Peat fires in Siberia contribute 14 times more to regional carbon emissions than previously estimated – a breakthrough finding made possible by a new level of precision in satellite-based fire monitoring. A team led by Amin Khairoun from the University of Alcalá (Spain) together with colleagues across Europe and China analysed Landsat satellite data across northern Siberia covering the period 2001 to 2023. They discovered that peatlands accounted for almost one-third of the 107.6 million hectares that burned and were responsible for 33% of all wildfire carbon emissions. This contribution had largely gone undetected by earlier monitoring tools.
The study was conducted within the framework of the European Space Agency's Climate Change Initiative (ESA CCI) Fire project, which provides global information on land affected by wildfire and has been instrumental in advancing fire monitoring capabilities worldwide.
"Siberian peatlands have long been considered stable carbon sinks – absorbing more carbon than they release," says Amin Khairoun, researcher in the ESA CCI Fire project. "But our findings show that peat fires contribute far more to carbon emissions than previously assumed. As fire seasons intensify, these ecosystems may be shifting from carbon stores to carbon sources."
Ancient carbon stores face record-breaking fires
Peatlands across the northern hemisphere store approximately 415 billion tonnes of carbon, accumulated over thousands of years. Although they cover just 2-3 per cent of Earth's land surface, they act as critical buffers against climate change. Yet with Arctic temperatures rising up to four times faster than the global average, this ancient carbon reservoir is increasingly vulnerable to fire.
Unusual hot and dry climatic conditions are the primary driver of extreme peat fire seasons. The years between 2019 and 2021 represent the most severe on record. In 2021 alone, fires consumed 10.6 million hectares, including a single megafire spanning 1.5 million hectares – the largest fire event ever recorded in Siberia.
A major contributor to these extensive fires were so called 'zombie fires' – fires from the previous year that survived underground through winter before re-emerging the following spring. After burning into carbon-rich peat soils during the 2020 fire season, these fires smouldered slowly beneath the snow and reignited when conditions warmed in spring 2021.
“Peat fires are different from surface fires. They burn deep into the soil and release carbon that took thousands of years to accumulate," says Khairoun. "A single peat fire can emit far more carbon per hectare than a forest fire. With conditions like those we saw in 2020 becoming more common, this is a growing concern – not just for emissions, but for the stability of the Arctic itself."
Burning peat threatens permafrost stability and accelerates warming
The vulnerability of Arctic peatlands to fire has implications far beyond immediate carbon emissions. When peat burns, it destroys the natural insulation that keeps permafrost frozen. Without this protection, permafrost thaws more rapidly, releasing additional greenhouse gases. This creates a dangerous feedback loop: fires release carbon, which accelerates warming, which in turn increases fire risk.
New high-resolution approach reveals full scale of peat fire emissions
Advances in satellite technology now allow scientists to map fires at a level of detail previously impossible. The ESA CCI Fire project provides global burned area data primarily based on MODIS imagery at resolutions of 250 metres or more. In this study, researchers went beyond the existing CCI products by developing two new high-resolution datasets: a burned area map using Landsat imagery at 30-metre resolution, and a separate map of peatland coverage at 90-metre resolution. Together, these enabled the team to detect up to 76 per cent more burned area than previous datasets had captured. By combining these two maps, the team could identify exactly which fires had burned through carbon-rich peat soils across northern Siberia over the 23-year period. For each peat fire, a machine learning model – trained on nearly 900 ground measurements – then estimated how deeply the fire had burned into the soil. The deeper the burn, the more ancient carbon is released.
Together, these advances deliver a far more accurate picture of where peat fires burn, how deeply, and how much carbon they release – providing the precise data that climate models and future monitoring efforts urgently need.
Refining the picture of Arctic fire risk
The sensitivity of northern peatlands to extreme weather – and how quickly stable carbon sinks can become emission sources – is starkly highlighted by these findings. Building on the foundation of ESA's CCI Fire datasets, this study demonstrates how advances in satellite resolution can deepen our understanding of fire-driven carbon emissions. By complementing established monitoring tools with finer-scale Landsat data, the researchers uncovered emissions from peat fires that were previously beyond the detection limits of available sensors. These findings can help refine climate models and strengthen future monitoring of Arctic fire risk. The ESA CCI Open Data Portal provides open access to Fire CCI datasets for independent verification and ongoing monitoring of fire-prone regions.