSAR airborne SAR data that were acquired in early May 2011 provide enhanced resolution and viewing geometries perpendicular to the satellite look directions. Following the eruption in March 2011 a pronounced post-diking transient is observed in the InSAR time series, most notably in the CSK data. In the case of the Kamoamoa eruption the pre-eruption inflation, co-eruptive fissure eruption, and the post-eruptive deformation time series demonstrate the important role that CSK data played in capturing both the detailed spatial pattern of the magmatic sources and the time-series behavior of the system. We present the InSAR data for these events and source models for the 2011 Kamoamoa fissure eruption and its pre-, and post-eruptive response and compare our modeling results to prior diking events in the ERZ .
We have studied coseismic and postseismic deformation for the sequence of earthquakes that have struck the South Island of New Zealand near the city of Christchurch using CSK and other SAR data. The Mw 7.1 Darfield earthquake was the largest (so far), rupturing several faults in the Canterbury Plain west of Christchurch on 4 September 2010 and causing some damage in Christchurch. A postseismic time series of CSK images maps the ongoing slip on some faults near the main rupture after the mainshock. On 22 February 2011, an Mw 6.2 aftershock struck beneath the city of Christchurch causing extensive damage in and around the city. We analyzed InSAR and pixel offset tracking or sub-pixel correlation on CSK SAR images to measure the coseismic deformation and assess damage due to the earthquake. The CSK pixel offsets provide two components of surface displacements even in areas not coherent in the interferograms. The high spatial resolution of the CSK stripmap images enhances the precision of the pixel-offset measurements. The CSK InSAR and pixel offsets, along with GPS data, indicate that at least two faults ruptured on 22 February or in the immediate aftershocks, with one very close to the city.
Another large aftershock struck Christchurch on 13 June 2011, with Mw 6.0, causing additional damage. Coseismic CSK InSAR, GPS data, and seismicity indicated that this fault rupture or ruptures were located to the east, offshore from Christchurch. CSK acquisitions imaged postseismic deformation after the June quake. Christchurch was hit yet again on 23 December 2011 with another large aftershock, this time measuring Mw 5.9, which caused some additional damage. Initial seismic locations for the December event were even further to the east, and we are now studying coseismic CSK interferograms to learn more about this event. The repetition of large magnitude ~6 events in the Christchurch area is complicating the InSAR time series analysis because of the discontinuities caused by each new event, but the CSK SAR time series will be essential for understanding the whole earthquake sequence.
A shorter sequence of large earthquakes struck in a sparsely populated area of southeastern Iran in 2010 and 2011, which fortunately caused little damage to human infrastructure. The first event was Mw 6.5 on 20 December 2010, for which we analyzed the ALOS PALSAR interferogram that was the only available satellite for the December coseismic deformation. After that event, we requested CSK acquisitions on both ascending and descending tracks to study postseismic deformation, which was interrupted by an Mw 6.2 aftershock on 27 January 2011. The coseismic deformation for the January event was imaged by CSK, ALOS, and TerraSAR-X (TSX) interferograms. Analysis of the two earthquakes showed that the Iran December and January events occurred on conjugate, perpendicular faults, a geometry similar to the Christchurch February and June quakes.
A large and damaging earthquake struck eastern Turkey north of the city of Van on 23 October 2011, with magnitude Mw 7.1. For this event, we are studying CSK InSAR and pixel offsets and interferograms from Envisat. Bayesian inversion of the InSAR, pixel offsets and GPS offsets indicates that the fault ruptured a blind thrust fault at depth beneath the mountains north of Van with little slip shallower than 6-8 km depth. Postseismic CSK and TSX interferograms show ongoing deformation on a number of faults in the area.