Open data show the first flood wave came within minutes, not from the later lake overflow

Why it matters

The wave arrived much sooner than the public alert, which came 38.8 min after the trigger. The author argues that monitoring glacier-covered slopes, not only glacial lakes, and sharing data across the China–Nepal border are priorities. The event hit the same site as the July 2025 Gyirong glacial lake outburst flood, 14 months earlier. This is a preprint and has not been peer reviewed.

What they did

The author used only open data: seismic records, Sentinel-1 radar backscatter change, Copernicus emergency mapping, flood extent mapping, ERA5-Land temperature data and official reports from both countries. These were combined into one timeline from source to downstream flood. Three explanations for the first flood wave were tested against the timing. Timing figures for the gauges come from another preprint, not from gauge records.

Key findings

• The seismic signal puts the avalanche start at 02:52:10 UTC. The debris flow then ran about 22 km in 6–7 minutes and destroyed the Gyirong port area.
• The first flood wave reached the first inhabited gauge 12.8 min after the trigger. The main barrier lake only began to overflow on 28 August, so that overflow cannot explain the first wave.
• Timing cannot separate the avalanche-emplaced debris flow from a rapid release by the debris dam. Gauge records from both countries would be needed.
• Radar change detection shows about 7.5 km 2 of strongly modified surface in the source sector. Assuming a 1–5 m affected layer, this gives a volume-equivalent scale of about 10 7 m3, which is not a direct estimate of the detached volume.
• No tectonic earthquake of M4 or larger preceded the event. July–August 2026 was only modestly warm, +0.59 K above 1991–2020, so the reported ‘warmest in two years’ claim is not strongly supported.

Limitations

• No post-event high-resolution elevation model or cloud-free optical imagery was available. Post-event radar coverage was a single scene, so volume is only a rough scale and seasonal surface change cannot be excluded.
• Gauge records from Nepal were not accessible. Timing figures rely on a non-peer-reviewed preprint, and casualty figures were still changing.
• ERA5 temperature anomalies are uncertain in steep terrain. The author’s timeline also conflicts with an earlier rapid assessment that dated the failure to 25 August.

Glossary

• Ice–rock avalanche: A fast slide of ice and rock detaching from a steep glacier-covered slope.
• GLOF: Glacial lake outburst flood: a sudden release of water from a glacial lake.
• SAR backscatter change: Differences in radar signal returned from the ground between two dates, used to spot surface changes such as erosion or deposits.
• Barrier lake: A lake formed when debris blocks a river valley.