What caused it
A glacier collapse, not an earthquake
The flood was triggered by an ice-rock avalanche near Langtang Lirung and Tsangbu Ri, where a section of glacier gave way into the Lhende Khola catchment above Rasuwagadhi. The collapse briefly dammed the river before releasing a fast-moving, sediment-laden flood downstream. The flood reportedly moved at up to 75 km/h, and reports describe the Trishuli rising roughly nine metres in thirty minutes at Galchhi, about 50 km from the source. It is the same failure-then-release pattern recorded in a smaller flood on this same river in July 2025.
A satellite-based rapid hazard assessment (HiRisk/AMAA/Stimson, RHA NP3) identifies the source more precisely: part of glacier RGI2000-v7.0-G-15-05732, together with an unknown depth of underlying bedrock, detached across roughly 1 to 1.3 km of width at about 5,200 m elevation, mobilising material down to around 3,000 m. The debris met deposits left by a smaller 2025 glacial-lake outburst flood in the Lhende Khola and dammed the river there, apparently for only a few minutes, before that dam failed and released the flood, with flow speeds above 30 m/s (over 100 km/h) estimated in the upper reach.
The collapse registered as a magnitude 5.2 seismic signal. USGS first treated it as a possible tectonic earthquake, then reclassified it as a landslide-generated signal: large mass movements can shake the ground as hard as a moderate fault rupture. A Taiwan landslide read M4.9 in 2025, and an Alaska landslide that triggered a local tsunami read M5.4 the same year. The exact failure sequence, and whether a pre-existing glacial lake played a role, remain under investigation; the International Charter Space and Major Disasters has been activated to support satellite-based impact mapping.
Downstream, the flood wave covered the roughly 168 km from the source to the Narayani at Devghat in about eight hours. A 9-metre rise in thirty minutes was recorded at Galchhi and a 7-metre rise at Malekhu; the flow peaked at Devghat around 16:00 at roughly 5,850 m³/s, some 20 million m³ above normal flow. Nepal's Department of Mines and Geology separately recorded a magnitude-4.4 tremor at 08:37, three minutes before the flood signal reached the Betrawati gauge; the Asian Disaster Preparedness Center's field assessment attributes the shaking to the avalanche's impact rather than a tectonic trigger, consistent with USGS's own reclassification. Source: DHM Flood Forecasting Division technical report; ADPC preliminary field assessment, 27 Aug 2026.
Independent satellite-radar monitoring had flagged this exact slope, near Tsangbu Ri above the Langtang valley, as a "site of potential glacier collapse" months before it failed. Line-of-sight ground-velocity measurements (mm/month) span 8 Jan–19 Aug 2026; red indicates motion away from the satellite. Data and analysis: Manoochehr Shirzaei, Virginia Tech. Basemap imagery: Google Earth / Airbus, CNES/Airbus, Landsat/Copernicus, Maxar Technologies.
