Laddie B. McDade could be intimidating, especially to a young explorationist like me.
He was a 1950s University of Oklahoma football player-turned-geologist and my boss at Mapco Oil and Gas in Tulsa.
I was abruptly called up to his office just days after my transfer to Tulsa from the Mapco office in Billings in the late summer of 1983. I feared the worst for my ego and career. Would I still be employed after our meeting?
To my great relief, he simply explained that my recent prospecting experience in the complex structures of the Bighorn Basin and western Montana was valuable. He then announced my immediate reassignment to explore south-central Oklahoma, an area in which I had absolutely no experience. McDade understood that the region, although already highly explored, was extraordinarily complex, both structurally and stratigraphically; it also afforded an opportunity to find multimillion-barrel fields.
Then he said, without smiling, “Dave, when you leave here today, if you do not go out and find Mapco a twenty-million-barrel field, I’m going to personally ‘kick your butt.’”
That’s a G-rated paraphrase of his actual words, but I left McDade’s office motivated by his final comment.
The Initial Work
Understanding that the search for oil begins in the library, during late 1983 and early ‘84 I developed an elementary working knowledge of southern Oklahoma with an extensive literature review. It soon became clear that structural complexity was the hallmark of the region. Unfortunately, the literature during the early ‘80s contained few detailed regional cross sections to reference. I began constructing several regional true-scale structural cross sections which generated a basic understanding of the complex structure. The complexity is illustrated by my regional structural cross section done several years later.
Armed with a basic knowledge of the structural complexity, I proceeded by reviewing Mapco’s in-house, widely spaced regional 2-D seismic database for possible leads. During the investigations in early 1984, one regional seismic line traversing southwest to northeast through section 30 of T4S-R1W in Carter County attracted special attention. It showed an unconformity between the base of the Pennsylvanian and the deeper Paleozoic section. Below the unconformity, some extremely faint anticlinal dips could be interpreted. I then constructed a sub-regional pre-Pennsylvanian subcrop map that suggested that a fold could be mapped beneath the unconformity in the southwestern quarter of T4S-R1W. Mapco named the early lead area “Walnut Creek.”
Choosing a Reservoir Target
Excellent petroleum reservoirs are plentiful in southern Oklahoma, but the vast majority of those from the Pennsylvanian down to the top Cambro-Ordovician Arbuckle had been highly exploited by the mid-1980s. Furthermore, geologists had long recognized that the Arbuckle itself, which attains a thickness approaching 8,000 feet in the southern Oklahoma aulacogen, was underexplored. I concluded that the Arbuckle, although high risk, would be a great target to pursue.
A 1968 article in the AAPG Bulletin by Sinclair geologist Jack W. Latham (Vol. 52, No. 1) caught my attention and confirmed our interest in the Arbuckle. Sinclair had established prolific oil production in 1960 from the Arbuckle at Healdton field in the Sinclair Myrtle Brown No. 1, on structural trend approximately 10 miles northwest of the Walnut Creek lead. Latham’s paper reported that the “Brown zone” at Healdton is a stratigraphic unit between 1,000 and 1,500 feet below the top of the Arbuckle in a vuggy and highly fractured section of the lowermost West Spring Creek and upper Kindblade formations of the Arbuckle Group. Cumulative production from the Brown zone at Healdton by 1984 was just more than 20 million barrels of oil.
At about this time an equally critical point of stratigraphic and reservoir control drew my attention: a subsurface test just two miles west of the Walnut Creek lead. The Conoco Denny Smith No.1, drilled in 1970, recovered 7,000 feet of saltwater and mud in a 35-minute drill stem test from the Arbuckle Brown zone. That flow rate equated to about 3,500 barrels of fluid per day. With this knowledge of the Brown zone performance in the immediate area, Mapco knew it was in an exceptionally good neighborhood to encounter excellent, high-volume Arbuckle Brown zone reservoirs in the Walnut Creek lead area.
Although there was some risk in revealing Mapco’s interest in the Brown zone, I phoned Latham, still working with Arco (Sinclair’s successor) in Dallas. This call proved to be especially beneficial, not only for the information gained, but for later drilling in the area. He reiterated the main points from his paper but emphasized the brecciated and vuggy nature of the Brown zone. The brecciation made coring almost impossible. Core barrels often jammed after no more than 10 feet of drilling and when core was retrieved, the drill floor personnel gathered pieces of it “by the bucketful,” not in one continuous core. Near the end of our phone conversation, Latham mentioned his plans to retire from Arco in about six months and that he might want to do occasional wellsite work from his retirement place near Gainesville, Texas. Several months after our conversation, he sent his resume. I put it in the bottom of my desk drawer, not knowing if it would ever be needed.
By mid-1984, Exploration Manager McDade and other managers determined that Walnut Creek was a firm lead but needed additional seismic to bring it to the prospect stage. Soon afterward, McDade left Mapco for a position with Mack Energy in Duncan, Okla. Head geophysicist Jim Lowden temporarily took over McDade’s position, and he brought in a talented consulting geophysicist, Michael Roberts, formerly of Cities Service in Tulsa, to review seismic and assist in future work on leads and prospects. Unfortunately, in the short term, changes were about to be made at Mapco Oil and Gas.
Transition in Companies, Management and Personnel
In early 1985, Mapco announced that the parent company intended to sell its oil and gas assets. We later learned the company wanted to concentrate on the more profitable refining and marketing segments of the oil business. Work on the Walnut Creek lead slowed to a halt. By midyear, rumors circulated that potential buyers for the Mapco properties included two majors. Employees widely believed that either company would strip Mapco’s producing assets and eliminate personnel. But, in late 1985 Mapco announced that effective January 1986, its producing assets were being acquired by Consolidated Natural Gas Producing Co., a subsidiary of utility Consolidated Natural Gas Co., headquartered in Pittsburgh. This was particularly good news: CNG wanted to keep all of Mapco’s Tulsa employees and told them to keep prospecting, sit tight and wait until the transition became finalized.
After the transition in early 1986, CNG upper management reviewed Mapco’s leads and prospects. The Walnut Creek lead survived the corporate shuffle intact; the company allocated funds for additional seismic and leases. The southern Oklahoma exploration team also began to come together. CNG hired Michael Lowrance to take over McDade’s vacated exploration manager position, and Tom Dimelow moved from Denver to fill the vice president of exploration job. Roberts became a full-time CNG employee. Jamie Vasquez, who joined CNG through the Mapco acquisition, became lead landman in southern Oklahoma. I continued as the southern Oklahoma team geologist. The final moves to develop the Walnut Creek lead were back on track.
From Lead to Prospect
From mid-1986 to early ‘87, CNG shot and processed, in-house, six dip-oriented and two strike seismic lines to better define the Walnut Creek lead. Roberts’ mapping confirmed the anticline I had recognized in section 30 on the regional seismic data in 1984. However, further refinement of the structural interpretation by Roberts pinpointed the crest of the Arbuckle anticline approximately 1.2 miles further southeast in the NW of section 32. Since section 32 was outside the original Walnut Creek leasing boundary, CNG gave the expanded prospect a new name, “Cottonwood Creek,” after a small stream in the immediate area.
From late 1986 into mid-1987, Jamie Vasquez purchased leases for $75 to $125 an acre with 1/8 to 1/6 royalties. By the time governmental spacing and pooling hearings were finished in late summer of 1987, CNG controlled approximately 3,200 net acres on the prospect. The drill site and four direct 80-acre offsets were leased 100 percent by CNG, as reported in the article “Field of Dreams” from the October 1989 issue of Oil and Gas Investor. In early September, the drilling authorization for expenditure was signed by management. The estimated dry hole and completion costs were $823,900 and $288,800, respectively. I then pulled Latham’s resume from my desk drawer and recommended that the company hire the retired Sinclair/Arco geologist to describe samples, interpret “shows” and recommend formation evaluation in the Arbuckle. CNG’s management agreed.
Drilling the Cottonwood Creek Wildcat
CNG spudded the Cottonwood Creek 32 No.1 on Oct. 16, 1987, in the northwest quarter of section 32, T4S-R1W. We planned a 12,500-foot well to evaluate both the Brown zone and the deeper Arbuckle Cool Creek formation potential. After setting surface casing at 1,005 feet, we drilled ahead and encountered the Pre-Penn unconformity (top Arbuckle) on Nov. 20 at about 8,000 feet, as expected. In the predawn hours of Saturday, Nov. 21, after drilling ahead approximately 300 feet, the bit encountered the Brown zone at 8,327 feet. This was about 800 feet higher than expected. The hole immediately lost circulation and the drillers lowered the bit through 24 feet of cavernous porosity to a depth of 8,351 feet where drilling stopped. Latham notified me immediately, and for a few overnight hours there were no sample or fluid returns from the Brown zone. By morning report time, both Latham and the mud loggers called to report oil on the pits and a sheen of oil over the location after a large gas bubble and oil came to the surface. The well soon began flowing oil through the surface casing to the pits and tanks. At this point, the well was marginally under control.
The situation became even more problematic three days later. While attempting to pull drill pipe, the bit became stuck at 2,712 feet, 1,717 feet below surface casing. CNG now had a well barely under control and approximately 5,600 feet of uncased Pennsylvanian section containing porous, lower-pressure sands between the stuck bit and the top of the Arbuckle unconformity at 8,000 feet. It became necessary to flow the well at a high rate and low enough flowing pressure to keep Brown zone oil from charging the shallower Pennsylvanian sands. In late November, CNG began flowing oil to the surface through ports in the stuck drill bit and surface casing at rates estimated at 4,000 barrels of oil and 3 million cubic feet of gas per day.
Since the high flow rate was far greater than the state allowable of 355 barrels of oil per day for the producing depth, the Oklahoma Corporation Commission became actively involved. The OCC issued a distressed well order on Dec. 2 allowing the well to flow at high rates until it could be brought under control. To further complicate the situation, on the evening of Dec. 3, an apparent static electric charge sparked a fire that destroyed three of seven frac tanks storing oil not far from the rig. The ensuing inferno sent flames as high as 200 feet over the location. It took 65 firefighters from Lone Grove, Ardmore and surrounding communities six hours to extinguish the blaze. CNG management soon contacted Red Adair Company to lend advice and assistance to keep the well from further control hazards and to recommend solutions.
Over the course of the next several months into early 1988, CNG drilling engineers, attorneys, Adair personnel and I were often in Oklahoma City offering testimony before the OCC. In early March, after flowing oil at high rates for more than three months, the Commission ordered CNG to curtail production to the allowable rate by April 17. On March 24, as a precautionary measure against a catastrophic blowout during the remediation operation, CNG commenced drilling a relief well to intersect the distressed well bore below the stuck drill bit and to kill the well, if necessary. On April 9, a top kill operation on the discovery well began employing a snubbing unit. CNG regained full control of the well by late May 1988 and set casing to 7,562 feet. The Arbuckle was eventually completed open-hole below casing. In the first five months of production, the discovery produced 350,000 barrels of oil.
Field Development
The first offset well, the Cottonwood Creek 32B-No.1, spudded in mid-December 1987 approximately 750 feet southeast of the discovery well. To avoid similar drilling problems, CNG set intermediate casing at 8,085 feet just below the top Arbuckle unconformity. Drilling out of casing, the well encountered no lost circulation problems in the Brown zone which was topped at 8,452 feet. Surprisingly, Latham described little dolomite in the Brown zone section, which was rather worrisome at the time because the drill bit had already penetrated a substantial Brown zone interval beyond the equivalent cavernous depth in the discovery well. However, Latham did observe some unexpected red Pennsylvanian shale deep within the Brown zone interval. Since the Pennsylvanian section was behind intermediate casing, we concluded that this material must have washed down to the Brown zone through the overlying more-than-350-feet of meteoric karsted limestone section during Pennsylvanian onlap of the anticline.
Near the base of the Brown zone interval, Latham soon described a few weak oil shows. Considering both the red Pennsylvanian shale samples which suggested possible karst in the lower Brown zone and the high oil production rate only 750 feet away, Latham and I agreed to recommend a drill stem test to management. There was some initial resistance because of the poor sample show quality and the lack of good sucrosic dolomite. Nonetheless, the DST was run and flowed oil to the surface. Later study of the caliper and the neutron density log indicated that the oil reached the well bore through about six feet of vugular and fracture porosity likely connected to the cavernous zone nearby. The Brown zone in the first confirmation borehole flowed at an initial test rate of 3,976 barrels of oil per day plus 2.8 million cubic feet of gas per day.
A description of the Cottonwood Creek field development appears in Read and Richmond (1993), Oklahoma Geological Survey Circular 95. A modified Brown zone structure map and structural cross section from that publication appear here. Integrating dip-meter and seismic data throughout the development process, by 1991 CNG completed 12 successful offsets with potential oil flow rates ranging from about 450 to 3,976 barrels of oil per day. Mack Energy also operated one successful well flowing 1,434 barrels of oil per day. Maximum field production occurred during early 1991 at around 6,000 barrels of oil per day from an area encompassing about 1,200 acres. CNG reservoir engineers estimated the Brown zone held 40 million barrels of oil in place from an oil column about 900-feet thick. Cumulative field production through August 2023 was 10.23 million barrels of oil and 15.5 billion cubic feet of gas. Field production declined to almost zero at the end of 2023.
Reservoir Ruminations and Postulations
The Brown zone reservoir throughout the field exhibits remarkable vertical and lateral heterogeneity. Two wells, the CNG 32F-No. 1 (included in the field stratigraphic section) and the discovery offset, the CNG 32B-No. 1 discussed earlier, are on the highest part of the anticline, yet both show the poorest Brown zone dolomite development. In later years, Latham and I reflected on the possibility that the Cottonwood Creek field might have easily been overlooked if the first offset location had been the first well drilled. The Brown zone would likely not have been drill stem tested because of the poor show quality and lack of good dolomite samples.
CNG’s geoscientists originally attributed the Brown zone cavernous porosity to Pennsylvanian-age meteoric karst. However, the Brown zone’s caves, heterogeneous dolomite distribution, and more importantly, the occurrence of high-temperature saddle dolomite within the Healdton cores and Cottonwood Creek samples, suggests the zone was more likely part of a hydrothermal karst system. In such a scenario, the dolomitization would have occurred during the Cambro-Ordovician, contemporaneous with Arbuckle deposition. Hydrothermal fluids likely originated in the upper mantle and moved through the Arbuckle along fault zones bounding the southern Oklahoma aulacogen. Later Pennsylvanian uplift and folding further fractured and brecciated the dolomites. Finally, the Arbuckle’s exposure at the base-Pennsylvanian unconformity undoubtedly overprinted some meteoric karst features on the reservoir.
Conclusions
The Cottonwood Creek discovery created an oil boom in the Oklahoma and north Texas oil patch. Newspapers published numerous articles as the field drilled out. Many articles claimed the new field was the largest Oklahoma oil discovery in 20 years. A national Business Week article a year after the discovery claimed it was the state’s largest in 40 years. No one at the time really knew the size for sure. There is one certainty: it was an exciting and rewarding experience to work as a CNG geoscientist exploring and drilling during one of the most severe downturns in the industry’s history.
Are there more Arbuckle fields left to be found? There was a 27-year hiatus between the Arbuckle discoveries at Healdton and Cottonwood Creek, even though many companies made a concerted effort during those years. In the 36 years since the Cottonwood Creek discovery, there have been no similar oil discoveries in the region. However, I believe future successes are still possible, especially after having the opportunity to review many square miles of 3-D seismic data along the Anadarko mountain front. The thrust structures there appear to be even more complicated than in the Cottonwood Creek area. Future successes will require careful integration of geologic cross sections and time-to-depth converted 3-D seismic data. It will also require a thorough knowledge of Arbuckle dolomite distribution and trends. A good deal of luck will also be required to find such a prolific, but fickle, Arbuckle reservoir. Mostly, a successful exploration effort will require an explorationist’s mindset, imaginative prospect generation, a team effort and a supportive company willing to take risks.
Acknowledgments
Many thanks to James P. Rogers, Carolyn Collie Read and Matt Silverman for their edits and useful suggestions that greatly improved this article. I am also indebted to consultant Gene Behrens in Oklahoma City for the updated Cottonwood Creek field production figures.