Climate Change Impact on Railway Track Buckling: A Probabilistic Risk Assessment for Taiwan

This study examines the risk of heat-induced rail buckling on the mainline network of Taiwan Railways Corporation (TRC) under historical and future climate conditions. As extreme heat events become more frequent under climate change, railway infrastructure faces increasing exposure to thermally induced instability. TRC has recorded the highest number of natural hazard-related incidents among Taiwan's railway operators, including recurrent buckling events and a major derailment in 2016. This study aims to estimate the temporal variation and spatial distribution of rail buckling risk. A physics-based probabilistic framework was developed for ballasted continuous welded rail by integrating analytical buckling models, Monte Carlo simulation, railway geometry, rail-air temperature regression, and climate projection data. Three analytical models-Meier, Sato, and Bae-were adopted for comparison. Future climate inputs for 2025-2100 were derived from the TaiESM model under the IPCC AR6 SSP1-2.6, SSP2-4.5, SSP3-7.0, and SSP5-8.5 scenarios, and the network was discretized into 20 m analysis units. Historical records indicate 75 buckling incidents during 1990-2023, more than 90 % of which occurred between April and September. Back-casting results show that the Bae model produced the closest agreement with Bayesian validation based on observed event frequencies. Future projections suggest limited change under SSP1-2.6 but increasing summer buckling risk under higher-emissions scenarios. Historical incidents were concentrated in northern and central Taiwan, and projected risk maps show clear spatiotemporal variation across years and scenarios. Sensitivity analysis indicates that curve radius and lateral ballast resistance are key factors affecting buckling probability.

Keywords: Climate change, Rail buckling, Risk of railway network, High temperature, TRC, Ballast resistance force, Monte Carlo simulation.