GEOPHYSICAL RESEARCH LETTERS, VOL. 20, NO. 18, PAGES 1903-1906, SEPTEMBER 15, 1993 OVERLAPPING RIFT-ZONE SEGMENTS AND THE EVOLUTION OF THE SOUTH ICELAND SEISMIC ZONE AgustGudmundsson NordicVolcanological Institute, Iceland Sigurdur Brynj olfsson Department of Mechanical Engineering, University of Iceland, Iceland Abstract. The South Iceland seismic zone is a 20-60 km-wide (north-south) andup to 70 km-long zoneof north andnorth- northeast trending Holocene arrays of en echelon tension fractures. Thesefracture arrays are related to dextralstrike- slip faults buried by Holocene lava flows. In this zone, majordestructive earthquake sequences occur at intervalsof 45-112 years, the largest events reaching magnitude 7 (Ms). We propose that this seismic zone is locatedbetween overlapping rift-zone segments (spreading centers), where the eastern segment has been propagating to the south during the past 3 Ma. We madea finite element study of this configuration with the segments modeled asmode I cracks loadedin tension. The results suggest that the South Iceland seismic zonein general, and the north and north- northeast trending dextral faults in particular, develop in responseto the shear stresses generated between the rift- zone segments. Introduction Since the settlement about 1100 years ago, most destructive earthquakes in Iceland have occurred in the South Iceland seismic zone (Figure 1). In this zone, major earthquake sequences (the largestearthquakes being of magnitude 7) occur at intervals of 45-112 years.Many authors havesuggested that the South Iceland seismic zone is an east-west trending, sinistral transform fault [Ward, 1971; Hackman et al., 1990]. This transform fault is supposed to connect the eastern volcanic zone in South Iceland with the western volcanic zone and, through theReykjanes Peninsula, with theReykjanes ridge. It is suggested that thetransform motion occurs in a 10-15km-wide and70 km-long zone of large earthquakes [Stefansson et al., 1993] running east-west at 64 ø N (Figure. 1); this is the central part of the South Iceland seismic zone asdefined in Figure 2. A transform-fault modelimplies that the part of the western volcaniczone north of the Reykjanes Peninsula became extinct long time ago. Thissection of the volcanic zone, however, has sustainedvigorous volcanic activity during Holocene and considerable seismic activityin recent years. Furthermore, some of themost active volcanic systems in Iceland, such as Katla andVestmannaeyjar, occur south of theproposed transform fault. The absence of east-west trending strike-slip faults, rotated andtilted blocks of the lava pile, transform-parallel grabens and normal faults, or otherstructural elements characteristic Copyright 1993 by the American Geophysical Union. Papernumber 93GL01888 0094-8534/93/93 GL-01888503.00 of oceanic transform faults, does not support the idea of a transform fault in SouthIceland. By contrast, the TjOrnes fracture zone in North Iceland has most characteristics of an oceanictransformfault [Gudmundsson, 1993]. The main surface deformation in the South Iceland seismic zone is associated with north and north-northeast trending dextral strike-slip faults, someof which reachthe bottom of the seismogenic crust [Bjarnasonand Einarsson, 1991]. To accommodate the proposed (hidden) east-west transform- fault deformation by the northandnorth-northeast trending strike-slip faults, thesefaults would have to be much more numerous and much longer than observed [Hackman et al., 1990]. A bookshelfversion of the transform-fault model would not account for the initiation of the north and north- northeast trendingstrike-slip faults, nor the east-northeast trendingfaults in the Pleistocene rocks (Figure 1). If the northtrending strike-slip faults (Figure 1) were initially northeast trending normalfaultsthat were rotated to their present direction, the accumulated displacement on these faultswould be about1 km [Bjarnason et al., 1993]. Thereis no evidence for such large displacements on these faults. In this paperwe explore the possibility of explaining the SouthIceland seismic zone, and associated faults, in terms of the stressfield generated between the overlappingrift-zone segments (spreading centers) in South Iceland. In Iceland a spreading center [Macdonald and Fox, 1983] can refer to either a volcanic systemor a volcanic zone. To avoid confusion, we usethe term rift-zonesegment, a concept that is similar in meaning to spreading center. Structural elements The main structural elements in South Iceland (Figure 1) are northeast trending dikes and normal faults, north to north-northeast trending normaland strike-slip faults,and east-northeast trending normal and strike-slip faults [Georgsson et al., 1988]. No detailed studies have been made of these fracture systems, butthere is evidence that some of the faults in the Pleistocene areas have slippedduring Holocene [Fridleifsson et al., 1980;Georgsson et al., 1988]. Also, manyof the north trending strike-slip faults aredextral, whereas at least some of the east-northeast trending faults are sinistral [K. Saemundsson, pers. comm., 1993]. Most dykes and northeast trending normal faults are presumably of an age similarto that of the Pleistocene lava pile, whereas many north to north-northeast and east-northeast trending faults areyounger [e.g.Fridleifsson et al., 1980]. The northeast trending faults and dikes were generatedin the western volcanic zone (Figure 2) and subsequently drifted outof it to thesoutheast. North and north-northeast trending arrays of Holocene 1903