2026/01/01 by Diego Pacheco-Ferrada, Thorsten Seehaus
Earth and Planetary Sciences · Environmental Science · #Cryospheric studies and observations #Climate change and permafrost #Landslides and related hazards
paper · doi:10.1017/aog.2026.10053
Abstract Glaciers in the Tropical Andes have experienced a significant and accelerated decrease over the last few decades, primarily driven by climate change. This study aims to generate updated and temporally consistent yearly glacier outlines for the Tropical Andes of Peru and Bolivia (2016–24) by implementing a fully automatic deep learning approach. Here, the Glacier-VisionTransformer-U-Net model was extended and trained on the most recent Peruvian glacier inventory to segment debris-free and debris-covered glaciers. The model accurately reproduced the overall glacier extent, particularly in debris-free areas, though performance decreased in debris-covered sections. Between 2016 and 2024, glaciers retreated by 201.6 plus or minus ± <mml:math xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:mnf="http://cambridge.org/core/manifest" xmlns:cup="http://contentservices.cambridge.org" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:m="http://cambridge.org/core/metadata" xmlns:core="http://cambridge.org/core" xmlns:c="http://cambridge.org/core/content" display="inline"> <mml:mrow> <mml:mi>±</mml:mi> </mml:mrow> </mml:math> 99 km 2 2 <mml:math xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:mnf="http://cambridge.org/core/manifest" xmlns:cup="http://contentservices.cambridge.org" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:m="http://cambridge.org/core/metadata" xmlns:core="http://cambridge.org/core" xmlns:c="http://cambridge.org/core/content" display="inline"> <mml:mrow> <mml:msup> <mml:mi/> <mml:mn>2</mml:mn> </mml:msup> </mml:mrow> </mml:math> (14.7%) at an average rate of 1.83% a minus 1 -1 <mml:math xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:mnf="http://cambridge.org/core/manifest" xmlns:cup="http://contentservices.cambridge.org" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:m="http://cambridge.org/core/metadata" xmlns:core="http://cambridge.org/core" xmlns:c="http://cambridge.org/core/content" display="inline"> <mml:mrow> <mml:msup> <mml:mi/> <mml:mrow> <mml:mo>−</mml:mo> <mml:mn>1</mml:mn> </mml:mrow> </mml:msup> </mml:mrow> </mml:math> , with losses concentrated at lower elevations and south- to west-facing slopes. Results suggest that glacier mapping in this region is best conducted during El Niño years, as these events minimize seasonal snow cover influence. Compared to the Randolph Glacier Inventory V7.0 (1998), 1261 glaciers have shrunk below 0.01 km 2 2 <mml:math xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:mnf="http://cambridge.org/core/manifest" xmlns:cup="http://contentservices.cambridge.org" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:m="http://cambridge.org/core/metadata" xmlns:core="http://cambridge.org/core" xmlns:c="http://cambridge.org/core/content" display="inline"> <mml:mrow> <mml:msup> <mml:mi/> <mml:mn>2</mml:mn> </mml:msup> </mml:mrow> </mml:math> , 436 of them between 2016 and 2024. Small size, low elevation and El Niño–Southern Oscillation-driven fluctuations in seasonal snow cover play an essential role in glacier persistence in the region.