Decadal Glacier Geometry and Mass Changes in Auyuittuq and Sirmilik National Parks, Canadian Arctic (1958–2024)

Long-term baseline data and interannual monitoring of glacier change in the Canadian Arctic Archipelago (CAA) remain sparse. This dissertation quantifies multi-decadal to recent interannual glacier changes in Auyuittuq National Park (ANP) and Sirmilik National Park (SNP), and presents a scalable framework for operational monitoring in data-sparse polar environments. Historical glacier geometry and elevation change were reconstructed from 1958/59 aerial photographs using photogrammetric processing, co-registered to contemporary digital elevation models (DEMs; ArcticDEM v4.1 and TanDEM-X) to derive geodetic mass balance for nine glaciers (1958–2022), with bias correction and uncertainty quantification. For interannual variability (2013–2024), a hybrid machine-learning approach—K-means pseudo-labelling followed by Random Forest classification—was applied to multispectral imagery to delineate snowline altitude (SLA) and snow cover ratio (SCR). An automated pipeline extracted continuous records of SLA, snow-covered area (SCA), and accumulation-area ratio (AAR) from 9,919 Sentinel-1/2 and Landsat 8/9 scenes, incorporating masking of clouds, shadows, and off-glacier areas. Validation was performed using in situ equilibrium-line altitude (ELA) from White Glacier. Widespread surface lowering and area loss occurred since 1958, most pronounced at low-elevation glacier tongues. Specific geodetic mass balances for six ANP glaciers ranged from −0.22 to −0.35 m w.e. a⁻¹ (1959–2021/22), while Fountain Glacier (SNP) averaged −0.35 ± 0.02 m w.e. a⁻¹ (1958–2022). Glacier length shortened by ~7–16 % and area decreased by ~9–28 % in ANP; in SNP, length loss was smaller (~2 %) with ~5–14 % area reduction. Co-registration yielded near-zero median elevation differences and NMADs of 4.07 m (ANP) and 2.08 m (SNP). The classifier achieved >99.9 % accuracy; remotely sensed SLA strongly correlated with ELA from 2019–2024 (r = 0.994; RMSE = 30 m). Late-summer SLA rose consistently from 2013 to 2024, co-varying with positive degree days (PDD) and modulated by glacier hypsometry. SLA sensitivity to PDD ranged from 0.67 to 2.56 m (°C d)⁻¹. SLA is shown to be a robust, transferable proxy for annual mass balance where late-season imagery is available. This study establishes a defensible long-term glacier monitoring baseline and an operational pathway for data-limited polar regions, highlighting vulnerability of low-elevation glaciers on Baffin and Bylot Islands.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.