The moment magnitude of the 1964 Alaska megathrust (Mw=9.2) was exceeded only by that of the 1960 earthquake in southern Chile (Mw=9.5). In both cases the intensity and duration of ground shaking caused widespread disrupted landslides and flows (terminology of Keefer, 1984) in upland areas nearly 500 km from fault traces. In lowland areas, however, lateral spreads and flows associated with liquefaction were the predominant landslide type: in Chile, they were localized and generally above the thrust; in Alaska, seismically induced subsidence and ground fracturing covered more than 260,000 km2 and extended more than 1000 km from the thrust. The anomalous extent of ground fracturing was due to seasonal ground ice, which acted as a heavy, rigid, impermeable, brittle pad, causing the ground above deforming sediment to break into rectilinear blocks, and liquefied sediment to be ejected from fractures, especially at their intersections.

The Alaska earthquake occurred on March 27, when seasonally frozen ground was near its annual limit of 1-3 meters. Ground failure was most severe where gravelly glacial outwash overlay thick, fine-sandy glaciolacustrine strata, where the frost line penetrated either to or near a shallow water table, and where the local slopes were negligible. In most cases, fracturing was caused by seismically-induced compaction of unconsolidated material which led to local extension and the upward ejection of pressurized water. Curiously, the outer bound of ground fracturing to the west and east coincided with the outer limit of seasonally frozen ground.

The Alaskan analog can be applied to an enigmatic and controversial set of ground fissures and fluidization structures at the Hain Quarry, in eastern Connecticut, where the stratigraphic and hydrologic conditions were identical to hundreds of severely ruptured sites in southern Alaska. These features (and others in New England) are reinterpreted to have formed under periglacial (17-20 ka) conditions when seasonal ground ice provided a rigid, impermeable upper layer and when deglacial tectonics may have provided the earthquake source. More generally, paleoliquefaction analyses, especially quantitative distance-magnitude formulae, must consider the presence/absence of seasonally frozen ground.