THE NEPAL FLASH FLOOD TO INDONESIAN WILDFIRES: WHEN NATURE STRIKES, WHO PAYS THE PRICE?

PHOTO: The site of the 2026 Nepal floods before the event, Sentinel-2 via Copernicus Browser on Wikimedia Commons

On the morning of August 26, 2026, the Trishuli River in Nepal rose without warning. In minutes, a wall of water and debris tore through the Rasuwa district, swallowing homes, a border crossing, and a highway that had connected Nepal to China for generations. Nearly 1,400 people are confirmed dead. Over 6,000 remain missing. Thousands of kilometres away, in the peatlands of Sumatra and Kalimantan, a different kind of disaster has been unfolding more slowly. Fires lit to clear land for plantations have burned for months, sending toxic haze across Indonesia and its neighbours, sickening tens of thousands and closing schools for children who now spend their days indoors, masked against air they cannot see. These two disasters present two differing perspectives of the impact of human decisions on ‘natural’ disasters; one presenting a ‘sudden’ disaster which tested government readiness, and the other a ‘gradual’ disaster which tested the long-term effectiveness of law enforcement.

From this, we can raise several questions: First, how much of what we call “natural” is actually a consequence of choices? Where we build, what we burn, whose warnings we heed, and whose we ignore? Neither began the day the water rose or the first hectare caught fire. Nepal’s flood swept through a corridor faster than anyone could react and there was no time for evacuation. But the severity is already reshaped by unplanned settlement in risky lands and early-warning infrastructure that is desperately in need of improvement. Indonesia’s fires were, in large part, caused by an economy built on access to drained peatland for palm oil, a pattern that repeats almost every year regardless of the weather. In both cases, the hazard was real and imminent.

Secondly, the consequences were not felt equally, so how were certain demographics more affected than others? The Nepal Food Security Monitoring System, NeKSAP, identified 940,000 food-insecure persons in the surrounding areas, with 300,000 needing life-saving food assistance. These same households – often too poor to relocate or rebuild elsewhere –  were the first to lose everything. Children have been pulled out of destroyed schools and elderly residents left with nowhere to go. In Indonesia, it is smallholder farmers and low-income families who breathe the worst of the haze, even though it is rarely their fires driving the crisis. This is borne out by Desia, a snack seller, in an interview with the South China Morning Post: “My house is surrounded by smoke … For more than a month, I haven’t opened the windows. It’s like we’re breathing toxin every day”. Additionally, children’s education has been hindered, as the school closures caused by the haze had forced 1.4 million students into remote learning. 

Lastly, were the warning systems in place adequate? Did emergency response reach affected communities quickly enough? And what happens once the immediate crisis fades from international headlines? Nepal’s early-warning infrastructure along the Trishuli corridor was limited, leaving little time between the glacier’s collapse and the flood reaching populated areas. Although Indonesia has passed peatland-burning bans before, the fires return most years regardless, raising the question of whether enforcement, not policy, is where the system is actually failing.

PHOTO: Imagery of the Indonesian Wildfires, NASA on Wikimedia Commons

 Why disasters are intensifying; vulnerability and land-use

Disasters are not exacerbated by accident, decisions made before, during and after a hazard determine whether it becomes a catastrophe. . In disaster studies, this is sometimes called the “pressure and release” model: a hazard – a flood, a fire, an earthquake – merely turns into a disaster when it reaches   an already vulnerable population. The hazard is often inevitable; however, the scale of the loss is not. For the Nepal flash flood, the trigger was a glacier failure high in Langtang National Park, a geophysical event that scientists had tracked for years. But the death toll was shaped by what lay downstream: settlements and infrastructure built into a known flood corridor, and years of underinvestment in disaster-risk planning. As Emily Wilkinson, Principal Researcher in ODI’s Global Risks and Resilience Programme,  told Mongabay, warming temperatures likely worsened the flood’s extent, but how badly households were hit came down to uneven development, weak investment in risk reduction, and social marginalisation.

However, in Indonesia, unlike Nepal, most of the fires across Sumatra and Kalimantan were not accidents of weather at all, but rather deliberately lit as a part of a land-clearing economy built on draining peatland for palm oil plantations. This is not a one-off event: it happens most years and is a recurring cost of an economy that treats cheap land clearance as worth the toxic smoke it produces. Whilst this year’s El Niño made the fires spread faster and burn longer, it did not create them.

The human dimension 

In both Nepal and Indonesia, the same groups keep appearing at the centre of the loss: children, the elderly, and low-income households. Children have borne some of the most disruptive costs. In Rasuwa, entire schools were destroyed, leaving families to rebuild not just shelter but also educational facilities. External observers such as Save the Children UK have also shed light on concerns regarding the spread of diseases amongst children amidst difficulty in accessing clean water. UNICEF estimated that millions of children around the world have had their schooling affected by the smoke in past fire seasons alone. In Indonesia, the disruption is no less serious: haze-related closures have kept students out of classrooms in Kalimantan for extended stretches. For families already stretched thin, missed schooling compounds into a longer-term loss that outlasts the disaster itself. The elderly face a different set of risks. In a flash flood, that risk is often literal, because of a lack of time and mobility to escape quickly. In a prolonged haze crisis, older residents with existing respiratory conditions are among the most likely to be hospitalized as air quality deteriorates over weeks, not hours. Indigenous and local communities carry a burden of displacement of not only the physical kind, but also cultural. As stated by Kyoko Yokosuka, UNDP Resident Representative in Nepal: “This is a devastating disaster for communities along the Trishuli corridor. Families have lost loved ones, homes and livelihoods, while entire communities are struggling to reconnect with essential services and one another.” In Indonesia, smallholder farmers, frequently blamed for fires that are, in large part, driven by industrial-scale plantation clearing, often find themselves both scapegoated and harmed. 

Underlying all these groups is a common thread: low-income households, across both countries, had the fewest resources to prepare for these disasters and the fewest resources to recover from them. Displacement has meant not just lost homes but lost income, savings spent on emergency needs rather than rebuilding, and in many cases, informal work that offers no safety net at all.

Government, policy, and response 

Then, was the damage foreseeable, and did the systems meant to prevent or reduce it actually work? Nepal’s flood exposed a structural blindspot in its warning system. The country’s flood gauges are built to track monsoon rainfall gradually raises river levels over hours – “a system which, for that purpose, works well”. It was never designed to detect an 80-metre-high wall of water and debris triggered by a glacier collapse upstream, arriving with no warning rainfall at all. The last functioning gauge on the river, at Devghat, registered what looked like an ordinary flood even as the disaster unfolded upstream. The equipment built to catch the real signal – the four monitoring systems at Langtang, Rasuwagadhi, Betrabati, and Syafrubesi –  had already been destroyed by the flood itself. This was not a one-off surprise: it was the second major flood in fourteen months, part of a pattern hazard scientists had already flagged.

As for the post-flood mobilization, government response was relatively quick: the government declared 15 municipalities disaster-affected areas, activated a Health Emergency Operation Centre with WHO, and deployed emergency medical teams to the region. But that  same destroyed infrastructure that eliminated the warning system also crippled responsive actions. With roads, bridges, and communications gone, officials in Rasuwa itself had to travel six hours to Nuwakot simply to relay information.

On the other hand, Indonesia’s fires present a different problem: existing policies that keep failing to change outcomes year after year. The government has some real capacity, deploying over 50 water-bombing aircraft over affected areas, resuming cloud-seeding operations, and taking legal action against roughly 30 companies over 2026’s fires, alongside the arrest of 72 individuals suspected of deliberate burning. However, the fires have now returned at scale nearly every year since a similarly severe season in 2015. Watchdog groups argue that law enforcement on peatland burnings has never functioned as a real deterrent, as well as warning that ordinary people and Indigenous communities risk being scapegoated, while the companies profiting from land conversion avoid consequences. Asia News argues that Indonesia’s state can mobilize dramatically after a fire starts, but preventing the conditions that make those disasters likely in the first place is a different story.

These cases represent mirror images of the same type of failures. Nepal’s institutions could not see the disaster coming because the hazard fell outside what the system was built to detect. Even with the hazard relayed by experts regarding the geographical and climate vulnerability in the area, the government failed to mitigate the risk through its infrastructure. Indonesia’s institutions see the disaster coming every single year, and still have not built a system that prevents it. These two cases demonstrate the flaws not in responsive actions, but in preventive mechanisms – specifically what happens in the years beforehand when the decisions that shape disaster severity are made.

A Silver Lining: Before the Water Rose, Before the Fires Lit

Nepal and Indonesia point to the same underlying truth from different perspectives. A flood and a fire are not, in isolation, disasters. They descend into catastrophe only after mismanagement and inefficient decision-making. In Rasuwa, despite the limited time available to respond and a flood-gauge system designed for an entirely different purpose, the disaster was still exacerbated by government policies and a lack of strategic infrastructural planning. While Indonesia presents a different set of challenges, the deliberate burnings of peatlands continues despite previous bans, driven by ineffective law enforcement. Both disasters demonstrate that hazards are intensified by land-use decisions; that costs are absorbed disproportionately by children and low-income households; and that institutions respond well once disaster is visible but fail in the quiet years beforehand. Global dialogue is intensifying as we approach COP31 in Antalya later this year, where Australia will lead the negotiations, including those on climate financing, which relate directly to government accountability. While natural disasters are not always avoidable, the large cost and severity are, and obscuring responsibility behind the word “natural” prevents accountability from forming. And if the causes of catastrophe are not random, then neither are the ones paying the price.