A steadily flowing glacier and no detectable precursor before the 11 October 2026 Budhi Gandaki rock–ice avalanche and flood, Nepal

Ajay Kumar Sah 

On 11 October 2026 at about 04:20 local time, a rock–ice avalanche at the head of the Ungilung Khola, on the southern slopes of Chamar (7,165 m) in Gorkha District, Nepal, became a debris flow and flood that travelled down the Syar Khola and the Budhi Gandaki. I reconstruct the event from free satellite data (Sentinel-1, Sentinel-2, Landsat 8/9, ITS_LIVE and NISAR), official records and my own reading of historical Google Earth imagery. Version 1 covered the first four days; this version adds the first clear Sentinel-2 images (21–26 October) and post-event radar. Each change test is compared with a control, where possible a fake event made from images taken before the event. The change mapped at the source on 11 October happened between 6 October 10:41 and 11 October 10:42 local time. A river gauge 23 km downstream recorded no flood pulse between 5 and 11 October. On 13 October, Landsat 9 shows a 0.31 km² block of snow loss at 5,020–5,470 m on an unnamed valley glacier, absent in the fake test, and a candidate avalanche track at 6,350–6,600 m above face B; face A is unchanged. The glacier had flowed at about 60–130 m yr⁻¹ for a decade, and no speed-up or other precursor was detected. Newly public L-band interferograms from NISAR kept face B within about ±2 cm along the line of sight in every pair from May to 7 October, and showed nothing unusual on the hanging glacier above it in the pair that ended 9.5 hours before the collapse. Against 1 October, the clear images show a 0.85–1.03 km² dark area at the head, centred on the 0.21 km² snow → grey patch of version 1, where bright snow had turned to a dark surface, most likely debris or dust (0.55 km² against the same day in 2025); face A was as in October 2025. About 20 km² of ground changed from the head to the Syar confluence, against 8.4 km² in the same weeks of 2025. Across the event, NISAR coherence fell on 13–18 % of the path below 4,000 m on all three tracks, against at most 6–10 % in earlier pairs, but it did not outline the source. Radar at 06:03 recorded a fresh rough surface, most likely deposit or debris-laden water, 0.5–1.5 km up the Syar Khola from its confluence with the Ungilung, where the blockage and a 1.5 km backwater were reported. The flood reached Arughat, 63 river km downstream, in at least three rises, the largest last. The release scar, its volume and any chain of small lakes remain unmeasured without high-resolution imagery or a post-event elevation model.