A World of Climate Data

When humans breathe, we take in oxygen from the air, transfer it into our blood, and breathe out carbon dioxide. With plants, it’s the reverse: they “breathe” by taking in carbon dioxide, depositing it into their tissues via photosynthesis, and then releasing oxygen. This process offers potential for the Earth’s ecosystems to take carbon out of the air, which is crucial as humans race to reduce the greenhouse gas emissions that are fueling the rapid advancement of climate change. 

Plants also release carbon dioxide when they decay, and when they burn in wildfires. Studies by the Inter-governmental Panel on Climate Change and others estimate that the uptake of CO2 exceeds these releases, and that plants may take up and store as much as a quarter of human-made carbon emissions. To better understand the capacity of ecosystems to act as long-term carbon sinks, scientists at Rausser College of Natural Resources and across the globe use micrometeorological towers fitted with sensors capable of measuring the ebb and flow of carbon dioxide and other greenhouse gases.

“When you breathe, a mass flow—air velocity multiplied by the concentration of gas—moves into and out of your mouth,” says Dennis Baldocchi, a professor of biometeorology in the Department of Environmental Science, Policy, and Management (ESPM). “We’re measuring that in ecosystems with a set of sensors placed over crops, grasslands, forests. We can see the eddies—the movement of air up and down—past our sensors, and the amount of gases they contain.”

Baldocchi helped pioneer this technology and methodology, called eddy flux, or eddy covariance. For two decades he led FLUXNET, a global network of scientists that shares open-source data, and his Biometeorology Lab has trained dozens of graduate students and postdoctoral researchers who now manage flux tower projects from the Bay Area and Arkansas to Panama and Estonia. 

“Grasslands and forests and every single kind of terrestrial ecosystem around the globe are ‘breathing.’ It would be very cool if, just the way we see our chests rise and fall, there was a way to see it,” says Margaret Torn, BS ’84 Conservation and Resource Studies; PhD ’94 Energy and Resources, a professor in the Energy and Resources Group and a senior scientist at Lawrence Berkeley National Laboratory. 

Torn coordinates the AmeriFlux Network, a network of towers in North, Central, and South America. “Having an eddy covariance tower to watch cycles in the data is maybe the closest we come to actually see the breathing of the biosphere.”

A global collaboration network

While researchers had measured fluxes of mass and energy since the 1950s, the eddy covariance technology developed in the 1980s introduced the capability to measure carbon dioxide flux, the rate at which CO2 moves into or out of a land surface. The approach allows scientists to take flux measurements above forest, crop, grassland, chaparral, wetland, and tundra ecosystems, attaching sensors to towers ranging from 15 to 225 feet high. In addition to carbon dioxide, water vapor, and other trace gases, the micrometeorology towers measure a range of environmental factors including sunlight, temperature, and soil moisture.

“In the early days of this technology, only four or five groups were able to make direct measurements of carbon dioxide fluxes,” says Baldocchi, who was working at the National Oceanic and Atmospheric Administration (NOAA) at the time. “The importance of understanding how ecosystems are consuming carbon dioxide through photosynthesis or producing it through leaf and soil respiration at ecosystem scales kick-started our desire to make continuous, year-round measurements.”

As a scientific community coalesced around measuring carbon fluxes, an international and multidisciplinary group of scientists, including Baldocchi, met in 1995 in La Thuile, Italy, for a workshop where some of the first continuous, years-long flux measurements were presented.

One outcome of the workshop was the first meeting in 1997 of FLUXNET, a loose “community of the willing,” with Baldocchi at the helm. FLUXNET compiled, archived, and distributed eddy covariance data globally and calibrated it to ensure consistency and comparability, while building the scientific community via workshops, conferences, and technical support. 
“The fun part was developing friendships and collaborations with people from all over the world,” Baldocchi says. 

Ecosystem responses to climate changes

Another important outcome of the La Thuile workshop was the launch of the AmeriFlux Network in 1996, with support from the US Department of Energy, NASA, NOAA, and the US Forest Service. 

Based at Lawrence Berkeley National Laboratory, the AmeriFlux Network now has more than 500 towers in the Americas making eddy covariance measurements at any given time. By 2025, the network had amassed 3,562 site years of data from 11 countries, which has been downloaded 42,141 times. “The importance of these networks is for people to be able to understand the global carbon cycle and how ecosystems are functioning around the world,” Torn says.

Torn leads the AmeriFlux Management Project, which launched in 2012 with ongoing support from the US Department of Energy. The project serves the robust AmeriFlux Network with technical support, data quality control and processing, assistance in ensuring continuity in the longest-running sites, and community events and training. In addition to validating the ability of ecosystems to act as carbon sinks, Torn says, research by the AmeriFlux Network has greatly contributed to scientific understanding of how ecosystems recover from disturbances such as logging, wildfires, and extreme weather events. 

“Ecosystems tend to change a lot, but they recover in one way or another,” Torn says. “A fire may cause a lot of erosion and degradation, but new plants come in or seeds resprout into trees. Recovery from disturbance is an important part of our overall carbon cycle and climate system.”

When Baldocchi arrived at ESPM in 1999, he organized the second international FLUXNET workshop in Marin County. That same year, sensors mounted on NASA’s Terra satellite began sending back Earth-scale data on carbon cycling, energy absorption, evaporation, and other climate-related factors. In addition to validating satellite data, eddy covariance researchers sought to understand carbon cycling at the ecosystem level.

“It’s very hard to tell from remote sensing what an ecosystem is actually doing, or how much carbon is being taken up,” says Trevor Keenan, an associate professor in ESPM and director of the FLUXNET Coordination Project. “Eddy covariance towers are literally taking the only observations available at the ecosystem scale.”

In 2001, Baldocchi and colleagues reported key findings from FLUXNET data, including that the age of a forest influences its ability to exchange carbon dioxide and water vapor, and that the efficiency of ecosystem carbon dioxide exchange (the amount of carbon assimilated per unit of sunlight) under cloudy skies was about twice the efficiency under clear skies. “The quality of light has a big influence on photosynthesis of a whole forest,” Baldocchi says. 
Over the years the scientific community using eddy covariance grew to about a dozen regional networks—including the ICOS (Integrated Carbon Observation System) in Europe and AsiaFlux networks—that are involved in the gathering and processing of open-source flux data. 

Baldocchi led FLUXNET for more than two decades and operated 15 towers in the San Francisco Bay-Delta Watershed and Sierra foothills, and his former students and postdoctoral researchers now operate more than 50 flux towers in California, nationally, and globally. For example, Alex Knohl of Germany’s University of Göttingen manages towers in palm oil fields and mountain tropical rainforests of Indonesia; Matteo Detto of Princeton University monitors towers in Panama’s tropical, broadleaf, and evergreen rainforest; and Kuno Kasak, currently a visiting researcher in Baldocchi’s Biometeorology Lab, operates towers in an abandoned peat-extraction site and restored wetlands in Estonia.

FLUXNET coordination project

Today, an estimated 3,000 flux tower sites operate worldwide, largely due to dramatic improvements in sensor technology and affordability as well as greatly expanded computing power and storage. About 500 sites currently share measurements with FLUXNET, Keenan says, and a few hundred more are joining this year.

The resulting long-term, continuous datasets are helping climate scientists understand the ability of various ecosystems to sequester carbon, the effectiveness of climate solutions such as wetland restoration and forest regeneration, and the climate impacts of extreme weather events. The number of scientific papers based on eddy covariance data published has grown from a handful during Baldocchi’s early days in the field to hundreds annually.

“We need to make these long-term measurements on ecological timescales, and in doing so we have to plan for generational succession so data collection can continue,” Baldocchi says. He’s passing the torch to Keenan, who now leads the FLUXNET Coordination Project, which received $2 million from the National Science Foundation in 2021 to facilitate international collaboration, offer training and research opportunities, and coordinate how global open-source flux data is shared.

In the past, FLUXNET periodically released comprehensive datasets compiled from contributing flux networks. “There was a massive effort to process and release the data, and that was it,” Keenan says. “We’re moving toward a model where the data will be dynamically updated and continuously accessible.” The first release of the new data system, being developed by ICOS, AmeriFlux, and other regional networks, is expected by the end of this year. 

Importantly, future versions of the FLUXNET data infrastructure will include measurements of methane, a more potent but less prevalent greenhouse gas that degrades about 10 times faster than carbon dioxide. “Methane sensors have proliferated over the past few years, and now many sites are making those measurements,” Keenan says.

Kyle Delwiche is deputy director of the FLUXNET Coordination Project and co-facilitator of the FLUXNET Community Council, which meets monthly in two time zones so that scientists from all over the world can participate. Delwiche’s research focuses on methane emissions from natural ecosystems, particularly over restored wetlands. 

“If we make changes today that reduce our methane emissions, we’ll see the effects of that far sooner than for CO2,” Delwiche says. “The more we understand the methane budget for the world, the more we can predict what’s going to happen with the climate.”

Studying nature on its own terms

As Baldocchi heads toward retirement this summer, his contributions to the science of eddy covariance have been recognized with top honors in the field. “Dennis has pushed the boundaries of the application of micrometeorology theory and technique,” Keenan says. “He’s been a cornerstone of the community, a huge ambassador for the science and the FLUXNET network, and the glue that has kept the regional actors together and moving forward.”

Baldocchi credits his too-many-to-name scientific mentors, colleagues, students, and research teams for keeping the eddy covariance data flowing. “We now have over two decades of data from hundreds of sites around the world,” he says. “We have to study nature on its own terms—measuring how it’s responding to all these multifaceted changes. The key to our success has been creating and sharing data so that the whole adds up to more than the sum of the parts.”