Economic models verify heavy oil profitability

These two steam injection wells and producing well are in California's Midway-Sunset field, which produces about 166,000 b/d of heavy (11-14° API) oil.

Major oil companies, as well as small independents, are now aggressively developing heavy-oil projects.

Until recently, the conventional perception of heavy oil (20° API gravity and lower) was that heavy oil production is uneconomic unless oil prices are in the $30/bbl range.

But heavy oil finding and development costs/bbl are significantly less than for conventional light oil. Also, operating costs now compare favorably with many secondary and primary recovery projects.

During 1997, a flurry of mergers and take-overs of public companies has involved both Canadian and U.S. heavy-oil producers. One estimate is that over the next 5 years, more than $20 billion will be invested in heavy oil and bitumen sand projects in Canada, Venezuela, and California.

Investment in heavy oil has been stimulated by new technologies that have significantly reduced costs and improved recovery. Also, heavy oil demand has increased because of the expanded refinery capacity for processing low-gravity, high-sulfur crudes.

The authors visited major and independent oil producers in the San Joaquin basin in California to gather data for modeling the economics for developing a hypothetical heavy oil field.

California

Most heavy oil is produced by thermal stimulation. In California, production is predominantly from large shallow reservoirs, generally in thick sand/shale sections (several hundred feet of net oil sand) with high porosities, permeabilities, and oil saturations.

Thermal recovery is most suitable for reservoirs with good continuous vertical permeability over substantial thicknesses that contain low-gravity oils (less than 16° API) having high viscosities at reservoir temperature.

The largest fields and current activity are in Kern County in the southern San Joaquin Valley, near Bakersfield.

Most of these large heavy oil fields were discovered around the turn of the century, developed on close well spacing, and produced on primary production for over 50 years.

Although much oil has been recovered, recovery efficiencies, as a percentage of original oil in place (OOIP), were low because of high oil viscosities and low pressures in the shallow reservoirs.

The largest fields with their cumulative oil production, start date, and current oil production, include Midway-Sunset (2.36 billion bbl, since 1894, 166,000 b/d), Kern River (1.57 billion bbl, since 1899, 135,000 b/d), and South Belridge (1.06 billion bbl, since 1911, 112,000 b/d).

Spurred on by the 1970s oil price shocks, operators refined thermal recovery technology and expanded its application. From 1970 to 1987, heavy oil production in California (16° API and less) tripled from about 200,000 to 600,000 bo/d. By the late 1980s, however, heavy oil prices plummeted and producing it became economically marginal at best.

But since the late 1980s, economics have been improved substantially by a combination of factors:

• Higher and more stable heavy-oil prices, including a narrower price differential between California heavy and West Texas Intermediate crudes.
• Lower fuel gas prices and the building of cogeneration plants for steam supply to the major fields. Steam cost is between one-third to one-half of the operating expenses. Fuel is by far the largest cost component, therefore, changes in gas prices and improvements in thermal efficiency in steam generation significantly impact operating expenses.
• More-refined operating techniques and technology that improved oil recovery efficiencies and producing rates, and lowered capital and operating costs.

Steam techniques

Steam injection proved to be the most effective. Steam heats the oil-bearing reservoirs by either cyclic stimulation or steam flooding.

In cyclic steam stimulation, wells are subjected to alternating periods of steam injection followed by production. Production lasts until the reservoir cools to the point that production reaches an economic limit. At that point, steam again is injected.

In steam flooding, steam is injected continuously into dedicated injection wells. The steam heats the reservoir and mobilizes oil produced from surrounding wells.

The oil recovery mechanism with either method is essentially gravity drainage. Mobilized heated oil flows gradually either downward in reservoirs where vertical permeability exists, or down-dip in steeply dipping reservoirs. This flow creates a desaturated zone, higher up in the reservoir.

In a steam flood, injected steam enters and expands that zone, thereby heating a larger volume of reservoir.

Because vertical flow dominates oil recovery with either technique, close well spacing is required for high recovery efficiencies.

Cyclic steam stimulation will generally have greater initial producing rates and will yield a lower steam injected-to-oil recovered ratio (SOR). However, producing rates will decline as the reservoir immediately around a well depletes with successive cycles.

Steam flooding, on the other hand, will heat a larger reservoir volume, and therefore provide a higher ultimate recovery. In practice, operators often begin with cyclic stimulation and then convert to a steam flood, if that is economic.

Operations

For the major fields, operators expect ultimate recoveries of 70% of OOIP, and even 80% in some cases, given favorable economics. Currently the fields have been depleted to about two-thirds of those levels. Therefore, the major projects should last for another 20 years or more.

In most existing steam floods, thick reservoirs are divided into several vertical horizons within which vertical communication exists. These horizons normally are flooded sequentially.

Some operators, however, are moving toward parallel steam floods in which two or more horizons are flooded coincidentally. This accelerates recovery, shortens field life, and lowers operating costs.

A number of evolving technologies have enabled operators to improve operating efficiency and ultimate oil recoveries. Some technologies include:

• Computer simulation for analyzing and predicting reservoir performance
• Combined-cycle cogeneration plants to obtain higher thermal efficiency in steam generation
• Control and measurement of two-phase steam flow in distribution piping
• More efficient separation and treatment of produced oil and water.

• Reducing rig time to drill and complete new wells. In some areas less than 48 hr are needed.
• Improving steam and heat management by varying the time and steam volume during each steam cycle.
• Selling excess produced water to agriculture instead of paying to dispose of it.
• Manufacturing marketable quality fuel oil at the lease site that can be used as steam boiler fuel.

Operators are experimenting with horizontal wells and related recovery processes such as steam-assisted gravity drainage (SAGD) to improve oil producing rates and ultimate recovery. However, drilling costs for vertical wells have been reduced so that it is difficult to affect further savings with horizontal wells.

Horizontal wells do produce at higher rates, thereby increasing payout and the project present worth. But ultimate oil recovery will likely be about the same as with vertical wells.

Kern River experience

The best wells produce at rates greater than 100 bo/d. The field's mean is 18 bo/d, and median is 7 bo/d. The economic limit is about 3 bo/d.

Produced crude has a 13° API gravity and contains 1% sulfur.

Discovered in 1899, production peaked in 1981 at 141,000 bo/d. Texaco Exploration & Production Inc. started steam injection in 1964. This original peak will likely be surpassed in 1997 as a result of recent redevelopment.

About 50% of OOIP has been recovered. Expected recoveries with new technologies may reach 80% and possibly even 90% of OOIP. In 1996, 7,874 wells produced oil in the field.

Texaco, is the largest operator in Kern River with 4,800 producing wells, 1,400 steam injection wells, and 350 temperature-observation wells. These numbers do not include Texaco's recent Monterey Resources acquisition.

Each year, Texaco drills about 50-100 new wells in the field.

In 1995, Texaco produced 85,948 bo/d and production has increased to over 90,000 bo/d as a result of workovers and better reservoir management. By 1998, production may peak at 100,000 bo/d.

With a 66% recovery of OOIP, Kern River will be depleted by 2030. With higher recoveries, field life will be extended.

Texaco has developed Kern River on 2.5-acre, inverted five-spot and 2.25-acre, nine-spot well patterns. Each development project involves a minimum 20 patterns.

Approximate wells cost are $65,000 for a typical 1,000-ft producing well, $50,000 for an injection well, and $30,000 for a temperature-observation well. A new grassroots pattern costs about $1 million, inclusive of associated field infrastructure for steam, produced fluids processing, electricity, and water disposal.

The average well produces only 10-20 bo/d. Average oil production per five-spot pattern is about 65 b/d.

Texaco, together with Mission Energy, operates two cogeneration plants that generate process steam and 600 mw of electricity for sale into the regional power grid.

For the five major operators in Kern River, 1995 operating costs ranged from $7.94 to $2.74/bbl of oil equivalent.

Economic models

• Demonstrate the potential rewards for investors.
• Test sensitivities of crude price and other variables.
• Place the economic potential of thermal recovery investments in context with those of conventional oil development.
• Illustrate why today's economic climate in the industry is improved.

The models do not include the purchase of proven reserves, but assume that the operator already owns the lease. Proven developed reserves may be purchased for $2.50-3.00/BOE.

A fourth economic model (GOMOil) compares a conventional oil development offshore in the Gulf of Mexico to the heavy oil project.

Table 1 [56,484 bytes] lists the key model parameters, and Fig. 1 [59,793 bytes] shows the production profiles.

Kern River models

The cases assume reservoir stratification, so that depletion requires multiple steam floods, resulting in long project lives of over 30 years.

The KRVirgin case models a steam flood from scratch in a virgin reservoir. This requires new wells and oil gathering and steam distribution pipelines, as well as facilities for oil treating and storage, water treating and disposal, steam generating, and feedwater treatment.

The 20, five-acre-pattern project assumes a 24 million bbl ultimate oil recovery. This case is unrealistic in that opportunities to invest in virgin heavy-oil reservoirs are unlikely to be found. The case is intended to illustrate the underlying economic strength of thermal recovery under today's conditions.

The second Kern River case, KRHalfDep, illustrates a more probable investment opportunity. This case has the same scope as KRVirgin, except that 50% of original recoverable reserves are assumed to have been previously produced by primary and cyclic steam operations.

This case, with a 12 million bbl ultimate recovery, assumes all wells and gathering/distribution piping are newly installed. But, steam generating and oil and water-treating facilities are assumed to already be available in the field.

In both models, a five-pattern pilot project would be implemented first. After producing for 1 year, the remaining work would be done step-wise at a rate of five new patterns per year for 3 years.

Fig. 1a illustrates the buildup in production. Field operators indicated that such a deliberate approach is typical.

This approach is partly driven by capital limitations, and therefore, the post-pilot phase could be implemented more quickly than modeled. This would accelerate production and improve present value profit and rate of return.

Also, to accelerate recovery and shorten field lives, operators could run steam floods in parallel in multiple reservoir intervals rather than sequentially. Our models do not include this option.

Both KR cases are assumed to be on property with a one-eighth royalty, although many major heavy oil fields in California are on fee land (company owned).

These models are not intended to be predictive of any actual reservoir situation or investment opportunity, but are meant to illustrate general economic potential for such investments.

The models are based upon general information and rules-of-thumb provided by Kern River field operators.

Thermal alpha model

The reservoir is deeper and thinner than Kern River, and prior operations produced about 40% of recoverable oil. Project investments included new steam generators and some new wells, but otherwise, existing wells with workovers and existing production facilities with minor modifications were adequate.

Fig. 1b plots the production profile for the Alpha case. No thermal pilot was proposed and implementation was aggressive, so that a rapid production buildup was expected. Consecutive steam floods were planned for three reservoir intervals.

Predicted ultimate incremental oil recovery was 13 million bbl, similar to the KRHalfDep case. Production rates also are similar to KRHalfDep, but are relatively more constant over a shorter project life.

The Alpha case updates the operator's original justification to today's conditions. The same physical model and performance outlook were used, but capital and operating costs and heavy oil prices were brought up to date. Royalty in this case was zero. The projected economics are based on the incremental increase over continued cyclic steaming.

Gulf of Mexico

Recoverable reserves of 12 million BOE were assumed, so that the case is comparable to the KRHalfDep and Alpha cases.

Fig. 1b shows a typical production profile for such a project. Rate would be expected to build rapidly to a peak at a much higher level than for thermal recovery projects.

Peak oil rate represents reserve depletion of about 12%/year. The brief peak would be followed by a rapid decline. This behavior is in marked contrast to thermal-recovery production profiles.

Peak rates, in those cases, represent reserve depletion of about 4.2-5.3%/year, but production rates are sustained for much longer.

The GOMOil case uses a 1,000 scf/bbl GOR.

The model assumes an investment in a new platform, development wells, production facilities, and pipeline connections to an existing nearby platform. Typical investments for such a project are about $4.00/recoverable BOE. This is substantially greater on a per bbl of reserves basis than for thermal recovery projects.

Operating expenses were modeled from actual experience with similar facilities. A normal federal royalty of one-sixth was assumed.

This case, also, is not intended to be predictive of any actual project, but illustrates typical economics for alternative investments to heavy oil.

Operating costs

Operating costs obtained from operators covered a wide range. All operators, however, indicated that direct operating expenses have been substantially reduced over the last few years. They expect this trend to continue.

Ammonite's operating cost model yields costs in the lower half of the range of actual operators' data, but costs were not reduced further for future projections.

All cash costs at the field level include local property taxes.

No specific information was obtained from operators on their intra-company charges for overhead and/or general and administrative (G&A) expense. Ammonite's models estimate these charges at 30% of direct expenses and 10% of capital spending.

Abandonment expenditures, in each model, were assumed to be in the last year of field life and equal to one-third of escalated initial capital investment.

Cash flows projects were made both before and after income taxes. Income tax calculations were not rigorous because they differ for specific investor situations. The generalized tax model assumed a 35% tax rate on taxable income after deducting intangible investments in the year made and depreciating tangible investments on a unit-of-production basis.

No tax bases were carried into the projects, and it was assumed that any tax credits could be realized in the year available.

Base cases

GOMOil assumes initial wellhead prices of $18/bbl for oil and $2/Mcf for gas. The thermal recovery cases discount the oil price by $4/bbl.

Historically, this differential has averaged about $5/bbl, but in 1995 and early 1996, the average was less than $4/bbl. This decrease apparently was due to substantial investment by West Coast refiners in heavy feedstock processing capacity and by the start of Alaskan crude exports.

During 1991, San Joaquin Valley heavy crude (13° API) average price was 57% of the average West Texas Intermediate (WTI) crude oil price. During 1996, it was 72% that of WTI.

During the third quarter of 1996 when crude oil prices in the main domestic producing basins strengthened, California heavy oil prices did not follow suit, and the differential to WTI rose to $7/bbl. This now has declined in the second half of 1997 with increased demand for heavy oil.

The bottom line in heavy oil economics is not the price differential, but rather, the price received per bbl crude sold. In October 1997, Kern River 13° API price reached $17/bbl when the WTI price was $21.50/bbl.

Base-case economics assume escalation of both product prices and operating expenses at 3%/year. Cash flow projections for all three thermal recovery models are attractive with internal rates of return after tax ranging from 23% for KRVirgin to 41% for Alpha.

The Alpha model economics are substantially better than those for KRHalfDep even though produced volumes and initial capital per bbl recovered are about the same. This is because of the faster production rate buildup and accelerated oil recovery, shorter project life, no royalty, and lower average operating costs.

For GOMOil, profitability and internal rates of return are attractive, although lower than KRHalfDep and Alpha. This is because the much higher initial capital expense are not entirely offset by the higher product prices and more rapid reserve depletion.

The relation of economics for this case to those for KRHalfDep is essentially determined by the assumption of initial capital requirements.

Fig. 3 [24,528 bytes] plots the cash flow streams. The differences between the rates of return for each case can be easily seen in the cash flow during the early years.

Heavy initial capital requirement drives negative cash flow for GOMOil to a level about three times that of the thermal recovery models. On the other hand, accelerated oil recovery in Alpha is reflected in the much more rapid payout and cash flow than in the Kern River models.

Alpha economics

Table 3 [76,101 bytes] compares the current economics with the original justification. The table also includes parameters for the KRHalfDep case rerun under the same conditions as the other cases, such as no royalty and unescalated prices and expenses.

It is apparent from the first two columns that today's higher oil prices and lower operating costs have substantially improved the economics of heavy oil development, in spite of relatively higher capital costs.

A major part of that capital cost increase is the cost of adding new steam-generation capacity under current emissions regulations.

The difference between the Alpha economics in the left-hand column and those for the KRHalfDep case is principally due to the accelerated recovery in the former case. This suggests the significant degree to which the economics of the KRHalfDep case might be improved by accelerating the project implementation and production buildup.

However, the trade off for the operator in pursuing such a course would be to risk investing in steam flood expansion into areas that offer less than the expected potential for oil recovery.

Model sensitivities

Table 4 [117,371 bytes] and Table 5 [70,397 bytes] shows the sensitivity to price and expense escalation and the impact of royalty. Fig. 4 [83,869 bytes] plots the sensitivities of oil and gas prices.

The most significant sensitivity of thermal recovery economics is to the future price of oil. The data indicate that $2/bbl lower crude prices would make all of these cases marginally attractive at best, and a $4/bbl reduction would essentially kill all three.

The sensitivities to oil prices assume a specific price over the project life. In the real world, prices are cyclic. Over the project life there will likely be several price cycles. A heavy oil project is better able to weather a down cycle than a conventional project because of its long life.

Note that the KRHalf Dep case, because of its longer life and more gradual production buildup, is substantially more sensitive to crude price than the GOMOil case.

The thermal recovery projects and Gulf of Mexico are sensitive to future gas prices in opposite directions. The Gulf of Mexico case assumes that gas is part of its production stream, so changes in gas price will impact its revenue.

In California, most steam generation facilities, including cogeneration plants, currently are gas fired, and this is expected to continue. Therefore, in those cases, gas prices affect operating expense.

For expense sensitivity in the thermal cases, an operator rule-of-thumb is used. This rule is that roughly one-half of direct operating expense is steam cost and roughly three-fourths of steam cost is gas fuel.

For the base cases, initial gas price in GOMOil was assumed to be $2/Mcf. For the sensitivity, an initial gas price to major thermal recovery customers in California is assumed to be $2.30/Mcf.

Sensitivity cases were calculated for gas price changes of ?$0.22 and $0.44/Mcf. These increments were chosen because they are proportional to the crude-price sensitivity increments of $2-4/bbl.

Fig. 4b shows that gas price changes would change after-tax rates of returns by only 2-3%.

Fig. 4c shows the impact on after-tax rates of return if oil and gas prices were to change together, in proportion. Oil price impact on rate of returns for the thermal recovery cases are moderated slightly by the additional gas price changes, but the basic sensitivity of all cases to crude price changes is not altered significantly.

Heavy oil future

Opportunities in California heavy-oil development are fundamentally different from those found in conventional oil field development. The maturity of the shallow heavy-oil plays in all of California's basins is such that discovery of new fields or acquisition of completely undeveloped reserves is unlikely.

Even if such opportunities could be found, it would be politically impractical to obtain the necessary regulatory permits to develop a new large heavy oil field and construct and operate the necessary thermal recovery facilities.

Rather, operators see their principal opportunities for growing their reserves and improving their financial results in their existing operations. They feel that development of new thermal recovery facilities in existing fields is politically feasible.

In some fields, existing steam generators or cogeneration plants may have surplus steam for sale. Barring that, it is very possible that new steam generators could be permitted in existing fields, at least in the San Joaquin Valley.

The permitting process is difficult and time-consuming, but it is well established and can be done. The key constraint is that any new source of combustion emissions will only be approved if the operator can demonstrate elimination of more than their equivalent in existing emissions.

Some operators have banked such emissions credits and may be willing to sell or trade them.

Specifically, operators cited the following as main areas for growth opportunities:

• Continued incremental improvements in operating efficiency and lowering of capital and lifting costs
• Continued improvement in thermal recovery project execution to recover additional oil
• Further development of acreage in existing projects to increase recovery efficiency where closer well spacing and/or more-efficient, more-aggressive steaming operations will lead to more-thorough heating of the reservoir
• Acquisition of producing properties from other operators, especially interests that might be peripheral to an operator's core business, and thereby, might be amenable to more-aggressive development and operation.

For example, each percent increase in the recovery factor of a lease with 100 million bbl in place equals 1 million bbl of additional recoverable reserves. A typical 5-acre, nine-spot well pattern in a partially depleted field represents recoverable reserves on the order of 600,000 bbl with zero exploration risk.

The Authors

G. Warfield Hobbs is a managing partner of Ammonite Resources Co., New Canaan, Conn., a firm of international petroleum geologists, engineers and economic advisors. He previously worked as an exploration geologist for Texaco Inc. in Ecuador, U.K., and Indonesia, and for Amerada Hess in New York City. Hobbs has a BS in geology from Yale College and an MS in petroleum geology from the Royal School of Mines, Imperial College, London. During 1993-1995, he was Secretary of the AAPG.

Robert O. Winkler is chief engineer of Ammonite Resources. He specializes in project planning, reservoir performance analysis, gas storage, and reserves-management systems. Prior to becoming a consultant, he was employed by Exxon Corp. in various worldwide locations. Winkler has a BS in mechanical engineering from Bucknell University and an MS in mechanical engineering from the California Institute of Technology. He is an SPE member.

Copyright 1997 Oil & Gas Journal. All Rights Reserved.