White hydrogen spotting becomes artificial intelligence’s latest mission

Researchers have developed a model to help identify underground sources of naturally occurring H2.

Ohio State University researchers have developed a new artificial intelligence model capable of scanning the surface of the Earth for signs that white hydrogen reserves are hidden below.

The deep learning algorithm was used to spot possible indicators that these natural H2 deposits were present.

The researchers applied the AI to narrow down the possible locations where semicircular depressions (SCDs) in the ground could suggest the presence of white hydrogen deposits below. SCDs and ovoids are common ground formations around deposits of natural H2. That said, they don’t always indicate that H2 is present, so the AI was used to help narrow down the results to those that were the most promising.

The AI was also used to help find the locations in the first place, as the SCDs frequently appear in areas of low elevation, which means that it is easy for vegetation or even agriculture to disguise them.

White hydrogen below SCDs have been found in the US, Namibia, Brazil, Russia, France and the US.

These discoveries suggest that naturally occurring H2 is more commonplace than had previously been thought. Not long ago, the common belief was that H2 reserves underground were exceptionally rare and would be too difficult to find and access to make them a worthwhile source of what is now being viewed as a clean carbon-free fuel.

The researchers who developed the AI were led by Byrd Polar and Climate Research Center postdoctoral scholars Sam Herreid and Saurabh Kaushik from Ohio State University. They combined data from their artificial intelligence model with global satellite imagery to spot SCDs that were potential indicators of white hydrogen deposits.The researchers accomplished this by compiling a list of the known locations of SCDs, so they could train the artificial intelligence search algorithm. Remote sensing data was used to conduct an analysis of how those sites appeared from above. Once they had that information, they used geomorphic and spectral patterns to identify the locations worldwide that were most likely to have SCDs that would also be indicators of the presence of white hydrogen below.

AI’s ‘unique ability’

The researchers discovered that AI had a “unique ability” for mapping expressions on the surface of the earth for their potential as markers for natural hydrogen reservoirs. The artificial intelligence models were also very helpful for establishing a baseline that can be used for future investigations of sites that have potential sources of H2 below the ground.

“Hydrogen in general is a very attractive energy source,” said principal investigator for the project Joachim Moortgat, an Ohio State associate professor of earth sciences. “If you burn it, its only by-product is water, and unlike wind or solar energy, hydrogen can be stored and transported, so there are all kinds of industries trying hard to make the switch.”

Renewable white hydrogen

Another advantage to this particular type of H2 is that it is continually produced within the Earth’s crust. As a result, it is renewable and can be tapped as a source of zero-carbon fuel that is obtained nearly without the production of greenhouse gas emissions. Many believe that in adequate quantities, this type of H2 source could entirely change the shape of the global energy landscape.

Still, Moortgat underscored that new exploration tools are still required in order to be able to identify enough sources of white hydrogen to make this possible.

“One reason [hydrogen sites] are difficult to find is that they probably occur in different kinds of geologies and locations than where you would find oil or gas,” explained Moortgat. “But with the AI tools we develop, we map everything that could potentially be an SCD.”

Ready to test your knowledge on the most abundant element in the universe? Take our fun and engaging Hydrogen Quiz now! [forminator_quiz id=”58712″]