Somewhere beneath the iron-rich bedrock of Minnesota, Kansas, and the southern Appalachians, water is reacting with ancient minerals and splitting off hydrogen gas. That process has been running for millions of years. Until recently, most geoscientists assumed the gas would leak away long before anyone could drill for it. Now the U.S. Geological Survey has published the first map that ranks every square mile of the lower 48 states by the likelihood that geologic hydrogen has pooled in extractable quantities, and the results point to corridors that energy companies can begin testing with existing technology.
The map, announced by the USGS in its May 2026 national news release, is not a production forecast. It is a probability surface, a county-by-county ranking of where the geology, geochemistry, and structural traps line up in hydrogen’s favor. But for an industry that has spent decades ignoring stray hydrogen readings in old well logs, it amounts to a treasure map with coordinates attached.
What the map actually measures
Every location in the conterminous United States receives a score on a 0-to-1 relative scale. Higher numbers mean a greater chance that hydrogen has accumulated underground in concentrations worth investigating. The USGS built that score by stacking multiple geologic and geophysical layers: rock types known to generate hydrogen through water-rock reactions (especially iron-rich ultramafic and mafic formations), structural traps such as faults and folds that could hold gas in place, and surface indicators including documented hydrogen seeps.
A companion online map explorer lets users toggle between individual input layers and the integrated output. A geologist in Salina, Kansas, can zoom into a single county and see exactly which factors drove its score, whether that is the presence of Precambrian basement rock, a mapped fault system, or an anomalous gas reading from a nearby well. The tool replaces guesswork with a transparent, layer-by-layer breakdown.
One design choice gives the map unusual empirical grounding. The USGS calibrated its model partly against real gas-composition analyses logged in the agency’s Energy Resources Program Geochemistry Laboratory Database, a catalog of samples drawn from wells across the country. Helium concentration data from a separate USGS dataset also fed the model, because helium and hydrogen frequently co-occur in deep geologic source zones. Where helium has already shown up in existing well records, hydrogen may be nearby. That correlation gives drillers a shortcut: they can narrow their search using data that already exists, without commissioning new seismic surveys from scratch.
The scale of what might be down there
The strongest quantitative anchor for the resource’s potential comes from USGS researchers Geoffrey S. Ellis and Sarah E. Gelman. Their peer-reviewed paper in Science Advances modeled global geologic hydrogen resources and arrived at a striking central finding: the estimated energy content of recoverable geologic hydrogen exceeds that of all proven natural gas reserves on Earth. The estimate carries wide uncertainty ranges and is modeled rather than measured, but even the low end of the range represents a resource class large enough to reshape long-term fuel supply assumptions.
That finding is what pushed the USGS to commit staff and computing resources to a nationwide prospectivity assessment. The agency has publicly acknowledged that conventional wisdom long held hydrogen would not accumulate in usable quantities underground. The new map is a direct institutional response to growing field evidence that contradicts that view, including hydrogen-rich gas samples in the geochemistry database and analogous discoveries reported at sites outside the United States, most notably the Bourakébougou field in Mali, where a well has produced hydrogen continuously since 2012.
Other federal entities are building on the USGS work. Sandia National Laboratories references the prospectivity map in its own geologic hydrogen research program, using the federal dataset as a baseline for designing experiments and identifying test locations. That cross-agency uptake does not validate any specific volume estimate, but it signals that multiple parts of the U.S. research establishment now treat geologic hydrogen as a serious candidate in the broader energy portfolio, not a fringe curiosity.
What nobody knows yet
A high prospectivity score is not a drilling permit, and it is not a production guarantee. The map identifies where hydrogen is geologically plausible. It does not say how much a given well might yield per day, at what cost, or whether a reservoir will sustain output over years rather than weeks.
The Ellis and Gelman resource paper provides only aggregate global estimates. It does not break down recoverable volumes by basin or by the specific high-scoring zones on the new U.S. map. That gap matters because geology varies sharply over short distances. Two wells drilled a few miles apart might encounter very different fracture networks, permeability levels, or microbial communities. Certain subsurface bacteria consume hydrogen as a food source, and if those populations are active near a wellbore, they could eat the resource before it ever reaches the surface.
Sandia’s public materials flag additional open questions: production mechanisms that have never been tested at scale, materials compatibility in wells exposed to hydrogen-rich fluids, and the basic challenge of designing completions for a gas that behaves differently from methane. No primary test data on hydrogen-specific well integrity under U.S. conditions appear in the federal sources released so far.
Regulatory frameworks add another layer of uncertainty. Existing oil and gas rules were not written with geologic hydrogen in mind. It is not yet clear how state and federal agencies will classify mineral ownership, structure environmental reviews, or assess surface impacts for hydrogen-focused wells. In regions where communities are already wary of new drilling, permitting timelines could stretch for years. Until those policy questions are resolved, the distance between a promising map score and a producing well remains wide.
What the map changes right now
The strongest contribution of this release is structural, not volumetric. The USGS has built a reproducible, transparent framework for ranking where hydrogen is most likely to exist and has made every input layer, scoring method, and output file publicly downloadable. Independent teams can stress-test the model, swap in alternative assumptions, or overlay proprietary datasets. That openness sets the map apart from the black-box exploration models used by private companies.
For energy companies weighing exploration budgets, the practical first step costs almost nothing: download the prospectivity data and cross-reference high-scoring zones against existing well databases and lease positions. Companies that already hold mineral rights in flagged areas can re-examine archived gas analyses for hydrogen content that was previously ignored or left unreported. That kind of desktop screening could identify candidates for targeted field sampling or recompletion of legacy wells long before anyone commits to a new exploratory drill.
For policymakers, the dataset offers a common reference point for coordinating federal research funding, state-level permitting reforms, and regional infrastructure planning. Because the map is national in scope, it can help agencies anticipate where interest in hydrogen leasing or pilot projects is likely to emerge and begin evaluating environmental and community considerations before applications arrive.
None of this guarantees a hydrogen boom. What it guarantees is that the subsurface potential is now mapped well enough to justify focused, well-by-well testing. The decisions made over the next few years, by drillers, regulators, and the communities living above these formations, will determine whether geologic hydrogen moves from a probability surface on a government server to an actual fuel flowing through American pipelines.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
