[Graeme]

Busting the Hubbert Myths Part I

Over the next several months, the M King Hubbert Tribute will publish a series of articles that discredit some popular myths surrounding the work of M King Hubbert - the father of peak oil theory – that are beginning to be accepted as fact and thereby confounding the peak oil debate.
The first myth is that Hubbert was stuck on one projection for world oil production.

Peter Jackson of Cambridge Energy Research Associates exclaims: “Despite his valuable contribution, M. King Hubbert’s methodology falls down because it does not consider likely resource growth, application of new technology, basic commercial factors, or the impact of geopolitics on production. His approach does not work in all cases – including on the United States itself – and cannot reliably model a global production outlook.” [1]

Jackson and other optimists contend that peak oil is several decades away.


The “one model” myth is false. Clearly, Hubbert recognized decades ago that world oil production did not fit a strict bell curve, and therefore peak estimates would need to be adjusted over time to account for actual production. The implication is that many different scenarios are indeed possible and they depend on (among other things) the planned production policies of producer nations and changes in world oil demand do to geopolitical circumstances and conservation efforts. Despite that we may not yet have reached world oil peak, conventional oil is limited, may peak within the next decade, and prospective alternatives (e.g., shale oil) will not replace oil. The idea of dismissing peak oil theory because Hubbert’s traditional model said oil would peak in 1995 is without merit. Let’s move the debate forward…

wordpress

energybulletin

[LastViking]

As a principle, the Hubbert Curve is sound when applied to Regular Conventional Oil. But most have missed MK Hubbert's purposeful exclusion of Alaska and non-Conventional oils in his studies. He mentions them ... and then excludes those categories.

His reason, as i interpret it, is based on his understanding that the differing flow rates of the several categories of non-conventionals disqualifies the general hypothesis that increased URR means an increased Peak Rate.

Since 1995, URR has doubled. But the Hubbert gaussian model is not robust enuf to reallocate increased non-conventional URR to the downslope. It mistakenly tries to increase the Peak. As URR approaches 4-Tb today (due to high oil prices), gaussian & parabolic models attempt to increase the Peak to the unprobable magnitude territory of 125 to 145 mbd.

As a predictive tool for the All Liquids profile, a blending of logistics, least squares or parabola is reguired for the Peak Rate and Peak Year with gaussian to define the Decline.

Parabola does a superior tracking of the actual two decade vicininity surrounding the Peak, but is deficient in being unable to flip over to concave for the tail. This results in absurd Decline Rates (exceeding 4%).

[seahorse]

Increasing reserves has not proven to delay the production peak. The US is a prime example and the North Sea also. The US peaked as predicted, despite the fact that US reserves have increased since that time.

[PraiseDoom]

Increasing reserves has not proven to delay the production peak.

Except in Ohio, where increasing reserves created a complete second peak, 70 years after the first. Except in Venezuela, where there are 2 peaks, perhaps 15 years apart. Except in US Natural gas, where production declines were completely reversed and a second peak was achieved in natural gas production perhaps 35 years after the 1st peak. Pesky reserves....doing things as terrible as that! Except in Montana, where recent unconventional production has created both a second peak and LARGER peak than the first. Same for North Dakota if I recall correctly. Did the North Sea have 2 peaks? I think it did, close together though. Russia? Same there I imagine.

Increasing reserves, depending on how, and the size, and when in the discovery process they are developed, can easily delay and CAUSE another peak in production. It is a unique answer per region. Making sweeping generalizations, while amusing, doesn't usually work in this particular area. Hubbert predicting that world crude production would never exceed 13 BBO a year anyone?

The US is a prime example and the North Sea also. The US peaked as predicted, despite the fact that US reserves have increased since that time.

The US is one of the few places where Hubberts method has followed his prediction. And it might be the ONLY large producer of crude which followed his model, in the general sense.

Increased reserves have, however, made certain that even in years when discoveries don't outpace consumption, that the "inventory" as it were, and they also make certain that the crowd who insists we're "running out" is, as usual, poorly informed.

[jerry_mcmanus]

I recently transcribed a long and detailed analysis of fossil fuels written by Hubbert in 1976 for an energy textbook. The Hubbert Tribute site graciously agreed to host the article here:

http://mkinghubbert.com/resources/documents/outlook

Hubbert applies three seperate analyses to oil, coal and nat. gas, and he also discusses shale oil, tar sands and heavy oil.

Most interesting to me is the analysis he calls "Cumulative Statistical Data" in which he models three curves: production, discovery, and reserves. Curiously, the discovery data does not appear to be backdated, and he shows only an 11 year gap between discovery and production.

He uses the US as an example in the article, but it would be fascinating to see this analysis applied to world data, if that's possible. I know Laherrere has made the case that reserves reporting is too "political", which I believe is the reason why discovery data is usually backdated.

Cheers,
Jerry

[WebHubbleTelescope]

I recently transcribed a long and detailed analysis of fossil fuels written by Hubbert in 1976 for an energy textbook. The Hubbert Tribute site graciously agreed to host the article here:

http://mkinghubbert.com/resources/documents/outlook

Jerry did yeoman's work in transcribing this article and it contains lots of interesting ways to think about Hubbert's work that the people at CERA can't even begin to comprehend. (or more likely choose not to)

[jerry_mcmanus]

Thanks WHT, I appreciate your support.

I know this is a long shot, but would you happen to know if Laherrere has seen the article? Clearly you and others (Khebab) have the math chops to revisit Hubbert's analysis, but Laherrere (or maybe Campbell) seems to have access to the world data, especially regarding reserves.

Cheers,
Jerry

[WebHubbleTelescope]

Thanks WHT, I appreciate your support.

I know this is a long shot, but would you happen to know if Laherrere has seen the article? Clearly you and others (Khebab) have the math chops to revisit Hubbert's analysis, but Laherrere (or maybe Campbell) seems to have access to the world data, especially regarding reserves.

Cheers,
Jerry

The last I noticed Laherrere is evaluating methane hydrate deposits
http://europe.theoildrum.com/node/3819

Before that he had a post on Arctic reserves:
http://europe.theoildrum.com/node/3666

I asked:

Yes, good to see someone here that essentially produces half the referenced and cited (and high quality) graphs concerning oil depletion.

The one pressing question I always have is how the interpolated and extrapolated smooth lines get drawn on these figures. We all know that the oil production curves tend to use the Logistic as a fitting function, but we don't have a good handle on what most analysts use for discovery curves and creaming curves. In particular I have seen several references to creaming curves being modeled as "hyperbolic" curves yet find little in fundamental analysis to make any kind of connection.

Based on statistical considerations I am convinced that the discovery and creaming curves result from a relatively simple model that I have outlined on TOD. I have a recent post where I make the connection from dispersive discovery to creaming curves here:
http://mobjectivist.blogspot.com/2008/0 ... persive....

In the following figure I apply the Dispersive Discovery function to one of the data sets on your graph. This function is simple to formulate and it produces a finite asymptote which you can use to estimate the "ultimate discoverable" (150 GBoe for NG in the following).

If in fact you use the same formulation for your extrapolation, I would be interested to know.

But I have a feeling you share the same frustration as I based on:

"But the USGS past approach since 2000 has been too optimistic, where only one geologist is simply guessing solely the number of discoveries and the size to be forecast (seventh approximation sheet) and then a Monte Carlo run (50 000 transforms) These wild guesses result into a beautiful distribution that looks real. A better approach should be hoped for, based on the complete past data reviewed by several geologists. "

This sounds like the best model out there is some horribly complex Monte Carlo run that ends up being a SWAG. I do indeed agree that a better approach is needed and think we need a good model to start with.

He responded with

Every time that I plot a creaming curve, I am amazed to see how easy it is to model with several hyperbolas, but this doesn't explain why, except that on earth everything is curved. Linear is just a local effect (horizontal with the bubble, vertical with the mead) being the tangent of a curve. I found the same thing with fractals: it is a curve, so I took the simplest second degree curve : the parabola.

For creaming, hyperbola is the simplest with an asymptote. But the most important is to use several curves because exploration is cyclical. But another important point is to define the boundaries of the area. If the area is too big, it may combine apples and oranges making it difficult to find a natural trend. If the area is too small it will have too little data to find a trend. The best is to select a large Petroleum System which is a natural domain. The Arctic area is an artificial boundary and not a geological one.

Best bet is to ask him next time he posts to TOD.

[jerry_mcmanus]

I spent the afternoon digging up links to some of Laherrere's pdf's and found what I was looking for, or close to it.

Hubbert's analysis shows two graphs, one for cumulative:



and one for annual rates:



Each shows an inflection point where reserves stop growing, a point that is about halfway between the peaks in discovery and production, and which also happens to be the same point at which the discovery and production curves cross in the annual rates graph.

I fond graphs for world data by Laherrere that show cumulative:



and annual rates:



neither of which has a curve for reserves, although the same document had this graph showing "political" vs. backdated reserves on the same scale as the cumulative graph:



A few minutes spent in Photoshop and I was able to transpose the backdated reserves curve (here in blue) on the cumulative logistic models:



Yep, it all checks out. In Laherrere's annual rates graph we can clearly see that the discovery and production curves did in fact cross in about 1980, just as the oil shocks hit, and in the cumulative graph this corresponds perfectly to the inflection point in the backdated reserves, also right about 1980.

Hubberts analysis was spot on. World discovery peaked in about 1965, 15 years before the inflection point, and if the oil shocks had not flattened production the "idealized" production curve would have peaked 15 years after inflection in 1995, just as Hubbert predicted.

Now that we know the inflection point lies halfway between the peaks in the discovery and production curves, and we know that the world inflection point was 15 years after the peak in discovery, we can clearly see in the cumulative graph that the oil shocks pushed the production curve out by about 15 years from what would have been the ideal curve, which puts the peak of world production in 2010.

Here's an interesting question: Is it my imagination or is there also a correlation between the inflection point and the point at which the data in a Linearization "settles down"?



Could this be a new means of determining at what point in time to begin the linear regression?

Cheers,
Jerry

[WebHubbleTelescope]

Now that we know the inflection point lies halfway between the peaks in the discovery and production curves, and we know that the world inflection point was 15 years after the peak in discovery, we can clearly see in the cumulative graph that the oil shocks pushed the production curve out by about 15 years from what would have been the ideal curve, which puts the peak of world production in 2010.

Interesting that a reserve peak can be mathematically proven.

Code: Select all

    Reserve = Cumulative(Discoveries) - Cumulative(Production)


.... given the above relationship, we can pinpoint the reserve peak precisely in time, and at the very least just by eyeballing the two original curves. If the reserve peak occurs when the derivative of the above relationship goes to zero, and dCumulative(x(t))/dt = x(t), then:

Code: Select all

    dReserve/dt = Discoveries - Production = 0


Or when the Discovery curve intersects the Production curve. That is where the Reserve peak occurs.

[jerry_mcmanus]

Interesting that a reserve peak can be mathematically proven.

Thanks for the link, I had a feeling the reaction might be "old news".

I just think it's a shame that Hubbert clearly had advanced his analysis well beyond what most people give him credit for. In four years of reading about peak-oil I don't think I've seen a single graph that had curves for all three factors: production, discovery AND reserves, and yet it's been a full 30 years since Hubbert published these analyses.

In fact, the more I think about it, the more I realize what a discredit it was to Hubbert, and a disservice to the entire peak-oil debate, that Campbell and Leharrere didn't use these more detailed analyses in their 1998 sci. am. article.

It might have been an entirely different picture for the last 10 years if so much of what has been written about Hubberts work, and the future of fossil fuels, wasn't needlessly stuck on those musty old 1956 graphs.

Cheers,
Jerry

[WebHubbleTelescope]

Interesting that a reserve peak can be mathematically proven.

Thanks for the link, I had a feeling the reaction might be "old news".

I just think it's a shame that Hubbert clearly had advanced his analysis well beyond what most people give him credit for. In four years of reading about peak-oil I don't think I've seen a single graph that had curves for all three factors: production, discovery AND reserves, and yet it's been a full 30 years since Hubbert published these analyses.

In fact, the more I think about it, the more I realize what a discredit it was to Hubbert, and a disservice to the entire peak-oil debate, that Campbell and Leharrere didn't use these more detailed analyses in their 1998 sci. am. article.

It might have been an entirely different picture for the last 10 years if so much of what has been written about Hubberts work, and the future of fossil fuels, wasn't needlessly stuck on those musty old 1956 graphs.

Cheers,
Jerry

I agree with this observation so much. There is a lot of mundane work that can be done just looking at equalities and relationships. I am afraid that empiricism tends to carry way too much weight. It's almost as if they think that oil depletion analysis is a branch of economics, where all you really have are empirical observations.

[jerry_mcmanus]

It's almost as if they think that oil depletion analysis is a branch of economics, where all you really have are empirical observations.

As opposed to deductive reasoning? I can't tell if you're being sarcastic or not.

I'm reading Living Within Limits by Garrett Hardin and he has a chapter on Hubbert in which he concludes:

"In the days when scholars thought that logic alone was enough for the discovery of truth, a postulation like Keynes's [that Jevons conclusions about coal were a result of his personal prediliction for hoarding] merited condemnation as an argumentum ad hominem -- an argument against the man who advances a view, rather than an argument addressed to the facts. But Freud has taught us to mitigate our logical purity: justified or not, critics' opinions of a man's doctrines are shaded by their evaluation of his personality. Different critics, different evaluations, different judgments.

"Pessimists are not given an easy time in this world. Prophets are subject to a double standard: optimists are permitted many mistakes; pessimists, none. One well-publicized mistake-and it may not even be a large one-and a doom sayer's words are heavily discounted from then on. People hunger for pleasant truths. This is understandable: but should this hunger be encouraged? If there must be a failure in prophecy, which is the more dangerous: the optimistic prophecy that is refuted by events, or the pessimistic prophecy that blessedly proves false? What is the true path of prudence? (That, however, may be a poor appeal to make: when was the last time you heard the word prudence used in public? In today's world many people are embarrassed to claim this virtue: Why?)

"Abandoning psychologism, can the frightening implications of Hubbert's pimple be escaped? This is the question that is now before the house."

Cheers,
Jerry

[WebHubbleTelescope]

It's almost as if they think that oil depletion analysis is a branch of economics, where all you really have are empirical observations.

As opposed to deductive reasoning? I can't tell if you're being sarcastic or not.

I "learned" about economics in college. An instructor actually wrote a long equation on the blackboard (describing some macro-economic law that I can't remember), and then promptly erased it, stating "That's the last equation you will ever see in this class". So yes, I am totally disillusioned with the rigors of economics as a quantitative science. OTOH, oil depletion analysis has the potential to become a rigorous science, if it hasn't already been tainted beyond redemption by profiteers making the data completely opaque to the lay-person.

[jerry_mcmanus]

I "learned" about economics in college. An instructor actually wrote a long equation on the blackboard (describing some macro-economic law that I can't remember), and then promptly erased it, stating "That's the last equation you will ever see in this class"

That's funny, I was just skimming through Herman Daly's Steady State Economics and he had a great snark along the lines that classical economics is an edifce of increasingly complex equations piled atop a "thin concrete" of actual observations.

Apologies for the confusion, but I thought that mathematical modeling of future trends based on observed statistical data is about as empirical as it gets, then it occured to me that most of the current depletion modeling, with the notable exception of your shock/dispersive model, is completely dependent on shot-in-the-dark SWAG's as to URR.

Cheers,
Jerry

[WebHubbleTelescope]

Apologies for the confusion, but I thought that mathematical modeling of future trends based on observed statistical data is about as empirical as it gets, then it occured to me that most of the current depletion modeling, with the notable exception of your shock/dispersive model, is completely dependent on shot-in-the-dark SWAG's as to URR.

Cheers,
Jerry

I think the problems with economics is why most of the mathematicians have split to econometrics.

"Econometrics is concerned with the tasks of developing and applying quantitative or statistical methods to the study and elucidation of economic principles. Econometrics combines economic theory with statistics to analyze and test economic relationships. Theoretical econometrics considers questions about the statistical properties of estimators and tests, while applied econometrics is concerned with the application of econometric methods to assess economic theories"

And this gets very deep and obscure very quickly. Fertile ground for Nobel prizes based on recent selections.

Like you say, the statistical predictors are strongly supported, but without a firm model basis, it turns into a crap-shoot.

[Veritas]

Greetings,

I was curious if the idea of oil production following a bell shaped curve has been disproved, as a colleague recently told me that it had been repeatedly shown that it is inaccurate.

Neither of us disputes that oil production will peak, and that it will happen sometime soon if it has not happened already. Where we disagree is whether the peak coincides with half of all oil being produced, and whether the right half of the curve will be a steep bell-shape or an undulating plateau.

While there are certainly repercussions in either case, if it is a bell-shape the implications are far more severe than if its a long slow decline.

This is important to know when people claim peak oil theory is false - for no thinking person can claim that oil production will never peak, but very intelligent people could argue that the peak will follow by very slow decline rather than a sharp decline. Is there any conclusive evidence one way or the other?

And if this is already covered in threads here I apologize, a link to said discussion would be great!

Thanks!

[Twilight]

The bell curve is employed as a mathematical model of the exploitation of a finite resource. It is just calculus. The real world does not perfectly follow a few lines of maths. Its greater complexity does not invalidate the principle however, namely that there will be a right hand side of the curve with a negative gradient.

I would object to any suggestion that the question of linear decline vs. exponential decline is a deal-breaker. The key word is decline, so immediately the difference is one of variety. And I am not sure that a linear decline is better than exponential. Exponential decline removes a fixed fraction in a fixed interval, which is bad at first, but at lower levels of production removes ever smaller quantities of production. Linear decline removes a fixed quantity in a fixed interval, which as time goes on, becomes proportionally more brutal. If your mitigation strategies involve long lead times, as ours surely would, you would prefer exponential decline.

[Twilight]

http://www.peakoil.com/fortopic35082.html

Cheers. Ninth post down with the grey graphs is the one.

[Twilight]

Here is an ideal illustrative model of how production lags discovery.

In reality it looks less tidy.

But the principle is the same - you have seen the dress rehearsal up to the final act, so you can roughly guess at what point you are halfway through the play on the one night that matters.