Monitoring Cryospheric Degradation in the Eastern Pamir: A Machine Learning-Based Analysis of the Kuljilga Glacier (2001–2025)

The Kuljilga Glacier, situated between 5000 and 5300 meters above sea level on the North Alichur Range in the Eastern Pamir, Tajikistan, serves as a vital indicator of the high-altitude cryosphere's response to climate change. As a component of the Alichur–Gunt River basin, the glacier’s evolution is central to regional water resource assessment. Given the ongoing warming and aridification of the Pamir region, regular monitoring of these ice masses is essential for understanding long-term cryospheric dynamics.

We employed an automated classification framework within the Google Earth Engine (GEE) platform to process Landsat 7, 8, and 9 imagery from 2001 to 2025. To mitigate seasonal snow interference, we focused on the annual maximum ablation period (July 1 – September 15). Clean ice was identified using a Normalized Difference Snow Index (NDSI) threshold of > 0.5, while debris-covered ice was classified using a combined algorithm of NDVI (< 0.3) and surface temperature (T < 300 K). Glacier volume changes were estimated using empirical power-law relationships calibrated for Central Asian glaciers.

The analysis reveals a significant structural transition: while the total glacier area remained relatively stable (averaging 4.71 km²), the extent of "clean" ice plummeted from 4.12 km² to a historic low of 1.71 km². Conversely, the debris-covered area expanded by over 50%, rising from 1.96 km² to 3.08 km². Estimated ice volumes fluctuated between 0.23 and 0.28 km³, with marked mass-loss episodes observed in 2004, 2008, and 2015, which coincide with extreme regional climatic anomalies.

This study demonstrates that monitoring glacier margins alone can mask significant internal degradation. The rapid transformation of the Kuljilga Glacier—characterized by the replacement of clean ice with a moraine shell—is a primary indicator of a persistent negative mass balance. This shift from a "clean" to a "debris-covered" type necessitates a revision of hydrological models for the Alichur River (upper reaches of Gund River) basin, as the expanding moraine cover fundamentally alters the ice-melt regime and future water availability projections.