Five decades of Abramov glacier dynamics reconstructed with multi-sensor optical remote sensing

Reference glacier sites with systematic in situ monitoring provide crucial information to understand trends in regional change. Throughout Central Asia, several sites have been established over the past 15 years, often restarting long-term time series interrupted after the Soviet Union's collapse. The region also features widespread ice flow instabilities, including surge-type glaciers. Unstable ice dynamics have been usually observed within large-scale remote sensing studies, with limited ground validation or historical observations. This hampers interpretation of the driving factors of glacier change, their interaction with mass balance, and regional representativity of single glaciers. Here, we reconstruct ice dynamics at the reference Abramov glacier using satellite-based optical remote sensing. The glacier, monitored in situ over 1967-1999 and again since 2011, experienced a well-documented episode of fast flow in 1972-1973. We compile a 55-year dataset of digital elevation models (DEMs) and orthoimages by processing raw and analysis-ready imagery from multiple archives, including declassified reconnaissance satellites, the Satellite Pour l'Observation de la Terre (SPOT), and RapidEye. Our estimates for glacier length and volume changes agree well with previous observational, remote sensing, and model studies. We describe at subseasonal scale a second pulsation over 2000-2005 - not observed in situ and poorly resolved by Landsat and ASTER products. We also measure the buildup to a third velocity pulsation, with doubling of mean annual velocity since 2011 despite a continued mass loss of -0.55 ± 0.06 m w.e. yr-1. The collected evidence indicates that Abramov is a surge-type glacier with a recurrence time of 20-30 years, challenging its representativity for regional mass balance. However, we also find a decreasing magnitude and increasing duration of the pulsations, suggestive of a potential ongoing transition towards more stable dynamics.