PALSAR interferometry and elastic dislocation modeling
Last update: 1 March 2011 at 13:06 UTC
Feb 28th, 2011
This analysis has been performed in the frame of The International Charter on Space and Major Disasters through the French “Cellule d’Intervention et d’Expertise Scientifique et Technique” (CIEST)
Data are from the L-BAND PALSAR sensor on board the Japanese Space Agency (JAXA) Advanced Land Observing Satellite (ALOS)
PALSAR data courtesy of JAXA
, though CNES
(PI on the Charter Call). Data copyright belongs to JAXA, METI. We are thankful to JAXA and CNES for providing the date within the frame or the International Charter on Space and Major Disasters.
This work is preliminary. The results presented here are meant to evolve in the next hours and days as new data are acquired, and need validation from other techniques as well as advice from the New Zealand scientific geophysical community. For this aim, we are in contact with the GNS
group lead by Dr John Beavan. All official scientific communications about the earthquake will be issued by the Scientific Authorities of New Zealand.
Data processing was performed at the Bureau de Recherches Geologiques et Minieres (BRGM) for the CIEST group with the GAMMA processor.
We combined PASAR data acquired on 2011-01-11 and 2011-02-25 (UTC) from ascending orbit track 335 with line of sight (LOS) angle of 34°.3, to form a differential interferogram. Topographic contribution to the interferometric phase is modeled and removed by the use of a Shuttle Radar Topography Mission (SRTM) digital Elevation Model. The interferogram shows a pattern of fringes each of which represents 11.8 cm of surface displacement in the LOS direction.
Elastic dislocation modeling
Elastic Dislocation modeling was performed at Ecole Normale Superieure (ENS) with the code Inverse3-5.
We picked 210 points on the fringes where we could see and count them clearly (download kmz file
of the fringes picking). We made an inversion (single fault plane dislocation with uniform slip in an elastic half-space and least square minimization) using the LOS value at the sample points. We used values close the CMT Harvard solution for the angles, without inverting them: azimuth 65°, dip 65°, rake 135°.
Best fit parameters of the model are:
- Coordinates of the centre of the upper edge of the fault 43.536 S, 172.716 E, depth 2km
- Fault length and width: 8km, 8km
- Slip: 1.63m
- Inferred geodetic moment tensor: 3.13 1018 Nm
- Inferred centroid depth: 5.6km
The modeled fault (download kmz file
showing projection of the fault at the surface) does not reach the surface but ends 2km beneath the surface, this depth is well constrained in the inversion by the spacing of the fringes north and north-west of the fault.
The geodetic moment tensor is 60% higher than the moment tensor from the CMT Harvard. Nota: we were able to find solutions with reasonably good fit to the data and moment tensor of ~2.5 1018Nm by reducing the fault width to 6km and inferred centroid to 4.7km
Marcello de Michele: