Weather
When a high-mountain slope failure cascades downstream: reconstructing the 26 August 2026 Gyirong mixed rock-ice disaster
Key Points
arXiv:2609.04563v1 Announce Type: new Abstract: On 26 August 2026, a high-elevation slope failure involving rock and glacier ice near the China--Nepal border developed into a channelised mass-flow and flood cascade that reached Gyirong Port/Timure. Using event-bracketing Sentinel-2, Landsat-9, PlanetScope and Sentinel-1 imagery, Copernicus DEM terrain analysis, and UNOSAT/Copernicus EMS rapid mapping, we reconstructed the source--pathway--receptor system. Nested multisensor envelopes...
arXiv:2609.04563v1 Announce Type: new
Abstract: On 26 August 2026, a high-elevation slope failure involving rock and glacier ice near the China--Nepal border developed into a channelised mass-flow and flood cascade that reached Gyirong Port/Timure. Using event-bracketing Sentinel-2, Landsat-9, PlanetScope and Sentinel-1 imagery, Copernicus DEM terrain analysis, and UNOSAT/Copernicus EMS rapid mapping, we reconstructed the source--pathway--receptor system. Nested multisensor envelopes constrained the post-event changed/source surface to a preferred 1.009 km$^2$ (0.491--1.841 km$^2$), while a terrain- and drainage-constrained representative route extended 21.843 km to Gyirong Port and descended about 3396 m. The pre-event Sentinel-2 sequence showed no unambiguous large-scale surface change; the last usable optical image preceded failure by 45.85 h. Neither GPM IMERG V07 Late nor ERA5-Land placed five prespecified precipitation windows in an unusually wet upper tail. In contrast, the 7 d ERA5-Land mean 2 m temperature (9.43 $^\circ$C) and positive-degree-day total (65.94 $^\circ$C d) exceeded all 25 matched years from 2001--2025, providing regional thermal context but not evidence of a trigger. Rapid mapping delineated a 37.353 km$^2$ downstream disturbed footprint containing 695 building points and 15.671 km of roads graded destroyed. These are preliminary mapped features, not a complete damage census. Our main contribution is an uncertainty-explicit, reproducible reconstruction of the full cascade and its observation limits: open Earth observation robustly constrained first-order source extent, propagation, environmental context and impacts, while the exact detached perimeter, source volume, short-lived pre-failure activity, initiation mechanics and source-scale hydrology remained unresolved.