Subpart 5B: GPS measurements of absolute displacements

The objective was to measure absolute plate motions near the South Iceland seismic zone (SISZ), in order to help with improved understanding of the mechanism of faulting in the seismic zone. The objective has been met by developing a new technique to combine GPS and satellite radar interferograms into three-dimensional motion maps, that give an unprecedented view of the absolute plate motions (Guðmundsson et al. 2000). The methodology has been applied to the western continuation of the SISZ, the Reykjanes peninsula.

The original plan was to operate one semi-continuous GPS station and interpret data from this station. Rather than doing that, we participated in the installation of network of continuously recording stations in S-Iceland. These results are described in Subproject 1. Additional work was conducted on two topics: Development of a new technique to combine GPS and satellite radar interferometry results (work lead by Sverrir Guðmundsson), and secondly, fault mapping was conducted in the Hengill region in order to advance further the understanding of interferograms collected in Subpart 5A (work lead by Amy Clifton).

The new technique developed to efficiently produce high-resolution three-dimensional surface motion maps, relies on combining information about motion of the earth's surface from interferometric observations of synthetic aperture radar images and repeated Global Positioning System geodetic measurements. Unwrapped interferograms showing pixel-wise change in range from ground to satellite, and sparse values of three-dimensional movements are required as input. The problem of finding the full three-dimensional motion field is separated into two two-dimensional problems. Initially the vertical component of the deformation field, and its horizontal component in the look-direction of the satellite are found. Later the look-direction component is resolved into north and east components. Initial values for the motion fields are assigned to each pixel of interferograms from ordinary kriging of available GPS observations. These values are then updated and optimized by comparison with the interferograms and the GPS observations. Additional constraint is the smoothness of motion field. Markov random field based regularization, and simulated annealing algorithm are used for the optimization. The technique has been applied to create surface motion maps for the Reykjanes peninsula (Figure 29). Although separate interpretation of GPS and InSAR data from the area (Hreinsdóttir et al. 2000; Vadon and Sigmundsson 1997) have shown the main components of deformation, the three-dimensional motion maps provide an unprecedented view of the three-dimensional deformation. The largest signals are plate movements causing large gradients in the east motion field, and circular subsidence centered on the Svartsengi and Eldvörp geothermal area. The north and east motion field images show also clearly that the subsidence is associated with horizontal movements towards the subsidence center, a pattern that is imprinted on the background plate movements. The motion maps form the basis of a future project on conducting a detailed study of strain accumulation in the area and how it correlates with seismicity.

A related effort was the study of faulting and surface rupture that has taken place in the Hengill volcanic area (Clifton and Sigmundsson 2000). It is the same area as studied by satellite radar interferometry under Subpart 5A. The Hengill and Hrómundartindur volcanic systems in SW-Iceland are considered to comprise the Hengill triple junction at which the oblique Reykjanes peninsula rift zone, the western volcanic (rift) zone and the South Iceland seismic (transform) zone meet. It is therefore experiencing both tectonic extension and left-lateral shear causing seismicity related to both strike-slip and normal faulting. Between 1994 and 1998, the area experienced episodic swarms of enhanced seismicity attributed to magma moving within the system. Activity culminated in a magnitude 5.1 earthquake on June 4, 1998, and a magnitude 5 earthquake on November 13, 1998. Geodetic measurements using GPS, levelling and InSAR detected uplift and expansion of the volcanic edifice above a point source of pressure beneath the Hrómundartindur volcanic system. Magma accumulation elevated the volcanic edifice 2 cm/year and is believed to have triggered motion along strike-slip faults that were near to failure due to tectonic stresses. A number of faults in the area generated small-scale surface breaks. Geographic information system has been used to integrate aerial photographs, field data and geophysical data to determine how the crust breaks in response to deformation along this plate boundary, and to see how much of the recent activity focussed on pre-existing weaknesses in the crust. Our data indicates that all surface rupture has occurred along or adjacent to old faults, several of which were previously unmapped. The most prominent surface breaks occurred along NE- and NNE-trending faults adjacent to the epicenter of the June 4 earthquake. Maximum opening observed along a single fault segment was 1.2 m. Styles of rupture include fresh rockfall into pre-existing fissures and along old scarps, subsidence along and possible widening of old fissures, tears in turf and gashes in soil, shattering of lava blocks and loosening of push-up structures related to strike-slip faulting (Figure 30). Although all geophysical data agree that rupture occurred along a shallow, right-lateral strike-slip fault, no clear evidence of lateral offset was observed at the surface. Foreshocks and aftershocks from the November 13 earthquake define a broad E-W zone that intersects the southern end of the June 4 fault. The only surface rupture observed in this zone was found at that intersection.

In addition to fault mapping in the Hengill area, then Páll Einarsson at UICE.SI worked as a subcontractor on preparing a digitized fault map of the South Iceland seismic zone, as described in the PRENLAB-2 first annual report. That work is still in progress.

 

Figure 29: Estimated one-year average motion field maps at the Reykjanes peninsula. (a), (b) and (c): Average vertical, east and north motion maps, respectively, inferred by the 1992-1993, 1992-1995, 1992-1996 and 1993-1995 InSAR data, and the 1993-1998 GPS data (Guðmundsson et al. 2000).

 

Figure 30: Different styles of surface rupture observed in the Hengill area in response to uplift, and the June 4, 1998, M=5.1 earthquake. a) Gashes in soil: Ellipital segments up to 1.5 m wide, 2 m long, more than 1 m deep, en-echelon and along-strike arrangement. b) Rock fall along old fissures: Unweathered surfaces of broken rock, free of vegetation, plant roots freshly exposed. c) Rock fall along old fault scarps: Large fallen blocks, exposing freshly torn plant roots and soil, large rocks dislodged from slope leaving gaps between rocks and soil. d) Loosened push-up structures: Piles of rock with gaps of up to several centimeters between rock and soil; individual rocks may show evidence of rotation. e) Shattered aa lava: Broken blocks up to 1 m in diameter, aligned along strike and at flow margins near the June earthquake epicenter (Clifton 2000).