TECHNOLOGY U.S. horizontal wells show varied production performance

Gary S. Swindell
Consulting Engineer
Dallas

Production performance analysis indicates that the Austin chalk trend is the only area in the U.S. with a significant amount of successful horizontal wells.

Results in other areas have been highly variable and even within the sizable Austin chalk trend, some wells have not found sufficient reserves to always make horizontal drilling profitable.

Rapid return of investment because of high initial production rates, ability to implement economies of scale, and the realization of statistical averages with multiple well programs are the driving force behind horizontal drilling in the Austin chalk trend.

Horizontal drilling

Horizontal drilling has become one of the most valuable technologies introduced to the upstream oil business. Along with other advances over the last 15 years, such as massive fracturing and 3D seismic, horizontal drilling has significantly impacted oil and gas production.

To date, the two top states, North Dakota and Texas, have produced over 285 million BOE (barrels of oil equivalent) through horizontal wells, a significant contribution to the U.S. oil supply.

A considerable number of articles and papers have discussed horizontal drilling, but the focus has been on the hardware and drilling side of the technology. Relatively little has been published on the reserves found.1

The lack of reserves and economic information is partly due to operators wanting to retain a competitive advantage. But it also is due to the difficulty in applying traditional reserves estimation methods to a situation that is not fully understood.

Volumetric analysis is complicated because of a lack of knowledge on the extent of the fracture systems. There are documented cases of horizontal wells 2 miles apart showing interference, and other cases where wells 1,500 ft apart show no interference.

Also the poorly understood reservoir character makes pressure transient analysis and reservoir simulation difficult. Analogies also have not been effective because there are too few examples.

Performance analysis is the remaining tool for estimating reserves. But only recently has there been enough history available to reliably examine the potential of U.S. horizontal wells.

At least 12 states have some horizontal production, with Texas having most of it. Together with North Dakota, these two states account for 89% of the producing horizontal leases drilled since 1987 (Fig. 1 [107799 bytes]).

Evaluation complications

One difficulty in evaluating horizontal well success is that production for Texas wells is reported on a lease basis and not by individual well. Dwight's EnergyData Inc. designates about 2,700 Texas leases as "horizontal." These include older leases that have had horizontal wells drilled around existing vertical production.

To reduce the effect of mixed vertical and horizontal drilling, the following analysis includes only leases that began producing after 1987.

In Texas, the 1987 limit excludes about 700 mixed leases from the statistics. Also in Texas, the leases included are mostly single well leases. Other states generally report production on an individual well basis, and the differentiation is less significant.

Another problem with summarized rate-vs.-time curves for reserves estimation is that the completions occurred over a number of years. To account for these effects, this analysis includes normalized production.

North Dakota

Horizontal wells in North Dakota have primarily targeted the Bakken shale. Billings and McKenzie counties account for most of the horizontal wells. Through mid-1995, horizontal well production totaled 11.3 million bbl and 18 bcf, or about 14 million BOE. The first wells were completed in late 1987.

In the Bakken shale (Fig. 2a [63344 bytes]) production began to fall in 1991, even as the total well count continued to rise. This is due to the sharp production declines often seen in the early life of fractured reservoirs.

Normalizing, the procedure of setting each well back to the same "time zero," provides a means for taking out the problem of time shown in Fig. 2a [63344 bytes]. When all the horizontal wells that make up the total horizontal production for North Dakota are normalized, a composite average well profile develops as shown in Fig. 2b [63344 bytes].

The average initial production from the typical horizontal well in North Dakota is 4,300 bbl/month with an early decline of about 65% and only a small hyperbolic factor that serves to flatten the decline over time. The relatively small hyperbolic nature of the curve compared with the Austin chalk wells may suggest poor contribution from the rock matrix compared to the fracture system.

An earlier study looked at every individual Bakken shale well in the state and projected an ultimate recovery using rate-vs.-time curves. The average vertical well was projected to ultimately recover 104,000 bbl.

Surprisingly, the average horizontal well's estimated ultimate recovery (EUR) was less than the vertical wells at 97,000 bbl. Perhaps the horizontal wells found fracture systems that had already been substantially depleted by vertical wells.

Of the larger fields, only Elkhorn showed higher per well EUR for horizontal wells. The advantage was only 15,000 bbl/well for horizontal drilling.

Poor overall economics have slowed the use of horizontal drilling in the Bakken shale of North Dakota.

Texas

More than 101 Texas counties have had some horizontal production (Fig. 3 [188214 bytes]). Statewide leases designated as "horizontal" (Dwight's EnergyData) with the earliest production after 1987, have produced 188 million bbl and 497 bcf (271 million BOE).

As in North Dakota, the total oil production began to decline after late 1992 even as well count continued to increase (Fig. 4 [43103 bytes]). When total horizontal well count generally stabilized in early 1994, the overall decline becomes considerably steeper.

Although scattered horizontal drilling has taken place in most Texas basins, the vast majority of the wells have been in the Austin chalk trend. The top ten Texas counties in horizontal drilling since 1987 contain more than 90% of the total horizontal leases (Fig. 5 [118814 bytes]).

Normalizing again provides a tool to remove much of the effect of differing dates of first production. Some leases in our study group did have additional wells added after the earliest production. Furthermore, some had wells that added a second horizontal leg, often in the updip direction.

Overall, the effect of added wells and reworks of existing wells does not appear to have a significant influence on the statistics developed in this study.

Austin chalk EUR

The average estimated ultimate recovery of horizontal wells, based on the normalized profiles of each of the top ten Texas counties, is shown in Fig. 5 [118814 bytes]. Average EUR ranges from a low of 80,000 BOE per well in LaSalle County to 200,000 BOE in Fayette County.

The average peak production can be obtained from the normalized rate-vs.-time curves. Usually this peak occurs in the first month of production but at times the second or third month is the highest level as wells clean up and produce back drilling and stimulation fluids.

The data are influenced by the fact that the first month of recorded production may not be a full 30 or 31 days.

Closer look at four counties

The rate-vs.-time profiles of four counties, Brazos and Fayette in the northeastern part of the chalk trend, and Frio and Zavala in the southwest, give a better understanding of the average reserves (Fig. 6 [160784 bytes]).

Brazos County shows fairly high peak production, 14,250 bbl/month, but very rapid decline. Over 68% of the average EUR (155,000 BOE) is produced in the first year, and over 80% in the first 2 years.

Fayette County shows similarly high initial rates and higher EUR than Brazos because of more flattening of the decline. Still, over 65% of the EUR is produced in the first year. The Giddings field is located in Fayette County.

Pearsall field is on the southwestern end of the chalk trend, covering portions of Frio, LaSalle, Dimmit, and Zavala counties.

The Frio County rate-vs.-time plot shows a significantly lower initial rate than the northeastern counties, but perhaps a stronger hyperbolic behavior. The first year accounts for 55% of the ultimate predicted recovery.

The Zavala County typical profile is a composite of 93 horizontal wells. Initial rate is fairly high at 12,400 bbl/month. Estimated ultimate recovery is nearly 200,000 BOE/well.

Hyperbolic behavior is not very evident after the first 2 years, possibly suggesting that the primary porosity is less a contributing factor than curves with more flattening. Other studies have concluded that both fracture size and fracture density are significantly lower in the Pearsall area-vs.-Giddings field.

Reference

1. Maloy, William T., Horizontal wells up odds for profit in Giddings Austin Chalk, OGJ, Feb. 17, 1992, p. 67.

Bibliography

Maloy, William T., Statistical relations predict horizontal well production, World Oil, April 1993, p. 55.

Beliveau, Dennis, Heterogeneity, Geostatistics, Horizontal Wells, and Blackjack Poker, Journal of Petroleum Technology, December 1995, p. 1068.

Thomas, Gilbert E., and Sonnenberg, Frank P., "Homing in on sweet spots in Cretaceous Austin Chalk," OGJ, Nov. 29, 1993, p 110.

The Author