Better Understanding Urged For Rapidly Depleting Reserves

C.J. Campbell
Petroleum Consultant
Milhac, France

In the worst days of the Battle of Britain, military commanders had to balance the need to commit fighter planes to defend the capital against having enough in reserve for later attacks. They had to have accurate information on the resources available, but it was easily obtained by a count that could be checked and double checked.

It is equally important that the world should know what its reserves of oil are because its economy depends heavily on oil-based energy, especially for transport and agriculture, which means food. It also needs to know where the reserves are. But unlike in the Battle of Britain, no one can go out to count the reserves because they lie far underground in geological traps whose precise nature can only be estimated.

The exact size of the reserves of a field will be known only on the day when it is finally abandoned; until then, some measure of uncertainty will always surround the number. We need to understand that uncertainty better, in the same way as we assess the probability of a hurricane striking a particular place at a given time.

For more than 60 years, the Oil & Gas Journal has put out an annual report of the world's reserves based on information provided to it by governments and industry. It has been a valiant effort, having the great merit of consistency. The numbers are widely accepted as a firm basis of what the reserve position is, being reproduced, for example, by British Petroleum Co. plc in its Statistical Review of World Energy.

In earlier years, when there was still a great deal of oil left to find, the issue of reserves was not particularly critical. They were little more than a pragmatic inventory. But now, when there is comparatively little left to find, the subject becomes vastly more important, especially when we want to use the trend of past discovery as a guide to estimating future discovery.

To date most of the oil produced has been what is called conventional, which means that it has a particular production profile, starting at zero, rising rapidly to a high peak or managed plateau, and then declining exponentially to zero, when it is exhausted. Probably most oil to be produced over the next 20 years or so will also be conventional. It is relatively cheap and easy to produce. In the past, this is what was in practice reported, and there was no particular need to identify it as such.

But as we look ahead there is an increasing need to distinguish it from non-conventional oil made up of heavy oil, tarsand oil, enhanced recovery, oil in hostile environments, etc. It has a very different depletion profile, rising only slowly to a long, low plateau, before eventually declining in the distant future. It is costly and slow to produce.

A third category of "oil" is natural gas liquid or condensate, which will increase as gas production grows in the future. It belongs to the gas domain and should be distinguished from conventional oil.

In short, instead of talking simply about reserves, we should be careful to make clear what we are talking about. I will confine my attention here to conventional oil.

Qualifying terms

The traditional and still widely used measure of the accuracy of the reserves is conveyed by the qualifying terms proved, probable, and possible with the meanings the words imply.

The proved category has a very high probability of being produced, say 80-90%. For many purposes, the classification is useful enough, especially where it is employed simply to describe a form of inventory. If, however, we want to know what is most likely to be produced in actuality, we need a different designation.

Unfortunately, the terms are cumbersome: median probability reserves; proved and probable reserves; P50 reserves; P, 2P, 3P; mini, midi, maxi; or low, mean, high case. It would be a great help if some professional body should set universal norms, so that everyone would know what they are talking about, but I do not propose to get into this arcane and highly charged subject here.

If the reserves of a field are initially defined conservatively as proved, it is obvious that they will grow over time as they approach the "median probability" value, which is the best estimate of what will be actually produced. It is important to distinguish this form of reserve growth, which is an artifact of definition and reporting procedure, from what is due to economic and technological factors. It is also important to understand recovery factors, which are sometimes assumed to improve over time from technological progress.

The initial estimate of reserves is based on calculating the oil-in-place of a trap and then applying a recovery factor based on the reservoir and oil characteristics. A fairly conservative number is normally applied. As the field is depleted, it becomes possible to determine the reserves and ultimate recovery from the decline curves of the wells themselves.

The new number often shows an apparent improvement in recovery when compared with the initial conservative oil-in-place value. Again, it is often more an artifact of definition and reporting procedure than a technological dynamic. Besides, no one will ever fully know exactly what the oil-in-place is because much of it stays forever in the ground.

In practice, companies tend to understate reserves for a variety of good reasons. By contrast, the U.S. Geological Survey tends to overstate. Its definition is high case (with a 5-10% probability) because it excludes economic and technological factors and has a very restrictive definition of "non-conventional." Soviet reserves were also defined on a comparable basis.

Some countries understate. The most glaring example is the U.K., which reports proved as 4.5 billion bbl, when the median probability number is thought to be almost three times higher: a seemingly excessive range for a shelf as well known as the U.K. North Sea.

Some countries overstate. Mexico has confessed to doing so, primarily by the inclusion of nonconventional reserves from the Chicontepec field.

In 1988, Venezuela doubled its reserves by including 20 billion bbl of heavy oil, which had been known for many years. It had the not necessarily intended effect of increasing its Organization of Petroleum Exporting Countries quota, which led other OPEC countries to retaliate with huge unsubstantiated increases, as has been recently confirmed by F. Barkeshli, an Iranian official. Although we don't really know if the new number was too high or the old number too low, the point is that nothing in particular changed in the reservoir in 1988: It was all in the reporting.

Lastly, many countries-43 in 1996-report unchanged numbers year after year, which is obviously implausible. In the absence of information to the contrary, it is therefore well to reduce such reports by the cumulative production of the unchanged period.

Backdating

For many purposes, it does not particularly matter if reserve revisions are taken on a current basis or backdated as if known on the discovery date of the fields containing them. But if we are to use the discovery trend of the past to extrapolate future discovery, it becomes vitally important to properly backdate.

In earlier years, it was normal to measure security of supply in terms of a reserve to production ratio, quoted in years. It was a good enough system when discovery exceeded production. This is no longer the case, and it is absurd to propose that a given reserve can support current production for so many years and then stop overnight.

Production declines in all oil fields as they approach exhaustion. It is much better to talk about depletion rate: the percentage of the reserves being produced each year. The world is currently finding less than 6 billion bbl/year on a falling trend and producing 24 billion bbl/year on a rising trend, evidently eating into its oil capital.

Ninety percent of all production comes from fields more than 20 years old. Discovery, and particularly giant discovery, peaked in the 1960s, despite all the subsequent technological advances and increases in worldwide knowledge. The industry becomes ever more efficient yet finds ever less.

It is not at all easy to estimate what the median probability reserves are without access to the authoritative database of Petroconsultants, but we can try to do so by evaluating each country, first removing spurious reports and then applying a factor to convert the adjusted number to "median probability" (Table 1- Requires Adobe Acrobat Reader). It is not an exact science, but it is better than accepting the reported numbers at face value.

The study shows that, as of the end of 1996, the world had produced 784 billion bbl and had reserves of 836 billion bbl, meaning that the total discovered was 1.6 trillion bbl. In Fig. 1 [46047 bytes], extrapolation of the discovery trend suggests a yet-to-find total of 180 billion bbl and an ultimate of 1.8 trillion bbl.

Fig. 2 [37745 bytes] shows a depletion model of conventional oil production based on these numbers. It assumes that peak production in each country coincides with the midpoint of depletion and a 2%/year increase in demand, save in the case of the Middle East gulf producers, which exercise a swing role.

A radical increase in price is expected when the swing share exceeds 30% of world supply, curbing demand. A resource-driven decline sets in when the share reaches 50%, by which time the gulf producers will be close to their midpoint of depletion, too.

As production declines, increasing amounts of NGL and non-conventional oil will no doubt be produced, but since the depletion profile of non-conventional oil is so different, it will become important only at the tail end of conventional depletion, having virtually no impact on peak.

End of cheap oil?

With about half the reserves and yet-to-find in just five Middle East countries, two currently vilified, it looks as if the end of abundant cheap oil supply is imminent.

At the risk of understatement, we may conclude that the world's political, economic, and political stability, which relies on an abundant supply of cheap oil, is in serious jeopardy. In these circumstances, it might be a good idea to determine what its reserves really are: It is more a political than technical problem.

Bibliography

Barkeshli, F., "Oil prospects in the Middle East and the future of the oil market," Oxford Energy Forum, August 1996, pp. 10-11.
Campbell, C.J., Laherrere, J.H., "World oil supply 1930-2050," Petroconsultants, Geneva, 1995.
Campbell, C.J., "The status of world oil depletion at the end of 1995," Energy Exploration & Exploitation 14/1, pp. 63-81.

The Author

C.J. Campbell holds a PhD in geology from Oxford University. He began work in the oil industry in 1958 as an exploration geologist in Latin America and Australia for Texaco and later worked for British Petroleum. In 1967 he joined Amoco in New York as a regional new ventures geologist, then became general manager for Shenandoah Oil Corp. in London in 1972 in a joint venture with Saga Petroleum. He was named exploration manager for Amoco in 1980 and executive vice-president for Fina in 1985, both in Norway.
In recent years, Campbell has been a consultant to industry and governments, specializing in oil resource assessments. He is an associate consultant with Petroconsultants and author of a book, The Coming Oil Crisis, to be published soon.