Content uploaded by Saud M. Al-Fattah
Author content
All content in this area was uploaded by Saud M. Al-Fattah on Apr 17, 2014
Content may be subject to copyright.
Copyright 1999, Society of Petroleum Engineers Inc.
This paper was prepared for presentation at the 1999 SPE Eastern Regional Meeting held in
Charleston, West Virginia, 20–22 October 1999.
This paper was selected for presentation by an SPE Program Committee following review of
information contained in an abstract submitted by the author(s). Contents of the paper, as
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at
SPE meetings are subject to publication review by Editorial Committees of the Society of
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper
for commercial purposes without the written consent of the Society of Petroleum Engineers is
prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300
words; illustrations may not be copied. The abstract must contain conspicuous
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.
Abstract
Natural gas is an increasingly important source of the world’s
energy. Estimating future supplies of this valuable commodity
is an important economic and strategic endeavor. This paper
analyses historical natural gas production trends for the 53
countries that produce virtually all of the world’s natural gas.
Using a multicyclic Hubbert method, we forecast the world’s
future supply of natural gas to the year 2050.
Our analysis showed that the world ultimate reserves of
conventional natural gas will be around 10,000 Tcf, of which
about 7,900 Tcf of gas reserves remains to be recovered at the
end of 1997. The world production of natural gas is expected
to peak by 2014 at a production rate extending from 2012 until
2017 of approximately 99 Tcf/yr. Based on the 1997 world gas
production and the results of this study, the world supply of
conventional natural gas will continue for 96 years with
reserves depletion rate of 1%/yr.
In his 1956, and later 1980, predictions of U.S. natural gas
production, M. King Hubbert1-4 used one complete production
cycle to forecast production and estimate ultimate recovery of
natural gas for the United States. Several authors have shown
that Hubbert’s model with one production cycle is generally
adequate for predicting crude oil production. However, this
study shows that, in the case of natural gas production, most
countries exhibit two or more Hubbert-type production cycles.
These additional cycles apparently result from changing
exploration technology, regulations, and economic and/or
political events. Using a Hubbert model with a single
production cycle did not allow for these factors. We found that
most of the 53 countries apparently exhibit multicyclic gas
production. To account for additional production cycles we
used a modified version of the Hubbert model which is
referred to as the “multicyclic Hubbert” model. A nonlinear
least-squares regression was used to determine the parameters
of the multicyclic model for each country. Exploration data,
when available, were used to calibrate country models with
production data. We also present a mathematical analysis of
the Hubbert model by deriving equations for determining the
production rates at inflection points and their time of
occurrence on the Hubbert curve. We will demonstrate a
graphical technique to verify the results.
Introduction
Several methods can be used to forecast future production of
fossil hydrocarbon fuels using either available determinations
of ultimate reserves or extrapolation of production history.
Among these models are the logistic and its derivatives (e.g.
Hubbert or Gauss model, normal), hyperbolic or creaming
curves, parabolic fractal, and stretched exponential,
econometric and statistical models.1-7
M.K. Hubbert1-4 made several estimates of the ultimate
recovery of natural gas in the lower 48 U.S. states, ranging
from 810 to 1,100 Tcf. In his 1956 study, Hubbert1 estimated
the U.S. ultimate recovery of natural gas to be 850 Tcf, with a
peak production rate of 14 Tcf/yr occurring in 1970. The
production rate of natural gas peaked in 1973 at a rate of 22.6
Tcf/yr. In the 1962 estimate, Hubbert2 based his analysis on
the ratio of gas discoveries to crude oil discoveries in
conjunction with prior estimates of the ultimate recovery of
crude oil. Two values of gas/oil ratio (6,250 ft3/bbl and 7,500
ft3/bbl) were used along with his estimate of crude oil ultimate
recovery of 175 billion bbl. The lower gas/oil ratio value gave
an ultimate recovery of 958 Tcf while the higher value gave
1,053 Tcf ultimate recovery. The production rate was
estimated to peak in 1977 at 18 to 20 Tcf/yr. With 10 more
years of available production and discovery data since his
1962 work, Hubbert updated his predictions for the U.S.
natural gas ultimate reserves again in 1972 using two methods
of estimation.4 The first method was based on the gas/oil ratio
along with a prior estimate of ultimate reserves of crude oil.
The second method was based on considering gas discoveries
to be a function of cumulative exploratory drilling. The
ultimate recoveries of U.S. natural gas obtained by the first
and second method were 1,000 and 1,100 Tcf, respectively,
with an average value of about 1,050 Tcf. It wasn’t until 1980
when Hubbert4 consolidated all his previous work and
graphical techniques and presented them in a more formal
SPE 57463
Analysis of Worldwide Natural Gas Production
S.M. Al-Fattah, SPE, and R.A. Startzman, SPE, Texas A&M University
2S.M. AL-FATTAH, R.A. STARTZMAN SPE 57463
mathematical structure. In the 1980 prediction of U.S. natural
gas production and ultimate recovery, Hubbert used five
different methods giving estimates of ultimate natural gas
reserves ranging from 810 to 900 Tcf. Hubbert concluded his
study by stating his best estimate of the ultimate recovery of
natural gas in the lower 48 states is 870 Tcf with plus or minus
30 Tcf of uncertainty. Table 1 presents a summary of Hubbert
forecasts of 1956, 1962, 1972, and 1980, for U.S. natural gas.
Masters and others8-9 of the US Geological Survey (USGS)
made a resource assessment of world crude oil and natural gas
at the petroleum basin level. Estimates of individual basins
were aggregated to the country level, then to the regional and
world levels. Quantitative and qualitative geological analysis
with historical data of discovery, exploration, drilling and
production were used for evaluating each petroleum basin or
province. They reported that the world ultimate resource of
natural gas is about 12,000 Tcf (modal value), of which about
5,500 Tcf is identified reserves and around 4,700 Tcf is
conventional undiscovered natural gas. This leaves around
1,700 Tcf for the world’s cumulative production by the end of
1992. The term identified reserves, as used by USGS, includes
proved, probable, and possible reserves (measured, indicated,
and inferred). Their estimates indicated that the world ultimate
resource of natural gas could be as low as 10,000 Tcf (95%
probability) and as high as 18,000 Tcf (5% probability).
Although the study aimed for estimates of world conventional
oil and natural gas resources, gas production from the North
American region includes substantial amounts of
unconventional gas which is mainly from coal-bed methane
(CBM) and tight-gas reservoirs.
Campbell and Laherrere10 also provided the basic data
results of their study for the world gas potential. The ultimate
gas reserves were estimated to be 9,250 Tcf, of which 7,050
Tcf of gas remained to be recovered as of 1996. The study
showed that the world’s natural gas supply would last for 56
years if the combined oil and gas consumption were to be
reduced to the 1996 oil consumption level.
The International Energy Agency (IEA) in their
publication of 1998 World Energy Outlook11 provided
regional forecasts of world gas supply and demand from 1995
to 2020. Projections of gas supply indicated that world gas
production will not reach its peak until after 2020. By 2020,
IEA predicted world annual gas production of 149 Tcf/yr, and
approximately 40% of the USGS9 estimated ultimate
conventional gas reserves will have been produced by that
time. Gas production from the Organization for Economic and
Corporation Development (OECD) countries was projected by
IEA to peak by around 2015 after which production starts
declining. Gas supply and demand of North America,
however, were assumed to be in balance throughout the
forecast period.
Hubbert Model
The Hubbert model8 was based on these three principles:
(1) Production starts at zero and it increases until reaches one
or more principal maxima,
(2) once the peak has been passed, production rate starts
declining until the resource is depleted, and
(3) the area under the curve of production rate versus time is
equal to the ultimate recovery as time approaches infinity.
The basic assumption of Hubbert’s approach is that the
production rate, q or dQ/dt, is a parabola in the cumulative
production, Q, domain. This quadratic relationship can be
expressed as
2
/bQaQdtdQq+== . (1)
Since production rate will be zero when cumulative
production is equal to ultimate recovery, Q∞, then from Eq. 1
we get
0
2=+ ∞∞ bQaQ (2)
which yields
∞
−= Qab /. (3)
By substituting Eq. 3 into Eq. 1 and rearranging, this gives
2dta
Q
Q
Q
dQ =
−∞
. (4)
To express the cumulative production as a function of
time, the left side of Eq. 4 is then integrated from Qo to Q
→Q∞ and the right side is integrated from to to t → t∞. This
will give
( )
1)( o
tta
oeN
Q
Q−−
∞
+
=, (5)
where
o
o
oQQQ
N−
=∞. (6)
Eq. 5 is called the logistic equation, which relates
cumulative production with time. Taking the derivative of
Eq.5 with respect to time gives the production rate as a
function of time, as given in Eq. 7:
[ ]
2
)(
)(
1o
o
tta
o
tta
o
eN
eaNQ
q−−
−−
∞
+
=. (7)
The maximum cumulative production is obtained by
differentiating Eq. 1 and setting it to zero:
2
max ∞
=QQ . (8)
This corresponds to the point of inflection of the S-shaped
cumulative production curve. The peak or maximum
production rate is then found by substituting Eq. 8 into Eq. 1
to give:
4
max ∞
=aQq. (9)
SPE 57463 ANALYSIS OF WORLDWIDE NATURAL GAS PRODUCTION 3
The peak time at which the maximum cumulative
production or the maximum production will occur can be
either read from the graph or computed from the logistic
equation, Eq. 5. Transforming Eq. 5 into a linear form and
substituting for Qmax at tmax from Eq. 8 give:
a)(Ntt oo ln
max += . (10)
The peak production rate and its corresponding time can
also be found using the rate/time equation, Eq. 7. Taking the
first derivative of Eq. 7, setting it to zero and solving for t will
give the time at peak production, tmax, the same as given in
Eq.10. Back substituting for t = tmax, Eq. 10, into Eq. 7 will
result in the maximum production rate as given in Eq. 9.
Alternative Form of Hubbert Equations. Eqs. 5 and 7 of the
Hubbert model each contain four parameters that are usually
determined by fitting the curve with one complete production
cycle. Having historic production data that exhibits more than
one production cycle requires more than four parameters to fit
and hence make the task difficult. To reduce the number of
parameters and its subsequent complexity, we can express the
cumulative production and rate of production in terms of
maximum production and maximum time. By solving for to
from Eq. 10 and substituting into Eq. 5, we get
(
)
[
]
1max
tta
eQQ −−
∞+= . (11)
The production rate is then obtained by taking the first
derivative of the equation above and substituting for Q∞ =
4qmax/a.
[
]
14)( 2
)()(
max maxmax
+= −−−− ttattaeeqtq. (12)
Expanding the quadratic term in the denominator and
simplifying, we get the same equation as presented by
Laherrere12:
(
)
[
]
{
}
maxmax cosh12)( ttaqtq−+= . (13)
Now, the logistic equation (Eq. 11) has three parameters as
do Eq. 12 and Eq. 13, making historic production data with
multiple production cycles easier to model.
Inflection Points of Hubbert Curve. We present here useful
equations for determining the inflection points on the Hubbert
rate/time curve for one complete production cycle. Starting
with the rate/time equation, Eq. 7, and taking its second
derivative, we get
(
)
( )
4
)(
)(2)(
2
)(
34
)(
o
ooo
tta
o
ttatta
oo
tta
o
eN
eeNNeNaQ
tq−
−−−
∞+
+−
=
′′ . (14)
Setting the second derivative to zero and solving for time,
t, we obtain the inflection points where changes of sign occur
on the curve with one complete cycle. The solution gives two
points of inflection called tifl1 and tifl2 which are given,
respectively, as
(
)
(
)
[
]
32ln1
1ooifl Natt −+= (15)
and
(
)
(
)
[
]
32ln1
2ooifl Natt ++= . (16)
The corresponding production rate of these inflection
points can then be obtained by substituting either tifl1 or tifl2
into Eq. 7. Thus, the production rate at the inflection points is
6
∞
=aQqifl . (17)
From qmax, Eq. 9, and qifl, Eq. 17, we can then find a new
relationship between the maximum production rate and
production at inflection points, given by
max
q
3
2
=
ifl
q. (18)
Another relationship is found using Eqs. 10 and 15 or Eqs.
10 and 16, relating the peak time and the time at the inflection
point for one complete production cycle on the bell-shape
curve. Considering the inflection point at the left side of the
curve, we get
317.1
1max a
tt ifl += . (19)
These relations state that the maximum production rate is
1.5 times the production rate at the inflection points at a
distance of (1.317/a) from the time of inflection points on
either side of the symmetrical bell-shape curve. Therefore,
knowing the rate of inflection on the curve, we can calculate
the peak or maximum production rate. Rather than solving
analytically, the production rate at the inflection points as
given in Eq. 17 can also be estimated from production data
and using a spreadsheet. The procedure is to calculate the rate
of change of production rate over evenly spaced time
intervals, i.e. ∆q/∆t. It is helpful to prepare a plot of these
computed values versus time as shown in Fig. 1. Smoothing
techniques such as exponential or moving-average methods
might be necessary to smooth the data.
Next, observe the signs of the computed values as they
change or the change of the curve up and down. The value of
∆q/∆t at which is zero or the sign alters (positive to negative
or vice versa) from that of the previous one is called the
inflection point on the rate derivative curve and the maximum
point on the rate curve. The maximum value (positive) of rate
derivative or ∆q/∆t corresponds to the inflection point on the
left side of Hubbert curve. Similarly, the minimum value
(negative) of ∆q/∆t corresponds to the inflection point on the
right side of the Hubbert curve. The corresponding time, tifl,
and production rate, qifl, can then be read from the production
data or from the graph. Using the value of qifl in Eq. 18 gives
the maximum production rate, qmax. The peak time is given by
the relation in Eq. 19, the time at the inflection point plus a
constant. The parameter a is usually determined by fitting the
production data using nonlinear regression analysis. Another
4S.M. AL-FATTAH, R.A. STARTZMAN SPE 57463
alternative is to compute the parameter a from either Eq. 9 or
Eq. 17 with a prior knowledge of the value Q∞. For this
purpose, estimates of the ultimate recovery, Q∞, from
independent studies can be used, or it can be roughly
approximated by adding published proved reserves to actual
cumulative production. Therefore, the peak time can be
expressed as
ifl
ifl qQ
tt ∞
+= 219.0
1max (20)
or,
329.0
max
1max qQ
tt ifl ∞
+= . (21)
If we choose to use the inflection point on the right side of
the curve in the calculations, then the positive sign of the
second terms in Eqs. 19, 20, and 21 must be changed to a
negative and tifl1 be changed to tifl2.
The advantage of this method is that we can forecast the
maximum production rate before it takes place. One limiting
condition of this procedure is that the production data must be
sufficient to reach the inflection point.
Example. To verify the equations above we used Thailand
production data as an example. Fig. 1 shows the production
rate fitted with one complete cycle of the Hubbert model. The
obtained results are: Q∞ =33.8 Tcf, a = 0.154 year-1, and No =
8.08 at to = 1997. Applying Eqs. 15 through 21, the inflection
points occur in 2002 and in 2019 at a production rate of 2.37
Bcf/D. The maximum rate of 3.56 Bcf/D can then be
determined using Eq. 18 and its corresponding maximum time
at around 2011 using Eq. 19, 20 or 21.
These results are consistent with the behavior of the rate
derivative as shown in Fig. 1, which shows identical results in
the time domain. The rate derivative curve has one maximum
at 2002 and one minimum at 2019, representing the two
inflection points on the Hubbert curve. The time at which the
rate derivative is zero (at about 2011) reflects the maximum
time on the Hubbert curve.
Multicyclic Model. Several authors4,13,14 have shown that
Hubbert’s model with one production cycle is generally
adequate for predicting crude oil production. However, this
study shows that, in the case of natural gas production, most
countries exhibit two or more Hubbert-type production cycles.
These additional cycles apparently results from changing
exploration technology, regulations, and economic and/or
political events. Using a Hubbert model with a single
production cycle did not allow for these factors. To account
for additional production cycles we used a modified version of
the Hubbert model which is referred to as the “multicyclic
Hubbert” model. Based on the number of cycles suggested by
the production data, we can sum up an equal number of
Hubbert-type production cycles using either Eq. 12 or Eq. 13.
Accordingly, the multicyclic model can be expressed, using
Eq. 12, as
( )
[ ]
∑∑ =
−−−−
=
+== k
ii
ttatta
i
k
iieeqtqtq
1
2
)()(
max
1
maxmax 14)()( (22)
or, using Eq. 13, as
( ) ( )
[ ]
{ }
∑
=−+= k
iii ttaqtq
1maxmax cosh12)( , (23)
where k is the total number of production cycles. The
parameters of the multicyclic model can be determined using a
nonlinear least-squares regression. Every complete production
cycle has its own value of ultimate recovery, Q∞, computed by
Eq. 9. The total ultimate recovery is then determined by
adding the ultimate recoveries for each production cycle.
The logistic curve or the cumulative production of the
multicyclic model can then be expressed as
( )
[ ]
∑
=
−−
+= k
ii
tta
ieaqQ
1
)(
max max
14 . (24)
Sources of Data
In this study, we use historical natural gas production data
acquired from Oil and Gas Journal (OGJ) database,15,16
Twentieth Century Petroleum Statistics,17,18 and the Energy
Information Admin. (EIA).19 We used discovery data of the
U.S. (1900-1997) from Refs. 19 and 20, and obtained
marketed gas production starting from 1918-1997 from the
Twentieth Century Petroleum Statistics. Annual production of
natural gas for all other countries was obtained from OGJ
database for the period 1971-1997. Annual gas production
data from EIA (1980-1997) were also used to validate
suspicious data and replace it if necessary. Proved reserves of
natural gas for all countries (1967-1998) were obtained from
OGJ.
Procedure
First, we examined historical production data for each country
and decided the number of production cycles based on initial
data examination. At a later stage of the modeling process, we
sometimes needed additional cycles to have a better fit.
Second, we set the problem up to solve the model, Eq. 22 or
Eq. 23, using a nonlinear least-square solver with initial
guesses for the parameters a, qmax, and tmax based on the
number of production cycles. For example, production data
exhibiting two production cycles require six parameters to
solve for the model. The optimal values of the parameters are
obtained by minimizing the root mean square, RMS, residual of
production rates. The root mean square is a measure of data
dispersion around zero deviation and it is defined as
SPE 57463 ANALYSIS OF WORLDWIDE NATURAL GAS PRODUCTION 5
( )
nqqRn
icalobsMS ∑
=−= 1
2 , (25)
where RMS is the root mean square, qobs the observed
production rate, qcal the calculated production rate from the
model, and n the number of observations.
Having determined the optimal values of the model
parameters for each production cycle, we computed the
ultimate recovery for each production cycle using Eq. 9 and
then summed them to get the total ultimate recovery. We
computed the cumulative production by adding annual gas
production from previous years. Since production data for
most countries start from 1971 and we wanted to account for
gas production from these countries before this date, we
calibrated the cumulative production to match the 1992 year-
end cumulative production data published by the USGS study8
in 1994. The USGS used the Petroconsultant database that has
historical discovery and production data based on fields level.
The difference of cumulative production was lumped as pre-
first year of available data (e.g. pre-1971). The cumulative
production from the multicyclic model is obtained from Eq.
24, or can be calculated using linear approximation as Σq ∆t.
The future recoverable reserves (FRR) of gas is
determined by subtracting the calibrated cumulative
production (as of 1997 year-end) from the estimated ultimate
recovery (EUR) obtained from the model. The future
recoverable reserves include proved, probable, and
undiscovered gas.
Goodness of Fit
We used the dimensionless root mean square (Drms) criterion13
to determine the goodness of fit for each country’s model. It is
obtained by dividing the root mean square, RMS, by the highest
peak production rate of the multicyclic model for each
country. The values of RMS and Drms for all countries are given
in Table 2. Fig. 2 shows a semi-log cumulative probability
plot of Drms having a mean of 5.5% and a standard deviation of
4.2%.
We categorized the goodness of fit of all models arbitrarily
as Good, Fair, and Poor. A country model has a Good fit if its
dimensionless root mean square is less than or equal to the
mean (i.e. Drms ≤ 5.5%), indicating that the model fits the data
very well. Thirty out of 53 models of all countries fell into this
category. A country model with a Fair fit falls in the range
5.5% < Drms ≤ 18.1%; that is, between the mean and the mean
plus three standard deviations. Twenty-two models which
showed slight deviations from the production data fell into this
category. A country model with a Poor fit has a Drms value ≥
18.1%, indicating great deviations from the data. Only one
country model fell into this category, the Libya model.
Analysis of Results
Since the discovery data for all countries, except for the U.S.,
were not found publicly available, our results are based on
modeling the production data of each country. However,
historical data of gas discovery (1900-1997), proved reserves
(1967-1998) and annual production (1918-1997) were used to
analyze and model U.S. natural gas. The following sections
present the analysis of U.S. natural gas, and for the world.
U.S.
Hubbert investigated the relationship that the cumulative
discovery is the sum of cumulative production and proved
reserves. It was shown by Hubbert that when proved reserves
reach their peak or maximum value, then the rate of increase
of proved reserves becomes zero, and the discovery rate will
be equal to the production rate. Fig. 3 shows the discovery
rate, production rate, and increment of proved reserves. This
figure suggests that proved reserves already passed its peak
and it is in a declining stage. In their review of U.S. gas
production, Wattenbarger and Villegas21 showed that the
Hubbert model with one complete production cycle is not
useful for U.S. gas production.
In this study, examinations of annual discovery and
production data suggest that the U.S. gas production exhibits
at least two production cycles, indicating a good candidate for
the multicyclic model approach. Using an over-lay matching
procedure, we correlated the annual discovery with the annual
production after a time-lag shift. The purposes of this
procedure are to estimate the maximum rate of discovery
corresponding to the maximum production rate, to determine
the time lag between the rates of discovery and production,
and to forecast the future production peak, if any. Fig. 4 shows
the rate of discovery superimposed on the production rate after
matching their corresponding maxima. Two principal maxima
were identified for the rate of discovery. The time of the first
principal peak of discovery rate was at around 1954 and the
corresponding time of peak production rate was at 1973, a
time lag of 19 years. The second principal maximum
discovery rate occurred in 1974. Therefore, the next peak of
production rate should have occurred approximately in 1993.
Unconventional gas reserves, which is not accounted for in the
discovery, fools and disturbs the correlation. However, we
determined that the optimum value for the second principal
peak of production is at 1998. Furthermore, two additional
production cycles based on discovery data were added at 1930
and at 2010, which fine-tuned the fit. Fig. 5 shows the
multicyclic model for the U.S. gas production trend. The
production model also fits the cumulative production data very
well, as shown in Fig. 6.
This approach gives ultimate conventional gas reserves of
1,200 Tcf and future recoverable gas (remained to be
produced) of 312 Tcf. The gas production had already passed
its peak of 22.5 Tcf/yr in 1973. Based on 1997 gas production
and current technology, the 312 Tcf of future reserves will be
produced for approximately 16 years and depleted at a rate of
6.4%/yr. This makes the U.S. the fastest country in depleting
its gas reserves. This analysis includes neither the reserves
growth in existing fields nor the unconventional gas (e.g. coal-
bed methane, tight gas, and hydrates) which we believe these
will play important roles in the additions of U.S. gas reserves
in the near future. In the 1995 assessment of U.S. gas
6S.M. AL-FATTAH, R.A. STARTZMAN SPE 57463
resources22 (excluding Federal offshore), USGS predicted 322
Tcf of reserves growth would be added to proved gas reserves
by the year 2071, and estimated 358 Tcf for unconventional
gas. The amount of proved and undiscovered conventional gas
was estimated to be 394 Tcf.
The World
Natural gas production model for each gas producing country
was constructed individually. For brevity, we present
examples of these models for most countries as shown in Figs.
1 and 5, and in Figs. 7 through 33. Countries with low gas
production in a specific region were lumped under the model
Others were assigned to their respective regions. Table 2
shows the classification of world gas countries and presents a
summary of the results.
About 80% (42 models) of the production trends of all gas
producing countries were modeled using the multicyclic
model. This approach is relatively simple and its data are
easily available. We found that only 10 production trends of
countries, most of which are from Asia-Pacific region, follow
the Hubbert model with one complete production cycle. These
models were constructed for countries of: Venezuela,
Hungary, Qatar, Syria, Egypt, Austria, India, Pakistan,
Thailand, and Other Asia-Pacific.
Individual-country models were aggregated to the region
level and then to the world level. Fig. 34 shows the world
aggregated gas model which fits the data very well. This gives
the ultimate conventional gas reserves for the world to be
around 10,000 Tcf, of which about 7,900 Tcf remains to be
recovered as of 1997 year-end. By year 2050, around 7,200
Tcf of gas will be produced, representing 71% of ultimate
reserves and 91% of future recovery of gas. Our results show
that the world conventional gas production will peak at an
approximate rate of 99 Tcf by the year 2014. The gas
production model of the world shows a flat production region,
called the middle region, extending from the year 2012 until
the year 2017 with about the same rate as the peak. After the
year 2017, the production rate starts declining steadily and the
curve gets flatter.
The depletion rate or P/R ratio for each gas-producing
country is calculated as the annual production divided by the
future reserves expressed in percentage, Table 2. Based on
1997 world gas production and current recovery techniques,
the world gas reserves is being depleted at 1%/yr. The United
States and Denmark have the highest depletion rates of
reserves with 6.4% and 6.3%/yr, respectively. Most countries
of the Middle East region have a depletion rate less than
0.5%/yr, the lowest worldwide. The distribution of the
ultimate and future recoverable gas shows that the top 10
countries (Former Soviet Union, U.S., Iran, Saudi Arabia,
Canada, Qatar, U.A.E., Venezuela, Mexico, and Algeria)
contribute about 78% of the total world ultimate conventional
gas, Fig. 35.
Fig. 36 is a log-log plot with a unit slope line comparing
proved gas reserves16 and future recoverable reserves obtained
from this study. Our results give higher values of reserves than
published proved reserves. This is expected since the latter do
not include probable and undiscovered reserves, and since the
gas industries in some countries have not yet reached maturity.
Figs. 37 and 38 illustrate the distributions, in semi-log
cumulative probability plots, of the future recoverable gas and
ultimate recovery, respectively. Both plots appear to be log
normally distributed with means of 148.9 Tcf for the future
recoverable reserves and 189.5 Tcf for the estimated ultimate
recovery.
Conclusions
In this paper, alternative forms of the Hubbert model were
derived with fewer parameters, making production data easier
to model. New relationships and procedures were presented
for determining the maximum production rate and production
at the inflection points on the symmetrical bell curve. This
study found that more than 80% (43 out of 53) of the countries
considered exhibit two or more Hubbert-type production
cycles. Gas production trends of these countries were modeled
with fairly good results using our modified version of the
conventional Hubbert model, the multicyclic modeling
approach. With this approach, we recommend the use of
discovery data, when available, to be correlated with
production data to aid in forecasting the future production and
its peak.
Our world gas model was developed by combining
individual models of all gas producing countries. The results
of this study estimated that the world ultimate reserves of
conventional natural gas is around 10,000 Tcf, and
approximately 7,900 Tcf remained to be recovered as of 1997
year-end. Our analysis also shows that the world production of
natural gas will have a flat production region of 99 Tcf/yr
from 2012 to 2017, giving the peak in 2014 at a rate of 99
Tcf/yr. The world gas reserves is being depleted at 1%/yr. The
U.S. and Denmark have the highest depletion rates at about
6.4%/yr, while most Middle East countries have the lowest
rates of reserves depletion (0.2%/yr).
Nomenclature
a = constant, 1/t, 1/yr
b = constant, 1/t-L3, 1/(yr-Tcf)
dQ/dt = production rate, L3/t, Tcf/yr
EUR = estimated ultimate recovery, L3, Tcf
FRR = future recoverable reserves, L3, Tcf
n = number of observations, n
No = dimensionless cumulative factor
q(t) = production rate as a function of time, L3/t, Tcf/yr
qifl = production rate at inflection point, L3/t , Tcf/yr
qmax = maximum or peak production rate, L3/t , Tcf/yr
Q = cumulative production, L3, Tcf
Qmax = maximum cumulative production, L3, Tcf
Qo = cumulative at an arbitrary time to, L3, Tcf
Q∞ = ultimate recovery of gas, L3, Tcf
RMS = root mean square, L3/t, Bcf/D
t = time, t, calendar year
tifl = time corresponds to qifl, t, calendar year
to = arbitrary time, t, calendar year
SPE 57463 ANALYSIS OF WORLDWIDE NATURAL GAS PRODUCTION 7
tmax = time at peak production, t, calendar year
Acknowledgements
S.M. Al-Fattah would like to thank Saudi Aramco for
supporting his PhD study at Texas A&M University. We
deeply thank J.H. Laherrere, consultant, for his valuable
comments and suggestions.
References
1. Hubbert, M.K.: “Nuclear Energy and Fossil Fuels,” Drill. & Prod.
Prac. (1956) 17.
2. Hubbert, M.K.: “Energy Resources,” Publication 1000-D, Natl.
Academy of Science/Natl. Research Council (1962).
3. Hubbert, M.K.: “Degree of Petroleum Exploration in the United
States,” AAPG Bull. (11 November 1967) 51, 2207.
4. Hubbert, M.K.: “Techniques of Prediction as Applied to
Production of Oil and Gas,” Proc., U.S. Dept. of Commerce
Symposium, Washington, DC (June 1980) 16.
5. Hotelling, H.: “The Economics of Exhaustible Resources,” The
Journal of Political Economy (April 1931) 137.
6. Campbell, C.J.: The Coming Oil Crisis, Multi-Science Publishing
Co. and Petroconsultants S.A., Brentwood, England (1997) 86.
7. Laherrere, J.H. and Sornette, D.: “Stretched Exponential
Distributions in Nature and Economy,” European Physical
Journal (January 1998) 525.
8. Masters, C.D., Attanasi, E.D., and Root, D.H.: “World Petroleum
Assessment and Analysis,” Proc., 14th World Pet. Cong.,
Stavanger, Norway (1994) 529.
9. Masters, C.D., Root, D.H., and Turner, R.M.: “World
Conventional Crude Oil and Natural Gas: Identified Reserves,
Undiscovered Resources and Futures,” U.S. Geol. Survey, Open-
File Report 98-468 (August 1998), Internet Home Page:
http://energy.er.usgs.gov/products/.
10. “The Global Hubbert Peak: Natural Gas,” Internet Home Page:
http://www.hubbertpeak.com/gas/index.html, 1997.
11. International Energy Agency: World Energy Outlook, 1998 ed.,
IEA/OECD, Paris, 1998.
12. Laherrere, J.H.: “World Oil Supply – What Goes Up Must Come
Down, But When Will It Peak?” Oil and Gas J. (February 1999)
97:5.
13. Al-Jarri, A.S. and Startzman, R.A.: “Analysis of World Crude Oil
Production Trends,” paper SPE 37962 presented at the 1997 SPE
Hydrocarbon Economics & Evaluation Symposium, Dallas, 16-18
March.
14. Al-Jarri, A.S. and Startzman, R.A.: “Worldwide Petroleum-Liquid
Supply and Demand,” JPT (December 1997) 1329.
15. Energy Statistics Sourcebook, 13th ed., OGJ Energy Database,
PennWell Pub. Co., Tulsa, OK (1998).
16. International Energy Statistics Sourcebook, 8th ed., OGJ Energy
Database, PennWell Publishing Co., Tulsa, OK (1998).
17. Twentieth Century Petroleum Statistics, 52nd ed., DeGolyer and
MacNaughton, Dallas, TX (1996).
18. Twentieth Century Petroleum Statistics, 54th ed., DeGolyer and
MacNaughton, Dallas, TX (1998).
19. EIA, Internet Home Page: http://www.eia.doe.gov/.
20. Attanasi, E.D. and Root, D.H.: “The Enigma of Oil and Gas Field
Growth,” AAPG Bulletin (March 1994) 78, 321.
21. Wattenbarger, R.A. and Villegas, M.E.: “Trends in U.S. Natural
Gas Production,” Advances in the Economics of Energy and
Resources, J.R. Moroney (ed.), J.A.I. Press, Greenwich,
Connecticut (1995) 9, 169.
22. Gautier, D.L. et al.: “1995 National Assessment of United States
Oil and Gas Resources-Results, Methodology, and Supporting
Data,” USGS, Digital Data Series DDS-30, Release 2 (1996).
SI Metric Conversion Factors
bbl x 1.589 873 E-01 = m3
ft3x 2.831 685 E-02 = m3
TABLE 1-HUBBERT FORECASTS OF NATURAL GAS FOR
LOWER 48 U.S. STATES
Year of estimate Ultimate
(Tcf) Peak prod.
(Tcf/yr) Peak time
(year)
1956 850 14 1970
1962 1000 19 1977
1972 1050 24 1975
1980 870 - -
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate (q), BCF/D
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
dq/dt, BCF/D
Inflection points
Maximum rate
1. Rate displacement = qmax/3
2. Time displ. = 1.317/a
3. Slope or depletion = aqmax/3.951
1
2
3
Fig. 1-Thailand production data, example for determining the
maximum and inflection points on the Hubbert curve.
Cumulative Probability, %
1 2 5 10 20 30 50 70 80 90 95 98 99
Drms, %
0.1
1
10
100
Fig. 2–Cumulative probability plot of dimensionless root mean
square.
8S.M. AL-FATTAH, R.A. STARTZMAN SPE 57463
TABLE 2- ANALYSIS SUMMARY OF WORLD CONVENTIONAL NATURAL GAS
Prod rate R/P Depletion R
MS
Drms
Country 1997 1997 2050 Time Prod
Proved
1/98 FRR EUR (P/R)
(Bcf/yr)
(Tcf)
(Tcf)
(year)
(Bcf/yr)
(Tcf)
(Tcf)
(Tcf)
(year)
(%/yr)
(Bcf/D)
(%)
Canada 6577 114.4 442.9 2011 10500 65.0 337.0 451.4 51.2 2.0 0.741 2.58
Mexico 1187 26.9 146.7 2036 2550 63.9 189.5 216.4 159.7 0.6 0.149 2.13
United States 19900 891.1 1202.3 1973 22540 166.5 312.0 1203.1 15.7 6.4 1.891 3.06
North America 27663 1032.5 1792.0 295.4 838.5 1871.0 30.3 3.3
Argentina 995 15.0 63.3 2010 1300 24.3 50.9 65.9 51.2 2.0 0.321 9.02
Bolivia 72 3.1 7.7 2016 155 4.6 4.7 7.7 64.6 1.5 0.038 9.04
Brazil 128 2.9 16.1 2031 295 5.6 18.2 21.2 141.9 0.7 0.040 4.96
Chile 45 3.8 9.7 1978 322 3.5 5.9 9.7 129.2 0.8 0.063 7.11
Columbia 171 3.8 15.4 2018 266 14.2 13.3 17.0 77.4 1.3 0.025 3.39
Ecuador 3.6 0.6 3.2 2055 125 3.7 6.6 7.2 >250 0.1 0.026 7.69
Peru 34 1.1 5.2 2070 175 7.0 15.2 16.3 >250 0.2 0.016 3.32
Trinidad & Tobago 328 4.6 22.2 2012 600 15.9 17.9 22.5 54.6 1.8 0.031 1.91
Venezuela 1003 20.1 142.6 2044 3015 143.1 231.3 251.4 230.7 0.4 0.171 2.06
S&C America 2780 54.9 285.5 221.9 363.9 418.8 130.9 0.8
Austria 50 1.6 2.5 1975 80 0.9 0.9 2.5 18.0 5.5 0.020 9.27
Denmark 267 1.7 6.0 2002 475 4.0 4.2 6.0 15.9 6.3 0.019 1.46
France 91 5.6 7.1 1983 371 0.5 1.5 7.1 17.0 5.9 0.059 5.80
Germany 738 17.7 39.3 2000 752 12.1 22.0 39.7 29.8 3.4 0.249 12.07
Italy 718 19.7 51.3 2006 858 10.5 32.9 52.6 45.8 2.2 0.191 8.12
Netherlands 3257 78.5 138.2 2000 3834 61.3 59.7 138.2 18.3 5.5 1.060 10.09
Norway 1601 20.0 170.7 2020 3500 52.3 174.2 194.2 108.8 0.9 0.523 5.45
United Kingdom 3250 46.0 111.9 2002 4021 26.8 65.9 111.9 20.3 4.9 0.485 4.40
Other 112 2.2 7.5 1985 150 1.9 5.8 8.0 52.0 1.9 0.037 9.02
W. Europe 10083 193.0 534.6 170.4 367.2 560.2 36.4 2.7
Albania 0.7 0.2 0.3 1988 25 0.1 0.1 0.3 139.9 0.7 0.004 5.95
FSU 23844 575.3 2365.3 2032 36000 1977.0 2753.0 3328.3 115.5 0.9 1.415 1.43
Hungary 154 3.7 9.3 1984 233 3.2 5.6 9.3 36.4 2.7 0.044 6.92
Romania 754 36.5 48.9 1984 1364 14.0 12.8 49.3 17.0 5.9 0.324 8.67
Other 248 19.4 23.9 1979 1246 8.9 4.5 23.9 18.2 5.5 0.429 12.57
E. Europe & FSU 25001 635.0 2447.6 2003.2 2776.0 3411.0 111.0 0.9
Bahrain 201 3.6 9.0 2004 266 5.1 5.3 9.0 26.5 3.8 0.061 8.40
Iran 1199 26.1 268.9 2076 13500 810.0 1130.2 1156.4 >250 0.1 0.914 2.47
Iraq 128 2.5 34.2 2073 2000 109.8 146.7 149.1 >250 0.1 0.090 1.65
Kuwait 210 5.5 35.7 2050 860 52.9 64.6 70.1 >250 0.3 0.165 7.00
Oman 165 1.9 29.3 2023 845 27.5 28.8 30.7 174.6 0.6 0.022 0.96
Qatar 611 4.6 259.9 2035 8534 300.0 301.9 306.5 >250 0.2 0.148 0.63
Saudi Arabia 1303 18.2 274.5 2045 7500 190.0 453.5 471.7 >250 0.3 0.488 2.37
Syria 145 1.3 9.3 2011 222 8.3 8.3 9.6 57.4 1.7 0.104 17.17
UAE 851 11.6 195.0 2035 5000 204.9 241.3 252.8 >250 0.4 0.188 1.37
Other 0.9 0.1 0.5 2125 850 17.1 52.3 52.4 >250 0.0 0.004 0.18
Middle East 4814 75.3 1116.3 1725.6 2433.0 2508.3 >250 0.2
Algeria 2473 31.3 195.5 2014 5300 130.6 167.7 199.1 67.8 1.5 0.405 2.79
Angola 20 0.8 3.5 2025 75 1.7 3.1 3.8 151.7 0.7 0.017 8.40
Egypt 444 4.7 35.5 2014 1027 27.6 31.5 36.2 70.9 1.4 0.084 2.97
Libya 212 6.2 30.1 2050 650 46.3 51.6 57.8 242.9 0.4 0.326 18.28
Nigeria 150 4.9 31.2 2085 1500 114.9 145.5 150.5 >250 0.1 0.085 2.06
Tunisia 23 0.7 3.0 2008 185 2.5 2.3 3.0 98.8 1.0 0.026 5.07
Other 10 0.7 4.6 2080 350 25.0 24.8 25.5 >250 0.0 0.110 11.43
Africa 3334 49.3 303.5 348.6 426.5 475.8 127.9 0.8
Australia 1061 13.5 91.1 2016 2200 19.4 81.3 94.8 76.6 1.3 0.089 1.48
Brunei 365 6.5 20.4 2007 592 14.1 14.0 20.5 38.3 2.6 0.124 7.66
China 741 18.4 127.2 2041 3000 41.0 179.2 197.6 241.9 0.4 0.126 1.53
India 800 7.0 37.9 2007 1450 17.4 31.0 38.0 38.8 2.6 0.128 3.22
Indonesia 2335 30.2 149.3 2011 4275 72.3 120.7 150.9 51.7 1.9 0.316 2.70
Japan 79 2.2 4.5 2000 81 1.4 2.3 4.5 29.1 3.4 0.026 11.62
Malaysia 1351 11.0 102.0 2012 3600 79.8 91.5 102.5 67.7 1.5 0.194 1.97
Pakistan 651 10.6 53.9 2018 1020 21.0 49.3 60.0 75.8 1.3 0.201 7.18
Thailand 502 3.7 33.7 2011 1300 7.0 30.0 33.8 59.9 1.7 0.063 1.78
Other 557 8.7 75.3 2030 1550 47.2 86.7 95.4 155.8 0.6 0.141 3.32
Asia-Pacific 8440 111.9 695.4 320.6 686.0 797.9 81.3 1.2
Total World 82115 2152 7175 2014 98840 5086 7891 10043 96 1.0
Cumulative prod Peak Reserves
SPE 57463 ANALYSIS OF WORLDWIDE NATURAL GAS PRODUCTION 9
-30
-20
-10
0
10
20
30
40
50
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000
Time, year
TCF/yr
Discovery rate
Reserves increase
Production rate
Fig. 3–Relationships of rate of discovery, rate of production, and
rate of increase of proved reserves for U.S. natural gas.
0
10
20
30
40
50
60
70
80
90
100
1910 1930 1950 1970 1990 2010 2030 2050
Time, year
Production Rate, BCF/D
0.0
3.7
7.3
11.0
14.6
18.3
21.9
25.6
29.2
32.9
36.5
Discovery Rate, TCF/year
Prod. Rate
Hubbert cycles
Discovery rate
Fig. 4–U.S. natural gas production correlated with backdated
discovery shifted 19 years, and multicyclic modeling process.
0
10
20
30
40
50
60
70
80
90
100
1910 1930 1950 1970 1990 2010 2030 2050
Time, year
Production Rate, BCF/D
0
200
400
600
800
1000
1200
1400
1600
1800
2000
Cumulative Gas, TCF
Prod. Rate
Rate Model
Cumulative
Cum. Model
Fig. 5–U.S. natural gas production multicyclic model.
0
10
20
30
40
50
60
70
80
0.0E+00 2.0E+05 4.0E+05 6.0E+05 8.0E+05 1.0E+06 1.2E+06
Cumulative Gas Production, BCF
Production Rate, BCF/D
Fig. 6–U.S. natural gas production rate vs. cumulative production.
0
5
10
15
20
25
30
35
40
45
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
50000
100000
150000
200000
250000
300000
350000
400000
450000
Cumulative Gas, BCF
Prod. Rate
Rate Model
Cumulative
Cum. Model
Fig. 7–Canada natural gas production model.
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
Cumulative Gas, BCF
Prod. Rate
Rate Model
Cumulative
Cum. Model
Fig. 8–Brazil natural gas production model.
10 S.M. AL-FATTAH, R.A. STARTZMAN SPE 57463
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
Cumulative Gas, BCF
Prod. Rate
Rate Model
Cumulative
Cum. Model
Fig. 9–Columbia natural gas production model.
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
2.4
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
22000
24000
Cumulative Gas, BCF
Prod. Rate
Rate Model
Cumulative
Cum. Model
Fig. 10–Trinidad and Tobago natural gas production model.
0
1
2
3
4
5
6
7
8
9
10
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
25000
50000
75000
100000
125000
150000
Cumulative Gas, BCF
Prod. Rate
Rate Model
Cumulative
Model Cum.
Fig. 11–Venezuela natural gas production model.
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
250
500
750
1000
1250
1500
1750
2000
2250
2500
Cumulative Gas, BCF
Prod. Rate
Rate Model
Cumulative
Cum. Model
Fig. 12–Austria natural gas production model.
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
1000
2000
3000
4000
5000
6000
7000
8000
Cumulative Gas, BCF
Prod. Rate
Rate Model
Cumulative
Cum. Model
Fig. 13–Denmark natural gas production model.
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
1000
2000
3000
4000
5000
6000
7000
8000
Cumulative Gas, BCF
Prod. Rate
Rate Model
Cumulative
Cum. Model
Fig. 14–France natural gas production model.
SPE 57463 ANALYSIS OF WORLDWIDE NATURAL GAS PRODUCTION 11
0
2
4
6
8
10
12
14
16
18
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
20000
40000
60000
80000
100000
120000
140000
160000
180000
Cumulative Gas, BCF
Prod. Rate
Model Rate
Cumulative
Cum. Model
Fig. 15–Norway natural gas production model.
0
2
4
6
8
10
12
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
20000
40000
60000
80000
100000
120000
Cumulative Gas, BCF
Prod. Rate
Rate Model
Cumulative
Cum. Model
Fig. 16–United Kingdom natural gas production model.
0
20
40
60
80
100
120
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
250000
500000
750000
1000000
1250000
1500000
1750000
2000000
2250000
2500000
Cumulative Gas, BCF
Actual Rate
Rate Model
Cumulative
Cum. Model
Fig. 17–FSU natural gas production model.
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
1000
2000
3000
4000
5000
6000
Cumulative Gas, BCF
Prod. Rate
Rate Model
Cumulative
Cum. Model
Fig. 18–Hungary natural gas production model.
0
5
10
15
20
25
30
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
50000
100000
150000
200000
250000
300000
Cumulative Gas, BCF
Prod. Rate
Rate Model
Cumulative
Model Cum.
Fig. 19–Iran natural gas production model.
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
5000
10000
15000
20000
25000
30000
35000
40000
45000
50000
Cumulative Gas, BCF
Prod. Rate
Rate Model
Cumulative
Cum. Model
Fig. 20–Iraq natural gas production model.
12 S.M. AL-FATTAH, R.A. STARTZMAN SPE 57463
0.0
0.5
1.0
1.5
2.0
2.5
3.0
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
5000
10000
15000
20000
25000
30000
Cumulative Gas, BCF
Actual Rate
Rate Model
Cumulative
Cum. Model
Fig. 21–Oman natural gas production model.
0
5
10
15
20
25
30
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
50000
100000
150000
200000
250000
300000
Cumulative Gas, BCF
Prod. Rate
Rate Model
Cumulative
Cum. Model
Fig. 22–Qatar natural gas production model.
0
5
10
15
20
25
30
1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
50000
100000
150000
200000
250000
300000
Cumulative Gas, BCF
Prod. Rate
Rate Model
Cumulative
Cum. Model
Fig. 23–Saudi Arabia natural gas production model.
0
2
4
6
8
10
12
14
16
18
20
1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
20000
40000
60000
80000
100000
120000
140000
160000
180000
200000
Cumulative Gas, BCF
Prod. Rate
Rate Model
Cumulative
Model Cum.
Fig. 24–United Arab of Emirates (U.A.E.) natural gas production
model.
0
2
4
6
8
10
12
14
16
18
20
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
20000
40000
60000
80000
100000
120000
140000
160000
180000
200000
Cumulative Gas, BCF
Prod. Rate
Rate Model
Cumulative
Cum. Model
Fig. 25–Algeria natural gas production model.
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
5000
10000
15000
20000
25000
30000
35000
40000
45000
50000
Cumulative Gas, BCF
Actual Rate
Rate Model
Cumulative
Cum. Model
Fig. 26–Egypt natural gas production model.
SPE 57463 ANALYSIS OF WORLDWIDE NATURAL GAS PRODUCTION 13
0.0
0.5
1.0
1.5
2.0
2.5
3.0
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
5000
10000
15000
20000
25000
30000
35000
Cumulative Gas, BCF
Prod. Rate
Rate Model
Cumulative
Cum. Model
Fig. 27–Libya natural gas production model.
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
5000
10000
15000
20000
25000
30000
35000
Cumulative Gas, BCF
Prod. Rate
Rate Model
Cumulative
Cum. Model
Fig. 28–Nigeria natural gas production model.
0
1
2
3
4
5
6
7
8
9
10
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
10000
20000
30000
40000
50000
60000
70000
80000
90000
100000
Cumulative Gas, BCF
Prod. Rate
Rate Model
Cumulative
Cum. Model
Fig. 29–Australia natural gas production model.
0
2
4
6
8
10
12
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
25000
50000
75000
100000
125000
150000
Cumulative Gas, BCF
Prod. Rate
Rate Model
Cumulative
Cum. Model
Fig. 30–China natural gas production model.
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
5000
10000
15000
20000
25000
30000
35000
40000
45000
50000
Cumulative Gas, BCF
Prod. Rate
Rate Model
Cumulative
Cum. Model
Fig. 31–India natural gas production model.
0
2
4
6
8
10
12
14
16
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
20000
40000
60000
80000
100000
120000
140000
160000
Cumulative Gas, BCF
Prod. Rate
Rate Model
Actual Cum.
Model Cum.
Fig. 32–Indonesia natural gas production model.
14 S.M. AL-FATTAH, R.A. STARTZMAN SPE 57463
0
2
4
6
8
10
12
14
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
0
10000
20000
30000
40000
50000
60000
70000
80000
90000
100000
110000
Cumulative Gas, BCF
Prod. Rate
Rate Model
Cumulative
Cum. Model
Fig. 33–Malaysia natural gas production model.
0
50
100
150
200
250
300
350
400
1960 1970 1980 1990 2000 2010 2020 2030 2040 2050
Time, year
Production Rate, BCF/D
Fig. 34–World natural gas production model.
0500 1000 1500 2000 2500 3000 3500
FSU
USA
Iran
Saudi Arabia
Canada
Qatar
UAE
Venezuela
Mexico
Algeria
TCF
Ultimate recovery
Future recovery
Fig. 35–Ultimate and future recoverable gas for top-10 countries
contributing 78% of world conventional ultimate gas.
0.1
1
10
100
1000
10000
0.1 1 10 100 1000 10000
Proved reserves, TCF
Future recoverable reserves, TCF
Fig. 36–Comparison between published proved reserves and
estimated future recovery of this study.
Cumulative Probability, %
1 2 5 10 20 30 50 70 80 90 95 98 99
Future Recoverable Reserves, TCF
10-1
100
101
102
103
104
Fig. 37–Cumulative probability plot of estimated future
recoverable reserves of natural gas.
Cumulative Probability, %
1 2 5 10 20 30 50 70 80 90 95 98 99
Ultimate Recoverable Gas, TCF
10-1
100
101
102
103
104
Fig. 38–Cumulative probability plot of estimated ultimate recovery
of natural gas.
