Evidence for two great earthquakes at Anchorage, Alaska and implications for multiple great earthquakes through the Holocene

Multiple peat–silt couplets in tidal marsh sediment sequences at Anchorage suggest up to five possible great earthquakes during the last 2500 years. Litho-stratigraphic, diatom and radiocarbon data of the youngest two couplets record the last two great earthquakes, AD 1964 and ∼950–850 cal yr BP. Similar multiple peat–silt couplets around Cook Inlet suggest recurring great earthquakes associated with marsh subsidence over large areas, comparable in extent to that recorded in AD 1964. Potential difficulties in interpreting evidence include lack of modern analogues for applying quantitative diatom transfer functions to parts of the fossil record and the observation that under certain sedimentary conditions co-seismic subsidence, in the case of AD 1964 of known magnitude, is recorded by a transitional peat–silt boundary. This differs in diatom stratigraphy from a phase of pre-seismic relative sea-level rise recorded for other great earthquakes. A field experiment to simulate co-seismic subsidence identifies limited mixing at the top of submerged marsh sediment and indicates how mixing must be shown to differ from evidence for pre-seismic relative sea-level rise. Analysis of diatom assemblages from ice, frozen intertidal sediment and melt-out sediment demonstrates the importance of winter processes in transporting diatoms and their interpretation in fossil sequences. Winter processes may also be important in transporting organic material that leads to significant differences, ∼120–2800 yr, between radiocarbon ages from plant macrofossils and bulk peat samples. This necessitates a new approach for establishing between-site correlations, recurrence intervals and spatial extent of great Holocene earthquakes in southern Alaska.