Investigations Undertaken:

This study was to determine the date of the most recent great subduction earthquake in the Cascadia region and the extent of area affected to estimate the magnitude.

Results:

The most recent date for an earthquake producing subsidence along at least 100 km of the Washington coast has been established by our tree-ring analysis as between the growing seasons of 1699 and 1700. This result is in agreement with the tsunami hypothesis of Sataki et al. (1996).

We examined the annual growth rings of old-aged trees, mainly Sitka spruce (Picea sitchensis) that were growing through the period when the earthquake(s) occurred according to radiocarbon dating and previous tree-ring work (Atwater and Yamaguchi, 1991 and Nelson et al., 1995). The trees sampled were growing at coastal lowland sites where small changes in land level could alter the growth environment due to flooding, saline water inundation, changes in water table, or other coseismic physical disturbance of the growth site or tree. Through application of tree-ring dating methods to account for false rings, missing rings, and breaks in tree-core samples; we assigned the exact year of growth to each ring and then examined the rings for evidence of trauma. We also determined the year of growth for the last ring of two dead but relict trees and roots that were likely to have been killed by the same coseismic environmental disturbances. The dates of disturbance or death at each site were established independently. Then inter-site comparisons were made and the year of initial disturbance showed agreement between sites.

We have been concentrating on a narrowed time period limited by previous tree-ring studies and radiocarbon dating (Yamaguchi et al, 1989; Atwater et al., 1991; Nelson et al.,1995; Meyers et al., 1996). This information indicates that a major earthquake is likely to have occurred between 1680 to 1720. There is also the possibility that there was more than one earthquake. It must be emphasized that although we focused on this period, each tree-ring sample was dated and studied for its entire length for any evidence of unusual growth. There are several disturbed growth periods in many individual samples but it is only at the turn of the century in 1700 that there were so many concurrent disturbances. The concurrence among trees indicates site disturbances rather than an event damaging a single tree (e. g. a lightning strike or storm damage). Intensive study yielding numerous samples covers sites along about 100 km in a north-south direction (Benson et al., 1994). This range extends from the Copalis River in Washington south to the Columbia River.

Tree growth was disturbed at several lowland sites and the timing of the disturbances lies within the range of radiocarbon ages of tsunami deposits at nearby locations. In the spring of 1996 the processing of samples taken in 1995 and earlier was completed and combined with results of previous years' work. Some cores of trees show initial response in 1700 and others in 1701 or later. Explanations for a delayed response include the small amount of subsidence, type of stress (flooding), and resiliency of trees to environmental stresses (Kozlowski et al., 1991). Such a delay is in agreement with our study of trees in Alaska disturbed by the Alaska earthquake of March 1964. The distribution of disturbed trees indicates that there was coseismic subsidence along at least 100 km of southwestern coastal Washington in the cooler season of 1699-1700.

We also determined that 1699 was the last year of growth for a root from a western redcedar (Thuja plicata) tree believed to be killed by the subsidence event. In the fall of 1996 another root from a different tree and location was also determined to have died after the growing season of 1699 but before the growing season of 1700. The trunks of both trees had lost their outer rings due to decay.

After being informed of our results in the spring of 1996, Atwater and Yamaguchi collected and dated several more stump and root sections of killed cedar trees over the summer of 1996. Their analyses confirmed the time of death as between the growing seasons of 1699 and 1700 (Yamaguchi et al., 1997).

Non-technical Summary:

Tree-ring analysis shows evidence for tree disturbance and death during the cold season of 1699-1700 at several lowland locations along over 100 km of the southwest coast of Washington State. With an estimated coseismic subsidence of 0.5 to 1.5 m, the disturbance signal in the tree rings is not always dramatic and instantaneous. It is the initiation of the changes in tree-ring anatomy that date the time of disturbance. The last year of growth for roots of two cedar trees killed by the event was 1699.

The tree-ring record and historical tsunami evidence from Japan (Sataki et al., 1996) indicate a large earthquake in early 1700. This finding is a significant addition to the seismic history of the region and aids in evaluating earthquake hazards.