The United States Geological Survey has made periodic assessments of the world’s
conventional oil and gas endowment since the oil shocks of the 1970s. Its last
study was presented at the World Petroleum Congress in Stavanger in 1994 (1)
and the next is due to be presented in Calgary in June of this year. The full
study is awaited but indications of its contents are becoming available through
a press release (2) issued on the eve of a critical OPEC meeting, on the Internet
(3) and in the Oil & Gas Journal (4). These indications suggest that there
are serious flaws in the study. It is a matter of grave concern because the
world is now approaching the peak of conventional oil production. It is important
therefore that the claims of the USGS be subject to close scrutiny, lest they
be given an undeserved credibility in the formulation of government policy.
Demand and price are influenced by the belief in the availability of future
oil and gas.
General Comments
The study gives estimates for the undiscovered amounts of conventional oil,
gas and Natural Gas Liquids (NGL), using a probabilistic approach, but the United
States is treated differently from the rest of the world. For the United States,
maxi ( P95), mini ( P5) and mean cases are quoted, whereas
for the rest of the world, P95, P50, P5 (5) and mean cases are given
by region, being in turn aggregated in a Monte Carlo simulation. Oil and NGL
are combined for the USA, but distinguished elsewhere.
The proposed mean value of undiscovered liquids is 939 Gb (billion barrels)
for the world’s Undiscovered, made up of 649 Gb oil, 207 Gb NGL outside
the USA, and 83 Gb for the USA. A further 730 Gb is proposed for what is termed
Reserve Growth, being made up of 612 Gb oil, 42 Gb NGL, and 76 for the
USA.
It is claimed that the numbers relate to what may be discovered and added to
the reserve base between 1996 and 2025, taking into account economic and
technological factors (6). Such a claim of adding more than 50 Gb/a is however
very difficult to accept in relation to the past discovery trend, which has
fallen from a peak in the 1960s to less 10 Gb a year for the 1990s (7). The
USGS estimate implies a five-fold increase in discovery rate and reserve addition,
for which no evidence is presented. Such an improvement in performance is in
fact utterly implausible, given the great technological achievements of the
industry over the past twenty years, the worldwide search, and the deliberate
effort to find the largest remaining prospects. It conflicts with the industry
viewpoint as expressed by the CEO for Arco who commented (8) "We've embarked
on the beginning of the Last Days of the Age of Oil".
Some of the many points of detail, on which the study displays flawed judgment
and procedure, are discussed below.
Aggregation of data
It is absurd to imagine that every field, basin or country will find only the
minimum estimate. The sum of the minimum values (such as P95) will give a number
smaller than the minimum value of the total. Statistical theory requires that
only mean probability values may be summed. We may add in passing that
most of reports on the world’s Reserves add the Proved Reserves of each
country to get the total, which is incorrect. Proved Reserves are defined
as having a P90 probability under the rules of industry authority (9), different
from the P95 (mini) used by the USGS.
Applying the Monte Carlo procedure gives a Low Case for the eight regions
listed in the study of 339 Gb and a High Case of 1107 Gb, compared respectively
with 179 Gb and 1289 Gb as derived from simple addition. Only the mean
value of 649 Gb is the same under both calculations. If the non-US values are
added using the Monte Carlo procedure, which is indeed the right way to do it,
why is it not applied to the world when adding the USA ? The failure to
use a consistent method means that the assessment of P95 and P5 values for the
world as a whole is fallacious.
The Monte Carlo method is used also for adding associated gas and non-associated
gas, but the total value for P95 for the non-US is given as 2298 Tcf when the
sum is 2312 Tcf . It is another flawed calculation as the total should
clearly be higher than 2312 Tcf.
There are domestic reasons for treating the United States differently from
the rest of the world, including the rules of the Security and Exchange Commission
(SEC) and the reluctance of most US experts to use the probabilistic
approach, which was rejected by the Society of Petroleum Engineers (SPE) until
recently, but it does not facilitate the analysis.
Furthermore, it is misleading to quote the values to an accuracy of more than
two digits, given the wide range of the estimates and the evident high level
of uncertainty.
Distribution of probability
The average ratio of the USGS Undiscovered values are by region:
|
Region |
P95/Mean |
P50/Mean |
P5/Mean |
|
Oil Gb |
0.27 |
0.93 |
1.99 |
|
Gas (associated) |
0.24 |
0.89 |
2.15 |
|
Gas (non-associated) |
0.25 |
0.92 |
2.06 |
|
NGL (associated) |
0.22 |
0.86 |
2.27 |
|
NGL (non-associated) |
0.23 |
0.89 |
2.18 |
|
Average |
0.24 |
0.90 |
2.13 |
Gas is shown as having almost the same probability distribution as oil. Yet,
the North America Region (Canada and Mexico, excluding the USA) has an average
approximately half that of the other regions.for the P95/mean. There is no explanation
for this apparent anomaly.
|
North America Region |
P95/Mean |
P50/Mean |
P5/Mean |
|
Oil Gb |
0.10 |
1.00 |
2.2 |
|
Gas (associated) |
0.13 |
0.93 |
2.3 |
|
Gas (non-associated) |
0.14 |
0.97 |
2.2 |
|
NGL (associated) |
0.12 |
0.89 |
2.4 |
|
NGL (non-associated) |
0.10 |
0.92 |
2.4 |
|
Average |
0.12 |
0.94 |
2.3 |
It is difficult to understand the extreme ranges shown for Mexico and Canada.
It suggests a motive to exaggerate North American quantities for diplomatic
reasons in relation to OPEC, meaning that the forecast is not a reliable basis
for determining general energy policy.
|
|
P95/Mean |
P50/Mean |
P5/Mean |
|
Regions : undiscovered |
0.24 |
0.90 |
2.13 |
|
total non-US : undiscovered |
0.48 |
0.92 |
1.79 |
|
US : undiscovered |
0.79 |
|
1.27 |
|
World : undiscovered |
0.51 |
|
1.72 |
|
non-US : field growth |
0.31 |
1 |
1.68 |
See in Annex 1: the log-normal distribution
It is curious that the distributions of the Undiscovered, which are
similar in seven regions, should be so different from those of North America,
especially since most of the US Undiscovered presumably lies in mature
basins. North America also differs markedly from the aggregated regions and
the world as a whole. It is evident that the study (as the previous studies
by Masters ) has been based more on subjective intuitive geological judgment
(termed the Delphi approach), without taking into account the study and the
extrapolation of past results. It moreover displays a certain national bias
by supposing that the endowment of oil in Nature is in some way different in
North America.
As it is well known, US reserves are reported on financial criteria to meet
the rules of the Securities and Exchange Commission, whereas those in the rest
of the world are based more on a technical assessment, with proper recognition
of the concept of probability. In general, a company has one set of values for
internal technical work, and another for external purposes. Thus, the first
step in such a study should have been to adjust the Proved Reserves
of the United States and Canada so as to match the Proved & Probable
values used in the rest of the world.
Comparison with earlier USGS studies
The previous USGS study by C.D.Masters in 1994, which was given with more data
in 1998 (10), was a much more serious contribution, insofar as the methodology
was explained and the details by basin given. Masters, who demonstrated a good
understanding of the subject, provided in addition the mode value, which
is the peak of the density of probability curve, or in other words, the most
likely case. Comparing the two studies in the following terms shows that the
probability distribution of the two studies is similar. The method is the same.
|
|
USGS 1994/98 |
USGS 2000 |
|
|
P95/Mean |
P5/mean |
P95/Mean |
P5/Mean |
|
Non-US |
0.4 |
2.0 |
0.5 |
1.8 |
|
US |
0.8 |
1.3 |
0.8 |
1.3 |
|
World |
0.5 |
1.7 |
0.5 |
1.7 |
Both used a Delphi method based on an iterative assessment by an un-named panel
of experts, who were asked to express their views of the range and reach a consensus.
It may well explain the extraordinarily high values for the P5 case, lying far
outside any reasonable extrapolation of the past. The high case in turn may
have influenced the mean and median values. It can be said that
studying cumulative discovery over time or against wildcats (method used in
the North Sea by the UKOOA), and applying modern methods of size distribution
analysis, along with other statistical techniques, will give better and more
convincing results than a Delphi committee expressing opinions.
We may compare the two studies as follows:
|
Undiscovered oil in Gb |
P95 |
Mean |
P5 |
|
USGS 1998 |
292 |
583 |
1005 |
|
USGS 2000 Non-US |
334 |
649 |
1107 |
In greater detail, the comparison of the Mean value for oil is:
|
|
USGS 2000 |
USGS 1998 |
|
2 Middle East & North Africa
|
230 |
157 |
|
1 FSU |
116 |
131 |
|
6 C&S America |
105 |
56 |
|
7 Sub-Saharan Africa & Antarctica
|
72 |
35 |
|
5 North America |
70 |
67 |
|
3 Asia Pacific |
30 |
66 |
|
4 Europe |
22 |
27 |
|
8 South Asia |
4 |
5 |
|
US with NGL |
83 |
44+12 |
It is entirely possible that the Middle East estimates may eventually prove
to be correct, but given the time frame, they have to be tempered by the low
level of exploration both in recent years and probably also in the future. The
USGS evidently gives weight to a map of uncertain authenticity that shows a
colossal potential in Iraq that has been circulating in the industry for some
time, and is generally dismissed as being unlikely. The estimates for the FSU,
Asia and Europe have all decreased, possibly reflecting the three years that
have elapsed between the studies.
A comparison of the distribution for North America (Canada and Mexico, excluding
the USA), which is so very different from the other regions, is given below:
|
Oil Gb |
Cumulative |
Identified |
Undiscovered Resources |
|
USGS 1998 |
Production |
Reserves |
P95 |
Mode |
P5 |
Mean |
|
Canada |
16 |
11 |
13.5 |
26.1 |
55 |
31.4 |
|
Mexico |
19.5 |
42 |
14.4 |
24.1 |
72.2 |
35.5 |
|
USGS 2000 |
|
|
|
|
|
|
|
Canada &Mexico |
|
|
6.8 |
70.8 |
157.5 |
70.5 |
To propose a new range from 7 Gb to 158 Gb for two well known, and in the case
of Canada, very mature, countries is utterly implausible. No credence at all
can be given to the P5 estimates. The earlier study of Masters was much more
realistic, albeit itself optimistic. The marked decrease in the P95 may reflect
the second downgrading in Mexico (11) (12), as Mexico reduced its reserves by
20 Gb in 1999 (13).
The mean value is about the same in both studies, changing from 67 Gb to 70
Gb. In annex 2 & 3, an estimate for the main basins of Canada and Mexico
using the extrapolation of the past discoveries gives values two or three times
less than the USGS values.
Discrepancies for cumulative production and remaining reserves
The Cumulative Production for combined oil and NGL in the United States
is reported at 171 Gb as of end 1995, whereas the industry database gives a
value of 205 Gb. Other sources show that the Cumulative Production of
oil and NGL is respectively around 170 Gb and 30 Gb, suggesting that what the
USGS reports as oil and NGL is in fact oil only. The discrepancies for US Cumulative
Production for oil and gas and Remaining Reserves vary from -9% to
+22%, compared with the industry database.
|
|
Cumulative Production end 1995
|
Remaining Reserves, end 1995
|
|
|
Liquids Gb |
Gas Tcf |
Liquids Gb |
Gas Tcf |
|
US |
|
USGS 2000 |
171 |
854 |
32 |
172 |
|
Industry database |
209 |
857 |
30 |
156 |
|
Discrepancy % |
22 |
0 |
-6 |
-9 |
|
World |
|
USGS 2000 |
717 |
1752 |
959 |
4793 |
|
Industry database |
807 |
2200 |
960 |
5083 |
|
Discrepancy % |
13 |
26 |
0 |
6 |
For the world as a whole, the discrepancy varies from 0% to 26%. It is remarkable
that the USGS has failed to comment on these major differences between their
values and those of the industry database for what has been discovered so far,
which clearly has an impact on what remains to discover in the future. One of
the difficulties is to know how the boundary between conventional and unconventional
hydrocarbons has been drawn. The USGS may have definitions very different from
those of the industry.
Field growth
As already pointed out, it is important to remember that the USA and Canada
report Proved reserves to meet SEC rules. They accordingly "grow" over
time to something close to their Mean value.
In this connection, it is worth inquiring into the practice of reporting both
in the USA and elsewhere. In principle, US companies are expected to follow
industry guidelines that define Proved (described as having "reasonable
certainty" in the SEC rules) as having a 90% probability of occurrence, but
in practice it appears that their estimates come closer to the Mode value.
For the last 20 years, the amount of positive revisions reported by the US Department
of Energy were double the negative values, meaning that the original estimates
had a probability of around 66%. That is close to the Mode value, which
in fact represents what the engineers consider the most likely.
The huge "field growth" of the United States is clearly a reporting phenomenon,
as only 6% of the additions over the past twenty years have come from new discoveries.
The rest-of-the-world reports Proved & Probable estimates, because
the industry has a greater need to know what the fields will actually deliver
when they plan costly offshore facilities or pipelines to remote areas.
It is worth discussing this critical issue in greater detail. The SEC rules
were formulated long ago when most US production was onshore, and required that
reserves for financial purposes were restricted to those being drained by a
producing well. Such onshore fields could be placed on production as soon as
the first well was drilled, meaning that the reserves grew as the field was
drilled up over time, in some cases tapping subsidiary reservoirs or new pools
at greater or lesser depth or in the immediate vicinity. In some cases, infill
drilling was also undertaken changing the reserve estimates.
The situation is different for offshore United States fields, because the options
for placing additional facilities to drain subsidiary late stage traps and reservoirs
were clearly less because of higher economic constraints than is the case onshore.
Thus whereas the initial reported reserves onshore understated the eventual
recovery by a factor of 6-7, offshore the factor is no more than about 4.
In the rest of the world, the reserves of a field are normally reported when
it goes into production, the earlier estimates being held confidential by the
company. Host governments take an interest to ensure that good oilfield practices
are used to maximize recovery, and in some cases revenue. The companies accordingly
have to compromise in their reporting between their general desire to under-report
and the demands of the host government. For these reasons, more valid estimates
are reported, although still often as much as half what the field is eventually
found to deliver.
It is evident therefore that most so-called reserve growth is a reporting phenomenon
and not due to technology as is so often claimed. That is not to dismiss technology
as irrelevant - far from it. Technology plays a vital role in increasing the
success ratio in exploration, reducing costs, the lead times between discovery
and the onset of production, and permitting the development of small satellites.
The scope for Reserve Growth, whether genuine or due to reporting practices,
certainly reduces over time as the field approaches exhaustion and eventual
abandonment. It may be that the USGS adds a reserve growth factor to its Undiscovered
estimates on the false belief that technology, even yet to be developed
technology, will inevitably increase the eventual size of the discovery. If
so, it has failed to realise that in practice the reserve estimates made prior
to the drilling of a prospect are almost always much higher than those initially
attributed to the discovery if it is success.
Analysis of the decline of the larger US fields is instructive.
It shows that new technology has had a negligible impact on the reserves of,
for example, the East Texas Field, where the decline increased from 5% to 10%
a year in 1990, and where the estimate of ultimate recovery was reduced from
6 Gb to 5.4 Gb in 1992, see Figure1.
Figure 1:
The Prudhoe Bay Field in Alaska tells the same story. The forecast by the Alaska
Department of Natural Resources up to 2010 shows no improvement in ultimate
recovery, see Figure 2. The Oil & Gas Journal estimates 13 Gb ultimate recovery,
which is close to the 12.7 Gb indicated by decline analysis.
Figure 2:
In short, the reserve growth claimed by the USGS is excessive. Other studies,
as for example that endorsed by the International Energy Agency in its World
Energy Outlook 1998, contemplate a reserve growth of about 200 Gb by 2025.
To claim 612 Gb of field growth in the rest-of-the-world, representing 44% of
the 1359 Gb discovered, is absolutely unrealistic, because the reserves have
been reported more accurately in the first place.
We should also not overlook what has been described as "political" reserves,
including the 300 Gb that certain OPEC countries added in the late 1980s as
they vied for production quota, which was partly based on reserves. In part,
such revisions corrected under-reporting by the industry prior to expropriation,
but in any event the revisions, whatever the correct number should be, have
to be backdated to the discovery of the fields containing them.
As mentioned above, reserve growth is often attributed to the application of
new technology, such as 3D seismic or horizontal drilling, forgetting that these
technologies are no longer new, having been routinely applied to most modern
fields. SheikYamani, for example, claims (14) that the recovery factor in Saudi
fields will increase to 30% to 45%, giving a huge increase in reserves. He evidently
has not realised that the estimated recovery factor in Ghawar, the world’s largest
oilfield, was estimated at 45%, already fifteen years ago, and that the field
is today being produced with horizontal wells, based on high quality 3D seismic
surveys. Its recovery factor is now 60%, with an Ultimate Recovery of 115 Gb.
Furthermore, the present USGS study contradicts Masters in the previous studies
who rightly rejected significant field growth outside the USA in the following
terms:
"we assume that many other countries are, in fact, reporting, effectively,
an Identified Reserve (Proved + Probable + Possible Reserves) or some major
part thereof. In particular, this is considered to be true for all OPEC countries,
the Former Soviet Union (FSU), China, and Mexico. The sum total of these major
producers accounts for more than 90 per cent of world oil reserves; therefore,
we have some confidence that the world value herein reported for Identified
Reserves is a reasonable maximum value for known fields and greatly exceeds
reserves developed for production."
We may also note in passing that another US government agency, the Minerals
Management Service forecasts (15) that the gas production of the Gulf of Mexico
will peak in 2005 and then decline, evidently rejecting the USGS claims of reserve
growth.
In short, the field growth claimed by the USGS in its present study is not
supported by the data.
Ultimate values
The detail of the Ultimate (Mean) values are as follows:
|
Mean |
Oil+NGL |
Gas |
|
Non-US |
Gb |
Tcf |
|
Cumulative Production (end 95)
|
546 |
898 |
|
Remaining Reserves (end 95) |
927 |
4621 |
|
"Field growth" |
654 |
3305 |
|
Undiscovered |
856 |
4669 |
|
Non-US Ultimate |
2983 |
13493 |
|
US |
|
|
|
Cumulative Production (end 95)
|
171 |
854 |
|
Remaining Reserves (end 95) |
32 |
172 |
|
"Field growth" |
76 |
355 |
|
Undiscovered |
83 |
527 |
|
US Ultimate |
362 |
1908 |
|
World |
|
|
|
Cumulative Production (end 95)
|
717 |
1752 |
|
Remaining Reserves (end 95) |
959 |
4793 |
|
"Field growth" |
730 |
3660 |
|
Undiscovered |
939 |
5196 |
|
Ultimate |
3345 |
15401 |
The US Ultimate is quoted as Oil and NGL combined, whereas NGL is separated
in the rest of the world. The proportion of NGL to total liquids in the USA
is approximately 25%, whereas it is only 5% elsewhere. Natural Gas liquids comprise
Condensate, which condenses at the wellhead separator, and such additional liquids
as may be extracted from gas by processing.
There is a huge difference between the two USGS studies in the estimate of
the world Ultimate, which for oil increases from 2.3 Tb to 3.3 Tb (trillion
barrels); and for gas from 12 Pcf to 15 Pcf (16). This increase is out of line
with other estimates made over the past thirty years (figure 3), which show
a gradual leveling off.
Figure 3:
According to this USGS study, only 50% of the oil has been discovered, with
22% and 28% yet to come from so-called Reserve Growth and new discovery.
It contrasts with a statement by, for example, Salameh (2000) of the World Bank,
who declares that almost 90% of the world’s conventional oil has been found.
-Assessment of FSU
The assessment of the former Soviet Union is burdened by the fact that the
Soviets, somewhat like the USGS, considered abstract geological potential, ignoring
the technological, economic and time constraints, as was later confirmed in
1993 by the Russian expert, Khalimov (17). Using these inflated numbers, the
fractal shows 2500 fields of over 1 Mb of reserves, with a total of 264 Gb.
A better approach in this case is to superimpose reported annual discoveries
and annual production with a time shift to give a close fit, which in this case
turned out to be fifteen years (Figure 4).
Figure 4:
Oil production peaked in 1988 a 4.5 Gb/a, with a Cumulative Production
of 100 Gb. Assuming that peak corresponds with the midpoint of depletion, it
means that the likely Ultimate is around 200 Gb, less than the 260 Gb
reported discovered by the Soviets. It suggests that the discoveries should
be reduced to two-thirds of their reported size to obtain a good estimate of
their Proved & Probable value.
The parabolic fractal, based on the inflated numbers, (Figure 5) shows that
Samotlor and Romashkino are Kings and that the larger fields were discovered
before 1980. The yet-in-the-ground Ultimate is 10 000 fields of >1Mb,
holding 330 Gb. The indicated 65 Gb Undiscovered has to be reduced to
two-thirds for the reasons explained, namely to at most 50 Gb. But time and
activity constraints will reduce this value to around 40 Gb.
Figure 5:
Plotting cumulative discovery over time is not useful because of the impact
of the fall of Communism and the resulting slowdown in exploration during the
1990s. So, a plot of cumulative discovery against cumulative wildcats has been
used to give a more valid picture (figure 6).
The curve is now almost flat, although offshore drilling with better equipment
(as in Caspian) may yet open a new chapter.
Figure 6:
Decline analysis confirms the exaggeration of reserves under the Soviet system.
In the case of the Samotlor Field, decline analysis (figure 7) gives an ultimate
value of 20 Gb, even including the planned drilling of 4500 horizontal wells
by Halliburton, compared with a reported value of 24 Gb.
Figure 7:
Masters’ forecast of the Undiscovered for the Former Soviet Union was
of 131 Gb, which has been reduced to 116 Gb in the new USGS study. That is still
more than double the 40-50 Gb indicated by the above analysis. There might also
be case for proposing negative reserve growth to correct the "strongly exaggerated"
present reserves. Field growth is not always positive!
Conclusions
The foregoing discussion demonstrates that the new USGS study has failed to
respect the evidence of past discovery both in terms of amounts and rates. Many
commentators have drawn attention to the remarkable technological achievements
of the industry over the past few years. The resulting performance in terms
of discovery is thus representative of what is achievable in the real world.
It is absurd for the USGS to claim that it is about to be revolutionised to
deliver vastly increased amounts of oil by new discovery and reserve growth
over the next twenty-five years.
The new study was awaited with anticipation in the hope that it would be a
definitive assessment, based on a very through evaluation of every petroleum
system, but the result has been a great disappointment. It is replete with contradictions
and downright errors.
In short, the new USGS report is misleading. It is also unrepresentative of
the normal standards of this highly respected organisation.
One is left to wonder if there is not a hidden agenda, given current US concerns
regarding oil supply, price, and growing OPEC control. The timing of the USGS
press release on the eve of a critical OPEC meeting would certainly be consistent
with such an interpretation.
