Tangguh is the liquefied natural gas production and export hub for multiple gas fields discovered during the 1990s in eastern Indonesia by Atlantic Richfield Indonesia Inc. (ARII), a subsidiary of ARCO. The name “Tangguh,” meaning “resilient” in Bahasa Indonesian, was chosen by Indonesia’s then-President Suharto in late 1997.
Tangguh turned out to be far more enduring than Suharto’s reign (he abdicated mid-1998), as 2024 marks the 30th anniversary of ARII’s drilling and testing of the discovery well, Wiriagar Deep-1. It is also the 15th year of LNG production. The WD-1 wildcat was beset with drilling crises, technical obstacles and initially, exploration disappointments. As one of ARII’s New Ventures exploration team members who was at the WD-1 wellsite, I offer this eyewitness account of the Tangguh discoveries.
Exploration Background
Our team’s focus was on Bintuni Basin, a foreland basin located in the West Papua Province (formerly named Irian Jaya), on the island of New Guinea. The island is bisected, with Papua New Guinea as the eastern portion and the western portion belonging to the Republic of Indonesia. Present-day Bintuni Bay is a semi-enclosed embayment fringed by extensive coastal swamps.
Occidental Petroleum operated the Berau production sharing contract, covering a large swath of offshore Bintuni Bay. Targeting deep oil plays, Oxy encountered gas-bearing Jurassic sandstones on three structures drilled from 1990 to 1992 (Roabiba-1, Ofaweri-1 and Wos-1). Roabiba-1 tested natural gas flowing 23.6 million standard cubic feet of gas per day from a mid-Jurassic Callovian sandstone several hundred feet thick. The gas volume at Roabiba was considered sub-commercial. Wos-1 and Ofaweri-1 also sampled gas from the Callovian sandstones via repeat formation testers, but neither well conducted drill stem tests.
Total, also targeting deep oil plays, drilled into the Permian with an onshore well in 1986 without encountering any oil, nor any Jurassic sandstones, and Sebyar-1 was deemed a dry hole. They relinquished the permit, which was subsequently acquired by British Gas and incorporated into its larger onshore and offshore Muturi PSC, in eastern Bintuni.
From 1977 to 1992 Conoco operated the onshore Kepala Burung Selatan A PSC west of BG’s Muturi PSC. Targeting deep oil, Conoco drilled Tarof-2 and Ayot-2, but both were dry holes. However, Conoco’s shallower exploration nearer the coast was more successful with the Wiriagar oilfield discovery (1981). This shallow onshore field produced small volumes of oil from 1989 through the 1990s from the Miocene Kais Limestone.
ARII farmed-in to the KBSA PSC joint venture in 1989, with ARII Vice President Gene Richards and New Ventures Manager Dick Gerard urging the JV to drill through the shallow Wiriagar oilfield to explore deeper targets. The JV balked at the estimated cost ($20 million in U.S. dollars per well), and the permit was relinquished in 1992, with Indonesia’s state-owned oil and gas company Pertamina operating the Wiriagar shallow oilfield production. ARCO and JV-partner Kanematsu were awarded the relatively small onshore ARII Wiriagar PSC in 1993 by Pertamina, with deep oil-plays as the exploration focus.
The Team’s Challenges
Suherman Tisnawidjaja replaced Gerard in 1992 as ARII new ventures manager while keeping in place the previous technical team of Larry Casarta as staff geologist and Sonny Sampurno as geologist. At VP Larry Asbury’s suggestion, Suherman and Casarta seconded me in 1993 as special exploration projects/senior wellsite operations consultant from ARII’s Offshore NW Java Sea Department. Geological specialist/reservoir engineer John Marcou was also utilized beginning mid-1994. This is the story of a great team.
The challenge facing PSC operators in West Papua was a pronounced lack of useable seismic data. The existing Bintuni Basin seismic in 1993 was limited to 2-D surveys with extremely poor resolution, the result of seismic energy scattering by vuggy and fractured shallow carbonates.
The top of the carbonates (Kais Ls) was the only definitive basin-wide reflector on seismic available to ARII. Casarta’s reprocessing of vintage KBSA 2-D seismic resulted in identification of an additional weak reflector correlative with the top Late Permian on a few regional wells. Unfortunately, the intervening Paleogene and Mesozoic reflectors were indistinguishable.
Casarta also interpreted the Late Permian reflector having rollover-closure below the Wiriagar PSC, indicating an anticlinorium steeply dipping southeast from Wiriagar toward Oxy’s Roabiba-1 offshore well. Importantly, there was also a significant thickening of sedimentary strata between the top Miocene and top Late Permian reflectors between the Wiriagar PSC and Roabiba-1. Therefore, a Callovian sandstone stratigraphic onlap and/or erosional unconformity should exist north of the Roabiba-1 well, given the absence of Jurassic sandstones at KBSA’s onshore Ayot-2 and Tarof-2 wells. This realization, combined with Sampurno’s regional well mapping and a 1992 paper published by Perkins and Livesey, suggested the Jurassic sandstone pinch-out or unconformity was near the present-day north shore of Berau Bay.
Late-maturation natural gas encountered at Roabiba-1 might have displaced early maturity oil, with the oil subsequently migrating updip toward Wiriagar Deep’s stratigraphic/structural crest. If the thinning Jurassic onlap, or unconformity, occurred below the Wiriagar shallow oilfield, then perhaps the Callovian sandstone at the structural crest was oil-charged with fractures providing conduits to the overlying shallow Wiriagar oilfield. Vexed by the lack of seismic reflectors, Casarta “creatively phantomed” lithostratigraphic top-depths between the Miocene and Late Permian from the Wiriagar PSC to Oxy’s Roabiba-1 well.
Drilling Blind
Arriving at WD-1 the first week of February 1994, I found myself in one of the world’s largest malaria-ridden, mosquito-plagued mangrove swamps, infested with saltwater crocodiles and snakes.
Unsurprisingly, the province is sparsely inhabited.
ARII utilized the Sebina-4, a drilling rig capable of being helicoptered into a compact area and light enough to be supported on wood logs and coconut mats in the Wiriagar Swamp to minimize the drill-site footprint and environmental impact. The derrick was connected to various work skid-units, the offices and the accommodation huts by a series of wooden catwalks elevated above the mire. Spudded Feb. 6, WD-1 drilled through the shallow Wiriagar oilfield with minor H2S. After the production interval was cased off, drilling of the12-and-a-quarter-inch hole commenced in the massively bedded Faumai limestone formation, with a partial loss of returns.
ARII had anticipated minor lost circulation and some minor H2S, as both problems had been encountered while drilling KBSA’s wells. Therefore, ARII had ample lost circulation material at the wellsite and had contracted Sperry-Sun H2S specialists with klaxon alarms and flashing red lights to alert drilling crews to H2S.
At a depth of 2,460 feet we completely lost returns. After days of pumping LCM downhole, the crew spotted “gunk” pills (a slurry of diesel, bentonite and cement) on bottom. But it was to no avail, so “blind drilling” began. ARII’s drilling manager Brett Crawford, drilling engineer Richard Leturno and Larry Casarta had a backup plan of “drilling blind” in case of incurable lost circulation. When drilling blind, swamp water, not drilling mud, would be pumped down the drill string with no fluid returns to surface. This would continue until the drill bit was below the loss zones, where casing could be set.
Gas Kick
I went to the Geoservices mudlogging unit at 6 p.m., telling the logging crew to have dinner and I would watch their unit. It was quiet in the usually bustling logging unit with no drilling fluid returns, no drill cuttings to examine and no gas readings to record. However, when the drill bit reached 2,559 feet, the first alarm went off inside the logging unit indicating return flow in the flowline, followed by the pit volume alarm as mud pits quickly gained volume.
There was no answer when I called the driller on the rig phone. Drill floors are noisy, so crews often don’t hear the phone ringing. I tried calling the ARII company man. No answer. Ditto for the rig superintendent. Where was everyone? At dinner, of course.
Then the mudlogging gas detector alarm went off, steadily rising to 320 units total gas. We were taking a gas kick when we weren’t supposed to have any returns. I called the galley for the company man and rig superintendent as I looked at the mudlogging unit’s H2S detector. Geoservices wasn’t contracted for H2S monitoring, but I had asked them to install it prior to drilling. I saw the “needle-gauge” at 50-parts per million, as I finally got through to the company man in the galley.
I explained we were taking a gas kick and now getting deadly sour gas. The company man was skeptical because the dedicated Sperry-Sun H2S alarms hadn’t sounded, and although the driller wasn’t answering the phone he could see roughnecks still working on the rig floor.
I rigged up my Scott air-pack to go to the rig floor, as the company man called from his office asking what the instrument readings were. I told him total gas was 400 units and H2S was 337 ppm. Before he could answer, Sperry-Sun’s klaxon began blaring with red-lights flashing.
I left the unit masked-up, watching in amazement as gas-water plumes shot 10 feet above the possum-belly and roughnecks scrambled down rig-floor steps without masks, some jumping into the swamp from the elevated catwalks. Then I noticed the kelly turning on automatic driller. The well hadn’t been shut in. As I approached the office/accommodation block using the plank shortcut, the company man and rig superintendent came out wearing 30-minute air tanks.
“Why didn’t the alarms go off before?” I asked.
“The engineers were both in the galley. Said they’d a blown fuse they didn’t know about. Grab a 30-minute air-tank so you can help shut-in the well,” said the company man.
What a way to start a wildcat. It took six days to get the well under control before drilling resumed with a rotating head, which allows blind drilling that stops fluids or gas from coming to the surface. With the rotating-head installed, we resumed blind drilling. But with no cuttings to surface and only phantomed formation-top depths, the question of where to stop drilling was the new conundrum. That hadn’t been included in the drilling prognosis, of course.
I suggested running the casing collar locator’s gamma-ray tool on wireline inside drill pipe and using an exaggerated 0-25 scale (rather than the usual gamma-ray log 0-150 scaling) to identify the Eocene and Palaeocene shales underlying the Faumai. The CCL/gamma-ray slim tools are used to locate collars inside casing strings and aren’t reliable open-hole logging tools. But in this case, it worked and, after blindly drilling an additional 2,000 feet, we found the base/Faumai and set casing at 4,500 feet.
Drilling a new eight-and-a-half-inch hole was uneventful as we approached the prognosed/phantomed Jurassic target depth of 6,540 feet. At 6,530 feet, the background gas shot up to 4500 units. The 9.4 pounds-per-gallon drilling mud-weight was gas-cut, and the well began flowing uncontrollably. We were taking a gas kick, requiring shutting-in the BOP.
On other Bintuni regional wells, the only pervasive lithological marker in drill-cuttings was a greenish Mesozoic volcanic layer. Conoco’s onshore wells and Oxy’s offshore wells all described blackish-green olivine-rich volcanics at the Base Cretaceous. WD-1 was flowing from an overpressured, thinly bedded, gas-charged sandstone at 6,530 feet. The bit was within 10 feet of the phantomed Jurassic prognosed depth, but there hadn’t been any green volcanics in the cuttings. Was this the Jurassic’s Callovian sandstone?
Taming the Wildcat
It took two weeks to shut-in the well, spot “kill-pills” of heavy-weighted barite drilling mud on bottom and hang an emergency seven-inch liner to case-off the kick zone. We began drilling the six-and-one-eighth-inch slim-hole with a 10.6 pounds-per-gallon mud-weight-in that was continually gas-cut to 9.6 ppg mud-weight-out. There were discussions in ARII/ARCO about pulling the plug on further drilling. But ARCO personnel in Texas sent instructions, relayed to wellsite by radio, to obtain whole core for shale analyses and mercury-injection capillary-pressure seal evaluation.
Coring proceeded smoothly from 6,786 to 6,811 feet, when another 2500-units gas-kick occurred. The zone was producing gas into the borehole, requiring an increased 11.6 ppg mud-weight-in. Recovering the core, I observed we had cored an active fault. The core captured a high-angle, slickensided fault acting as a conduit for overpressured gas from deeper strata. Coring the fault was serendipity, but it later provided an important piece of geological evidence in the WD-1 wildcat puzzle. After the well was brought under control, a second core was cut, capturing a 20-foot sandstone at depth 6,830 feet.
Mud-weight-in was increased to 12.4 ppg as drilling resumed with more thinly bedded sandstone stringers encountered. I recommended cutting a third core in overpressured sandstones at 7,165 feet. The core contained a gas-charged 20-foot dolomitic sandstone. But ARII Jakarta was disappointed. No oil, only gas. If this was our Jurassic target, then this wildcat and the oil play were a bust.
ARII sent word to drill to the maximum approved depth of 8,000 feet if gas levels and hole conditions allowed, run wireline logs to total depth, then plug and abandon the well. We eventually penetrated a massively bedded carbonate, with significant gas shows at 7,600 feet. The team now suspected that the Jurassic interval’s onlap/erosional truncation might be south of WD-1.
However, Casarta reported that new laboratory analyses in Jakarta had identified only Palaeocene spores/pollen in the three cores, with no Jurassic palynomorphs. Then greenish-black nodules appeared in drill cuttings from 7,740 to 7,750 feet. It wasn’t greenish volcanics, but green sandstone composed of 50-75-percent glauconite and 50-25-percent quartz grains with dolomitic cement. I thought, “this glauconitic sandstone is the base/Cretaceous. We haven’t penetrated the Jurassic yet.”
We circulated out as I radioed Jakarta to recommend coring the next sandstone. But the disappointing reply came. “No, we’re close to the AFE’s approved maximum depth. Just continue drilling to 8,000 feet and run TD wireline logs.”
We drilled sandstone yielding a 2100-unit gas peak at 7,880 feet. The mud was gas-cut again, requiring increased mud weight to 12.7 ppg. We had encountered our Jurassic sandstone target, but it was only 16-feet thick. Finally, authorization came to extend TD to 8,500 feet, but we’d already penetrated Permian interbedded coals and sandstones.
WD-1 reached TD at 8,500 feet on May 2, 1994, and wireline logging commenced. Afterward, I returned to Jakarta and congratulated the team on the discovery. The office was downcast. “It’s a disaster!” they said.
The wildcat had been an oil play but there had been no oil, only gas shows. The deep oil play concept was a bust, and petrophysics indicated no moveable hydrocarbons in the Permian and Jurassic intervals. There was little appetite for testing possibly discontinuous, thinly-bed Palaeocene sandstones at an estimated cost of $1 million per DST.
I presented to the team a pressure profile of the wildcat well, which suggested overpressured gas reservoirs in the Paleogene and Mesozoic intervals. Reservoir overpressures and borehole characteristics seemed due to gas-buoyancy, and from estimated pressures I extrapolated a vertical natural-gas column height of 2,200 feet within the Jurassic reservoir.
Ultimately, the team approved DSTs of five intervals: Late Permian sandstone, Jurassic sandstone, Cretaceous limestone, and the three best Palaeocene sandstones with testing and sampling. Testing was scheduled for late May; however, it wasn’t long before the WD-1 well had big problems, again. While preparing for testing, the well took a gas kick requiring closing the BOP. The seal was lost while stripping out, and Marcou called ARII Jakarta from the wellsite by radiotelephone, “There’s gas discharging from the well with a cloud of mist reaching the crown, everyone is running from the drill-floor. I have to go.”
The casing string had collapsed downhole. High formation pressures, combined with H2S metal fatigue and excessive drill-string wear against a casing dogleg had caused another downhole disaster. Wild Well Control Inc. brought in a snubbing unit to control the well, safely patch the casing, and then run new casing prior to testing.
It was late July when testing designed and supervised by Marcou began. Testing conditions were far from ideal, he told us. All the test intervals were perforated overbalanced and tubing restricted (two-and-seven-eighths-inch tubing), yet the Jurassic, Cretaceous and three Palaeocene intervals provided decent flow rates. By late August well testing was completed with the following results:
- The Late Permian had no flow.
- The Mid-Jurassic sandstone flowed 3.9 MMSCFGPD.
- The Cretaceous carbonate flowed 9.8 million.
- The deepest Early Palaeocene sandstone flowed 0.9 million.
- The shallowest Early Palaeocene sandstone flowed 11.5 million.
- The Late Palaeocene sandstone flowed 3.8 million.
Implications
The final well cost had ballooned to $20 million during seven months of operations, the result of numerous downhole problems and challenges, far exceeding budgeted time and cost. Testing had been successful at 30 million per day from five combined zones, but there hadn’t been any oil, only lean natural gas with only four-to-six barrels of condensate per million cubic feet. Management considered relinquishing the permit.
Feeling that gas-buoyancy best fit the overpressure data, I calculated a potential 2,200-foot gas-column height. That meant the gas reservoir extended far down-dip on the structure, implying the anticline was filled from the Wiriagar axial crest down to the structure’s spill-point with Oxy’s offshore Roabiba structure. Wiriagar Deep’s Jurassic gas field might be more than 20 kilometers, or 12 miles long.
To supplement the WD-1 overpressure analysis, I plotted the temperature gradient of the borehole with depth, showing a decreased temperature in the overpressured transition caprock: a classic overpressure gas effect. In support of the giant gas field size, I generated regional north-south and east-west geothermal gradient cross-sections extrapolated from shallow, intermediate and total depth wireline-logging temperatures on all existing Bintuni Basin wells. The cross-sections suggested a basin-wide temperature profile closely resembling the unique geothermal profiles noted at other giant gas fields.
Marcou did an independent evaluation of WD-1 reservoir pressures from DSTs as well and concluded a gas-column height in excess of 2,000 feet was reasonable. Furthermore, he demonstrated that WD-1’s gas reservoirs fell on a common gas-gradient, indicating they were probably in communication. The geological importance of the cored Palaeocene fault (also lying on the common gradient) acting as a gas-charging conduit was now apparent.
Exploration teams routinely consider hydrocarbon sourcing, generation, migration, reservoir, seal, trap, and timing, but rarely are borehole pressure and temperature data utilized as exploration tools. WD-1’s P-T data were instrumental in the evolution from a small, deep oil play to a much larger deep gas play.
The team agreed that WD-1 was a significant gas discovery, but the challenge was to convince ARCO management to farm-in to Oxy’s Berau PSC for the appraisal well (where 80 percent of the reserves were); and to keep WD-1’s evaluation tight-holed, lest Oxy reap the rewards of ARII’s discovery. Indonesian well data is turned over to the government, but Suherman and Sampurno assured the team that WD-1 interpretations could be kept confidential while still complying with regulations.
It took the expertise and experience of the entire five-member new ventures team to make the wildcat exploration a success. But WD-1 was only the start of the Tangguh story because none of us at the time realized that the appraisal well would lead to additional gas reservoirs and major new discoveries. That saga is for Part II of this tale next month.
The author thanks John Marcou for editing and permission to publish personal photos. Special thanks also to Larry Casarta, Suherman Tisnawidjaja, Sonny Sampurno and Matt Silverman for comments and suggestions.