Could you clarify this? If I am not mistaken, commercial US crude oil stocks were around 20 million barrels higher in the Eighties in some years. Are you including the SPR?
Yes.
In fact, I just posted the following at The Oil Drum, which represents the dates of the 10 highest levels of U.S. crude stocks this century (listed in chronological order).
There were some higher levels many years ago, but there were also more refineries years ago. So what the absolute levels were 20 years ago don’t really tell you whether they were higher or lower than normal, and level alone is not the entire story anyway. You need a level and a trend. When the Saudis cut they had a high level and it was trending higher. So while price may have given them a motive to keep production steady (at least until it started to fall), the inventories denied them the opportunity.
I think that the approximately two-thirds increase in average world crude oil prices after 5/05, relative to the 20 months prior to 5/05, versus the cumulative shortfall in crude oil production relative to 5/05, is telling us what is going on in world oil markets.
But does the fact that oil prices have fallen well off their highs, despite the Saudi cuts, tell you anything? It tells me that the world would have been oversupplied had they not made the cuts. Which is precisely why I believe they made them.
Think about this. Had the Saudis not made the cuts, where would all of the oil have gone? It would have plummeted the price. I don't see any other option.
While I was at lunch, I realized what you meant when you said "This Century."
As best that I can tell, the all time record high US commercial crude oil inventory was 372 million barrels, for the week ending 2/25/83, when the crude oil throughput was 10.9 mbpd, a Days Supply of 34 days.
For 6/16/06, we had 347 million barrels (7% less than 2/25/83), versus crude oil throughput of 16 mbpd, for a Days Supply of 22 days (a 35% reduction relative to Days Supply for the week ending 2/25/83).
But in neither case does it tell us what was going on in areas like Africa.
Also, what about the theory of refiners going to a Just In Time inventory plan, because of the SPR?
Think about this. Had the Saudis not made the cuts, where would all of the oil have gone? It would have plummeted the price. I don't see any other option.
The question really is, what would have happened if world oil production kept increasing in 6/05, instead of declining. The price spiked followed the production decline, starting in 6/05.
If oil production had kept increasing, I think that Yergin would have been correct, oil prices probably would have stayed at about $38 per barrel, which was the average monthly Brent spot crude oil price in the 20 months preceding 5/05.
Instead, as Deffeyes predicted, world crude oil production has fallen, relative to 5/05, and the approximately two-thirds average increase in oil prices following 5/05 was necessary in order to balance reduced supply against reduced demand--much in the same way that my example of "Rich Town" would be well supplied with food while "Poor Town" had to get by on reduced rations.
The problem is that the post-Peak Oil environment is dynamic--not static.
We will see a period of equilibrium, and then production--and especially exports--drop again, leading to another round of bidding, and a new definition of "rich" and "poor."
In any event, note that you are forced to ask a question which is contrary to fact, to-wit, "What if Saudi production had increased," when in fact it declined--as predicted by the HL model.
In any event, note that you are forced to ask a question which is contrary to fact, to-wit, "What if Saudi production had increased," when in fact it declined--as predicted by the HL model.
Well, I didn't ask if they had increased, I asked what if they hadn't decreased. Given that price fell even though they cut production, I think the answer to the question is obvious. Then the question becomes: Was the world just fortunate enough to have Saudi peak just as the market was becoming oversupplied, or did KSA see that the market was oversupplied and make a preemptive move?
But you keep saying "as predicted by the HL model." Yet the HL model didn't predict that it would decline in 2006, did it? 2004 would have been consistent with the HL, 2003 years ago would have been consistent, 2008 would have been consistent with the HL model. So let's put last year's "prediction" in perspective. You could have a span of many years that would work. It just happened to decline last year - and whether it is voluntary or not you have latched onto that as "HL predicted it."
But at least be honest enough with yourself to look at the HL and admit that if production had started declining in 2010, it would have been consistent with the HL and you could still say "as predicted by the HL."
To be clear, I want the readers to understand that the HL did not predict that KSA would decline in 2006. It predicted, according to what you have said, sometime between 50 and 60% Qt. And given that the % Qt has been moving backward for a couple of years, we don't really know how long the prediction was good for. As I have indicated, I looked at Texas and a Texas decline at any point between 1960 and 1977 was consistent with the HL prediction.
In fact - and I think this is a very important point - I bet the last Saudi decline - which was voluntary - was within the range of the HL prediction. I need to check this out, but I bet I am right.
http://www.energybulletin.net/16459.html
Published on 24 May 2006 by GraphOilogy. Archived on 25 May 2006.
Texas and US Lower 48 oil production as a model for Saudi Arabia and the world
by Jeffrey J. Brown & "Khebab"
In summary, based on the HL method and based on our historical models, we believe that Saudi Arabia and the world are now on the verge of irreversible declines in conventional oil production. While there will be massive efforts directed toward unconventional sources of oil, we predict that unconventional sources of oil will only serve to slow and not reverse the decline in total world oil production.
True or false: A Saudi peak in 2003 would have been consistent with the HL?
True or false: A Saudi peak in 2008 would be consistent with the HL?
Both cases are true, and you have inspired me to make this the topic of my next essay: HL's Throughout History. The fact is that your parameters are such that you have a very, very broad range of years in which you could claim a "hit." Why is this so difficult for you to acknowledge? You are really overstating your case every time you say "as predicted by the HL", because "as predicted by the HL" doesn't define a precise time."
This is about like me predicting in 2000 that based on his age, Ronald Reagan was on the verge of dying. Four years later, I say "Just as I predicted." Or, had he died in 2001, "Just as I predicted."
I was typing an edit when you blocked me out, with your response.
What we had in 2005 in Saudi Arabia was a high level of production at a fairly advanced stage of depletion, so I knew that the decline would be past 58% of Qt.
Of the HL plots of large producing regions that I have looked at, Texas showed the latest peak, as a percentage of Qt (57%).
Since Texas was also the prior swing producer and since I could find no examples of large producing regions showing sustained steady increases in production past the 55% to 60% of Qt range (and given Matt Simmons' book), I thought that 2006 was the most likely year for a decline, but I agree that it had to be by 2008.
So, the Lower 48, North Sea, Russia, Mexico and the world fit the 50% of Qt model.
Texas and Saudi Arabia fit the 55% to 60% model. What Saudi Arabia has demonstrated is that we still have no examples of large producing regions showing sustained, steady increases in production past the 60% of Qt mark.
What Saudi Arabia has demonstrated is that we still have no examples of large producing regions showing sustained, steady increases in production past the 60% of Qt mark.
What you will soon be surprised to learn - as I was - is that this claim is not remotely true. Stay tuned. :-)
OK Jeffrey, I am going to do something I should have done a long time ago. I am going to do something that every fan of the HL method should have done a long time ago. I am going to investigate the precision in my next post.
Here is what I would request from you. Define for me the parameters in which you would say "That region has peaked." I presume this is not a problem. I am thinking 50-60% Qt with some defined intercept. But I am going to let you define the parameters, and your reasoning. Then I will proceed with the analysis. Texas and Saudi will be my test cases. My hypothesis is that the precision is going to be very poor.
I think it's quite stupid to hope for the best and do nothing because the precision is bad and the peak could be later. It's like driving a little bit to fast in fog and hope for the best because everything has gone well before.
Although I personally believe we are at or past peak, Robert has stated many times that even if peak is further down the road as he believes, that it is critical that we start preparing now. He has never said, "Relax, be happy, go for a drive."
The Texas/Lower 48 article has the HL data and production data.
Just a suggestion though. Before going to great lengths to continue to attack the HL method (even as recent data support the HL models), you might want to wait for some evidence of rising crude oil production in Saudi Arabia and the world.
I could be wrong, but I am willing to bet that many people on the website--at this point--would be willing to pay us to not further debate this topic, at least until we have data contradicting the HL models.
i'd be quite happy for you to continue this debate (although i might not go as far as paying you to do it :-)
even it doesn't help with Saudi Arabia, it might teach us something about the Russian HL plot for instance.
i'm starting my own HL work to try and get my head around how good it is (or isn't) as a predictive tool. we have some good production and reserves data for several sufficiently large basins here in Oz. i expect that the data will show that HL isn't very useful once you have large new basins coming onstream within a country that already has declining production from other regions. HL could not have anticipated the very recent rise in Australian production, and similarly in Russia now. HL would have predicted Russia to go back into decline for several years already (after post-Soviet recovery). not sure what i'll find as i dig into it..
If I recall correctly, RR said his belief in peak is 2012 +/- 3-4 years putting next year within his guesstimate zone, so why all the bickering... is it just to score debating points or is it to score bragging rights as to who correctly called Peak first?
It seems to be a tad pointless to obsess on pedantic details when all agree the ship is sinking!
This is exactly the reason that I need to evaluate the precision. I think this is what you are failing to grasp. Do the data "support the HL models" because the error bands are so wide that any data over the course of 20 years would have supported the models? That is a key question, especially when you are using these models to tell the world "Saudi has peaked."
I am writing up my results now. They are pretty shocking. All I can say is: HL, RIP. The precision is horrible beyond belief. I will have the essay up ASAP.
Stuart did a bunch of analyses last year about how well curve-fitting techniques such as HL work, when applied to different data sets. I think he used the US as an example. For example, fitting a curve based on 1930-1960, when would you have predicted a peak? How about 1920-1970? Etc.
He came up with all these graphs showing when the peak would be as a function of the two years that defined the beginning and ending of the curve-fitting, and then compared different types of curves to see which produced the most stable peak predictions. I think his preferred method, a quadratic based curve, came out better than Hubble's linear method, which corresponds to the logistic curve.
Unfortunately it would be hard to find the posts right now, but there was plenty of information for someone who wants to look at how accurate Hubble-style predictions of the peak tend to be.
The issue of methodology, predictability and resolution were also evident in FH's skepticism of the Texas and Bakhtiari studies several weeks ago (which have vanished).
This does not apply to RR's KSA upcoming effort, due to its current nature, but the essence wrt to Texas would be that using today's corrected EIA database would not be entirely indicative of what analysts would have deducted in 1969. I have addressed this very recently.
The corrected data of past years has discrepancies of as much as 0.5 mb/d in some provinces and 4 mb/d globally. Using the 2007 database will be an interesting exercise, but only by performing the same task with the *then* available 1970 or so database will it be shown what MK Hubbert and others were viewing at that time.
This will become apparent with RR's test if it (Texas) does not show Hubbert's curve apex if one exists. In 1956, Hubbert's data revealed that the Texas Peak would be 1.15 GB/d (3.15 mb/d) in 1962 assuming a URR of 60 GB. I do not recall if he attempted further calls. Westexas posted in January that the actual Texas peak was 3.5 mb/d in '72.
Okay, here are some of Stuart's posts on the issue. It appears that Khebab was actually the first to use the method of varying the start and end year in looking at HL stability, but I can't find his posts on that right now. Khebab produced this graph, which is very informative if you take the time to study it:
Turns out that contrary to my memory, the HL method is the most stable in terms of predicting things pre-peak. However as you can see in the plot above, the URR estimate varies from about 160 to 240 depending on what years you pick.
As you, I believe that we should stick to the stuff that matters (risk assessment, mitigation) and talk about things we can have a reasonably certainty of.
Thanks for the link. I am working on it now, but I keep getting distracted. I need to stay away from TOD for a while. It takes a lot of data-mining to come up with this information, and when I have to address someone's misconceptions every half hour it doesn't help matters.
I'm sticking my oar in here because I wasn't convinced by Stuart's explanation of the SA production curve. I'm backing RR in this debate. It's too early for the SA peak.
First, demand. The global economy is slowing - witness Alan Greenspan's comments last week on the risk of a US recession in Q4. It takes up to 2 years for rising interest rates to effect the economy according to the Swiss central bank.
Then supply. Not only is oil demand slowing but a number of large projects came on stream last year. I think we are seeing SA cutting production because the demand isn't there, just as they say.
Why try too hard? The Saudi's don't need the money right now. They have seen the world economies cope with $60/bbl so that is likely to be their preferred price. If the futures price is in contango you make more money by deferring production.
IMHO Stuart is reading too much into his SA chart. There are two variables at work - capacity and demand. I don't see how you can tell them apart. Surely, prior to global peak, the oil production curve is a demand curve. There has been spare capacity for most of the last century. If demand had had a different shape (World War III, bird 'flu, asteroid strike, anything) then production would have looked different. And that would blow HL out of the water.
Bottom line - we are looking through SA's swing producer role. We don't know what we are seeing. Moreover, my bet is that SA will produce as little oil as possible from now on.
Does anyone know how I can change my username to DemandSide? ;-)
Thanks for the quicker than expected reply I quickly scanned your two references and will read all three of the postings(Westexas has added more) tonight, but at a quick glance these two things struck me.
From your discussion with Stewart,
Those questions are difficult to answer definitively. Hurricane Katrina really shook up the oil markets by exposing just how little excess capacity presently exists in both oil production and refining capacity. A few short years ago, several million barrels of spare capacity were available. But by the time demand reached 85 million bpd, there was little excess capacity. This makes the markets nervous, and so geopolitical events have a disproportionate impact on oil prices.
About the "85 million bpd with little excess capacity", where was Saudi Arabia as a 'swing producer' at this time?
And in this current article,
Having said that, I do think world oil production will peak around 2012, plus or minus 3 or 4 years.
I am no more competent to say this is any more correct than to say that PO has already occurred, but would you consider qualifying that statement by saying that this small difference in dates is immaterial in considering the necessity to 'get on with the job'. (Not much point worrying who is in front with a cement truck bearing down).
Actually, I do believe Robert has said that it makes no difference whether he is right or WesTexas is right regarding the year of actual PO-- that we need to prepare now as the time difference is essentially meaningless. Robert, correct me if I am wrong.
Cheryl
No, you are absolutely correct. And I am doing my best to prepare as if peak is upon us; I am making changes in my life and trying to set an example for those around me. My main concern - explained above - is that if peak is still 4 years away we are going to completely lose our voice because we called peak now, and the 4 years are going to be lost time.
Yes in your debate you do state that response should be immediate. Sorry about that bad oversight of mine.
But I do not see why, if on this point you and Westexas are in close agreement, a statement to this effect shouldn't be broadcast, with force, while explaining the limits of the PO forecast.
It seems to me that a stumbling block to this forecast is the inability to access the Saudi figures necessary to a quite definite prediction? The Saudis must hold accurate knowledge of their reserves and I would think this information would be available to the US government as well. Oil reserve figures for SA are a US security matter that would be of foremost concern and given the political relationship between the SA and the USA this information would be common knowledge between them.
If the US Government holds this information would it not be an idea to request it directly if only for public relations value for TOD in MSM .
Further,
SamuM (above, I think) gives the following link,which I will repeat here:
I couldn't copy and paste the bit that interested me so will quote it as follows: from the article by Dr. Marcel Shoppers and DR. Neil Murphy. Pleasantly for us lay guys they end with afterthought #!:
2.5% chance the peak occurred by 2003
15% chance the peak will occur by 2006
50% chance the peak will occur by 2009
85% chance the peak will occur by 2012
95% chance the peak will occur by 2015
2.5% chance the peak will occur after 2015
This sort of approach if validated, I think,would stifle the 'wot year is this supozed to hapen?' sort of criticism, while limiting 'Cry Wolf' concerns.
BTW Robert, Still beavering away at all the material you loaded on my plate...thanks a bunch..I think. And if you have anything on how you are preparing for APO I would appreciate hearing about it, but only at your leisure.
re: "But I do not see why, if on this point you and Westexas are in close agreement, a statement to this effect shouldn't be broadcast, with force, while explaining the limits of the PO forecast."
Thank you. I second this (as I've been making similar comments for a while).
I very much like the Shoppers/Murphy way of presenting it. Add what you say above, (summarize the analyses that give rise to one or more predictions) - and this should do it. Something rational and scientific (yes?) upon which to base positive action.
If it is 4 years away, and don't get me wrong, I wish it was 20 yrs away, I believe they will not care if it was called early. They will be saying, "oh sh*t, what do we do now?", and turn to anyone who seemed to know something about it. They certainly won't be listening to CERA.
I just wanted to respond (though no one may be reading by now) - To me, the issue, as I've said above, is to lay out, with care, the parameters and methods of any and all predictions - for all to see.
It should be doable to come up with a statement that summarizes the range of methods and outcomes.
The reason this matters, as I see it, have to do with the following:
1) There's been much discussion about how things might unfold once peak is upon us (or even if it has already begun). The evidence of "peak" per se - ie., "peak"-as-the cause-of-the-outcomes-of-peak, might not be at all evident. We need a clear statement about what we are saying re: world supply of oil, in order to be able to draw connections between events and what we view as the cause (major cause) of those events.
2) Most people (in my experience) have very little of the relevant knowledge that might enable them to grasp the significance of peak. So, in other words, they might not have anything at all resembling the response you imagine. They may not know; they may not be able to know. They may draw the wrong conclusions, for example, the recession is temporary. The Depression is temporary. Things will return to "normal"...just as an example. Hence, it's important to be able to make extremely well-reasoned, well-evidenced "predictions" - or even background statements.
3) Even people with relevant background may not understand how oil functions (or energy for that matter) in the economy. and finally...
4) My experience is that the ability to grasp (let alone respond in a constructive fashion to) the idea of "peak" has almost nothing to do with anything at all - education, etc. There is something quite uniquely difficult on the emotional side about this topic, in other words.
This is why I believe the statement Black B makes (I have made, also) is important. It does matter. To be able to give a coherent summary matters.
Hirsch et al did this (in a way) in the first Hirsch report, by simply listing the various "calls". We need to take this further.
About the "85 million bpd with little excess capacity", where was Saudi Arabia as a 'swing producer' at this time?
Oh, I think up until they started making the cuts, they were producing all the could. In other words, I do believe they lost their ability as swing producers during the time of high demand. Now, is that because they were peaking, or was it because they failed to properly anticipate demand? Given that they still have lots of projects in the pipeline that will bring new oil online, I would say they failed to properly anticipate demand (and supply, for that matter) and have these projects coming on later than they should have.
this small difference in dates is immaterial in considering the necessity to 'get on with the job'.
I think you are kind of new here, so I will repeat my position. In no way do I think we should relax. We need to get on with the job. If we peak in 2012, we are still going to be in dire straits, because we won't be prepared.
Here is my concern. Many people have gone on record now and said "Saudi has peaked." Some of them are using the HL model to back them up. Some very influential people are picking up on this and repeating it. Therefore, it is important to understand the precision of that model.
Here is the worst case scenario in my mind, and the whole reason I argue these points with Jeffrey. What if production starts to rise next year? He, and others who have been calling peak last year or in 2005 are going to be finished prognosticating. The influential people who repeated these claims won't be taken seriously on this subject.
Yet we may still have 4 or 5 years in which to convince the policy-makers that we have a serious problem, and it is nearly on top of us. But they aren't going to listen, and those 4 or 5 years will be wasted as well. Folks like CERA are going to point to yet another failed prediction of peak, and credibility will slide. But I feel we will need our credibility more than ever in the next few years, and I am determined to fight anything that may chip away at it. If the methodology for calling peak in 2006 is sound, so be it. If not, I want to make sure that is understood.
It is better to have 2 or 3 years to prepare than no years. Maybe only marginally better, but for some people this may be the difference between life and death. I fear that if Jeffrey is wrong because he overstated his case, we are going to end up with no years to prepare because people are going to start viewing these predictions as "end of the world" religious beliefs that continue to be wrong.
Yes, just a bouncing baby meathead who finally has figured out how subthreads work...I think.
What if production starts to rise next year? He, and others who have been calling peak last year or in 2005 are going to be finished prognosticating. The influential people who repeated these claims won't be taken seriously on this subject.
By that logic he is out in right field and and you in left, while the ball is somewhere in the middle. Westtexas takes a risk in loosing credibility while you with caution hold back and risk losing a lot of important years. Each of you strike me as intelligent people but, in your individual rigors, perhaps both going to miss the ball. I would suggest a large net that the two of you hold up for a possibly, maybe, sure 'out' on.P.O. (of course nets don't always succeed...I've been out and about for a mighty long time).
and now to tackle a greased piggy-wigg...the Parent subthread...whatever that ferocious looking creature is.
Westtexas takes a risk in loosing credibility while you with caution hold back and risk losing a lot of important years.
That's just the thing, though. I am not saying we should hold back with respect to preparations. I think we should prepare as if a meteor is bearing down on the planet. But I think the message is going to lose a lot of its urgency if those who forecast peak last year turn out to be wrong as production starts to rise. We may have 4 or 5 years to keep beating the drum: Start preparing for event.
I do like that approach that you copied above where they give probabilities. Those probabilities are not too far from what I would suggest. That approach really emphasizes the urgency without the risk of credibility loss.
Comments can no longer be added to this story.
The contents below are paid advertisements. Their appearance does not imply an endorsement by The Oil Drum.
Could you clarify this? If I am not mistaken, commercial US crude oil stocks were around 20 million barrels higher in the Eighties in some years. Are you including the SPR?
Yes.
There were some higher levels many years ago, but there were also more refineries years ago. So what the absolute levels were 20 years ago don’t really tell you whether they were higher or lower than normal, and level alone is not the entire story anyway. You need a level and a trend. When the Saudis cut they had a high level and it was trending higher. So while price may have given them a motive to keep production steady (at least until it started to fall), the inventories denied them the opportunity.
I think that the approximately two-thirds increase in average world crude oil prices after 5/05, relative to the 20 months prior to 5/05, versus the cumulative shortfall in crude oil production relative to 5/05, is telling us what is going on in world oil markets.
But does the fact that oil prices have fallen well off their highs, despite the Saudi cuts, tell you anything? It tells me that the world would have been oversupplied had they not made the cuts. Which is precisely why I believe they made them.
Think about this. Had the Saudis not made the cuts, where would all of the oil have gone? It would have plummeted the price. I don't see any other option.
While I was at lunch, I realized what you meant when you said "This Century."
As best that I can tell, the all time record high US commercial crude oil inventory was 372 million barrels, for the week ending 2/25/83, when the crude oil throughput was 10.9 mbpd, a Days Supply of 34 days.
For 6/16/06, we had 347 million barrels (7% less than 2/25/83), versus crude oil throughput of 16 mbpd, for a Days Supply of 22 days (a 35% reduction relative to Days Supply for the week ending 2/25/83).
But in neither case does it tell us what was going on in areas like Africa.
Also, what about the theory of refiners going to a Just In Time inventory plan, because of the SPR?
The question really is, what would have happened if world oil production kept increasing in 6/05, instead of declining. The price spiked followed the production decline, starting in 6/05.
If oil production had kept increasing, I think that Yergin would have been correct, oil prices probably would have stayed at about $38 per barrel, which was the average monthly Brent spot crude oil price in the 20 months preceding 5/05.
Instead, as Deffeyes predicted, world crude oil production has fallen, relative to 5/05, and the approximately two-thirds average increase in oil prices following 5/05 was necessary in order to balance reduced supply against reduced demand--much in the same way that my example of "Rich Town" would be well supplied with food while "Poor Town" had to get by on reduced rations.
The problem is that the post-Peak Oil environment is dynamic--not static.
We will see a period of equilibrium, and then production--and especially exports--drop again, leading to another round of bidding, and a new definition of "rich" and "poor."
In any event, note that you are forced to ask a question which is contrary to fact, to-wit, "What if Saudi production had increased," when in fact it declined--as predicted by the HL model.
In any event, note that you are forced to ask a question which is contrary to fact, to-wit, "What if Saudi production had increased," when in fact it declined--as predicted by the HL model.
Well, I didn't ask if they had increased, I asked what if they hadn't decreased. Given that price fell even though they cut production, I think the answer to the question is obvious. Then the question becomes: Was the world just fortunate enough to have Saudi peak just as the market was becoming oversupplied, or did KSA see that the market was oversupplied and make a preemptive move?
But you keep saying "as predicted by the HL model." Yet the HL model didn't predict that it would decline in 2006, did it? 2004 would have been consistent with the HL, 2003 years ago would have been consistent, 2008 would have been consistent with the HL model. So let's put last year's "prediction" in perspective. You could have a span of many years that would work. It just happened to decline last year - and whether it is voluntary or not you have latched onto that as "HL predicted it."
But at least be honest enough with yourself to look at the HL and admit that if production had started declining in 2010, it would have been consistent with the HL and you could still say "as predicted by the HL."
To be clear, I want the readers to understand that the HL did not predict that KSA would decline in 2006. It predicted, according to what you have said, sometime between 50 and 60% Qt. And given that the % Qt has been moving backward for a couple of years, we don't really know how long the prediction was good for. As I have indicated, I looked at Texas and a Texas decline at any point between 1960 and 1977 was consistent with the HL prediction.
In fact - and I think this is a very important point - I bet the last Saudi decline - which was voluntary - was within the range of the HL prediction. I need to check this out, but I bet I am right.
Based on Saudi production data through the end of 2005:
http://static.flickr.com/55/145186318_27a012448e_o.png
That was a complete dodge.
True or false: A Saudi peak in 2003 would have been consistent with the HL?
True or false: A Saudi peak in 2008 would be consistent with the HL?
Both cases are true, and you have inspired me to make this the topic of my next essay: HL's Throughout History. The fact is that your parameters are such that you have a very, very broad range of years in which you could claim a "hit." Why is this so difficult for you to acknowledge? You are really overstating your case every time you say "as predicted by the HL", because "as predicted by the HL" doesn't define a precise time."
This is about like me predicting in 2000 that based on his age, Ronald Reagan was on the verge of dying. Four years later, I say "Just as I predicted." Or, had he died in 2001, "Just as I predicted."
I was typing an edit when you blocked me out, with your response.
What we had in 2005 in Saudi Arabia was a high level of production at a fairly advanced stage of depletion, so I knew that the decline would be past 58% of Qt.
Of the HL plots of large producing regions that I have looked at, Texas showed the latest peak, as a percentage of Qt (57%).
Since Texas was also the prior swing producer and since I could find no examples of large producing regions showing sustained steady increases in production past the 55% to 60% of Qt range (and given Matt Simmons' book), I thought that 2006 was the most likely year for a decline, but I agree that it had to be by 2008.
So, the Lower 48, North Sea, Russia, Mexico and the world fit the 50% of Qt model.
Texas and Saudi Arabia fit the 55% to 60% model. What Saudi Arabia has demonstrated is that we still have no examples of large producing regions showing sustained, steady increases in production past the 60% of Qt mark.
What Saudi Arabia has demonstrated is that we still have no examples of large producing regions showing sustained, steady increases in production past the 60% of Qt mark.
What you will soon be surprised to learn - as I was - is that this claim is not remotely true. Stay tuned. :-)
OK Jeffrey, I am going to do something I should have done a long time ago. I am going to do something that every fan of the HL method should have done a long time ago. I am going to investigate the precision in my next post.
Here is what I would request from you. Define for me the parameters in which you would say "That region has peaked." I presume this is not a problem. I am thinking 50-60% Qt with some defined intercept. But I am going to let you define the parameters, and your reasoning. Then I will proceed with the analysis. Texas and Saudi will be my test cases. My hypothesis is that the precision is going to be very poor.
I think it's quite stupid to hope for the best and do nothing because the precision is bad and the peak could be later. It's like driving a little bit to fast in fog and hope for the best because everything has gone well before.
Although I personally believe we are at or past peak, Robert has stated many times that even if peak is further down the road as he believes, that it is critical that we start preparing now. He has never said, "Relax, be happy, go for a drive."
The Texas/Lower 48 article has the HL data and production data.
Just a suggestion though. Before going to great lengths to continue to attack the HL method (even as recent data support the HL models), you might want to wait for some evidence of rising crude oil production in Saudi Arabia and the world.
I could be wrong, but I am willing to bet that many people on the website--at this point--would be willing to pay us to not further debate this topic, at least until we have data contradicting the HL models.
hey robert & westexas,
i'd be quite happy for you to continue this debate (although i might not go as far as paying you to do it :-)
even it doesn't help with Saudi Arabia, it might teach us something about the Russian HL plot for instance.
i'm starting my own HL work to try and get my head around how good it is (or isn't) as a predictive tool. we have some good production and reserves data for several sufficiently large basins here in Oz. i expect that the data will show that HL isn't very useful once you have large new basins coming onstream within a country that already has declining production from other regions. HL could not have anticipated the very recent rise in Australian production, and similarly in Russia now. HL would have predicted Russia to go back into decline for several years already (after post-Soviet recovery). not sure what i'll find as i dig into it..
cheers
phil.
Westexas, all you have to do is STOP.
I just scroll by now anyway.
As I have been saying, IMO it's a question of whether the Titanic sinks in two hours or four hours.
Exactly!
If I recall correctly, RR said his belief in peak is 2012 +/- 3-4 years putting next year within his guesstimate zone, so why all the bickering... is it just to score debating points or is it to score bragging rights as to who correctly called Peak first?
It seems to be a tad pointless to obsess on pedantic details when all agree the ship is sinking!
Thats $0.02 from a dedicated fan of the Oil Drum.
(even as recent data support the HL models)
This is exactly the reason that I need to evaluate the precision. I think this is what you are failing to grasp. Do the data "support the HL models" because the error bands are so wide that any data over the course of 20 years would have supported the models? That is a key question, especially when you are using these models to tell the world "Saudi has peaked."
From my perspective WT & RR are actually very close in their respective positions. RR doesn't refute HL methology, only the precision of the results.
It's like one person saying that that fish is 12.4 inches long and the other person saying it's about a foot, plus or minus an inch or two.
In the end it's still a fish about one foot long.
B.W.
I am writing up my results now. They are pretty shocking. All I can say is: HL, RIP. The precision is horrible beyond belief. I will have the essay up ASAP.
Stuart did a bunch of analyses last year about how well curve-fitting techniques such as HL work, when applied to different data sets. I think he used the US as an example. For example, fitting a curve based on 1930-1960, when would you have predicted a peak? How about 1920-1970? Etc.
He came up with all these graphs showing when the peak would be as a function of the two years that defined the beginning and ending of the curve-fitting, and then compared different types of curves to see which produced the most stable peak predictions. I think his preferred method, a quadratic based curve, came out better than Hubble's linear method, which corresponds to the logistic curve.
Unfortunately it would be hard to find the posts right now, but there was plenty of information for someone who wants to look at how accurate Hubble-style predictions of the peak tend to be.
The issue of methodology, predictability and resolution were also evident in FH's skepticism of the Texas and Bakhtiari studies several weeks ago (which have vanished).
This does not apply to RR's KSA upcoming effort, due to its current nature, but the essence wrt to Texas would be that using today's corrected EIA database would not be entirely indicative of what analysts would have deducted in 1969. I have addressed this very recently.
The corrected data of past years has discrepancies of as much as 0.5 mb/d in some provinces and 4 mb/d globally. Using the 2007 database will be an interesting exercise, but only by performing the same task with the *then* available 1970 or so database will it be shown what MK Hubbert and others were viewing at that time.
This will become apparent with RR's test if it (Texas) does not show Hubbert's curve apex if one exists. In 1956, Hubbert's data revealed that the Texas Peak would be 1.15 GB/d (3.15 mb/d) in 1962 assuming a URR of 60 GB. I do not recall if he attempted further calls. Westexas posted in January that the actual Texas peak was 3.5 mb/d in '72.
Okay, here are some of Stuart's posts on the issue. It appears that Khebab was actually the first to use the method of varying the start and end year in looking at HL stability, but I can't find his posts on that right now. Khebab produced this graph, which is very informative if you take the time to study it:
Stuart's final and most informative post:
Projecting US Oil Production
Earlier posts:
Four US Linearizations
Predicting US Production with Gaussians
Linearizing a Gaussian
Turns out that contrary to my memory, the HL method is the most stable in terms of predicting things pre-peak. However as you can see in the plot above, the URR estimate varies from about 160 to 240 depending on what years you pick.
There are a few posts on PeakOil.com:
How Reliable is the Hubbert Linearization Method?
How Reliable is the Hubbert Lin. Method? the world case
I've started an article on wikipedia in order to try to clarify what is an HL and its properties:
http://en.wikipedia.org/wiki/Hubbert_Linearization
RR, great if you do this.
I'm not sure exactly what you were planning, but this may come in handy:
Uncertainty in Peak Oil Timing (Marcel Schoppers)
http://www.cge.uevora.pt/aspo2005/abscom/ASPO2005_Schoppers.ppt
As you, I believe that we should stick to the stuff that matters (risk assessment, mitigation) and talk about things we can have a reasonably certainty of.
Thanks for the link. I am working on it now, but I keep getting distracted. I need to stay away from TOD for a while. It takes a lot of data-mining to come up with this information, and when I have to address someone's misconceptions every half hour it doesn't help matters.
I think the time and effort you spend on this site is a fantastic service to all. Please, don't burn out on it though! Let the minor stuff slide...
I'm sticking my oar in here because I wasn't convinced by Stuart's explanation of the SA production curve. I'm backing RR in this debate. It's too early for the SA peak.
First, demand. The global economy is slowing - witness Alan Greenspan's comments last week on the risk of a US recession in Q4. It takes up to 2 years for rising interest rates to effect the economy according to the Swiss central bank.
Then supply. Not only is oil demand slowing but a number of large projects came on stream last year. I think we are seeing SA cutting production because the demand isn't there, just as they say.
Why try too hard? The Saudi's don't need the money right now. They have seen the world economies cope with $60/bbl so that is likely to be their preferred price. If the futures price is in contango you make more money by deferring production.
IMHO Stuart is reading too much into his SA chart. There are two variables at work - capacity and demand. I don't see how you can tell them apart. Surely, prior to global peak, the oil production curve is a demand curve. There has been spare capacity for most of the last century. If demand had had a different shape (World War III, bird 'flu, asteroid strike, anything) then production would have looked different. And that would blow HL out of the water.
Bottom line - we are looking through SA's swing producer role. We don't know what we are seeing. Moreover, my bet is that SA will produce as little oil as possible from now on.
Does anyone know how I can change my username to DemandSide? ;-)
Hi Robert,
Thanks for the quicker than expected reply I quickly scanned your two references and will read all three of the postings(Westexas has added more) tonight, but at a quick glance these two things struck me.
From your discussion with Stewart,
About the "85 million bpd with little excess capacity", where was Saudi Arabia as a 'swing producer' at this time?
And in this current article,
I am no more competent to say this is any more correct than to say that PO has already occurred, but would you consider qualifying that statement by saying that this small difference in dates is immaterial in considering the necessity to 'get on with the job'. (Not much point worrying who is in front with a cement truck bearing down).
Actually, I do believe Robert has said that it makes no difference whether he is right or WesTexas is right regarding the year of actual PO-- that we need to prepare now as the time difference is essentially meaningless. Robert, correct me if I am wrong.
Cheryl
Robert, correct me if I am wrong.
No, you are absolutely correct. And I am doing my best to prepare as if peak is upon us; I am making changes in my life and trying to set an example for those around me. My main concern - explained above - is that if peak is still 4 years away we are going to completely lose our voice because we called peak now, and the 4 years are going to be lost time.
Hi Robert,
Yes in your debate you do state that response should be immediate. Sorry about that bad oversight of mine.
But I do not see why, if on this point you and Westexas are in close agreement, a statement to this effect shouldn't be broadcast, with force, while explaining the limits of the PO forecast.
It seems to me that a stumbling block to this forecast is the inability to access the Saudi figures necessary to a quite definite prediction? The Saudis must hold accurate knowledge of their reserves and I would think this information would be available to the US government as well. Oil reserve figures for SA are a US security matter that would be of foremost concern and given the political relationship between the SA and the USA this information would be common knowledge between them.
If the US Government holds this information would it not be an idea to request it directly if only for public relations value for TOD in MSM .
Further,
SamuM (above, I think) gives the following link,which I will repeat here:
http://www.cge.uevora.pt/aspo2005/abscom/ASPO2005_Schoppers.ppt
I couldn't copy and paste the bit that interested me so will quote it as follows: from the article by Dr. Marcel Shoppers and DR. Neil Murphy. Pleasantly for us lay guys they end with afterthought #!:
2.5% chance the peak occurred by 2003
15% chance the peak will occur by 2006
50% chance the peak will occur by 2009
85% chance the peak will occur by 2012
95% chance the peak will occur by 2015
2.5% chance the peak will occur after 2015
This sort of approach if validated, I think,would stifle the 'wot year is this supozed to hapen?' sort of criticism, while limiting 'Cry Wolf' concerns.
BTW Robert, Still beavering away at all the material you loaded on my plate...thanks a bunch..I think. And if you have anything on how you are preparing for APO I would appreciate hearing about it, but only at your leisure.
Hi Black B.
re: "But I do not see why, if on this point you and Westexas are in close agreement, a statement to this effect shouldn't be broadcast, with force, while explaining the limits of the PO forecast."
Thank you. I second this (as I've been making similar comments for a while).
I very much like the Shoppers/Murphy way of presenting it. Add what you say above, (summarize the analyses that give rise to one or more predictions) - and this should do it. Something rational and scientific (yes?) upon which to base positive action.
If it is 4 years away, and don't get me wrong, I wish it was 20 yrs away, I believe they will not care if it was called early. They will be saying, "oh sh*t, what do we do now?", and turn to anyone who seemed to know something about it. They certainly won't be listening to CERA.
Hi Cid,
I just wanted to respond (though no one may be reading by now) - To me, the issue, as I've said above, is to lay out, with care, the parameters and methods of any and all predictions - for all to see.
It should be doable to come up with a statement that summarizes the range of methods and outcomes.
The reason this matters, as I see it, have to do with the following:
1) There's been much discussion about how things might unfold once peak is upon us (or even if it has already begun). The evidence of "peak" per se - ie., "peak"-as-the cause-of-the-outcomes-of-peak, might not be at all evident. We need a clear statement about what we are saying re: world supply of oil, in order to be able to draw connections between events and what we view as the cause (major cause) of those events.
2) Most people (in my experience) have very little of the relevant knowledge that might enable them to grasp the significance of peak. So, in other words, they might not have anything at all resembling the response you imagine. They may not know; they may not be able to know. They may draw the wrong conclusions, for example, the recession is temporary. The Depression is temporary. Things will return to "normal"...just as an example. Hence, it's important to be able to make extremely well-reasoned, well-evidenced "predictions" - or even background statements.
3) Even people with relevant background may not understand how oil functions (or energy for that matter) in the economy. and finally...
4) My experience is that the ability to grasp (let alone respond in a constructive fashion to) the idea of "peak" has almost nothing to do with anything at all - education, etc. There is something quite uniquely difficult on the emotional side about this topic, in other words.
This is why I believe the statement Black B makes (I have made, also) is important. It does matter. To be able to give a coherent summary matters.
Hirsch et al did this (in a way) in the first Hirsch report, by simply listing the various "calls". We need to take this further.
About the "85 million bpd with little excess capacity", where was Saudi Arabia as a 'swing producer' at this time?
Oh, I think up until they started making the cuts, they were producing all the could. In other words, I do believe they lost their ability as swing producers during the time of high demand. Now, is that because they were peaking, or was it because they failed to properly anticipate demand? Given that they still have lots of projects in the pipeline that will bring new oil online, I would say they failed to properly anticipate demand (and supply, for that matter) and have these projects coming on later than they should have.
this small difference in dates is immaterial in considering the necessity to 'get on with the job'.
I think you are kind of new here, so I will repeat my position. In no way do I think we should relax. We need to get on with the job. If we peak in 2012, we are still going to be in dire straits, because we won't be prepared.
Here is my concern. Many people have gone on record now and said "Saudi has peaked." Some of them are using the HL model to back them up. Some very influential people are picking up on this and repeating it. Therefore, it is important to understand the precision of that model.
Here is the worst case scenario in my mind, and the whole reason I argue these points with Jeffrey. What if production starts to rise next year? He, and others who have been calling peak last year or in 2005 are going to be finished prognosticating. The influential people who repeated these claims won't be taken seriously on this subject.
Yet we may still have 4 or 5 years in which to convince the policy-makers that we have a serious problem, and it is nearly on top of us. But they aren't going to listen, and those 4 or 5 years will be wasted as well. Folks like CERA are going to point to yet another failed prediction of peak, and credibility will slide. But I feel we will need our credibility more than ever in the next few years, and I am determined to fight anything that may chip away at it. If the methodology for calling peak in 2006 is sound, so be it. If not, I want to make sure that is understood.
It is better to have 2 or 3 years to prepare than no years. Maybe only marginally better, but for some people this may be the difference between life and death. I fear that if Jeffrey is wrong because he overstated his case, we are going to end up with no years to prepare because people are going to start viewing these predictions as "end of the world" religious beliefs that continue to be wrong.
Yes, just a bouncing baby meathead who finally has figured out how subthreads work...I think.
By that logic he is out in right field and and you in left, while the ball is somewhere in the middle. Westtexas takes a risk in loosing credibility while you with caution hold back and risk losing a lot of important years. Each of you strike me as intelligent people but, in your individual rigors, perhaps both going to miss the ball. I would suggest a large net that the two of you hold up for a possibly, maybe, sure 'out' on.P.O. (of course nets don't always succeed...I've been out and about for a mighty long time).
and now to tackle a greased piggy-wigg...the Parent subthread...whatever that ferocious looking creature is.
Westtexas takes a risk in loosing credibility while you with caution hold back and risk losing a lot of important years.
That's just the thing, though. I am not saying we should hold back with respect to preparations. I think we should prepare as if a meteor is bearing down on the planet. But I think the message is going to lose a lot of its urgency if those who forecast peak last year turn out to be wrong as production starts to rise. We may have 4 or 5 years to keep beating the drum: Start preparing for event.
I do like that approach that you copied above where they give probabilities. Those probabilities are not too far from what I would suggest. That approach really emphasizes the urgency without the risk of credibility loss.