Statistical and deep learning approaches for mapping 3-D ground ice properties

Ground ice is a crucial component of permafrost, and mapping its distribution and properties is critical for understanding the implications of climate change on local environments. This project investigates creating a 3D ground ice map of Canada based on borehole data in the Inuvik-Tuktoyaktuk highway region. The research aims to develop methods to integrate borehole vertical profile point-based data with land surface spatial raster data, such as satellite remote sensing and geomorphometry, by investigating the correlations between surface and subsurface characteristics, and developing predictive models for ground ice parameters.