Updated : read the ICIMOD October preliminary synthesis guide, NDRRMA's official English report through 3 October and its unresolved category difference, and four September–October source-area observations. Earlier figures below retain their own dated cutoffs.
On the morning of 26 August 2026 a Himalayan glacier snout collapsed north of Langtang Lirung. Ice, rock and till ran the Lhende Khola (Donglin Tsangpo), hit Gyirong Port and Rasuwagadhi (also written 热索桥 / 热索瓦加迪), then the Bhote Koshi and Trishuli through Timure, Syabrubesi, Betrawati and Devighat.
Gyirong Port and Rasuwagadhi, 26 August 2026: a public before/after map
This was a landslide signal, not a tectonic earthquake
The U.S. Geological Survey revised event us7000tbwb to an M5.2 landslide at (08:37 Nepal Time, 10:52 China Standard Time). Depth 0.0 km. Early notices called it M4.4. USGS Key Points issued at 18:44 UTC that day state that long-period waves, nearby stations and satellite images show the energy came from a glacial collapse and debris flow: no tectonic earthquake occurred.
A second landslide signal, us7000tc90, is still an M4.2 landslide at on 26 August 2026. The USGS catalog on estimated 28.3281°N, 85.4798°E from satellite images, with a horizontal error of about 20 km. This site does not draw that estimate as a second source, and does not merge it with the Watanave snout at 28.2877°N, 85.52723°E or with us7000tbwb at 28.271°N, 85.515°E. us7000tbwb itself is unchanged: an M5.2 landslide at 02:52:10 UTC. Reuters, citing GFZ, put the Gyirong CCTV strike about seven minutes after the seismic origin, which matches travel time down the valley.
Three coordinates that must not be merged
- Ice-tongue snout (Hidenori Watanave, University of Tokyo, PlanetScope + Landsat): 28.2877°N, 85.52723°E.
- USGS us7000tbwb (seismic / satellite location): 28.271°N, 85.515°E, on the north face of Langtang Lirung, not the valley floor.
- Gyirong Port CCTV (Reuters visual verification): 28.28032°N, 85.37761°E. The Friendship Bridge / Rasuwagadhi temporary span sat near 28.2787°N, 85.3781°E and was destroyed on the 26th. Do not confuse the port with Gyirong town about 15 km north.
This site keeps each label on its own source. The us7000tc90 estimate above stays off the map. The red line on the 3D map follows OpenStreetMap waterway centrelines: Chhochen Khola → Guobaxiaqu → Donglin Zangbo (Lhende) → Bhotekoshi → Trishuli. It does not climb the USGS pin, Langtang Lirung summit, Dhunche town (about 810 m above the same-latitude channel), the Tupche hillside, or the Galchhi gaunpalika centroid. Flooded markets named in reporting are Timure, Syabrubesi, Haku Besi, Mailung and Betrawati.
What this copy adds besides Watanabe’s Cesium story
Hidenori Watanave published the Cesium Ion story Nepal - Tibet Border Flood 2026 at , credited to Planet Labs PBC / Landsat / Vantor. Open the original at ion.cesium.com/stories/viewer. This page is an independent reconstruction for non-commercial use. It is not an official product, mirror or review from that laboratory. The story JSON last write checked here is , 34 slides. On 29 August 2026 it had been 25 slides. The viewer is still a JavaScript shell. The snout is still 28.2877°N, 85.52723°E, and there are no new pins. His 9 September post reads the 8 September Sentinel-2 frame as a new slope pond about 240 m long, and says the confluence pond he had described is gone. That is his visual read. It is not an official water level.
Where a reduced recreation would stop at two slides (source and border), this site:
- Audits the red line against OSM rivers and ridge tests (Dhunche, Tupche, Galchhi).
- Places eighteen corridor sites from the ice tongue through Gyirong / Rasuwagadhi to Trishuli Bazar, Devighat, Galchhi, Malekhu and Mugling.
- Drapes Vantor ~30 cm OpenAerialMap tiles and PlanetScope 27 May 2026 on MapLibre terrain. The default post-event drape for Gyirong–Syabrubesi can use Vantor LG03 from , cloud about 18%, CC-BY-NC. That strip does not reach 85.527°E. Vantor LG06 on , cloud about 6–7%, is the first relatively clear strip whose polygon covers the snout. No tile service is published for it, so the 3D map still uses PlanetScope and GeoEye-1 there. Older tiles from 16 October 2021 and 27–28 August 2026 stay in the catalog.
- Keeps a filmstrip of public before/after sliders along the swept valley.
Open the map at a named place: Gyirong Port · Rasuwagadhi · Syabrubesi · glacier source. Hash-only links such as #rasuwagadhi are for the live camera; crawlers should use the HTML pages and ?site=.
Licenses, sensors, and what we do not copy
New follow-up observations: the 29 September Pelican scene was acquired at 05:13:36 UTC, with catalog values of 1% scene-wide cloud cover and 0.57 m GSD; the 28 September SkySat scene at 09:34:20 UTC reports 17% cloud cover and 0.66 m GSD. Both footprints contain the site's source-area point. Both use 0.5 m output pixels; that pixel spacing is distinct from GSD. Scene-wide cloud cover does not guarantee clear pixels at the source, and these metadata do not confirm a terrain change. Both collections are © Planet Labs PBC, CC BY-NC 4.0. See the observations page for the observation record and the imagery notes for asset limitations.
Landsat 9 is treated as public domain with USGS credit. Sentinel-2 may be redistributed with Copernicus attribution. Planet Crisis Response and Vantor Open Data are CC-BY-NC 4.0: attribution required, non-commercial only. This site does not dump those pixels as public-domain stock. Catalogues: Planet STAC on source.coop, Copernicus Browser, USGS EarthExplorer, OSM / Vantor Nepal Floods 2026, International Charter activation 1052 / ISRO Resourcesat-2A. Step-by-step catalogue notes are on the imagery page.
Planet Pelican at is a same-swath repeat of 27 August. The north frame is about 47% cloud and clearer at Rasuwagadhi, under CC-BY-NC. Its east edge is about 85.436°E. There are no 3D tiles. The homepage filmstrip pair is pelican-0901-border. The source Sentinel-2 comparison stays on the 27 August scene, about 78% cloud, and source previews remain monsoon cloud. Downstream, tile 45RUL on is about 16% cloud over Betrawati, and 45RTL on is about 13% cloud over Galchhi. Copernicus attribution still applies. Landsat 9 path 141/040 on 3 September and 11 September is cloudier than 26 August, so this site does not swap that scene.
Planet released additional 28 August snout scenes by 29–30 August 2026 (Guardian, 1 September). Size estimates must stay labelled uncertain: Watanave’s 1 × 2 × 1.2 km envelope is a collapse outline, not ice volume; Dan Shugar’s ~0.2 km² is area; NRSC’s “two-thirds of a glacier body” is an area fraction; UNOSAT’s ~20 km² mudflow in the Rasuwa analysis box is unverified in the field. USGS landslide findings dated 27 August, still the text cited on 1 September, leave open whether the first failure was a landslide that incorporated glacier ice or a glacial collapse. Science on 28 August quoted Cook and Ekström arguing a rock landslide initiated the cascade and the glacier “came along for the ride.” Those readings are listed side by side, not merged.
October synthesis: a new reading of the evidence
ICIMOD's preliminary synthesis, published 13 October with a v2 repository record dated 14 October, interprets bedrock failure involving glacier ice and a cascading debris flood. The dominant trigger remains unresolved. Printed pages 10 and 21 report up to about 40 cm of Sentinel-1 line-of-sight displacement from 7 July to 24 August while retaining glacier motion, atmospheric and geometric explanations. This is neither a vertical-displacement measurement nor evidence that the exact failure time could have been predicted.
The report also calls for checking whether warnings were received, understood and actionable. Conflicting border-arrival timings within the report are not combined here, and its September casualty citations do not replace the 3 October NDRRMA snapshot below. See the bilingual analysis guide for printed-page references, methods and source links.
NDRRMA report through 3 October: a category difference remains
NDRRMA's report through 3 October 2026, hosted by NSET, lists 1,455 in its total-deaths field, which includes partial human remains. On PDF page 5 (printed page 4), the female (336), male (585) and partial-remains (533) categories sum to 1,454; district counts sum to 1,455. The unexplained difference of one was still present when checked on . This site preserves the reported figures and does not interpret the total as a verified count of distinct deceased people.
The report lists 5,285 missing after deducting 112 identified bodies, including 636 foreign nationals. The register remains under verification and may contain unidentified bodies, duplicate entries or people already rescued. These are not final counts and are not added to Tibet's independently dated figures. The report also provides a minute-by-minute warning and flood-propagation timeline; this is NDRRMA's retrospective reconstruction, not an independently measured timeline produced by this site. Further context is in the bilingual FAQ.
The official English edition, published 8 October and checked 19 October, retains the same discrepancy on PDF page 5, printed page 5. The Nepali mirror's printed pagination differs. NDRRMA also lists a missing-person information verification procedure dated 12 October; that administrative source is not a new casualty total.
Earlier casualty snapshots
Death and missing counts still move. Nepal Police at 07:00 NPT on reported 469 recovered bodies. NDRRMA’s sixth situation report on 27 August listed 389 recovered and 977 missing. BBC Japanese on 28 August cited at least 469 dead in Nepal and 910 missing, plus 3 confirmed dead and 558 missing in Tibet. NDRRMA at 18:00 NPT on reported 1,050 recovered and about 3,916 missing, including 583 foreign nationals; its 09:00 NPT snapshot the same day was 987 recovered and 3,916 missing. NDRRMA at 19:00 NPT on reported 1,451 recovered, a count that includes 530 remains pieces, and about 5,705 missing. The missing figure already deducts 112 handed-over bodies and includes 636 foreign nationals. That bulletin is not a final total and may still contain duplicates. Tibet authorities as of 18:00 CST on reported 16 dead and 546 missing. As of 18:00 CST on the Tibet bulletin is 43 dead and 519 missing, 261 of them foreigners (Reuters, CGTN, Global Times). English reports through mid-September still cite the 5 September figures. Nepal and Tibet missing lists can overlap at Gyirong / Rasuwagadhi (pilgrims, drivers, border staff). This site does not add the two sides into one global total.
Chinese interface and 3D map: gyirong-rasuwa-flood.yilaoweather.org. Place pages: Gyirong Port, Rasuwagadhi. Methods: about. Timeline: brief.