A deadly flood and the questions that preceded it
When a glacier collapse sent a wall of water, mud and rock hurtling down the Nepal-China border on August 26, it killed nearly 800 people in Nepal and Tibet and left more than 3,000 missing, by the latest counts. The deluge damaged about 19 bridges, some 40 kilometres of highway, and knocked out more than 10 percent of Nepal's power-generation capacity, according to preliminary assessments. Recovery will cost an estimated tenth of Nepal's annual economic output, according to a report by Reuters.
But the questions around the catastrophe began months earlier. On May 27, ten Nepali officials and about a dozen Chinese officials, including representatives from Tibet, met in Kathmandu to discuss strengthening cooperation on floods, weather and glacier-related hazards, according to an agenda prepared by Nepal's side. The participants called for joint risk-reduction measures, including inventories of glacial lakes and hazard mapping.
Nepali officials at that meeting sought more detailed information from China on glacier movements, water levels, avalanches and other hazards. "What we actually wanted from them, and still want, is if there are early movements of glaciers and early movements around that area, we would want to know a bit early," said Sauhardra Joshi, a hydrologist at Nepal's Flood Forecasting Division, who attended the May meeting. Binod Parajuli, another government hydrologist, said China began sharing heavy-rain forecasts for Tibet through WeChat and email about a month before the disaster, but the officials said they received little on glacier risks, water levels, avalanches or extreme events.
Dharam Raj Uprety, chief executive of Nepal's National Disaster Risk Reduction and Management Authority, told Bloomberg that in the latest floods, "People even didn't get five minutes of lead time." He said Nepal lacks weather stations in high mountain areas because of what he called "the lack of trans-boundary data seeding mechanisms," or limited sharing of weather and climate data across national borders, calling that one of the biggest challenges. Nepal has monitoring systems at lower elevations, but large sections of the highest mountains remain poorly observed; satellite analysis has indicated the disaster was triggered by the collapse of bedrock beneath a glacier in a relatively unmonitored area.
China's response and the May meeting
China has rejected suggestions that it failed to provide Nepal with adequate information. The Chinese Embassy in Nepal said the two countries had maintained "close communication on transboundary disaster management" and made "substantial progress" during the May meeting on cross-border disaster information-sharing. In a statement on X on August 29, the embassy said Beijing and Kathmandu had kept in close contact, and that China had "always made every effort, within the limits of existing technologies, to conduct monitoring, assessment and analysis, and have shared important information with the Nepali side in a timely manner."
China's World Meteorological Centre emailed Nepali officials twelve days before the disaster, warning of enhanced rainfall for Aug 14 to 19, and possible "secondary and cascading hazards," including landslides and flash floods. The Chinese embassy also said China has been sharing information, including monitoring data on a lake formed by debris from last week's calamity.
At the May meeting, according to Parajuli, the Chinese side said they did not have the mandate to sign an agreement to meet Nepal's requests, asking for further discussions. Nepal raised similar concerns last year, after a flood in the same area triggered by a glacial lake outburst in Tibet killed at least nine people in Nepal, officials said.
The path forward: a proposed data-sharing pact
Nepal's authorities will propose a data-sharing agreement with China modelled on an arrangement Nepal already has with India, which provides updates every 10 minutes covering glacier movement, water levels, avalanches and glacial lakes. Officials describe deeper cooperation with China as a priority.
Not everyone views the bilateral data gap as central to this disaster's toll. Austin Lord, an environmental anthropologist at the Stimson Center, said the shortcomings in Nepal-China data-sharing were not a significant factor in the response to this particular disaster, though he acknowledged there is very limited exchange between Nepal and China.
Scientists, meanwhile, have cautioned that it is too early to directly link the specific glacier collapse to climate change. The Hindu Kush Himalayas, which span Nepal and China, contain the world's largest volume of ice outside the poles, with more than 63,700 glaciers storing up to 5,700 cubic kilometres of ice — a vast, fragile landscape where the risks were known, and the warnings, for some, came too late.
In the minutes after the disaster, the limits of monitoring became starkly clear. At 8:32 am Nepal time, the deluge blasted through a border crossing in Tibet; 26 minutes later, Parajuli got a call alerting him to a flood on the Trishuli River. At 8:40 am, the border monitoring station stopped transmitting data, and Parajuli's team sent a public alert to mobile phone users at 9:13 am — about 41 minutes after the first report, and well after the flood had already begun its destructive path downstream.