Earth’s breath and sweat uncovered - everywhere and continuously

Breathing of the Earth (continuously) gauged

March 11, 2025

Thanks to the next generation of data driven, AI-based earth system models, scientists can now see the Earth’s metabolism at unprecedented detail – assessed everywhere on land and every hour of the day. The Earth’s “breath” is monitored as the gas exchange of land ecosystems with the atmosphere, its “sweat” as the exchange of transpired water vapor and energy. This will open the door to a more accurate understanding of how natural systems function from local up to global scales. These models are instrumental in the monitoring of greenhouse gas budgets and ultimately in decision making with respect to climate protection.

Multiple small satellites orbiting Earth, captured against a backdrop of clouds and the blue atmospheric edge, viewed from space. — With satellites, we can observe the Earth's metabolism from space. However, these data are neither complete in space nor in time. That is why researchers are combining observational data with machine learning.

© NASA Johnson

With satellites, we can observe the Earth's metabolism from space. However, these data are neither complete in space nor in time. That is why researchers are combining observational data with machine learning.

© NASA Johnson

The new modelling environment that makes this possible is called FLUXCOM-X. It has been developed in a global inter-disciplinary collaboration at the Max Planck Institute for Biogeochemistry, Jena. By using the new model, scientists are able to critically evaluate how the estimated breathing and sweating activity of ecosystems is impacted by methodological decisions in the modelling process.

The modelling approach is based on a network of stations which measure the exchange of CO₂, water and energy worldwide. From small patches of land at the stations, scientists can now ‘zoom out’ to a ‘bird or even satellite view’ of larger regions and even the globe. They achieve this by combining the networks’ measurement data with satellite observations and machine learning. Thereby, they obtain an estimate of all places where no station measurements are available. This upscaling approach allows to quantify, for example, how much CO₂ a forest has taken up in a specific year, or how strongly ecosystems have suffered from dry conditions. It also helps to better understand the observed variability of the CO₂ concentration in the atmosphere. This approach to modelling the biosphere of the Earth is based solely on continuous observational data.

Color-coded map of Europe indicating net ecosystem exchange levels from blue (low) to red (high) — Example for a simulated net ecosystem exchange as produced by Fluxcom-X-Base (July 7th 2010, 6-7AM). Blue colors represent a net movement of CO₂ from the land to the atmosphere (release), red colors mean a net movement of CO₂from the atmopshere into the ecosystem (uptake). The differences are mainly a result of different light conditions. Further to the East, the sun has already risen, and plants are taking up CO₂ through photosynthesis. Further in the west, where it is still darker they are still in “night-mode”. Smaller, superimposed patterns are caused by partial cloud cover, also limiting photosynthesis.

© Fluxcom-X-Base

Example for a simulated net ecosystem exchange as produced by Fluxcom-X-Base (July 7th 2010, 6-7AM). Blue colors represent a net movement of CO₂ from the land to the atmosphere (release), red colors mean a net movement of CO₂from the atmopshere into the ecosystem (uptake). The differences are mainly a result of different light conditions. Further to the East, the sun has already risen, and plants are taking up CO₂ through photosynthesis. Further in the west, where it is still darker they are still in “night-mode”. Smaller, superimposed patterns are caused by partial cloud cover, also limiting photosynthesis.

© Fluxcom-X-Base

The methodology in principle was already established many years ago and has led to the publication of widely used global datasets. The major novelty of the new implementation FLUXCOM-X is that it facilitates relatively easy experimentation with strongly automated workflows. Jake Nelson, one of the lead scientists developing the model, remarks that “the work has strongly profited from the pioneering experience and expertise of Martin Jung, and from the diverse expertise of the international FLUXCOM-X team and colleagues from the Max Planck Institute for Biogeochemistry.” This large team brought together expertise in such diverse fields like ecophysiology, micro-meteorology, remote sensing, machine learning and scientific programming. The work also fundamentally relies on a big number of global researchers and technicians who operate hundreds of measurement stations, measure, process, and openly share their data in several flux networks such as FLUXNET, ICOS, and AmeriFLUX. “The timely availability and inclusion of global satellite observations is essential” adds Sophia Walther, co-lead researcher behind FLUXCOM-X.

For the first time, the new modelling environment covers fluxes of water vapor worldwide, based on measurements and calculations of meteorological and evapotranspiration data. The models will become instrumental in the global monitoring of greenhouse gas budgets. Combined with atmospheric inversions it will allow scientists to better detect sources and sinks of greenhouse gases. This will ultimately allow them to provide better policy guidance with respect to climate protection.

Earth’s breath and sweat uncovered - everywhere and continuously

Other Interesting Articles