SURFACE ROUGHNESS ANALYSIS OF 6 BEDROCK LANDSLIDES IN WASHINGTON’S CASCADE RANGE: A MODEL FOR DETERMINING REGIONAL LANDSLIDE CHRONOLOGY AND LANDSLIDE DRIVING MECHANISMS IN WESTERN WHATCOM COUNTY, WA

The high relief, wet climate, unique tectonic setting, and geology of the western Cascade Range, Washington, USA sets the stage for hundreds to thousands of deep-seated and shallow landslides each year. However, understanding their primary driving mechanisms at regional scales remains a challenge, mainly because landslide timing is poorly known. In this study, we define a novel empirical relationship between landslide deposit surface roughness and age for bedrock landslides in the glacially sculpted valleys of western Whatcom County, WA. We then use that age-roughness model to predict the ages of all deep-seated, bedrock landslides in the county recently mapped in detail by the Washington Department of Natural Resources. To define the age-roughness model, we quantify the roughness of 6 bedrock landslide deposits with previously determined ages by applying a two-dimensional continuous wavelet transform with a 20 m Mexican hat wavelet to high resolution (~0.9 m) light detection and ranging data. Age is significantly correlated with roughness over ages ranging from the youngest landslide, the Bonneville landslide, occurring ~530 years before present (B.P.), to the oldest landslide, the Damnation Creek landslide, occurring 7855 years B.P. Those two landslides have the highest and lowest mean roughness values at 0.112 and 0.078 m^-1, respectively, while the remaining four landslides (the Day Lake, Van Zandt, Church Mountain, and Racehorse Creek landslides) have intermediate mean roughness values of 0.091 – 0.108 m^-1. The 20 m length scale roughness explained 85% of the variance in age, which was higher compared to the 10 and 5 m length scales, demonstrating its ability to capture large scale deposit features contributing to the overall roughness such as hummocks and displaced blocks. Our age-roughness model differs significantly from those developed for landslides in glacial deposits in the Puget Lowlands, WA, and for bedrock landslides in the Oregon Coast Range, highlighting the importance of calibrating such models to specific study sites. Future work efforts are aimed at deriving a more accurate equation for predicting landslide chronology to investigate landslide driving mechanisms throughout western Whatcom County, WA by comparing spatiotemporal trends to major paleo earthquakes on nearby crustal faults.