Musings: Trying To Solve Mystery Of Missing Marcellus Resource

A tenant of America's gas shale revolution is that shale is ubiquitous and uniformly spread under our oil and gas producing basins.  That belief has translated into growing estimates of the resource's potential and how it has radically changed the long-term outlook for America's, and potentially the world's energy future.  Is it possible this tenant has been knocked into a cocked hat by the latest estimate of the resource potential of one of our largest gas shale basins – the Marcellus Shale? 

The recent assessment by the U.S. Geological Survey (USGS) that the Marcellus Shale contains 84 trillion cubic feet (Tcf) of undiscovered natural gas and 3.4 billion barrels of undiscovered natural gas liquids was greeted with both joy and consternation.  The joy came from the recognition that the USGS estimate was a huge increase from its prior assessment made in 2002 that said there was only about 2 Tcf of gas reserves in the shale formation that stretches from Alabama to New York.  The consternation stems from the assessment being about 80% less than an estimate promoted earlier this year by the Energy Information Administration (EIA) that there was 410 Tcf of gas in the basin.  Talk about a gap wide enough to drive a truck through, how about a whole fleet of pickups?

First, it is important to understand that the USGS estimate is the mean of various estimates the agency prepared.  Each estimate was assigned a confidence level based on how sure the agency was that the estimated volume actually is present.  The estimates ranged from a very highly confident (95%) estimate of 43 Tcf to the estimate with the lowest confidence (5%) of 144 Tcf.  The 50% confidence scenario estimated total gas reserves of 78.7 Tcf, or somewhat below the mean estimate the agency decided to publish.  Second, it is important to understand that these estimates reflect a view that the resources are technically recoverable, which, to quote from the agency's press release, means "are those quantities of oil and gas producible using currently available technology and industry practices, regardless of economic or accessibility considerations."  The USGS went on to say, "…these estimates include resources beneath both onshore and offshore areas (such as Lake Erie) and beneath areas where accessibility may be limited by policy and regulations imposed by land managers and regulatory agencies."

Importantly, the USGS attributed the increase in its undiscovered resource estimate to the "new geologic information and engineering data, as technological developments in producing unconventional resources have been significant in the last decade."  Clearly, the USGS was referring to the improvements in horizontal drilling and hydraulic fracturing, which the petroleum industry has embraced wholeheartedly in driving the gas shale revolution. 

In doing periodic Google searches of news articles discussing the new USGS estimate, we found that they overwhelmingly focused on the new estimate being 42 times greater than the 2002 estimate.  In fairness, the successful marriage of horizontal drilling and hydraulic fracturing that has driven the success of the gas shale revolution had barely begun when the previous assessment was prepared, so it is not surprising the new estimate jumped.  For people counting on the Marcellus Shale to rewrite the nation's plan for powering the Northeast region, there was concern about how low the USGS estimate is compared to the recent estimates made by academics and consultants, one of which was officially embraced by the EIA. 

In Exhibit 1, we show a table taken from the Scope and Results section of the report, "Review of Emerging Resources: U.S. Shale Gas and Shale Oil Plays" published in July by the EIA and based on a report it had commissioned.  The agency contracted with the consulting firm INTEK, Inc., "to develop an assessment of onshore Lower 48 States technically recoverable shale gas and shale oil resources."  As can be seen from the table, the Marcellus Shale is assigned an estimated 410 Tcf of "undeveloped technically recoverable shale gas and shale oil resources remaining in discovered shale plays as of January 1, 2009."  Both the description of what the 410 Tcf represents and the date of the estimate are important and we will come back to them in a minute.

The Marcellus Shale Coalition, fearing the USGS estimate might be misinterpreted, undertook a "myth dispelling" effort in a press release.  The Coalition's release was titled, "Myth vs. Fact: USGS/EIA Marcellus Data."  The thrust of the release was to counter those news reports that focused not just on the huge increase in the USGS resource estimate, but how it clashed with the EIA's assessed resource estimate.  The relevant section of the press release is quoted below.

Exhibit 1.  EIA Estimate of Gas Shale Resources

"Over the course of the past week, while several flawed media reports surfaced regarding the new USGS data, straightforward facts ultimately prevailed. Below are several of these claims and corresponding facts. Most importantly, the latest production estimates remind us that, more than ever before, the dialogue about the Marcellus Shale continues to be rightfully focused on the region's leading role in providing clean-burning, job-creating American energy for our nation. (Coalition emphasis). 

"CLAIMS: Citing Philip Budzik, an Energy Information Administration (EIA) bureaucrat who touts the work of anti-shale gas groups like the Natural Resources Defense Council (NRDC), Bloomberg News claims new USGS data will cause "U.S. to Slash Marcellus Shale Gas Estimate 80%." (Coalition emphasis). 

"The U.S. will slash its estimate of undiscovered Marcellus Shale natural gas by as much as 80 percent after an updated assessment by government geologists. The formation, which stretches from New York to Tennessee, contains about 84 trillion cubic feet of gas, the U.S. Geological Survey said today in its first update in nine years. That supersedes an Energy Department projection of 410 trillion cubic feet, said Philip Budzik, an operations research analyst with the Energy Information Administration. (Bloomberg News, 8/23/11) 

"The New York Times' Ian Urbina, in his story entitled "Geologists Sharply Cut Estimate of Shale Gas," writes that the new USGS data "is drastically lower than the 410 trillion cubic feet that was published earlier this year by the federal Energy Information Administration," and also cites Mr. Budzik. (New York Times, 8/24/11) (Coalition emphasis). 

"FACTS: USGS notes that "it's important to do an apples-to-apples comparison here" – something Bloomberg News and the New York Times fundamentally failed to do. "Hold off on those shale gas obituaries," accurately reports the Washington Post's Brad Plumer. (Coalition emphasis). 

"As it turns out, many of those stories may have been premature — and appear to be based on a slight misunderstanding of the USGS survey. … As Brenda Pierce, program coordinator for the USGS energy resources program told me, it's important to do an apples-to-apples comparison here. The USGS and EIA aren't measuring the same thing, exactly: The USGS offered an estimate of undiscovered resources that can be recovered with current technology, whereas the EIA report looked at both "active" and "undeveloped" reserves together. "Ours is additive to what's already in production," explains Pierce. … So the revision could end up being much less stark than early media reports implied. (Washington Post, 8/26/11) 

"As [Kenneth Medlock, an energy fellow at Rice University's Baker Institute] explains, the USGS figures refer only to "undiscovered" resources — which implies that gas that producers have already discovered cannot be included in the USGS assessment. The terminology favored by the EIA, meanwhile, implies that it can include resources that producers have already discovered. The variation can be significant. … The USGS estimate would theoretically exclude such reserves because they have already been discovered, while the EIA estimate could include them. … Medlock said that anybody who points to the latest USGS assessment as an indication that the Marcellus is less prolific than initially thought should take note of the magnitude of the increase. "They went from 2 Tcf to 84 Tcf. That's a pretty big signal," Medlock said. He also noted that shale has gone from accounting for just 1% of domestic gas production to 25% in less than ten years. "That doesn't happen if it's a Ponzi scheme," Medlock said. (Energy Intelligence, 8/28/11)"

So what do we know about these two estimates?  According to the USGS spokesperson, their estimate is purely of "undiscovered resources" so it should be added to what reserves are already known, although we're not sure whether that means actually proven reserves or merely those reserves "guessed at."  Secondly, we know, according to Dr. Medlock of Rice University, the Marcellus is not a "Ponzi scheme," dredging up a pejorative phrase that originated with The New York Times gas shale articles in June and that would be best left out of discussions.  But Dr. Medlock does appear to credit the Marcellus with single-handedly accounting for the dramatic rise in U.S. gas shale production by pointing out how much shale gas now accounts of our gas production.  That may be a loose reading of his statement, but it highlights why people discussing topics such as this need to be more accurate in their statements because of how they may be interpreted. 

A chart presented by the Coalition in its press release, and taken from the EIA's web site, was headlined "Pennsylvania drives Northeast natural gas production growth."  In the explanation accompanying the chart, the EIA quotes Bentek Energy LLC, an energy consulting firm, saying that Pennsylvania and West Virginia production now accounts for nearly 4 Bcf per day of gas production, and that represents more than five times the average daily production of 2004-2008.  Together, Pennsylvania and West Virginia account for 85% of total Northeast production, according to Bentek.  It is interesting that the EIA doesn't have its own data to demonstrate this point, as the last year for individual state production data that includes Pennsylvania is 2009. 

Exhibit 2.  Gas Production For Northeast States

One of the questions is what does the Bentek estimate represent?  According to EIA data, we know Pennsylvania and West Virginia were producing natural gas before the Marcellus Shale emerged as a hot spot, so how much of the estimate Bentek reports is from conventional gas and how much from shale?  The EIA has shale gas production by state for 2007-2009 and shows that Pennsylvania shale produced 1 Bcf in 2007, 2 Bcf in 2008 and 76 Bcf in 2009.  Since the EIA doesn't have more recent data, we are forced to turn to the Pennsylvania Department of Environmental Protection's web site that shows the amount of gas production from the Marcellus.  The problem with the state data is that it is only reported for blocks of time, now six month intervals, but before it was 12 months.  This makes it impossible to know on a real-time basis actual production.

What we know from the state web site is that for the 12-month period of July 2009 through June 2010, average monthly production was 16.2 Bcf.  For the six-month period of July 2010 through December 2010, it averaged 45.3 Bcf per month and increased to 60.8 Bcf for the first half of 2011.  Clearly this is exemplary performance in growing Marcellus shale gas production.  But that doesn't answer the key question about the resource's potential. 

If we start with the EIA assessment of 410 Tcf of gas shale resources, we need to keep in mind that the estimate was based on data as of January 1, 2009.  Then, according to the EIA, the Marcellus Shale had proven reserves of 102 Bcf.  During 2009 proven reserves increased by 4.4 Tcf.  Unfortunately, 2010 data won't be available until the end of 2011.  The EIA's proven reserve estimate comes from its Form EIA-23, "Annual Survey of Domestic Oil and Gas Reserves," which surveys 1,200 domestic operators, both public and private companies.  This survey is more comprehensive than the data collected from companies reporting reserves to the Securities and Exchange Commission (SEC), because those are only companies with a minimum of $10 million in assets and whose securities are held by more than 500 owners.  There is also a difference in that the SEC asks only about reserves owned by a company, and not those operated.  As a result, the EIA captures more of the total reserves, especially those held by individuals and very small companies who are not operators.

Exhibit 3.  Gas Shale Production And Reserves

To better understand how INTEK prepared its estimate, we turned to the section following its table of shale resource estimates by basin entitled "Methodology."  We copied the relevant section of their methodology with emphasis added.

"The resource estimates shown in Table 1 were developed by

INTEK from publicly available company data and commercial databases for wells and acreage currently in production. The estimates of technically recoverable resources shown in Table 1 are based on the area, well spacing, and average expected ultimate recovery (EUR) for each shale play or subportion of the play. An effective recovery factor has been applied which reflects: (a) a probability factor that takes into account the results from current shale gas activity as an indicator of how much is known or unknown about the shale play; (b) a recovery factor that takes into account prior experience in how production occurs, on average, given a range of factors (including mineralogy and geologic complexity) that affect the response of the geologic play to the application of best-practice shale gas recovery technology; and (c) resources in the play that have already been produced or added into proved reserves.

"Estimates of technically recoverable shale gas resources are certain to change over time as new wells go into production and new technologies are developed. For example, the gas resource estimates in the INTEK shale report are predicated on the assumption that natural gas production rates for current wells covering only a limited portion of a play are representative of an entire play or play sub-area; however, across a single play or play sub-area there can be significant variations in depth, thickness, porosity, carbon content, pore pressure, clay content, thermal maturity, and water content. As a result, individual well production rates and recovery rates can vary by as much as a factor of 10. (Emphasis added.)

"There is considerable uncertainty regarding the ultimate size of technically recoverable shale gas and shale oil resources, including but are not limited to the following:

"• Because most shale gas and shale oil wells are only a few years old, their long-term productivity is untested. Consequently, the long-term production profiles of shale wells and their estimated ultimate recovery of oil and natural gas are uncertain. (Emphasis added.)

"• In emerging shale plays, production has been confined largely to those areas known as "sweet spots" that have the highest known production rates for the play. If the production rates for the sweet spots are used to infer the productive potential of entire plays, their productive potential probably will be overstated. The INTEK shale report mitigates this problem by differentiating the productivity of a play's sweet spot from the productivity for rest of that play." (Emphasis added.)

Based on the first highlighted text, it appears INTEK relied on estimates of the production and ultimate recovery from wells and applied that information to larger shale areas.  They did correctly warn of the risk of this method, especially in the second highlighted text.  The last paragraph (highlighted) is the more troubling one considering that the report was based on data as of 2008.  We aren't sure the idea of "sweet spots" was clearly delineated at that time.  Then the industry was embracing the blanket formation and factory development approach to shales, both of which are questionable assumptions today.  What we do know is that the USGS worked with various state geological associations and had more recent data to work with.  Now that still doesn't mean their estimate is gospel, but we need to see what additional information they publish later.

Exhibit 4.  Proven Reserve Changes For 2009

If we look at the estimates and apply what we have found out, here's what we know.  The EIA says there are 410 Tcf of shale gas in the Marcellus Shale.  If we subtract from that estimate what the USGS says is undiscovered (84 Tcf), we have 326 Tcf of reserves that supposedly are known or should be known by now.  Of course, it is possible that the 84 Tcf of USGS reserves should be added to the EIA estimate, bringing it to nearly 500 Tcf. 

Exhibit 5.  Individual State Shale Gas Reserves

If we look at the total gas discoveries (Exhibit 4) for the nation in 2009, 47.6 Tcf, and attribute 100% of them to the Marcellus, that only represents 15% of our mystery reserves.  Unfortunately that isn't the case as we see when we look at the chart in Exhibit 5, which shows that Arkansas and Louisiana added more shale gas reserves in 2009 than Pennsylvania. 

While our article in the last Musings discussing the USGS estimate and its discrepancy from the EIA number may have been among those the Marcellus Shale Coalition was targeting as spreading myths, an examination of the data suggests there is a mystery that needs to be solved.  As we said at one point in that article, "It [the USGS assessment] will force everyone to question what the USGS sees, or doesn't see, that everyone else assumes as gospel."  Unfortunately, solving this mystery will not be as "Elementary" as Sherlock Holmes would have suggested to us, assuming we are merely another Dr. Watson.  Our hope is that the gas industry and the nation finds out sooner rather than later the answer to the missing reserves as they account for approximately half the EIA's total shale gas resource estimate for the country.  Shale gas is an important energy resource and a necessary one.  Billions of dollars are being wagered on shale gas being America's energy salvation.  If there is a problem with this assumption, we need to figure it out soon in order to make a mid-course correction if needed.