FOCUS: PRODUCTION Deepwater Gulf of Mexico more profitable than previously thought

Michael J.K. Craig, Steven T. Hyde
Spirit Energy 76
Lafayette, La.

Four factors contributing to the changed viewpoint are:

First, deepwater reservoirs have proved to have excellent productive capacity, distribution, and continuity when compared to correlative-age shelf deltaic sands. The turbidite sands characteristic of the deepwater GOM have demonstrated recoveries from a well in excess of 15 million bbl of oil and 150 bcf. Wells have produced in excess of 13,500 bo/d and 75 MMcfd, with rates of 25,000 bo/d predicted with horizontal wells and larger tubing.

These productivities far exceed those in the deltaic, discontinuous, compartmentalized reservoirs typical on the shelf. The productivity is possible because these sands can have laterally extensive 30% plus porosities and Darcy plus permeabilities.

Second, improved technologies and lower perceived risks have lowered the cost of floating production systems (FPSs). Fig. 1a [41160 bytes] illustrates cost reduction over time for a GOM FPS in 3,000 ft of water producing from a 125 million bbl of oil equivalent field through a 60,000 bo/d process train. Costs relative to a normalized Auger field development have been cut in half, and there are opportunities for further reductions.

Unocal Corp.'s estimates in the early 1990s, now, and in the near-term future, are also shown in Fig. 1a [41160 bytes]. This figure excludes the impact fabrication capacity shortages could have on future costs.

Third, projects now get on-line quicker. Fig. 1b shows approximate timelines from project approval to first production for the same generic play in 3,000 ft water. Times to first production have also been cut in half, due in part to greater confidence and learning curve advantages.

Fig. 1 reflects the effects of improved contracting strategies from conventional, to design/build partnering, to true alliance contracting.

Fourth, a collection of other important factors are:

• Reduced geologic risk and associated high success rates for deepwater GOM wells due primarily to improved seismic imaging and processing tools (3D, AVO, etc.)

• Absence of any political risk in the deepwater GOM (common overseas, and very significant in some international areas)

• Positive impact of deepwater federal royalty relief.

This article uses hypothetical reserve distributions and price forecasts to illustrate indicative economics of deepwater prospects. Economics of Shell Oil Co.'s three deepwater projects are also discussed.

Unocal's analysis

In bidding for prospect leases in the biannual deepwater Gulf of Mexico sales, Spirit Energy 76, a business unit of Unocal Corp., generates prospect economics and ranks sale prospect portfolios by investment efficiency.

The process involves:

• Interaction of a small, multidiscipline team, with critical third party input

• Identification of each geologic prospect's reserves distribution, location, and product probabilities

• Identification of optimal development scenarios and associated investment schedules for discrete points on the reserves distributions

• Generation of risked economic parameters for each prospect, based on its specific investment and revenue profiles.

Unocal's deepwater interest

In the early 1980s, Unocal was leading the industry's GOM deepwater efforts by installing in 935 ft of water the two largest single-piece jackets, Cerveza and Cerveza Liguera, in the East Breaks area. During this time, Unocal also acquired interests in 24 deepwater blocks, some which it no longer controls.

For a decade since, in the face of what appeared to be mediocre reservoirs and exorbitant facility costs, Unocal's interest in the deepwater waned, as the deepwater was being redefined by industry in increasingly greater water depths.

In the mid-1990s, Unocal's interest in the deepwater GOM resurged. Its strategy includes a 50/25/25 deepwater Area of Mutual Interest (AMI) association between Unocal, Petrobras America Inc., and Fina Oil & Chemical Co., respectively.

Unocal and its partners have actively participated in the past four deepwater lease sales, acquiring 24 blocks over 10 prospects (Fig. 2 [45910 bytes]). Unocal is aggressively pursuing other deepwater opportunities. Drilling on one prospect could begin in the fourth quarter of 1997.

In bidding for deepwater lease positions, about 50 geologic prospects have been evaluated as potentially economic. That is, for at least two reserve sizes at each prospect, an optimal development scenario with associated, all-inclusive investments, timelines, and production rates have been identified.

In the case of mixed oil/gas prospects, two sets of scenarios, investments, and spending rates have been identified, again for at least two points on each product's reserves distribution curve.

These investments and spending rates combined with geologic element risk, projected production rates, and a price forecast produced risked economic indicators for each prospect.

Different, discrete reserves sizes are picked along the reserves distribution curve because of the risking process used in the economic model. The points are chosen to simplify integration or risking of economics across the full spectrum of reserves sizes.

Only those points with economic reserve sizes will require the full-blown computation of size-specific, life-cycle investments and economics. This is typically two points for most deepwater reservoir size distributions.

Sale economics teams

Within Unocal, there are basically two teams involved in the deepwater GOM prospect evaluation phase for lease bidding: a team of geologists, geophysicists, and landmen (the G&G team), and the engineering and economics team (the E&E team) comprised primarily of a scenario planner/facilities engineer and an economist/reservoir engineer.

The planner and economist work closely together. Each must understand the other's skills because their tasks are very interdependent. They also require input from in-house drilling personnel and significant support from out-of-house contractors.

The input to the E&E team is prospect water depth, reserve distribution, probability of success, and product type (all oil, all gas, or a mix with a specified ratio). The all-important reserve distribution is assumed to be lognormal, such that it plots as a straight line on semilog or probability paper. These data are generated for each prospect by the G&G team.

The output from the E&E group is:

• Mean reserve size for leads that are a minimum for commerciality, for both oil and gas plays

• Fully risked economics on all prospects submitted by the G&G team for evaluation, inclusive of life-cycle investments, cash flow, rate of return, ultimate development cost, etc.

This output is then combined with other strategic criteria to establish bids on up-for-lease land positions over these prospects.

Sale economics process

The starting point for sale prospect E&E evaluations is to define two typical generic deepwater base cases: one for oil and another for gas.

For each base case, a water depth and reserve distribution are defined. The prospect's location is assumed to be a certain distance from the product sales point on or near the shelf edge.

Base case economic indicators include:

• Discounted, after tax net present value (DAT NPV)

• Discounted, after tax profit to investment (DAT P/I) ratios

• After tax rate of return (AT ROR)

• Undiscounted ultimate development cost (UDC).

The inputs to the economic model that result in these economic measures are then modified, on a prospect-by-prospect basis, to reflect the product characteristics, reserves distribution, water depth, anticipated per-well recoveries and flow rates, proximity to sales point, etc. These are specific to a given prospect, because each prospect is different.

Multiprospect evaluation is easier with detailed base cases that, in general, can be readily modified to reflect prospect-specific criteria (Fig. 3 [22277 bytes]).

As discussed, minimum, unrisked, mean reserves are also computed at the start of the leasing process to assist the G&G group with cutoff limits below which leads are uneconomic and not worth pursuing further. Those leads with estimated reserves greater than the cutoff limits are promoted into bona-fide prospects, or later discarded on the basis of high geologic risk. The promoted prospects are in turn evaluated fully with development plans and risked economics.

Economic risk models

Fully risked economics means that the investments and revenues are integrated across the full reserves spectrum to produce a set of risked economic indicators for the prospect. How the integration is performed and what the set of economic indicators look like is a function of the economic model.

Semideterministic economic models identify discrete points on the reserves distribution, and generate associated economics for the reserve size associated with each point on the curve. These discrete-point economics are combined by brute force using simplistic combination ratios and the prospect's probability of success (POSP1), to produce a set of semideterministic, risked economic indicators.

The risked economic indicators are discrete numbers, having no distribution.

POSP1 is the probability of finding reserves that are on the assumed distribution such that they exceed the one percentile reserves size, as opposed to the POS commercial (POSC) which is the probability of finding reserves that are commercial. POSC reserves have a magnitude that when developed should make money in excess of threshold economic limits. POSC reserves are typically less than half the POSP1 reserves.

Fig. 4a [44202 bytes] illustrates reserves size probabilities and economic risking for a hypothetical deepwater oil reserves distribution. In it, three reserve sizes are identified that are associated with the 10th (P10), 50th (P50 or median) and 90th (P90) probabilities of exceedance, respectively. Also shown is an assumed POSP1.

Economics are evaluated at these P10, P50, and P90 values, and simplistically combined using the risk combination factors.

Alternatively, instead of using the POSP1 with the full reserves distribution to generate economics, the distribution can be truncated at the economic limit and the economics evaluated, again semideterministically, using the (lower) POSC.

Fig. 4b [44202 bytes] shows the hypothetical reserves distribution of Fig. 4a truncated at 40 million bbl of oil. Economics are evaluated at the revised P10, P50, and P90 values, and simplistically combined using the same risk combination factors.

Unocal has found that the two methods (one using the full reserves spectrum and the higher POSP1, the other using a truncated distribution and POSC) give comparable results, despite the fact that the ends of the truncated distribution are not particularly lognormal (Fig. 4b). The curve is bent at the bottom due to commerciality truncation and at the top due to natural capacity limitations.

New, more-sophisticated, probabilistic economic models1 are emerging in which investments, timelines, pricing, and in fact just about any set of parameters (in addition of course to the reserves distribution) are described in full probabilistic format. In that format, each parameter has a probability distribution associated with it, and their distributions are combined rigorously through Monte Carlo simulation to produce a set of truly risked economic indicators, each with its own distribution.

These models allow for greater insight into the economic sensitivity than one number, semideterministic economic indicators. This is very beneficial for evaluating the merits of large capital projects.

The probabilistic program also requires much greater computing horsepower. Only about four to eight combinations are required per prospect with a deterministic model, but about 20,000 combinations can be required by the probabilistic model.

In the few cases that Unocal has run with both semideterministic and probabilistic models and the same full reserves distribution, the probabilistic model tends to predict lower economic indicators than the deterministic model. This is due to the majority of input distributions for the probabilistic model being skewed to the low side. For example, there is a greater likelihood of overspending by 20% than underspending by 20% the target costs for FPSs.

The same goes for most other variables. Such modeling sophistication is probably unwarranted at this phase of the process.

Risked economics

Fig. 5a [75068 bytes] illustrates a semideterministic four-point risk model or risk tree with the simple risk combination ratios for the hypothetical, single-product prospect. The four points are P90, P50, P10, and failure. Also shown are the hypothetical deepwater oil reserves and investments, which are combined with the risk ratios, the prospect's POSP1 and a flat price forecast, to produce a set of risked economic indicators.

All hypothetical prospect and estimated project economics are based on an assumption of flat, real pricing of $18/bbl and $1.80/Mcf. These are not decreased for inflation.

Fig. 5b is similar to Fig. 5a except for a mixed product case of 80% oil and 20% gas producing a seven-point risk tree. The gas size distribution is again hypothetical and used for illustration purposes only.

The unrisked investments from Fig. 5 excludes operating and abandonment costs. The economic indicators are all discounted, after tax (DAT), except for the ultimate development cost (UDC, undiscounted investments divided by reserves).

The risk combination factors of 0.3, 0.4, and 0.3 are associated with the P10, P50, and P90 probabilities of exceedance, respectively, of the prospect's reserve distribution. These values are an attempt to simplistically integrate the economics associated with the probabilities of finding reserves across the full distribution of possible sizes.

This has been shown to provide a reasonable approximation to more rigorous but impractical integration techniques. In the hypothetical case shown, the P10 case is not commercial after drilling the exploratory well, inclusive of an up-dip or down-dip sidetrack. A POSP1 of 47% is assumed.

Risked DAT NPVs and AT RORs of $100 million plus and 35+%, respectively, with UDC values almost down to $2.50/BOE, are the kinds of economic numbers for the deepwater GOM that are making people sit up and participate. This is the bottom line.

Using the flat pricing defined earlier and publicly available data, even the Auger prospect is economic, based on our estimates, despite being burdened with a huge learning curve and risk management costs (Fig. 1). Its estimated project AT ROR is about 16%, with a DAT NPV of about $130 million. Mars' ROR is estimated at a whopping 30%, generating an estimated cumulative DAT cash flow of over $2.5 billion (even at $18/bbl flat). The ROR for the Mensa gas play is estimated at 28% based on a flat $1.80/Mcf.

Table 1 [17743 bytes] summarizes the estimated project and hypothetical prospect economics, assuming a 6:1 gas/oil BOE conversion and the set of other assumptions noted previously.

Unocal's economics for a best shot development of the Auger reserves, using today's costs for a nonupgraded trussed spar supporting a 100,000 bo/d process train and 12 producing wells, along with true alliance contracting, fast-track scheduling, and assuming royalty relief, show a remarkable ROR of ~60% and a DAT NPV of over $500 million.

Scenario Planning

Investments, timelines, production volumes, rates, and price forecast are the key ingredients that result in these healthy economic indicators.

These data can either come from commercially available programs or from in-house data bases created and maintained through the help of key contractors such as those participating in new and ongoing deepwater projects. Additional data on deepwater developments can come from scouring newsletters, publications, press releases, and the Internet.

Unocal's experience is that the latter data base has proven reliable and flexible, and allows us to maintain an edge on evolving technologies. The commercial programs provide a useful foil, whose output forces us to justify our development systems and economics which are invariably somewhat different to the programs' recommendations.

As alluded to earlier, for lease sale bidding, the accuracy of the investments and the absolute values of the economic indicators are not that significant. What is more significant is identifying the best way to develop the prospect, with associated ball-park costs and timelines, and the relative ranking of each of the prospects in the lease portfolio that will be taken to the sale.

The real issue with lease sale bidding is the current competitive environment.

Generally, deepwater GOM oil plays are best developed with surface trees on a floating production platform (FPS), like a trussed spar. Production is separated and processed on-board at site and exported through dedicated oil and gas pipelines to the shelf-edge sales points.

Deepwater GOM gas plays are generally best developed with subsea completions where the multiphase production is manifolded at the site and exported through a piggable, looped, multiphase line to a shelf-edge host platform for production processing.

In the drilling area, the oil play is typically developed by exploration and appraisal wells drilled from a floater, with the remaining wells drilled from the FPS. The optimal number of FPS based-vs.-semisubmersible wells (such as the amount of prospect predrilling) is evaluated by rigorous economic comparisons, and is obviously driven by current market rig rates and FPS costs. In today's market, predrilling is costly.

The gas play is developed by wells that are all floater-based.

Investments, timelines

Major life-cycle investments are:

• Land/G&G costs

• Drilling costs for exploration, appraisal, and development wells

• Facility costs for the FPS, topsides equipment, subsea trees and manifolds, host modifications, and flow line

• Operating expenses, inclusive of special well remediation work

• Well abandonment costs.

Timelines to first production start with lease purchase, but start in earnest on successful logging of the exploratory well.

For the hypothetical oil prospect with P50 and P90 unrisked reserves of 80 and 400 million BOE, respectively, in 4,000-ft water in the GOM, typical major investments are:

• Drilling-About $21 million for the exploratory well ($24 million if it is completed) and about $6.5 million for completed development wells drilled from the FPS. Eight to 25 wells are required for the assumed hypothetical reserves range.

• Facilities-The P50 FPS with production risers supporting eight wells and a 50,000 bo/d process train costs about $215 million. A 50,000 bo/d train and all other topsides equipment (quarters, cranes, etc.) costs about $35 million, and two 35-mile flow lines to sales points cost about $43 million. The P90 costs are about 25 to 40% higher.

• Total oil-Total drilling/completion and facilities costs for the P50 oil case are roughly $115 million and $300 million, respectively. Total oil P50/P90 life-cycle costs are about $460/$860 million (Fig. 6 [41198 bytes]).

• For the hypothetical gas play with P50 and P90 unrisked reserves of about 180 and 680 bcf, respectively, in 4,000 ft of water, typical major investments are:

• Drilling-Costs are the same as for oil with about four to seven subsea wells.

• Facilities-Costs are $70-100 million for subsea completions and manifolding hardware, host platform modifications, and dual multiphase pipelines.

• Total gas-Total gas for the P50/P90 life-cycle costs are on the order of $175-300 million (Fig. 6b [41198 bytes]). Note in Fig. 6 the smaller pie for the gas development relative to the oil pie. But the gas development has a greater per well drilling costs.

Fig. 7 [41804 bytes] illustrates the approximate timelines for the P50 reserves on the above oil and gas prospects, respectively. Time from successful appraisal well logging to first production for the hypothetical oil case is about 2 years. For the gas case, it is about 11/2 years. Note, times from lease acquisition or first exploration well to first production can be substantially longer.

It is important to note that these timelines assume adequate yard capacity and vessel availability, which in these heady times of numerous large deepwater developments may not be valid, at least in the near-term. Costs are also assumed to be uninflated by a supply/demand imbalance, which may also be optimistic.

Cash flows

Fig. 8 [50228 bytes] illustrates investment and DAT cumulative cash flow profiles for the P50 and P90 oil cases, respectively. Investments per year and the cash flow stream are laid out over the full life-cycle of each hypothetical development.

Evidently, lower drilling and facilities costs, as well as higher product prices and greater production rates, profoundly change development economics.

Fig. 9 [40800 bytes] illustrates explicitly that the parameter sensitivities change the bottom-line economics.

It is clear that focusing on capital efficiency (delivering production on-line for the capex costs and in the time frame on which the prospect economics and subsequent project AFEs are based) is critical, both for the drilling and facilities. It is also clear that low-cost production, keeping operating expenses low, is not a significant factor. Opex has little influence on gross project economics.

With respect to minimum economic reserve size, for the hypothetical oil play described in 4,000 ft of water with the flat pricing, to meet a reasonable economic threshold, the minimum reserves for an economical oil development are about 40 million bbl of oil, and about 130 bcf for a gas development. This assumes royalty relief (as do all the economics described here).

Without royalty relief, oil reserves of less than about 52 million bbl of oil would remain undeveloped for the same play and pricing. To reach NPVs of zero, the reserves sizes are quite a bit smaller.

Future trends

It was alluded to earlier that costs and design-to-installation times may be subject to upward pressure due to limitations of yard space and crane barge availability. Countering this pressure are the future impacts of technological innovation, applicable across the full spectrum of deepwater technologies. Some of the innovations thought to have a major impact on improving deepwater prospect economics are:

• Less expensive FPSs such as trussed spars, tension raft jackets (TRJs), mini-TLPs (Atlantis, Moses, Seastar, buoyant leg satellites), spec-built spars, purpose-built FPSs, dynamically positioned FPSs (all current ones are moored), leased tankers, taught moorings, polyester moorings, flow lines as part of the mooring system, suction anchors, and synthetic fuel conversion (the conversion of gas into diesel or fuel oil).

• More friendly and reliable subsea oil systems such as diverless, guidelineless systems, rig-less well intervention (from a swath or workboat), wax and hydrate control by temperature, magnetics, acoustics, early warning systems, heat-generating chemicals, or roto-rooting through steel catenary, better finite element analysis (FEA) of multiphase slugging due to uneven bottom topography, subsea multiphase metering and pumping (downhole on the tree or at the base of the shelf-edge host riser, remote electric sumbmersible pumps), and subsea separation (no hydrates, no multiphase pumping).

• Less expensive pipeline installations, such as bundled, towed pipelines (5+ miles long), improved pipeline connections to floating platforms and subsea facilities, diverless repairs, noncontact support of free-spanning pipelines, valving made from composite material, strain-based, limit state pipeline design procedures, faster (cheaper) installation methods (J-lay, deeper water S-lay, reel and tow, more rapid welding flashbutt and more promisingly, forged) and NDE (nondestructive examination) techniques, better route planning (high-resolution shallow hazard surveys).

• Less expensive, more-productive wells such as extended reach wells, horizontal wells with 5-in. tubing, slimhole wells, multiple lateral wells (especially helpful in the appraisal phase), and better completions (including high-rate water packs).

It is difficult to quantify the magnitude of these new technologies. Fig. 1 suggests small percentage reductions to the future costs, sizable FPSs, but there is probably much greater potential for reduction than shown, provided yard and vessel supply/demand remains balanced.

Reducing capital investments (drilling and facilities) by 25% results in a huge 60% increase in investment efficiency (P/I), from about 1.0 to 1.6 for the hypothetical play described (Fig. 9 [40800 bytes]). A drilling facilities combined 25% reduction in capex should be possible in the future through the application of the innovations described previously, provided, again, that the capital is used efficiently.

In executing these projects, good project management principles and effective contracting strategies will be critical for achieving the healthy economics described here.

Acknowledgments

We thank John Forbes, Spirit Energy 76's trend leader of the deepwater GOM G&G team, for providing the information discussed. Thanks also to Jim Friberg and Dave Johnson (Spirit Energy 76's new vice-president of exploration) for their critical insight. Tom Miller provided excellent support in helping prepare this article.

Reference

1. Hightower, M.L., David, A., "Portfolio Modeling: A Technique for Sophisticated Oil and Gas Investors," Paper No. SPE 22016, SPE Hydrocarbon Economics and Evaluation Symposium, Dallas, 1991.

The Authors