We present a comprehensive average velocity field for Alaska, based on repeated GPS surveys covering the period 992-2007, and review the major results of previously published papers that used subsets of this data. The spatially and temporally complex pattern of crustal deformation in Alaska results from the superposition of several processes, including postseismic deformation after the 1964 earthquake, spatial variations in plate coupling/slip deficit, translation and rotation of large crustal blocks or plates, and a large slow-slip event in Cook Inlet. Postseismic deformation from the 964 earthquake continues today, mainly caused by viscoelastic relaxation, and causes trenchward motion. The behavior of the shallow seismogenic zone along the Alaska-Aleutian megathrust is characterized by dramatic along-strike variability. The width of the inferred seismogenic zone varies over along-strike distances that are short compared to the width. The alongstrike distribution of locked and creeping regions along the megathrust is consistent with the persistent asperity hypothesis. A large slow-slip event occurred in upper Cook Inlet in 998-200, and a smaller event in the same area in 2005-2006. No sign of slow-slip events has been found in segments that are dominated by creep, which suggests that creep there occurs quasi-statically. The overriding plate in Alaska is subject to considerable internal deformation, and can be described in terms of the independent motions of at least four blocks: the Bering plate, the Southern Alaska block, the Yakutat block, and the Fairweather block.

GPS velocity observations from the Kenai Peninsula and Kodiak Island, Alaska, display a pattern of spatial variability suggesting the presence of multiple active processes on the underlying Pacific—North American plate interface. Velocities of sites on the eastern Kenai Peninsula and Kodiak Island are consistent with elastic strain accumulation at the plate interface, whereas velocities of sites on the western Kenai Peninsula and Cook Inlet are oriented in the opposite direction and suggest an ongoing postseismic response to the 1964 Alaska earthquake. We can reproduce the observed velocities using a three‐dimensional elastic dislocation model to estimate the variation in coupling between the Pacific and North American plates. Two different inversion schemes are examined to check the robustness of the results. We find that the GPS data can be satisfied by the presence of a locked area near southwest Prince William Sound, a locked area near southwest Kodiak Island, and an area of postseismic reverse slip beneath and north of the western Kenai Peninsula. The shallow plate interface trenchward of the western Kenai Peninsula does not appear to be locked. The locked areas correspond to the Prince William Sound and Kodiak Island asperities that ruptured in 1964, and the area of postseismic slip lies downdip of the 1964 rupture area. The correspondence between the present shallow coupling distribution and the 1964 slip distribution suggests that the locked regions repeat from one earthquake cycle to another.

No abstract