By 11 September Indonesia's fires had released 76.2 megatonnes of carbon this year, according to Copernicus satellite estimates. That is behind the catastrophic 2015 season at the same point, and about two-fifths of 2019's total. The season is not over.
Indonesia's 2026 fire season began in earnest in late July, with the first large fires in West and Central Kalimantan on Borneo, followed by South Papua and Sumatra. The Copernicus Atmosphere Monitoring Service (CAMS), the EU's atmospheric monitoring service, said in August that the fires were more intense than in recent years.2
CAMS estimates fire emissions with its Global Fire Assimilation System (GFAS), which converts the heat that satellites detect from active fires into emissions. On those estimates, as published on the service's Fire Emissions Watch portal, Indonesia's fires released 34.83 megatonnes of carbon (MtC) between 1 January and 26 August 2026. Over the same dates in 2015, the figure was 47.68 MtC.16 On that measure 2026 stood at 73% of 2015.
The gap narrowed in early September. In the week of 1–7 September, Indonesia's fires emitted 19.7 megatonnes, against 21.7 in the same week of 2015, and accounted for more than a third of the world's fire emissions that week.7 Mark Parrington, a senior scientist at CAMS, described early indications as running at pace, if not slightly above those previous El Niño years
.10
By 11 September the 2026 total had reached 76.2 MtC, according to CAMS data reported by The Jakarta Post.8 That is the most recent cumulative CAMS figure Noodle News could find in print. It is equivalent to about 279 megatonnes of carbon dioxide.
The two benchmark seasons ended far higher. CAMS estimated that Indonesia's 2019 fires emitted at least 708 megatonnes of CO2 between August and the end of November that year, or about 193 MtC.4 For 2015, the Fire Emissions Watch portal puts the full dry season, which ran into November, at more than 380 MtC.16 On those figures 2026 had reached about 39% of 2019's total and no more than about 20% of 2015's.
Before August the year was quiet: 8.25 MtC from January to the end of July.6 The remaining 67.95 MtC of the 76.2 total, about 89%, was released in the six weeks from 1 August to 11 September.
Dividing each period's emissions by its length shows the escalation. The daily average rose from 0.039 MtC in January to July, to 1.02 MtC in the first 26 days of August, and to 2.59 MtC between 27 August and 11 September, about 66 times the early-year rate. The single week of 1–7 September averaged 2.81 MtC a day.7
Large fire years in Indonesia have coincided with El Niño and related dry conditions: CAMS lists 1997, 2005–06, 2015, 2019 and 2023, and said 2026 appeared to be no exception.2 They are not uniform. In 2023, an El Niño year, CAMS found Indonesia's August-to-October emissions were below average overall.5 Much of the country's burning is on peatland, where fires can smoulder underground and are hard to put out once established.2
GFAS is not the only estimate. The Global Fire Emissions Database (GFED), a separate inventory built partly on burned-area mapping, put Indonesia's 2026 fire carbon at 76 MtC by 7 September, and its 2015 total at 333 MtC, according to an analysis by Carbon Brief.9 GFED therefore reached roughly the same 2026 total four days earlier than GFAS, while putting 2015 lower. Rankings between years can shift depending on which inventory is used.
Both methods depend on what satellites can see. Parrington told Reuters that peat fires burning at low temperature or below ground can fall under sensor detection limits, so the figures are probably an undercount.7 CAMS's comparison in 2019 found that the 1 August to 18 September window produced about 360 megatonnes of CO2 that year against 400 in 2015, a reminder that 2019, not just 2015, was a severe season.3
How large the 2026 season has become since 11 September. The CAMS figures are published through an interactive portal that Noodle News could not query directly; every 2026 and 2015 figure here is as read and reported by news organisations, and the latest cumulative reading is nearly four weeks old. Parrington said emissions could keep rising until the end of October, as in 2015.7 Whether 2026 will end above 2019, or near 2015, is therefore open. GFAS has also been revised since 2019, so the 2019 bar is not a strictly like-for-like comparison, and the share of peat burning missed by satellites is not known.