CORRELATING EARTHQUAKE-KILLED TREES FROM THE FIRST MILLENNIUM CE BY MEANS OF RADIOCARBON DATING AND DENDROCHRONOLOGY IN THE PUGET LOWLAND AND SOUTHEAST OLYMPIC MOUNTAINS, WASHINGTON

We investigated subfossil forests in several locations using radiocarbon dating and tree-ring analysis to better constrain the ages and regional effects of shallow crustal faulting in western Washington State. We used saws and increment borers to sample in three areas: along Johns Creek, a tributary of the Hamma Hamma River on the west side of the Hood Canal, Washington; in lower Dry Bed Lake and Lena Lake, dammed by rockslides in the southeast Olympics within a few tens of kilometers of Johns Creek; and in Mill Pond, a small lake dammed by the Tacoma Fault ~25 km southwest of Johns Creek and 20 km south of the Seattle Fault. All dates noted herein are from bark-bearing trees except that from the Mill Pond. A Johns Creek tree had been dated previously to ~900 CE (Contreras and others, 2012). Wiggle-match radiocarbon dating and provisional tree-ring analysis show that additional Johns Creek trees likely died the same year as Seattle-area trees that were victims of the Seattle Fault rupture of 900–930 CE (Jacoby and others, 1992; Atwater, 1999). New wiggle-match radiocarbon and tree-ring results for a previously-dated tree in lower Dry Bed Lake (Schuster and others, 1992) are consistent with death ages like those of the Johns Creek trees. Tree-ring analysis of a subfossil tree at lower Lena Lake previously dated by Logan and others (1998) also yields a provisional correlation with the Johns Creek trees and thus with rupture of the Seattle Fault. At Mill Pond we were unable to collect samples from a bark-bearing tree in Mill Pond, but we did obtain a radiocarbon age on wood about 80 rings from the outermost wood. The age implies tree death after about 800 CE, consistent with age estimates for the Tacoma Fault of 770–1160 CE obtained by Sherrod and others (2004) and 770–1011 Logan and Walsh (2007) and on subsidence in southern Puget Sound at 860–940 CE by Sherrod (2001). The results of these investigations show that improved and accurate wiggle-match radiocarbon results for the Seattle Fault require subfossil tree specimen victims have enough annual growth rings (>275) to sample the steep portion of the radiocarbon calibration curve between 650–675 CE. Longer tree-ring records also improve confidence in tree-ring cross correlations that can better characterize the nature and extent of strong ground motion associated with earthquakes.