THE GOOD, THE BAD, AND THE NOT-SO-UGLY IMPLICATIONS OF NEW 10BE EXPOSURE AGES FOR ROCK AVALANCHE DATING AND DYNAMICS, NOOKSACK DRAINAGE, WA

A suite of new ¹⁰Be exposure ages from boulders in the debris fields of several large, catastrophic rock avalanche deposits in northwest Washington provides some intriguing new insights into the potential for using cosmogenic radionuclide (CRN) analyses to investigate prehistoric landslides. We collected samples from 16 large (>3 m diameter) boulders from the debris fields of four post-glacial rock avalanches in the Nooksack River basin east of Bellingham, WA. The landslides are all derived from the Eocene Chuckanut Fm., a highly deformed unit dominated by thick beds of quartz-rich sandstone. The debris fields all possess abundant boulder fields, providing what would appear to be excellent targets for ¹⁰Be-CRN dating: the boulders are all large, stable, show little signs of significant post-emplacement weathering, and are relatively unaffected by snow, forest cover, or other sources of cosmic ray shielding. Furthermore, the geometries of the source zones for the landslides indicates that the percentage of material that might have experienced pre-landslide exposure should have been minimal (~4-8%).

Despite these conditions, our results are anything but simple. The good: two of the four landslides have at least two CRN ages that agree within analytic uncertainty of independent ¹⁴C age control for those slides (Van Zandt: ~1600-2100 yr BP; Racehorse Creek: ~3800-5100 yr BP). The other two landslides, which are less well constrained by independent ages, each have two adjacent CRN ages that are analytically indistinguishable (Maple Falls: ~3380-3440 yr BP; Middle Fork Nooksack: ~3650-3680 yr BP). We take the close correspondence in these ages to indicate actual emplacement ages. The Bad: all landslides have some gross outliers in model ages that appear to be either too old or too young relative to independent age constraints. The Not-So-Ugly: although our sample numbers are too small to be conclusive, the spatial distribution of the outliers indicate that only a couple (of the younger ages) are related to post-emplacement weathering; the others appear to record either multiple events not obvious in the landslide morphologies, or actual pre-slide cosmogenic inheritance. The lattermost idea, if correct, suggests such large landslides may experience laminar flow dynamics during emplacement, preferentially keeping near-surface rocks at the surface.