We can now tell how much CO2 in the air is due to fossil fuel burning

A way of distinguishing between natural carbon dioxide emissions and those from burning fossil fuels could help cities and countries monitor their progress in cutting emissions

By Michael Le Page

A way of directly measuring the carbon dioxide released by burning fossil fuels could help cities and countries monitor their efforts to reduce emissions in near real time.

“We are in a shrinking window of time to do this, so I think we really need to know what the situation is as quickly and as accurately as possible,” says Penelope Pickers at the University of East Anglia, UK.

At present, governments and research organisations estimate countries’ overall emissions based on data such as how much oil or gas has been sold. While initial estimates are often made fairly quickly, it can take years to fully compile this information and estimates can vary substantially.

Measuring fossil fuel emissions directly would help confirm the accuracy of these inventory-based estimates and reveal more quickly if emission-reduction policies are working or not. It could also enable us to track how much specific regions or cities are emitting.

But such measurements are extremely difficult, because plants take up or release varying amounts of CO2 as the seasons shift and weather changes. It is like standing on a beach and immediately trying to tell whether the tide is going in or out, as waves are constantly coming and going.

So, while the long-term global rise in atmospheric CO2 due to human activity – from around 280 parts per million before the industrial revolution to nearly 420 ppm today – is crystal clear, the short-term, regional picture is much less so.

Researchers have tried various ways of directly measuring fossil fuel emissions. One is to determine what proportion of CO2 is in the form of the radioactive isotope carbon-14, which isn’t found in fossil fuels because it decays over time, and oil and gas supplies are millions of years old. But this requires the collection of samples in flasks, so continuous measurement isn’t possible. What’s more, some types of nuclear reactors emit carbon-14, obscuring the picture.

Pickers’s team has used an alternative approach based on measuring both atmospheric oxygen and CO2 simultaneously. When plant matter is used as food, or it decays or burns, the ratio of the oxygen lost from the atmosphere to increased CO2 is around 1.1. For coal, it is around 1.2 and for gas it is 2.

The researchers used measurements taken at the Weybourne Atmospheric Observatory on the Norfolk coast to calculate emissions from the southern UK since 2020. They used machine learning to estimate how changes in weather and wind direction affect oxygen and CO2 levels in the area.

The team was able to detect falls in fossil fuel emissions during the first and second covid-19 lockdowns in England. “Covid has been a great example of a quite sudden, abrupt change,” says Pickers.

With around four observatories, it would be possible to measure emissions from Britain, she says. More would be needed to monitor individual cities.

The study makes a strong case that the method is effective, says Brad Weir at the NASA Goddard Space Flight Center in Maryland. But building monitoring stations around the world would take a lot of time and money, he says.

“If we are going to have a fossil-fuel-carbon-monitoring system, it’s going to start with satellites,” says Weir.

His team reported last year that falls in carbon dioxide emissions due to the pandemic were detectable using existing CO2-sensing satellites, and there are plans to launch more missions focused on detecting this gas in the coming years.

Pickers says the problem with satellites is that they can’t detect CO2 through clouds and can’t distinguish between biological and fossil fuel emissions. Instead, satellite estimates rely on computer models of natural processes to determine fossil fuel emissions.

But these models are informed by data and can be highly accurate, says Weir, who points out that Pickers’s team relies on the “black box” of machine learning.

Ultimately, the best results may come from using all the different methods. “We should go for a combined approach,” says Pickers.

“We are going to have to integrate all of these observations,” says Weir.

Journal reference: Science Advances, DOI: 10.1126/sciadv.abm3952

Sign up to our free Fix the Planet newsletter to get a dose of climate optimism delivered straight to your inbox, every Thursday