Depletion patterns show change due for production of conventional oil

C.J. Campbell
Petroleum Consultant
Milhac, France

Last summer, the industry, led sur -prisingly by the Chinese, bid large sums for the right to reenter old fields in Venezuela under tough terms. The year also saw the leasing of deepwater tracts in the Atlantic off Europe. Brazil, too, had a good response to its offer to open its deepwater shelf, much in water depths of more than 6,000 ft.

Does it mean that there are some nuggets left out there? Or does it rather mean that there are not easier places left to explore?

At any rate, it suggests that there is little left to find in China, where demand is growing fast.

Economists tend to say that oil resources are infinitely large and that there is no possibility of absolute shortage-only shortage at a price. In a certain sense, they are right: homo sapiens won't ever find the last thimblefull of oil hidden out there somewhere; besides, there are huge amounts of heavy oil and bitumen to go at.

But this is not the stuff that has fueled economic growth and prosperity for the past 150 years and provided much of the energy that allowed the world's population to expand sixfold over the same period. That was mainly light oil above 20° API, which flowed naturally from the reservoir into the wellbore. About 70% of it occurs in 300+ giant oil fields (larger than 500 million bbl), whose discovery peaked in the 1960s. It was relatively easy and cheap to both find and produce, and it flowed at a high rate. It is commonly called conventional oil and has a characteristic depletion pattern, starting at zero, rising to a peak or managed plateau, before declining exponentially to zero when it is exhausted.

The family of hydrocarbons is a large one, and each member has its own and very different characteristics and depletion profile. Some heavy oil and bitumen is already in production. Some deeper-water oil is coming in, having stretched technological skills to the limit. Some polar regions are in production. And some enhanced recovery procedures are already in place.

Collectively these other types are sometimes called nonconventional oils, and generally they are difficult, expensive, and very slow to produce.

Oil discontinuity

History is full of discontinuities, when periods of tranquility and steady growth were overturned by sudden changes. Empires, built over hundreds of years, fell in sudden war and invasion. After a century of growth, the builders of steam locomotives found themselves eclipsed by the diesel engine or electric motor. The fish canning industry of Norway was ruined by the advent of the domestic refrigerator.

So the proposition that the world is about to experience the end of the cheap abundant supply of oil, which has created what has been called Hydrocarbon Man, should come as no surprise: It is a pattern well-founded in history. Things change-and suddenly.

Such a fundamental oil discontinuity will likely arise from two factors: first, the physical peak of production at about the midpoint of depletion, which is driven by the physics of the reservoir; and second, the control of the market around peak by a few richly endowed countries.

Oil is unevenly distributed because it was formed only very rarely in time or place in the Earth's long geological history.

What oil will peak? Clearly, the relatively easy conventional stuff. It has been responsible for almost all oil produced to date and will dominate supply for the next few decades until long after peak production. Increasing amounts of other types will gradually be brought into production to soften the landing at the end of conventional depletion, but they will have a negligible impact on peak itself.

The world has been very extensively explored with the help of sophisticated technology and great advances in petroleum geology. It is accordingly almost inconceivable that any large provinces-that is to say, with a potential sufficient to supply the world for more than, say, a year or two-have been missed. It means that most of the future supply will have to come from known basins, most of which are very well known.

Various robust statistical techniques can be used to extrapolate such past discovery to predict the size and number of fields yet to find. They include the creaming curve, which plots cumulative discovery against cumulative wildcats or over time; the parabolic fractal, which evaluates size distributions; and the bell curve and the multiple bell curve, which model production and discovery patterns. You have to find oil before you can produce it.

Naturally, such studies depend on valid data. It is extraordinary that an industry which prides itself on its science and technology is content to operate in such a medieval manner when it comes to defining and reporting its reserves. What we need to know are P50 (probability 50%) reserves, for which the risks that the estimates will be above or below actual production are evenly matched.

As it is, some companies initially report P90 numbers, giving the illusion of later reserves growth. And some countries report P10 numbers partly for political reasons. In 1996, 46 countries reported unchanged numbers, which is clearly implausible. Table 1 [49,890 bytes] attempts to unravel all this and provide a realistic assessment of the world's endowment of conventional oil for end-1996.

The important line is the rounded 1 trillion bbl yet to produce, which is the ultimate less the produced. (A listing of the distribution by country and region appeared in OGJ, Apr. 7, 1997, p. 51).

Depletion models

Theoretical and empirical evidence shows that peak production in any country more or less coincides with the midpoint of depletion, save where production is artificially restrained, as for example by OPEC quota. A depletion model can therefore be built on three groups of countries:

1. Postmidpoint countries, where production is set to decline at approximately the current depletion rate (which in fact is normally close to the midpoint depletion rate).

2. Premidpoint countries, where production is yet to peak.

3. Swing countries: a group of richly endowed countries that are at an early stage of depletion.

Fig. 1 [274,739 bytes] and Fig. 2 [304,318 bytes], built on end-1996 data, plot the depletion profiles of representative countries in the first two categories. Ideally, they should show total discovery, but here we will have to be content with published giant field reserves, shown as bars. They show a characteristic time lag between peak discovery and peak production. The amount produced at depletion midpoint (half the ultimate) and the date when it is reached are noted.

Given the atrociously weak database, no one should expect great accuracy. The aim is to identify the climate, not the weather.

Many claims are made for the impact of technology, but it is well to remember that the main result, if not the objective, of the remarkable technological achievements, of which the industry can be justly proud, is to extend plateau production, which makes economic sense. In some cases, these efforts may indeed defer peak beyond depletion midpoint, but at a cost-the subsequent decline will become a cliff, making a delayed crisis even more severe.

Other claims are made for "reserve growth," but in reality most revisions simply reflect the shift from P90 to P50 numbers as the field's size can be assessed with greater confidence as it approaches exhaustion. Such revisions should of course be back-dated to the discovery of the field containing them, because nothing was actually added.

Still other claims are made about improving recovery. It is said that recovery was previously 30% but now is up to 50%, thanks to technology.

This doesn't bear close analysis either. In earlier years, few fields had been abandoned, and no one knew what percentage would be recoverable: the 30% was little more than a pragmatic inventory number. The very concept of a recovery factor in fact is useful only in the early stages of development; later, recovery itself is calculated from well decline curves and material balance studies. Many apparent increases in recovery factor simply mean that the oil-inplace estimate is also due for upward revision.

Postmidpoint countries

The post-midpoint countries are the easier ones to model, as they show a consistent exponential decline over many years which may be extrapolated with confidence. Drilling rate has in most cases steadily declined as fewer and fewer prospects remain to test.

Extrapolating the trend commonly puts an end point to exploration drilling within a decade or so. The discovery plot is now almost flat at the tail end of its hyperbola: only small fields being found either now or in the future. Most small fields were found in the course of looking for the large ones. So, in practice exploration will likely end before all of them have been found.

How many wildcats are drilled today around Titusville, Penn., where production peaked in the last century? Things do grind to a stop

Good examples are Austria, Germany, the U.S., Canada, Argentina, Trinidad, Venezuela, Egypt, Indonesia, and Australia (Fig. 1). In some cases, the depletion is essentially unfettered and follows a natural curve, reflecting the discovery curve before it. In other cases, the opening of the offshore or the constraints of OPEC quota or underinvestment may intervene.

In the countries with the smaller and more concentrated endowment, such as Germany or Austria, peak comes before depletion midpoint, whereas those with a large population of fields, such as the U.S., tend to have more symmetrical profiles.

The shoulders on the U.S. profile incidentally reflect proration and the late discovery of Alaskan oil, which nudges the nonconventional. There may be another late blip on the tail for the deep offshore.

Trinidad's profile reflects an early onshore phase, followed by two offshore campaigns, all of which are now mature.

Venezuela is a special case, with the saddle of the 1980s representing OPEC quota and perhaps underinvestment. Now, production is rising to a second peak, built largely on what would naturally have been produced during the saddle. It may well rise above what the model shows, in which case the subsequent decline will be steeper. With most of the country's oil found before 1930, it does not seem likely that current readmission of the international companies will lead to any radical change in the conventional resource base. Much of Venezuela's reported reserves is heavy oil that should be treated as nonconventional, having the corresponding depletion profile.

The late postmidpoint peak of Canada's profile is surprising for a mature country.

Australia's curve, with its rather steep imminent decline, may be distorted by the inclusion of large amounts of NGL from the North West Shelf gas fields.

Although the long term model is robust enough, there are always anomalies around the inflection point, not least from the gross weakness of the public database.

Premidpoint countries

Many countries, and especially those with a second offshore lease of life which opened during the 1970s, are now approaching their midpoints. Production accordingly is set to rise for a few more years (Fig. 2).

The model either extrapolates the past trend or uses a default of a 5% annual increase. The assumptions could be improved with short term actual forecasts where known, but since they apply only for a few years it is not a critical factor.

One of the more important is the former Soviet Union (FSU). Its reserves are here assessed at 98 billion bbl; published reports range from 57 billion bbl to 190 billion bbl. Peak discovery was followed by peak production. Then came the post-Communist collapse but the prospect of a second subsidiary peak around 2010 for Caspian production. It could be higher.

China is a very thoroughly explored country, expected to peak around 2000. Its reserve data are, however, suspect, the reports having been unchanged for many years at 24 billion bbl. There seems little scope for reserve revision as many of the fields are produced on a close spacing.

Mexico and Colombia will peak in a year or two, the former's reserve reports being suspect. Norway and the U.K., where the resources have been depleted at high rates following the early discovery of giant fields, are also close to peak production. If efforts to extend plateau production are successful, their already steep declines will be even steeper. Norway is a major exporter, and its decline will have a major impact on the swing production discussed below.

Algeria, Libya, and Nigeria all demonstrate a production saddle related to OPEC quota and are likely to reach the second peak around 2000.

India, Malaysia, Syria, and Oman are other examples of countries about to peak.

Swing producers

The model assumes that five Middle East countries-Abu Dhabi, Iran, Iraq, Kuwait, and Saudi Arabia (including the Neutral Zone)-act as swing producers making up the difference between world demand under alternative scenarios and what the other countries can produce. How they exert their swing control will largely determine the price of oil.

Together, they own about half the world's yet-to-produce volume of conventional oil. Here, it is assumed that they will impose radical price increases when they supply more than 30% of the world's demand, leading to a plateau of production until their share reaches 50%. By then, they will be close to their midpoint of depletion, and world production will have to decline at the then depletion rate of about 3.3%.

How to allocate the swing production to the member countries is a second issue. Here, it is assumed that Iraq will be in full production by 2000 and that from then onwards each swing country's production will equate to its share of the aggregate swing yet-to-produce until it reaches its midpoint. Saudi Arabia then picks up the balance.

Fig. 3 [74,525 bytes] illustrates the swing share in relation to oil price.

The model therefore suggests that the world faces a two-stage oil crisis. The first will come when the swing share reaches a level to impose much higher prices, which is expected to happen before 2000. That should curb further demand. But 10 years later will come the second stage, when physical shortages begin to appear. By about 2015, the Middle East swing producers themselves will be past midpoint and into permanent decline (Fig. 4 [55,721 bytes]).

Fig. 5 [67,092 bytes] shows that the depletion profile of the nonswing countries, as modeled, comes very close to the theoretical bell curve because production was relatively unfettered, being free of OPEC quota.

Gas, unconventional oil

After the oil age will come the gas age. Gas is more abundant in nature than is oil, but it also needs much better seals-salt or permafrost-to hold it in the reservoir.

The ultimate endowment is just short of 10,000 tcf, and production is likely to peak around 2020 at 120 tcf/year. In terms of oil equivalent on a current value basis, this amounts to 1 trillion bbl of oil equivalent (boe), compared with 1.8 trillion boe for conventional oil.

Heavy oil and bitumen production, mainly from Canada and Venezuela, will no doubt rise in the future. More deepwater oil and gas (greater than 500 m water depth) will come in. Perhaps there will be some contributions from new finds in the Arctic, and enhanced recovery may help out.

Furthermore, nonconventional gas, especially coalbed methane, will be important in North America and perhaps Europe, China, and Australia. Fig. 6 [178,434 bytes] is a tentative forecast of production from all hydrocarbons.

Are these numbers wrong? Yes, of course they are, given the atrocious state of the database, but they are not that wrong. With the 1997 reserve numbers released in this issue, we can begin to fine-tune the model.

The message they convey is sufficiently serious for the issue to be addressed with some urgency, given the long lead time to develop alternative energy and find ways to use less.

It is not a Doomsday message: quite the contrary, because a more sustainable world may be a better place in which to live. The difficulty is the transition.

The Author