Science1 distinct publisher2 min readPublished
CAMS is comparing this season to 2015 and 2019, both exceptional El Nino years. For anyone planning around regional haze, the more durable input is the drained peatland that turns a dry spell into weeks of smoke downwind.
The Scientist · Science desk

Follow any of these and your For You feed starts watching them — no settings page required.
science
Indonesia's peat fires already match big El Nino years, and the forecast peak is October1 distinct publisher
science
Indonesia scales cloud seeding against a fire season it cannot yet count1 distinct publisher
science
Indonesia seeds clouds over a haze that has already reached 5 million people1 distinct publisher
science
Kalimantan's fires: El Nino is the multiplier, the drained peat is the base rate1 distinct publisher
Compiled by The ScientistSomething wrong?How this is made
Seven days, Aug. 28 through Sept. 3, produced more than 23 million metric tons of carbon from Indonesian fires, roughly 3.3 million tons a day [3][2]. The Copernicus Atmosphere Monitoring Service logged 20.7 million tons over the same week of 2015, so this year's week ran at least 11 percent above its strong El Nino analog [4][1]. Mark Parrington, a senior scientist at the service, told AFP the emissions are "really extreme" and pointed to "the extent of the smoke and the haze and the air quality impacts" [6].
That comparison is between levels, not between seasons: this week burned harder than the matched week of 2015, but burned area and duration for the two years are not in the same evidence, so where the year finishes is a separate question.
On causation, the two analysts AFP quotes are careful. David Gaveau, who founded the Nusantara Atlas portal, said El Nino "does not ignite fires" but "creates the dry conditions that allow human-started fires to escape and become very large" [8]. Wahyu Perdana of Pantau Gambut said that when peat is drained, its stored organic material is converted into fuel [11]. Peatland is also why the fires resist extinguishing once lit [9]. El Nino describes the weather and the drainage describes the fuel, and only one of the two arrives with a forecast attached.
The impact counts deserve the same care as the emissions ones. An education ministry official said more than 1.4 million students had moved to remote classes as schools closed [13]. No enrolment denominator is published alongside it, so the figure counts disruption without telling you what share of the school system is involved. Emissions, for their part, measure combustion and not deposition. The evidence that the smoke arrived downwind sits in the air quality readings: a week of "very unhealthy" and "severely unhealthy" PM2.5 in the Philippines, where officials closed schools in cities and provinces [15], and an "unhealthy" reading in Singapore's central region [16].
Four countries are currently dealing with the effects at once [17][3]. The comparison years are 2015 and 2019, and the UN said this week that the current El Nino could turn out to be the strongest in four decades [5][7]. What carries into an ordinary year is the drained peat, and that is the input a haze plan can act on before the dry season starts.
Ranked by verification strength, evidence, and original report placement.
Thick smog has blanketed parts of Indonesia as fires have burned large swaths of Borneo and Sumatra for weeks; most fires are concentrated on Borneo and on Sumatra.
Nearly 300,000 hectares (741,000 acres) of land have burned in Indonesia so far this year, according to an independent estimate from the Nusantara Atlas portal.
Fires in the Indonesian archipelago released more than 23 million metric tons of carbon emissions between Aug. 28 and Sept. 3, according to the CAMS Fire Emission Watch portal.
That was higher than the 20.7 million metric tons of emissions from fires over the same period in 2015, which was a strong El Nino year.
Indonesia saw an uptick in carbon emissions since late July, and by the end of last month levels were comparable to those in the previous El Nino years of 2015 and 2019, Copernicus Atmosphere Monitoring Service senior scientist Mark Parrington told AFP.
Parrington described the fire emissions as "really extreme" and referred to "the extent of the smoke and the haze and the air quality impacts".
Distinct publishers with included, body-backed reporting in this cluster.
1 article · September 5, 2026
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Named owner for every number, one reading each
Each quantity here has an institution attached to it: Copernicus for the 23 million tons, Nusantara Atlas for the 300,000 hectares, Pantau Gambut for the 210,000 peat hotspots, Singapore's environment agency for its own air reading, and the deputy health minister for the 5 million exposed. Two soft spots keep this off the top of the scale. The 1.4 million remote students come from an education ministry official who goes unnamed, and no quantity in the piece is corroborated by a second measurement.
Four governments already responding
The most solid material in this story is institutional behaviour rather than science. Sarawak is under an emergency with school closures scheduled, Indonesian classes have moved online at scale, Philippine schools shut and are reopening, and Singapore's own agency has logged the reading in its central region. Those are dated actions and instrument readings, which is why response scores higher than the atmospheric picture behind it.
Precise in the wire, loose one step downstream
Phys.org's account is careful about what it is comparing: seven days in 2026 against the same seven days in 2015. The overstatement risk sits in what a reader does with that. Parrington's own words are 'comparable to' those El Niño years, softer than the single-week figure sounds, and the season-to-date burned area is printed with no 2015 counterpart, so the year-scale question stays open while the week-scale answer travels.
Mostly disinterested monitors, two interested suppliers
Copernicus and the Nusantara Atlas publish these series whichever way the numbers fall, and their figures are the ones doing the comparative work. Pantau Gambut campaigns on peatland drainage and supplies both the hotspot count and the explanation that drainage creates the fuel, so its data and its argument point the same direction. Indonesia's health and education ministries supply the exposure and schooling figures they are themselves measured on. AFP attributes each of them cleanly and weighs none against the others.
One wire report carrying all of it
The underlying attributions are good enough that the figures are unlikely to be wrong in direction, but a single AFP explainer means one set of editorial choices about what to include, and no outside check on the weekly emissions total that the story's comparison depends on. Near-real-time fire emission estimates are published while a season is still burning, and this reporting gives no revision history for them.