The Nepal Flood: What We Know

On Wednesday morning, the 26th of August at 8.37am local time, a mass of rock and ice slid off the north face of Langtang Lirung in the Himalaya, near the border between Nepal and Tibet. As of this morning, Nepal police have recovered 359 bodies with 1400 people still missing, many of them tourists, and that number is still rising. Here's what we know so far about the events that took place yesterday, what's still open and what satellite data is revealing. I'm Sean, I'm a hydropower engineer, and here's where things stand one day after this tragedy. The first record of the event was seismic. Stations picked up shaking near the Nepal-Tibet border and the automatic solution logged it as a magnitude 4.4 earthquake. The United States Geological Survey revised that the same day. The shaking wasn't a fault, it was a magnitude 5.2 glacial collapse and debris flow, and the agency located it from satellite images rather than from the seismic network. Water levels in the Bote-Koshi Basin began rising rapidly around 9 o'clock according to the International Center for Integrated Mountain Development. Downstream on the Trishuli, the gauge at Galchi rose 9 meters in 30 minutes. Maleku rose 7 meters over the same period. Several other gauges along the river were damaged or washed away during this event. There was no rain in Nepal that morning. The chief district officer in Rissua said there was no signs of this coming. The source is the north face of the Langtang Lirum. The peak sits in Nepal, but the debris entered the river on the Tibetan side in the upper Lende-Kola. From there, it ran down the Lende into the Bote-Koshi. The first place in its path was the Zhirong border crossing on the Tibetan side. Zhirong is the main land route between Nepal and Tibet, and more than 100,000 people passed through it in the first half of this year. The flow reached it and continued. Back across the border into Nepal, the next community downstream is Siapru-Besi, a town in Rissua that sits where the trail up towards Langtang begins. Then, Rissua-Gadi and Timure, close to the border. The Bote-Koshi runs south from there and joins the Trishuli. The Trishuli carried the flow past Beitrawati and down to Trishuli Bazar in Nuwakot, which is the farthest point where destruction has been independently verified so far, about 40 miles from the border. From there, the river ran on through Dading and beyond. The reason this stayed destructive for so long was the shape of the valley. It's steep and it's narrow, with nowhere for the water to spread out. When a flood hits a floodplain, it usually thins out and slows down. This one stayed deep and fast because the walls held it in. Every one of those places I mentioned are towns where people live. 13 months ago, in July of last year, a flood came down the same channel. That one started out as a supraglacial lake on the Purepu Glacial System, about 16 kilometers upstream of where Wednesday's debris entered the river. It drained and the water came down the Lande Kola and hit Rissua-Gadi. So I pulled the satellite data for this area, and because it's monsoon season, there's a lot of clouds in the data, but some of the satellite bands allow us to observe the extents of the flood. These are Sentinel-2 scenes from the 24th of August and the 27th, two days before the event and one day after. The band I'm using measures water content in vegetation. It separates green cover from bare ground cleanly. Vegetated slopes read at one end of the scale and barren soil and rock at the other. I'm using it as a proxy for scour. Where the flow went, it stripped out everything rooted and laid down sediment, and ground that read as vegetated on the 24th reads as barren on the 27th. At the point where the debris entered the river, the barren corridor was 290 meters across. After this flood, it's measuring 1.4 kilometers. That's five times wider. So that's my measurement from the satellite data. There's some error in it due to cloud cover, but essentially what it's showing us is the deposit footprint, the ground that the flow tore up and buried. It's not the wave itself, and it can't tell you how deep the water was, but it's a good indicator of the magnitude of this landslide. The mechanism as currently understood is that a mass of rock and ice detached from the north face and fell about a kilometer down into the valley. It broke apart on the way down and became a slurry of rock, ice, and sediment moving fast enough to destroy everything in its path. Volume estimates are wide. A rapid assessment provided to the government of Nepal put the initial mass in the tens of millions of cubic meters, with the upper end an order of magnitude above that. Whether there was one source area or two is still being worked out. Alton Byers, who has studied floods in this region for around 30 years, told USA Today it'll be days and possibly weeks before anyone can lay out the exact sequence. Settling it takes ground survey and helicopter reconnaissance, and part of the source area is inside Tibet, which means it takes permission from China. Also, it's important to point out that there's an active hazard right now that directly follows the events from Wednesday. The debris that came down is blocking the river. A lake is rising behind it. Nepal's disaster authority has warned residents and rescue workers to stay away from the riverbanks, and satellite information from China indicates the impoundment is growing. On the Tibetan side, there are evacuation orders at Girong. China's Ministry of Water Resources has modeled the volume and says that with continued rain, there is a high risk of a breach within days. Regarding the climate, researchers caution that attributing a single mountain collapse to climate change will require detailed modeling. However, glaciologists and geomorphologists agree that the frequency and intensity of compound mountain hazards in the Himalaya are trending upwards. Glacial hazard monitoring in this region is organized around an inventory of glacial lakes, where they are, how big they've grown, and how stable the ice or moraine holding them back is. That's excellent for a monitoring program, but it only works when the hazard is a lake. A slab of rock and ice coming off a peak face isn't in that inventory. There's no lake to measure, no outlet to watch, no growth curve to track. Two weeks ago, a paper on the Puraipu system, 16 kilometers up this same valley, recommended continued high frequency monitoring and named the Rasuagadi bridge and hydropower project as the exposed infrastructure downstream. That paper was absolutely right about the corridor, but the failure ended up coming from somewhere that wasn't being monitored. So everything I covered today remains provisional. The volume figures will move, the sequence may get refined, and even the initiation mechanism might be revised once field teams have a chance to get on the ground. When updated data becomes available, I'll cover it here. Search and rescue operations are still underway, so the figures I mentioned earlier were current at the time of this recording, but will change by the time you watch this. My heart truly goes out to the people of Nepal and everybody impacted by this tragedy. I'm Sean, and this has been The Hydraulic Record.