Evidence for possible precursor events of megathrust earthquakes on the west coast of North America

Megathrust earthquakes in western North America may be preceded by a precursor phase several years prior to megathrust, induced earthquakes. For example, on 27 March 1964, a 9.2 magnitude (on the Richter scale) earthquake occurred on the coast of Alaska. Changes in foraminifera and diatom assemblages at Girdwood Flats, Alaska, provide evidence of a precursor to this earthquake, thereby detailing a previously unknown sequence of events. We describe further evidence of precursor phases from marshes in Turnagain Arm, Alaska, United States, and farther south in Netarts Bay, Oregon, United States; this is the first time that two widely spaced locations have been examined for earthquake-related precursor stages. The Alaska earthquake offers the possibility to compare a modern sequence (A.D. 1964) of events with the geologic record. The Netarts Bay marsh has experienced no modern earthquake that could be used for comparison, but the nature of megathrust zones implies that the modern and ancient events should be physically similar. New cores examined from Turnagain Arm include both the 1964 earthquake and an event identified and dated at 1800 yr B.P. Prior to the 1964 event, the foraminifera and thecamoebian assemblages changed from a forest phase to a mildly brackish stage; this sequence was dated as being 15 yr or less in length, using [Pb.sup.210] and [Cs.sup.137] dating techniques. The event at 1800 yr B.P. was also associated with a similar precursor stage, indicating a small subsidence prior to the megathrust earthquake-related subsidence event. In Netarts Bay a new core was taken at a previously cored site to make use of carbon14 dates and the assurance that at least four events were known to have occurred over the past 3000 yr. The new core contains four transitions, each identified by a mineralic deposit, often sharply bounded below and gradationally overlain by marsh peat. Transition in this context refers to the section of core analyzed, ~15 cm above and below each of the four sand layers. Foraminiferal assemblage analyses indicate that these units represented a high marsh prequake phase followed by a lower marsh precursor stage, the earthquake-related deposition (sand layer), and then a rebound back into postquake marsh deposits. Sand deposits with either no or few foraminifera are inferred as tsunami/ earthquake-related deposition stages. These transitions in two widely separated geographic areas (Alaska and Oregon) indicate that similar mechanisms operate for large megathrust earthquakes at subduction zones from Alaska to northern California in the Cascadia Subduction Zone, thus implying that precursor events also occur and can be detected by foraminiferal zonation all along this area. In a recent article (Dragert et al., 2001), scientists from the west coast suggested that 'slow' or 'silent' earthquakes they measured with continuous geographical positioning systems might be indicators of megathrust earthquakes. The transitions we document may be the prehistoric representations of these 'silent' quakes. Foraminiferal evidence may help by allowing, more accurate positioning of seismometers along the west coast of North America and thereby lead to more precise and timely earthquake prediction methods. Keywords: precursor, megathrust earthquake, foraminifera, thecamoebian, marsh/ forest transition.