D41 Spatial-temporal characterization of faults and sources of inflation/deflation at Hengill – IMO+NVI+CNRS.DTP

NVI:

This deliverable is presented in several papers that use the nonlinear inversion algorithm (Deliverable D40, Pedersen et al., (2003)) to study several sources of deformation in active areas in Iceland. Since the RETINA project started, the activity in Hengill volcano has ceased, but other volcanic centers have shown signs of unrest. We have therefore used the nonlinear algorithm successfully to study more active areas of deformation in Iceland, as described below.

The subglacial volcano Eyjafjallajökull had two intrusive episodes (in 1994 and 1999) which have been observed with GPS and more recently with increased spatial coverage using InSAR. Studies of these deformation events form part of a PhD thesis (Pedersen, 2004), which includes a GRL paper on the 1994 intrusion (Pedersen and Sigmundsson, 2004) as well as a paper on the 1999 intrusion (Pedersen and Sigmundsson, subm. to Bull. Volc. 2004). These studies find that both the 1994 and 1999 intrusions in Eyjafjallajökull can be explained by sill like bodies intruded at 4 to 6 km depth, with maximum opening of 1 m. The total volume of the intrusions is estimated to be about 0.047 km³. In both cases, the northern edge of the sill intrusions correlates with locations of earthquakes observed at the time of the intrusions, but spatially offset from the area of maximum uplift. The authors suggest that the seismicity is related to the feeder dikes for the intrusions, located at the northern rim of the sill.

Another study combining different types of geodetic data (InSAR and GPS) using the nonlinear algorithm (D40) has recently been submitted to the JVGR (Pagli et al., 2004). This study used InSAR and GPS data collected at Askja volcano from 1992-2000. The center of the caldera at Askja volcano has been subsiding by a nearly constant rate of 5 cm/yr (Sturkell and Sigmundsson, 2000). The InSAR data add spatial information, showing a clear signal of subsidence over the fissure swarm, north of the Askja caldera. The joint inversion of InSAR and GPS data tested models with different geometries (point source and ellipsoidal sources) to explain the subsidence signal. Using an ellipsoidal source allows one to estimate the dimensions of the magma chamber as well as the pressure change. The Askja study found that a shallow (2.6 km deep) ellipsoidal source and a deeper (~11 km deep) point source (Mogi) best explain the data. The subsidence is associated with a pressure decrease of about 1 Mpa/yr and the maximum volume change is 0.0021 km³/yr (Pagli et al., 2004).

The nonlinear algorithm has been used to study sources of triggered earthquakes on the Reykjanes Peninsula on June 17, 2000. The results have been reported in a paper by Árnadóttir et al. (2004). They use GPS and InSAR data to estimate fault parameters of three M>5 earthquakes and use the sources to calculate the cos-seismic stress changes. The study shows that all three events ruptured N-S right-lateral strike slip faults, with the largest geodetic moment released on a previously un-recognized fault near lake Kleifarvatn. Discrepancies between seismic and geodetic moment for this event suggests that most of the energy was released aseismically, within a few hours after the June 17 main shock in the south Iceland seismic zone. The Kleifarvatn event increased the Coulomb stress change at the location of the third triggered event (Nupshlidarhals), as well as on the Hvalhnúkur fault. These calculations have implications for future seismic hazard assessment on the Reykjanes Peninsula.

Furthermore, the distributed slip model obtained from joint inversion of GPS and InSAR data for the June 2000 main shocks in the south Iceland seismic zone (Pedersen et al., 2003) has been used to calculate the co-seismic Coulomb failure stress change in south Iceland (including the Hengill area). The results of this study were published in Árnadóttir et al. (2003). We find positive stress changes at the June 21 main shock hypocenter, due to the June 17 main shock, hence promoting failure of the second fault. Our calculations show positive stress changes in the area west of the June 21 rupture, increasing stress on potential future failure planes in the area.

IMO:

Completion of the spatial-temporal characterization of faults at Hengill, depends on input from WP 3200 (D20: A catalog of relatively located earthquakes at Hengill), which is delayed until M42. Delivery of the report is therefore delayed until M42.

Because of a recent volcanic event in Iceland, we have extend our study area to include analysis of EQ-VO coupling observed in the 1 November 2004 subglacial eruption in Grímsvötn volcano. The volcano is situated under 150-250 m of ice in the Vatnajökull ice cap in southeast Iceland. This latest eruption came on the heels of a subglacial flood (jökulhlaup) from the Grímsvötn caldera lake, and both were successfully predicted by IMO because of proper identification of seismic and geodetic (GPS) precursors. The National Civil Protection Agency, the Volcanic Ash Advisory Center in London, and the Icelandic Oceanic Area Control Center were all issued timely warnings of the impending jökulhlaup and eruption, allowing for precautionary measures to be taken on land and in the air. Real-time monitoring of eruption progress was achieved by using both seismological an meteorological observation tools, enabling real-time hazard assessment to aviation and monitoring of flood hazard. The eruption lasted 5 days; the jökulhlaup over a month. Based on earthquake locations under Grímsvötn and the horizontal displacement of the CGPS station SKRO, 51 km west of Grímsvötn, an estimate of the volume of magma involved in the eruption is 0.03 km³, sourced from a depth of ~4 km. An article has been submitted to EOS (Vogfjörd et al., 2005) and an IMO report is in preparation, intended as a general handbook for future volcano monitoring in Iceland, including interaction with civil and aviation authorities during volcanic crises. Four abstracts have been submitted to the EGU General Assembly in Vienna, in April 2005 (Jakobsdóttir et al., 2005a, 2005b; Arason, 2005a, 2005b)

CNRS.DTP:

CNRS-DTP is responsible for this deliverable in collaboration with NVI and IMO. It depends in part on the of relocated earthquakes, fault planes, and subsurface faults (D20 in WP3200) that is behind schedule. Working with the data we will modify D41 slightly to cover “Spatial-temporal characterization of seismicity and sources of inflation/deflation in active volcanoes in south Iceland”. This will be delivered slightly behind schedule, before the RETINA workshop at M39.