Extreme positions are not succeeded by moderate ones, but by contrary extreme positions
Last week I took the view that The Oil Situation Is Really Bad as we look out 5, 10 or 20 years from now. My article was prompted by the whistleblowers story published by the UK newspaper the Guardian in which current or former anonymous International Energy Agency (IEA) employees asserted that the agency is covering up the precarious oil situation to appease the Americans and prevent panic in the oil markets.
These warnings are plausible to anyone familiar with the current state of world oil production. Once the current spare capacity—I believe it is approximately 4 million barrels-per-day—is worked off after demand rises at some unknown point in the future, it is hard to see how oil production can rise much thereafter.
This week the Guardian followed up on their initial story by highlighting the far more pessimistic Uppsala University study The Peak of the Oil Age by Kjell Aleklett, Mikael Höök, Kristofer Jakobsson, Michael Lardelli, Simon Snowden, and Bengt Söderbergh. (Also see the Guardian’s Peak Oil: what does the data say?)
The Guardian stories compare and contrast the 2008 IEA and Uppsala studies, thus creating for its readers a stark choice between them (Figure 1).
Figure 1 — The views of the IEA (top) and Uppsala (bottom). The IEA sees oil production rising smoothly to about 104 million barrels-per-day in 2030. In contrast, Uppsala sees oil production falling below 80 million barrels-per-day immediately and staying there throughout the forecast period. Oil production in their scenario ends up at ~75 million barrels-per-day in 2030. (I don’t know why the Uppsala graph starts out at a slightly lower initial production level than the IEA graph. Look at Figure 4 below for a closer look.)
In my view, the Uppsala study is unduly pessimistic, implying an immediate crisis (in 2010 and thereafter) which is not in accord with reasonable expectations about future production levels both within OPEC and outside the cartel. In alerting the public to the peak oil issue, the Guardian is doing good work. But not knowing any bettter, they picked the wrong study in my view. The false choice the Guardian offers us between the IEA and Uppsala amounts to a kind of all or nothing proposition.
It is not the barrels-per-day figure for 2030 arrived at by the Uppsala team that I find objectionable. In fact, I think that ~75 million barrels-per-day in 2030 is probably too optimistic! However, I must remind you (and myself) of a fundamental principle about forecasts, including oil production forecasts—
- The further forward in time a forecast goes, the more worthless that forecast becomes.
We can not believe any number anybody gives us for 2030. The converse in time is also true—a study should be reasonably accurate within shorter time frames for it to have any value whatsoever. This latter consideration depends on the subject under study. For example, interest rates can change quickly (in a matter of days or weeks) whereas global oil production can not under normal circumstances (outside big, exogenous shocks to the system, e.g. a devastating act of terrorism or war).
The Uppsala reference forecast calls for global oil production to drop to a very low level in 2010 (approximately 77 million barrels-per-day, Figure 1, Figure 4) and decline over much of the next decade. I don’t believe that such a sharp drop-off in world productive capacity is possible next year or for some years thereafter. I do believe that Uppsala’s short-term views, in alignment with other catastrophic forecasts, are used by naive observers to support End of the World prophesies I don’t subscribe to.
My view does not reflect that of ASPO-USA, which is easy to say in so far as ASPO-USA does not have an official view. Also, I do not mean to single out Uppsala. There is no dearth of alarmist forecasts or Chicken Little predictions for true believers to choose from. The chances that oil will hit $200/barrel in 2010, as some have predicted, are vanishingly small, absent a geopolitical shock. I picked the Uppsala study because the Guardian featured it.
And one more caveat: do not mistake demand for oil with oil production capacity. This distinction seems simple enough, but in my experience people sometimes confuse these issues. Suppose, for instance, that the global economy Crashed Big Time, and oil consumption fell to 79 million barrels-per-day. That would certainly have a devastating effect on future oil production via low prices & investment, but it would not mean that the world oil production capacity had fallen to 79 million barrels-per-day.
Methodology & Uncertainty in the Uppsala Study
The goals of the 2008 Uppsala study are to overturn the IEA forecast and offer an alternative. These are good goals. To accomplish this, the Uppsala team employed the methodology described in #1-#3 below for the crucial contribution from yet-to-be developed oil fields. This is a bit technical, so bear with me.
- They “unpacked” (digitized) the IEA oil production graphs to get annual oil production numbers for yet-to-be-developed fields. This process yields the production curve shown in Figure 3 below. The Uppsala team also revised downward the IEA’s forecasts for gas liquids, unconventional (tar sands) production and enhanced oil recovery (EOR). I will not be dealing with those topics today, but I don’t discount their importance in the Uppsala forecast.
- They applied the depletion-rate-of-recoverable-resources production parameter shown below to the IEA data.
In plain language, the depletion rate dδt measures how much you are depleting an oil field (or a set of fields, a region) with known ultimately recoverable reserves R0 (URR, in billions of barrels Gb) as it is produced over time t (e.g. 3% per year).
- Historical production in various countries or regions—the North Sea is their main example—shows that depletion rates never exceed 7%. But IEA projections for yet-to-be-developed fields call for depletion rates well in excess of 7%, i.e. the production rates are much higher than historical experience would suggest is possible given the URR, which the IEA puts at 257 Gb. See Figure 2 below.
Figure 2 — Uppsala’s Figure 5. Look at the caption in the figure for the explanation of the graph as explained in my #3 above. The IEA assumes unrealistically high depletion rates compared to the North Sea, which already had the highest depletion rate in the historical cases the Uppsala team examined.
I have left out some theoretical niceties they discuss, but another crucial observation in the Uppsala model is that depletion rates rise until peak production occurs and then flatten out. (You can see some flattening of the normalized North Sea depletion rate in Figure 2 at the end, roughly coincident with its production peak year 2000 in Uppsala’s Figure 4, which is not shown here.)
Based on historical experience, the Uppsala team sets flattened, post-peak depletion rates at 2-3% and close to 6% for non-OPEC offshore production for yet-to-be-developed fields. This contrasts sharply with the much higher rates assumed by the IEA in Figure 2. The revised depletion rates for allow them to conclude that future production from these fields will fit the profile shown in Figure 3 below.
Figure 3 — As above, read the Uppsala Figure 7 caption. Fields-to-be-developed will never attain a peak production rate of 29.33 million barrels-per-day as predicted by the IEA. Instead, production will climb to a lesser peak at a much slower growth rate as shown. I have put in an alternative Why not this? (in red) scenario for contrast.
This sums up the Uppsala team’s essential argument for future production from yet-to-be-developed fields. As theoretical arguments go, I’ve seen worse. The Uppsala team’s argument is sound. For yet-to-be-developed fields, it is clear that the IEA’s production rates are exaggerated to the high side. I have no quarrel with that.
The flaw in the Uppsala argument is their assumption that the volume of oil that will ultimately be recovered is fixed. Regrettably, the IEA specifies that there are exactly 257 Gb of new reserves in fields-to-be-developed, so the Uppsala team runs with that number.
Arguments based upon fixed ultimate recoverable reserves (URR) numbers have a very low probability of being correct. As former BP executive Jeremy Gilbert told the recent ASPO-USA conference, the key words in reserve/supply forecasting are uncertainty, uncertainty, and uncertainty.
Although it is not all it’s cracked up to be, reserves growth does occur. This is especially true in new oil developments in which reservoir trends are further delineated and new, associated oil pools are discovered. If you base your result on a moving target, as Uppsala does, you must prepared to revise production levels up if the recoverable reserves grow over time.
Major discoveries occur too, albeit rarely nowadays. Brazil will no doubt make a large upward revision in their proved reserves as they delimit the potential of Tupi and the other pre-Salt ultra-deepwater fields in the Santos basin. This revision may be in excess of 65 Gb as detailed in Rose Anne Franco’s presentation to the recent ASPO-USA conference. Brazil has ambitious plans to produce 1.8 million barrels-per-day from the pre-Salt fields by 2020. Petrobras is a serious organization that is very good at what they do. Again, this would require a revision to Uppsala’s projections. Even in the North Sea, Euan Mearns reports that—
There are signs that the decline trajectory has already been influenced
by a third cycle of giant field development with the Buzzard oil field
and Ormen Lange gas field both coming on stream in 2007.
Although these new hydrocarbons will only lift slightly the tail end of the declining North Sea production curve, it does illustrate the hazards of holding URR fixed, even in mature oil production regions. Oil production is, at minimum, a function of geology, price, technology and investment. Uppsala’s model takes only geology into account, and incorrectly fixes that variable because it neglects the other factors determining oil production.
Beyond keeping recoverable reserves numbers constant, I have more serious concerns about Uppsala’s reference forecast. Look at Figure 4 for a closer look.
Figure 4 — The burnt-orange line shows Uppsala’s reference case. Oil production falls off a cliff. Production (capacity?) is only 77 million barrels-per-day by the end of 2010. The fast and slow cases illustrate their two scenarios for production of yet-to-be-developed fields, but neither scenario boosts the total number of barrels produced as implied in Figure 3 (faster production sooner results in lower production later) because the URR is fixed. I have staked out the middle ground in my moderated forecast (red line). Oil as defined by Uppsala and the IEA in keeping with Figure 1 includes gas liquids, tar sands and conventional crude.
If I am to take the Uppsala’s 2008 world oil outlook seriously, I would also have to believe that the world currently has little or no spare oil production capacity. No serious oil analyst believes this, and lately OPEC has been putting more oil onto the market because they fear that prices are too high for a weakened global economy to handle. OPEC has lost control of the situation, however, because the continuing devaluation of the dollar has overwhelmed the supply & demand fundamentals in the oil markets.
As to the crucial question of Saudi capacity and reserves, both Ray Leonard and Sadad al-Husseini believe that Saudi capabilities are not overstated, at least in the medium term. Leonard is a world class geologist who presented his views to the recent ASPO-USA conference. Leonard also believes that Russian production can be maintained at about 10 million barrels-per-day for quite some time to come. Sadad al-Husseini worked at every level of Saudi Aramco for decades, and recently gave his views to ASPO-USA. Both oil industry veterans strongly support a peak oil scenario, but not in the extreme form presented by Uppsala.
Figure 5 shows al-Husseini’s supply & demand forecast.
Figure 5 — al-Husseini’s forecast for supply & demand out to 2020. This scenario depends on sufficient investment in production capacity, so it may viewed as an assessment of production potential based on al-Husseini’s assessment of recoverable reserves. This view is similar to my red line alternative in Figure 4.
If my choice is between Leonard and al-Husseini on the one hand, and a well-intentioned group of (mostly) Swedish academics on the other, I’m going to go with the seasoned industry professionals every time. That’s just the way it is.
Staking Out The Middle Ground
The oil situation is very bad. How bad is it? The Uppsala team’s forecast posits a nightmarish scenario as early as next year. This position is not supported by theoretical considerations or by current realities. Recoverable reserves uncertainty undercuts their arguments. If “fast” production of yet-to-be-developed fields is accompanied by reserves growth in those fields, the production might follow the red line I showed in Figures 3 & 4. Existing spare capacity belies their short-term forecast for 2010-2012.
Sadad al-Husseini’s forecast represents the middle ground on peak oil. I endorse his position, which represents a kind of “best case” which is still not very good.
I also do not believe that serious oil analysts—think PFC Energy, not CERA—would find Uppsala’s forecast credible. And neither do I. Oil production in the years 2010-2012 will reveal who is wrong and who is right. The fact that oil production is mostly a demand-side story over the next few years complicates the situation. And of course there is always the problem of figuring out what the Saudis will do.
As I said last week, being taken seriously by the world-at-large has been an enduring, insurmountable problem for those concerned about the problems created by peak oil. It’s Cassandra’s curse. Still, our being marginalized is not helped when untenable forecasts are presented to the public.
The Guardian unwittingly does the public a disservice when they present a choice between the IEA, who are very likely wrong on the high side, and the Uppsala team, who are very likely wrong on the low side. Highlighting one study and not other less extreme views implies that the peak oil group is monolithic, which is to say that everyone worried about peak oil speaks with one voice. Nothing could be further from the truth.
The Uppsala study does not represent my views. Many of those who attended ASPO-USA’s recent conference also would agree that the situation is not so dire. That is not to say that the situation is not very serious as we look out 5 or 10 years—it is. The goals of the Uppsala study are admirable, but I think their results miss the mark.
Contact the author at email@example.com