Session: Track 1-5B: Track Safety

Paper Number: 122455

122455 - Enhancing Railway Safety Through Satellite-Based Monitoring for Rockfall Potential 

Rockfalls pose significant challenges and hazards to railways, as they occur when rocks, boulders, or debris fall onto or near railway tracks. Recognizing and monitoring rockfall- and landslide-prone areas is crucial for effective risk mitigation. It has been reported that small-scale surface deformations take place ahead of a rockfall event, and monitoring a site for such deformations enables effective risk management. Current practices monitor for rockfall potential along the railway right of way only after an area is characterized as high risk, and even then, a network of sensors is typically installed at discrete points covering a limited area of the site.  In contrast, satellite radar images, cover large areas of the earth’s surface and is an innovative approach in detecting precursors of shallow geohazards along the railway right of way that can significantly improve early warning and mitigation strategies[i]. 

Small Baseline and Subset (SBAS), an advanced subset of the Interferometric Synthetic Aperture Radar (InSAR) technique, can measure small displacements and effectively detect landslides. However, relying solely on a single deformation reading increases result uncertainty and the potential for incorrect judgments in landslide recognition. This paper introduces a novel approach that implements SBAS with threshold stacking to address this concern. This technique involves stacking multiple SBAS analyses that exceed a displacement threshold, thereby identifying areas with substantial activity. When combined with SBAS timeline analysis, which monitors regions using a set number of satellite images, this approach provides effective tools for rockfall recognition.  This approach was tested in the case of a rockfall event in Sandstone, WV, where, On March 8, 2023, a freight train derailed after colliding with a large rock on the track. The accident site is approximately one mile from Sandstone, West Virginia. The railroad track stretches along the New River, between the riverbank and a cliff. The cliffside is densely covered in flora and the entire area is characterized by the lack of Persistent Scatterers, making standard InSAR techniques ineffective in monitoring the site for displacements. The proposed SBAS threshold stacking methodology successfully identified two distinct clusters along the track about 100 meters apart, highlighting areas with significant activity. Upon comparison with optical images, the two identified areas are the ones that experienced the rockfall event. To further validate the approach, the methodology was also employed in a rockfall event in Maupin, OR, and found consistent results. Additionally, we incorporated precipitation data to monitor the influence of weather on rockfall activity. In both cases, the monitored regions exhibited a gradual increase in cluster activity leading to the landslide event. The SBAS threshold stacking, coupled with the timeline approach, proved effective in predicting geohazard events based on sudden increases in displacement activity. This study demonstrates the potential of satellite-based monitoring for geohazard early warning systems, aiding in disaster mitigation and preparedness.

[i] European Space Agency, "Railway infrastructure monitoring," 23 February 2023. [Online]. Available: https://space4rail.esa.int/projects/theme/railway-infrastructure-monitoring.

Presenting Author: Sumanth Varma Byrraju University of South Carolina

Presenting Author Biography: To be submitted at a later time

Authors:

Sumanth varma Byrraju University of South Carolina
Dimitris Rizos University of South Carolina
Micheal Sutton University of South Caroline
Ning Li University of South Carolina