Characterizing Permafrost Distribution Along the Hudson Bay Railway Using Temperature-Calibrated Electrical Resistivity Tomography

Abstract

As the climate warms, understanding permafrost distribution along linear infrastructure is a key aspect of assessing, mitigating, and adapting to the risks associated with permafrost thaw. Electrical resistivity tomography (ERT) is becoming an increasingly common tool for evaluating permafrost distribution at the infrastructure scale. ERT yields a subsurface resistivity model which is influenced by the phase state of pore water, with frozen soils typically displaying substantially higher resistivity than unfrozen soils. However, the electrical properties of earth materials vary widely. Therefore, many studies use ERT in a mostly qualitative way, where general interpretations are made that do not incorporate additional ground truth or validation data to delineate permafrost boundaries. In order to quantitatively establish thresholds to identify frozen versus unfrozen ground, other data types must be incorporated. Here, we present a series of ERT surveys collected along the Hudson Bay Railway (HBR) where borehole temperatures were used to establish site-specific resistivity-temperature relationships and delineate permafrost bodies. We also examine the influence of the choice of inversion parameters (regularization parameter, error model, and starting model) on the ERT results to assess uncertainty in our interpreted boundaries. In our study region, ERT models reveal the relationships between permafrost extent and surface conditions, including vegetation cover, surface water, and the railway embankment. These efforts to delineate permafrost boundaries, their variability, and their uncertainty are an important step toward monitoring and modeling permafrost, predicting hazards, and maintaining resilient infrastructure along the HBR and in permafrost regions globally.