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.