Why the Haze Comes Back Every Year | Indonesia's Fires Explained

Every dry season, it comes back. Kuala Lumpur. Singapore. Pekanbaru. Palangka Raya. Different countries, different air-quality scales, the same smoke. Tens of millions of people, breathing fires that are burning hundreds of kilometres away.

The region has a very full calendar. Chinese New Year. Hari Raya. Deepavali. Christmas. The school holidays. And one more festival that comes back every year. No public holiday. No lights. Just this.

And every year, a kid somewhere stands at a window and asks: why does this happen every year? The usual answer is, someone, somewhere, is burning a farm. That's not wrong. But it's not the whole truth either.

The real answer is stranger. It involves a fire you can't see, a swamp that was never supposed to burn, the shampoo in your bathroom, and a treaty that took one country twelve years to ratify. So I followed the smoke. Here's the truth behind the truth.

This film is brought to you by myna, the family app with a helper called Aunty. More on her at the end. Chapter one. Where it comes from. Start in orbit. This is the neighbourhood, from NASA's satellites: Malaysia, Singapore, Brunei, Indonesia. Now, here is every fire the satellites detected in one bad season: 2015.

They cluster in two places. Sumatra, across the Strait of Malacca. And Kalimantan, on Borneo. Which one you breathe depends on where you live. And the wind handles the delivery.

From June to September, the southwest monsoon blows across the region. It carries Sumatra's smoke to the Malay Peninsula and Singapore, and Kalimantan's across Borneo, to Sarawak, Brunei and Sabah. Smoke doesn't need a passport. It only stops when the monsoon turns, usually around October or November, and the rain comes back.

But look closely at where the fires are. Not up in the mountains. Not deep in the rainforest. Along the flat, swampy coasts. Because underneath those coasts is something most people never see. Peat. In 2015, across Sumatra, Borneo and the Malay Peninsula, peat was about 12 percent of the land, and had 53 percent of the fires the satellites spotted.

Peat is what you get when a forest is too wet to rot. Leaves, roots, whole trees fall into the swamp, and instead of breaking down, they pile up, over thousands of years. In Indonesia it can be up to twenty metres deep. That's a seven-storey apartment block of ancient, soggy forest.

And there's a lot of it. Southeast Asia has about 250,000 square kilometres of tropical peat. Let's measure that in a unit everyone in the region knows. Bali. That's roughly forty-three Balis. Minus the beaches. And worldwide, peat holds more carbon than all the world's forests put together.

Here's the thing. Wet peat doesn't burn. Normally, it's simply too wet. So how do you set a swamp on fire? You drain it first. Dig a canal, the water runs out, and the peat dries. It even sinks, about five centimetres a year. And dry peat is just fuel. Metres and metres of it.

When it catches, it does something a normal fire doesn't. It goes underground. A peat fire smoulders through the soil at just a few centimetres an hour. Slower than a snail. And almost impossible to stop. It can burn for months, even through rain.

You can't just hose the surface. Firefighters have to soak the ground itself, metre by metre, while the fire moves underneath their feet.

And because it smoulders, it is filthy. A forest fire is a bonfire. A peat fire is a damp mattress, smouldering in a cupboard. For months. In 2015, burning peat produced about two-thirds of all the smoke particles from Indonesia's fires.

Then add El Niño. In El Niño years the rains over Indonesia fail, the dry season drags on, and the peat dries even deeper. 1997 and 2015, two of the worst haze years: both strong El Niños. In 2019, it was a warm swing in the Indian Ocean instead. So when the forecast says dry year, the haze season goes on the calendar.

Chapter two. Is it slash and burn? So who drains the swamps? Mostly, people growing two crops. Oil palm. And acacia, the fast-growing trees that become paper. By 2015, industrial plantations covered more than a quarter of the region's peatland.

And the biggest drain of all was a government project. In 1995, Indonesia ordered a million hectares of Central Kalimantan peat turned into rice fields. Thousands of kilometres of canals were dug. A leading peat scientist later wrote that not one blade of productive rice was ever grown there. In 1997, it burned.

Why fire? Because it's cheap. In 2015, a CIFOR researcher put the cost of clearing a hectare by fire at about seven US dollars. By machine, about a hundred and fifty. Other studies found a smaller gap. But fire always wins on price. Seven dollars. For a patch of land bigger than a football pitch. The matches are extra.

Then there's the land market. In Riau, researchers found burned land sold for 856 US dollars a hectare. Cleared by hand: 665. Planted with oil palm: 3,800. The World Bank put it bluntly: a few hundred businesses and a few thousand farmers profit, while tens of millions of Indonesians pay with their health.

So who lights them? This is contested, so here are both sides. In Central Kalimantan, one study found most fires started outside oil palm concessions, on degraded and community land. In Sumatra in 2013, another found half the burned area inside concessions, though communities lived on much of it. And under Indonesian law, a concession holder can be liable for fire on its land, whoever lit it.

So, take them to court? Indonesia has. Between 2013 and 2023, its government sued 22 companies over fires. Fourteen were found liable. One had paid in full. None had restored the burnt land.

Now follow the oil. Indonesia grows more than half of the world's palm oil, and Malaysia about a quarter. About a third of Indonesia's comes from smallholders, families on small plots. It's milled, refined, shipped and traded, and several of the biggest palm and paper groups are run from Singapore and Malaysia.

You may have heard palm oil is in half of everything at the supermarket. That number is a little hazy itself. When scientists actually checked 1,600 products, about eight percent listed it outright, though more can hide under other names. It's in cooking oil, instant noodles, biscuits, soap and shampoo. And the paper in your tissue box may start as acacia on drained peat.

So, who's burning the farm? The honest answer is: a system. Cheap land. Cheap fire. Weak enforcement. And a global appetite for cheap oil and paper. Including yours. And mine.

Chapter three. Who pays. Start closest to the fires. In 2015, in Palangka Raya, in Central Kalimantan, daily particle levels reached more than 1,800 micrograms per cubic metre. The scale's hazardous line starts at 350.

Indonesia's disaster agency counted more than 500,000 cases of acute respiratory infection in just a few months. Schools across Sumatra and Kalimantan shut, affecting more than four and a half million students. In Pekanbaru, the city moved newborn babies into an air-conditioned city hall, because many homes had the same air as outside. In 2019 it came back: more than 900,000 cases, and some two thousand schools shut in Central Kalimantan and Riau alone.

The smoke crossed borders too. In September 2015, Malaysia closed more than four and a half thousand schools, for over two and a half million students. In October, it reached southern Thailand, where particle levels in Hat Yai hit three times the Thai safe limit. And on the 25th of September, for the first time ever because of haze, Singapore closed every primary and secondary school. A day off school that nobody wanted.

So how many people did the 2015 smoke kill? Here it gets uncomfortable. Indonesia's official count: 24. A Harvard and Columbia model: about 100,000 early deaths across three countries. Another study: 44,000. Another: about 12,000.

They're all measuring the same smoke. One counts death certificates. The others model the strokes and heart attacks that smoke quietly adds, months later. The models disagree on how deadly it is, and governments around the region questioned them. But even the lowest estimate is hundreds of times the official count.

And it lasts. Research on the 1997 fires estimated about 15,600 infants and young children in Indonesia who did not survive. Babies who were in the womb during that smoke grew up measurably shorter by seventeen. The smoke follows a child for life.

It isn't just people. Scientists following orangutans in Central Kalimantan found that during the smoke, they rested more and travelled less. Like everyone else, they cancelled their plans. And in Singapore, a recorder listening to the dawn chorus caught the birds going quiet. Weeks after the haze lifted, they still hadn't fully come back.

And the planet pays. In 2015, 2.6 million hectares burned. About four and a half Balis. For two months, Indonesia's fires released more CO2 every day than the entire European Union. The World Bank put the bill to Indonesia at 16 billion US dollars. More than twice the cost of rebuilding Aceh after the 2004 tsunami.

Chapter four. What's been tried. This isn't new. In 1997, during a monster El Niño, fires burned across Indonesia, by some estimates as much as ten million hectares. That's up to seventeen Balis. In Kuching, Malaysia's air index passed 800, and Sarawak declared an emergency. The region promised: never again.

In 2002, ASEAN signed a haze treaty. It asks countries to share information and help when asked. Disputes are settled, quote, amicably, by consultation. Its definition of zero burning even says it may allow some burning. And Indonesia, where most of the fires burn, ratified it last. In 2014. Twelve years later.

The treaty also created a regional haze centre. Its office finally opened in April 2026. Twenty-four years later. And ASEAN's goal of a haze-free region by 2020 quietly became: by 2030.

In 2013, Singapore's three-hour PSI hit a record 401. Across the strait in Muar, Malaysia's index hit 746, and an emergency was declared. Singapore's Prime Minister called for a permanent solution, to stop it recurring annually. Two years later came 2015.

So in 2014, Singapore passed its own law. Companies anywhere that cause haze in Singapore can be fined up to 100,000 dollars a day, capped at two million. In 2015, its environment agency served legal notices on six Indonesian companies. Two cases were closed. The other four were still open in 2021. As of 2026, public records showed no prosecution that had reached a verdict. Malaysia proposed a similar law in 2019, and in 2023 decided not to table it.

To be fair, a lot has been tried. Big growers pledged: no deforestation, no peat. Indonesia set up a peat restoration agency, which says it restored about 1.6 million hectares by 2024. Almost three Balis. And it worked, partly. In the dry year of 2019, restored areas had up to 22 percent fewer hotspots. And from 2020 to 2022, helped by wetter La Niña years, the region's skies stayed mostly clear.

So why does it keep coming back? Because dry years keep coming back, and every strong El Niño dries the peat again. Restoration needs upkeep: the agency's mandate ended in 2024, and one study found 70 percent of canal blocks damaged. And the rules keep changing: in 2026, Indonesia's president ordered an end to the old exemption that let small farmers burn, and his government suspended dozens of company licences. Which brings me to the real answer.

Here's the truth behind the truth. The haze isn't a fire problem. It's a water problem.

Wet peat doesn't burn. Keep the swamps wet, and you take away the fuel. One study estimates that protecting peatlands alone would prevent about two-thirds of future haze deaths in the region. The tools are almost embarrassingly simple. Block the canals. Let the water rise. Grow crops that like wet feet.

None of it is glamorous. A concrete dam across a canal. A village fire brigade walking the dry months. Sago, grown on wet peat, instead of palms on dry peat. Blocks break, and they need fixing. But where the water stays up, the fires stay down.

And from a family kitchen, in Kuala Lumpur, Singapore or Jakarta? Look for the RSPO certified palm oil label. It isn't a magic shield: one study found certification cut deforestation, but not peat fires. Still, it tells companies someone is checking. And markets do move. In 2015, supermarkets in Singapore pulled products off the shelves.

When the haze comes back, check your local air-quality index before heading out: the PSI in Singapore, the API in Malaysia, the ISPU in Indonesia. Above 100 on any of them, the air is unhealthy, and kids should cut back on long, hard play outside. Keep inhalers in school bags. Close the windows, run a purifier if you have one. And follow your health ministry's advice, not the family group chat's.

The rain will come. The monsoon turns, the birds come back, and that kid at the window gets their sky back. But the real fix was never in the sky. It's in the ground, in the water, in a swamp a few hundred kilometres away. Keep it wet. And maybe, one day, the haze comes off the calendar.

Until then, a word from the sponsor of this film: myna, the family app we make. Knowing what to do on a haze day is the easy part. Remembering it on the morning is harder. So share the school's notice to Aunty in myna, and it lands on the family calendar. On the day, she reminds the kids out loud, on their phones and on the family hub. Your helper gets the family's messages and tasks in her own language. Aunty can't fix the peat. But she can remember the inhaler. myna. Free to try on the App Store and Google Play, at myna-app.com. Save this for your family chat, and subscribe for the next one.