The Great Alaska

Annual Global Positioning System (GPS) campaign surveys since 1995 in the area of the 1964 Great Alaska earthquake (Mw 9.2) show evidence for a large aseismic slip event beginning in late 1997 or early 1998 and apparently ending in 2001. Prior to 1998, velocities of sites in Anchorage and the area to the north were oriented toward the NNW, consistent with strain associated with a locked subduction interface to the south. Between fall 1997 and summer 1998, velocities of GPS sites in an area at least 150 km by 100 km in size changed by as much as 25 mm/year. North of Anchorage, the change in site velocities was large enough that sites changed direction, from NNW-directed motion to SSE-directed motion. One permanent site in the area, installed in late 1998, shows a clear time-dependent signal, moving rapidly to the south shortly after installation and then slowing down over the next two years. A preliminary evaluation of data from summer 2001 suggests that the anomalous southward motion has ended or nearly so. These observations are consistent with the sudden activation (taking less than several months) of some process that causes southward motion of the sites, and a slow decay of that process over a span of 3-4 years. We can explain this change in velocities by a model of increased creep on a large section of the plate interface downdip of the 1964 rupture zone. In this model, slip on the interface increased from roughly the rate of plate motion to roughly double that, decaying back to roughly the rate of plate motion after 3+ years. This event is different from the recent creep event observed in Cascadia, as it extends well downdip of the seismogenic zone and appears to have affected a very large area simultaeously rather than propagating along strike. The westward extent of the zone of anomalous creep is uncertain due to a lack of data prior to 1997, but this zone is inferred to be at least 100 km by 100 km, lying at a depth of 35 km or more. The change in velocities was accompanied by a significant change in the rate of microseismicity in two volumes within the subducting Pacific plate, which lie on the edges of the inferred creeping zone. We infer that these changes, in one case a reduction in the rate of seismicity and the other an increase, had the same root cause as the creep event.