In 1970, the oil reserves, published by the oil companies, represented 30 years of consumption of the time, which amounted then to 2,4 billion tonnes of oil per year. The published reserves therefore amounted to 72 billion tonnes of oil back then.

In 2005, after having consumed oil for 35 years, for about 110 billion tonnes, that is much more than the known reserves in 1970, there was still roughly 160 billion tonnes of reserves, not taking into account what is called "non conventional oil" that comes on top, and for which the estimation of the potentiel is all but simple (see oil formation).

How can we have witnessed this apparent miracle, which is that reserves have grown and multiplied along with an increasing consumption, when earth is finite ? All the answer to this question lies in the definition of a reserve, which is not a physical notion, but a physico-technico-economical notion.

Besides the definition of reserves is not the same depending on countries !

Americans define reserves as "oil that we reasonably consider to be able to extract in the future from the know physical resources, with the known techniques and in the present economic conditions". They are also called "proven reserves".

In the rest of the world, "reserves" generally correspond to the addition of proven reserves, defined as above, and a part of other reserves, said to be probable or possible, that correspond either to the extra reccoverable fraction of resources yet discovered, or to the recoverable fraction of oil resources not yet discovered but whose future discovery is considered more or less probable.

The proven reserves, which are the only one to be published, therefore do not designate what remains below the surface, but only the fraction of that oil that we think we will be able to extract in the future, with the techniques available today (or in a near future), and in the current economic conditions (in other words the limit is that the cost of extraction must not be superior to the market price). A proven reserve is therefore a subjective notion by nature.

Once someone has understood what is a reserve, one can realize that there are at least 4 ways to "make them grow" :

we can discover new resources underground. Without changing the technical or economical conditions, this leads the reserves to larger amounts. However, in such a case we had better talk of a better knowledge of what earth contains than of an "increase of the ressources". Indeed, as hydrocarbons require millions of years to form, we can consider that the initial stock is given once and forever. We can therefore not "increase" the resources, just discover those that we still do not know of.

we can improve the technique, which, practically, means that we increase the proportion of the oil enclosed in the reservoir that we are able to extract (or consider that we are able to extract !). If the recovery factor increases enough, it can increase very substancially the reserves (this recovery factor is around 35% today : any 1% increase - that is going from 35% to 36%, then 36% to 37%, etc - would increase the reserves of one to two years of consumption - 2000 being a reference).

economic conditions may change : if the market price of oil is $ 20 dollars per barrel, it is meaningless for oil companies to try to extract oil if it costs $ 25 dollars per barrel, even if the quantities that they would be able to extract at his price are potentially huge. If the barrel jumps to $ 60 (we are not there yet, but as long as our civilization is heavily dependant on oil all we have to do is wait), then the reservoirs where the extraction cost is $ 25 dollars per barrel will become included in the reserves, for the recoverable fraction of course.

the share value for public oil companies being proportional to the amount of the reserves they declare owning, and the production quotas of the OPEC countries being proportional to the reserves they publish, it is also easy to understand that these reserves can vary - for example go from a high estimate to a low one, or the reverse - without the modification of any technical or physical parameter, just because the appreciation of what is "reasonnable" has changed.

Now we have gone through all the various ways to "increase" the reserves, the good question is : what combination reasons of has allowed this "growth" for these 35 years ?

The first idea that comes to mind, of course, when one observes that the reserves have increased in spite of an evergrowing consumption, is that many new oilfields have been discovered, and that these discoveries more than compensate our consumption. Well.... for the last 20 years, this is totally wrong.

In green, annual discoveries of cenventional oil, in billion of barrels, and in blue, the world annual consumption of oil products (in billion barrels also !). Since 1980 (some authors even consider this as true since 1970), we consume each year more oil than the one found in the new reservoirs (the curve refers to all oil discovered below the ground, not the fraction that we will extract, that will be notably inferior). Source : Exxon Mobil, 2002

Same curve than on the left, except with annual data, 20 year running mean, and published by another oil company. Source : Shell/IHS Energy, 2005

In particular, almost no "giant oilfield", that is those who "make a difference" for the world total of the resources, has been discovered since 1980.

Breakdown of the cumulated discoveries, for the whole world except US and Canada, in billion barrels, depending on the class of size and the year (Mb means million barrels ; 1 barrel = 159 litres). We will note that discoveries still happen for "small" oilfields (blue and red curves go on rising, but almost no "giant oilfield", over 2 billion barrels apiece, has been discovered since 1980, and almost no "large oilfield", between 500 million and 2 billion barrels apiece, has been discovered since 1990. Besides small oilfields are less easy to exploit than large ones. Source : Jean Laherrère, 2003

To illustrate the fact that the "major" discoveries" are now pretty ancient, one can observe that most oil production for the Middle East - which totals 2/3 of the world reserves, and a third of the world production - comes from oilfields discovered a very long time ago.

Saying that the world annual discoveries are inferior to the world annual consumption, is saying that the reevaluation, or even the keeping to a constant level, of the reserves, must come from "something else" than discoveries. In short, reserves did not mainly increase because of new discoveries. We than have to examine the economic conditions, technolgy, and...the rest. Actually technology and economy are a little linked together : if technolgy becomes more performing, the extraction price diminishes a little "everything remaining the same otherwise".

At the beginning of oil production, all we could do was to drill into the oil reservoir, then wait for the oil to kindly spring up to the ground under the pressure of the gas generally associated to oil in the same reservoir. Of course, as soon as the gas pressure becomes too low, nothing comes out any more. Besides, an oil reservoir is not like a fuel tank, with a pretty fluid liquid freely circulating from one end to the other without meeting any obstacle. A better picture of what an oil reservoir is would be pumice stone impregnated with oil (like olive oil !), and then it is easy to understand that it is not necessarily obvious to "suck" the oil through thick layers of more or less porous stone !

Hence proceeding by "stupidly" driving a vertical tube in the ground, like it was done at the beginning of the oil era, does not frequently allow to recover properly the oil in all the parts of the reservoir. Since that time the drilling and extracting techniques have considerably improved : it is possible to drill "sideways", horizontally, with several branches....(a couple of examples below).

It is also possible to inject water, vapour, or pressurized gas, in the reservoir to extract a larger fraction of the initial oil. The "recovery factor" (which is really known only when the oilfield is abandonned) can vastly vary from one oilfield to another, with a median value located around 35% (graph below).

Recovery factor observed or supposed (vertical axis) for 3300 oilfields in the word, depending on the size of the reservoir (that is the amount of oil physically underground) expressed in million barrels (horizontal axis). 1 barrel = 159 litres. Source Jean Laherrère, Petroconsultants, 1997

Even without discovering new oilfields, an increase of this recovery factor for the existing resources increases mechanically the proven reserves, and oil companies willingly explain that this recovery factor went from 25% to 35% during the last 30 years, what generates, even with no new oilfield discoveries, an increase of almost 50% of the proven reserves. There is of course a limit to the possible reevaluation that proceeds from the technical progress, because recovery factors do not only depend on the methods used, but also - and mainly, say the geologists - on the physical characteristics of the reservoirs (size and shape of the pores, for example) and of the oil (viscosity for example).

The factor presently observed is only 3% for some reservoirs said to be "compact fracturated", where the oil circulates with great difficulties, and technical progress won't enable to increase this figure a lot, but can jump to 80% when the rock that contains the oil is very porous, and the oil itself pretty fluid, as in Lybia or Canada.

Besides some studies suggest that technical improvements do not allow to significantly increase the recoverable fraction of a reservoir, but mainly to extract it faster. As estimates of the resources - and therefore of the reserves - are sometimes based on the flow coming out of the wells already drilled (and the estimates of what lies underground increase with this flow, of course), one understands better that this little discussion is not without importance !

We will find here another reason to the "growth" of the resserves : the price of the barrel has significantly increased since the first oil shock, making worthy the extraction of resources that were too costy before (deep water for example).

Price per barrel in dollars of the day (red) and 2001 dollars (blue) since 1860. Source : BP Statistical Review, 2007

This being said, the observation of historical trends show that the doubling of the price of the barrel increased the reserves by a couple % only for conventional oil. Economists use to say that the amount of reserves is has a very low elasticity to the price of the barrel, contrary to what can be witnessed for other mineral resources. There is a very good reason to have this : for metal ores, when the metal price rises, it becomes possible to exploit ores that have lower grades, even if it requires an additional energy spending, and therefore reserves increase (it is notably the case for uranium). There is of course a limit, but as the product extracted is not among the inputs required to allow the extraction (among which energy), the limit will be economical before being physical : the price of the inputs - partially driven by their scarcity, of course - must remain below the price of the output, and there will always remain something under the ground (it might be very small, though !) when extraction stops because it is not profitable any more.

For liquid or gaseous hydrocarbons, and more generally and energetic resource, there is also a purely physical limit (that can be reached or not before the economical limit is reached) : as energy is used to extract... energy, the physical limit is reached as soon as it requires more than 1 kWh (or toe or MBTU or whatever) to extract and refine 1 kWh (or toe or MBTU or whatever) of oil (or gas or coal).

This physical limit drives the fraction of the oil that will be ultimately extracted from a conventional oil reservoir (extraction becomes physically harder when the residual amount of oil decreases) : when it is necessary to spend more energy than what is enclosed in the amount extracted, there is no "primary energy reserve" any more.

If an increase of the price of the barrel has a marginal effect on the short term increase of the volume of the reserves, the lasting of such a price increase will nevertheless turn profitable the extraction of resources other than conventional oil (extra-heavy oil, part of the shale oil, etc), than can be used to supply the same final energy products than conventional oil, but have higher operational costs. Part of these resources wil never be "energetically profitable" (it is notably the case for most of the shales) but for others it is just a matter of price.

One should note that for coal, the "energy" limit is almost never reached, because the energy used for extraction is mostly mechanical, and the mechanical energy required to extract a fuel which is solid and already "concentrated" does not represent much of the calorific value of the fuel extracted.

As we have seen at the top of this page, a proven reserve is a simple declaration of the operator, and corresponds to the oil that the company is certain to be able to extract from the reservoirs it owns, given the geologicel, technical and economical data available. It is thus a sub-category of the physical resources, and if the proven reserves increase it is that "something else" is decreasing somewhere else.

To understand clearly what is going on, we will have to use a new definition, which is that of the ultimate reserves. They are defined as all the oil that will be extracted from Earth from the beginning to the end of the history of oil. It is therefore the sum of :

what has already been consumed,

what is enclosed in the proven reserves,

what is enclosed in the probable and possible reserves. In other terms, we are talking of the extractible fraction of the oil not discovered yet, or already discovered, considered as not extractible for the time being, but that will become extractible sometime in the future, when technology has progressed or prices have increased, as explained above.

Well during the past decades :

What has already been consumed has increased (most obvious, my dear Watson !)

what is enclosed in the proven reserves has increased, as we have seen,

There are thus only two possible outcomes with this information (sorry, it's maths !) :

either the probable and possible reserves have also increased, which means that estimates regarding all the oil that will eventually get out from Earth have increased,

or these estimates on the ultimate reserves did not increase, and then it means that the estimates on probable and possible reserves have decreased.

It happens that ultimate reserves have long been the subject of regular estimates, and the median value of these estimates, for the last 30 years, has been...stable.

The logical conclusion of all this is what follows : if the proven reserves have increased during the last decades, it is just because probable and possible reserves have decreased much faster than consumption rose.

Strictly based on the knowledge of the ultimate reserves by oil specialists, the perspectives for 2040 were not very different in 1975 than what they are today !

Actually, in spite of what is written above, that is technical and economical conditions always more favourable, it's been 10 years now that reserves, expressed in years of consumption (for the year of publishing), decrease.

World reserves expressed in years of production (and therefore of consumption, roughly) since 1975, according to BP Statistical Review, june 2002

There is something normal there : production increases, when no much oil is dicovered, and when the recoverable fraction of the reservoirs is not increasing rapidly. Reserves, expressed in years of production, therefore decrease. Actually, the interesting question is not that much on reserves than it is on production. Indeed, it is this production that, on the short run, allows to satisfy a demand. And the good question is : may we then have an idea of the date when the decline of the production will begin, not for conjunctural reasons, because a war has damaged a pipeline here, or because there is a transcient little recession there, but for structural reasons, because there isn't enough left underground to satisfy the demand ?

There are of course two other possibilities :

that we limitate voluntarily our oil consumption because of the climate change that derives from its use,

that humanity starts its decrease (that will eventually happen, the only question being to know when !) before the oil reserves get depleted or before climate change has major consequences, because another limiting factor will have produced its effects.

However for the rest of this page we will assume that the will to mitigate climate change will not be the first limitating factor for the consumption, no more than any other cause.

This topic of the "production peak" is the ground for a large discussion on figures between experts, but the important information is finally this one : between the optimistics and the pessimistics, there are today only 15 to 20 years. In other words, when the pessimistics see the peak within 5 to 10 years from today, the optimistics see it "only" within 20 to 30 years.

The "pessimistics" base their views on the model exposed below, that allowed a perfect prediction for the United States, and that can be stated as follows : the production curve reproduces, with a constant shift (35 years in the case of the USA), the discovery curve. After all, it is understandable : before producing oil, it has to be discovered !

In green : curve of the discoveries in the USA, in billion barrels, smoothed (5 year mobile average). In black, gaussian curve reproducing approximatively the rythm of discoveries. In red, annual oil production in the USA, in billion barrels. In thin red, the black gaussian curve shifted by 35 years. We observe that it almost perfectly fits the production curve. In other words, a good approximation of the production curve can be obtained by shifting the gaussian modelling the discoveries, what allows to "foresee" the production from the discoveries, and therefore the production peak when there is a good visibility over the discoveries. Source : Jean Laherrère, 2003

The first man to have noted this correspondance was a american geologist, King Hubbert, who, basing himself on this observation, predicted in 1959 that the production in the US would cuminate during the early 70's then would decrease thereafter (because of this story the production peak is often called the "Hubbert peak"). It turned out that he had made a perfect guess, as the red curve above shows. May we then extrapolate this reasoning to the whole world ?

In green : curve of the annual discoveries in the world, in billion barrels, and in blue, the same curve smoothed by using a 5 year mobile average. We also recognize the general shape of a gaussian curve in these discoveries. The small "rebound" after 1990 corresponds to deepwater discoveries. In red, annual oil production of the world, in billion barrels. We note, just as for the US, that it more or less reproduces, with a shift of a couple decades, the curve of the discoveries. Source : Jean Lahérère, 2003

What the above graph says it that if the "physical" data on the discoveries is reliable (which is the main question, see discussion below), and if the model is applicable, then we are not very far from the production peak, hence from the beginning of the structural decrease of the production. Actually, if we prolongate the trend for the reserves as they are declared by the operators, we go towards an increase, but if we base the future on "technical" data on the resources, published or estimated by geologists or experts (such as Petroconsultants), then the decreasing trend observed for the past 10 years should go on.

Evolution since 1950 of the reserves, expressed in billion barrels remaining to produce, depending on the sources. The "remaining to produce" reserves represent a combination of proved, probable and possible reserves (see first box of the page), with relative weights that vary depending on the organism that publishes the data. The "technical backdated mean" curve corresponds to an estimate made by the author of the graph. "Backdated" means that what we consider being able to extract from a given oil reservoir is dated in totality with the year the first well is drilled in the said reservoir. If several wells have been drilled, and that each of them has been followed by an estimate of reserves associated to the well, the "backdated" method reaffects the integrality of the "reserves by well", that are published along with the drilling of the wells, to the year of drilling of the first well. It has the aim to affect to each reservoir discovery the date of drilling of the first well in this reservoir. This curve has already begun to decrease, contrary to the curve of the proven reserves published in the specialized magazines of the profession. How to interprete this ? As the discovered reservoirs all experienced at least one drilling (!), the reserves calculated the "normal" way may increase while the "technical backdated mean" decrease if the new reserves are associated to new wells that are drilled in existing reservoirs. What is the right curve ... ? Source : Jean Lahérère, 2003

In short, expressing the reserves in years if constant consumption, as the press is accustomed to do (we have "40 years of oil ahead of us"), leads to the impression that we are prevented from trouble for at least the number of years mentionned, which is triply misleading :

as the consumption is constantly increasing by 1,5% to 2% per year, the reserves expressed in years of rising consumption amount to 30 years instead of 40, roughly,

trouble will begin when the production is not able any more to grow as fast as the demand would like to. Of course, when this happens, any "hiccup" (war in a major producing country, strike, sabotage, downgrading of the reserves of a major operator, etc) will generate a surge in the oil price, going from anything from a little increase to a major oil shock, equivalent to that of 1974 or even greater. If there is no voluntary decrease of the consumption that would go faster than the decrease of the production (because of climate change, for example), this will happen, at the latest, when the world oil production goes through a maximum, the said maximum being a mathematical certainty (this comes from an old theorem on integrals with an upper limit !). This maximum of oil production, sometimes called "peak oil", or "Hubbert Peak", should happen between 2010 and 2035, depending on the sources.

World oil production in million barrels per day. The curve indicates the real production up to 2002, then various predictions depending on the sources : The ASPO curve represents the maximum possible production in the future according to this association. "ASPO" means Association for the Study of Peak Oil, and gathers retired oil geologists and various experts still professionnally active. It organizes an annual meeting discussing the remaining ultimate reserves (you can refer to their web site). "Shell" refers of course to a scenario issued by this oil company ; this curve is also giving the maximum possible production in the future, that considers that the begining of the peak (actually a plateau) should happen around 2025 ; it is also the approximative date mentionned in the 2003 annual report of Total (2020 to be precise), "IEA" means "International Energy Agency", and the curve shown refers to a prediction on demand, not production. What this graph clearly shows is that the "forecast" on demand by the IEA (which is, in a way, the consumer's desire) is not backed by information given by oil companies themselves. Source : Total, 2004

As we will now see below, much before the "end of oil" happens, the remaining reserves will concentrate more and more in a small number of countries, not necessarily very stable, what can lead to a potentially unpleasant situation for the consuming countries.

Actually we have discussed so far just as if there were just ONE reserve, ONE production, and ONE consumption. But all this is dispatched in the world : there are reserves in several places, and consumers in several places also ! And the interesting point is that the reserves - therefore the production - are mainly...elsewhere than were the consumers are. The Middle East, that consumes "only" 6% of the world total, owns the two thirds of what "remains to produce".

Share of the various regions in the world for the proven reserves (at end 2005), the production (in 2005), and the consumption (in 2005). We observe that countries with "low" reserves (Europe, Northern America) still have production shares that are not ridiculous, but that may well change within 10 to 20 years (see below). However these figures do not take into account the non conventional oil reserves, located for a large part in Canada (the figure of 15% of the world reserves is sometimes put forward for Canada if we take non conventional oil into account, that would give this country the second place in the world, after Saudi Arabia). Source : BP Statistical Review, 2007

If one knows, in particular, that Saudi Arabia owns a fourth of the planetary reserves, and that Irak has the second rank in the world, with 10% of the total, the geopolitical situation of this early 21st century becomes pretty obvious ! Indeed, the reserves by zone, expressed in years of present concumption, show that the US, now self sufficient for 40% of their demand (that is producing 40% of their demand), will have to find "elsewhere" the precious liquid in not much time.

Reserves at end 2005 of the various regions of the world, expressed in years of constant 2005 production. The lower the figure, the lower the share in the future world production. So with a constant production (but actually it decreases) Northern America could produce for 12 years only, Europe for 9 years, etc. Before the world runs out of oil, there will mostly be an increased dependancy to Latin America for a small part, and to the Middle East for a large part. Source : BP Statistical Review, 2007

The US have otherwise very small hopes to discover significant new resources, as all land has been explored in all ways (the situation is more controversial for Canada, where there are large resources of bituminous sand, but whose potential are subject to large discussions, because they are uneasy to exploit, and very "dirty" regarding greenhouse gases emissions).

In other words, being presently dependant to "remote" and "unstable" countries (Venezuela, Middle East, and, pardon me, Russia) for 40% only of their present supply, USA would become dependant to the Middle East for 75% of their supply in 15 to 20 years. Isn't there is here an interesting explanation to the suddently acute interests in the suffering of the irakian people, that didn't seem to be such a problem 10 or 15 years ago, when Bush (father) was an open ally of Saddam ? We should pity Pakistanese people, Northern Koreans, inhabitants of various african countries, etc, that have a dictator but no oil !

If we now try to see who is "optimistic" and who is "pessimistic" in this discussion on the reserves, we note something interesting : "pessimistics" are mainly geologists, that is people that deal with the technical field data. "Optimistics" are rather economists, that is people who willingly prolongate the past trends to "predict" the future ("discoveries and technical improvements have always compensated the consumption, so it will remain the same in the future"). Of course pessimistics will eventually prove right, and optimistics wrong. The only point is to know when !

The weakness of the reasoning of the economists is obvious : it is based on a credo that contradicts the fact that the world is finite. But the reasoning of the geologists also suffers from a weakness, not on the ultimate evolution, but on its precise date, because the data on the oil resources that are underground are not always published by the oil companies.

The only published information is the declaration, by every company listed on the New York Stock Exchange, of the proven reserves it owns, that is the oil that the company considers being able to extract without doubt (publishing this information is required by the Securities and Exchange Commission). But the resources, that is oil still underground, are not subject to any obligation of publishing.

Each expert (for example Campbell, Laherrère, Perrodon....to quote those that are the most known in France) works here with his own data set, obtained from his own experiences, and as there is no public reference, you and me are not in the position of easily knowing who is right and who is wrong, apart from the obvious conclusion that oil will have an end. And we must not forget that plenty of oil would have a serious inconvenient : the more oil we have, and the higher the risk of changing our climate increases....

Since there is no more oil, let's turn to gas ! In the first place, let's recall that burning gas also leads to carbon dioxide emissions, that is a greenhouse gas. Calling massively on gas, even before we discuss the problem of the reserves, is a "false good idea" on the long term, just as it is for the other hydrocarbons (that also include coal).

this being said, the reserves for gas have exactely the same defintion as for oil : they don't correspond to what remains in the earth, but to the fraction that we will get out. There is a main difference, though, it's that the "natural springing ratio" for oil, that is the fraction of the oil enclosed in the reservoir that "naturally" springs out under the pressure of the associated gas is roughly of 20% on average, when for gas we are closer to 80%.

Recovery factor of the gas enclosed in the reservoir for gasfields (of gas alone, with no associated oil) larger than 30 bilion cubic feet. It is easy to observe that the recovery factor is much higher than for oil. Source : Jean Laherrère, 1998

It means, practically, that the possibilities of reevaluation of the reserves as a consequence of technical improvements on extraction methods, or of price raises, are marginal : improving the recovery factor so that is goes from 80% to 81% mechanically adds a little more than 1% to the reserves, and that's it. Here, to increase the reserves, we must find new gasfields....or realize that the initial estimates on the size of the already discovered ones were too conservative !

Regarding discoveries, it happens that the situation is pretty close to what it is for oil : the annual discoveries went through aa maximum 30 years ago, and have been decreasing since then.

Annual discoveries for oil and gas since 1900. Source Exxon Mobil, 2002

If we combine these decreasing discoveries with a "spontaneous yield" of gas reservoirs which is already of 75% on average, we realize that gas does not allow this "reserve of optimism for the future" that oil gave us, with a constant reevaluation of the proven reserves in spite of a volume of ultimate reserves remaining about the same. But on the opposite gas shares a major characteristic with oil : the reserves are neither better dispatched, nor more illimited.

Dispatching of the gas reserves by geographical zone. Iran owns more than a third of the reserves of the Middle East, and Quatar owns an amount almost equal. Source : BP Statistical Review, 2007

Should we hope for a miracle making possible the tripling of the known oil and gas reserves ? Such an event would take us, in a way, from Charybde to Scylla : the first problem that the world would gave to face would not be that of the availability of fossil fuels, but of the consequence resulting from their use... Trying to shift as fast as possible from oil to coal is alike : it relieves the constraint on resources, only to increase the burden associated with climate change. No perfect world !

More about the Hubbert Peak (beware, it's pretty technical...) : Oilcrisis.com