TREE-RING ANALYSIS OF SUBFOSSIL TREES FROM THE BONNEVILLE LANDSLIDE DEPOSIT AND THE "SUBMERGED FOREST OF THE COLUMBIA RIVER GORGE" DESCRIBED BY LEWIS AND CLARK

Recent radiocarbon ages on a subfossil Douglas fir tree (WILL1) buried by the Bonneville landslide deposit suggest that this large rock slide-debris avalanche dammed the Columbia River to form "the bridge of the gods" of Native American oral history between AD 1550 and AD 1760, several centuries later than previously thought (Pringle and Schuster, 1998). A study of lichens growing on the landslide (Reynolds, in press) has narrowed the calendric-age estimate to AD 1670 to 1760. We located slabs of two specimens of "the submerged forest of the Columbia River Gorge" (SFC) cut in 1934 (Lawrence and Lawrence, 1958) and measured the annual ring widths and latewood widths of sanded wedge samples from each slab. Cross-correlation tests of these data using the dendrochronology program Cofecha show that a bark-bearing SFC tree (WYE1) from Wyeth, Oregon, 13 km upstream of the landslide, likely died the same year as tree WILL1 that had been buried by the Bonneville landslide. This match shows that the recent radiocarbon data for WILL1 dates the main damming event and not a localized slope failure. These results also confirm that a temporary lake impounded by the landslide extended upstream at least as far as Wyeth, and probably as much as 90 km to the east portal of the Columbia River Gorge, as suggested by Lawrence and Lawrence. A review of historical Columbia River discharge records suggests that this lake could reach as high as 97 m asl (maximum allowed by the blockage) within several months. Because the size and nature of the Bonneville landslide suggests it was likely caused by a major earthquake, because its age is similar to that of the great AD 1700 Cascadia earthquake, and because shallow fault structures have been identified or postulated nearby, our ongoing work includes construction of a master tree-ring record that should assist in the determination of the landslide's calendric age. We will also study the residence time of the temporary lake as well as continue to document the downstream effects of its draining (O'Connor and others, 1996).