Why the Nepal flash flood was almost impossible to predict
Human Health & the Environment
CBC News, 09 02, 2026
Video (11:52)
Description
A collapsed glacier in Nepal caused catastrophic flooding that killed more than a thousand people, while thousands more are still missing. Andrew Chang breaks down what accelerated the flood to such a violent degree and how factors like the scale of the Himalayan mountains and the region’s climate make prediction and monitoring so challenging.
Intro - 00:00
A cascading disaster - 1:34
A blind spot? - 5:13
CORRECTION (September 3, 2026): The correct unit of measurement for the debris described going down the mountain slope at 2:39 is metres.
(Photo credits: The Canadian Press, Reuters and Getty Images)
»»» Subscribe to CBC News to watch more videos: http://bit.ly/1RreYWS
For breaking news, video, audio and in-depth coverage: https://www.cbc.ca/news
Follow CBC News on TikTok: https://www.tiktok.com/@cbcnews
Subscribe to CBC News on Snapchat: https://bit.ly/3leaWsr
Subscribe to CBC News on WhatsApp: https://www.whatsapp.com/channel/0029Vb6MSjH9Bb63S9tbQa1Y
Download the CBC News app for iOS: http://apple.co/25mpsUz
Download the CBC News app for Android: http://bit.ly/1XxuozZ
»»»»»»»»»»»»»»»»»»
For more than 80 years, CBC News has been the source Canadians turn to, to keep them informed about their communities, their country and their world. Through regional and national programming on multiple platforms, including CBC Television, CBC News Network, CBC Radio, CBCNews.ca, mobile and on-demand, CBC News and its internationally recognized team of award-winning journalists deliver the breaking stories, the issues, the analyses and the personalities that matter to Canadians.
Transcript
Intro
0:00 · If you ever needed convincing that minutes matter, this was it.
0:04 · Quick thinking on courage saved the lives of over 1,600 school children [music] as devastating flash floods hit Nepal on Wednesday.
0:11 · A wall of earth, ice, and water comes crashing down a mountain thousands of meters high. All that debris and flash flooding barreling through the valleys of Nepal, and this school is about to be wiped off the map.
0:25 · An entire river valley along the Nepal border wiped out in seconds.
0:29 · Because absolutely everything [music] is getting destroyed.
0:33 · A monster wave crashing down from the mountains sweeping away everything in its path.
0:38 · Entire towns washed away.
0:45 · But this school, which was also destroyed, best as I can tell, not one student died.
0:52 · I was inside my classroom.
0:53 · One of my colleagues went around and came here, and he said that the flood is coming, the flood is coming. All areas washed away.
1:01 · The principal you just heard figures he had about 13, 14 minutes to get nearly a thousand students out of school and to higher ground. [music] In the chaos of the evacuation, he even had the foresight to reroute buses full of [music] younger children that were on their way to school, telling them, “Don’t come. Get somewhere safe.” This is one incredible outcome in a disaster full of terrible ones, but it also raises a question.
1:28 · If someone knew this flood was coming, why didn’t everyone get that call?
A cascading disaster
1:39 · Even if you knew nothing about Nepal, I could situate this for you by saying this disaster happened about a hundred or so kilometers west of Mount Everest in the Himalayas. [music] And when we think about floods in this part of the world, especially during monsoon season, which [music] it is now, we’re often thinking torrential downpours of rain. That’s not what this was.
2:00 · What we really saw was that the the lower part of a tongue of a glacier completely sheared off across a width of 1,300 m and it’s really like someone took a knife and basically cut the lower part of a glacier and dropped it off the mountain cliff.
2:14 · Yes, this wasn’t initially water, but rather the earth and its ice that came crashing down. It’s here on the northern face of the Langtang Range where this massive glacier is believed to have collapsed, detaching from the mountainside more than 5,000 m above sea level, sending an enormous amount of ice, rock, force, and energy screaming thousands of kilometers down a steep slope.
2:40 · The avalanche would ultimately tear through the twists and turns of the valley, slamming into a river, sending an immense amount of water, mud, and debris into nearby Rasuwagadhi, a gateway between Nepal and China, destroying it in seconds. The flash flood continued south, gathering more water, scarring the riverbanks, swallowing entire settlements along the way. Timure and surrounding towns basically [music] disappeared.
3:09 · Further south, Sabrubesi, it was considered a hub of the region. Thousands lived on either side of the river and tourists came here, too. If you wanted to trek through Nepal’s mountainside, you might start here. But the whole area absolutely ravaged. Dams, buildings, highways, bridges washed away. And the destruction this flood carved through the country would eventually reach more than 100 km away from where it started.
3:41 · It all happened very quickly, too quickly. There are warning systems in place, but they’re not really designed to track this kind of disaster.
3:50 · So, one of the only monitoring systems that we have in place are river gauges that that can track when there’s a sudden increase in the water level. That can trigger a siren downstream that can hopefully get to the communities that need it in time for them to evacuate.
4:10 · Water [music] is a major threat in these mountain ranges, but flooding is usually driven by rain or by melt, things you can track in real time. So, river and lake levels, [music] for example, can rise during monsoon season, or you can have glacial meltwater pool and then burst its banks.
4:30 · Point is, it tends to happen gradually in a way that these river gauges can actually measure over time and then [music] relay that data downstream to where it’s useful. Except in this case, it would have been like measuring the direction of a hurricane by holding out a finger. The speed of the flood was so overwhelming, the sudden flow so catastrophic, in some cases it destroyed the monitoring systems before they could even send their alerts.
4:59 · So, we went from, you know, real-time data to all of a sudden there was no information.
5:06 · So, how do you warn people about a disaster that moves faster than the warning system itself?
A blind spot?
5:16 · [music] The thing is, there’s a lot we understand about how these disasters happen.
5:23 · This is absolutely an event of a type that occurs in the Himalaya. It has occurred many times in recent recorded history.
5:31 · And even in those cases where people are in danger, there are success stories, like in Switzerland. Now, don’t get me wrong, what happened in the small Swiss mountain town of Blattem last year was tragic.
5:45 · [music] An entire community was buried under an avalanche of ice and rock.
5:49 · Authorities saying 90% of the village of Blatten has been covered by thousands of tons of ice, [music] mud, and rock.
5:56 · But nearly every single person survived because the town had been evacuated [music] more than a week earlier. Hundreds of people fleeing the area well in advance because of what locals and scientists had been noticing for years.
6:13 · In the European Alps, uh the mountainside collapse in Blatten, Switzerland, uh a year or so ago, uh that was known to be an issue well before it happened.
6:24 · These are photos from a Swiss natural hazard monitoring company showing radar and camera setups pointed directly at the [music] mountain, which they had believed at the time might give way at any moment. This type of precise [music] measuring equipment can reveal even the tiniest details about a mountain, its cracks, its movement, the slightest [music] changes in position. And what that equipment detected was this.
6:48 · What you’re seeing here is a 26-hour time-lapse [music] from the same Swiss company, GeoPrevent, just before the glacier collapse. Maybe seemingly slow, but alarmingly fast [music] for a mountain. The Earth was moving 10 m a day.
7:09 · With this kind of information, you [music] can’t predict the exact timing of the collapse itself. But this kind of observation is what allowed scientists to recognize [music] the increasing chance that something bad might happen at some point. The danger [music] was progressive and predictable.
7:27 · It’s easier to access some of those cliff sides in Switzerland, get sensors and monitoring equipment in those locations so that we can keep get real-time data. Here in Nepal, um the extreme elevations we’re looking at, there’s just there’s no trail to the top, there’s not a cable car there.
7:43 · The capacity and the resources necessary are going to be very difficult to replicate in places like the Himalayas.
7:50 · It’s one thing when you’ve been noticing changes over a long period and you decide to point a whole bunch of measuring devices at a very specific spot. But the problem with the Himalayas is scale.
8:01 · The physical scale of the range, the difficulty in working with it, the fact that it’s often covered in clouds as it is due to the monsoon season now, it it presents tremendous challenge for prediction and monitoring.
8:14 · This part of the world isn’t just a mountain range. It is an entire wall of mountains stretching thousands of kilometers containing some of the highest peaks and deepest valleys on Earth.
8:28 · Glaciers and the lakes they melt into are a key concern because they can overflow. But there are thousands of them across Asia and [music] scientists do their best to track every single one with reports like this one from 2020 creating a meticulous inventory of the most dangerous sites like one giant catalog of risk. There’s promise there, maybe, you think. And looking in hindsight at what happened in Nepal, they are indeed finding clues that suggest maybe these types of events are predictable.
9:02 · A new report published after the collapse refers to how movement of the ice surface sped up just prior, how meltwater had turned visibly brown suggesting mountain sediment was shifting. Also, noticing a crack propagating in the days leading up to the disaster. Even learning to better track events as they occur, noticing that the mass movement caused a seismic signal at a magnitude larger than four. So, maybe there’s promise in deploying more seismic sensors.
9:33 · And imagine tying all that knowledge together, pushing it to the limit using AI, combining susceptibility maps [music] with real-time data to inch closer and closer to predicting the unpredictable. But then you remember, once again, the problem of scale. I think one Washington-based think tank put [music] it best. The Himalayan region offers more slopes capable of failing at this magnitude [music] across greater vertical relief than anywhere else on Earth.
10:04 · Then it goes on to this point of how hard it can be to gauge risk even when you think you’ve found it. Of roughly a dozen extreme flow events in the region in the [music] past 10 years, only one originated in a place already known to be a risk. Every other one of these events also came from somewhere no one was watching.
10:26 · It’s very challenging to predict exactly where and when these rock collapses are going to occur.
10:33 · It would be terrific if you could put sensors in in all the locations where events like this potentially were going to happen, but it’s just not practical to monitor so many different locations all at once. As remote sensing technologies improve, as computing improves, we’ll be more and more able to watch the landscape and detect small amounts of change, but that’s still a work in progress.
11:01 · And that’s really the key to all of this. Right now, there is no crystal ball, just small incremental ways scientists are learning how to buy precious seconds, maybe even minutes. But trying to prevent or even predict the inevitable, National Geographic referred to these glaciers as cryogenic time bombs, a reference to our changing climate and how frozen water that makes up this kind of glue that binds the Earth together, the cryosphere, it’s melting.
11:33 · Is that to blame in Nepal?
11:36 · We may never know directly, but it is why scientists [music] are so certain something like this will happen again. It’s just a question of when and where.
11:46 · [music]