Environment news · Development: 10 Nov 2026
An early warning model maps where El Niño may weaken land carbon sinks
An ESA-supported study forecasts potential carbon-sink and fire-risk hotspots for the second half of 2026. Its authors’ preprint remains under review, so the projections are risk estimates to test against later observations.
Global; research reported by ESA during COP31

Researchers have published an early-warning estimate of how the developing 2026 El Niño could affect the land carbon sink. The ESA-supported analysis combines seasonal climate forecasts, ecosystem models and predictions of burned area to map possible regional risks before the season is complete. Its estimated global weakening is a forecast, not a measured annual result; the paper is a preprint under review. That difference is central to reading the headline responsibly. [1] [2]
From climate outlook to carbon-risk map
The study begins with a 51-member ensemble of seasonal forecasts from the European Centre for Medium-Range Weather Forecasts. Researchers used those climate inputs to drive two land-surface models, ORCHIDEE-MICT and JULES, and combined their output with machine-learning estimates of burned area. The intended product is a prospective map of where climate stress and fire could affect carbon absorbed or released by ecosystems during July to December. [1] [2]
Each link in that chain answers a different question. The climate ensemble represents a range of plausible weather patterns. The ecosystem models translate temperature and moisture conditions into simulated carbon exchange. The burned-area component estimates another pathway for carbon release. Joining them can help identify where to look, but a model output does not turn a forecast into an observation. It is a conditional estimate dependent on the inputs, model structure and assumptions.
A large central estimate with a wide range
The preprint estimates that global net biome productivity could be lower by 0.23 petagrams of carbon in the second half of 2026, with an uncertainty of 0.21 petagrams. ESA summarizes this as roughly 230 million tonnes of reduced net biome productivity. The range is substantial relative to the central estimate: it signals that the size of the projected change is uncertain, even as the analysis identifies regions where several risk indicators converge. [1] [2]
Net biome productivity is a modeled balance of carbon uptake and release across land ecosystems. A reduction means the land sink is projected to absorb less carbon overall within the study’s accounting. It does not mean every ecosystem becomes a source, that the same amount is emitted directly to the atmosphere, or that the forecast has already been confirmed. The authors compare South America as the largest projected contributor to the weakening and identify elevated fire risk in equatorial Southeast Asia. Those are places to monitor, not a verdict on observed losses.
What independent observations can tell us
NOAA’s Climate Prediction Center reported on 12 November that El Niño continued to strengthen and that a strong-to-very-strong event was likely through early 2027. That assessment supports the climate context for a risk forecast; it does not verify the study’s regional carbon estimates. NOAA also cautions that even a very strong El Niño does not guarantee impacts consistent with the usual pattern. [3]
A useful check would compare the forecast with observations collected as the season unfolds: rainfall and soil moisture, vegetation condition, fire detections and emissions, and carbon-flux estimates. Each measures something different. A satellite can detect a fire or surface change, while estimating the net exchange of carbon across a broad region requires additional measurements and analysis. A later evaluation should report what data were available, how the forecast was scored and whether the pre-registered or original model was adjusted after seeing outcomes.
Why early warning can still be useful
A forecast need not be exact to help agencies decide where to strengthen monitoring or prepare fire and drought response. Its value depends on communicating probabilities and limits clearly enough that users do not mistake a risk map for a loss map. The 2026 framework is especially useful as a test of whether seasonal climate information can be translated into carbon-cycle warnings with enough regional detail to guide attention.
The study is still under review, and the forecast period runs through December. Readers should therefore treat its numerical estimate as provisional and revisit it when the authors publish a reviewed paper or evaluation. If later measurements differ, that would not make the early warning meaningless; it would show where forecasts, ecological models or fire estimates need improvement. The practical question is not whether a map looks precise, but whether it helps observers ask better questions and then checks its predictions against evidence.
Follow the evidence
Primary reporting and background sources. Event dates and publication dates are shown separately above.
- Du et al.: Forecasting carbon-risk hotspots ahead of the extreme 2026 El Niño (arXiv preprint, under review) ↗
- European Space Agency: Predicting El Niño’s impact on carbon sinks ↗
- NOAA Climate Prediction Center: ENSO Diagnostic Discussion, 12 November 2026 ↗
Read our editorial standards. Have a correction? Identify the article and relevant source in your feedback to the site owner.
