Co-seismic slip from the 1995 July 30 M_w=8.1 Antofagasta, Chile, earthquake as constrained by InSAR and GPS observations
We analyse radar interferometric and GPS observations of the displacement field from the 1995 July 30 M_w= 8.1 Antofagasta, Chile, earthquake and invert for the distribution of slip along the co-seismic fault plane. Using a fixed fault geometry, we compare the use of singular-value decomposition and constrained linear inversion to invert for the slip distribution and find that the latter approach is better resolved and more physically reasonable. Separate inversions using only GPS data, only InSAR data from descending orbits, and InSAR data from both ascending and descending orbits without the GPS data illustrate the complimentary nature of GPS and the presently available InSAR data. The GPS data resolve slip near GPS benchmarks well, while the InSAR provides greater spatial sampling. The combination of ascending and descending InSAR data contributes greatly to the ability of InSAR to resolve the slip model, thereby emphasizing the need to acquire this data for future earthquakes. The rake, distribution of slip and seismic moment of our preferred model are generally consistent with previous seismic and geodetic inversions, although significant differences do exist. GPS data projected in the radar line-of-sight (LOS) and corresponding InSAR pixels have a root mean square (rms) difference of about 3 cm. Comparison of our predictions of vertical displacement and observed uplift from corraline algae have an rms of 10 cm. Our inversion and previous results reveal that the location of slip might be influenced by the 1987 M_w= 7.5 event. Our analysis further reveals that the 1995 slip distribution was affected by a 1988 M_w= 7.2 event, and might have influenced a 1998 M_w= 7.0 earthquake that occurred downdip of the 1995 rupture. Our slip inversion reveals a potential change in mechanism in the southern portion of the rupture, consistent with seismic results. Predictions of the satellite LOS displacement from a seismic inversion and a joint seismic/GPS inversion do not compare favourably with the InSAR observations.
Additional Information© 2002 RAS. Accepted 2001 November 27. Received 2001 October 5; in original form 2001 April 16. Article first published online: 19 Jul. 2002. We acknowledge critical reviews from P. Segall and an anonymous reviewer. We are grateful to S. Husen, P. Ihmlé and L. Ortieb for access to their data, and thank D. Sandwell for suggesting using the satellite clock to help find missing lines. We thank E. Fielding for guidance in generating the DEM, made using ROI PAC. The ERS SAR imagery has been acquired under research user category from Eurimage, Italy. GMT mapping software (Wessel & Smith 1998) was used to prepare figures. MEP was partly supported by an NSF Graduate Research Fellowship. Contribution number 8841 of the Division of Geological and Planetary Science, Seismological Laboratory, California Institute of Technology.
Published - Pritchard_2002b.pdf