Critics of our position on shale gas plays have focused on methods of decline-curve analysis, and the projections of estimated ultimate recovery (EUR) that result. The problem with this debate from all sides is that we are uncertain about how to apply decline models to newer shale plays because there is insufficient production history to satisfy all of our questions. I will, therefore, focus on some stubborn facts about Barnett Shale cumulative production and approaches to play development.
Major operators claim that their average Barnett EUR will reach 2.2-3.3 Bcf/well. Figure 1 shows that those levels of EUR are unlikely to occur in an economically meaningful timeframe based on cumulative production to date. Figure 2 shows that well performance has been erratic since operators began drilling horizontal wells, though the trend has been improving slightly in recent years. This is probably due to drilling outside of what are now known to be the core areas. The "manufacturing" paradigm that is prevalent in shale plays has led many companies to assume that all areas in the Barnett Shale and other plays are uniformly attractive.
Shale plays typically begin with a leasing frenzy whereby major players accumulate hundreds of thousands of acres, often at astronomical bonus prices. Next, a drilling campaign ensues driven more by lease expiration schedules—typically in the 3- year range—than by science. Only after considerable capital has been destroyed in this manner are the core areas recognized. This "Braille method" is completely opposite to the customary approach to E&P projects, where a cautious approach based on science is used to high-grade focus areas.
The methods used to obtain decline rates and reserve estimates for shale plays presented in this column employ best practices in the petroleum industry. Yet a group of professionals believe that some shale plays are exceptions to the methods of decline-curve analysis established by peer-reviewed papers published by the Society of Petroleum Engineers (SPE). It does not seem logical that type-curve methods should be more reliable than individual well decline-curve analysis. If the pattern of well decline is empirically exponential, it makes no sense that it should be treated as hyperbolic for conceptual reasons or because of a preference based on production from higher permeability reservoirs that are not comparable to those in the Barnett or other recent shale gas plays.
We recognize that it may take many years before true pseudo-steady-state flow is reached. But in the Barnett, decline trends are well developed in thousands of wells, and we must forecast reserves based on those trends, and not on some future, model-driven expectation of flattening decline rates. Let me be clear. We do not dispute the volume of gas resources claimed by operators. We do question the reserves that, by definition, must be commercial on a full-cycle economic basis.
The time has come for the companies that operate in the shale plays to show the data that supports their optimistic forecasts for natural gas supply in the US. The economic viability of shale gas is a serious issue with profound implications for capital investment, alternate energy research funding and national policy. To simply say that those that have doubts about shale plays are wrong will no longer satisfy the many intelligent people who follow this debate.
February 7, 2011
Doubts About the Shale Revolution
Doubts about the cost, efficiency and reserves of shale plays, writes Arthur Berman, "are shared among many petroleum industry scientists and financial analysts despite the enthusiasm for these plays by large public companies."