by C.J. Campbell on July 7th 1999
The title of my talk is The Imminent Peak of World Oil Production. I would like to provide the evidence. It is of course a very large subject. There are colossal economic and political consequences. Indeed the very future of our subspecies – Hydrocarbon Man – is at stake. But I think that you are better qualified than I to assess these matters. I will therefore concentrate on the technical assessment.
My qualifications for taking up your time are that I have spent the last 45 years studying the subject both directly and indirectly. I have evaluated hundreds of oil prospects around the world. I have drilled many dry holes and even made a few discoveries. I have observed the oil industry from many angles, including its senior management. I have published two books and several papers on oil depletion.
I will start by giving you a short explanation of petroleum geology.
If you go to the coast near the village of Kimmeridge in Dorset, you will find a black clay which smells petroliferous and sometimes even catches fire. It was deposited 140 million years ago near the end of the Jurassic Period.
Chemical analysis shows that it contains up to 10% organic material, which itself contains various compounds characteristic of plankton and algae.
It is a truly remarkable clay, only about 100 metres thick, which was responsible for almost all the oil in the North Sea. It was deposited under unique conditions not found before or since in NW Europe over 600 million years of recorded history.
Geochemical and geological advances over the past 20 years have made it possible to understand the origin of hydrocarbon source rocks such as this Kimmeridge Clay. It was deposited in warm sunlit waters that allowed prolific blooms of algae. The Jurassic was a period of global warming. Britain was then closer to the tropics than now, due to plate tectonic movements.
Normally, the remains of the algae are destroyed as they sink to the seabed. But great rifts, analogous with the Red Sea, were developing in what is now the North Sea as the North Atlantic began to open. Stagnant conditions developed in the depths of these rifts, which preserved the organic material. Furthermore, relatively little other sediment was being washed in, so that the organic material was concentrated.
Let’s now look at samples from this same Kimmeridge Clay taken from a borehole in the eastern side of the North Sea. We will notice a difference. The organic material now contains admixtures of other organic substances that have the isotopic signature of plants. The area was closer to the Jurassic coast and more vegetal material was being washed in.
Tests in the laboratory show that if you heat the algal material you get oil, whereas if you heat the other stuff with plant remains you get gas. Furthermore, if you heat the oily rock too much it breaks down into gas.
Let’s now track the later history of this Kimmeridge Clay in the North Sea. The rifts were covered by several thousand metres of younger sediment which were laid down in a broad basin.
As the Kimmerdige Clay was buried, it became heated by the Earth’s heat flow. At a certain point, chemical reactions occurred similar to those observed in the laboratory. They led to the conversion of the organic material into oil and gas.
The conversion involved expansion so that the each droplet of oil and gas was born under a very high pressure. The pressure increased as the rocks continued to subside. Finally the droplets burst and the oil and gas forced its way upwards to zones of lesser pressure. It was an episodic movement. It occurred only briefly in geological time when the temperature and pressure thresholds were passed. In the North Sea, oil generation commenced when the source rock was buried to about 2000 m
We now have to study where it moved to. The subsidence I have described was periodically interrupted by earth movements, even volcanic activity. The basin was folded and faulted into complex structures. They can now all be mapped in extreme detail thanks to advances in seismic surveying. It is as if we had a very high quality X-Ray of the earth, showing every feature in three dimensions.
We have to unravel geological conditions of great complexity. I will simplify to consider three different situations.
- First, let’s consider the case where the source-rock is buried under a thick sequence of uniform clay. The oil and gas will move upwards along the hair line fractures until pressure equilibrium is reached. The result is an un-exploitable disseminated deposit.
- Second, let’s look at the case where the oil, as it moves upwards, encounters a porous sandstone, rising to the surface at the edge of the basin. The pores in this carrier bed contain water. The oil is lighter than the water, and floats upwards to the surface where it escapes and is lost.
- Lastly, we come to the third case where the carrier bed has been folded or faulted. The oil floats upwards, but this time it cannot escape, being trapped in the highest part of the structures. Such traps form oilfields.
- But we are not quite home yet, because the trap has to be sealed by the rocks above it. Such seals are never perfect so that the oil and gas leak over time.
This slide is an illustrative map of an oil province. It explains why oilfields depend on a very rare combination of circumstances. There are many structures shown as white blobs. The yellow belt shows where there is a reservoir. The purple belt shows where there are generating source rocks. Only structures where the essential ingredients coincide can contain oil. It is of course still more complex than this.
I think I can summarise the position into a few key points.
We now have a comprehensive understanding of petroleum systems. It has become relatively easy to identify and map them, once the critical information has been gathered from seismic surveys and exploration boreholes.
The prolific generation of oil and gas was a very rare event in both time and place in the geological past. Furthermore, much that was formed was lost by leakage. As I said, in the North Sea, we rely on one unique event 140 million years ago.
The world has now been very thoroughly explored with the benefits of this new understanding and the high resolution seismic surveys. About 90% of the world’s oil endowment lies in just 30 major petroleum systems.
This shows where oil was generated in the North Sea. It was generated in the coloured areas and nowhere else. Similar maps are now available for virtually the entire world. All such provinces have been subject to at least some exploration. All the promising areas have been thoroughly explored. There are good reasons why other areas receive little attention. There are of course details to fill in, but the general picture has become very clear. Some oil economists claim that if all the basins of the world were drilled as intensively as Texas, they would yield a huge amount of new oil. They are utterly mistaken for well understood scientific reasons.
Now, I would like to go to Norway and explain the pattern of discovery there. I choose Norway, because the Norwegians publish reliable information.
This plots cumulative discovery against cumulative wildcats. These are the exploration wells that either do or don’t find new fields. By plotting discovery against wildcats we overcome the distortions from external factors such as government policy. The large fields were found first and discovery follows a hyperbolic model fairly closely. Exploration will end when the discovery curve becomes almost flat. This plot suggests that Norway’s total is just under 28 Gb.
This is a plot of the size distribution of the fields. We can see the close correlation with the theoretical parabolic fractal distribution. It gives a slightly lower total of 26 Gb
These models – and there are other statistical techniques – suggest that the total of Norway’s known basins is between 26 and 28 Gb. But I think that some of the more recent fields will prove a little bigger than currently estimated. I accordingly use judgment to slightly modify the theory, giving Norway an Ultimate Recovery of 29 Gb. Note how critical it is to have good data of past discovery both by quantity and date.
Similar plots have been made for every productive area.
It is also interesting to plot a company’s record
This shows Shell’s record. It has drilled about 3600 wildcats since 1885, finding 60 Gb oil. The fit with the hyperbolic model is remarkable. It suggests that if Shell drilled as many wildcats again, it would find only about 16 Gb. The falling discovery is despite a large budget, the best available technology and a deliberate policy to find the largest fields. The short explanation is that there has been progressively less left to find.
We might look at another company, Amoco. It was much less successful because it was late into the international arena. It found about 15 Gb with 900 wildcats, but hardly anything with the last 600. It is no surprise that it was forced to sell out to BP.
The point is that all discovery curves are flattening.
Looking at the world as a whole, we see this growing deficit. Discovery peaked in the 1960s with a 60 Gb surplus. But that has given way to a deficit of almost 20 Gb We now find one barrel for every four we consume.
The general situation seems so obvious. Surely everyone can see it staring them in the face. How can any thinking person not be aware of it? How can governments be oblivious of the realities of discovery and their implications? How is it possible, given the critical importance of oil to our entire economy.
There are several possible explanations, including denial, but one of the main reasons relates to the reporting of reserves. It is a huge subject, but I will try to reduce it to three cardinal issues
First, we need to distinguish all the many different categories of oil and assess their endowment in nature, their characteristics and above all depletion profile. Obviously, a Middle East well flowing at 20 000 b/d is very different from extracting a few barrels a day from a tar deposit. Deepwater and polar oil is different from oil in Texas. We need to know how each category can contribute to peak.
Second, we need to define the probability ranking of our estimates so as to get the best estimates of what a field will actually deliver, such that statistically any revisions will be neutral.
Thirdly, we have to backdate any reserve revisions to the discovery of the fields containing them.
Unfortunately the public database is extremely unreliable. I might here refer to BP’s Statistical Review which many people consider a valid source of information given its reputable author. It is in fact exceedingly unreliable. It simply reproduces data from a trade journal and does not reveal the Company’s own considerable knowledge. It is very unfortunate that a company of this standing should put out such misleading information. I don’t know why it does so.
There are several reasons why the public database is so unreliable and misleading..
- First, companies systematically under-report the size of discoveries for a host of good regulatory and commercial reasons. They refer to what they call Proved reserves which are much less than what the field will eventually deliver. They treat reserves as an inventory to be booked as suits their commercial and financial needs.
- Second is the unreliable reporting by the major OPEC countries. This table shows two obvious flaws. First is the huge increase in the late1980s, underlined in red. The sudden increase resulted from the so-called quota wars: quota being partly based on reserves. The second is shown in green. It is absolutely implausible that reserves should have remained so constant subsequently given the high production rates. Incidentally, 60 countries reported unchanged numbers last year. No right thinking person could accept such a data-set, yet it is embodied un-noticed in the public database.
…..But there is a third less obvious, yet much more important, factor. Clearly, nothing happened in the reservoir in these countries in the late 1980s. The reserve revisions — whatever the right reserve number is — should be backdated to the discovery of the fields containing them. They had been found up to 50 years before. Backdating has a huge impact on the discovery plot.
This illustrates the popular image of growing discovery, based for example on BP’s public numbers. It fails to backdate revisions and has misled many analysts. They see reserves growing which they attribute to technology. In reality the growth was just in the reporting.
This shows the true position with more realistic numbers properly backdated. It shows that discovery is flattening not growing.
It is fashionable to accuse oil companies of conspiracy, but I think the deception has more to do with differing mind-sets and objectives than conspiracy. Oil companies are in business to make money not plan the world’s future.
Backdating is obvious to an explorer. But the engineers who develop the fields over a long period of time soon forget the original discovery and report the revisions when they occur having no particular reason to backdate. The management takes the revisions as best suits their commercial image. Reserve growth makes a much better story than a dwindling asset, which is the reality.
Mostly, this lax reporting does not matter. But if we want to use the record of the past to extrapolate future discovery, we need to insert the right amounts and the right dates.
I should say something about getting more oil out the reservoir. You will hear a great deal about that. Now that falling discovery is widely recognized, hopes are pinned on getting more out of what has already been found.
I am sorry to say that reports of improved recovery are largely an illusion reflecting the reporting procedure rather than any particular technological achievement.
I might illustrate this by Prudhoe Bay. It is by no means exceptional. This plots annual production against cumulative production to show the end point. The operator internally estimated its reserves in 1977 at 12.5 Gb. But for a host of good commercial and regulatory reasons reported 9 Gb. The depletion curve has been as straight as an arrow since 1991. It shows that the field will barely make the original estimate, despite all the considerable technology that has been applied. Yet many analysts are misled by the rise in the official numbers from 9 to 12.5 Gb.
Of course it is possible to go back to an old field developed long ago with poor technology and extract a little more oil from it by a range of well known methods, such as steam injection. But this is a phenomenon of the dying days of old onshore fields of the United States, Soviet Union and Venezuela. Most modern fields are developed efficiently from the beginning. In any event the addition contributes little in global terms and has no impact on peak.
Technology serves mainly to hold production rate as high as possible for as long as possible. That obviously makes the most profit. But it adds little to the reserves themselves and clearly accelerates the rate of depletion. The high depletion rate of Norway shows how efficient they have been at extending plateau production. The decline slope now becomes a cliff.
I hope I have explained why it is so difficult to obtain valid numbers and dates. It is hugely complex subject.
If we hired a detective to somehow penetrate the smokescreen, I think he would come up with numbers like this for what we can call “narrowly defined conventional oil”. This is the stuff that has provided most oil to-date. It will also continue to dominate supply until long past peak. It excludes oil from
coal and shale;
heavy oil and tar,
deepwater and polar areas,
synthetic oil and
natural gas liquids.
The contribution of these non conventional categories has to be added to get total supply, but their impact on peak is small.
The bottom line is that there is a rounded one trillion barrels left to produce.
It is most unevenly distributed with about half lying in just five Middle East countries, due to that region’s unique geology, including particularly the widespread occurrence of salt which seals the reservoirs, preventing the escape of oil.
No one can dispute that you have to find oil before you can produce it. The curve of discovery clearly has eventually to control the curve of production that follows it after a time lag.
Experience has shown that the exploration of a new province follows a standard pattern. It starts with the learning period. Boreholes have to be drilled and seismic shot to find out if the new area has the necessary geology. If it does not, it will never deliver no matter how much investment and technology is thrown at it.
But if it does have the ingredients, the larger fields are found first. They are too large to miss. These large fields give a natural peak.
This natural discovery curve has to be reflected in the production curve that follows. Production, like discovery, starts and ends at zero, and reaches a peak around the midpoint of depletion when half is gone.
The US-48 is a good example. Discovery peaked in the 1930s and production in 1971, despite ample technology, money and incentive. In the North Sea, peak discovery around 1980 is now being followed by peak production. It should surprise no one.
We can use this general curve to model depletion in most countries, adjusting for any special circumstances.
But it is not quite a simple as that because we have to recognize the swing role of the Middle East countries around global peak. They have large reserves and a low depletion rate. So they can make up the difference between world demand and what the other countries can deliver within their resource base.
Swing share was 38% at the time of the first oil shock in 1973. It fell to18% by 1985 as flush production from new provinces, such as the North Sea, that had already been found before the shock, was dumped onto the market. I say dumped because had the major companies not lost control through expropriation, they would probably have managed things better.
Since then, share has been rising to around 30% to-day. This time it is set to continue to rise because there are no new provinces ready to deliver flush production, save perhaps the Caspian and that seems to be turning sour.
This allows us to develop some scenarios about future production. I will describe in a moment what seems a reasonable base case.
But before coming to that, I should say something about the International Energy Agency. It was established in the aftermath of the oil shocks of the 1970s, under a treaty signed by most OECD governments. It has a mandate to study oil supply and alert the OECD governments to dangers to supply. Note that the governments are treaty-bound to react and coordinate their response.
It is naturally a highly political institution. But nevertheless it did succeed this year to deliver a coded message. I am glad to say that the media is decrypting the message, as I gather the IEA intended.
I would say that this message from the IEA, the highest world authority on the subject, is dynamite. I am sorry to say that the European Commission remains in blissful ignorance; and I don’t have any confidence in the DTI either.
I will now explain what I think is a reasonable scenario
1) Oil demand will grow at 1.5% a year – slightly below the IEA estimate of 1.8% – until Swing Share reaches about 35% in 2001.
2) The Middle East countries will then have the confidence to impose much higher prices, realising that they have no competition. They may even get such confidence sooner.
For example, they might read an official report showing that Norway’s production is set to halve by 2006. Norway is the world’s second largest exporter. The impact on Swing Share is obvious.
3) I think prices may briefly soar to very high levels due to the working of the market that sets prices on the marginal barrel. I believe that the market itself may be manipulated by hedge funds and similar insiders, who are in a position to talk price up and down. They must have made huge fortunes when prices recently rose 80% over a few weeks. Most forecasts now predict falling stocks by the last quarter as the insiders talk price up again.
4) I think that a price shock around 2001, if not before, from Middle East control is inevitable and will probably trigger a stockmarket crash
5) I think that demand does become elastic above about $30/b, reacting to normal market forces, so higher prices may curb demand.
6). Nevertheless, I think it will be a time of great political and economic tension as Europe, America and Japan vie for access to Middle East oil. More missiles can be expected. The third world will be badly hit, being unable to afford imports. Agriculture is very dependent on oil.
7). But I expect that somehow a plateau of production, however volatile, will unfold around $30 a barrel. But the end of the plateau will soon come into sight.
8) It may have a fundamental impact on investment. Up til now, the investment community has believed in perpetual growth on which cycles are superimposed. The bottom of each cycle has been higher than its predecessor making capital appreciation the primary goal of investment. But the tensions of the oil shock and related events, including the colossal financial transfers to the Middle East, may create a new view.
After the many years of growth we may then experience a new downward trend, however cyclic. Share prices may sink to more realistic levels as the main focus will be on yield not growth. Capital will be destroyed.
9) The plateau has to come to an end by around 2008 when Swing Share will have passed 50% and the Swing countries in the Middle East will be approaching their depletion midpoint too. Production will then start its inevitable long term decline at about 3% a year. Increasing shortages will develop, and agriculture and transport will be seriously affected. The global market will come to an end because of high transport costs.
That is a scenario. There are of course many alternatives, but the range of possibility is limited given the resource constraints. These constraints are facts not scenarios. If by some miracle we could add 500 Gb of reserves – more than half as much as produced so far – it would delay peak by only ten years.
One indisputable fact stands out. Discovery peaked 30 years ago. It takes no feat of intellect to conclude that we now face the corresponding peak of production.
But there are solutions. Gas and non-conventional oil production can be stepped up giving an overall peak around 2015. Gas will be useful but needs special management because it depletes very differently from oil on account of its greater mobility. The controlled plateau ends abruptly. The open market is not designed to deal with depletion.
Nuclear energy and coal production can be stepped up. The environmental hazards can presumably be solved by technology if efforts are made..
Renewable energy and above all energy savings can make really important contributions.
One problem is that all of these things take time to implement. Nothing is likely to be done without proper fiscal or price incentives.
I think it is absurd that the management of the depletion of the world’s supply of its most important fuel should be left to a few feudal families controlling the Middle East. The consuming governments should recognize where their interests lie.
They could for example ameliorate the tensions by introducing a Depletion Protocol such as I proposed last year.
But it is easier for them to react to a crisis than anticipate one. My hope therefore is that this talk will have helped you understand the nature of what is about to strike even if in practice you are unable to prepare.
I close with a simple diagram. Think of it as two tanks. The top contains what is Yet-to-Find and it is flowing into the second tank at a falling rate of about 6 Gb/a. The second tank contains what it left from past discovery. This tank is being drained at about 23 Gb/a rising : much faster than it is being filled. We take out four barrels for every one we put in.