Comments by Jean Laherrere
jean.laherrere at wanadoo.fr
on the paper "Challenges and Concepts for Long Term Oil and Natural Gas Supply Modeling"
presented by Vello A. Kuuskraa (Advanced Resources International, Inc, vkuuskraa at adv-res.com) at PEW Workshop on Oil and Gas Markets Stanford Energy Modeling Forum #19 Snowmass, CO August, 2001
Addressed to Vello Kuuskraa on Sept. 20, 2001
Your paper has some goods points, but others are questionable and I will discuss in detail some pages, bringing new data and a few graphs to support my views.
Page 6: Hubberts outlook for US natural gas
You compare 1974 Hubbert forecast around 6 Tcf/a for 2000 with the present current annual production (all US with unconventional) at 19 Tcf/a, but Hubbert forecast was only for conventional Lower 48 production.
The most important was in 1972, when dry production had been rising from 6 Tcf/a in 1950 to 21 Tcf/a in 1970, forecasting a decline towards 15 Tcf/a in 1985. US gas production did peak in 1972 and was about 16 Tcf/a in 1985. Hubbert missed the second cycle. It is the main weakness of Hubbert modeling to use only one cycle when in reality there are several cycles. Hubbert mentioned the possibility of several cycles, but he did not display any example or model with two cycles. So he was right from 1972 to 1985 and wrong from 1985 to 2000! No one is right all the time!
Page 7: Why Hubbert was pessimistic!
Hubbert was not the only one to be wrong on gas, but he was right on oil. Your arguments to be wrong on gas by not taking into account technological progress as 3D and horizontal drilling, should also apply for oil! So the reason is not in technology progress.
In his book "Alternative energy resources" in 1976, James Hartnett gave on page 35 his forecasts on oil & gas US production until 2020. He forecasted for oil a continuous decline of the Lower 48, a new cycle with Alaska Prudhoe Bay since 1975 and another cycle due to EOR (Enhanced Oil Recovery), expecting 2.1 Gb/a in 2000 for conventional oil and another 1 Gb/a for EOR. In fact oil US production was at 2.1 Gb/a, but the dreams of EOR did not come true. But for natural gas he forecasted also Alaskan production in addition to declining Lower 48 with a total of 15 Tcf/a in 2000 and an additional 12 Tcf/a for new techniques (27 Tcf/a in 2000), but all declining since 1985. In fact North Slope gas is not there, total production for 2000 was about 18 Tcf (dry), but rising since 1985. Hardnett was all wrong.
Why has gas production been increasing since the trough of 1985? It comes in part from gas wells but also surprising in part from oil wells, despite the decline of oil. As shown in the following graph, gas production from oil wells follows the oil production from 1950 to 1985, but from 1985 oil declines and gas rises. Why? It is well known that at the end of an oilfield, associated gas production rises. The increase of gas production from oil wells seems to indicate the near end of most US oilfields. This behavior was not forecasted by most experts.
The technology of the 70s already included 3D and horizontal drilling. These techniques have increased drastically the way to produce, but hardly the way to discover. I challenge you to show me a giant oilfield which is not visible on 2D seismic. Furthermore in offshore, 3D is compulsory when using seismic ship with more than a dozen streamers.
Page 9: Natural gas resource pyramid
The (US) resource pyramid adds reserves and resources: they are not the same thing; it is as adding oranges and apples! What about the reserves pyramid?
It is interesting to notice that the reserve growth is on one side of the pyramid, but at the same level as undiscovered resources, as being different. Reserve growth is mainly due to the US's poor practice of reporting only proved reserves. USGS with Chuck Masters was not reporting proved reserves but inferred reserves, preventing reserve growth.
USGS with Ahlbrandt is now counting on large reserve growth on the US proved data and extending it also to the rest of the world which uses proven+probable, despite their stating that they do not have any reliable study on it. Furthermore USGS estimates growth from very old fields developed with old techniques and wants to apply to new discoveries found and developed with new technology.
USGS (US DOI USGS fact sheet FS-119-00http://pubs.usgs.gov/factsheet/fs119-00/fs119-00.pdf "Reserve growth effects on estimates of oil and natural gas resources")
Take the best example of reserve growth, the oilfield of Midway-Sunset discovered in 1897 and producing heavy oil (13°API) (under Campbells definition it is unconventional oil). On the next graph, Midway-Sunset production has increased since the beginning with the number of producing wells, up to 60 Mb/a with 10000 wells during the 90s. It is interesting to note the ultimate reserve grew versus the production or the number of producers, showing that it is a poor estimate! The ultimate is presently estimated at 2.7 Gb but it is likely that the real value is about 3.5 Gb.
It is obvious that the reported value has nothing to do with the technical data and is very conservative as estimated with future low oil price. An oil company in large fields wants to keep some potential of growth for future bad years. This kind of growth should not be used to model the growth of US new proved discovery (as deepwater), and furthermore in the rest of the world to new proven+probable discovery. The USGS reserve growth estimate on the rest of the world is presented as if the world reserves will increase if they adopt the US wonderful practice (which leads to such poor evaluation with 90% of new additions being revisions of past discoveries).
"Exporting U.S. technology and know-how should raise the yields in other countries by about 700 billion barrels, says Thomas Ahlbrandt, the head of the USGS study."(US News & Report Sept 2001 David Lagesse). The rest of the world oil industry is incompetent, but thanks to US technology, their reserves will grow by 700 Gb!
The estimates of initial reserves of the gasfields in the GOM have been studied on the 12 gasfields (over 1 Tcf) found from 1948 to 1960 from the MMS data and the evolution from 1975 to 1998 taken as a percentage of the last published value (1998) The next graph shows that on average the 1988 value was already at 1999 value, then diminished to get back to 1998 value. This oscillation means that growth should be over.
Page 10: Expansion of natural gas resource base: the resource estimate by the USGS
There was no increase from 1981 to 1989 and the increase in 2000 is due to the inclusion by Tom Ahlbrandt of reserve growth when his predecessor Chuck Masters did not include any reserve growth when estimating inferred reserves instead of proved reserves. Masters, being retired and replaced by Ahlbrandt, however issued an official USGS report OFR98-468 giving a different view from USGS 2000 (http://greenwood.cr.usgs.gov/energy/WorldEnergy/DDS-60/ESpt4.html) which is in fact for the US study of 1998
|Masters 98-468||Ahlbrandt 2000|
|Ultimate||1438 Tcf||1908 Tcf|
What is the report referenced as USGS/MMS 1995-2000 giving the value of 2492 Tcf?
The authors of these estimates show a poor understanding of the concept of accuracy by giving 4 significant digits when only 2 should be given, as 2500 Tcf and not 2492 Tcf (furthermore there are adding proved reserves, which is wrong, as Capen 1996 stated about "illegal addition of proved reserves"). It was the same with the ridiculous 3003 Gb oil estimate from USGS, meaning that it is not 3004 or 3002!
I am afraid that you compare estimates of only conventional gas with estimates including unconventional gas as coalbed methane. You are comparing apples and oranges!
In the report USDOE/EIA AR 1999 the total gas resources including unconventional (not included in the 1981 & 1989 estimates) are reported as 1245 Tcf, with around 1000 Tcf already produced, the ultimate is still short of the 2500 Tcf of your 1995-2000 USGS/MMS estimate.
Page 17: fixed and dynamic resource base
This concept is mainly "wishful thinking" from economists. Geologists know more about change and evolution with time than economists do, as they study them on periods of hundreds of millions years. Explorers and producers are always looking how to improve discovery and production, nothing is fixed. Technology allows one to produce cheaper and faster, seldom to increase the conventional reserves (but on the contrary for unconventional oil and gas). The best proof is that the depletion of gas wells has been increased drastically over the last 20 years, being now over 50%/a in the GOM.
What is wrong is the US practice of estimating reserves because the SEC rules of omitting to report probable reserves as is done in the rest of the world. This is why from 1988 to 1999 gas new field discoveries represent only an average of 5% of the total reserves additions, and the revisions of the past discoveries represents 95%, meaning only that the past estimates are pretty lousy. US practice for reporting reserves is very poor (as obsolete as the punched cards of the last presidential elections!). This poor practice leads to a strong reserve growth wrongly attributed to new technology when in fact it is due to poor methodology. But such practice allows oil and gas companies to report growth even without any discovery and they love it.
Page 18: the resource base might continue to expand by including the unconventional gas, but in your presentation in IIASA June 2001 your forecast for unconventional gas is not very optimistic as the average is almost flat for the next 30 years at less than 5 Tcf/a compared to 4.5 in 2000. Kuushraa V, D.Beecy, A.Kydes "the Energy Modeling Forum study on technology and global climate change policies" IIASA-IEW June 19-21 2001
Page 23: decrease in finding costs
Why take as a start 1980 before the oil shock? Data are available since 1960. It is unfair to take as a starting point the highest value.
Gas finding costs depends highly on gas drilling costs, which depends almost entirely upon the wellhead gas price of the previous year as shown in the following graph.
Page 25: lower 48 drilling cost per well:
Why only 1970-1989? And as above what about the relationship with oil & gas wellhead price?
Page 28: large volume of oil and natural gas exist in moderate size and small fields.
In the world outside US+Canada the volume of oil +condensate in fields up to 50 Mb (called medium fields in the US) represents only 6% of the total discovered but 82% of the number of oilfields, in fields up to 500 Mb (giants) the volume is 25% but 97% in number. The volume of gas in fields of 300 Gcf (50 Mboe) is only 7% of the total discovered but with 86% of the number of gasfields, in fields up to 3 Tcf (500 Mboe) the volume is 28% but 98% in number
So outside the US, there is not a large volume of oil or gas in moderate and smaller sized fields.
Furthermore many of these uneconomical discoveries will never be developed if they are not connected to larger fields before the depletion of the basin.
In the US, the last published distribution was in 1991: "Global oil, gas fields, sizes tallied, analyzed" OGJ April 29, 1991 by Ivanhoe & Leckie. It is as follows as number and the volume is estimated from the average size of the interval:
|size||ultimate Tcf||number||volume Tcf|
|very small||Tcf 0.01-0.1||2 927||90|
|tiny||Tcf 0.001-0.01||5 298||16|
|insignificant||Tcf 0-0.001||6 512||3|
The graph in percentage of total discoveries shows that the fields up to small size (maxi 0.25 Tcf) represents only 26% of the volume but 97% of the number. Up to the medium maxi represents 38% of the volume and 98% of the number, far from being a large part if large means the majority.
Page 29 & 30: the case for giant fields:
It is written: "so-called "creaming curves" (Laherrere 1999) using volume of discovery versus number of fields": This is wrong: the creaming curve (introduced by Shell) displays the cumulative discovery versus the cumulative exploration activity, being the number of New Field Wildcats (NFW). It is better than a graph of cumulative discovery versus time as it erases the stop & go of exploration. It is a graph showing the well-known "law of diminishing returns" in mineral exploration. Usually the creaming curve displays one or several hyperbolas (with several cycles of activity). It is well shown by this Norwegian graph where there is only one cycle of explorarory activity:http://www.isv.uu.se/aleklett/Iva/sld018.htm
The UK graph from Michael Lynch on page 30 called a creaming curve is not a creaming curve, showing his poor understanding of exploration concepts.
I sent a mail to Michael mentioning some wrong points in his paper, with many data as proofs, but I did not get any answer.
The real UK offshore oil creaming curve is as follows, having discovered 26 Gb with about 300 NFW. If the number of discoveries does not show any sign of bending, the volume of discoveries slows down. To model it, the best fit is with 2 hyperbolas, one for the period 1966-1985 and the second for 1985-2000. The annual discovery versus time displays 2 exploration cycles with a strong trough between; the second hyperbola must correspond to a different objective and/or technique. Another 2000 more NFW would give only an additional 3 Gb, but another cycle (the Faroes?) could be added.
The US creaming gas curve gathering a very large number of gasfields displays only one hyperbola but a little bumpy (the data is less reliable because the correction to "mean" values). It took more than 300 000 NFW to find around 1150 Tcf and the model shows that another 300 000 NFW would bring only less than 300 Tcf!
Page 32: distribution of Frio Strandpoint play
Good example of "diminishing returns", unfortunately old data, only up to 1985. What about the new technology impact?
Page 38: Is there evidence of resource depletion?
It is written: "Natural gas displays no level of depletion for either time or effort: reserve additions are up from the 1970s and 1980s and productivity (per well or foot drilled) is up by two fold" Of course when reserves are reported with a poor practice as proved reserves, neglecting the probable, 88% of additions for the last 20 years come from old discoveries. When measured with "mean" (or 2P) reserves and backdated to discovery year, the new discoveries of the 90s are better than the 70 & 80s, but they are on the model, and they are not as bright as the 30s as shown in the above creaming curve, furthermore these are very unreliable as grown with an obviously overestimated model of reserve growth.
Page 41: Examples of technology progress:
San Juan basin CBM
CBM is interesting as the US has the most experience. Most of the production in the past was in the San Juan basin where special techniques were developed, and everyone thought that this new technology (for unconventional gas) would do miracles outside, but it was found that the Powder River basin, with a different kind of coal, needs a different technique.
In a paper "The next gas crisis" August 20, 2001 http://www.canadianbusiness.com/magazine_items/2001/aug20_01_thenext.shtml
Andrew Nikiforuk wrote: "US analysts typically sum up the situation this way: "It's the geology, stupid.""
Geology is the most important parameter in oil and gas production; techniques have to be adjusted to each kind of field. Technical progress is also important but related to geology. The complexity of what is left to produce needs more scientific thinking. The decrease in the number of geoscientists (and students) is one of the problems in dealing with future discovery and production
I am not pessimistic about supply meeting demand, as it will be regulated by price, but future demand is far too optimistic.
It is difficult to model supply but the best is to keep it simple: what is produced has been found in the past. The best model is to use the "mean" discovery and to fit it to the production with a certain shift. The shifted discovery provides the forecast of future production. The problem is that the reported discovery, as the US proved data, is far from the "mean". A certain amount of correction is needed. The mean gas discovery for North America gives a fairly ggod fit with production.
But the future for the next 10 years is a significant decline!
I look forward to discussing all these points; sorry for my broken English!