Developing Hyperspectral Imaging Workflow for Ice Core Analysis

Impurities trapped within glacial ice serve as unique archives of past environments. This study presents results from imaging ice core samples collected from Antarctica, Greenland, and the Arctic using the IceSpec (VNIR) hyperspectral imaging (HSI) system. Image processing algorithms, developed with open-source Python libraries (e.g., numpy, photutils, scikit-image, and SPy) enable the quantification of trapped air bubbles, dust content, and other impurities. This work expands parameterization of ice core physicochemical properties. HSI offers a robust, fast, high resolution and automated method that enhances traditional ice core analyses while introducing new capabilities. A key advantage is its non-destructive nature, which preserves full spectral information for subsequent impurity fingerprinting, chemical characterization and sample archiving.This work was supported by National Science Foundation (NSF) grants 2149518 and 2149519, and by the Center for Oldest Ice Exploration (COLDEX), an NSF Science and Technology Center funded under grant NSF 2019719. We also acknowledge the logistical support provided by the NSF Antarctic Infrastructure and Logistics Program, the US Ice Drilling Program (supported by NSF Cooperative Agreement 1836328), the NSF Ice Core Facility, and the Antarctic Support Contractor.