The devil is in the details
      —German Proverb

Last week the U.S. Geological Survey (USGS) released its long-awaited reassessment of the undiscovered recoverable oil potential in the Bakken Formation of North Dakota and Montana. The USGS estimated "mean undiscovered volumes of 3.65 billion barrels of oil," which sounds like a lot. Senator Byron Dorgan (D-ND), who commissioned the study, was delighted with the result, saying "this is great news, this is 25 times the amount of the previous assessment."

Oil is now being produced in the Elm Coulee section (graph left) of the Bakken and more oil will be produced there and elsewhere as the years go by. Although one can quibble about the USGS numbers—and we will—the more important question is what will the oil flows be from the Bakken?

A proper understanding of the likely conversion rate of Bakken recoverable oil to flows (measured in barrels per day) reveals that this unconventional play will have little effect on peak oil concerns or America's perilous oil dependency. This understanding comes from an examination of the geology and production practices in the Bakken.

Bakken Geology and Oil Production

Often called the Bakken "shale," the productive sandstone lithofacies lie within the formation's middle member sandwiched between the shale source rocks above and below, as shown in the graph (left) from Julie LeFever of the North Dakota Geological Survey. Landmark's Bakken Horizontal Best Practices Review indicates that the highly pressurized productive (pay) layers have a net thickness of 6-15 feet, low porosity (8-12%), low permeability (0.05-0.5 millidarcies) and a water saturation of 15-25%. The oil produced is light, with an API gravity of 42°. The pay zones have only minor natural fracturing according to Landmark and other sources.

All of these reservoir characteristics combine to make the Bakken play unconventional, which for the non-specialist means that it is challenging to lift this oil out of the ground.

What makes the Bakken play possible? As LeFever says, "technology has finally caught up to the Bakken Formation. The ability to fracture stimulate these horizontal wells is what makes this play work." Operators must drill down vertically about 10,000 feet and then "kick out" almost as far (± 9500 feet) horizontally through the productive sandstone layers. These long-length laterals maximize wellbore contact with the reservoir. Accordingly, wells must be widely spaced (e.g in 320 acre parcels).

The reservoir must be stimulated along the lateral drill length by the injection of fracturing fluids combined with a proppant to hold the fractures open to improve the permeability—the ability of a rock's ability to transmit fluids to the borehead.

It takes great human and technological skill to steer the drill bit horizontally through the pay layer. A large number of drilling and well completion design decisions—open hole or cement liner?—must be made to maximize well productivity. Many well designs have been unsuccessful. A "lessons learned" interview with Halliburton's Tom Lantz reveals the complexities of drilling at the Bakken. Operators must study spectral log data and carry out sophisticated well completion diagnostics. Sometimes older or unsuccessful wells can be revived using this sort of data.

All of this is very expensive. An AP report North Dakota shale oil recoverable tells the story—

According to Jim Ehrets, a Denver-based geologist with Headington Oil Co. of Dallas, it costs about $5 million to drill a well tapping the middle Bakken, and companies need crude prices of at least $50 a barrel to make it economical. Even with crude prices now double that, "there still is a ton of risk," he said.

Given the expense and complexity of drilling in the Bakken, it is no wonder that there is "a ton of risk" for operators—success is not guaranteed. The risk is so large that "drillers began sharing technology about two years ago on how to recover the oil" according to the AP report.

What about the successful wells? Unfortunately, a successful well at the Bakken will produce a few hundreds, not thousands, of barrels of oil per day. Landmark's 3000 foot horizontal well simulation (graph left) reveals that in the best case, with a permeability after hydraulic fracturing of 0.66 milladarcies, the well would produce 150,000 thousand barrels (averaging 375 stock tank barrels per day) after 400 days or so (purple line). As the curve shows, cumulative production declines dramatically thereafter, leveling off at approximately 370,000 barrels over the well's lifespan of 18+ years.

Elm Coulee provides a real-world snapshot of Bakken production as it stood in 2006. Headington Oil Company has made some useful data available that helps us evaluate how things might go in the future. Here are the pertinent facts & figures—

• The cumulative oil production 6 years after discovery is 32 million barrels. Production from about 350 wells had reached 1.6 million barrels in March, 2006. This works out to 53,000 barrels per day with an average well producing about 152 barrels per day.
• About 520 square miles are under development and there were 20 drilling rigs working continuously as of November, 2006.
• Headington estimates ultimate recovery from the area will be somewhere in excess of 250 million barrels, assuming ≅ 500,000 barrels will be ultimately recovered per square mile.

Combining Headington's Elm Coulee data with the Landmark simulation, we can draw the following conclusions—

• Well productivity drops off rapidly after the first year or so of production. If other parts of the Middle Bakken are as productive as the drilled parts of Elm Coulee, and constant large investment in drilling activity in the western Williston Basin continues, we might see peak production somewhere in excess of 100,000 barrels per day. This is an educated guess, but this estimate is not off by an order of magnitude, i.e. we are talking about peak production rates in the very low hundreds of thousands, not millions, of barrels of oil per day.
• It will likely take another 5-10 years to ramp Middle Bakken production up to its peak level as described just above.

To put this in perspective, the ultra-deepwater Thunder Horse field in the Gulf of Mexico is supposed to produce at a peak level of 200,000 barrels per day in fairly short order should British Petroleum get its act together and put the field on-stream. Once Thunder Horse produces at peak rates for a while—assuming it reaches BP's target production capacity—declines thereafter are likely to be fairly steep. The middle Bakken, which likely will never produce as much as Thunder Horse at peak production, will also have a much shallower decline curve. Production will go on for decades in the Bakken, but thousands and thousands of wells will be required to extend the play over those years.

It's the size of the oil tap flow, not the size of the ultimately recoverable oil tank, that matters for peak oil calculations. But how much oil will eventually be recovered from the Bakken? Let's take a brief look at the USGS numbers.

Skepticism About the USGS Estimates

With oil near $114/barrel today, surely we can all agree that the era of cheap, abundant oil is over. The USGS has played a significant role in leading us to the present crisis in so far as their estimates have abetted complacency about future oil supply problems.

The U.S. Department of Energy uses the Geological Survey's extravagant estimates directly (e.g. the IEO 2006) to predict future discoveries and reserves growth numbers. These putative reserves predictions are then used to dismiss peak oil concerns. Over the last several years, the EIA's demand-driven model, which implicitly assumes that these huge recoverable reserves volumes will be produced as needed, has consistently served up erroneous predictions about future oil prices.

The inflated USGS estimates for the Bakken are no exception to the historical pattern, although some people are catching on that salvation does not lie just around the corner. As noted above, the USGS came up with a mean (≅ F50) estimate of 3.645 billion technically recoverable barrels of oil in the 5 assessment units (AU) of the Bakken-Lodgepole total petroleum system (TPS, graph left). Here are the tabulated estimates and some key text from the USGS fact sheet—

The assessment of the Bakken Formation indicates that most of the undiscovered oil resides within a continuous composite reservoir that is distributed across the entire area of the oil generation window (fig. 2) and includes all members of the Bakken Formation. At the time of this assessment, only a limited number of wells have produced from the Bakken continuous reservoir in the Central Basin–Poplar Dome AU, the Eastern Expulsion Threshold AU, and the Northwest Expulsion Threshold AU. Therefore, there is significant geologic uncertainty in these estimates, which is reflected in the range of estimates for oil (table 1 [graph above left]).

Few will remember the geologic uncertainty. Most will remember the large numbers. Here are some telling—damning?—observations and comments on the USGS results—

1. Operators in Elm Coulee, which forms part of the Elm Coulee-Billings Nose AU, were skeptical about the USGS estimates. The International Herald Tribune reported that "Donald Kessel, vice president of Houston-based Murex Petroleum Corp., said he believes the Geological Survey's assessment of how much oil can be recovered in the Bakken may be a little on the high side. "That's a lot of zeros," Kessel said Thursday [about the 3,645,000,000 barrels].
2. The Elm-Coulee-Billings Nose AU has the smallest mean estimate, coming in at 410 million barrels, yet that is the only part of middle Bakken play undergoing extensive development. Headington's guess was somewhere in excess of 250 million barrels for "the area", but it is not clear whether this is just for Elm Coulee, or also includes some other parts of the Survey's assessment unit.
3. The USGS indicates that its largest estimates are for the areas where, by its own admission, "only a limited number of wells have produced." This includes the Eastern Expulsion Threshold AU, for example, where the Survey estimates an astonishing 973 million undiscovered and unrecovered barrels. There's always more oil where you haven't looked according to the USGS.
4. Middle Bakken expert Tom Lantz (link op. cit.) has this to say about the geology of some of the undeveloped but supposedly oil-rich assessment units covered by the USGS—
   

   Certainly the thrust of the Middle Bakken play is moving north and east into the deeper sections of the basin in North Dakota. Since the lithology changes from higher quality dolomitic facies to a lower quality sandy/silty facies, the reservoir does not seem to be nearly as forgiving in North Dakota ... as it is in the original play area [Elm Coulee]. The ultimate extent and duration of the play will hinge on how effective we are in understanding these changes and tailoring the drilling and completion designs to exploit this lower quality reservoir. [emphasis added]

That's enough—you get the idea. It's hard to swallow the Geological Survey's story.

The Middle Bakken play is nevertheless a success story. Given the difficult geology of the oil-bearing reservoirs, sophisticated lateral drilling and well completion technology has been applied to allow the recovery of relatively large amounts of high-quality oil. Congratulations to those intrepid operators who are lifting oil out of the Middle Bakken. Despite this success, the Middle Bakken is clearly not the answer to our peak oil problems and dependency on foreign oil.

Contact the author at [the original article at the ASPO-USA website]