Using Spatial Regression as a Tool for Permafrost Hazard Assessment: A Case Study of the Hudson Bay Railway
The Hudson Bay Railway (HBR) has faced increasing instability and rising maintenance costs due to permafrost thaw, a process accelerated by climate change over the past three decades. Geotechnical investigations have identified the Herchmer Subdivision as the most severely impacted area, with histor...
| Main Author: | |
|---|---|
| Other Authors: | , , |
| Format: | Master Thesis |
| Language: | English |
| Published: |
Graduate Studies
2025
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/1880/120980 https://doi.org/10.11575/PRISM/48570 |
| Summary: | The Hudson Bay Railway (HBR) has faced increasing instability and rising maintenance costs due to permafrost thaw, a process accelerated by climate change over the past three decades. Geotechnical investigations have identified the Herchmer Subdivision as the most severely impacted area, with historical and contemporary data revealing that previously stable ground is becoming unstable and that the permafrost boundary is shifting northward. As permafrost degradation continues, there is a pressing need for accurate predictions of thaw-related hazards to support infrastructure resilience and maintenance planning along the HBR. To address this challenge, we employed Multiscale Geographically Weighted Regression (MGWR) to identify the key variables contributing to sinkhole formation along the railway. This spatial modeling tool enables the assessment of multiple climatic and ecological factors influencing permafrost degradation while determining their statistical significance. In our MGWR model, sinkholes—measured by the number of surface depressions along the railway—served as the dependent variable. In contrast, the independent variables included ground ice abundance, snow depth, surface temperature, and organic carbon content. Regression coefficients derived from the MGWR model were used to calculate variable weights, which were then applied in a weighted sum analysis in ArcGIS to generate a hazard map. This hazard map incorporates projected datasets for the 2030–2039 period, offering insights into the evolving risk landscape along the railway. By illustrating the shifting permafrost boundaries, the map enhances our understanding of the railway’s vulnerability to damage caused by the thawing permafrost. The findings from this study will serve as a crucial tool for guiding resource allocation and mitigation strategies, ensuring the long-term stability of the HBR in the face of climate change. |
|---|