Twelve Years of Slush and Supraglacial Lakes on Petermann and C.H. Ostenfeld Glaciers from Supervised Classification

Published in M.Sc. Thesis, Department of Physical Geography, Stockholm University, 2026

Supraglacial meltwater collects as slush and supraglacial lakes on Greenland’s northern outlet glaciers and is a key control on the region’s accelerating mass loss. Petermann Glacier hosts one of the largest remaining ice tongues in northern Greenland, while the ice tongue of neighboring C.H. Ostenfeld Glacier collapsed in 2003. Despite this, comparative study of these glaciers remains a gap in the literature, and no detailed inventory of surface meltwater on Ostenfeld exists.

This thesis adapts a random-forest classifier from Antarctic ice-shelf methods and combines it with a physically based lake-volume formulation to map supraglacial lake and slush area, lake-feature elevation, lake volume, and rapid drainage events on both glaciers. Classification runs across twelve melt seasons (2014–2025) using Landsat 8/9 imagery composited into 2-day windows, returning 90.2% overall accuracy (κ = 0.85; macro F1 = 0.90) on 2,556 holdout pixels.

Selected findings:

  • Melt area shows strong interannual variability, exceeding baselines by factors of 13.3 at Petermann and 4.8 at Ostenfeld, with both glaciers peaking in 2023.
  • Slush dominates the meltwater area, with median slush-to-lake ratios of 8.8 at Ostenfeld and 4.9 at Petermann.
  • In 2024, despite near-baseline melt area, Petermann produced the highest aggregate drainage volume on record (34.5 Mm³), reflecting carryover from the 2023 melt season.
  • This is the first detailed slush-area and lake-volume dataset for Ostenfeld Glacier.

Supervised by Dr. Nina Kirchner and Dr. Abhay Prakash. The analysis code is openly available on GitHub at MSc_ThesisScripts, which also includes a public Google Earth Engine drainage-event viewer.

Recommended citation: Urso, L. (2026). "Twelve Years of Slush and Supraglacial Lakes on Petermann and C.H. Ostenfeld Glaciers from Supervised Classification." M.Sc. Thesis, Department of Physical Geography, Stockholm University.
Download Paper