ENSONINAID26 is the system dynamics model behind this paper. It extends ENSOID26, the model behind Fire on the Horizon, with a La Nina phase, links between the two phases, an adaptation layer, and a dated quarterly layer. Set the El Niño and La Nina intensities, the timing of each peak and the policy levers to see how the cycle propagates. Opens in a new tab.
Abstract
Fire on the Horizon, published in June 2026, modeled the economic effects of an El Niño that was still forming. The El Niño has since become one of the strongest on record, and the fire season it drives is now under way. This paper updates that account with the events of February to mid-October 2026, checks the June prognosis against them, and extends the analysis to the La Nina that frequently follows a strong El Niño.
It uses ENSONINAID26, which adds a wet chain of extreme rain, cyclones, floods, and landslides to the dry chain of drought, fire, and haze, and links the two phases through nine cross-phase channels. Run from weak to record intensity, the model puts the net damage of an El Niño year followed by a La Nina year at 3.5 percent of gross domestic product (GDP) under a strong cycle, 4.6 percent under a super cycle, and 7.1 percent under a record cycle, or $47 billion to $97 billion (Rp759 trillion to Rp1,558 trillion).
Growth falls by 1.1 to 2.1 percentage points over the two years, the deficit widens by 1.1 to 1.6 percentage points of GDP, 2.8 to 5.2 million livelihoods come under stress, and 3.2 to 7.7 million people fall below the poverty line at the peak. Damage in the dry year raises damage in the wet year, and this compound effect adds about a tenth to the cost of the cycle.
The first wave has arrived as modeled. Official data show 202,011 hectares (ha) burned by July, more than in the same months of 2019 and 2023, alongside hydropower shortfalls and stalled river freight. Satellites recorded more fire detections in August 2026 than in August 2015. Prices and growth have moved less than the June prognosis expected, because rice stocks are high and the harvest effect lags into 2027. The 2026 event is therefore more severe than past events in its ocean forcing, below 2015 in burned area so far, and potentially the most costly cycle on record if a strong La Nina follows.
An integrated adaptation portfolio would avoid close to half of the net damage at a benefit–cost ratio of 1.7 to 2.5. Early warning and peatland rewetting return the most per rupiah. The figures are scenario outputs, not forecasts.
The system dynamics model on which this paper is based, ENSONINAID26, is available here. It extends ENSOID26, the model behind Fire on the Horizon (June 2026).