Cordilleran Section - 115th Annual Meeting - 2019

Paper No. 26-2
Presentation Time: 9:00 AM-6:00 PM

PAIRING INTERTIDAL SEDIMENTARY EVIDENCE WITH OUTBURST-FLOOD MODELS FOR THE BONNEVILLE LANDSLIDE, COLUMBIA RIVER


DAVIS, Elizabeth1, MOREY, Susannah2 and PODHAISKY, Max1, (1)Earth and Space Sciences, University of Washington, Seattle, WA 98105, (2)Department of Earth and Space Sciences, University of Washington, Seattle, WA 98195

The largest flood on the Columbia River in the past millenium is evidenced by intertidal deposits that overlap in age with the Bonneville Landslide. This slide, in the 15thcentury CE, dammed the river upstream of Portland, Oregon, at the site of the modern-day Bonneville Dam. We plan a combination of field, lab, and modeling studies to examine how the dam was breached.

Field evidence for the inferred flood, described by other scientists in the 1990s, stands out as a layer of gray silt and clay bounded by the brownish gray mud and peaty mud of freshwater tidal marshes and swamps. This gray layer, mostly 1-6 cm thick, is exposed at low tide in river banks 30-60 km inland from the Pacific coast, between Clatskanie and Knappa, Oregon. It commonly contains horizontal laminae, and as many as two fining-upward layers have been observed. It is distinctively paired with an overlying waterlaid tephra derived from the 1479 Wn eruption of Mount St. Helens.

To build upon existing studies, we plan to quantify grain size distribution and image sedimentary structures of the silt deposit. The results of this lab work will be compared with output from outburst flood models. Questions to address include: (1) Did the landslide dam fail once, or in multiple stages? (2) How may tides influence flood flows?

The modeling will use the open-source program GeoClaw to solve the depth-averaged shallow water equations and will simulate multiple scenarios of landslide dam failure and flooding. First, we will add a two-stage dam failure mechanism to GeoClaw that forces the landslide to fail from a full extent (90 m asl) to a level noted by previous studies (40 m asl), then to a local pre-landslide base level (13 m asl). We will compare flood hydraulics and inundation patterns of this two-stage partial landslide failure to an instantaneous failure. Second, we will simulate high and low tide levels by filling a reservoir in the Columbia estuary DEM to two different levels. Hydraulics and inundation patterns can indicate the size of sediment that could be transported by the different failure scenarios, which we will compare with the sedimentary evidence.