When the Haze Becomes a Carbon Event
credit: Photo by Rifqi Ali Ridho on Unsplash
What Indonesia’s 2026 fires mean for the climate
By Thiru Jallendran, Director - Pure Environment
AT A GLANCE
|
In 1997, during the upheaval of the Asian Financial Crisis, I remember driving back to Kuala Lumpur with friends from the east coast of Peninsular Malaysia beneath a sky that had turned an unsettling dirty grey. I had started my career as a graduate structural engineer only the previous year, on 2 May 1996. We were deeply anxious about what the economic turmoil might mean for our jobs and livelihoods.
I also remember feeling exasperated. Surely, I thought, recurring problems of this scale demanded collective action. The haze made me think more deeply about the environmental consequences of the choices made by societies, governments and industries, and about the work I was doing within the construction industry. It did not prompt an immediate career change, but, looking back, those questions probably played some part in my eventual decision to pursue sustainability as a career.
The exact date of that drive is uncertain. Thailand moved the baht to a managed float on 2 July, while Malaysia's official chronology places the principal 1997 haze episode in September-November. It also records serious Indonesian-fire haze in 1991 and a still more severe recurrence in September 1994. The lesson is that 1997 was not the beginning; it was the year the recurring crisis became impossible for many of us to ignore.
To me, the haze has therefore never been an abstract air-quality issue. As someone with allergic rhinitis, I often became unwell during severe episodes. Since moving to Australia, I am no longer exposed in the same way, but I have not forgotten what it felt like when the air itself made people sick.
Severe El Niño drought amplified the 1997 fires linked to forest clearing, peat drainage and deliberate burning for agriculture and plantation development. Oil palm is an important part of that history, particularly in Kalimantan, but so are pulpwood, logging, smallholder burning and the governance failures that allow fire to escape across degraded landscapes.
Then came Australia’s 2019 - 20 Black Summer. The scale of the smoke, the human suffering and the devastation of wildlife led me to a broader question: what is the global-warming impact of these great fire events, and how does it compare with the emissions we routinely count from countries and industries?
A satellite-constrained study later estimated that southeast Australia’s fires released about 715 million tonnes of CO₂ between November 2019 and January 2020. Separate research attributed an estimated 417 excess deaths to bushfire smoke in eastern Australia, while almost three billion native vertebrates were present in the forests and woodlands affected by fire. Carbon, health and wildlife are different measures—but they belong to the same story.
That story became painfully immediate again when Sempurna, an adult male orangutan rescued in West Kalimantan, died on 30 August 2026. Indonesia’s Forestry Ministry reported lung abnormalities suspected to be associated with smoke exposure and acute oxygen deprivation. It did not claim that haze was the sole cause, but the case is a stark reminder that the damage does not stop with people, property or atmospheric carbon.
For weeks, smoke from Indonesia’s forest and peatland fires has moved across Kalimantan, Malaysia, Brunei and Singapore. Schools have closed. Flights have been disrupted. Respiratory illness has surged. The haze is immediate, visible and deeply human.
The cost measured in lives
I have often wondered how much life has been taken from people living repeatedly beneath these smoke-filled skies. There is no scientifically defensible answer in the form of a single number of months or years lost by every resident. What researchers can estimate is the increase in premature mortality across an exposed population.
During the 1997 haze, a high-pollution day in Kuala Lumpur was associated with an approximately 19% increase in all-age mortality the following day. For people aged 65-74, the effect was substantially larger and persisted for several days. The extreme 2015 haze was later estimated to have caused about 100,300 excess deaths across Indonesia, Malaysia and Singapore.
A longer-term study covering 2013-2017 estimated that peat-fire PM₂.₅ caused, on average, around 33,100 premature adult deaths and 2,900 infant deaths each year in Sumatra and Kalimantan alone. These are modelled population estimates, not individually identified deaths, and they should not be added to the 2015 figure because the periods and methods differ. Even with those qualifications, the scale makes clear that haze is not simply an inconvenience that disappears when the sky clears.
But behind the smoke is another consequence that is easier to miss: a large pulse of carbon entering the atmosphere. That is the question this article sets out to make tangible.
Scale of the carbon pulse
The Global Fire Emissions Database estimates that Indonesian fires released 76 million tonnes of carbon between 1 January and 7 September 2026. Converting the carbon mass using 44/12 gives approximately 278.7 million tonnes of CO₂. GFED's current-year data are beta and subject to revision. A separate WALHI estimate based on Copernicus data reported 65.18 MtC for January to early September, equivalent to 239.0 MtCO₂; the different methods should not be merged.
The 278.7 MtCO₂ GFED carbon-mass estimate is roughly equivalent to:
218 days of Australia's net national greenhouse gas emissions;
310 days of Malaysia's emissions excluding land use and forestry;
95 days of Indonesia's emissions excluding land use and forestry;
around 3.7 years of New Zealand's gross emissions; or
about 2.7 days of global energy-related CO₂ emissions.
A NECESSARY CAVEAT These are scale comparisons, not claims that the inventories are methodologically identical. Fire carbon, national CO₂e accounts and land-sector accounting use different boundaries. |
The rolling figures show that the fires intensity accelerated again in early October. CAMS estimated 19.7 MtC for 5-11 October, equivalent to 72.2 MtCO₂ by carbon-mass conversion. This was 4.3 MtC, or about 28% above the 15.4 MtC recorded during 28 September–4 October. Because the periods are consecutive and non-overlapping, the comparison indicates a genuine increase in weekly intensity. Indonesia accounted for about 47% of global fire carbon during the week of 5-11 October and was about 308% above its 2003–2025 seasonal mean. The CAMS week falls after the GFED 7 September cut-off, but the estimates use different methods and should not be combined into a single total.
On 8 October, SiPongi reported 387 high-confidence hotspots nationally, down from 1,176 a day earlier. The sharp daily movement illustrates why hotspot counts must be interpreted cautiously: they are satellite-detected heat anomalies, not fires, burned area or emissions, and cannot be added to the GFED or CAMS estimates. SiPongi had not published a dated national 2026 emissions total by 15 October.
Carbon, CO2 and CO2e are not the same thing
This distinction matters. CAMS often reports total carbon, written as MtC. Multiplying by 44/12 converts that carbon into the mass of CO₂ containing the same amount of carbon. It does not automatically create a formal CO₂-equivalent (CO₂e) inventory.
Fire smoke contains CO₂, carbon monoxide, methane, nitrous oxide, organic particles and many other compounds. Methane adds extra warming before it oxidises. Smoke particles can temporarily cool or warm the region depending on their composition. These effects operate over different timescales and should not be folded casually into one headline number.
Why peat changes the climate story
A forest fire and a peat fire can look similar through the haze, but their carbon consequences are different.
Where above-ground vegetation regrows, some emitted carbon can be taken back from the atmosphere over years or decades. That recovery is not immediate, and it may never occur if burned forest is converted to another land use or burns repeatedly.
Peat is different. Tropical peatlands store carbon accumulated over centuries to millennia. Once drained, they become more flammable. Fires can smoulder underground for weeks, consuming carbon that cannot be rebuilt on a meaningful net-zero timeframe. Drainage can also keep releasing CO₂ after the visible fire is extinguished.
There is a second loss as well. A damaged forest is no longer absorbing carbon at the rate it might have. The climate cost therefore includes what went up in smoke, what continues to oxidise and what the landscape will no longer remove.
Indicative global temperature effect
Not in a way that could be isolated in next month's temperature record. If the 278.7 MtCO₂ cumulative estimate became a net, long-lived addition, an IPCC carbon-temperature relationship gives an indicative contribution of about 0.00013°C. That number is a carbon-budget screen, not a forecast of a measurable temperature jump from this fire season.
It would be a mistake, however, to read ‘small temperature increment’ as ‘unimportant’. Climate change is the accumulation of billions of individual emission events. Every additional tonne adds to the atmospheric stock; repeated peat-fire seasons consume carbon budget and weaken one of the region’s most valuable natural carbon stores.
The warnings of 1997 and 2015
A landmark study estimated that Indonesia’s 1997 peat and vegetation fires released between 0.81 and 2.57 billion tonnes of carbon, approximately 3.0 to 9.4 billion tonnes of CO₂ by carbon-mass conversion. The wide range reflects uncertainty in burned area and peat depth, but even the low estimate was extraordinary. It confirmed that peatland conversion could turn a regional haze emergency into a globally significant carbon event.
In 2015, fires across maritime Southeast Asia released an estimated 884 million tonnes of CO₂ over the year, with 97% originating in Indonesia. When methane, nitrous oxide and oxidised carbon monoxide were included, the impact reached as high as 1.2 billion tonnes of CO₂e. During September and October, daily fire emissions exceeded the European Union’s fossil-fuel CO₂ emissions at the time.
The 278.7 MtCO₂ GFED estimate through 7 September is already about 31.5% of that 2015 annual CO₂ estimate. The latest CAMS weekly figure shows renewed acceleration into early October, and the 2026 fire season remained active when this article was updated. GFED placed the early-September trajectory close to 2015, but that is not a prediction that the final total will match it. Weather, suppression, peat conditions and the duration of the dry season will determine the final outcome.
What turns drought into disaster
El Niño creates hotter and drier conditions across much of Indonesia, but it does not light the match. Ignition, forest clearing, drainage, land management and enforcement determine whether dry conditions become a regional emergency.
It is tempting to reduce the story to palm oil, and oil-palm development is unquestionably an important driver. But the evidence is more complex. Fire has also been associated with pulpwood and logging concessions, smallholder land clearing, abandoned or contested land and deeply degraded peat. Accountability therefore requires transparent concession and supply-chain data rather than a single convenient culprit.
Recent research shows that El Niño substantially increases the probability of extreme fire weather in Borneo, but major fires remain possible under neutral and even La Niña conditions. Long-term resilience therefore cannot rely only on seasonal forecasting.
The most effective actions begin before smoke appears:
protect intact peatlands and avoid new drainage;
re-wet degraded peat and manage water tables before the dry season;
provide practical no-burn options for landholders;
combine hotspots with near-real-time emissions to target response; and
Make companies and land managers accountable through transparent concession and fire data.
Looking beyond the haze
The immediate priority is protecting people, controlling the fires and restoring safe air. But the climate accounting cannot stop when visibility returns.
We need to know how much peat burned, how deeply it burned, whether the land will recover, whether drainage continues and whether the same ground will burn again. Gross fire emissions and net land-carbon impact should be reported side by side. Neither tells the complete story alone.
credit: Photo by David Kristianto on Unsplash
There is a broader lesson here. Environmental crises are often managed through the indicator that is easiest to see: smoke, hotspots, hectares or hospital cases. Good climate governance asks what remains after the visible emergency has passed.
The haze will clear. Much of the carbon consequence will not.
Research assistance disclosure
Deep research, source discovery, cross-checking, calculation support and drafting assistance for this article were provided using OpenAI Codex in ChatGPT between 4 September and 2 October 2026. The author reviewed the cited evidence, determined the interpretation and accepts responsibility for the published content. The AI system is not an author and did not generate or independently validate primary data.
Selected sources
Page et al. (2002): carbon released by Indonesia’s 1997 fires
Marlier et al. (2015): fires in oil-palm, timber and logging concessions
Indonesia Forestry Ministry: the death of orangutan Sempurna
Indonesia Forestry Ministry: national hotspot update, 9 October 2026
Medical Journal of Australia: Black Summer smoke-health burden
Malaysia Department of Environment: chronology of haze episodes
Hein et al. (2022): long-term health impacts of Indonesian peat fires
Comments