Special report  /  Climate and responsibility 2 September 2026  /  20 min read
LATEST 2 September, 09:00 — Nepal's disaster authority revised the confirmed death toll to 1,114. The figures throughout this report are the 1 September bulletin, the last one published with a district-level breakdown. Numbers were still rising at publication.
08:37, 26 August 2026 ice and rock leave the mountain

The mountain that fell

Nepal produces four hundredths of one percent of the world's carbon dioxide. In August, a Himalayan glacier collapsed and killed a thousand of its people. This is the arithmetic of that sentence.

KATHMANDU / RASUWA — A data investigation into emissions, responsibility and the flash flood of 26 August 2026, compiled from the Global Carbon Project, the World Meteorological Organization, ICIMOD, the IMF, the UNFCCC, Ember, Nepal's disaster and electricity authorities, and the international press. Verified to 2 September 2026. Casualty numbers were still rising at publication.

1,003confirmed dead in Nepal, 1 September
3,916still missing, including 583 foreign nationals
11,884people rescued
0.04%Nepal's share of global fossil CO₂
748 MWof hydropower damaged in one morning
$5 bnpreliminary cost of rebuilding, about 11% of GDP

At thirty-seven minutes past eight on the morning of Wednesday 26 August, seismometers in the Himalaya registered what was first reported as a magnitude 4.4 earthquake near the border between Nepal and China. It was not an earthquake. It was a mass of ice and rock leaving the side of Langtang Lirung, a peak of 7,227 metres, and hitting the valley floor hard enough to shake the ground at a magnitude of 5.2.

What came down afterwards was not a flood in the ordinary sense. It was a debris flow — ice turning to water under the energy of its own collapse, mixed with saturated sediment, river water, boulders and mud. It ran roughly a hundred kilometres down the Lhende Khola into the Bhote Koshi and then the Trishuli. In places it reached eighty metres above the normal river level. It destroyed the Rasuwagadhi border post, the customs house, thirty-two bridges and the entire forty-two-kilometre road linking Betrawati to the Chinese frontier. It carried the bodies of some of its victims two hundred and forty kilometres, across the border, into India.

Nepal is a country of about 29.6 million people with almost no heavy industry, no oil, gas or coal of its own, and one of the cleanest electricity systems on Earth. It emits roughly one tonne of carbon dioxide for every two and a half thousand the world emits. Seventy-three percent of the new cars it bought last year were electric — a higher share than anywhere on the planet except Norway. Its forest cover has nearly doubled since 1992. The average Nepali uses about 460 kilowatt-hours of electricity a year; the average American uses twenty-eight times as much.

Nepal is also, on the long-term Climate Risk Index, the tenth most affected country in the world.

The world emits Nepal's entire annual carbon dioxide output in three hours and forty minutes.

That line is the argument of this report. What follows is the evidence — and, because a document meant for publication should survive a hostile reading, a clear account of where the argument is more complicated than a slogan.

CHAPTER ONE

Ninety seconds, one hundred kilometres

The disaster began on a mountain almost nobody lives on, and killed most of its victims in districts a hundred and fifty kilometres away.

INTERACTIVE — RECONSTRUCTION

How fast it travelled

Published reconstructions put the first 22 kilometres of the flow at about 6 minutes 50 seconds — an average of roughly 193 km/h, faster than almost anything that moves on the ground in Nepal. Press play to run that segment against the clock. Nobody downstream had a warning system that works on this timescale.

DISTANCE DOWN THE VALLEY 0 km22 km 50 km75 km100 km Langtang Lirung Trishuli valley Chitwan and Nawalparasi 22 km in 6 min 50 s the flow continued roughly 100 km in all — later timings were never published The horizontal axis is distance travelled, not a map. Only the shaded first segment has a documented duration.
ELAPSED0:00
DISTANCE0.0 km
AVERAGE SPEED0 km/h

Sources and caveats: the 22 km / 6 min 50 s figure comes from published reconstructions of the flow summarised by Britannica and EarthSky. A separate rapid analysis by the Stimson Center reports the wave reaching the Rasuwagadhi border facilities about 7.5 minutes after the collapse and gives an average of roughly 150 m/s over the run; the two reconstructions are not fully consistent with each other, so only the first is animated here. The impact itself registered as a magnitude 5.2 seismic signal. Copernicus Emergency Mapping and the International Charter on Space and Major Disasters were both activated within days.

Where it started, and where the dead were found schematic map of Nepal; not to scale Langtang Lirung, 7,227 m collapse origin, on the Tibet border Kathmandu Thame — glacial lake burst, Aug 2024 Rasuwa Nuwakot Dhading Gorkha Tanahun Chitwan and Nawalparasi rivers continue into India — bodies found 240 km downstream Bodies recovered, by district NDRRMA, 1 September 2026 — total 1,003 Chitwan 321 Nawalparasi E216 Nawalparasi W170 Nuwakot 95 Gorkha 65 Dhading 57 Rasuwa 41 Tanahun 38
Sources: National Disaster Risk Reduction and Management Authority bulletin, 1pm, 1 September 2026; US Geological Survey landslide hazards programme; UN OCHA ReliefWeb. The district where the flood began recorded 41 dead. The district a hundred and fifty kilometres downstream recorded 321.

INTERACTIVE — ONE DOT, ONE PERSON

The 1,003 confirmed dead, by the district where they were found

Every dot below is one person whose body had been recovered and confirmed by 1 September. Dots run north to south, downstream from the collapse: blue for the five upper districts, red for the three southern ones. The mountain fell in Rasuwa — the 41 pale dots in the first row. The red block that fills the lower half of the grid is Chitwan and the two Nawalparasis, between a hundred and two hundred kilometres away. Select any district to isolate it.

Source: NDRRMA bulletin, 1pm, 1 September 2026. District totals sum to exactly 1,003. Nepal's Ministry of Health reported that only about 4% of recovered bodies had been formally identified at that point. On the morning of 2 September the NDRRMA revised the confirmed total upward to 1,114; this grid shows the 1 September district breakdown, which is the last figure published with a district split.

The geography of the death toll is the first thing to understand. Rasuwa, where the mountain fell, is thinly populated. The wave gathered mass and speed as it descended, and by the time it reached the wide, flat, densely settled valleys of Chitwan and Nawalparasi it was still lethal. Nearly seven in ten of the bodies recovered were found in the three southernmost districts on that list.

Among the 3,916 people still missing in Nepal on 1 September: 1,558 from Nuwakot, 865 from Rasuwa, 639 connected to hydropower construction sites, and 583 foreign nationals from thirty-eight countries — pilgrims and trekkers heading for Mount Kailash at the height of the season. Sixteen more died in Tibet, where a further 546 are missing, and the Gyirong border checkpoint on the Chinese side was destroyed outright.

By 1 September, Nepal had mobilised 21,314 security personnel: 9,199 from the army, 7,912 police, 4,203 armed police. Relief reached Rasuwa and Nuwakot mostly by helicopter, because the road no longer existed. At the Rasuwagadhi hydropower project, rescuers were digging towards a tunnel entrance blocked seven metres short of the powerhouse. At the 22 MW Chilime plant, the executive director told the Kathmandu Post that the tunnel is filled with mud and it is impossible to enter. Indian tunnel-rescue specialists and Chinese teams joined the operation. Qatar flew in humanitarian aid on 31 August.

One valley, two mornings illustration of a typical Bhote Koshi settlement, based on published damage assessments 25 August 26 August
Officials confirmed 32 bridges and roughly 40 kilometres of road destroyed. Satellite imagery released by the US Geological Survey showed settlements at Syaprubesi and Rasuwagadhi flattened and the ground surface smoothed away. Upper settlements, higher on the valley walls, largely survived.
How a warming mountain becomes a wall of mud the mechanism, in four stages 1. Ice holds rock glacial ice buttresses the slope beneath and behind it 2. The ice retreats warming thins the glacier; the rock face is left hanging 3. Collapse millions of tonnes fall; impact energy converts ice to water 4. Debris flow water, rock and sediment run 100 km down a settled valley
No glacial lake burst here. Researchers believe the water came from ice converted under the energy of the collapse, water held in the sediments the flow picked up, and the river itself. An international scientific team began work within days.

Still unstable

A rapid hazard assessment published days after the event warned that potentially unstable masses of rock and ice remain on the glacier, that lakes have not drained from the base of the valley, and that smaller follow-on events are possible. Slopes downstream have been undercut, and an abundance of loose sediment will raise the threat from any future flash flood in that valley for months or years. Heavy rainfall is the specific trigger to watch. The monsoon was still running when this report went out.

CHAPTER TWO

The clock

Nepal emits roughly 16 million tonnes of carbon dioxide a year. Here is how long the world's major emitters need to produce the same amount.

Time taken to emit what Nepal emits in a year based on 2025 shares of global fossil CO₂; a shorter bar means a faster emitter The whole world China United States India European Union 3 hours 41 minutes 11 hours 30 minutes 1 day 4 hours 1 day 22 hours 2 days 13 hours
Calculated from Global Carbon Budget 2025 country shares of 38.1 GtCO₂, against Nepal's territorial fossil and industrial CO₂ of roughly 16 Mt. On an all-greenhouse-gas basis, which includes Nepal's substantial agricultural methane, the world takes about ten hours rather than under four.

LIVE — THE ATMOSPHERE, WHILE YOU READ

The counter that makes the argument

Global fossil CO₂ reached 38.1 billion tonnes in 2025 — 1,207 tonnes a second. Nepal's roughly 16 million tonnes a year works out at half a tonne a second. Both counters started the moment this page loaded.

TIME ON THIS PAGE  00:00:00

The world38.1 bn tonnes a year
0TONNES CO₂
Nepalabout 16 mn tonnes a year
this bar is drawn wider than the data
0TONNES CO₂
×2,381 Whatever the numbers on the left read, the ratio between them never changes. The world needs 3 hours 41 minutes to emit what Nepal emits in a year — and has already emitted a full Nepal-year since you opened this page.

Method: 38.1 Gt ÷ 31,557,600 seconds = 1,207.3 t/s; 16 Mt ÷ the same = 0.507 t/s. Both are steady-rate approximations of an annual total, not a live measurement of the atmosphere. The two bars are a fixed proportional comparison of those rates: at true scale Nepal's would be about a third of one pixel wide, so it is drawn at three. Sources: Global Carbon Project, Global Carbon Budget 2025; Our World in Data Nepal profile.

A different way to see the same thing: Nepal's share of world population is 0.36%. Its share of world carbon dioxide is 0.04%. It contains about one person in every 277 alive, and produces about one tonne in every two and a half thousand.

The reason this matters is that the atmosphere is a single, shared, well-mixed reservoir. There is no version of the physics in which Nepal's restraint protects Nepal. A country can decarbonise completely and still lose a valley.

CHAPTER THREE

The ledger

Climate change is a stock problem. Carbon released in 1910 is still warming the planet today — which makes the question not only who is emitting now, but who has emitted since industrialisation began.

Between 1850 and 2025, humanity put about 2,651 billion tonnes of CO₂ into the atmosphere from fossil fuels, cement and land-use change. Those emissions have consumed more than 95% of the carbon budget compatible with holding warming to 1.5 °C, and they are the reason the planet is now roughly 1.4 °C hotter than in pre-industrial times. They were not distributed evenly.

Every tonne of CO₂ emitted since 1850, by who emitted it billions of tonnes, from fossil fuels, cement, land use and forestry United StatesChina European Union (27)Russia Nepal 542336 ~318185 20.4% of everything ever emitted, from 4% of the world's people 12.7%12%7.0% under 0.1% — this bar is drawn wider than the data
Source: Carbon Brief analysis of Global Carbon Project, CDIAC, Our World in Data, IEA and Carbon Monitor figures, January 2026. Nepal's share of global CO₂ from fossil fuels and industry since 1751 is 0.01% (Our World in Data). Counting all greenhouse gases and land use, its historic warming contribution is roughly 0.1% (Climate Change Tracker).

Britain, where the Industrial Revolution began, was responsible for more than half of all cumulative emissions on Earth until 1882. Until around 1950, more than half of everything ever emitted had come from Europe alone. That is what "developed nations industrialised first" means in practice: wealth in Europe, North America, Japan and Australia was built across roughly two centuries of unpriced, unrestricted fossil combustion. The atmospheric bill for that period is still being paid — but not, for the most part, by the countries that ran it up.

Share of the world's people, against share of the world's carbon population today; CO₂ share of the cumulative 1850–2025 total people carbon United StatesEuropean Union ChinaNepal 4.0% 20.4% 5.5% 12% 17.5% 12.7% 0.36% under 0.1%
The United States has emitted five times more carbon than its population share would suggest. China, the largest emitter today, has emitted slightly less than its population share across history.

And who is emitting now

History explains the warming already locked in; current emissions decide how much more is coming. In 2025, global fossil CO₂ reached a record 38.1 billion tonnes, with coal, oil and gas all contributing.

Nepal sits between the world's largest emitter and its third largest. China accounts for about 32% of global fossil CO₂ and India about 8%: together, the two countries wrapped around Nepal produce two fifths of the world's carbon dioxide. This is not abstract for Nepal. Black carbon — soot from cooking, transport, brick kilns and industry across South Asia — settles on Himalayan snow and ice, darkens it, and makes it absorb more sunlight. It is a regional pollutant with a direct physical line to Nepal's glaciers.

But volume is not the same as responsibility, and a report that blurs the two is easy to dismiss. Per person, the picture inverts.

Carbon dioxide per person, 2024 tonnes of fossil CO₂ per inhabitant per year QatarUnited States ChinaWorld average IndiaNepal 41.314.28.7 4.72.2 about 0.5
Source: Our World in Data and the Global Carbon Budget, 2024 data. The average American emits more carbon dioxide in about thirteen days than the average Nepali emits in a year. The average Qatari does it in under five.

India's 2.2 tonnes per person is less than half the world average — lower than Portugal, lower than Argentina, lower than every country in western Europe. China's 8.7 tonnes are well below Australia's, Canada's or America's. Both countries are also deploying more solar and wind capacity than anyone in history: Chinese emissions grew just 0.4% in 2025 and India's 1.4%, both far below their own decade-long trends, while emissions rose 1.9% in the United States and 0.4% in the European Union after years of decline.

So the honest framing is not that China and India are doing this to Nepal. It is this. A small number of wealthy countries created the problem across two centuries and some are now reversing their own progress. Two very large developing countries have become major current emitters while remaining modest per-person ones. And a very small country between them, which did neither, is being handed the physical consequences first.

INTERACTIVE — THE SAME COUNTRIES, THREE LEDGERS

Who is responsible depends entirely on which book you open

Switch between the three measures. The ranking moves; the country at the bottom does not.

Sources: cumulative shares from Carbon Brief's January 2026 analysis of historical responsibility; per-person figures from Our World in Data and the Global Carbon Budget, 2024. The current-emissions shares are this report's own arithmetic from the Chapter Two clock — if the world needs 221 minutes to emit Nepal's annual total and China needs 690, China's share is 221÷690, or 32%. That method returns 32% for China and 8% for India, matching the published figures, and yields 13.2% for the United States and 6.0% for the European Union.

Where the emissions actually go

Territorial accounting — which country the chimney stands in — flatters rich consumers. A large share of Chinese and Indian industrial emissions produces goods bought in Europe and North America. On a consumption basis, which counts carbon embedded in imports, the wealthy world's footprint grows and the manufacturing world's shrinks. Neither method moves Nepal off the bottom of the table.

CHAPTER FOUR

What Nepal actually did

The moral weight of this report rests on a factual claim worth testing: that Nepal has not merely failed to pollute through poverty, but has actively built a low-carbon economy. The record supports it — with one important qualification.

Electricity

Nepal's installed generating capacity reached about 3,878 megawatts by mid-2025, after adding 631 MW in a single year. Essentially all of it is hydropower and solar; fossil-fired thermal amounts to roughly 53 MW, a rounding error. More than 95% of domestic electricity supply is renewable. The country has no oil, gas or coal production of its own, and now exports surplus power to India and — since 2024, and in US dollars — to Bangladesh.

Transport

In 2019, electric vehicles were 8% of new car sales in Nepal. By 2025 they were 73% — second in the world behind only Norway, and ahead of Denmark, Sweden, Germany, the United Kingdom and the United States. It was achieved not by wealth but by policy: tax concessions, reduced customs duties, charging-corridor investment, and the plain logic that a country with hydropower and no oil should not be importing petrol. Around 490 charging stations were operating nationwide by early 2026.

Electric share of new car sales, 2025 battery-electric and plug-in hybrid, percent of new passenger cars NorwayNepal DenmarkSweden IcelandFinland United KingdomGermany United StatesIndia 97%73% 69%61% 57%56% 33%29% 10%5%
Source: International Energy Agency and Ember, compiled 2026; Nepal Department of Customs. Nepal's income per person is about $1,180 a year — roughly one hundredth of Norway's.

The qualification: Nepal is clean, and also energy-poor

An argument for climate justice that presents Nepal purely as a green success story is missing half the picture, and the missing half is the more important one. Nepal's emissions are low partly because Nepal is virtuous and partly because Nepal is poor. Its people use a fraction of the electricity available to citizens of the countries that caused the warming.

Electricity used per person, per year kilowatt-hours, 2025 NorwayCanada United StatesChina World averageIndia Nepal 27,81816,090 13,0607,460 3,8601,420 about 460
Sources: Ember and national electricity consumption data, 2025; Nepal figure calculated from the Ministry of Finance total of 13,699 GWh against a population of 29.6 million. Norway, the only country ahead of Nepal on electric-car adoption, uses about sixty times more electricity per person. Nepal's clean-energy record was set at a level of consumption no wealthy country would accept.

INTERACTIVE — MEASURE ANY COUNTRY AGAINST NEPAL

Pick a country

Each comparison is drawn only from the datasets already charted in this report, so nothing here rests on a figure the sources above do not carry. Where a country is absent from one of those datasets, that measure is simply left out.

Sources: carbon dioxide per person — Our World in Data and the Global Carbon Budget, 2024. Electricity per person — Ember and national statistical agencies, 2025; Nepal calculated from the Ministry of Finance total of 13,699 GWh against 29.6 million people. Electric share of new car sales — IEA and Ember, 2025, with Nepal from the Department of Customs.

This matters for the policy argument. Nepal's energy use will and should rise as its people gain access to cooking, heating, refrigeration and industry. The right ask is not that Nepal stay poor and clean. It is that Nepal be financed to grow clean — and compensated for damage it did not cause while doing so.

Forests

Between 1992 and 2016, Nepal's forest cover rose from 26% to about 45% of national territory, and stands near 46% today. The near-doubling came from community forestry: more than 22,000 community forest user groups, comprising around three million households, managing roughly a third of the country's forest. NASA's Earth Observatory documented the recovery from Landsat records. In November 2025 Nepal received its first results-based payment of $9.4 million from the World Bank's Forest Carbon Partnership Facility.

Nepal's forest cover, as a share of national territory each square represents the whole country 26% 1992 the World Bank had predicted the hill forests would be gone by the 1990s 46% today 22,000 community forest user groups; around three million households
Nepal is one of very few countries on Earth to have meaningfully increased its terrestrial carbon sink in the modern era. Conservationists note the gains skew towards pine monoculture and have been a mixed result for biodiversity.

Commitments

Nepal's Long-term Strategy commits it to net-zero emissions by 2045 — five years earlier than the date most wealthy nations have set for themselves, and fifteen years earlier than China's. Its third Nationally Determined Contribution, filed in May 2025, targets a 17.1% cut in net greenhouse gases against business-as-usual by 2030 and 26.8% by 2035, with battery-electric vehicles at 90% of private passenger sales by 2030.

Nepal did the thing it was asked to do. It decarbonised early, cheaply, and from a standing start.

CHAPTER FIVE

What the warming is doing to the mountains

2025 was the second or third warmest year in the 176-year record, at about 1.44 °C above the 1850–1900 average. The eleven years from 2015 to 2025 are the eleven warmest ever measured. For the first time, a full three-year period — 2023 to 2025 — averaged more than 1.5 °C above pre-industrial levels.

High mountains warm faster than the global mean, and the Hindu Kush Himalaya is coming apart faster than almost anywhere outside the poles.

65%faster glacier loss across the Hindu Kush Himalaya in 2011–2020 than the decade before
rate of ice loss since 2000, per ICIMOD's March 2026 assessment
12%of total glacier area in the region lost between 1990 and 2020
27 mof ice thickness lost from monitored glaciers since 1975
¼of Nepal's own glacier surface area lost between 1977 and 2010
75–80%of remaining Himalayan ice volume that could go by 2100 on current emissions

Three quarters of the region's glaciers are small — under half a square kilometre — and small glaciers vanish fastest. As ice retreats it leaves meltwater lakes dammed by loose moraine, and it unbuttresses rock faces that ice had held in place for millennia. ICIMOD and UNDP's joint inventory identifies 47 potentially dangerous glacial lakes in the river basins feeding Nepal. Twenty-one lie inside Nepal. Twenty-five lie inside Tibet — upstream, across a border Nepal cannot survey, on rivers that run straight into Nepali valleys.

ICIMOD's 2026 report carries a warning about the evidence base itself: of 38 monitored glaciers across the entire range, only seven meet the World Glacier Monitoring Service's benchmark standards. Large parts of the Himalaya are, in the authors' term, blind spots.

A tightening sequence major water- and ice-driven disasters affecting Nepal, 2021–2026 2021 Melamchi floods August 2024 Thame glacial lakes burst, Everest region September 2024 Kathmandu valley floods and landslides, at least 236 dead July 2025 Lhende Khola outburst destroys the Rasuwagadhi border bridge — same valley 26 August 2026 Langtang Lirung ice-and-rock collapse 1,003 dead, 3,916 missing
The same river valley flooded catastrophically twice in fourteen months. A rock-and-ice avalanche had also come down on the other side of Rasuwagadhi in July 2024.

Nepal's own instrumental record shows mean maximum temperatures rising about 0.056 °C every year. Sixty-eight of its seventy-seven districts are landslide-prone. Roughly 80% of the population lives exposed to climate-induced hazard. Losses in the water and agriculture sectors alone have been estimated at 2–3% of GDP a year, before any single catastrophe is counted.

Was this event climate change?

Formal attribution takes months, and scientists have been careful. What can be said now:

Other researchers have cautioned against declaring this specific collapse a climate event before the international investigation reports, and this report states both positions. The consensus is not that warming pulled this particular cliff down on this particular morning. It is that warming is systematically loading the dice, and that this is the category of catastrophe the Himalaya will now produce more often.

CHAPTER SIX

What one morning cost

What was lostScale
Hydropower projects damaged14 projects, about 748 MW
— operating plants taken off the grid9 plants, 354 MW
— projects under construction5 projects, 394 MW
Road destroyed, Betrawati to Rasuwagadhithe entire 42 km
Bridges swept away32
Schools damagedat least 20
Electric vehicles destroyed in transit at customsabout 1,000
Preliminary damage to roads and bridgesNPR 200 bn (US$1.3 bn)
Finance ministry's preliminary rebuilding estimateup to US$5 bn

Two lines in that table should be read together. Around a thousand electric vehicles — the physical substance of Nepal's clean-transport transition, queued at the border waiting to enter the country — were destroyed by a flood caused by the warming those vehicles exist to avoid. And roughly a fifth of Nepal's entire generating fleet was damaged, because run-of-river hydropower must be built on riverbanks, and the riverbanks are exactly where the mountains are now failing.

Nepal's clean energy system and its climate exposure occupy the same valleys. Its virtue and its vulnerability have the same address.

The market registered the point immediately. On the day after the flood, shares in Rasuwagadhi Hydropower fell 15%, Mailung Khola 14.99%, Trishuli Jal Vidyut 12.95%, Chilime 10.12%. Private hydropower projects are generally insured; government-owned infrastructure often is not.

The $5 billion rebuilding estimate is roughly eleven percent of Nepal's entire annual economic output. Scaled to the United States, that is the equivalent of losing about $3.3 trillion in a single morning.

CHAPTER SEVEN

The bill, and who is paying it

Nepal knows what it needs. Its National Adaptation Plan costs out resilience to 2050 at $47 billion, of which the government could commit $1.5 billion. The rest was always assumed to come from international climate finance.

The adaptation priorities in Nepal's 2025 NDC — early warning systems, integrated watershed management, and specifically glacial lake outburst flood risk reduction — need $18–20 billion between 2025 and 2035. Set against that, here is the entire global machinery for compensating climate harm.

What has been paid, against what is needed billions of US dollars, all four bars on the same scale Total ever paid into theglobal Loss and Damage Fund Damage to Nepal's roadsand bridges, one flood Rebuilding afterthat one flood Nepal's adaptation planto 2050 $0.45 bn $1.3 bn up to $5 bn $47 bn for every developing country in the world, combined eleven times everything that fund has ever received
Sources: Fund for Responding to Loss and Damage reporting, 2026; Nepal Ministry of Finance preliminary estimates, August 2026; Nepal National Adaptation Plan, 2023. Pledges to the fund total roughly $730–768 million; the amount actually paid in is about $448 million.

The Fund for Responding to Loss and Damage was agreed at COP27 in 2022, after more than three decades of campaigning by vulnerable nations. Four years on, the money actually deposited in it — for every developing country on Earth, for every flood, drought, cyclone and failed harvest — is less than a tenth of what Nepal alone needs to rebuild from a single morning.

The same sums, measured against what the world spends on fossil fuels

The IMF calculates that governments spent $725 billion on explicit fossil fuel subsidies in 2024 — direct fiscal transfers holding fuel prices below supply cost. That is about $2 billion a day, every day. Counting implicit subsidies, which price in the unpaid costs of air pollution and climate damage, the figure rises to $6.7 trillion.

The same figures, priced in fossil fuel subsidy how long the world's explicit fossil fuel subsidies — about $2 bn a day — take to equal each sum Everything ever paid into theglobal Loss and Damage Fund Nepal's road and bridgedamage, 26 August Nepal's entire floodrebuilding bill Nepal's adaptation planto 2050 about 5 hours 16 hours 2 and a half days 24 days
Sources: IMF, Underpriced and Overused: Fossil Fuel Subsidies Data 2025 Update; other figures as above. The world could fund every adaptation measure Nepal has costed out to 2050 by pausing fossil fuel subsidies for twenty-four days.

And against the budget that is left

The carbon budget for 1.5 °C roughly 2,800 billion tonnes of CO₂ in total, counting everything since pre-industrial times 2,651 bn tonnes already emitted — about 95% of the budget what remains roughly four years at 2025's rate of 38.1 bn tonnes a year
Carbon budget estimates carry large uncertainty and are revised as the science develops; this is an order-of-magnitude illustration, not a countdown clock. Sources: Carbon Brief and the Global Carbon Project.

The wider finance settlement is no more reassuring. At COP29 in Baku, developed countries agreed a new goal of $300 billion a year by 2035, with a broader call to mobilise $1.3 trillion. Developing countries had asked for the $1.3 trillion as the core commitment. Meanwhile the largest single historic emitter — the United States, with 542 billion tonnes to its name — has moved to withdraw from the UN climate treaty entirely.

The status trap

Nepal was scheduled to graduate from the United Nations' Least Developed Country category on 24 November 2026 — twelve weeks after the flood. Graduation removes access to LDC-specific concessional finance, preferential trade terms and dedicated climate funds, at the precise moment the reconstruction bill arrives.

In May 2026 the government formally asked the UN Committee for Development Policy to defer graduation to 2029, citing geopolitical tension, supply-chain disruption and remittance pressure. The request predates the flood. It has not become less necessary since.

CHAPTER EIGHT

What this report does not claim

An argument about fairness is only as strong as its willingness to state the other side. Five qualifications, each of which a critic would otherwise raise.

Nepal is not a zero emitter

Nepal emitted around 59 million tonnes of CO₂-equivalent in 2024, about 0.1% of the global total. Its largest single source of warming impact is methane, chiefly from livestock, with carbon dioxide close behind; agriculture accounts for over half of national emissions. Emissions are rising roughly 1.7% a year, against the 4% annual global decline needed for net zero by 2050. Nepal's contribution is negligible, not nil, and its own targets acknowledge that.

The specific event has not been formally attributed

An international team is still working on Langtang Lirung. This report links the disaster to climate change through an established mechanism — accelerating glacier retreat destabilising high mountain slopes — and through the explicit statements of cryosphere scientists, while noting that formal single-event attribution is pending.

Local decisions shaped the damage

Settlements, roads, customs infrastructure and more than a dozen hydropower projects were concentrated in a narrow, known-hazardous river corridor. Early warning coverage was thin. The same valley had flooded catastrophically thirteen months earlier. None of this caused the collapse, but it converted a geophysical event into a mass-casualty one, and Nepal's own institutions carry some responsibility for that.

Blaming China and India is the wrong frame

They are Nepal's neighbours and, in absolute terms, the first and third largest emitters. But per person they emit 8.7 and 2.2 tonnes — India's figure is under half the world average — and their cumulative historical shares are far below those of the countries that industrialised first. Both are decelerating faster than the rich world. Proximity is not culpability. The relevant asymmetry runs between the wealthy, high-per-capita, high-cumulative world and everyone else, and Nepal sits at the far end of that axis regardless of who its neighbours happen to be.

Hazard crosses borders that responsibility does not

More than half the dangerous glacial lakes threatening Nepal's rivers sit inside Tibet. Nepal cannot survey them, instrument them or drain them. Whatever the emissions ledger says, managing this particular risk requires cooperation Nepal cannot compel.

CHAPTER NINE

What justice would actually look like

The point of a climate justice argument is not to assign blame but to specify a remedy. Six things follow from the evidence above.

  1. Loss and damage finance must be real money, delivered fast, as grants. A fund holding $448 million cannot answer a $5 billion event. Nepal should not be borrowing to rebuild from other countries' emissions. The pathway campaigners have proposed — $50 billion a year from 2027, rising thereafter — is the minimum order of magnitude that matches the physics.
  2. Adaptation finance should follow exposure, not economic size. Nepal's $47 billion plan is costed, published and mostly unfunded. Funding it is cheaper than funding the disasters it would prevent, and equal to twenty-four days of global fossil fuel subsidy.
  3. Redirect the subsidy, not just the aid budget. The $725 billion a year that governments spend holding fossil fuel prices below cost is the single largest available source of climate finance, and spending it is a choice made annually.
  4. Cryosphere monitoring is a global public good and should be paid for globally. Only seven of 38 monitored Himalayan glaciers meet international benchmark standards. Early warning systems on the highest-risk lakes, and transboundary data-sharing with China and India, would have cost a fraction of one flood.
  5. Countries that decarbonise early should not be penalised for it. Nepal reached the world's second-highest EV share and near-total renewable electricity on a hundredth of Norway's income and a sixtieth of its electricity use. That should attract concessional finance and technology transfer — and it is an argument for deferring Nepal's LDC graduation, not accelerating it.
  6. Historic emitters should stay in the treaty they built. The largest cumulative emitter leaving the UNFCCC does not change the physics. It changes only who is at the table when the bill is discussed.

The next date that matters

COP31 opens in Antalya, Türkiye, on 9 November 2026 — sixty-eight days after this report was published, and fifteen days before Nepal is currently scheduled to lose its Least Developed Country status. Loss and damage implementation and delivery of the $300 billion finance goal are both live on the agenda. Türkiye holds the presidency; Australia leads the negotiations.

Nepal's case is not unusual in kind. It is shared by small island states, by the Sahel, by every country whose contribution to the problem rounds to zero. It is unusual only in clarity. Here is a country that produces four hundredths of one percent of the world's carbon dioxide, runs on hydropower, buys electric cars faster than almost anyone, has doubled its forests, and uses a twenty-eighth of the electricity an American takes for granted. On 26 August, a mountain that had been warming for forty years came down on it, and a thousand of its people died.

The atmosphere does not check where the carbon came from before it decides where the ice will fail.

RESPONSE

Where help is going

The IFRC has launched an emergency appeal of CHF 25 million. UNICEF is appealing for $17.2 million, having assessed more than 17,000 children as in urgent need, and has delivered hygiene kits, jerrycans, water purification supplies and disinfectant for around 19,000 people in Rasuwa and Nuwakot. CARE, World Vision and other agencies have declared national responses. The Government of Nepal has opened a relief fund, released NPR 10 million each to Rasuwa and Nuwakot, NPR 5 million to Dhading and NPR 67.5 million to fifteen affected local government units, and declared the flood-hit areas disaster zones for three months.

Readers wishing to contribute should give through established agencies with an operational presence in Nepal, or through the government's official disaster relief fund, and should verify bank details against each organisation's own published channels. Fraudulent appeals follow disasters of this scale as a matter of routine.

METHOD AND SOURCES

Where every figure came from

All figures were verified on 2 September 2026. Casualty and damage numbers from the August 2026 flood were provisional and changing daily; where sources conflicted, the most recent official Nepali government bulletin is used and the discrepancy noted. Death and missing totals are from the National Disaster Risk Reduction and Management Authority bulletin issued at 1pm on 1 September 2026. Derived figures — the emissions clock, the subsidy-days chart, the rebuild-to-GDP ratio — are this report's own calculations from the sources named, and the underlying arithmetic is stated in each caption so readers can check it.

Global emissions and shares. Global Carbon Project, Global Carbon Budget 2025; Carbon Brief analysis of the 2025 budget; CSIRO summary, November 2025.

Cumulative historical emissions. Carbon Brief analysis of historical responsibility, January 2026, drawing on Jones et al. (2023), Lamboll et al. (2023), the Global Carbon Project, CDIAC, Our World in Data, the IEA and Carbon Monitor.

Per-capita emissions. Our World in Data country profiles and the Global Carbon Budget, 2024 data.

Nepal's emissions. Our World in Data Nepal profile; Climate Change Tracker national pages; national inventory compilations.

Electricity consumption. Ember and national statistical agencies, 2025; Nepal Ministry of Finance total electricity consumption series (13,699 GWh, 2025).

Global temperature. World Meteorological Organization, confirmation of 2025 temperatures, January 2026, and State of the Global Climate 2025; Copernicus Climate Change Service.

Himalayan glaciers. ICIMOD, HI-WISE report (2023); ICIMOD, HKH Glacier Outlook 2026 and accompanying assessment, March 2026; ICIMOD and UNDP inventory of potentially dangerous glacial lakes.

Nepal energy and transport. Nepal Electricity Authority; Nepal Department of Customs; IEA and Ember EV sales-share data for 2025; Kathmandu Post and Climate Home News reporting; US Department of Commerce country commercial guide on Nepal's hydropower sector.

Nepal's forests. NASA Earth Observatory analysis of Landsat records, 1992–2016; RECOFTC; World Bank Forest Carbon Partnership Facility; Mongabay reporting on biodiversity outcomes.

Nepal's climate policy. Nepal's third Nationally Determined Contribution (May 2025); Long-term Strategy for Net-zero Emissions (2021); National Adaptation Plan (2023); UNDP Climate Promise.

The 26 August 2026 flood. US Geological Survey landslide hazards programme event page; UN OCHA ReliefWeb rapid situation overview, 27 August 2026; NDRRMA and Nepal Police bulletins; Nepal Electricity Authority damage assessments; Department of Roads; independent rapid hazard assessment, 28 August 2026; UNICEF, CARE and IFRC situation reports; reporting by NPR, NBC News, CBC, ABC News, Al Jazeera, the Christian Science Monitor, Bloomberg, the Guardian, Reuters and the Kathmandu Post.

Climate finance. Fund for Responding to Loss and Damage reporting and the Barbados Implementation Modalities; Climate Policy Initiative; World Resources Institute analysis of the New Collective Quantified Goal; NRDC and E3G climate finance tracker.

Fossil fuel subsidies. International Monetary Fund, Underpriced and Overused: Fossil Fuel Subsidies Data 2025 Update, Working Paper 2025/270.

LDC graduation. UN Committee for Development Policy; Nepal Ministry of Foreign Affairs deferral request, 13 May 2026; UNDP Nepal Smooth Transition Strategy.

Nepal economy and vulnerability. World Bank Nepal climate change technical note and Nepal Development Update, April 2026; World Bank national accounts.

Picture desk

Every graphic and illustration here is original vector work created for this report, carrying no third-party rights. The file can be republished exactly as it stands, with no licensing cost and no broken image links.

To add photography, use these sources in order of ease of clearance. US Geological Survey and NASA Earth Observatory satellite imagery of the Rasuwagadhi and Syaprubesi destruction is public domain in most jurisdictions. Copernicus Sentinel imagery is free to reuse under the Copernicus licence with attribution. Wikimedia Commons holds a Langtang Lirung category and extensive Nepali landscape, hydropower and Kathmandu transport imagery under Creative Commons — check each file's individual licence and credit the photographer as stated. Ground photography of the response is available through UN OCHA ReliefWeb, UNICEF and IFRC media libraries under their published terms. Photographs from Reuters, AP, AFP, Getty and Bloomberg require paid licences and must never be hot-linked.

Download any image you use, host it yourself, and place it as a standard figure immediately after the relevant paragraph, with photographer and licence in the caption. Suggested placements: the Langtang Lirung massif after the opening paragraph; satellite before-and-after at Rasuwagadhi alongside the two-panel illustration in Chapter One; a Kathmandu charging station or electric microbus in Chapter Four; a mid-hills community forest beside the forest-cover graphic.

Reuse and citation

This report is offered for free republication. Please attribute data to the original sources named above rather than to this document, and check current NDRRMA bulletins before republishing casualty figures.

Suggested citation: The mountain that fell: Nepal, the world's carbon, and the flood of 26 August 2026. Special report, 2 September 2026.