Levels of molecular hydrogen (H₂) in the atmosphere have surged in modern times due to human activity, according to new research.

When scientists analyzed air samples trapped in drilled cores of Antarctica's ice, they found atmospheric hydrogen had increased 70 percent over the course of the 20th century.

Even as recent air pollution laws have sought to curb fossil fuel emissions, hydrogen emissions have continued to surge with no signs of slowing down. And there's a chance that leakage is to blame.

Molecular hydrogen is a natural component of our atmosphere due to the breakdown of formaldehyde, but it is also a byproduct of fossil fuel combustion, especially from automobile exhaust, and biomass burning.

While hydrogen doesn't trap heat in the atmosphere on its own, it can indirectly impact the distribution of methane and ozone. After carbon dioxide, these are the two most important greenhouse gases, which means global hydrogen levels can also perturb the climate.

Nevertheless, the sources and sinks of atmospheric hydrogen are rarely studied. We don't even have a good estimate of how much humans have emitted since industrial times.

The current study is the first to offer up a solid figure. Between 1852 and 2003, air samples from near the South Pole of Antarctica suggest atmospheric hydrogen jumped from 330 parts per billion to 550 parts per billion.

"Aging air is trapped in the perennial snowpack above an ice sheet, and sampling it gives us a highly accurate account of atmospheric composition over time," explains Earth scientist John Patterson from the University of California Irvine.

"Our paleoatmospheric reconstruction of H₂ levels has greatly enhanced our understanding of anthropogenic emissions since the beginning of the industrial revolution."

The news isn't exactly good. As it turns out, we may have been significantly underestimating our hydrogen emissions.

Some tailpipe emissions have been mitigated in recent years with the use of catalytic converters, and ideally we would have seen hydrogen emissions decrease or even plateau as well.

Yet hydrogen levels have continued to rise in the atmosphere, almost uninterrupted.

"[W]e are likely underestimating nonautomotive sources of the gas," says Patterson.

Instead, there must be another rapidly increasing source that is offsetting our progress in the automobile industry – we just don't know where it's coming from.

This isn't the only dataset to identify such a discrepancy. Prior research has also shown a consistent rise in hydrogen from 2000 and 2015, distinct from trends in other forms of exhaust pollution.

In terms of human-caused emissions, hydrogen emissions are thought to mostly come from automobile exhaust, but hydrogen leakage from industrial processes is rarely considered.

No one has directly measured how much hydrogen leaks from these processes, but initial estimates suggest it could be significant. 

A 10 percent leakage rate between 1985 and 2005 would account for roughly half the rise in recent hydrogen emissions, researchers estimate.

They can't be sure this is where the hydrogen is coming from – hydrogen emissions from coal combustion are also seriously understudied – but the authors argue it's worth investigating more.

Especially since green hydrogen processes, which split hydrogen from water to create carbon-free power, could also result in substantial leakage if they are one day scaled up, as some climate scientists and environmentalists hope they will be.

This isn't a new worry. It's a concern scientists have been pointing out for years now.

If hydrogen one day leaks from industrialized hydrogen gas plants, experts are troubled it could increase the lifetime of methane in our atmosphere, a greenhouse gas 20 times more potent than carbon dioxide.

Although, even with a small percentage of leaks, a global hydrogen economy would likely have far lower climate impacts than our existing fossil fuel-based energy system, researchers estimate.

Scientists are now on the hunt to find the mysterious source of hydrogen we seem to have been missing all along. If at least some of it turns out to be leakage, the future of green hydrogen might have a problem in need of solving.

The study was published in PNAS.