Good forecasts/bad forecasts: how does the US DOE/EIA come out?
Forecasts are only of value if they are reasonably accurate. In that regard, I decided to look at an analysis I made in 2001, “Analysis of the IEO2001 Non-OPEC Supply Projections”
In the document, I analyzed the U.S. Department of Energy/Energy Information Administration’s (US DOE/EIA) International Energy Outlook 2001 (IEO2001) and compared their forecast to my forecast for selected non-Organization of Petroleum Exporting Countries (non-OPEC).
The following update compares the accuracy of the US DOE/EIA's forecasts with the accuracy of my own forecasts that were presented in that document five years ago and an analysis I made of North Sea oil written in 1999.
Future North Sea Oil Production
In the IEO2001, the US DOE/EIA stated the following concerning future North Sea oil production:
In the IEO2001 forecast, North Sea production reaches a peak in 2006, at almost 6.6 million barrels/day (mb/d). Production from Norway, Western Europe’s largest producer, is expected to peak at almost 3.7 mb/d in 2004 and then gradually decline to about 3.1 mb/d by the end of the forecast period with the maturing of some of its larger and older fields. The United Kingdom is expected to produce about 3.1 mb/d by the middle of this decade, followed by a decline to 2.7 mb/d by 2020.
I had stated in my analysis that a peak as late as 2006 followed by a minor decline from 2006 to 2020 was totally unrealistic. I stated that there should be a significant decline between 2001 and 2006. I further stated that North Sea crude + condensate production would probably peak in 2000. According to US DOE/EIA data, I was off by 1 year: production peaked in 1999 at 5.948 mb/d.
How has production been going in the North Sea lately and how does that compare to the US DOE/EIA’s IEO2001 forecast as well as my forecast? Although the US DOE/EIA doesn’t make it clear, the IEO2001 analysis is for total petroleum production rather than oil. I consider oil to be crude oil + condensate. Total petroleum production includes crude oil, condensate, natural gas liquids, refinery gain and other hydrocarbon liquids. Crude oil + condensate (oil) made up 92.4% of total petroleum production for Norway and 88.9% for the U.K. in 2004.
Table 1 contains data of IEO2001 projections, my projections based upon a paper I wrote in 1999 entitled “The Impact of Declining Major North Sea Oil Fields Upon Norwegian and United Kingdom Oil Production” and actual production figures. IEO2001 projections are in terms of oil, calculated by multiplying the total petroleum production value by the fractional value of oil production for Norway and the U.K. North Sea values are in terms of total petroleum production.Table 1: Projected and Actual Production Data (mb/d) for Selected Yearsb
|Actual||IEO-2001||Actuala||IEO-2001||My forecast||My forecast||IEO2001|
|a Through the first 9 months of 2006|
b Actual production figures are from the US DOE/EIA
According to US DOE/EIA production data, in the first 9 months of 2006 North Sea total petroleum production was 4.77 mb/d, or ~1.8 mb/d less than the IEO2001 forecast for 2006 (~38% error relative to actual production rate). Norwegian oil production peaked in 2001 at 3.226 mb/d. In 2004, Norwegian oil production was down to 2.95 mb/d whereas the US DOE/EIA forecast was for a production rate of ~3.42 mb/d (~15% error relative to actual production rate). U.K. oil production peaked in 1999 at 2.68 mb/d and was down to 1.65 mb/d by 2005. The IEO2001 projected that U.K.’s oil production would be ~2.76 mb/d in 2005 (~67% error relative to actual production rate).
Based upon 2006 actual production values, it appears likely that my 2010 projection may be a little low for Norway and a little high for the U.K. The Buzzard field, in the U.K. sector of the North Sea, will slow the rapid decline of U.K. production and may make my 2010 projection error minor as would an increasing production decline rate for Norway (If the decline rate of the last 2 years were to continue through 2010, Norway oil production rate in 2010 would be 1.8 mb/d rather than 1.6 mb/d). Note that the actual oil production figures in 2006 are considerably less than the IEO2001 projections for 2020. There was sufficient field data for North Sea fields back in the 1990s to make reasonably good projections for future North Sea production. Obviously the US DOE/EIA doesn’t use field data when making projections of future oil production.
Future Mexican Oil Production
The IEO2001 stated the following concerning Mexico’s future oil production:
Mexico is expected to adopt energy policies that encourage the efficient development of its vast resource base. Expected production volumes in Mexico exceed 4 mb/d by the end of the decade and show little decline out to 2020.
I stated that Mexico’s oil production relied heavily on production from the Cantarell complex and that production from Cantarell was poised to decline at a rate of 10%/year or more. I stated that the decline of Cantarell production would drag down Mexican oil production in the future. I further stated that it was likely that Cantarell would peak in 2002. Cantarell actually peaked in 2004 so I was off by 2 years on that account.
Recently the chief executive of Pemex, Luis Ramirez Corzo, stated that Cantarell production would decline 14%/year from 2007 to 2015. This year, Cantarell’s production is projected to be ~330,000 b/d less than in 2004. A 14% decline rate from 2007 to 2015 would mean that Cantarell’s production would decline ~1.7 mb/d from 2004 to 2015.
Although there has apparently been rapid development of the Ku-Zaap-Maloob project and exploration and development spending by Pemex has increased dramatically since the 1990s, it appears that Mexican oil production peaked in 2004. Oil production through the first 9 months of 2006 was down ~76,000 b/d from the 2004 average.
Although Mexico has prevented a significant decline in total petroleum production up to this time, it will be increasingly difficult to keep production up. If Cantarell declines at 14%/year, its production would decline ~800,000 b/d from 2006 to 2010. It appears unlikely that Mexico’s total petroleum production will reach 4 mb/d this decade and the production decline by 2020 will be substantial.
Future Oil Production in Colombia
The IEO2001 stated the following concerning Colombia’s future oil production:
Colombia’s current economic downturn has somewhat delayed its bid to join the relatively short list of worldwide million barrel/day producers, but its output is expected to top a million barrels/day within the decade and show little decline for the remainder of the forecast period.
I stated that the 3 largest fields in Colombia (Cusiana, Cupiagua and Cano Limon), which were producing ~60% of Colombia’s oil in 2001, were in rapid decline and the idea that Colombia would become a million barrel/day producer in this decade were baseless. For the first 9 months of 2006, Colombia’s average oil production rate was 533,000 b/d, which would make Colombia’s total petroleum production ~548,000 b/d, only a little more than half of the 1 mb/d production level the US DOE/EIA was predicting for sometime this decade. The production rate this year is up slightly over 2005 but no one should expect any sizeable increase in Colombian oil production during this decade, or any future decade. Colombia’s oil production peaked in 1999 at 816,000 b/d.
Future Oil Production in Argentina
The IEO2001 stated the following concerning Argentina’s future oil production:
Argentina is expected to increase its production volumes by at least 100,000 b/d over the next 2 years, and by the middle of the decade, it could possibly become a million barrel/day producer.
I had stated that the IEO2001 statement concerning Argentina was totally unrealistic. From 2001 to 2003 Argentina’s total petroleum production declined from 876,100 b/d to 865,800 b/d. Argentina’s oil production has declined from 802,000 in 2001 to 698,000 for the first 9 months of 2006, which would make Argentina’s total petroleum production ~783,000 b/d. Argentina will not become a million barrel/day total petroleum producer.
Future Oil Production in Canada
The IEO2001 stated the following concerning Canada’s future oil production:
Canada’s projected output is expected to increase by more than 200,000 b/d over the next 2 years, mainly from Newfoundland’s Hibernia oil project, which could produce more than 150,000 b/d at its peak sometime in the next several years. Canada is projected to add an additional 600,000 b/d in output from a combination of frontier area offshore projects and oil from tar sands.
It reads as if they expected an increase of more than 200,000 b/d from Atlantic Canada and an additional 600,000 b/d from other areas of Canada, mostly from tar sands, during the period 2001 to 2003. I’ve since looked at the IEO2003 statement for Canada and the US DOE/EIA had inserted a few words in the first sentence to state “Canada’s projected conventional oil output …”, so they did mean that the 200,000 b/d was basically from Atlantic Canada, where there was rapid development of the Hibernia, Terra Nova and White Rose fields.
I had stated that a 200,000 b/d increase, which I was assuming was for Atlantic Canada, appeared unattainable. Atlantic Canada’s oil production did increase 188,000 b/d between 2001 and 2003 but since then it has declined from 336,885 b/d to 288,629 b/d for the first 10 months of 2006. The 188,000 b/d increase was achieved by producing from Hibernia and Tera Nova at higher peak/plateau rates than were originally projected. The disadvantage of producing at higher peak/plateau rates is that the fields have started to decline sooner than they should have. Atlantic Canada’s oil production is in long-term decline.
If the 600,000 b/d increase beyond Atlantic Canada (assuming total petroleum production) was meant to be for the 2001 to 2003 period, it was not close to being realized. The total petroleum production increase for all of Canada from 2001 to 2003 was 297,000 b/d, so the increase beyond Atlantic Canada was ~109,000 b/d. Since 2003 the total petroleum production increase has slowed even with intense development of tar sands in Alberta. The increase from 2003 through the first 9 months of 2006 is ~50,000 b/d based upon US DOE/EIA data.
Future Oil Production in China
The IEO2001 stated the following concerning China’s future oil production:
In China, oil production is projected to decline to 3.0 mb/d by 2020.
I had stated that the 3 largest fields (actually the 3 highest producing oil regions) would experience significant production declines in coming years that would drag down China’s future oil production and that a production level of 3.0 mb/d in 2020 appeared very optimistic. Those three regions have been experiencing declining production since 2001 but there has been intense offshore oil development as well as intense oil development in western China.
China’s total petroleum production has increased from ~3.4 mb/d in 2001 to ~3.8 mb/d in 2006. It appears likely that China’s oil production will peak around 2008. When oil production from offshore regions and western China start declining, it will add to the decline from Daqing, Shengli and Liaohe oil regions. I will state again that a total petroleum production rate of 3 mb/d appears very optimistic for 2020.
Oil Production in India
The IEO2001 stated the following concerning India’s future oil production:
India is expected to show some modest production increase early in this decade and only a modest decline in output thereafter.
I had stated that production from India’s largest field, the Bombay High field, was declining and that it would drag down India’s oil production in the future. Since 1997 India’s oil production has stayed relatively stable with production averaging between 640,000 b/d and 685,000 b/d. That relatively stable production rate is due basically to two factors: the number of producing fields has increased significantly and Phase II development of the Bombay High field has reduced the rate of decline for the field. These admirable efforts can not be expected to prevent a decline in India’s oil production for too much longer. By 2020, India’s oil production will be considerable less that it is at present, probably around 300,000 b/d.
The US DOE/EIA doesn’t perform realistic analyses for the future production capabilities of countries. Their objective appears to be to provide happy projections that fit their assessment of future global consumption.
The mass media in the U.S. reports on forecasts and assessments by the US DOE/EIA and the United States Geological Survey (USGS) as if they represent the best possible forecasts and assessments possible.
My experience is that the media has little interest in reporting evidence that contradicts the forecasts and assessments of the US DOE/EIA and USGS, or that of happy-talking economists, no matter how well documented the evidence may be.