What's new in exploration

	Vol. 228 No. 11
ARTHUR BERMAN, CONTRIBUTING EDITOR, bermanae@gmail.com with Clinton R. Carson

Revisiting the Barnett Shale. Earlier this year, I wrote in this column that many Barnett Shale wells would be uneconomic (World Oil, April 2007). That conclusion was based on a high-level statistical evaluation of average annual decline rates for a relatively small number of horizontal wells, and a larger number of vertical wells. My colleague Clint Carson and I have since done individual decline curve analyses for nearly 2,000 horizontal wells. Our findings are more optimistic than the earlier assessment but still suggest that many wells in the trend will be marginally commercial at current natural gas prices.

A total of 1,966 horizontally drilled producing wells from the Barnett Shale were evaluated to determine commercial gas reserves using standard decline methods. Based on this analysis, approximately 28% of Barnett Shale wells should realize revenues that meet or exceed drilling, completion and operating costs in the most likely case based on assumptions incorporated into a 10% net present value (NPV₁₀) economic model.

The economic model includes per-well drilling and completion costs of $3.0 million, a wellhead gas price of $6.25/MMbtu (the average to date for 2007), 75% net revenue interest, 7.5% Texas severance tax, and $1.25/Mcf lease operating and overhead cost. Model assumptions were based on published cost and price information about the Barnett Shale and other onshore gas-producing plays in the United States. These assumptions are consistent with information published by a UBS consortium of independent gas producers that include key Barnett Shale operators Chesapeake, Devon, EOG and XTO, Fig. 1. We further believe that these assumptions are supported by Chesapeake Energy’s recent curtailing of gas production in the Barnett Shale and other producing areas because of unfavorable economics at current natural gas prices.

Rate- vs. time-decline curve analysis was used to evaluate Barnett Shale reserves. Most wells were hyperbolically declined using an economic cut-off of 2,300 Mcfg/month (76 Mcf/day) based on monthly revenue of $15,000 per well-spacing unit. This is an amount commonly used by onshore gas producers and supported by gas well abandonment rates.

The model requires per-well cumulative production of 1,520 MMcfg over 10 years to reach an economic threshold. These volumes fall within recoverable per-well reserves for the Barnett Shale stated by operators. Based on the model and estimated economic recoveries from decline curve analysis, 547 of the 1,966 wells we evaluated should meet or exceed the economic threshold, about 28% of the sample group.

It was difficult to obtain specific Barnett Shale drilling and completion costs from operators’ quarterly reports to shareholders, analyst presentations, or 10-Q filings, possibly because well costs vary considerably both by operator and by area across the trend. We used $3.0 million based on AFEs we were shown confidentially by operators. Economic sensitivities were also run using completed well costs of $2.0, $2.5 and $3.5 million, in addition to the most-likely case of $3.0 million. These additional sensitivity cases yielded threshold success rates of 46%, 36% and 22%, respectively.

Fig. 1. UBS consortium data of US onshore gas producers Anadarko, Apache, Chesapeake, Devon, EOG, Newfield, Nobel Plains and XTO.

The economic model used in this evaluation is generous, because it does not incorporate significant operating costs such as Depletion, Depreciation and Amortization (DD&A); actual wellhead prices in the Barnett play (the US average wellhead price for September, 2007 was $5.45/MMbtu); gas pipeline tariffs and other fees; or additional fracture stimulations after the initial treatment. In addition, it includes hyperbolically projected ultimate recoveries that are probably unrealistically high based on history matching.

Available data at this writing does not anticipate a near-term increase in natural gas prices that might improve Barnett Shale economics. Natural gas prices have been falling since June because of declines in fuel demand for heating and cooling in 2007, and ample gas supplies. Most analysts predict that gas prices will remain low at least through 2008, after a short, seasonal price rally anticipating the winter 2007-2008 heating season.

Some Barnett Shale operators consider production of oil or condensate to be an important component of their economics. About 43% of the wells (838/1,966) in the sample group had some liquid production. While this may be important to some operators, it was not found to greatly increase the overall number of wells in the play that meet the economic threshold.

Some believe that the Barnett Shale is an example of the triumph of technology over E&P risk. Our analysis suggests that considerable risk exists in the play. Fewer than 30% of wells are predicted to be economic in the most-likely case under current gas pricing assumptions, and using a generous economic model. Sensitivities that reduce capital costs naturally improve success rates but still result in a substantial percentage of marginal wells.

Core areas of the play are certainly profitable and the Barnett Shale is commercially attractive for many operators. Our findings are not intended to detract from the play’s success, and we expect that operators will continue to find ways to reduce cost and optimize production rates and reserves.

At the same time, the technology of horizontal drilling and fracture stimulation is costly, and Barnett producing rates are relatively low in relation to capital expenditures. Natural gas prices are lower than they were when many operators entered the play, and most service costs have increased. The hope for a play that seemingly defies technical and economic risk has tremendous appeal to the E&P industry. In the end, however, even resource plays are subject to a changing price and cost environment, the technical risks of the petroleum system, and net present value economic models. 

We gratefully acknowledge IHS for providing the production data and Petra software used for the decline analysis and other computations used in this article.