Message from Nepal

THREE features of the catastrophic flood that tore through Nepal on Aug 16, 2026, do not fit Pakistan’s mountain disaster strategy: it started without rains; it originated across the border; it had huge compounding factors. None of these three is a one-off. They describe a disaster category that has been forming across the region for over a decade, growing more frequent, widespread and severe. But Pakistan’s climate planning has stayed focused almost entirely on the monsoon.

A mass of rock and glacier ice broke away from an altitude of about 5,200 metres in the upper mountain range, generating seismic energy equivalent to a 5.2-magnitude earthquake. The fall produced enough friction to melt the ice on the way down, turning a dry avalanche into a slurry of mud, ice and boulders moving at up to 167 kilometres an hour at a height of 9m. The debris choked the river and formed a dam. When that gave way, a second surge followed; both hit settlements and the hydropower works below.

It started above the cloud levels. Slopes at this altitude are held up by permafrost gluing the rock and glacial ice buttressing the valley walls. Warming removes both: thawing permafrost raises pore pressure and cuts shear strength, while retreating glaciers stop bracing the slopes from outside. The rock fails on temperature’s schedule, not the monsoon’s, which is why Chamoli in Uttarakhand in India went down under clear skies in 2021 and Langtang followed five years later.

The record before both confirms a trend, not two isolated events. The 2015 avalanche off Langtang Lirung, the 2016 Gongbatongsha outburst crossing from the Tibet area into Nepal, and the 2024 Thame outburst in the Everest region struck without warning, each widening the geography of a hazard.

High-altitude disasters underscore the need for Pakistan to go beyond monsoon planning.

In Pakistan, the 2010 Attabad collapse dammed the Hunza River and submerged 24 kilometres of the Karakoram Highway. The Shisper Glacier’s ice-dammed lake breached without rain in 2022 and destroyed the highway bridge at Hassanabad. Each event has moved higher in elevation, wider in geography, and heavier in what it destroys, from one buried village in 2015 to a cascading event spread over several kilometres.

Pakistan keeps no comparable dataset on its own landslide activity. But a study in the National Science Review, covering more than 8,000 landslides across the eastern Himalayas, found a five-fold increase in high-elevation slope failures over three decades, tracking a doubling of Hindu Kush Himalaya glacier mass loss since 2000. Failures are also rising and entering the 5,000m-6,000m band that was once too frozen to fail.

The Langtang collapse struck Gyirong Port, before the debris continued into Nepal’s own river systems. The direction of travel is not new. Nearly half of the more than 25 major glacial lake outburst floods recorded in Nepal since the mid-20th century have originated across the same border, arriving from a stretch neither country monitors jointly. This event was not itself a Glof — no lake burst before the collapse — but the debris dam it left behind and the discharge that followed sit inside the same unmonitored corridor.

Pakistan belongs to the same geography, through the Kyagar Glacier’s periodic surges in the Shaksgam Valley. But China is not just Nepal’s or Pakistan’s upstream neighbour. It is South Asia’s upper riparian. It is the source plateau for the Indus, the Sutlej and the Brahmaputra — the headwaters of most of South Asia’s major rivers. That makes bilateral arrangements structurally insufficient. What this event calls for is a South Asia-wide early warning architecture, built and shared across every state whose rivers begin on that plateau, not a patchwork of separate deals each covering one border and missing the rest.

None of Nepal’s responses fixed the underlying failure. Built for a slow monsoon rise, its flood sensors were destroyed before they could warn of a moving slurry. Existing bilateral arrangements track slow glacial lake outbursts, not sudden ice-rock collapse. Nepal had no real-time alert before the surge. Pakistan, sharing the same upstream geography, has the identical gap.

What Pakistan needs: Dasu, Diamer-Bhasha and Mohmand are located in river corridors engineered for a monsoon flood profile, not the debris slurry that trapped 500 workers in Nepal. The Karakoram Highway has already been severed once at Attabad and once at Hassanabad. A third closure is not hypothetical.

Nepal rewrote its school-siting rules before rebuilding, not after. It accepted specialised help without inviting chaos. Pakistan’s community-based disaster risk management networks in Gilgit-Baltistan and KP already have the local trust that saved lives on the Trishuli corridor before state machinery arrived. What they lack is a local government signal to act on before the water reaches them.

Four steps are needed. First, convert the existing weather arrangement with China into automated, real-time information sharing on slope deformation and barrier-lake formation in the Shaksgam and Xinjiang zones — the exact category the Nepal-China arrangement missed.

Second, focus on joint satellite monitoring with the Chinese Academy of Sciences, ICIMOD and Suparco, using synthetic aperture radars across the Karakoram, Pamirs and Hindu Kush structured from the outset as a Saarc-wide facility that India, Bhutan and Bangladesh can draw on since all sit downstream of the same plateau.

Third, install seismic and hydrological sensors along the Indus and Shyok corridors calibrated for a slurry, not a rising river, so that downstream infrastructure gets lead time measured in hours instead of minutes.

Fourth, join the consortium undertaking primary technical monitoring, satellite radar analysis, and field assessments, led by ICIMOD, the United States Geological Survey, Integrated Research on Disaster Risk, Nepal Environment Society, and academic glaciologists from such institutions as the University of Michigan. They are combining real-time remote-sensing satellite tracking with ground-level hydrological stations. These entities are quantifying the initial collapse volume, utilising synthetic aperture radar to monitor adjacent unstable slopes, assessing secondary barrier-lake outburst risks, and advancing regional hazard-mapping protocols — areas of critical importance to Pakistan.

This catastrophe combines rapid- and slow-onset — the compound hazard now falling inside the mandate of the Fund for Responding to Loss and Damage.

The writer serves on the Board of the Fund for Responding to Loss and Damage.

Published in Dawn, September 10th, 2026