Subsidence in rift zones : analyzing results from repeated precision leveling of the Vogar Profile on the Reykjanes Peninsula, Southwest Iceland

Abstract: The mid-Atlantic ridge connects with land at the Reykjanes peninsula, Southwest Iceland. Iceland extends well over the whole ridge and allows for representative surface observations of sea-floor processes. Across the Peninsula the plate boundary is characterized by a rift zone, which encompasses leaky transform fault zone characteristics. Across a fissure swarm, connected with the Reykjanes volcanic system, the Vogar leveling profile extends. Precision leveling was conducted along the line in 1966, 1968, 1969, 1971 and 1976. The profile was remeasured in 2004 and it was aimed to observe various aspects of subsidence in the rift zone. Subsidence along the profile is the result of the effect of two main components. The first being influence from the central axis of the plate boundary causing the southeastern end of the profile to subside at a higher rate then the northwestern end, averaging 1.1mm/yr. The second being the influence from the local fissure swarm causing the central part of the profile to subside due to movement on two boundary faults. Annual rates of subsidence are not constant but are related to the size of appreciable fault movements. 1966-1968 and 1971-1976, annual subsidence rates were higher and in 1976-2004 lower than average. Prominent fault movements are earthquake related. 1976-2004 was a calm period following the stress release by several earthquake swarms between 1971 and 1976. One prominent, unconfirmed, fault movement occurred between 1976 and 2004, presumably triggered by the two large earthquakes in the South Iceland seismic zone in 2000. At several faults local progressive faulting is occurring with subsidence of a single benchmark on the down-throw side of a fault. This occurs at most faults and is the result of movement of semi-detached fault blocks within the ‘disturbed’ zone sliding down along the fault plane. Modeling the subsidence results for the Vogar profile in relation to the central plate axis proved too complex for the simple linear pressure decrease model when comparing with the GPS data obtained for a benchmark along the profile. This GPS data indicated a minimum subsidence of this benchmark of 3mm/yr. The best fit for points affected only by the regional component gave an annual subsidence at the rift of 5.5mm/yr and a locking depth of 3.8km, which did not correlate with GPS data. Presumably the effect of local component overprints the regional component or the GPS data for the last 11 years is not representative of the last 38 years.