Swechha SearchAct

Active situation ·Nepal ·Glacial lake outburst flood

Active situation. Being tracked now. Figures move while you are on this page.

Nepal: glacial flood

547 Confirmed dead Preliminary Outlets report 160–547 · 5 sources
over750 Missing or unaccounted for Preliminary Outlets report 750–1,468 · 2 sources
1,944 Injured Media report 1 source, agreeing
about320 Indians missing or uncontactable Preliminary Outlets report 170–320 · 6 sources
Every figure, and the words it was read out of

No number on this page is this site’s own. Each one was read out of a published headline and is printed here beside that headline’s own wording, so a figure and the sentence it came from can be checked against each other. Where outlets disagree, all of them are below — nothing is averaged and nothing is reconciled.

Confirmed dead

  • 547 India Today — “toll hits 547 as fresh flood”
  • nearly160 News On AIR — “claims nearly 160 lives over 750”
  • 469 Maktoob — “rises to 469 1,944 injured India's”
  • 500 The Times of India — “toll nears 500”
  • 472 NDTV — “Rise To 472 Over 170 Indians”
  • 392 The Times of India — “toll hits 392 288 Indians among”

Missing or unaccounted for

Injured

  • 1,944 Maktoob — “to 469 1,944 injured India's MEA”

Indians missing or uncontactable

  • about320 A News — “says around 320 Indians still missing after Nepal flood”
  • 320 ThePrint — “320 Indians still unreachable India prepares to”
  • 320 ANI News — “320 Indian nationals remain uncontactable in flash”
  • 290 News On AIR — “Ministry informs 290 Indian nationals in Nepal remain uncontactable”
  • nearly300 News On AIR — “Nearly 300 Indian tourists missing after flash flood”
  • 290 Maktoob — “MEA says 290 Indians missing”
  • over170 NDTV — “472 Over 170 Indians Missing Amid New Threat”
Occurred The hour the first report carried, not a verified onset time.
Last updated Feeds re-read every 30 minutes. This moves only when the evidence does.

Nepal floods: 6 ways to help victims of the glacial collapse that left hundreds dead or missing — News On AIR, the most corroborated headline of 308 read. This site writes no summary of a live disaster; it prints the reporting and names it.

Where it happened

547 Confirmed dead160–547 reported 750 Missing or unaccounted for750–1,468 reported 93.2 Rain, 7 daysmm over the region 08/28 SatelliteTerra · 29.6% cloud

One reading, one label. Every source behind them.

Where

Nepal — and every place the water reaches after it. Positions are real coordinates; the extent of the event is not known.

The area the satellite image below covers Koshi — in the path downstream Bihar — in the path downstream Assam — in the path downstream West Bengal — in the path downstream Nepal — the region named in the reporting, not a located event site Nepal Koshi Bihar Assam West Bengal

Region named in the reportingDownstream, in the pathThe satellite frame below

Frame 1,481 km wide. The ring is an AREA, not a point — this page knows the region the reporting names and does not know where inside it the event was. Positions of the downstream places are true.

  1. Origin Glacial lake, high Himalaya
  2. Direct impact Moraine dam fails
  3. In the path Headwater river
  4. Downstream Koshi / Gandak / Brahmaputra
  5. Under watch Bihar, Assam, north Bengal

Upstream of India — shared river system Nepal’s rivers drain into Bihar and eastern Uttar Pradesh. An outburst or a landslide dam on the Koshi, Gandak or Karnali becomes an Indian flood within days.

What is broken

Physical damage, where an authority has counted it.

Damage assessment is ongoing. No count of buildings, bridges, roads or cropland has been published in a form this page can attribute. In a Himalayan disaster the first credible infrastructure figures typically follow the first aerial survey, which follows the weather. This band fills itself when they exist; it will not fill itself with dashes.

What IS counted — people — is at the top of this page, with every outlet that reported each figure.

What caused it

Candidate mechanisms for a glacial lake outburst flood, each with what is actually known about it here. Nothing on this list is established for this event.

Under investigation Glacial lake breach A moraine-dammed lake failing and draining in minutes.
Under investigation Ice or rock avalanche A collapse into the lake, or a debris flow with no lake in it at all.
Under investigation Extreme rainfall Rain overtopping the dam, or a cloudburst producing the flood by itself.
Under investigation Landslide blockage and collapse A slide dams the river, the blockage fills, and then it fails.
Not established Thawing permafrost Frozen debris holding the dam together losing its ice.
How would we know?

Each of these leaves different evidence, and the evidence arrives at different speeds. That is why the status words above are what they are.

Glacial lake breach
Satellite imagery of the lake before and after: a lake that has emptied is visible at 10 m resolution, and its area change is measurable. This is normally the first thing settled, within days of the first clear pass.
Ice or rock avalanche
A detachment scar on the slope above, and seismic records: a large rock-ice collapse registers on regional seismometers and can be timed to the second. This is how Chamoli 2021 was reclassified from a glacier burst to a rock-ice avalanche.
Extreme rainfall
Gauge or radar rainfall over the catchment in the hours before. A true glacial lake outburst has no rain in it, so a heavy-rain record is evidence AGAINST this being one.
Landslide blockage and collapse
A temporary lake upstream in imagery, and a discharge record showing a rise, a fall and then a second larger rise. This one is often established only after the second flood.
Thawing permafrost
Ground temperature records and modelling of the specific slope. Rarely established for an individual event; it is a condition rather than a trigger, and it was the finding at Sikkim in 2023.

The mechanism in general, researched in advance

Immediate trigger. Something displaces the water or weakens the dam: a rockfall or ice avalanche into the lake, a surge wave that overtops the moraine, or thawing permafrost loosening the frozen debris holding it back. The failure is minutes long, and there is no rain in it.

Environmental conditions. Glaciers retreat and leave meltwater ponded behind loose moraine — rock and ice debris, never engineered. Mass loss across the Hindu Kush Himalaya rose 65% between the 2000s and the 2010s, so the lakes are more numerous, larger and younger than the record they are judged against.

What science says. Peer-reviewed work projects outburst risk across the Third Pole roughly tripling by 2100, with lakes rated high or very high risk rising from 1,203 to 2,963, and the eastern Himalaya remaining the hotspot under every scenario. ICIMOD's own caveat is that higher hazard has not yet produced a higher rate of floods.

Human factors. What turns an outburst into a disaster is what was built in the path. The 2023 Sikkim flood destroyed a 1,200 MW dam and about 25,900 buildings along the Teesta. India has no end-to-end glacial lake warning system commissioned, and Nepal's Bhote Koshi system has roughly six minutes of lead time because its sensors are all below a border the water starts above.

This is the mechanism for this kind of event, not a finding about this one. The classification above is this page’s reading of how the event was first reported. Himalayan disasters are often reclassified once fieldwork is done.

What the satellite sees

Imagery published here, not linked to. The same public NASA layers a newsroom would use, over the region the reporting names, on the dates either side of the event.

The same region before the event, 2026-08-23
The same region after the event, 2026-08-28
Before · 2026-08-23 After · 2026-08-28

Drag to wipe between the two dates. Both frames are the same layer, the same box and the same satellite, so what changes between them is the ground. 10.5% cloud before, 29.6% after.

Frame 25.79–30.99°N, 81.165–87.075°E. A representative point for the region named in the reporting, not a located event site. The frame is what was imaged; the flood may be anywhere inside it.

NASA Worldview Snapshots / NASA EOSDIS GIBS. Imagery courtesy of NASA EOSDIS Global Imagery Browse Services, rendered through Worldview Snapshots. Free to use, no key required.

How it developed

1 development this page can date and attribute. Not a news feed — only moments where something changed.

  1. 13:57 28 Aug First reported in the sources below. Maktoob

Assembled from the feeds this page reads. A development nobody published does not appear here, and the absence of an entry is not the absence of an event.

What Swechha says

Vimlendu Jha has been on the record on this hazard before. These are the items themselves, not a summary of them — this page does not write anybody’s opinion for them.

Vimlendu Jha

Vimlendu Jha

Founder and Executive Director, Swechha

Who he is

Where he posts

Vimlendu on X @vimlendu

Swechha on X @swechhaindia · the organisation’s account

Swechha on YouTube @swechhaindia · the organisation’s account

Video ·Swechha, YouTube Uttarkashi Flash floods | What is happening? On the Uttarkashi flash flood — the same mechanism, the same range, the Indian side of it. Video ·Swechha, YouTube The real question isn’t just WHAT happened, it’s WHO’S responsible? On accountability after a Himalayan flood, rather than on the weather. Television ·Television interview Vaisho devi Flooding: "This isn’t just extreme weather, it’s climate change in motion" The title is the quotation, as the channel published it. Video ·Swechha, YouTube Mother Nature has been constantly warning us. But about What? #dharali #dharalidisaster #shorts

Titles are the platforms’ own, verbatim. What he said is in the item; this page does not paraphrase it.

It has happened before

4 comparable events, with what each one actually cost. Every toll below is an official or peer-reviewed count, not a headline.

2023 South Lhonak Lake, Sikkim 55 dead, 74 missing and other credible counts run to 102; there is no settled figure Official estimate Science
2013 Chorabari Lake, above Kedarnath 169 confirmed, 4,021 missing for the whole June 2013 Uttarakhand floods; the state later put it above 6,000, and no final count exists Official estimate National Institute of Disaster Management
1985 Dig Tsho, Nepal 4 or 5 deaths — the figure of twenty or more in circulation is unsourced Official estimate ICIMOD
2021 Chamoli, Uttarakhand — and not a GLOF 204 missing, presumed dead Official estimate Science
What happened in each, and what was learned
South Lhonak Lake, Sikkim · 3–4 October 2023
South Lhonak Lake, Sikkim. 14.7 million m³ of frozen moraine collapsed into the lake, raising a wave around 20 m high that breached the dam and released some 50 million m³ of water. It destroyed the 1,200 MW Teesta-III dam at Chungthang and travelled 385 km down the Teesta into Bangladesh, damaging about 25,900 buildings and 31 major bridges. There was no functioning early warning system; what warning reached people came informally from border police upstream.
Chorabari Lake, above Kedarnath · 16–17 June 2013
Chorabari Lake above Kedarnath, Uttarakhand, overtopped its moraine and drained in five to ten minutes — the second phase of the disaster, after a rainfall-driven debris flow the day before. A genuine lake outburst, though the lake was rain-fed rather than glacier-fed.
Dig Tsho, Nepal · 4 August 1985
Dig Tsho, Nepal. An ice avalanche into the lake overtopped and cut the end moraine; six to ten million m³ drained in about four hours, destroying a nearly-complete hydro plant 11 km downstream and 14 bridges. It struck during a lull in the trekking season.
Chamoli, Uttarakhand — and not a GLOF · 7 February 2021
Chamoli, Uttarakhand — commonly filed as a glacial lake outburst and not one. About 27 million m³ of rock and glacier ice fell from Ronti Peak and became a debris flow. No lake was involved; the peer-reviewed account was written specifically to correct the early reporting.

What these counts leave out

  • Lakes below the monitoring threshold. CWC monitors above 10 hectares; the July 2025 Rasuwa flood came from a supraglacial lake in Tibet that had not been noticed at all.
  • Anything upstream of the border. India has no gauge above the Tibetan frontier on the Brahmaputra system, and Nepal's Bhote Koshi warning system has a lead time of about six minutes because its sensors are all on the Nepali side of a catchment that begins in Tibet.
  • The distinction between hazard types, in early reporting. Rock-ice avalanches, landslide-dam breaks and true lake outbursts are routinely reported as one another in the first days, and the correction usually arrives months later in a journal.

What happens next

Two horizons. A level is printed only where a published threshold and a number exist to put something against; everything else is named as a watch, not graded.

Next 24–72 hours

Moderate More rain over the region Peak 28.6 mm in one day across the next 3, against IMD's heavy-rainfall threshold of 64.5 mm and very heavy at 115.5 mm.
Not assessed Whether an official agency confirms the mechanism. Himalayan disasters are routinely reported as one thing and found months later to be another — Chamoli 2021 was called a glacier burst and was a rock-ice avalanche.
Not assessed River levels at the Central Water Commission's downstream gauges, and whether any cross a warning or danger mark.
Not assessed Whether Copernicus Emergency Mapping is activated, which would produce the first authoritative extent maps.
Not assessed Whether a landslide dam formed in the flood path. A blockage that fills and then fails produces a second flood, often larger than the first.
Not assessed The casualty figure settling. Early counts in Himalayan disasters move by an order of magnitude over the first week, and the missing are counted differently by different agencies.

Coming days and weeks

Not assessed Sediment and debris moving down the river system for months. The 2023 Teesta flood was traceable 385 km downstream and into Bangladesh.
Not assessed Slopes destabilised by the flood failing again in later rain. The valley walls a debris flow has cut are the next monsoon's landslides.
Not assessed Roads and bridges out for a season rather than a week, which in a Himalayan valley means the supply route as well as the tourist route.
Not assessed Displacement that does not reverse. Villages in a newly-cut channel are not rebuilt in the same place, and the count of people who did not go home is not published anywhere.
Not assessed Hydropower and irrigation infrastructure in the path, and whether it is rebuilt in the same place with the same assumptions.

The rainfall row is computed from a forecast model over a representative point for the region — Open-Meteo forecast API, keyless; ECMWF-driven forecast model, not a gauge — not from a gauge at the event. The ungraded rows are the mechanisms that follow this hazard, drawn from the standing research pack. They are things to watch, and this page does not pretend to know their probability.

Is this a climate signal

What is established about this KIND of event, at scale, with citations. Read the two statements at the foot of this band together; neither is complete on its own.

25,614 Glacial lakes mapped across the Hindu Kush Himalaya Confirmed ICIMOD
2,485above 10 ha Glacial lakes CWC monitors Confirmed Central Water Commission
195 Lakes on India's high-risk list Confirmed National Disaster Management Authority
766events, 1533–2025 Recorded outbursts across High Mountain Asia Confirmed ICIMOD
15 million People exposed worldwide Confirmed Nature Communications
+65% Glacier mass loss, one decade against the last Confirmed ICIMOD

What science can say

Across the Third Pole, outburst risk is projected to roughly triple by 2100, with lakes rated high or very high risk rising from 1,203 to 2,963 and the eastern Himalaya the hotspot under every scenario. ICIMOD's own caveat travels with it: a higher hazard has not yet produced a higher rate of floods.

What has not been attributed

Nobody has established that climate change caused THIS event. Single-event attribution takes months of modelling and for a glacial or landslide-driven flood it is often not possible at all. A rising hazard across a region is not a finding about one flood in it, and this page will not print one as though it were.

The full standing pack for this hazard (12 sources)

A glacial lake outburst is a flood with no rain in it. Meltwater collects behind a moraine or an ice dam, something displaces it — a rockfall, an avalanche, a thaw — and the dam fails in minutes. The Himalaya holds tens of thousands of these lakes, India monitors a fraction of them, and the warning systems are almost entirely unbuilt.

25,614 — Glacial lakes mapped across the Hindu Kush Himalaya. A 2005 Landsat baseline, not a current count — lakes have grown and formed since, so this understates today's number. India's own basin atlas counts 28,043 lakes above 0.25 ha, but across whole river basins including Pakistani, Chinese and Nepali territory, so it is not a count of lakes in India. ICIMOD

2,485 above 10 ha — Glacial lakes CWC monitors. Plus 358 other water bodies, monitored June to October on Sentinel imagery. The portfolio grew from 902 in 2025. Of these, 681 glacial lakes lie inside India: Arunachal 324, Ladakh 179, Jammu & Kashmir 76, Sikkim 72, Himachal 17, Uttarakhand 13. Central Water Commission

195 — Lakes on India's high-risk list. Himachal 48, Sikkim 41, Ladakh 34, Arunachal 32, Jammu & Kashmir 27, Uttarakhand 13. By type, 103 are moraine-dammed. A 2026 parliamentary reply put 56 of them in the highest risk class. National Disaster Management Authority

766 events, 1533–2025 — Recorded outbursts across High Mountain Asia. About 59 of them in India, per the peer-reviewed version of the database. ICIMOD's own caveat: the record is an underestimate, and the apparent rise over time is mainly better observation rather than more events. ICIMOD

15 million — People exposed worldwide. Living BOTH within 50 km of a glacial lake AND within 1 km of a potential flood path — the two conditions together, which is the part most often dropped in quotation. India holds the largest exposed population of any country, about 3 million. India leads on exposure; on the study's danger index, which weights vulnerability, China and Pakistan rank higher. Nature Communications

+65% — Glacier mass loss, one decade against the last. From −0.17 to −0.28 metres water equivalent a year, between 2000–2009 and 2010–2019, with high confidence. Losses are steepest in the eastern Himalaya. ICIMOD

1,203 → 2,963 — Lakes at high or very high risk, now and by 2100. Outburst risk across the Third Pole is projected to roughly triple, with the eastern Himalaya — the transboundary basins above Bihar and Bengal — remaining the primary hotspot under every scenario. Nature Climate Change

none — End-to-end warning systems commissioned in India. The government's own position as of 2026, under a ₹150 crore four-state programme approved in July 2024, of which ₹27.73 crore of ₹135 crore had been released by mid-2025. Two automatic weather stations are confirmed installed, in Sikkim. Press accounts describing twenty or more instrumented lakes contradict the government's own answer. Ministry of Home Affairs / NDMA

Figures deliberately not published here

  • 'About 200 glacial lakes in the Hindu Kush Himalaya are dangerous' — the figure appears in an ICIMOD press release but nowhere in the report it is attributed to, and its likely ancestor is an inventory built partly on topographic maps from 1960–1982.
  • '178 dead at South Lhonak' — contradicted by every readable source; the disputed range runs from 55 dead with 74 missing to about 102.
  • 'India has the highest GLOF danger in the world' — India has the highest exposed population. On the danger index, which weights vulnerability, China and Pakistan rank above it.
  • '67 Indian glacial lakes have expanded by more than 40%' — absent from the ISRO page it is attributed to. ISRO's published finding is that 130 lakes in India expanded notably since 1984.
  • '47 potentially dangerous glacial lakes in Nepal' — the 47 is a three-country figure. Nepal's share is 21, China's 25 and India's 1.

India

Nepal is outside India. This band is why an Indian site is carrying it, and it separates what has been counted from what has not.

Confirmed India impact

about320 Indians missing or uncontactable Preliminary A News, ThePrint, ANI News and 4 more

Potential, and under watch

Nepal’s rivers drain into Bihar and eastern Uttar Pradesh. An outburst or a landslide dam on the Koshi, Gandak or Karnali becomes an Indian flood within days.

Downstream: Bihar and eastern Uttar Pradesh on the Koshi and Gandak; Assam and north Bengal on the Brahmaputra and Teesta. The Koshi basin is 2.6% of the Himalaya's southern slope and has produced 51.6% of its recorded glacial lake outbursts.

India has the largest population exposed to glacial lake outburst floods of any country — about 3 million people live both within 50 km of a glacial lake and within 1 km of a potential flood path. The hazard is also inherently transboundary: the Koshi basin is 2.6% of the Himalaya's southern slope and has produced 51.6% of its recorded glacial lake outbursts, and its upper catchments are in Tibet and Nepal. India's own high-risk lakes sit in Himachal, Sikkim, Ladakh, Arunachal, Jammu & Kashmir and Uttarakhand.

This column is a mechanism, not an outcome. These are the basins and districts the hazard reaches when it reaches India. Nothing here is a report that it has.

Data and sources

Everything this page is built from, counted by kind. Nothing here is quoted in this site’s own voice.

24 published reports
125 independent publishers
View all 24 sources

Reported

A headline is evidence that something was said. It is never evidence that it is true — which is why every figure at the top of this page carries the outlet that printed it and, where they disagree, all of them.

What this page does not know
  • Casualty and damage figures are not established here. Any number in the headlines below belongs to the outlet that printed it.
  • The mechanism is this page's reading of how the event was first reported, not a finding. Himalayan disasters are often reclassified once fieldwork is done.
  • The extent of the affected area has not been verified against satellite or official mapping.
  • Downstream consequences for India are a known mechanism for this hazard, not a confirmed outcome of this event.

Method: assembled automatically from 308 items across 125 publishers by scripts/detect-climate-events.mjs, which scored this event 22 against a publication threshold of 14. Assembled automatically from published headlines and official alert feeds. Every figure is a count this script performed; no claim about the event is made in its own voice.

Feeds last read 18:53 IST, 28 Aug 2026, and re-read every 30 minutes.

This is one event

The standing picture — how much extreme rain India actually gets, what IMD counts as extreme, who the rain kills, and what is changing across twelve cities — is on the situation page this belongs to. That page moves slowly on purpose. This one moves every half hour.

India’s extreme rain

Every reading, kept

Each reading keeps its own address, with the source that produced it, when it was observed, and the limit it was judged against. Nothing is overwritten when it improves and nothing is quietly restated when it gets worse. An empty day stays empty — a gap in the record is a gap in the record, never a zero.

Cite this page

Reuse freely — CC BY 4.0. Every figure carries its source and its cadence in how the number is made, and every figure carries whether it was counted or modelled on the rule beneath it. If you quote a number from here, quote the kind with it. The grant covers this page; each upstream source keeps its own terms, which is why every figure names one.

This is what we are doing.
It is not enough.

The work against this reading

Mostly the written record, and going to look. Three of the essays on this site were reported out of flood years, and Gram Anubhav takes groups into villages living with this, run with over a hundred grassroots partners in the villages themselves.

And why it is not a reply

Publishing and going to see are not resilience. Swechha does not build drainage, and nothing here reduces the 3,594 deaths this page counts. What the work does is make the count harder to ignore — a smaller claim than it sounds, and the only one the evidence supports.

Explore the workFifteen pages, four kinds Who is SwechhaWorking since 2000, and who runs it Take partGive, volunteer or partner

Once a month, what these numbers did.

One email. What the six readings did that month, what moved and what did not, and what we did about it. Nothing else, ever — no appeals between digests, and your address is never shared with anybody.

Confirm by email before anything is stored. Every message carries a one-click link to stop them.