Crustal deformation near Hengill volcano, Iceland 1993-1998: coupling between magmatic activity and faulting inferred from elastic modeling of satellite radar interferograms

Kurt L. Feigl CNRS UMR 5562, Observatoire Midi Pyrenees, Toulouse, France. (e-mail Kurt.Feigl@cnes.fr, Jerome Gasperi UMR 5562, Observatoire Midi Pyrenees, Toulouse, France , Freysteinn Sigmundsson Nordic Volcanological Institute, Reykjavik, Island, Alexis Rigo CNRS UMR 5562, Observatoire Midi Pyrenees, Toulouse, France

Tectonic activity in the Hengill volcanic area in southwestern Iceland accelerated in July 1994 when an unusually persistent swarm of moderate-sized earthquakes began. The two largest events in this sequence had magnitudes of 5.2 and 5.0. Yet upward crustal deformation at the rate of approximately 2 cm/yr since at least 1992 indicates that most of the activity is related to inflation of a magma chamber at depth. To monitor this activity, we analyze synthetic aperture radar (SAR) images acquired by the ERS-1 and ERS-2 satellites between 1992 and 1998. By combining these phase images in interferometric pairs, we calculate interferograms which record the change in satellite-to-ground range along the line of sight between the acquisition epochs of the two images. Images acquired during the snow-free summer months remain coherent on Holocene lava flows, even after four or five years. Some of the interferograms show a discontinuity in the fringe pattern, which we interpret as 8 mm of (aseismic) dip slip on a 3-km-long segment of a N5 E -striking normal fault, part of which had been mapped previously. This slip must have occurred between July 31 and September 3, 1995 (inclusive) and has been confirmed by observations in the field. The predominant signature in all the interferograms spanning at least one year, however, is concentric fringes centered just south of the Hrómundartindur volcanic center. These we interpret as mostly vertical uplift caused by increasing pressure in an underlying magma source. The number of fringes is roughly proportional to the time interval spanned by the interferograms, suggesting that the uplift rate is relatively constant at approximately 2 cm/year. This result confirms that the uplift and seismicity observed since July 1994 continue through at least the summer of 1998. We model this signal as the deformation caused by the inflationary volume increase of a source buried in an elastic half-space. This source stresses the country rock, increasing the Coulomb stress on favorably oriented vertical faults by at least 1 bar, sufficient to trigger earthquakes. The volume source model which best fits the observed interferograms lies at 7 +/- 1 km depth and remains in the same horizontal position at between 1992 and 1998 to within about 2 km. It yields 19 +/- 2 mm/year of uplift between June 1992 and September, 1998, which is equivalent to increasing the volume of the source by ~ 106 m3/year. Accumulating as elastic strain energy in the country rock, this deformation has a volcanic moment rate of ~ 1017 N.m/year, an order of magnitude faster than the seismic moment release rate. The same model indicates that the Coulomb failure stress increases by more than 0.6 bar in an area which includes some 84 percent of the earthquakes recorded between June 1992 and September, 1998. Under our interpretation, magma is injected at approximately 7 km depth, just below the seismogenic zone formed by colder, brittle rock. There, the inflation induces stresses that exceed the Coulomb failure criterion and trigger earthquakes, possibly in a cyclical fashion.