A study recently published in the journal Nature finds plants have been absorbing 31 percent more carbon dioxide from the atmosphere than previously assumed and modeled.

The international team of researchers, led by Jiameng Lai at Cornell University, examined gross primary production (GPP), the largest carbon flux (carbon sink and cycling) in the biosphere. The GPP is calculated in petagrams of carbon per year, with one petagram equaling about 1 billion metric tons, approximately the amount of carbon dioxide (CO2) emitted annually from 238 million internal combustion engine vehicles.

The research team used a new, integrated model they developed to track the cycling of carbonyl sulfide, or OCS, “from the air into leaf chloroplasts, the factories inside plant cells that carry out photosynthesis,” SciTechDaily reports. OCS is a good proxy for CO2 and is more easily measured. Science Direct writes:

The research team quantified photosynthetic activity by tracking OCS. The compound largely follows the same path through a leaf as CO2, is closely related to photosynthesis, and is easier to track and measure than CO2 diffusion. For these reasons, OCS has been used as a photosynthesis proxy at the plant and leaf levels. This study showed that OCS is well suited to estimate photosynthesis at large scales and over long periods of time, making it a reliable indicator of worldwide GPP.

To fill in the model’s parameters on plant growth and GPP, the scientists used plant data from a variety of sources. To verify the accuracy of the inputs, the researchers compared them with high-resolution data from environmental monitoring towers instead of satellite observations, which can be hampered by cloud cover, especially in tropical regions.

Whereas historically, the GPP has been thought to be approximately 120 petagrams of CO2 per year, their results suggest the Earth’s plant life is removing between 157 and 175 petagrams of CO2 per year. What is the difference? They write,

Our global GPP is higher than satellite optical observation-driven estimates (120–140 PgC yr–1) that are used for Earth system model benchmarking. This difference predominantly occurs in the pan-tropical rainforests and is corroborated by ground measurements, suggesting a more productive tropics than satellite-based GPP products indicated. As GPP is a primary determinant of terrestrial carbon sinks and may shape climate trajectories, our findings lay a physiological foundation on which the understanding and prediction of carbon-climate feedback can be advanced.

It seems that persistent cloud cover in tropical regions often distorts satellite mapping of plant life.

“Figuring out how much CO2 plants fix each year is a conundrum that scientists have been working on for a while,” Lianhong Gu, a coauthor of the study and a distinguished scientist with the Environmental Services Division of the Oakridge National Laboratory (ORNL), said in a press released quoted by SciTechDaily. “The original estimate of 120 petagrams per year was established in the 1980s, and it stuck as we tried to figure out a new approach.

“It’s important that we get a good handle on global GPP since that initial land carbon uptake affects the rest of our representations of Earth’s carbon cycle,” Gu said. “We have to make sure the fundamental processes in the carbon cycle are properly represented in our larger-scale models. … For those Earth-scale simulations to work well, they need to represent the best understanding of the processes at work.”

Peter Thornton, who leads the ORNL Earth Systems Science Section but did not participate in the research, agrees properly understanding and accurately modeling the carbon cycle is critical to an analysis of climate change and its impacts.

“Nailing down our estimates of GPP with reliable global-scale observations is a critical step in improving our predictions of future CO2 in the atmosphere, and the consequences for global climate,” Thornton told SciTechDaily.

This piece originally appeared on the newsletter Climate Change Weekly and has been republished here with permission.