I doubt that one can trust BP for accurate data on reserves. The following is from the first issue of the Hubbert Center Newsletter http://hubbert.mines.edu
---------------------------------------------
OIL RESERVES AND SEMANTICS
by
L.F. Ivanhoe
The prestigious British newsmagazine “The Economist” stated in its 10/07/95 feature article ‘The Future of
Energy’:
“Fossil fuels supply over three-quarters of the world’s energy needs...In the energy needed to
move transport, oil is still king, supplying 97% of the fuel used...Truly modern renewables such
as solar and wind power provide less than 1% of world electricity...Proven reserves of oil are
now enough to supply the world for 43 years at current rates of production, compared with 35
years during the 1970s, according to BP, one of the oil giants...But even given the growth in
demand, the proven reserve figures probably underestimate the longevity of fossil fuels... ‘I
don’t believe that we will be out of oil in 40 years’, says Sean O’Dell, chief economist of the
International Energy Agency.”
Such optimistic statements are typical of the views of most economists in financial newspapers and magazines.
Unfortunately, their opinions are often based on research quoting questionable data. For example, the much-quoted
annual ‘BP Statistical Review of World Energy’s’ tables and graphs on ‘Distribution of oil reserves in 1994’ contains a
fine-print footnote:
“...(reserve) estimates contained in this table are those published by the Oil and Gas Journal
(O&GJ) in its issue of 26 Dec. 1994, plus an estimate of natural gas liquids (NGL) for North
America.”
O&GJ RESERVES
One must go back to the O&GJ annual year-end reports to check each of the oil producing nation’s oil and gas
reserves / production. On close examination, one soon runs into problems with the O&GJ numbers. The O&GJ merely
compiles the RESERVE information provided annually by each country’s government source. There is no way for the
O&GJ to check on the accuracy of foreign reports. To many foreign ministries, the O&GJ requests for reserve data was
either a sensitive state secret or a nuisance chore that no one was critically concerned about. Due to lake of guidance or
ignorance, a common answer was apparently “same as last year”. (52% of the 67 nations listed in O&GJ’s 25/12/95
report gave the identical oil reserve numbers as on 26/12/94; many for several years.)
Some nations’ numbers are obviously gross approximations. IRAQ doubled its reported reserves from 47 Bbo to
100 Bbo (Billion barrels of oil) in 1987, the number that IRAQ still lists for its reserves. Who could prove IRAQ wrong?
And what difference did it make to IRAQ if the world’s economists and planners were misled? By 1989 all of the OPEC
nations had raised their reported reserves to maintain everybody in line for their “oil quotas”. All of this increase was
creative bookkeeping “political reserves” rather than new oil discovered. These are the figures that bolster the apparent
change as reported by The Economist, that the world potential of reserves / production (R/P) increased from 35 years in
the 1970’s to 43 years in 1995.
Operating oil companies pay only limited attention to details of the O&GJ reserve numbers which are the domain
of academic planners. It is simply impossible to check foreign reports. Caveat Emptor! If the USA / UN will not trust
IRAQ’s Saddam Hussein to correctly report details of his military might - how can they trust his report on IRAQ’s OIL
RESERVES (his economic might)? The same question might be raised for all nations who get foreign government and
World Bank grants and loans based on large reported RESERVES. In summary - don’t trust the O&GJ (or BP’s) reserve
numbers implicitly - some 300 Bbo of political reserves may need to be subtracted therefrom.
In short - OPEC nations’ claimed reserves are intended to increase their “OPEC PRODUCTION QUOTAS” -
not to evaluate their oil as in other countries. Consequently, relating OPEC’s with other nations’ oil reserves is like
comparing apples with oranges (see Figure 1).
USGS OIL RESOURCES
The possible exaggerations in the O&GJ reserve numbers don’t bother me nearly as much as the economists’
treatment of the U.S. Geological Survey (USGS) “RESOURCE” numbers. These are commonly added to the O&GJ
RESERVES to produce a grand total of each nation’s “OIL ENDOWMENT”. The sum of two (unknown) values gives
huge numbers which are routinely and incorrectly called “RESERVES” - rather than “undiscovered RESOURCES”. The
equation should be:
(known) RESERVES + (undiscovered) resources = (unknown) resources.
The two technical terms (known) RESERVES and (unknown) RESOURCES are virtually synonymous to the
public - but not to the oil industry. Known RESERVES are calculated by the petroleum reservoir engineers and
conservative bankers will lend money on such RESERVES. RESOURCES, however, are estimates by scientific
geological committees, of the total amount of oil they feel may be present but still undiscovered in any area. Conservative
bankers will not lend money on undiscovered RESOURCES.
USGS geologists now prepare RESOURCE numbers for all nations, based on elaborate computer programs and
global data bases. Nobody else could do a better job. Unfortunately, the basic problem - like earthquake predictions - is
beyond the ability of any scientific committee to solve. The USGS uses a statistical formula that gives estimates of the
PERCENTAGE PROBABILITY OF NEW DISCOVERIES (oil volumes) and their reports have lots of guesses,
assumptions and fine print qualifying their methodology. Their final computer printouts are very impressive. However,
few operating oil companies concern themselves much with the USGS RESOURCE numbers, which are designed for
global planners/economists rather than working oil people who have different problems. Government and academic
scientists are always much more optimistic about the chances of new discoveries than are their oil company colleagues
who are regularly humbled by dry holes drilled on the very best of data.
The 1994 USGS undiscovered world oil RESOURCE numbers were:
95% probable (sure things) = 292 Billion barrels oil
50% probable (likely) = 582 Billion barrels oil
5% probable (unlikely) = 1005 Billion barrels oil
But the actual oil discovered in new global fields discovered during the 1982-1991 decade was only 91 Bbo, at
the declining end of a global finding curve (see Figure 2). A straight extrapolation of the 1982-1991 finding rate would be
180 Bbo during the next 20 years...far less than the USGS 95% (sure thing) numbers! This USGS discrepancy may be the
result of errors in their basic computer model assumptions - or the lack of any TIME limits in their computations. Oil to
be found more than 20 years in the future- is effectively worthless today.
If “Reserves” are “apples”, and OPEC “quota reserves” are “oranges”, then USGS “Resources” might be
“olives.” Combining them results in “fruit salad” - not “apples.”
SUMMARY
Unfortunately, planners who trust the O&GJ and USGS numbers base their economic projections on overly
optimistic data. It is now a question of WHEN-not WHETHER the foreseeable world oil supply shortage will confront us
sometime between 2000-2010 AD when the global demand will exceed the world’s oil production...to the surprise and
dismay of those who believed the superoptimistic economists.
L.F.Ivanhoe #HC96/1-2
7/18/96

Although BP provides regional totals, I want a little more flexibility. In a future version I'd like to allow users to select any combination of countries they want. "North America" is a fine grouping, but how about "English Speaking Nations" or BRIC or Commonwealth, etc.? And you'd want to see each country in the group colored by import/export. That's the plan at the moment.

It seems like BP's 2007 energy report came out in June last year. Perhaps the 2008 report (showing 2007 data) will be out shortly.

I noticed in looking at the report that Iran has been a net importer of natural gas for several years. Also, Russia's natural gas exports have been approximately flat for quite a few years. If they add lots of new pipelines, will there be more gas to go around?

The EIA web page is very slick and contains many more countries than the BP dataset but I would submit that their presentation of the data makes it difficult to understand the interaction of production, consumption and natural decline.

Their data does include 2007 but only goes back to 1980 so that it's harder to see the shape of the production curve in each nation. That is the real story behind these data and I believe it should be made as visible as possible.

I think the problem is that what is being presented is NET imports/exports. i.e. it looks to me like the source data is simply production and consumption. The difference between the two is calculated and presented as either imports (red) or exports (green).

http://www.eia.doe.gov/pub/oil_gas/petroleum/data_publications/company_l...

shows the U.S. as importing about 1.8 million barrels of crude oil a day from Canada, and I think substantially all oil produced in Canada is exported to the U.S.. If net exports are 1 million barrels per day (estimated from mazamascience graph), then 0.8 million barrels per day is imported into eastern Canada and 1.8 million barrels per day are shipped (piped) to the U.S. from western Canada.

How about some more real life case histories? We (my idea, Khebab did the heavy lifting) looked at two radically different producing regions, the Lower 48 and Russia, and generated prediction production curves based only on the data to 1970 and to 1984 respectively. The Lower 48 post-1970 and Russian post-1984 cumulative production was quite close to what the HL models predicted. And I did use the method to warn of near term production declines by both Saudi Arabia and Russia.

Why not extend this analysis (use the data through the peak to predict post-peak cumulative production) to some other post-peak regions, such as Indonesia, the North Sea and Mexico? The key of course is whether or not the data set shows a relatively stable pre-peak HL plot.

Well done!

Damn. I am going to have to sit and think about that one for a while.

Very nicely done, WHT!

Good work, clearly a solid foundation has been established. What a shame that WHT is either unable or unwilling to write for a general audience, or at least an audience that does not have an advanced degree in calculus and statistics, which would be just about everybody.

I ran across an article that does a reasonably good job of explaining some of the basics of mathematical modeling:

http://www.causascientia.org/math_stat/Tutorial.pdf

although still fairly advanced for those who didn't get much further than college level algebra, even so it's worth a read for those who would like to shed some light into the impenetrable gloom of "probability density functions".

Cheers,
Jerry

Per capita is one of the very next scheduled improvements. I intend to use the data available at the US Census Bureau IDB unless anyone has a better suggestion.

"off topic: One reason for the industrial decay of the US is the non metric system."

Are you serious? In the age of the computer?

Besides, metrics are used when it matters, like in engineering and chemistry/physics (9.81m/s2 pops to mind right away..)

Cheers from Munich, Dom

I would love to get a dataset that went back to 1900 so that we could see all the nations that once produced and their complete production profiles. If anyone knows of a suitable dataset, please let me know.

And no disagreement about the metric system. However, in order to communicate unfamiliar concepts to people one must use familiar language. And bbls is the defacto standarad unit for oil in the English language. When I translate the databrowser into other languages we can talk about appropriate units or even allow for unit conversion.

As to Germany's production -- Yes, it is a producer. But not in enough quantity to register in BP's dataset. Have a look at the EIA statistics for Germany ~ 150,000 barrels per day.

Very nice idea. Do you know of a dataset that has things organized this way?

World Total is, of course, one of the rows in the BP Stastical Review worksheet. The production and consumption numbers are almost exactly equal resulting in no net exports/imports and the same production curve we've seen in many places. It's not that compelling a graphic but should definitely be included for completeness.

The reason this and other groupings are not in the databrowser yet revolves around programming issues more than data issues. The data are all there in the dataset. But I need to be careful how I build this functionality into the system. If I design things properly I should be able to allow several desirable features:

• a world map colored by importer/exporter
• named groupings like "OPEC" or "EEC"
• user-selected groupings of anything they want with individually colored countries

These are all very desirable but I need to worry about the following as I build this functionality into the databrowser:

• how to pass lists of countries around in the code
• how to calculate a single summary prod/cons/import/export graph yet color individual countries
• how to construct reasonable length image file names from long lists of country names

It is my hope that patience with my slow progress will be rewarded with a useful and stable tool that withstands the test of time.

By all means include the charts wherever you see fit. You may choose whichever level of attribution you wish though my preference is the one at the top of the list:

1. Brief comment in the HTML text mentioning the databrowser, why you like it and giving its URL
2. URL to the databrowser only
3. Just include the image
4. Just include the image but carefullly crop off the bottom pixels where it says: "Data: BP Statistical Review 2007 Graphic: mazamascience.com" ;-)

This databrowser is my way of advertising my skills and the more people that see the data graphics and appreciate why they are good, the better.

If you click on an image you will download a .pdf version of the image If you have Adobe Acrobat or some other .pdf reader installed in your system you will get a vector graphic that can be scaled to whatever size you want. They also print out very nicely as hardcopy.

Keopele,

The data in the databrowser comes from the BP Statistical Review 2007. It is not surprising that there are differences between the BP data and the EIA data. No doubt there would be further differences if you compared these data against the IEA data and the OPEC data. I am not really qualified to address the specific reasons why these different organizations came up with different numbers.

Would it be desirable in the future to display the data from all four organizations on top of each other to get a sense of the spread? Folks running weather and climate models do this all the time and call the resulting plots spaghetti plots. Hopefully, our multi-dataset plots would look like uncooked spaghetti still in the package as opposed to the cooked spaghetti the weather models spit out.

Perhaps it might be a mistake to condemn all of the data based on one country and one year?

There is also the chance that BP will revise the 2006 data, when they issue their 2007 annual data (the EIA revised their 2006 data, when they released their 2007 data).

When BP releases their 2007 data, perhaps you could do a detailed comparison of the top net oil exporters for each data set (and even better look at the variance over a five period).

In any case, my preference is the EIA, because of the amount of data that they provide: http://tonto.eia.doe.gov/country/country_time_series.cfm?fips=IR

Having said that, it's obvious that the data sets are not perfect, but the net export trend is pretty clear.

Keopele,

I understand your frustration. I have worked for over a decade with all kinds of data while employed at NOAA -- oceanographic and atmospheric, measured and modeled, station data and satellite data. There are a million reasons why we always include error bars on the absolute measurement of some value. Various groups will argue that their method is better but for the outsider it can be disconcerting.

That's why it's important to present data that is at least self-consistent. I have presented only the BP data in the databrowser and not mixed in data from other sources. I trust (hope really) that BP was as careful as they know how to be in assembling this dataset.

Secondly, it's important not to get too fixated on specific numbers and instead look at data trends. That is what I have tried to do with this data graphic -- focus on the longer term trends to make the interaction of production, consumption and natural decline more apparent. It's basically just westexas' Export Land Model with real data.

I'll try to get some EIA data up in the next month or two so folks can decide for themsleves whether these two datasets are telling the same or different stories regarding export/import trends.