Towards integrated management of rockfall risk along the access roads to Yosemite Valley (California, USA)

Yosemite National Park attracts millions of visitors each year that arrive to enjoy views of the iconic 1000-m-high granitic rockwalls. This setting, and the access roads to the park, are coincidentally prone to rockfall hazards due to their geology (e.g., exfoliating granite) and complex geomorphological features (e.g., glacially sculpted landscape). Because U.S. National Park policies limit engineering mitigation on natural slopes, rockfall hazard management along roadways typically rely on traffic management strategies informed by local risk assessment.

The access roads to Yosemite Valley (El Portal Road, Big Oak Flat Road, and Wawona Road), where most visitors travel, pass through areas characterized by a variety of rock types (generally variations of Cretaceous granitic rock) and geomorphological settings, such as high-relief glacial valleys with steep rock walls and talus deposits, as well as areas with lower local relief characterized by gentle, subdued topography, intense weathering, and thick granular soils. These characteristics influence the nature and severity of rockfall hazard and risk along access roads. To assess rockfall hazard and risk along the park’s roadways, a probabilistic risk analysis was conducted to estimate annual probability of loss of life for visitors on the three entrance roads to Yosemite Valley. The analysis was based on 3D rockfall simulations performed using the Hy-STONE rockfall runout modeling software and on rockfall event and vehicle traffic data collected by the National Park Service. Rockfall runout simulations leveraged high-resolution data (1-m LiDAR-derived DEM and canopy height models, geology, and vegetation maps), a unique database of rockfall events (1857-2023), and focused field surveys to map slope deposits, rockfall evidence, and potential source zones.

A probabilistic rockfall hazard analysis (PRHA) was performed to determine the kinetic energy that could be exceeded in N years for each 10-m-long segment of road, for each travel lane (inbound and outbound from Yosemite Valley) on the three access roadways. This analysis considered different rockfall volume scenarios (0.01-100 m³) and model uncertainties. By combining these expected kinetic energies with annual rockfall frequency and an exposure analysis based on vehicle speed and size, the study calculated the dynamic annual probability of loss of life considering weekly and seasonal variations.

The results indicate that, depite vehicle traffic conditions, rockfall risk is lower in high areas with low local relief, where rockfalls are frequent but tend to be small in size and have limited runout distances. In contrast, areas with high local relief (i.e., Yosemite Valley and adjacent Merced River gorge) exhibit higher rockfall risk, due to larger, more frequent rockfalls with greater hazard potential. These findings highlight the importance of considering the specific characteristics of each area when assessing and managing rockfall risk. Adopting an approach using detailed modeling of all park access roads provides a more complete and integrated understanding of rockfall risk, with potential applications in risk management and land use planning. Consequently, this study will offer park managers valuable tools to make adaptive, datadriven decisions for managing risk in response to dynamically changing conditions in space and over time.