HABITAT: High-resolution Arctic Built Infrastructure and Terrain Analysis Tool

Accurate monitoring of Arctic permafrost landforms and thaw disturbances is critical for understanding environmental change and assessing risks to human-built infrastructure. Although the Arctic has been extensively imaged by very high spatial resolution (VHSR) satellite sensors, such as Maxar over the past two decades, pan-Arctic geospatial products derived from this rich imagery remain limited. The complexity of remote sensing data, combined with the growing demand for large-scale permafrost modeling, highlights the need for scalable, GeoAI-driven analysis methods. We introduce the High-resolution Arctic Built Infrastructure and Terrain Analysis Tool (HABITAT), a novel GeoAI pipeline that integrates deep learning models with high-performance computing (HPC) to enable automated, extensible, large-scale VHSR image analysis tasks across heterogeneous Arctic landscapes. HABITAT has successfully been deployed on multiple HPC resources for generating rst-of-its-kind pan-Arctic geospatial products. This work demonstrates the potential of AI-powered work ows to accelerate permafrost science and inform infrastructure resilience across the Arctic.