ODAC Guest Commentary – Response to Peter Odell

February 29, 2008

Last week in the ODAC Newsletter I presented four background reasons why Professor Odell could write what he did in response to the article ‘The great fuel folly’ by Jeremy Leggett (The Guardian, February 15th and 5th 2008, respectively). The main reason was the existence of two very different data sets for how much oil the world has discovered – the oil industry’s proved and probable (‘2P’) data, and the very misleading proved reserves (‘1P’) data to which Odell has access. (See below for why 1P data can be so treacherous).

The other reasons mentioned were: lack of detailed discussion between the parties, absence of a complete reference source for the topic, and the strength of the standard ‘economic view’ of oil depletion.

Here I examine the remarks by Odell in more detail. Odell’s article has eight paragraphs; I deal with these in turn, citing remarks by Odell, and then giving a response.

Paragraph 1.
Odell: “I am utterly perplexed by Jeremy Leggett’s attempt to foretell the impending demise of the oil industry”.

This sentence has three ideas: ‘utterly perplexed’, ‘Leggett’s attempt to foretell’, and ‘demise of the oil industry’.

a). ‘utterly perplexed’.
It is a serious matter that those who understand peak have not yet managed to explain the idea to those who don’t. For a region producing conventional oil, its resource-limited production peak occurs some while after discovery has begun to fall off, and specifically at the point when decline in production from the large early fields is no longer compensated – because discovery has become so sparse – by bringing on-stream the smaller later fields of the region.

This seems straightforward enough. But if one does not have access to reliable discovery data peaking is counter-intuitive. This is because just before peak there is little sign that the peak is about to occur: production in the region has always been rising, the region has large reserves, exploration is still finding new fields, and technology is raising the recovery factor in the fields already discovered. Without good discovery data, the peak comes as a surprise.

Empirically, the peak frequently occurs when about half of the recoverable oil in a region has been produced. There are now a hundred or so large regions in the world past their resource-limited peak in conventional oil production, so there are many good examples to illustrate the mechanism.

The UK is one such example. UK offshore discovery started with a small find in 1969, but over the next seven years virtually all the major fields were discovered. From this discovery trend it was not difficult to estimate the total amount ever likely to be found. In 1976 the UK government released its first figure for this total, about 35 billion barrels. They then fitted this amount to the production trend and calculated that UK production would not reach its resource-limited peak until around the mid-1990s. The government, however, recognised the implication of such a peak: it would likely herald oil supply difficulties for the UK, as the UK peak was expected to be only a little in advance of the world peak. (See: ‘Oil forecasts, past and present’. R.W. Bentley. Energy Exploration & Exploitation vol. 20, pp 481-492, Multi-Science, 2002. Copyright of this article belongs to the publisher, so it must not be put up on the web.)

Over subsequent years the UK government released revised estimates for this likely total of offshore recoverable oil in its annual ‘Brown Book’. These estimates were given as quite a wide a range, but the simple average of this range – while fluctuating somewhat – has stayed pretty close to the 35 billion barrel figure over this 30 year period. It takes the same value today. Resource-limited peaking of conventional oil in a region is not hard to calculate, provided good data are available.

But in understanding ‘utterly perplexed’, we need to see what the UK proved reserves data indicated over this same period. Since 1986 the UK’s proved reserves have stayed at around 5 billion barrels; roughly five years’ worth of production. So although it was estimated that the UK offshore contained about 35 billion barrels of recoverable oil, the ‘secure’ proved reserves – oil in fields sanctioned by government and close to market – stood each year at only 5 billion. As the years passed, and this secure 5 years’ worth was continually eaten up by production, magically there was still always 5 years’ of reserves left. As the discovery rate by now was low, the standard explanation by the UK government, the IEA and elsewhere was that better knowledge and better technology were generating these new reserves; and that this was a process with no clear end in sight. That few analysts tied together the fairly static UK estimate for total likely offshore oil with this evolution of proved reserves was an error that has cast a long shadow.

The same line of thinking developed in the US – where proved reserves are equally uninformative about how much oil has been discovered – and this led to the widespread acceptance that any forecast assuming a fixed stock of oil was fatally flawed.

But as the UK and many other countries now show (and as Hubbert showed for the US, provided ‘grown’ proved reserves were used) a ‘fixed stock’ model, though approximate and missing some undoubted economic feedbacks, is generally accurate. In essence, the scientists who assumed a fixed stock of oil, and who used the ‘mid-point peaking’ rule to draw a bead on the date of peak, were correct.

Is fairly certain, however, that this abhorrence of a ‘fixed stock’ model is what – in major part – lies behind Odell’s phrase ‘utterly perplexed’.

b). “Leggett’s attempt to foretell”.
As Leggett has made clear in his book ‘Half Gone’ and elsewhere, he does no forecasting himself. He refers to the ‘early toppers’ (those who calculate the peak as now, or soon), and the ‘late toppers’ (those who see peak as being typically at 2030 or beyond). What has changed is that the early toppers used to be a handful of scientists with access to oil industry data and a concern to get the modelling right. But there are now – as Leggett points out – a growing number of senior oil people on record who see the early peak as probable. It is true, as Odell says, that this group does not yet include Exxon nor the US’ EIA (nor the UK’s BERR for that matter), but increasingly a wide range of experts – including all too belatedly the IEA – are now beginning to understand the issue.

c). ‘demise of the oil industry’.
In Leggett’s article it is not fully clear if ‘demise of the oil industry’ refers only to the private oil companies, or to the entire industry including the nationalised oil companies (NOCs). What is known is that the private oil companies, as Leggett points out, while currently producing large profits have, in general, and for some years now, not been able to raise output. LBST, drawing on company annual reports, has produced a good summary graph of this. For output from the whole oil industry, see the forecasts mentioned in last week’s commentary, as well as the information below.

Odell: ” … highly favourable to his solarcentury company.”

As Leggett’s book ‘The Carbon War’ makes clear, his earlier fight was in bringing the scientists’ fears of rapid climate change – as opposed to ‘global warming’ – to public attention, and it was this concern that ultimately led to the setting up of solarcentury.

Despite his academic background in petroleum geology, Leggett came to oil peaking much later. Though aware of the topic for some time, it was Shell’s dramatic reserves revisions that led him to examine the evidence. (There is an irony here: Shell’s Sir Philip Watts was an explorationist, and the disputed oil – mostly as condensate – really had been found. It was SEC rules, however, that did not permit the booking of these as proved discoveries, and hence the trouble in which Shell found itself. For a good description of this see David Strahan’s The Last Oil Shock.) Today Leggett is one of the still relatively few scientists who see the urgency of both climate change and oil depletion.

Odell: “But Leggett seriously understates the ability of the many oil-producing countries to sustain provision of oil and natural gas.”

As mentioned above, Leggett’s concern is based primarily on the detailed forecasts by the ‘early toppers’, listed last week.

Paragraph 3:
Odell: “Countries outside the OECD, together with OECD members Norway, Italy and Austria are now at the forefront of the world’s oil and gas industries.

Gas is a separate issue (but see the forthcoming LBST study on this as well as other consultancy reports; the gas picture is far from rosy).

Certainly OPEC still has a lot of oil, but the oil data from Norway, Italy and Austria tell a familiar, dismal story. Norway peaked in 2001, at 3.4 Mb/d. By 2006 production had fallen to 19% below this peak. Italy peaked in 2005 at 0.12 Mb/d; by 2006 production was 5% below peak. Austria peaked over 40 years ago, in 1955 at 0.07 Mb/d; by 2005 production had fallen to 73% below its peak. (And if Austria was a typo. error for Australia, the latter peaked in 2000 at 0.8 Mb/d; by 2006 production was 33% below this peak.)

Odell: “And countries outside OPEC – such as China, India, Brazil and Malaysia – not only have large proven and probable reserves, but are intent on enhancing production at as high a rate as possible.”

The story here also is not good, and Odell really ought to know this.

Though China is not yet at peak, its peak is expected soon: between about 2010 and 2015. But more importantly, China became net importer 15 years ago. Its current imports stand at nearly 4 Mb/d (behind only the US and Japan). These imports are expected to continue to increase. Its gently rising production to peak thus helps to satisfy China’s demand only a little, and helps demand in the rest of the world not at all.

India peaked in 2007, driven primarily by its single very large field of Bombay High. Brazil and Malaysia are both at some distance from their predicted peaks (in 2016 and 2018, respectively, according to Energyfiles), but like China they see rapidly growing demand at home. Brazil has always been an importer, and though the deep offshore fields will continue to grow, her internal demand will grow also. Malaysia’s consumption is well over half her moderate production level of 0.8 Mb/d.

No-one should look to any of the countries Odell lists above for significant help with future oil supply.

Odell: “There is a similar situation in some of the former countries of the Soviet Union, most notably Russia, Kazakhstan and Azerbaijan, .. which have significant potential for increased production.”

Here, likewise, Odell’s views do not reflect what is in the 2P data.

Oil industry 2P discovery data show that oil discovery in Russia peaked many years ago. Russia’s main production peak was back in 1987, at 11.5 Mb/d. Following the Soviet collapse when production fell to 6.1 Mb/d, production has now climbed back to 9.8 Mb/d. Some think it might just exceed the earlier peak; but from the 2P data it is clear that Russia has produced far past the mid-point of its likely total discovery, so no very large future production increases should be expected. Moreover, Russian demand looks to be edging up, so its scope for increasing exports will soon enough be under threat of reversal.

Kazakhstan’s peak is indeed some distance off (2016 according to Energyfiles), and with low consumption its oil will indeed come on to the market. Azerbaijan’s peak, however is next year according to Energyfiles. (You can see Energyfiles’ forecasts of peak from the ‘Depletion Atlas’ map at www.oilshock.com. More recent forecasts can be obtained from the company directly.)

Odell: “Meanwhile all member countries of OPEC continue to increase their annual production.

This statement is true if applied to OPEC as a whole. But it bears examination.

Several large OPEC members are past their resource-limited peaks. Iran peaked over 30 years ago in 1974 at 6.1 Mb/d; in 2006 she produced 72% of this. Kuwait’s production also peaked more than 30 years ago, in 1972 at 3.3 Mb/d; in 2006 it was at 81% of peak. Libya peaked nearly 40 years ago, at 3.4 Mb/d; in 2006 she pumped 55% of this. Indonesia peaked in 1991 at 1.7 Mb/d; in 2006 produced 64% of this.

Several more OPEC countries are expected to peak fairly soon, including Algeria and Nigeria.

In considering the above data it is absolutely essential to realise that these peaking dates cannot be known from production data; and also nor from proved reserves. Iran is a case in point. If you download the statistical data for Iran from the BP website you can see that Iran did indeed have a peak in production in 1974, but also that production has been rising strongly in recent years. So how can one know that the 1974 peak was resource-limited in the sense that future production will not pass this level? Only by examining the oil industry 2P discovery data; by combining these with creaming curves vs. wildcat wells drilled, and by use of geological knowledge about the remaining prospective areas. Odell’s situation – and that of many other analysts – is of only knowing what can be seen from the public-domain data.

In looking at this question it is also important to realise that most of the current detailed models mentioned last week are bottom-up by-field (or by field-group) models. With peak so close, these models no longer need to assume either an estimate for ‘ultimate’, nor that peaking occurs at a ‘mid-point’. These models make estimates for yet-to-find by region, and this is the final, but important, criticism to make of Odell’s piece.

It makes no sense to simply list countries or regions with likely future oil as a counter to the detailed forecasts. This is because such forecasts – as explained above – include quantitative estimates for future oil likely to be found within the forecast timeframe. So what critics of such models – such as Odell (and the UK’s BERR) need to show is not that oil is likely to be found in future – recognised by everyone – but that future discoveries will significantly exceed what is assumed in the models. BERR, for example, recently quoted the large new Brazilian find of Tupi as a proof that peak must be far away. But such a find – as BP’s Miller points out – merely moves oil from the forecasters’ ‘yet-to-find’ category into ‘discovered’.

Thus these comments by BERR, and here by Odell, only serve to highlight the extraordinarily poor understanding of oil peaking among many recognised analysts in the field. Hopefully, what I have written here may raise this level of understanding a little.

Dr. Roger W. Bentley is Visiting Research Fellow, Department of Cybernetics, University of Reading


Tags: Fossil Fuels, Oil