Natalie Pekney and Jim Sams are hunting across the United States for a treasure they’d rather didn’t exist: undocumented oil and gas wells, some so old that foliage has overgrown them and hidden the environmental hazards they pose. Like any good explorers, the team needs reliable maps—ones they’re constructing, in part, with historical documents.
Old photos and drawings, paired with new data gathered by remote sensing, help Pekney, an engineer with the U.S. National Energy Technology Laboratory (NETL), and Sams, a geologist with the information technology company Leidos, navigate as if they were some of the country’s earliest oil and gas prospectors themselves. “When you go out into the field and attempt to find wells, it can be really overwhelming,” Pekney says. The better their maps, the easier the search. “It increases confidence that if I walk to the coordinates that I have here, I can find a well.”
There are likely hundreds of thousands of wells that are both unplugged and unregistered with governments, according to the 21 state agencies that replied to the 2020 Interstate Oil and Gas Compact Commission survey. The unsealed wells they and their team from the NETL find, which still could be emitting the potent greenhouse gas methane or threatening to pollute groundwater, help create more complete datasets for state environmental agencies deciding which wells are causing the most damage, and which to seal up first.
Pekney and Sams belong to one of many government and university research teams helping to ferret out these abandoned wells. Methane leaks into the atmosphere are dangerous over time, and the EPA estimates that each unplugged well releases over 100 kilograms annually. A more immediate health concern are the fluids and gasses that could migrate into groundwater deposits, says Mary Kang, a civil engineer at McGill University. From 1983 to 2007, the Ohio Division of Mineral Resources Management identified 41 instances of groundwater contamination from leaking orphaned wells, for example, while the Railroad Commission of Texas confirmed 30 similar issues between 1993 and 2008.
“I would say groundwater is probably something I'd be concerned about more,” Kang says. “By no means do we understand the extent of methane emissions from these wells, but we know even less about groundwater impacts."
The forgotten wells came to national attention early this year when President Joe Biden’s infrastructure bill set aside $1.15 billion for states to properly seal them.
How to find a hidden well
For the first century of oil and gas drilling in the U.S., the practice of well plugging essentially didn’t exist. “There was no environmental regulatory structure of any kind in the United States. It simply wasn't a thing,” says Ron Bishop, a hazardous materials specialist at SUNY Oneonta. New York, for example, in 1879 passed the first legislation requiring wells to be plugged, but the law had no enforcement, save an amendment three years later that gave citizen whistleblowers a cut of the levied fine. If a drilling team did comply—and that was a big if—they might shove a tree trunk or bowling ball down the shaft, far from an effective seal for leaking gasses.
It wasn’t until the 1950s that cement well plugs became common. New York didn’t give any state agencies the authority, staff, and funds needed to track industry growth until the early 1970s with the New York Department of Environmental Conservation, Bishop says.
Research has yet to draw conclusions about how the age of a well influences how much methane it releases. The limited data available shows conflicting results, says Kang, who has assessed which well qualities lead to more methane emissions. Additionally, “a well could be a high methane emitter, but that doesn’t tell you anything about its impact to groundwater. A well could emit no methane and really be bad for groundwater,” Kang says. But plugged wells on average release less of the greenhouse gas into the atmosphere, she says.
If states want to know which wells pose the largest threats to the environment, they need to know how many drillings exist in the first place. To determine that, Pekney and Sams create and assemble maps of a study area. The goal of the multi-step process is to draw out as many clues as possible about where unplugged, forgotten wells might be. First, the team gathers the current-day maps of the study area that federal and state agencies have available. The United States Geological Survey topographic maps mark terrain, roads, and paths, for example, while counties or townships can offer recent maps of property lines.
Next, the NETL team flies drones outfitted with special survey equipment over the study area to make its own maps of the region. An electromagnetic field detector identifies metal well remnants, which show up as lumps on the landscape. To find wells fitted with wood pipes—used at the very beginnings of the oil and gas industry in the early 1800s—or those missing their metal bits, the NETL team scans for tell-tale depressions. A LiDAR survey, which showers the ground with beams of light and records how long it takes for each emission to bounce back to the device, highlights every divot under the greenery. A lollipop-shaped indentation might indicate a circular collapsed well and the long, narrow well pad alongside it.
Finally, Sams searches for old maps, photos, and illustrations relevant to the area.
Some reliable resources for older materials, Sams finds, are the archives assembled by federal documentation programs. The United States Geological Survey, for example, has digitized its collection of topographical maps it’s been creating since 1884. And in the 1930s, the United States Department of Agriculture began building its own vault of birdseye property photos, sometimes capturing well infrastructure and service roads. A simple internet search helps, too. That’s how Sams came across the David Rumsey Map Collection and an assortment of maps at the Library of Congress, which also show property lines, buildings, paths, and more from different eras.
When the collecting and surveying is done, Peckney and Sams insert all the maps they have accumulated into a software program that stacks everything like a sandwich, allowing the data to be viewed as if from above. Whichever map elements the team makes visible in the software appear on top of or alongside one another as if transposed into one document.
Since part of the data trove includes photos that capture infrastructure on the ground, Sams will rely on man-made features in the images, like a road intersection, to orient the pictures correctly on today’s landscape. From there, he can assign modern-day geographic coordinates to items of interest that might not otherwise be visible on a hike. Old service roads might be drawn into maps or even glimpsed in photographs taken when the forest had yet to cover the ruts, for example. With help from the LiDAR surveys that detect overgrown transportation furrows, Pekney and Sams have often followed old paths directly to rusted-out well pipes.
A rare trove of photos
Nowhere have historical records guided the pair better than in Oil Creek State Park. In 1859, that patch of northwestern Pennsylvania is where Edwin Drake became the first prospector to tap a pipe deep enough into the ground to release gushes of oil. His success inspired a frenzy of copycats. A local history written around that time claimed that in a four-month period in 1861, prospectors dug 665 new oil wells into the area.
In their desperation to exploit new oil pockets, those looking to get rich left tree trunks jutting out of the soil amongst rigs and ad-hoc housing. Anyone can see the destruction today thanks to John Mather. The photographer dragged his equipment out to boomtowns in northwest Pennsylvania in 1860, capturing workers and the changed landscape. “At Oil Creek, John Mather was photographing everything,” Sams says. “That was quite amazing work that he did.”
While on a project identifying wells in the area, Sams got in touch with the Drake Well Museum, which shared its archives with the team, including the Mather photos. “I wonder if he realized how helpful that would be 100 some years later,” Sams says, “that we could actually see oil derricks on a hillside and go and search those out from his photographs.”
Hiking through former drilling hotspots and seeing remnants for themselves is how Pekney and Sams verify that suspected wells exist. With their layered maps and Mather’s photos, the NETL team confirmed 245 wells in and around Oil Creek State Park. Testing for methane at 210 of them showed 21 were releasing the gas. All but one lacked concrete plugs.
Pekney and Sams know the trove of information available at the site of the first oil boom in the country is rare. But a search for other helpful historical documents is part of the process they’re deploying across the country. The team is or soon will be working on similar surveys in Kentucky and New York. Colleagues at other national labs are set to search for wells elsewhere in the country, as the Bipartisan Infrastructure Law set aside $30 million for a research consortium dedicated to the effort.
Some people might choose to deploy limited plugging funds differently and put a lower priority on identifying every well in the landscape. “I would probably try to do something about the wells that we know where they are,” Bishop says. But if the national lab team can streamline their protocol, then maybe state environmental agencies can adopt the process to more quickly identify wells and build a list of the ones that are most important to plug, Pekney says.
No matter the management strategy, it’s impossible to plug a well no one knows to exist. That is part of why the NETL team accepts all kinds of hints as to potential well whereabouts. The lab maintains a submission portal for any tips from the public, and they’re open to other information sources, too. Earlier this year, Pekney says, the landowner of a New York property the team is examining found some maps of the area while cleaning—records older than anything the team already had.
“Maybe those maps are turned into various state agencies,” Pekney says, “But I imagine homeowners with really old properties that go back 100 years—maybe they have some of this information.”