Natural gas is now flowing so fast into U.S. pipelines that the big question seems to be what to do with it all: Engineer cars to run on methanol? Reopen shuttered chemical plants that rely on gas for feedstock? Export liquefied gas by tanker? With about two-thirds of U.S. states thought to hold natural gas reserves, many take President Barack Obama seriously when he calls the United States the "Saudi Arabia of natural gas."
But just how much natural gas does the United States have?
A close look at the assessments shows that even the experts disagree. Most dramatically, the U.S. Energy Information Administration (EIA), the government's own analytical team, last month slashed in half its estimate for a key and large subset of reserves: the amount of gas in shale rock formations across the country.
Although the government's new estimate for total U.S. natural gas resources—2,214 trillion cubic feet (tcf)—is a third higher than its 2008 estimate, the shale gas markdown underscores the uncertainties around this new supply source. In an interview with National Geographic News, the EIA has offered a sneak preview of the more detailed explanation it will publish in April on why its shale gas estimate plummeted.
But with other geologists convinced that EIA's new numbers are too conservative, it is certain that there will be plenty of debate ahead on the size of the energy windfall from shale gas.
Ever since Texas gas producers proved less than a decade ago that they could stimulate production from seemingly impermeable and ubiquitous shale rock through a combination of hydraulic fracturing and horizontal drilling, shale gas production has been booming across the United States.
(See interactive: "Breaking Fuel From Rock")
Shale gas now makes up about 23 percent of U.S. production, helping to offset a fall in natural gas supplies coming from conventional wells. Shale production could double or triple over the next 25 years, EIA projects, estimating it will make up nearly half of U.S. production by 2035.
As a result, natural gas prices in the United States are now at their lowest point in a decade. Throughout the northeastern United States, homes long reliant on oil for heating have been switching in droves to cheaper natural gas. Gas now seems so abundant that the United States, which only a few years ago was contemplating major imports to address a perceived shortfall, now is contemplating new expensive port facilities to chill and liquefy gas to export it by tanker overseas. (The EIA now projects U.S. liquefied natural gas exports starting in 2016.)
The U.S. chemical industry, which depends on natural gas as a feedstock, was fleeing offshore amid high prices only a decade ago; now it is contemplating new factories in the heart of shale country.
Tom Ridge, Pennsylvania's former Republican governor and the first U.S. secretary of Homeland Security, who has worked as a consultant to the natural gas industry, co-authored an op-ed in last week's New York Times urging Congress to consider requiring carmakers to design vehicles to run on 100 percent methanol, an alcohol that can be derived from natural gas. Even though methanol has half the energy content of gasoline, he argued that the money saved compared to today's oil prices, and the security of relying on a domestic resource, would make the switch worthwhile.
But the wisdom of any of these new policy choices for an era of U.S. natural gas abundance all depends upon how much natural gas there actually is.
New doubts seemed to emerge on January 23, when the EIA issued the "early release" of its Annual Energy Outlook. A key estimate, the amount of natural gas in "unproved reserves"—areas that have not yet been drilled—was lowered dramatically. EIA's new estimate for how much gas these areas might yield was just 482 tcf, a bit more than half the 2011 estimate of 827 tcf.
The most drastic change was for the Marcellus shale formation, underlying Pennsylvania, New York, and West Virginia. The EIA used the new analysis from the U.S. Geological Survey (USGS), along with the known histories of production from wells in the area, to create a new estimate.
(Related: "Natural Gas Stirs Hope and Fear in Pennsylvania")
The 2011 estimate figured "unproved reserves" in the Marcellus were 410 tcf, but the new estimate plummeted by two-thirds, to 141 tcf.
Because shale gas production in the Marcellus is still relatively new, having catapulted to prominence only in the past few years, researchers are still working to understand just how much gas each well might produce over the long run.
In the "core area" of the Marcellus—where most of drilling activity has taken place so far—the EIA had assumed before that each well would, in time, produce 3.5 billion cubic feet of gas. But with new data on how much gas wells have produced to date, the EIA's calculation came out lower, with each well expected to produce about 2 billion cubic feet, a drop of about 40 percent.
Proving Their Worth
And the issue is more complicated still. There is a bewildering array of classifications for energy reserves and resources, with no standard methods or terminology for most types of estimates.
Of all the types of reserves, the most certain class of reserves are known as "proved reserves," which publicly traded natural gas producing companies in the United States have to estimate and report to the U.S. Securities and Exchange Commission (SEC). There are strict rules about how these reserves must be estimated, since these numbers are often taken as a measure of a company's success-and so can affect their stock prices.
According to the EIA's latest data, in 2011 the U.S. had 60 tcf of proved reserves of shale gas.
(There may be room for doubt even over this figure, since the SEC is investigating whether companies overstated their proved reserves.)
On top of these "proved reserves" are a variety of other types of estimates from the EIA, the USGS, and nongovernmental groups, which include "inferred reserves" and "unproved reserves," as well as "undeveloped discovered resources" and "undiscovered resources."
"We've gotten a lot of questions in the past about these different estimates," said energy analyst John Staub of the EIA.
"We're trying to make it more intuitive," using a simpler system, he added. Instead of using a variety of different estimates, "we've switched this year to just talking about proved reserves and 'unproved' being everything else."
In addition to the changes in the Marcellus estimate, to estimate the "unproved reserves" for four of the other major shale gas regions—the Eagle Ford, Haynesville, Fayetteville, and Woodford formations—the EIA used estimates published last year by the USGS, which led to some drops compared to the EIA's 2011 estimate.
For these four areas, the EIA used the USGS estimates straight off the shelf, Staub said. Although the two government bodies use different language, the EIA's "unproved reserves" are "essentially the same as the 'undiscovered resources' that the USGS talks about," he said.
With the new assessments from the USGS, the EIA estimates that, outside of the Marcellus, the "unproved reserves" are lower than it estimated in 2011, dropping from 417 to 341 tcf.
Adding up all the proved and unproved reserves around the country, the EIA now estimates there is 542 tcf of shale gas available—roughly half what it estimated in 2011.
The Early Stages
However, geologist Terry Engelder of Pennsylvania State University argued that the recent EIA estimate is too conservative, given the data on which the agency based its conclusions.
Engelder is often given credit for spurring the shale gas rush in the Marcellus with early estimates that the formation held large amounts of natural gas. In his most recent published estimate, from 2009, he figured the Marcellus could in the long run yield 489 tcf, a number in the same ballpark as the EIA's 2011 estimate.
More recently, he has obtained production data from leaseholders for a small number of wells, to see how much they produce, and to update his estimate.
Engelder's earlier estimate was based on very limited data available at the time. "That's a pretty challenging thing to do," he said, "to take just 50 wells and try to project what a field might do that might ultimately end up having 100,000 to 300,000 wells."
With new data on production from 16 Pennsylvania counties, Engelder has updated his estimate. Though he has yet to publish the results, "half the counties are doing better than predicted, and half of them are not doing quite as well as predicted," he said. "But on average, it is just right where we were with that 2009 estimate."
Engelder said the geological analysis by the USGS—which was a crucial input for the EIA's reassessment of the potential for Marcellus shale gas—is problematic.
A key problem, Engelder said, was that the USGS assessment broke up the Marcellus into thousands of parcels, and then assumed that only 37 percent of them would yield significant natural gas. Engelder thinks that a lot more of the parcels will be productive.
Sorting out this disagreement may have to wait until more data comes in.
The Marcellus "is a very large geographic area," said Staub of the EIA; for the 2012 estimate, the agency had access to production data from about 3,000 wells.
With longer histories from individual wells, their potential over the long term should become clearer. "Three thousand wells is not that many," Staub added. "We're still in the early stages of development."
(Related photos and video: "Faces of the Gas Rush")
This story is part of a special series that explores energy issues. For more, visit The Great Energy Challenge.