| In June 2022, a rain-on-snow event led to a 500-year flood event in Yellowstone National
Park, causing widespread weather-related landslides. Trails, bridges, and roads were destroyed,
highlighting the need to assess regional landslide susceptibility and establish a localized method
for slope monitoring, especially in relation to park infrastructure. In response, we developed a
remote sensing-based approach involving both susceptibility modeling and landslide monitoring.
Two primary methods were applied: 1) park-wide scale susceptibility modeling, 2) local-scale
change detection. For landslide susceptibility modeling, we applied the multiscale model to
model cloud comparison (M3C2) method to detect surface changes between 2020 and 2023
manned aircraft LiDAR datasets, identified landslides, and extracted their pre-failure conditions
as inputs in a logistic regression model. Previous susceptibility input approaches have relied on
professional opinion due to limited landslide inventories and temporal coverage; our model
leverages pre-landslide conditions to extract data-driven input parameters. The result is an
empirically grounded susceptibility map representing landslide susceptibility under severe
weather conditions in Yellowstone. For landslide monitoring, we investigated two local slide
areas adjacent to major road infrastructure using 2020 and 2023 manned aircraft LiDAR and
2025 Unmanned Aerial Systems (UAS) LiDAR. We compared the LiDAR datasets using M3C2
to determine topographic changes across post-flood and current conditions. We matched the
change detection results with inclinometer measurements to evaluate the airborne and UAS
LiDAR comparisons. By integrating UAS-based monitoring with susceptibility mapping, this
xv
offers Yellowstone geohazard managers a comprehensive method for managing landslide hazards
under escalating extreme weather events.
Keywords: Landslides, LiDAR, Susceptibility, Yellowstone, UAS |