Strongly underestimated Sources of Methane: Water Bodies in the Amazon Rainforest

Comprehensive aircraft measurements reveal that methane emissions from Amazon wetlands are significantly higher than previously estimated.

In a nutshell:

• Rivers are a major source: Emissions of the highly potent greenhouse gas methane in areas along rivers and tributaries are up to four times higher than previously calculated.
• Discrepancy between model and measurement: The largest discrepancies between model predictions and measurements occur in large river deltas, reservoirs, and regularly flooded areas across the Amazon Basin. In river deltas alone, methane emissions were underestimated by up to 26%.
• For reliable climate projections: The findings underscore the need to strengthen observational infrastructure in tropical regions and to calibrate climate models with real-world measurements.

Methane (CH₄) is a potent greenhouse gas whose concentration in the atmosphere has risen sharply in recent decades. Wetlands are the largest natural source of methane to the atmosphere, but large uncertainties remain about how much methane comes from wetlands and how these emissions may be increasing in response to a changing climate. Tropical wetlands, including in the Amazon, produce substantial amounts of methane, but accurately estimating their emissions sources and magnitudes remains difficult. One reason is the lack of measurement data, especially in the tropics, where extensive cloud cover interferes with satellite observations and ground-based measurements are sparse. 

An international research team led by the Max Planck Institute for Chemistry has now determined, using aircraft measurements, that actual emissions in the Amazon region were significantly higher than those estimated by climate and Earth system models. In certain areas, methane emissions are even four times higher than previously calculated. The findings, published in the journal Geophysical Research Letters, have important implications for the accuracy of climate projections.

Established models are reaching their limits

“The discrepancies between model calculations and reality motivated me to find out where the missing methane comes from,” says Linda Ort of the Max Planck Institute for Chemistry. The atmospheric chemist and lead author of the study notes that at altitudes of six kilometers or higher, the measured and modeled data align quite well. However, the closer to the Earth’s surface, the greater the discrepancy became. On average, the measured methane levels were about twice as high above the background level as the corresponding model values. 

At the time of the measurements in December 2022 and January 2023, the background level in the atmosphere was about 1,907 parts per billion (ppb), but it has risen again since then. The unit ppb (parts per billion) is used to precisely describe very low mixing ratios, such as those found in trace gases in the atmosphere. Ort attributes this to the fact that methane mixes well in the atmosphere at high altitudes and the models accurately represent the transport of air masses and mixing effects. However, the models fail at lower altitudes.

A closer examination of the sources revealed that in some areas of the Amazon wetlands release up to four times more methane than previously assumed. Broken down by wetland type, the values were, for example, 26 percent higher at river deltas, 19 percent higher at reservoirs, and 13 percent higher in regularly flooded river areas.

The researchers collected their data over a two-month period between December 2022 and January 2023. This time window falls precisely during the transition from the dry season to the wet season—a phase in which biomass burning typically does not occur, meaning the measurements were not strongly impacted by human activity. While methane emissions are even higher during the actual rainy season, they are lower during the dry season, which roughly balances them out. Thus, the two months during which the measurements took place roughly correspond to the annual average of wetland emissions.

Extensive measurement flights between the dry and wet seasons

Using sensors aboard the HALO research aircraft, Ort and the international team measured methane concentrations at various altitudes over a vast area of the Brazilian rainforest. Flight altitudes ranged from just 200 meters above the treetops to over 14 kilometers. At more than 7,000 measurement points, the scientists collected data using an absorption spectrometer specially developed for HALO, which precisely detects the trace gas methane even at low air pressures, such as those found at high altitudes.

For their analyses, they divided the entire Amazon region into grid cells measuring 0.1 degrees × 0.1 degrees—a high resolution for emission maps. To link methane measured in the air to its sources on the ground, the researchers used an atmospheric transport model to trace air masses backward in time to the respective grid cell on the ground. Using a well-established NASA ensemble model estimating wetlands and a complex numerical method, they then calculated the actual amounts of methane released. The NASA ensemble model indirectly estimates methane emissions based on satellite measurements of land surface characteristics such as moisture, vegetation, and temperature.

“Our results show that there are still many underestimated sources of methane in tropical wetlands such as the Amazon Rainforest,” says Linda Ort, who participated in numerous measurement flights during the research campaign. According to Ort, more measurement data is needed to better identify these methane sources—some of which vary strongly—and to improve climate and Earth system models.

Methane Emissions: 35% from Natural Sources, 65% Anthropogenic

About 65% of global methane emissions come from anthropogenic sources such as agriculture, fossil fuel production and use, and waste management. The remaining 35% come from natural processes. Methane is produced in large quantities when organic material, such as dead plants and leaves, is decomposed underwater by microorganisms. One example is dams, when large areas of forest are flooded as a result. Although these man-made structures are primarily used for hydroelectric power generation, they are also considered natural sources of methane.

“These results underscore how important it is to better understand the processes that control methane formation in wetlands,” says Eric Kort, director of the Atmospheric Chemistry Department at the Max Planck Institute for Chemistry and a co-author of the study. “To reliably assess the global methane budget, we need significantly more measurements—not only in the Amazon region, but also in other data-poor tropical regions such as Central Africa and Southeast Asia.”

Participating Institutions

The study was conducted by an international team led by the Max Planck Institute for Chemistry in Mainz. Scientists from Johannes Gutenberg University Mainz, Forschungszentrum Jülich, the Cyprus Institute in Nicosia, Cyprus, the California Institute of Technology in Pasadena, USA, as well as the University of Washington and the University of Michigan, USA, also contributed.

Background: CAFE Brazil research Campaign

The aircraft measurements of methane emissions were part of the scientific campaign CAFE BRAZIL (Chemistry of the Atmosphere: Field Experiment in Brazil). Using the HALO research aircraft, scientists investigated natural chemical processes over the Amazon rainforest in December 2022 and January 2023; balloons, drones, and the ATTO research station complemented the measurements from the ground. 

One goal was to understand how the forest and its plant emissions affect atmospheric processes such as cloud formation. Collaboration with local partners in Brazil was key to the success of the research mission. Researchers from INPA (National Institute of Amazonian Research) and the University of São Paulo contributed their on-the-ground expertise.