A probabilistic methodology to estimate site-scale thaw settlement in permafrost terrain under climate change

Abstract

In permafrost terrain climate change poses a severe threat to infrastructure. Deterministic methods for predicting soil temperature profiles struggle to account for inherent uncertainties in soil properties and surface conditions such as spatial and temporal variations and heterogeneity in surface material characteristics. This paper addresses this limitation by developing a probabilistic thermal analysis model using Monte Carlo simulations in Python, integrated with TEMP/W software. The model provides an estimate of site-scale thaw depth and associated thaw settlement of permafrost sediments in Hudson Bay Railway region under the worst-case climate scenario predictions for 2023–2100. The results of this study indicate that understanding the initial ground temperature conditions is critical for realistic predictions of both short-term and long-term thaw depths and thaw settlement variability. This research reveals that climate warming trends will likely accelerate the rate and depth of permafrost thaw, as evidenced by the increasing variability of possible thaw depth and settlements, which become more diverse and exhibit multiple probabilities as climate warming intensifies throughout the century. The methodology was also used to understand the sensitivity of input parameters and identified moisture content and thawing and freezing indices as the key drivers influencing the magnitude and variability of estimated thaw settlement, respectively. The methodology presented in this study provides valuable information on the distribution of potential outcomes when climate change is incorporated into thaw prediction. This research builds on existing knowledge of uncertainties in permafrost modeling with climate change scenarios and contributes by providing a probabilistic framework that integrates these uncertainties into infrastructure resilience, serviceability, and maintenance assessments.