Peak Oil III: Forecasting Future Oil Discoveries

Much of the work presented below was performed and published by Jean Laherrere, co-author of a seminal paper "The End of Cheap Oil", published in Scientific American in 1998.  Mssr. Laherrere is a member of the Association for the Study of Peak Oil.   Many thanks for the availability of his work.  Links are shown below.

Abstract:

Historical trends in oil exploration ("creaming curves") provide a basis for forecasting the volume of future oil discoveries.  Historical trends strongly indicate diminishing returns with time and exploration effort (number of exploration wells).  Extrapolation of regional and global trends provides good agreement, that about 170 billion barrels of new reserves will be discovered by the year 2030, or about 200 billion barrels including heavy oil.   By comparison, projected production and consumption greatly exceed the forecast of new discoveries, with about 770 billion barrels expected to be produced and consumed.   Production by the year 2030 will exceed the volume of new reserves discovered by about 4-fold.  There is no reasonable extrapolation of historical discovery trends that would allow reserve replacement on a global scale.

-----

The history of oil exploration shows certain statistical trends which can be eused to predict future results.  Historical results reflect the intersection of markets, engineering, available capital, political and geographical boundaries, and the geologic occurrence of oil in the earth.  Despite huge improvements in engineering and exploration technology; increased exploration drilling; and the exploration of deep water, remote, and arctic regions, the total volume of new oil discoveries has been declining since the 1960's.  Although new technologies have improved the rate of exploration success; the size of newly discovered fields has become smaller, in a 50-year continuous trend.  

New technologies for oil exploration and production will undoubtedly emerge.  But will the new technologies of the current generation have any greater impact than the new technologies of the previous generation?  It seems to me that the volume of new oil discoveries continue to decline, reflecting the geologic availability of oil, and limits on the accumulation of physical capital.  Let's look at the data, and see what we can predict according to the extrapolation of historical trends in oil exploration.

There are a couple of basic laws of oil exploration, which are virtually never violated. 

First law:  Within any given area, there are very few large fields, and very many small fields. (Geologists say that oil field sizes follow a log-normal or parabolic fractal statistical distribution.)   The existence of a large field implies the existence of small fields nearby.   In my experience in oil development in the Gulf of Mexico,  smaller satellite fields always occurred in the neighborhood of a large field.  As a rule of thumb, I found that the small, satellite fields surrounding a large field would add an incremental 40% or 50% to the reserve potential of the large field.  Like fractals, oil fields occur in self-similar patterns independent of scale.

Second Law:  Within a given area, fields are generally discovered in rank order according to size.    Large fields are easier to find, and geologists usually find the biggest fields first.  Here is a figure showing the history of gas discoveries in shallow water of the Gulf of Mexico.   

Figure 1.  Gas fields discovered in shallow water (U.S. waters) of the Gulf of Mexico, by discovery year.  Bubble size indicates field size (ultimate recoverable reserves).

Large fields were found beginning in the earliest days of offshore exploration, and diminishing field sizes followed.  A dramatic fall-off in the size of new field discoveries occurred in the mid-1980’s.  The industry’s continuing efforts in shallow-water exploration were essentially wasted, considering the expense of the exploration effort and the results achieved.   The single large recent discovery in the shallow water gulf of Mexico is the Davy Jones discovery at a depth of about six miles.   The energy return on energy invested (EROI), and the economic return in exploring for such targets remains in question, in my opinion.

Another figure from the Gulf of Mexico shows the decline in oil-field size over time, even as more fields are discovered.    Although more fields are being discovered, the total volume of newly discovered hydrocarbons diminishes.  The energy return and economic return from those discoveries are usually subject to diminishing returns.

Figure 2.  Number of discovered fields (bar graph) and mean field size (line) by discovery year; Gulf of Mexico.

The global plot of average field size over time shows the same thing.   With increasing maturity and technology, success rates rise, but field sizes diminish.  Overall, discovery volumes diminish with time, as we have previously seen in discussions of peak oil.

Figure 3.  Global average oil field size and exploration success ratio, by discovery year.

Another presentation of the diminishing field sizes is a cumulative graph of reserves discovered.   The graph is called a “creaming curve”, and cumulative reserves are graphed against time, the number of exploration wells, or the number of fields discovered.   Here is the world creaming curve, presented as a function of number of fields discovered.   The inflection of the curve shows that new field discoveries are progressively smaller and smaller.

Figure 4.  The Global Creaming Curve.  Horizontal axis is the number of fields discovered; the vertical axis is volume of reserves found.  The dark green curve includes heavy oil; the light green curve does not.  Red is gas; the blue curve is the cumulative count of exploration wells.

The creaming curve can also be presented in terms of time.  The global curve shows less of an inflection, but it should be noted that extending the trend of new oil reserves requires progressively more effort for every barrels.  I will present the creaming curve by discovery year, in order to forecast future discoveries by a given date.  Let's look at the globe, region by region.

Figure 5.   Creaming Curve for the Middle East.

The Mid-East shows a very mature creaming curve.  The flattening of the cumulative discovered reserves shows a serious decline in the size of newly discovered fields.  The light blue line, showing the number of exploration wells, shows that exploration effort has increased rather than decreased with time, but with progressively diminishing results.  (Excuse me, I obscured the scale to modify the chart.  There are about 4400 exploration wells.)   The historic trend suggests that about 10 billion barrels of new reserves will be discovered in the Middle East by the year 2030.

Figure 6.  Creaming Curve for Africa

Extrapolation suggests that 35 billion barrels can be discovered by 2030.  Prospective areas include deep water in the Atlantic offshore, central African rift basins, deep water Mediterranean and frontier desert basins in Libya and Egypt.

Figure 7.  Creaming Curve for Asia (without the former Soviet Union).

Extrapolation suggests 35 billion barrels of new reserves by 2030.  Prospective areas include the South China Sea (currently the subject of territorial disputes between China and other SE Asian countries), deep water Indonesia, offshore Pakistan and India.

Figure 8.  Creaming Curve for the Former Soviet Union.

Extrapolation suggests that 15 billion barrels will be discovered in the Former Soviet Union by 2030.   Prospective areas include the Caspian region, central Asian countries (e.g. Turkmenistan), eastern Barents Sea, Kara Sea, eastern Siberia, sea of Okhtosk, and Siberian arctic and offshore.   The large diversity of frontier geologic targets suggests to me that results might well exceed the historical trends.  But available capital, transportation issues, and markets may impede the exploration of these remote places.

Figure 9.  Creaming Curve for Europe.

Europe shows a very mature creaming curve.  A relatively minor 5 billions barrels can be expected to be found by 2030. Future potential may be found in the deepwater Mediterranean, the western Barents Sea, west of Shetlands Atlantic margin, and in shale oil on the continent.

Figure 10.  North American Frontier Creaming Curve.

Extrapolation suggests 17 billion barrels of new reserves by 2030.   Prospective areas include the Eastern Gulf of Mexico (offshore Florida), deep water offshore southern California, Mexican and Cuban deep water Gulf of Mexico, the Alaskan arctic offshore, including the Chukchi Sea.  Political and environmental issues are likely to prevent access to these areas for exploration and production for the foreseeable future.

Figure 11.  Latin and South America Creaming Curve

South America has seen some of the largest discoveries in recent years, with supergiant oil fields recently discovered in the sub-salt play of Brazilian deep water, following on giant oilfields discovered in the deepwater Santos Basin in the previous two decades.  Extrapolation of the current trend suggests that 50 billion barrels of new oil will be discovered by 2030.  Some moderation of the trend is expected, as massive capital will be needed for the development of existing discoveries.  Capital will primarily flow to development projects, rather than to additional exploration.  Still, some exploration drilling will take place, and the knowledge gained from existing discoveries will aid in future exploration.

The total forecast of new discoveries by region adds up to 167 billion barrels by the year 2030.

Middle  East                                       10

Africa                                                 35

Asia (ex-Soviet Union)                       35

Former Soviet Union                          15

Europe                                                 5

      Eastern Hemisphere Subtotal                      100 Billion Barrels

North America Frontier                      17

Latin America                                     50

      Western Hemisphere Subtotal                       67 Billion Barrels

Figure 12.   The Global Creaming Curve.   Extrapolation indicates 170 billion barrels of new conventional reserves by the year 2030; or 200 billion barrels including heavy oil.
There is good agreement between forecast future discoveries by region and the total extrapolated from the global creaming curve.

The principle flaws or criticisms of extrapolation from creaming curves include:  1)  The creaming curve represents only past experience, and doesn't include potential from new regions.  Large oilfields might be found offshore Greenland, in parts of the Pacific, the Black Sea, the Arctic, or even in the Eastern Gulf of Mexico, 2)  The creaming curve doesn't account for possible new exploration or production technologies, or 3) The creaming curve doesn't account for large volumes on unconventional resources, such as shale oil or tar sands.
It is possible that any of these, or some combination of these, may result in a reversal of historical trends, and that large and economic volumes of oil will be found and brought to market.
However, it is worth remembering that the historical record ALREADY includes discoveries and production from new regions, notably the revolutions of offshore exploration, deep water exploration, and arctic exploration.  Substantial improvements in technology are also already part of the existing historical record.  And while unconventional resources are only beginning to have a substantial impact on global production, realizing large gains from these resources will require the slow accumulation of financial and physical capital, and may be subject to limits in the availability of enabling resources, such as natural gas and water.
In sum, historical trends of exploration results (particularly on the global scale) form a legitimate basis for forecasting the volume of future oil discoveries.

As a final point, we should compare our forecast of new discoveries to the year 2030 to the volume expected to be produced through the year 2030.  Global production in 2010 was about 31.7 billion barrels (about 87 million barrels per day).  Assuming the IEA (International Energy Agency) estimate of 1.4% annual growth in production and demand, production in the year 2030 will total 41.9 billion barrels.  And cumulative production from 2010 to 2030 will be 768 billion barrels.  Production to the year 2030 will exceed the volume of new reserves discovered by about 4-fold.   There is no reasonable extrapolation of historical discovery trends that would allow reserve replacement on a global scale.

http://www.hubbertpeak.com/laherrere/bibliography.pdf
http://aspofrance.viabloga.com/files/JL_Sophia2010_part1.pdf
http://aspofrance.viabloga.com/files/JL_Sophia2010_part2.pdf

Peak Oil II: Oil Production by Hemisphere

Daniel Yergin, author of “The Prize” and “The Quest”, chairman of Cambridge Energy Research Associates, recently wrote an interesting opinion for the Washington Post.  Yergin speaks of a new world oil map, centered not on the Middle East, but centered in the Western Hemisphere.   Yergin cites production from Canadian oil sands, shale oil in the United States, and deepwater discoveries in Brazil, as remaking the global oil map.

http://www.washingtonpost.com/opinions/daniel-yergin-for-the-future-of-oil-look-to-the-americas-not-the-middle-east/2011/10/18/gIQAxdDw7L_story.html

Is Yergin right?   Does this make any sense?

In a nutshell, by my review of the data it does not make sense, at least in the near term.  New sources of oil in the Western Hemisphere will require massive amounts of capital, and long lead times for the accumulation of physical capital (rigs, processing plants, pipelines, etc.).   Oil sands production will be limited by available water and gas resources, and oil shale production will be constrained by land use issues and environmental opposition.   All of these sources of production will have low Energy Return on Energy Invested, meaning that there is little excess value to encourage investment, or to allow government support or taxation.  (See my earlier post regarding EROI http://dougrobbins.blogspot.com/2011/09/energy-return-on-investment.html).

Let’s look at some data.

This figure is available as an interactive graphic on-line:  http://chartsbin.com/view/wyw

The Middle East is the largest petroleum producing region.   Production in 2008 was over 26 million barrels per day; production has also been increasing faster than any other region for the past two decades.    On average, Middle East oil production has increased 2.6% annually for over two decades.  The Eastern Hemisphere produces 62 million barrels per day (2008) or 75% of global supply.  The Western Hemisphere produces 20 million barrels per day, or 25% of global supply. 

Now let's look at the new production sources cited by Daniel Yergin.

Canadian oil reserves are reported to be 2^nd largest in the world, behind Saudi Arabia.  Canadian reserves consist of 5 billion barrels of conventional oil, and 173 billion barrels of oil sands.  Properly speaking, the oil sands are not reserves, but contingent resources.  The capital and resources required for production must be obtained, before the oil sands can be classified as reserves.   Resources are not production.  Canadian geologist David Hughes estimates that oil sands production will be limited to 2.5 million barrels per day, due to limits on natural gas and water required for production, and light oil dilutants required for transportation.  http://www.aspo-usa.com/fall2006/presentations/pdf/Hughes_D_OilSands_Boston_2006.pdf

Canadian production is currently 3.2 million barrels.   Production is expected to increase by about 1 million b/d by 2020.

North Dakota currently produces 350,000 barrels per day; there are estimates that production could double within a decade, and pipeline capacity is being constructed to carry 1 million barrels per day.

It is likely that success can be duplicated in other regions (Eagle Ford shale, Texas; Niobrara shale, Great Plains; Green River Shale, Wyoming).  According to a US Government report, an aggressive goal for shale oil development in the United States would be 2 million barrels per day by 2020. http://fossil.energy.gov/programs/reserves/npr/publications/npr_strategic_significancev1.pdf

Brazil produces 2.6 million barrels per day (2009 data).  Brazil’s production has been increasing at about 4.5 % per year, and could add another 1.3 million barrels per day in ten years.  New exploration discoveries are located in deep water offshore Brazil's southern coast, and are found under a thick layer of salt.  Although new reserves may total 40 billion or 50 billion barrels, these will be very expensive and difficult fields to develop.  http://lucian.uchicago.edu/blogs/bric/files/2011/05/Afonso-H.M.-Santos-Brazilian-Energy-Overview1.pdf

To sum up, expected incremental production in the Western Hemisphere is about 4.3 million barrels per day by 2020.   If all other production remains constant, the Western Hemisphere share of global production would increase from 25% to 28%, a noticeable increase, but in no way does this re-center the world oil map on the Western Hemisphere. 

It is true that Canadian oil sands, U.S. oil shale, and Brazilian deepwater discoveries represent large oil resources.  But development will require massive amounts of capital, long lead times, and are subject to limits of necessary resources.   Much of this development will also be subject to strong environmental opposition.

What was Yergin thinking?   He is focused on contingent resources, rather than proved, production-ready reserves.  He may be focused on potential investment opportunities for Western oil companies.  In the very long run, it may be possible to significantly increase Western Hemisphere oil production, but only with high investment costs and long delay, and it is unlikely to shift the geopolitical balance of Western dependence on oil from the Eastern hemisphere.

Yergin poses a good question.  Where will future oil production come from?   I'm preparing two additonal posts; one regarding the likely location and size of future oil discoveries, and another regarding oil production forecasts to the year 2030.
----------------
Postscript:
Posts written after this article considered other aspects of Peak Oil.  In each, the Eastern Hemisphere shows potential from more future production than the Western Hemisphere.  The Eastern Hemisphere shows greater potential for future oil discoveries; greater potential for increasing recovery from existing oil fields; and greater potential from discovered, but undeveloped new fields.

Peak Oil III:  Forecasting Future Oil Discoveries:
http://dougrobbins.blogspot.com/2011/12/forecasting-future-oil-discoveries.html
Historical trends show that about 100 billion barrels of new reserves will be discovered in the Eastern Hemisphere by the year 2030, while only 67 billion barrels will be discovered in the Western Hemisphere.

Peak Oil IV:  Recognition Lag and Reserve Growth

http://dougrobbins.blogspot.com/2012/02/peak-oil-4-recognition-lag-and-reserve.html

Mature giant oil fields in the United States have produced about 35% of original oil in place, and are approaching the technical limit of recovery.   Globally, and particularly in the Middle East, giant oil fields have only produced about 22% of original oil in place.  There is greater potential for reserve growth in the Eastern Hemisphere than the Western Hemisphere, as re-development, secondary and tertiary recovery methods are applied to these fields.

Peak Oil V:  Discovered, Undeveloped Reserves in the Middle East
http://dougrobbins.blogspot.com/2012/02/peak-oil-v-discovered-undeveloped.html
Production from the Middle East has been dominated by a small number of giant and supergiant fields.   The large fields imply the existence of many smaller fields which have not yet been developed or produced.  About 1300 fields without significant production are indicated in summary tables; these may contain about 300 billion barrels.   Again, the advantage is to the Eastern Hemisphere.
-------------------------------------------
References and Credits:

Many thanks to Carleton '78 classmate Anne Croley, who brought the Daniel Yergin article to my attention.

http://www.theoildrum.com/node/5395

http://www.aspo-usa.com/fall2006/presentations/pdf/Hughes_D_OilSands_Boston_2006.pdf

http://www.capp.ca/getdoc.aspx?DocID=147850&DT=PDF

http://www.eia.gov/state/state-energy-rankings.cfm?keyid=28&orderid=1

http://www.thisismoney.co.uk/money/markets/article-1707785/BG-set-to-raise-Brazilian-oil-find-forecast.html
http://www.adn.com/2011/01/02/1629025/north-dakota-oil-production-is.html
http://www.msnbc.msn.com/id/43085246/ns/business-oil_and_energy/t/bubbling-crude-americas-top-oil-producing-states/
http://www.eia.gov/state/state-energy-rankings.cfm?keyid=28&orderid=1
http://www.eia.gov/countries/

Peak Oil

Peak Oil is a concept proposed by geologist M. King Hubbert, in a paper written in 1956.   In that paper, he famously predicted the peak of US oil production, which strikingly occurred as predicted fourteen years later   (figure 1).   

Figure 1.  Hubbert’s amazing prediction!

The concept of Peak Oil is highly contentious.  A number of smart people claim that a critical economic resource will soon decline, and other smart people claim that there are no limits to growing oil production.

Among the Peak Oil advocates are Colin Campbell and Jean Laherrere, authors of “The End of Cheap Oil”, Scientific American, 1998; Matthew Simmons (RIP, 2010), author of Twilight in the Desert; Ken Deffeyes, professor emeritus at Princeton; participants in the on-line forum “The Oil Drum”; and a group of distinguished geologists and engineers who formed the Association for the Study of Peak Oil.   

On the other side of the debate, smart people argue that Peak Oil will not occur for many years.   This group includes the highly regarded BP Statistical Review of World Energy; the International Energy Agency;  Daniel Yergin and the Cambridge Energy Research Associates (CERA), economic advisors to the Saudi government and many major oil companies.   Daniel Yergin is the Pulizer-winning author of The Prize and The Quest, and head of the CERA.   In the view of this group, new production from natural gas liquids, tar-sands,  the Mid-East, the Caspian region and Brazil will more than offset the natural decline of existing production.

Hubbert predicted global peak oil in the year 2000.   We are now eleven years past that date, and production is still growing.   So, who is right?  Is Peak Oil around the corner?   How did Hubbert make his predictions, and will he be right again?

Figure 2.  Hubbert predicted global peak oil about the year 2000.

Hubbert’s Peak

Hubbert’s original idea is ingeniously simple.   We have to find oil before we produce it, and we usually have a pretty good idea of how much we found several decades before that oil is produced.  The rate of production is assumed to be a bell-shaped curve, so the half-way point, when half of the reserves have been produced, is peak production.

Geologists find most of the oil in early exploration, because the largest fields are easiest to find.   After the largest fields are discovered, there is a predictable decline in field sizes.  The graphical display of the decline is sometimes called “the creaming curve”.   As geologists explore the globe, they proceed from the easiest areas to more difficult areas, leaving remote and difficult areas for last.  So if we look at the peak of discovered oil and make some assumptions about the decline of new discoveries, we can forecast the production peak.

United States’ Peak Oil

Let’s look at US oil discoveries and production as a case history.

We can only produce as much oil as we have found.   About 200 billion barrels have been discovered in the United States to date.   This is the area under the discovery curve (figure 3).   We can see that US discoveries peaked in the 1930’s, and declined thereafter.   

Figure 3.  US Oil Discoveries, 1900 – 2008.

And so far, we have produced 181 billion barrels (Laherrere and Tverberg).  The number is somewhat larger if natural gas liquids are included.  This is the area under the production curve (figure 4).   Remaining proved reserves are about 21 billion barrels (Wikipedia, Oil Reserves in the United States).   Production peaked in 1970 and declined in turn, exactly as Hubbert predicted in 1956.

Figure 4.  US Oil Discoveries and Production.   Production volumes must follow discovery volumes.

Since oil must be discovered before it is produced, it follows that the shape and size of the oil discovery graph must be reflected in the subsequent graph of oil production.   The area under the production curve has to match the area under the discovery curve (figure 5).

Figure 5.   US Oil Discoveries and Production; areas under the curves must ultimately match. 

Hubbert took the volume of known oil discoveries and estimated the total volume of oil that would eventually be discovered.   He then made a bell curve that fit the history of production, and fit the total volume of oil under the curve.   Hubbert assumed that the peak would be symmetrical; i.e., that Peak Oil would occur when we have produced one-half of the discovered oil.    Hubbert tried two models: one with 150 billion barrels of ultimate reserves, and another with 200 billion barrels of ultimate reserves.    By 1970, we had produced 94 billion barrels, very close to one-half of our original reserves endowment.  And that was the production peak, about 40 years following the exploration peak.

Global Peak Oil

So, then we move from US history to the global picture.   Hubbert was able to predict the peak of US oil production, because data on discovered oil volumes were well-defined and accurate.   The problem in predicting global peak oil is that data for discovered oil volumes are not well known, distorted by various national policies related to OPEC production quotas and national security interests.    Nevertheless, it is well known that global oil discoveries peaked in the mid-1960’s.  

Figure 6.   Global Oil Discoveries peaked in the 1960's.  Data courtesy of Colin Cambell.

The sum of global oil discoveries through 2005 is about 1500 billion barrels (data from C. Campbell).  However, it seems likely that if the global oil endowment was only 1500 billion barrels, oil production would have already peaked.  The USGS has estimated the global endowment of conventional oil as a range between 2250 billion barrels and 3900 billion barrels, with a mean of 3000 billion barrels.  

Global oil production is now over 86 million barrels per day, or about 32 billion barrels per year.    Cumulative global production through 2010 is about 1285 billion barrels.  If the USGS mean estimate of global oil of 3000 billion barrels is correct, we will reach cumulative the half-way point of 1500 billion barrels in 2016, and can expect to see peak conventional oil.   If the high estimate of 3900 billion barrels is correct, peak oil will occur about 2026, keeping in mind that the USGS places only a 5% probability on the high-side resource estimate.

Figure 7.  Global oil production is still rising in 2010 (BP Statistical Energy Review).

Problems with Peak Oil Theory

    Critics say Hubbert’s method fails to take into account discoveries in new areas, unconventional sources, or applications of new technology.   Newer thinking about peak oil modifies Hubbert’s model with consideration of unconventional resources, new technology, economic substitution, arctic resources, and geopolitical considerations.   

US Gas Production History

US gas production provides a dramatic example of how production decline can be reversed through new technology and unconventional resources.  US natural gas production also peaked about 1970, in agreement with Hubbert’s prediction, but new sources of gas resulted in an extended plateau, rather than a decline from the peak.   Although conventional gas production peaked in 1970, exactly as Hubbert predicted, gas in unconventional reservoirs (coal and shale) became commercial through the application of horizontal drilling and hydro-fracturing technology.  The new supply rejuvenated US gas production, and produced a production plateau, rather than a peak.  Gas prices, which spiked to about $12 per thousand cubic feet in 2008, declined and stabilized at about $4 per thousand cubic feet as a result of the new supply.

Figure 8.   US Conventional gas peaked about 1970, as Hubbert predicted.

Figure 9.   Gas production from unconventional reservoirs has sustained total gas production on a plateau, rather than declining as expected in Hubbert's theory.

Per Capita Oil Production

But from a self-centered viewpoint, the question is not how much oil exists in the world, but how much there is for me to put in my car.  And proportionally for every other person on earth.   World economic data show, without exception, that per capita energy and oil consumption is proportional to economic productivity.   Reduction of oil consumption per capita would cause higher prices and economic disruption until global economies adjusted to a higher level of conservation.

Although there is great uncertainty about the trend of future oil production, there is little uncertainty about future population growth.   Global population recently passed 7 billion, and will approach 9 billion by 2035.   Let’s consider several production scenarios, and the oil available for per capita consumption.

Figure 10.  Global population will continue to grow.

The IEA (International Energy Agency adopted an optimistic scenario termed “New Policies” as their base scenario.   The scenario assumes 1.2% annual increases in oil production to the year 2035.   About 55 million barrels per day of production (of the total 96 million) is forecast to come from fields not yet discovered or not yet developed.   There is substantial risk in this forecast.  My experience in the petroleum industry suggests that such unqualified success in new production is unlikely to occur.   

Figure 11.  In the IEA "New Policies" forecasts, a large wedge (55 million bpd) of forecast production is expected to come from fields not yet discovered or not yet developed.  

Even if the optimistic result of the "New Policies" forecast is realized, per capita consumption will rise only slightly by 2035, a gain of about 7.5%.   

Figure 12.  The IEA "New Policies" forecast would permit growing per capita consumption through 2035.

A second scenario might be a production plateau, as suggested by Daniel Yergin and CERA.   Assuming a plateau beginning in 2010, the result would be a per capita decline of 21% by 2035. 

 Figure 13.   A production plateau would still result in declining per capita consumption.

An alternative scenario termed “450” is presented by IEA.  This would be a production forecast limited by policies to stabilize atmospheric CO2 concentrations at 450 ppm.  Under this scenario, production would peak about 2019, and per capita consumption would decline 26%.

Figure 14.  The IEA "435" forecast would result in a significant decline in per capita production.

A final scenario considers a symmetrical Hubbert Peak in 2015.   This scenario reflects the classic Hubbert theory, that production will peak and decline when we have produced one-half of the USGS mean estimate (3000 billion barrels) for the global oil endowment.  Global per capita production would decline 36% by 2035.   It is worth noting that the American share of per capita production can be expected to decline substantially more than the average, given economic growth in developing nations.

Figure 15.   A symmetrical Hubbert's Peak in 2015 would result in more than 35% reduction in per capita oil consumption.

Conclusions

I think it is likely that peak oil will occur before 2020, with a consequent rise in oil prices.   Although tar-sands, shale-oil, synthetic oil and other substitutions will moderate the decline, I think it is unlikely that the “Shale-gas Revolution” will be replicated in oil.  Oil has a much higher viscosity than gas, and is simply more difficult to extract from rock.  Any unconventional oil source will have a lower EROI (Energy Return on Investment*) than conventional oil, and will be developed more slowly, and at a higher cost.   New sources of conventional oil will also be more difficult and more costly than previous sources.   

As a policy recommendation, I think it is prudent to anticipate higher prices and reduced oil supply, for personal and national planning.

References

My Ideas

* http://dougrobbins.blogspot.com/2011/09/energy-return-on-investment.html

http://dougrobbins.blogspot.com/2010/12/energy-and-productivity.html

http://dougrobbins.blogspot.com/2011/01/us-oil-production-and-oil-imports.html

http://dougrobbins.blogspot.com/2011/08/wealth-of-nations.html

Various Peak Oil Sites

http://online.wsj.com/article/SB10001424053111904060604576572552998674340.html

http://www.mnforsustain.org/oil_peaking_of_world_oil_production_appendix.htm

http://www.oilcrisis.com/us/eia/oilsupply2004.htm

http://www.oilcrisis.com/campbell/endowment.htm

http://www.npc.org/Study_Topic_Papers/10-STG-Geologic-Endowment.pdf

http://pubs.usgs.gov/dds/dds-060/ESpt2.html

http://www.theoildrum.com/node/2409

http://dieoff.com/page140.pdf

http://www.theoildrum.com/

http://www.peakoil.net/

BP Statistical Energy Review

http://www.bp.com/sectionbodycopy.do?categoryId=7500&contentId=7068481

IEA

 http://www.iea.org/

CERA

http://www.ihs.com/products/cera/index.aspx

United States EIA; lots of good data

http://www.eia.gov/

GSwindell; more good data

http://gswindell.com/stats.htm

Global Population

http://www.census.gov/population/international/data/idb/worldpopinfo.php

IEA Documents

http://www.worldenergyoutlook.org/docs/weo2010/factsheets.pdf

http://www.iea.org/weo/docs/weo2010/WEO2010_ES_English.pdf

http://www.iea.org/weo/docs/weo2010/weo2010_london_nov9.pdf

http://www.iea.org/weo/docs/weo2010/key_graphs.pdf

On Progressive Risk-Taking

When I was a young driver, I discovered, to my delight, that the car was narrower than my perception from the driver’s seat.   I found that the car could fit through narrow spaces; I could thread the needle past obstacles in alleys and parking lots.  My confidence grew as I successfully navigated through difficult places.  Until, at last, the inevitable happened.  I tried to drive between a lamp-post and a pickup truck.  The car did not fit.

That’s the nature of progressive risk-taking.   In some situations, we receive no incremental feedback; no warning that things are about to go disastrously wrong.   Outcomes are discrete values of positive or negative, and repeated positive experience only encourages more risk-taking.   [Example: “Well, you didn’t get pregnant the LAST time….”]

A beautiful example of progressive risk-taking is the TV game show “Deal—or No Deal”.   There are 32 suitcases bearing prizes, and a highly skewed statistical distribution of prizes.  In the early rounds, it is a virtual certainty that the contestant will increase the mean (expected) value of his prize by guessing and eliminating suitcases of low value.   Players gain confidence from their early success.  But abruptly, and without warning, the game changes.   Suitcases with high value are inevitably eliminated.   The mean value of the prize declines.   The player, filled with confidence from the early rounds, persists.  Generally, he plays to destruction, to the absolute end of the game where the statistical chance of a substantial prize is very low.   Early success is seductive, and it encourages poor decisions in absolute defiance of all evidence and reason.

Progressive risk-taking is exacerbated when operating outside of the envelope of previous experience.

On a cold Florida morning in 1986, mission controllers decided to launch the Space Shuttle Challenger, despite objections from project engineers.     With a history of 50 successful launches: "What could go wrong?"   But the temperatures that day were without precedent in the shuttle program, and outside of the range for which components had been tested.   Outside of the envelope of experience, the prior record of success is meaningless; the probability of failure cannot be calibrated.  The result was a disaster which scarred a generation, and devastated the American space program.

Humanity is now engaged in a great experiment, outside of the envelope of previous experience.   Near the end of October, 2011, world population will exceed 7 billion people, as compared to about 1.6 billion people in the year 1900.   To date, human ingenuity and technology have enabled mankind to feed and shelter the growing population.   Within the past 50 years, in fact, standards of living, lifespan and health have improved dramatically, coincident with industrialization and rising per capita GDP.  

However, the forecast population is outside of the envelope of previous experience.  Some of the technologies that have enabled the current population, such as irrigation from fossil aquifers, are not sustainable.   We cannot simply assume that technology will always solve the problems of growing population, simply because we have been successful in the past.   There will come a point where we can no longer squeak through the gap.

The Limits to Growth

I just finished reading "The Limits to Growth", 1972 edition, which I should have read a long time ago.  It's an excellent book.   The principal authors were Dennis and Donella Meadows, Carleton College '64 and '63, respectively (yay, Carleton).

The book presents the results of a world economic and social computer simulation.  At the time of the 1972 model development, the authors did not claim that the model was either predictive or quantitative.  Instead, the focus was simply to construct the model and observe the behavior of the simulation.  After the "standard run" base case was created, the model was perturbed in various ways, and the variation between runs was considered.  Model inputs matched historical data from 1900 to 1970, and matched the exponential rates of growth observed through most of the 20th century.  The number of parameters was very small:  Population, Food Supply, Arable Land, Non-renewable Natural Resources, Industrial Production, Pollution, etc.  Feedback loops were created which established relationships between the parameters, and provided for limits on exponential growth.

In almost every case considered, population, food supply, and industrial production rise to a peak, and then fall, sometimes catastrophically, before the end of the 21st century.

According to the 1972 model, world population was forecast to reach 7 Billion by about 2017.   According to current UN estimates, population will reach 7 Billion earlier, sometime in October, 2011.  The LTG estimate was pretty good for a model that was not intended to be predictive or quantitative.  However, we can note some changes in the current picture, as compared to the simulation from 1972.   First, global birth rates have declined much faster than assumed in the model runs.  Secondly, lifespan has increased faster then the model, and I believe that global wealth has risen higher and faster than imagined in the model.  I also believe there we have a much greater ability to expand production of non-renewable resources than considered in the models, although I expect diminishing returns from that expanded production (as described in my blog post regarding Energy Return On Investment).
http://dougrobbins.blogspot.com/2011/09/energy-return-on-investment.html
And the model generalizations regarding "pollution" appear to be unfounded, or uncalibrated.  But there is a good correspondence between atmospheric CO2 and the pollution parameter of the LTG models.

Graham Turner of CSIRO presented a 30-year lookback on the LTG findings, presented here:
http://www.csiro.au/files/files/plje.pdf.
Matthew Simmons, author of "Twilight in the Desert" (R.I.P., 2010), also presented a look-back on the 1972 LTG.  http://www.greatchange.org/ov-simmons,club_of_rome_revisted.pdf

So we are left with the question:  Will food, wealth and population collapse in the 21st century, as predicted in the 1972 world model?  Time will  tell.

Energy Return On Investment

Here is one of the most important graphs I’ve ever seen.  The slide was prepared by Dr. Charles Hall, of the College of Environmental Science, SUNY.

It’s a busy slide, so let’s spend a little time understanding what it says, before we decipher what it means.

EROI

Dr. Hall is a biologist by training, and a professor in the School of Ecology at SUNY.   His early research involved how trout must expend energy in order to eat (i.e., to obtain energy).   In different environments and by using different strategies, animals maximize their survival chances by maximizing their EROEI, their Energy Return On Energy Invested -- expending a minimum amount of energy to obtain a maximum amount of food.   Stated more simply as EROI, the equation is as follows:

If EROI is high, the animal is efficient in obtaining what it needs to survive.   If EROI is low, the animal is inefficient, and susceptible to harm in adverse circumstances.  
By analogy, the same is true of societies.  By maximizing the efficiency with which we obtain energy, we have more time, capital, and energy available to produce food, build housing, and manufacture goods.  If we are inefficient in how we obtain energy, we have less time, capital, and energy available for those things.

Dr. Hall introduced the concept of EROI (Energy Return on Investment) into debates on Peak Oil and our energy future.  It concept is disarmingly simple:  Energy options can be ranked according to how much energy investment is required, in order to produce energy available to society.    The following chart shows a number of energy sources, ranging from hydroelectric, which returns about 100-fold energy to society for the energy invested, to corn ethanol and biodiesel, which essentially are break-even propositions, consuming as much energy as they produce.

High numbers are highly efficient: in the 1930’s, spending one barrel of oil (energy equivalent) to drill an oil well returned the investment 100-fold.  That’s why oil millionaires (e.g. Jed Clampett, Fred Astaire’s “Daddy Long-legs”, J.R. Ewing, and real millionaires, such as the Hunt and Koch brothers) became part of our folk culture.  However, exploration maturity and the law of diminishing returns nibbled away at the excess return.   By the 1970’s, domestic oil had about 30-fold return on investment, and currently, the number is less than 10.  (Keep in mind that there are other financial costs; this is just the return of energy produced compared to energy invested.) 

On Dr. Hall’s chart, the vertical axis is EROI.  This represents the efficiency of a particular enterprise in returning energy to society, compared to the energy consumed in the process.  The horizontal axis represents the volume consumed in the U.S. economy.   The entire United States uses about 100 Quads (quadrillion BTUs).    Note that coal represents about 30% of our current energy budget (about half of our electrical generation), and is very efficient in terms of EROI.  This cheap energy produces economic benefits throughout the economy, but with corresponding environmental costs, including those related to mining, air quality, and CO2 emissions.

Let’s look at how different energy sources appear on the chart, through some of our history.   First, domestic oil in 1930 yielded about 100-fold return on energy invested: for every barrel-equivalent of energy used to drill a well, the well would produce 100 barrels.   Domestic oil production produced about 5 quads annually.

By 1970 (US peak oil) production had increased to about 22 quads annually, but efficiency had fallen.  The EROI for domestic oil was about 30-fold.  

By 2010, production rates had fallen to about 12 quads, and EROI had fallen to 10-fold.  

New oil, and future oil from current exploration, will be located in very deep water, in the Arctic, or extracted with difficulty from low-quality shale reservoirs.  New and future oil is expected to have an EROI of about 3 to 5, which Dr. Hall considers to be the minimum required to sustain civilization (Dr. A.S. Hall, 2010 Complex Systems Lecture, UAA).

Renewable Energy

Now suppose that we want to replace our fossil fuel consumption with renewable energy.    Evidence for climate change is overwhelming, and there is scientific consensus that CO2 emissions from fossil fuels is the cause. 

But currently, Wind energy produces less than one quad of our 100 Quad total demand .   Solar produces less than 1/10 of one Quad.  Despite huge rates of growth, these renewable will not soon make a significant contribution to our energy mix.  

Further, there are likely to be significant limits or barriers to the growth of renewable energy.  For example, electrical system instability limits the use of wind power to about 20% of capacity of installed power, on a name-plate basis (Paul Morgan, electrical engineer with GVEA, personal communication).   And since even the best wind sites have about a 30% capacity factor (annual utilization), the limit of wind energy in a particular system is about 6% .  The availability and cost of rare-earth elements is likely to be another limit to growth, for both Wind and Solar energy.   It’s interesting to note that there are about 150 pounds of rare-earth elements in every large wind turbine.


Biofuels deserve special mention. Ethanol production from corn is supported by Federal subsidies, with are estimated to have cost $17 Billion from 2007 to 2010, and will cost $53 Billion by 2015, if the subsidies and mandated volumes are not repealed earlier.  In theory, corn ethanol contributes to American energy independence and reduces carbon emissions.   But, does it really?  Estimates for the EROI of corn ethanol are very weak, ranging from a maximum of 1.6 to a minimum of 0.9.   (Dr. Hall comments "....as compared to real fuels, which have EROI of 30 or 40).  If we assume the mean, about 1.35, it means that 135 gallons of fuel must be produced in order to deliver 35 gallons of fuel to society.  And in doing so, producing corn ethanol removes food from global markets, raising food prices in a hungry world; consumes large amounts of water; causes soil erosion and degradation of water runoff.  Corn ethanol is an inefficient fuel, and supporting corn ethanol is bad public policy.

So here is the dilemma, graphically.   How do we move renewable energy from the lower left-hand corner of Dr. Hall’s chart, to the far right?  

The short answer is: We can’t.  Or as Dr. Hall puts it:  At the present time, there is no renewable energy solution that is efficient, scalable, and timely. 

Any alternative to oil and coal must be Efficient, Scalable, and Timely.  1)  Alternatives must have a high EROI.   This is really easy to recognize, because energy is a large component in every investment dollar.   Phrased another way, alternative energy must be substantially profitable, without subsidy.   2)  Alternative energy must be scalable, in order to provide meaningful contribution to total energy supply.   3)Alternative Energy must be Timely.  It must be possible to implement the alternative before serious economic contraction begins, reducing the capital available to build the alternative.

So to make the required reductions in CO2 emissions, and provide a significant amount of power from renewable sources, we have to move the goalposts. 

Large-scale conservation is essential. 

We can start by driving smaller cars.    (Amory Levins, Winning the Oil Endgame http://www.oilendgame.com/)  Redesigning our cities to favor carpools, changing our lighting, and other steps can reduce our energy consumption.    But even with the most dramatic conservation measures imaginable, fossil fuels will be a necessary part of our fuel supply for a long time to come.

Conventional Fuel Supply

EROI  has implications for conventional fuel supply, as well.  As old oil fields are depleted, exploration and technology bring new supplies to the market.   For the last 2 decades, and for the foreseeable future, new oil supplies will come from deep water, in various parts of the world, from the Arctic, and from low-quality unconventional reservoirs, which can be produced by new production technology.    However, all of the new sources of oil have lower EROI than conventional sources.  

This brings up something that I call the Economist’s Fallacy.    Standard economic theory assumes that as price rises, additional production will become commercially viable, increasing supply.  

 So if oil can be extracted from oil at $150 per barrel, at today’s prices, economists assume that expanded supply will become available when oil prices rise to that level.  

However, the cost of extraction will rise as the price rises, because energy is a large component of the cost required to extract the oil.    If something requires more energy to extract than it yields, it will not be commercial at any price!  

Major sources of new oil are approaching EROI of 3 to 5, which is the minimum needed to justify production.  Oil which is more remote (e.g. the Kara Sea, the Chukchi Sea, or deepwater subsalt offshore Brazil) will have even lower EROI, and may not be viable at any price.  More energy is required to extract each barrel of oil, consuming what might have been available for other purposes in the economy.   This progression, requiring disproportionately more energy to produce more energy, may become a factor which will exacerbate Global Peak Oil.  

http://www.wesleyan.edu/econ/seminar/2009s/hall.pdf

http://www.theoildrum.com/node/3786

http://www.theoildrum.com/node/3800

http://teotw.net/energy_EROI.html

http://www.eia.gov/renewable/
http://www.ewg.org/files/EWG-corn-ethanol-energy-security.pdf

http://onlinelibrary.wiley.com/doi/10.1111/j.1749-6632.2009.05282.x/full