Megaprojects Predict Decline Of Oil Production

I’ve written a number of posts about megaprojects analysis, the most recent being in early July.   That post did not extrapolate a full supply model, which is one thing I want to do in this post.  But another post in early June did analyze the next seven years of oil supply based on megaproject work.  It is interesting to see how that analysis differs from the one below and yet posits similar conclusions as discussed below.

I think megaprojects analysis gives us the best insight on near-future oil supply levels.  Some other observers prefer to to predict future supplies using the formulas of Dr. M. King Hubbert which are based on original oil in place and decline rates.  But Hubbert’s math was done before the current technologies for enhanced oil recovery were available.  These new techniques and technologies change the timing of oil extraction over the life of a field resulting in more oil being produced more rapidly than could be done when Hubbert was writing.  That puts Hubbert’s mathematical model into question, I believe, as a near term predictor of peak oil.

There is nothing theoretical about megaprojects analysis, only a bunch of best guesses as to actual production from new oil fields measured against estimated near term declines in production from old oil fields and estimated changes in global demand for oil, all by year.  Those guesses may be close to the truth or not but they are not based on theory.  They are based on announced production intentions and historical facts of decline and demand used to estimate future decline rates and demand.  All three areas, when projected into the future, will differ from the eventual reality, but if they are fairly close - even just in the right ballpark - the results may give us a meaningful estimate of the timing of peak oil.

“Megaprojects analysis” refers to the best guess about production from new fields. It is simply a compilation of public announcements of new oil production scheduled to flow each year for the next seven years from new fields. Since large projects are announced up to 7 years before they actually start producing oil in many cased, these announcements provide a basis for predicting with some reasonable accuracy the visibility of future oil supplies. After adjustment, it can give us a decent bottom-up understanding of how oil supply is likely to unfold over the next 5 -7 years.

The public announcements of new oil field development by private companies and national oil companies are compiled by individuals who volunteer their time to scour the world’s public documents and post the information cooperatively through Wikipedia.  The Oil Drum, where many of the people who work on the Wikipedia Megaprojects analysis frequently post, recently ran an update on it.  My analysis is based partly on that update and partly on my own ways of adjusting the data to reflect a more nuanced reality than the raw data can provide.

Adjusting Megaprojects Data

Megaprojects data is constrained by several tight definitions.  First it includes Crude oil + concentrates including unconventional oil such as from the Canadian oil sands.  It does not include natural gas liquids or non-petroleum based fuels such as ethanol or bio-diesel.  Secondly, it counts a project only once, assigning all the new production from the field at its projected maximum output to the single year that it’s first production is scheduled to flow.  Thirdly, it counts only projects with production greater than 50 kb/d.  Clearly, then, a simple recap of the megaprojects’ production of new oil flow by future years needs to be adjusted to account for the fact that:

1.  many projects, particularly the larger ones, are delayed and do not start to produce for some time after the most recent target date, and

2.  some projects, particularly the larger ones, take a period of years to build up to their ultimate maximum production rates.  For example, the Tupi field in Brazil is expected to start flowing oil in 2012 at about 100 kb/d but will not reach peak production until nearly 2020 at over 1 mb/d, and

3.  by focusing only on the “mega” projects the analysis fails to count the huge number of smaller projects that in combination will provide substantial new oil flows.  The importance of smaller projects is particularly important as the price of oil rises, since with a higher price more small projects are profitable.

Two areas that are not adjusted are natural gas liquids and changes in global inventory level.  Both of these numbers are important compared with total “all liquids” production and availability but only their changes from year to year make a difference in terms of this analysis.

Here is a graph of the current megaprojects projection showing the amount of oil production scheduled each year through 2015 by project size:

One of the first adjustments needed is to add in more smaller projects in the out years, since they are often announced only two or three years ahead of production.  The megaprojects data contain a lot of small projects in 2008 and 2009 compared with prior years and especially compared with later years.   This makes sense because it takes a few years for high oil prices to motivate producers to exploit formerly unexciting smaller projects.  The higher prices of oil only began in 2004.  Thus it makes sense that a good number of smaller projects would come on stream in 2008 - 2010 as shown above.    On the other hand, it does not make sense that there would be so few in 2011 and beyond.   Therefore, we should adjust the contribution of small projects upward starting in 2011.

A second adjustment, as noted by many commentators, is that there is a strong tendency for projects to become delayed, particularly large projects.  Delays are caused by the very size and thus complexity of a megaprojects and delays are compounded by the present environment of shortages of oil drilling equipment and personnel.  The Kashaghan debacle, a mb/d field that has been delayed from 2005 to 2013, is a prime example.  A report by CIBC World Markets says they expect about 5 mb/d of megaprojects scheduled in 2008 - 2012 to be delayed beyond those periods.  I have assumed 3.5 mb/d are delayed, but only by two years.  This is clearly a more optimistic picture than CIBC’s.

A third adjustment relates to decline rates for existing fields.   A conservative 4.5% decline rate resulting in about 3.8 mb/d of reduced production from oil fields every year.   This tends to be a consensus number.  But more recent data indicates that 5.2% may be a more realistic number which would result in an additional 600 kb/d of decline each year.   In implicit assumption contained in the decline rate is the increase in enhanced oil extraction techniques that speed production from old fields and thus are in effect netted against the decline rate.    I have shown net decline rates of both the  4.5% and the 5.2% so that readers can see the difference and may use their own best judgment.

The real ringer in the decline analysis is Ghawar, the giant Saudi field that has generated 5 mb/d for nearly 60 years, but only with much more drilling equipment and artificial pressurization being utilized in recent years.  Ghawar requires millions of barrels of water to be injected into the field every day.  That water pushes the oil upward and into the collecting apparatus.   At some point, the water will push up to the surface and will displace nearly all of the oil.  This is likely to happen first in limited parts of Ghawar and then spread to the whole field over time.   Then the amount of oil produced every day from Ghawar will start to decline.

When Ghawar declines it may well follow the example of another giant old field that has had extremely high artificial pressurization efforts, Cantarell.   That Mexican field has been in decline for a couple of years and is now declining at the alarming rate of 36% per year.  The current estimated global decline rate - whether 4.5% or 5.2% - assumes flat production at Ghawar or makes no specific assumption about Ghawar.  But when Ghawar begins to decline the global decline rate will begin to rise and the rate at which it rises is likely to increase markedly over time as has been the case with Cantarell.  By simply showing both the 4.5% and 5.2% decline rates I make the implicit assumption that Ghawar keeps on producing at the 5mb/d rate through 2015.  My own guess is that this assumption is optimistic.

Another adjustment relates to non-petroleum based fuels.  In recent years ethanol has made a significant contribution, especially in the U.S.  I am assuming no additional contribution from ethanol or bio-diesel going forward for two reasons.  First, the big increase came when the U.S. mandated that MBTL ethyl additives be removed.  The substitution of ethanol for MBTL has now taken place.  Secondly, increased understanding that food-based ethanol production is ineffective in creating additional real energy and helps raise global food prices is making grain-based ethanol expansion much less popular.

Eventually technology may be able to create fuel from cellulose-based ethanol.  If/when that happens, there may be another source of fuels that will require further adjustment to the megaprojects analysis.   Given ramp-up constraints, I doubt that could begin to have a significant influence earlier than 5 years from now.

Finally, with plug-in hybrid electric cars (PHEV’s) we are starting to see electricity substitute for petroleum based fuels.   The first production model PHEV will be sold in 2010.  It has been estimated that by 2015 there may be 2 million such cars on the road globally.   That will not make a discernable difference in global gasoline usage.  2015 is the end point of this analysis, so additional PHEV’s in future years will help, but not until beyond the purview of this analysis.

Adjustments

Below is an adjusted Megaprojects graph from an analyst who posts on The Oil Drum under the name “Ace”.  He is one of the key people working on the Wikipedia Megaprojects analysis.  It shows an adjustment to the chart shown above to reflect likely delays in scheduled projects.

Lets look at the raw numbers that come off this chart compared with the original.

Year Oil Production - original Oil Production - adjusted for delays

2008          5.5 mb/d                                       5.5 mb/d

2009          4.75 mb/d                                     5.2 mb/d

2010          5.25 mb/d                                     4.25 mb/d

2011          3.5 mb/d                                       3.25 mb/d

2012          3.5 mb/d                                       2.75 mb/d

2013          1.3 mb/d                                       2.25mb/d

2014          1.1mb/d                                        2.0mb/d

2015            .75 mb/d                                     1.75mb/d

I’m not sure what methodology was used above to adjust the original Megaprojects numbers, but I suspect it fails to adjust for all the factors discussed above. I want to try my own approach to this adjustment in accordance with a formula that recognizes that project delays are becoming more extended.  Thus, from 2009 onward I will transfer 25% of the originally projected oil  from the original year to two years later.  12.5% of 2008 projected oil will be deferred.  A second adjustment I will make is that a substantial amount of oil from small projects - which generally do not have the early visibility of large projects - will be added starting in 2010.  The chart below shows these adjustments:

New Oil Supply Adjusted for Delays and Uncounted Small Projects

I then adjusted the numbers to account for the fact that it takes time to ramp up a field to achieve its ultimate capacity.   To adjust for that I took 25% of the production from each year and shifted it to two years from startup.  That is obviously a gross attempt at the adjustment and I make no claim for accuracy other than the fact that I think it reflects reality better than the number without any adjustment for ramp-up would.

Adjusted New Oil Supply Further Adjusted for Ramp-up

As you see, the numbers are much flatter than the original Wikipedia Megaprojects numbers but they still show a significant slowdown in oil production starting in 2013.  They also indicate less production in 2008 - 2010, which might account for the continuing tight oil market in which oil prices have stayed above $100 a barrel.

Finally, I added the two other factors that put new oil production into the perspective of its adequacy.  First is the decline rate for existing (older) fields that reduces oil production each year.  Second is an estimate of new demand for oil from an expanding global market - or what might be called Gross Global Product.  I’ve shown two different rates of decline, the “standard” 4.5% and a higher rate based on recent discussion of 5.2% or even higher and based on the expanding declines of Cantarell which is now running at 36%.  Therefore I show a 5% decline rate.  In both cases I increase the decline rate by .1% each year because there is a clear trend toward higher declines, a trend that is supported by our understanding that the fields to which EOR methods of artificial pressurization have been applied always decline at faster rates than other old fields.  Another reason to use a slight annual increase in decline is to account for the fact than my model has no specific assumption of Ghawar starting to decline.

Here is an excerpt from the recent IEA report on global decline in their latest Medium Term Oil Market Report:

Since the 2007 MTOMR, significant downward revisions have been made to both non-OPEC supplies and OPEC capacity forecasts. Project delays averaging 12 months, coupled with global average decline of 5.2% - up from 4% last year – are the factors behind these revisions.

For global demand increase I used rather low numbers compared with recent years’ averages of about 1.5 mb/d per year.  I reduced that number in the near term to account for the double impact of higher oil prices and lower global growth rates.  This combination has resulted in lower OECD and lower Chindia oil consumption changes than in prior years.  I assume oil consumption increases for a couple of years as the global economy limps back to health.  Then I assume it falls off again as lower net production causes much higher oil prices which should damp both oil demand and economic activity once again.  In any event, there is less oil being produced than is demand in all years after 2008.

Net Annual Oil Sufficiency Analysis, Best Guesses

Conclusions

What does it all mean?  First, if these estimates are at all closely correlated with reality we should expect oil markets to be tighter than at present through 2012 followed by even more need for demand destruction from 2013 through 2015.  The degree of global oil shortage will depend greatly on whether the rate of decline in old field production is closer to the 4.5% number or whether it goes higher.   But we are only talking about the degree of stress.  There will be unremitting stress in the oil price going forward and prices will escalate after 2010.  As noted by Chris Skrebowski, there may well be seasonal shortages of oil even in years when the total supply is close to meeting global needs.

Let’s remember that these numbers relate to millions of barrels of oil per day.  Even in the easier decline case for 2011 when the expectation is for .84 mb/d shortfall, that amounts to over 300 million barrels of oil less than is desired by the world for the full year.   I would think that would be enough to drive the price above $200 by 2011.   By the time we get to 2015 there seems to be little doubt that a price in excess of $500 a barrel would be needed to continue the demand destruction required, even with no new demand assumed for 2013 - 2015.

For the sake of comparison,  here is a table showing the net balance of demand and supply with the same decline and demand assumptions but using the simplified assumption of supply shown by the the original “smoothed”  version of the Megaprojects production flow shown above (the curved line).

Net Oil Sufficiency Based Only on Wikipedia Smoothed New Production

Note the dramatic differences that result between these “unadjusted” numbers above and those in the prior chart that I adjusted for delays in startup, delays in reaching peak production, and addition of new small projects in the out years.  These “unadjusted” numbers show ample supply of oil through 2009 followed by shortfalls in 2010 and very important shortfalls starting in 2011 - possibly even catastrophic shortfalls in 2012 and thereafter if decline rates are more severe than 4.5%.  In other words, the adjustments I have made to the megaprojects data result in a far milder oil shortfall than a simple smoothing of the data does.

It is interesting to compare the differences between the unadjusted Wikipedia annual new field flows available in my June 3rd post, the smoothed Wikipedia projections of new field flows available now, and the more systematic adjustments that I made above to the current Wikipedia estimated new field flows.  How do they change the message of the raw data?

Here is the raw data available in early June followed by my adjustments to that data

Here is a comparison of the three models:

All of the models show an increasing problem of oil supply adequacy going forward from 2008 and a fairly sharp drop in 2013.  Interestingly both the early June model and the my current model show a declining net supply in 2009 and a slightly less problematic situation in 2010 and both show more substantial problems starting in 2013.  It is tempting to prefer the Wikipedia smoothed model because it shows a steady decline in oil sufficiency in each year.  Possibly that will happen, but I suspect reality may well be lumpy as shown in both of my models just because the world tends to be a lumpy sort of place.

For one final data point, let me remind you that in March, 2008, Chris Skrebowski made his most recent prediction that I have seen about peak oil based on his own megaprojects work.   The report that I posted on his work said: “Chris Skrebowski, a researcher for the Energy Institute in Britain, told delegates that the oil supply will peak in 2011 or 2012 at around 93 million barrels a day, that oil supply in international trade will peak earlier than the oil production peak, and he forecast: “There will be supply shortfalls in winter before peak.”

In continue to feel that Skrebowski has as good a feel for this complex set of related oil flows as anyone.  My model tends to reflect his predictions.  Absent any consideration of oil demand - so looking only at supply - I show oil peaking in 2010, but 2011 and 2012 are nearly break-even and clearly the major problems start in 2013.  But I love Skrebowski’s statement that “oil in international trade will peak earlier than production will peak”.   What that means is that oil exporters will reach peak oil exports before global oil production peaks - in other words before 2012 or 2013.  That would seem to be the practical bottom line.

A major uncertainty in the above analysis relates to Natural Gas Liquids.  These flows have been increasing substantially over the past five years and have served to mitigate what would have been an even tighter oil market.  Some observers believe that NGL flows have peaked.  I make no prediction but would be interested to learn of others’ views.

Subscribe by RSS Subscribe by Email