SPE 81010

Performance of Trinidad Gas Reservoirs (Cassia, Immortelle, Flamboyant, Mahogany,

Amherstia and Teak)
Richard Hallam, SPE, and Maria Lum Kin, SPE, BP Trinidad and Tobago

Copyright 2003, Society of Petroleum Engineers Inc.

This paper was prepared for presentation at the SPE Latin American and Caribbean Petroleum
Engineering Conference held in Port-of-Spain, Trinidad, West Indies, 27–30 April 2003.

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
The Cassia, Flamboyant, Immortelle, Mahogany, and
Amherstia fields currently produce over 1.4 bcf/d for the
Trinidad domestic gas market and Atlantic LNG Projects Figure 1: Location of the bpTT Gas Fields
(Trains I and II). By early 2004 the gas rate should increase to
about 2 bcf/d when two new projects start-up. These fields, The Cassia, Immortelle, Flamboyant, Mahogany and
plus the older Teak field are located in the Columbus Basin off Amherstia fields currently supply about 1.4 bcf/d of gas to the
the southeast coast of Trinidad. domestic market and to the Atlantic LNG Trains I and II. The
production will increase to about 2 bcf/d by early 2004 after
The six fields have produced gas from 35 reservoirs, which are two new projects come on stream: the Atlantic LNG Train III
at depths from 5,084 ft tvdss to 13,000 tvdss, and with initial Project and the Atlas Methanol plant. The Kapok field will
pressures from 2305 psi to 7420 psi. The production data start production in 2003 in order to meet the additional gas
show that the gases are condensates, with yields that vary from sales requirements.[1]
1 b/MMscf up to 50 b/MMscf.
The reservoirs exhibit a number of production mechanisms,
The reservoirs exhibit a number of production mechanisms, including volumetric, volumetric with compartmentalization,
including volumetric, volumetric with compartmentalization and water-drive (weak to strong aquifers). The paper reviews:
(due to faulting), water-drive (weak to strong aquifers) and • the recovery mechanisms,
waterdrive with compartmentalization. These reservoirs • production trends for different drive mechanisms,
generally produce at high initial rates (50 – 247 MMScf/d). • gas and condensate recovery factors,
• resource development
The paper shows examples of each recovery mechanism and o reservoir size,
reports the gas and condensate recovery factors. In addition, o well rates,
the paper documents the reserve distribution for the reservoirs, o number of wells per reservoir.
the number of wells required to develop each reservoir, and
the well rates. Geology and Field Description
The Teak, Cassia, Flamboyant, Immortelle, Mahogany, and
Introduction Amherstia gas fields are located in the Columbus Basin, which
The Teak, Cassia, Flamboyant, Immortelle, Mahogany, and is a sediment rich depositional center, off the southeast coast
Amherstia gas fields are operated by bp Trinidad and Tobago of Trinidad. The sediments were deposited from the Orinoco
(bpTT) and they are located between 20 and 40 miles off the river system in the Upper Tertiary and Quaternary periods.
southeast coast of Trinidad (see Figure 1). The initial gas The basin contains over 20,000 feet of sand, silt and shale
reserves from these fields were close to 10 Tcf of gas. sediments. The gas is trapped against large northwest-
2 SPE 81010

southeast trending faults where they are intersected by NE-SW The Immortelle field was discovered in 1968 and production
trending anticlinal ridges. started in 1994. This field contains at least 15 hydrocarbon-
bearing horizons (13, 14, 15, 16, 18, 19, 20, 21, 22, 23, 24, 25,
The Cassia field was discovered in 1973 and initial production 25N, 30, 32) that are trapped up against the A4 fault (see
started ten years later. The field contains four reservoirs (22, Figure 2). Four of these horizons (14, 16, 21 & 22 sands)
23, 25 and 30 sands) that are trapped up against the E fault contain a thin oil leg. The Immortelle field shares some of the
(see Figure 2). The 22 Sand is a water-drive reservoir, the 25 reservoirs (21, 22, 23, 24 sands) with the Amherstia field. The
Sand is volumetric and the 23 and 30 Sands are volumetric deep reservoirs in Immortelle (25N and 30) are volumetric,
reservoirs that are compartmentalized. These sands have while the shallower reservoirs (23 and shallower) appear to be
produced over 1.4 tcf and have remaining reserves of about waterdrive. The Immortelle reservoir had initial gas reserves
0.4 tcf. The properties for these reservoirs are reported in of 1.6 tcf. The properties for the Immortelle and Amherstia
Table 1. reservoirs are reported in Table 1.

Figure 4 Trapping mechanisms in the Amherstia Area

Figure 2: Cross Section of Cassia to Immortelle/Amherstia Fault The Amherstia field started production in 2000. The field
Block currently produces from 7 reservoirs (13UT, 13DT, 22, 23L,
23U, 24M, 24U Sand) (see Figure 4). The 22 Sand contains a
The Flamboyant field was discovered in 1986 and put on thin oil leg. (2,3) The Amherstia field had initial gas reserves of
production in 1993. The field contains 6 sands, in 5 fault 2.2 tcf.
blocks and there are at least 12 separate reservoirs (see Figure MAHOGANY FIELD REGIONAL DIP CROSS-SECTION
3). The main reservoirs are the MP50 and G60. The MP50 is B A
volumetric and the G60 is a compartmentalized waterdrive
reservoir with a strong aquifer. The properties for these
reservoirs are also reported in Table 1. The Cashima
discovery lies to the east of the Flamboyant platform.
II I
III
Flamboyant Platform Cashima Well V
IV
15 Sand 15 Sand
15 Sand 15 Sand

VI
17 Sand 17 Sand
17 Sand
17 Sand
132
18/18A Sands
18 Sand
19 Sand 19 Sand 18A Sand
MA-3
MA-7 MA-7X MB-3 MB-7
MA-5 MB-8
19 Sand
21 Sand 125
19 Sand
MB-9X
23 Sand (SUSPENDED) MB-4ST2 MB-6 MA-6 MA-15 MA-10 MB-5 MA-9 MB-10 MA-14 MA-13 MA-12 MA-11
MB-2 MA-4st1
23 Sand 20 Sand
EM-5
24 Sands
115 21 Sand
24A Sand 23 Sand
110
24A Sand
MA-8 24 Sands 100
25 Sand MB-9
23 Sand
25 Sand MB-1 MB-1ST MA-1

24A Sand
MA-2

VII 95
EM-2
24 Sands
65 90 EM-5Xst2
25 Sand
EM-3
24A Sand
EM-4
80

G60 25 Sand
BRUCE EGGERTSON NOV. 2000

Figure 5 Trapping mechanisms in Mahogany

MP50
The Mahogany field was discovered in 1968, but was not
produced until 1998 when it started supplying gas to the first
Atlantic LNG project (Train I). The field contains seven fault
Figure 3: Cross Section of Flamboyant Field and Cashima
SPE 81010 3

blocks and 10 gas-bearing sands: 15, 17, 18, 18a, 19, 20, 23,
7000
24, 24a and 25 (see Figure 5). The field also has two gas
reservoirs with a thin oil column and they are the 21 Sand 6000
Fault Block V and the 21 Sand Fault Block I. (4) The field 5000
contained initial gas reserves of 2.3 tcf. The properties for

P/z (psi)
these reservoirs are also reported in Table 1. 4000 Flamboyant MP50
3000
The Teak field contains both oil and gas reservoirs. The gas
2000 Cassia 25
reservoirs are found in some of the 2, 4, 6, 7 and 7al sands.
Figure 6 shows the field cross-section. Gas production started 1000
in 1974. The original gas reserves were about 1 tcf. The abandonm ent P/Z w /o com pression
0
properties for these reservoirs are also reported in Table 1. 0 50 100 150 200 250 300 350 400
Cum ulative Produced Gas (Bcf)

Figure 7 Straight-line P/Z plot for Volumetric Reservoirs

The compartmentalized volumetric sands recognized to date

are:
• Cassia 23,
• Cassia 30 sand,
• Immortelle 25N sand,
• Immortelle 30 sand.

The P/Z plots for the compartmentalized volumetric reservoirs

are shown in Figures 8-9. Both figures show a non-linear
relationship between P/Z and cumulative produced gas
volumes. The non-linear trend observed in the Immortelle
25N reservoir (Figure 8) was caused by faulting, that impeded
Figure 6 Cross Section of the Teak Field flow between the different fault blocks. The dashed lines in
Figure 8 show the size of the compartment around the well.
The produced gas from all the fields is processed offshore on These were determined using either pressure data (30 sand) or
the platforms. The separators on these platforms operate at calculated reservoir volumes (25N sand). The faulting caused
about 90 oF with an operating pressure of about 1000 psi +/- the reservoirs to behave like multiple connected compartments
100 psi. This gas is transferred to the buyer, while the liquids or tanks with flow being impeded between compartments. In
are transported to shore and processed. the Immortelle 25N reservoir the volume produced from the
well is significantly greater than the fault block containing the
Recovery Mechanisms well.
The production and pressure data from the 35 Trinidad
reservoirs have demonstrated that there are a number of 6000

different recovery mechanisms, which include: 5000

• Volumetric
• Volumetric with compartmentalization 4000
P/z (psi)

• Weak waterdrive 3000

• Strong waterdrive
• Waterdrive with compartmentalization.
2000

1000
abandonm ent pressure
The simple volumetric reservoirs recognized to date are: w /o com pression

• Cassia 25 sand,
0
0 50 100 150 200 250 300
• Flamboyant MP50 sand. Cumulative Wet Gas (Bcf)
Immortelle 30 sand Immortelle 25N sand
The P/Z plot for the Cassia 25 sand and Flamboyant MP50
sand is shown in Figure 7. These plots show the typical Figure 8 Non-linear P/Z plot for Immortelle Compartmentalized
straight-line relationship between P/Z and cumulative Volumetric Reservoirs
produced gas volumes for a volumetric reservoir. The
reservoir abandonment pressure without compression for the The non-linear trend observed in the Cassia sands (Figure 9)
Cassia 25 sand and Flamboyant MP50 sand are 1660 psi and was probably caused by a combination of faulting,
1730 psi, respectively. The abandonment P/Z values for these stratigraphy and the fact that all the wells were drilled at one
reservoirs are 1880 psi and 1940 psi, respectively. end of each reservoir. All these factors lead to reservoirs
behaving like multiple connected compartments with
4 SPE 81010

restricted flow from one compartment to another. For these

reservoirs the produced volume is close to the total gas-in-
4600 6000
place initially sensed by the wells. 18 fb II 23 sand FB V
4400 5500
19 sand FB II
7000 5000

P/Z (psi)

P/Z (psi)
4200
6000 4500
4000
5000 4000
P/z (psi)

3800
4000 3500

3000 3600 3000

0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
2000
Gas Recovery Factor Gas Recovery Factor
1000 abandonm e nt press ure
w /o com pre ssion Figure 11 Non-linear P/Z plot for Mahogany Waterdrive Reservoirs
0
0 200 400 600 800 1000 1200 1400

Cumulative Wet Gas (Bcf) The observations that can be made from all the reservoirs
Cassia 30 sand Cassia 23 Sand
developed with known recovery mechanisms are:
• Only 8% of the reservoirs (2 reservoirs) were simple
Figure 9 Non-linear P/Z plot for Cassia Compartmentalized volumetric reservoirs. These reservoirs contain about
Volumetric Reservoirs
8% of the reserves.
• All the volumetric reservoirs represent about 20% of
Recent log data from some horizontal wells drilled in the
the reservoirs and 52% of the reserves.
Immortelle and Amherstia 22 sand, showed that gas existed at
the same height as oil across a fault. Thus the faults acted as • Many reservoirs are compartmentalized by faults that
seals under static conditions. This statement is further act as barriers under static conditions but which allow
supported by the different gas-water and gas-oil contacts flow under dynamic conditions.
logged by Amherstia wells. Production from these reservoirs • The pressure behavior (P/Z) of compartmentalized
and others has indicated that under dynamic conditions these volumetric reservoirs are non-linear.
faults allow fluids to flow between compartments. • Early time pressure data may give a misleading
indication of the total size of the reservoir as it may
The confirmed waterdrive reservoirs have been: only see a small compartment around the well.
• Cassia 22 sand,
• Flamboyant G60 sand (compartmentalized), By analogy, many of the newer reservoirs are expected to be
• Teak 2, 4, 6, 7 &7al reservoirs, waterdrive and the percentage of volumetric reserves under
• Immortelle 18 sand, current development should drop to about 40% once further
• Mahogany 17, 18, 18a and 19 sands in Fault Block II, performance data is collected. Table 1 lists the reservoirs that
are expected to be waterdrive but the early data from these
• Mahogany 18a in Fault Block III,
sands is not conclusive.
• Mahogany 23 sand in Fault Block V,
• Mahogany 24, 24a and 25 sand in Fault Block VI. The volumetric reservoirs have tended to be the deeper
reservoirs within each field. However, the depth at which the
The P/Z plots for some of the waterdrive reservoirs are shown reservoirs mechanism changes from waterdrive to volumetric
in Figures 10-11. Both figures show the non-linear is different for each field (see Table 1). Also, the relationship
relationship between P/Z and cumulative produced gas is not a function of the reservoir being over-pressured (see
volumes for a waterdrive reservoir. Figure 12). Further work is underway to try and understand
why certain reservoirs have active aquifers and others are
6000 volumetric.
5000
0
Reservoir Depth (ft TVDSS)

4000
P/z (psi)

2000
3000
4000
2000 Gas In Place: 6000
1000 Flamboyant G60 = 133 bcf volumetric
8000
Cassia 22 = 261 bcf waterdrive
0 10000
0 50 100 150 200 250 300 12000
Cum ulative Wet Gas (Bcf) 14000
Flamboyant G60 sand Cassia 22 sand 0 2000 4000 6000 8000
Initial Reservoir Pressure (psi)

Figure 10 Non-linear P/Z plot for Flamboyant and Cassia

Figure 12 Reservoir Pressure Versus Depth
Waterdrive Reservoirs
SPE 81010 5

Gas Recovery Factors The relationship for the waterdrive recovery factor is:
To date 35 reservoirs have been produced, of which 12 have
been fully exploited and a further 6 have sufficient history to RF = 1 – (Pf/Zf)/ (Pi/Zi) [1 – Ev (Sgi – Sgt)/Sgi]
predict the gas recovery factor (see Table 2). Recovery factors
for 6 volumetric reservoirs and 12 waterdrive reservoirs have Where:
been estimated. RF = gas recovery factor
Pf = final or abandonment reservoir pressure
The recovery factor for the volumetric reservoirs is a function Pi = initial reservoir pressure
of (1) the initial reservoir pressure, (2) abandonment pressure, Zf = gas Z factor at pressure Pf and reservoir temp.
and (3) gas properties. The abandonment pressure is in turn a Zi = gas Z factor at pressure Pi and reservoir temp.
function of the (1) platform pressure, (2) reservoir depth, (3) Ev = sweep efficiency
pressure drops in the well completion and reservoir at Sgi = initial gas saturation
abandonment production rates. A prediction of the volumetric Sgt = trapped gas saturation
recovery factor is therefore relatively straightforward because
the only unknown variable is the abandonment pressure and The Trinidad reservoirs have shown a large variation in
this can be easily estimated. abandonment pressure (see Table 3). The Flamboyant G60
sand had an abandonment pressure of over 4500 psi that is
The six volumetric reservoirs have recovery factors between 75% of its initial pressure. The Mahogany data has also shown
0.61 and 0.74 (see Table 2). The relationship between existence of moderate to strong aquifers as the abandonment
recovery and reservoir pressure is shown in Figure 13. The pressures have varied from 1800 psi to 3600 psi. The Teak
lines shown on Figure 13 represent the expected recovery reservoirs have shown moderate and extremely weak aquifers.
factor based on typical reservoir depths, reservoir Three of these reservoirs were abandoned at pressures
abandonment pressures and gas properties. between 1400 and 1950 psi, which is close to values for
volumetric reservoirs.

0.9 The laboratory-measured trapped gas saturations range from

0.8 0.07 to 0.38. Most of these measurements lie between 0.15
0.7 and 0.25.
Gas Recovery Factor

0.6
0.5 The reservoirs developed to date typically only have 1 or 2
0.4 wells per reservoir. The water contact movement can
0.3 occasionally be determined from wells being drilled to deeper
0.2 horizons and from the producing wells. However, due to the
0.1 scarcity of data it is not possible to accurately determine the
0 sweep efficiency. Once the recovery factor and abandonment
0 2000 4000 6000 8000 pressure is estimated, then the sweep efficiency can be back
Reservoir Pressure (psi) calculated by assuming values for the trapped gas saturation.
The back-calculated sweep efficiencies are reported in Table
Figure 13 Recovery versus Pressure for Volumetric Reservoirs – 3. These efficiencies are useful references to forecast ultimate
dotted lines show expected range for recovery factor
recovery in new analogue water-drive gas reservoirs.
The recoveries from these reservoirs can be increased by
0.9
approximately 10-15% by reducing the platform operating
pressures by about 400 psi. As such, installation of 0.8

compression is viable means of increasing the recovery factor. 0.7

Gas Recovery Factor

Production of the Cassia 23 and 30 sands utilizing 0.6

compression will start up in late 2003. This could increase the 0.5
recovery for the 23 Sand from 0.62 to 0.78 and for the 30 Sand
0.4
from 0.70 to 0.83.
0.3
The recovery factor for the waterdrive reservoirs is a function 0.2
of the (1) initial reservoir pressure, (2) abandonment pressure, 0.1
(3) sweep efficiency of the aquifer, (4) trapped gas saturation,
0
and (5) to a lesser extent the gas properties. Some of these 0 2000 4000 6000 8000
variables depend upon the reservoir depth, reservoir properties Reservoir Pressure (psi)
and well placement. A prediction of a waterdrive recovery
factor is therefore more complex because there are three Figure 14 Recovery Factor for Waterdrive (blue symbols) and
unknown variables - abandonment pressure, sweep efficiency Volumetric Reservoirs (red symbols)
and trapped gas saturation - that can vary considerably.
6 SPE 81010

The twelve waterdrive reservoirs have recovery factors that Figure 16. The volumetric reservoirs had a condensate
varied from 0.53 to 0.83 (see Figure 14). The high recoveries recovery range of 0.29 to 0.40, with an average of 0.34.
observed in the Teak reservoirs (0.73 to 0.83) resulted from
the weak water drive in these sands. The Cassia, Flamboyant The waterdrive reservoirs on Cassia, Flamboyant and
and Mahogany reservoirs had an average recovery of 0.65 Mahogany had a condensate recovery factor range of 0.31 to
with a range of 0.53 to 0.80. The aquifer strength in these 0.68, with an average of 0.50. The values for these water-drive
reservoirs varied from ‘weak’ through to ‘strong’. reservoirs were higher than the volumetric reservoirs because -
as a consequence of the aquifer support - the reservoirs were
Condensate Recovery Factors abandoned at a higher pressure.
The gas in the Trinidad reservoirs [Teak, Cassia, Immortelle,
Flamboyant, Amherstia and Mahogany] is a mixture of The Teak waterdrive reservoirs had condensate recovery
biogenic gas and thermogenic gas. These two different sources factors in the range of 0.16 to 0.42, with an average of 0.26.
then migrated to the reservoirs, forming different mixtures in Teak, even with the aquifer pressure support, had lower
each reservoir. The data indicate that there is (1) an areal condensate recovery values than the volumetric reservoirs.
variation, (2) a general increase in the yield with reservoir Jemmott et al ascribe this anomaly to the different fluid
pressure or depth, and (3) the yields vary from 1 b/MMScf to characteristics of the Teak reservoirs. (5)
50 b/MMScf. The condensate yield data for each field shows a
correlation when plotted against the reservoir pressure (or
depth) (see Figure 15). Well Performance Trends
Many of the wells in the volumetric reservoirs had similar
behaviors and they showed three distinct periods during their
60 producing lives:
• A constant rate period due to tubing limitations
50 (erosional limits),
• An exponential, or near exponential decline period,
40
Yield (B/MMscf)

and
• Reduced rate periods caused by market demands.
30
An example of this behavior is shown in Figure 17 for the
Flamboyant Fa-01 well in the MP50 sand.
20

10 120
Gas Rate (MMScf/d)

100
0
2000 3000 4000 5000 6000 7000 8000 80
Initial Reservoir Pressure (psi) 60
40
Cassia Immortelle Flamboyant Mahogany Teak
20
Figure 15: Initial Condensate Yield versus Reservoir Pressure 0
0 50 100 150 200 250
5
Cumulative Gas Production (bcf)

4 volumetric waterdrive
Figure 17: Gas Rate data for Flamboyant well Fa-01 in the MP50
Number of Reservoirs

sand
3

A well completed in a waterdrive reservoir showed three

2 distinct periods during its producing life:
• a constant rate period due to tubing limitations
1 (erosional limits),
• a declining period controlled by the reservoir,
0 • a period with a high decline rate after water
0.1-0.2 0.2-0.3 0.3-0.4 0.4-0.5 0.5-0.6
breakthrough,
Condensate Recovery Factor
and these periods can be very varied.

Figure 18 shows the rate data for the Mahogany well Ma-05.
Figure 16: Condensate recovery factor
This well essentially produced at its maximum rate until water
To date, it has been possible to estimate the condensate breakthrough occurred, and then the well loaded up within a
recovery factor for 15 reservoirs. These values together with few months. This well produced 94% of its reserves before
the average produced yields are reported in Tables 2-3 and water breakthrough.
SPE 81010 7

volumes of gas after water breakthrough whereas other

100 500
reservoirs (wells) stopped producing shortly after
80 400
breakthrough. On average, 78% of the reserves produced per
well is recovered before water breakthrough, and the range
Gas rate (MMScf/d)

Water Rate (b/d)

60 300
varied from 56% to 100%.

40 200 10

Number of Reservoirs
8
20 100
6
0 0 4
0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0
2
Cumulative Gas Production (bcf)
0
Figure 18: Gas rate data for Mahogany well Ma-05 in the 19 Sand 0.4-0.5 0.4 -0.5 0.6-0.7 0.7-0.8 0.8-0.9 0.9-1
Fault Block II Fraction produced from a well
before water breakthrough
Figure 19 shows the rate data for the Flamboyant well Fa-02.
This well produced 81% of its reserves before water
Figure 21: Fraction of reserves produced before water
breakthrough. breakthrough

80 3500
Every attempt is made not to shut-in waterdrive gas reservoirs
due to market constraints because this could reduce the
70 3000
Gas Rate (MMScf/d)

recovery factor and reserves as the aquifer encroaches.

60
Water Rate (b/d)

2500
50
2000 Market and Well rates
40 The Trinidad gas fields discussed in this paper [Teak, Cassia,
1500
30 Flamboyant, Immortelle, Mahogany and Amherstia] have
1000
20 been brought on stream from 1974 through current times.
10 500 Originally, some of the gas was used to supply gas lift gas to
0 0 the fields but most of it was sold to the domestic market. In
Nov-93

Nov-94

Nov-95

Nov-96

Nov-97

Nov-98

Nov-99

Nov-00

Nov-01

Nov-02

1999 Atlantic Train I was added, and in 2002 Atlantic Train II

started-up. Atlantic Train III will start-up in mid 2003
followed by the Atlas Methanol Project in late 2003 or early
Figure 19: Gas Rate data for Flamboyant well Fa-02 in the G60 2004. Further expansion projects are likely over the 2005/2006
sand timeframe. The gas sold by bpTT to these markets is shown in
Figure 22.
Figure 20 shows the rate data for the Teak well Td-05. This
well produced at rates from 60 MMscf/d to 25 MMscf/d
before water breakthrough after which the rate dropped off 2.5
quickly to around 5 MMscf/d.. This well produced 93% of its actual projected

reserves before water breakthrough. 2

Gas Rate (bcf/d)

1.5
60 800

50
700 1
600
Gas Rate (MMScf/d)

40 0.5
Water Rate (b/d)

500

30 400 0
300
96

97

98

99

00

01

02

03

04

05

06

20
19

19

19

19

20

20

20

20

20

20

20

200
10
100

0 -
Figure 22: bpTT Trinidad gas sales
- 10 20 30 40 50 60 70 80 90
Cumulative Produced Gas (bcf) The average well rates have generally increased with each
field that has been developed (see Figure 23). The maximum
Figure 20: Gas Rate data for Teak well Td-06 in the Teak 6 sand
well rates have also generally increased over time (see Figure
The performance of the waterdrive reservoirs was variable 24). These rates have increased from 50 MMscf/d in Teak
(see Figure 21). Some of the reservoirs produced moderate (Tb-12), through 75 MMScf/d in Cassia (Ca-01), then 100
8 SPE 81010

MMscf/d in Flamboyant (Fa-01), 150 MMScf/d in Mahogany permeability (3) compartmentalization and (4) drive
(Mb-02 & Ma-02), 227 MMscf/d in Amherstia (A-06) and mechanism. All of the volumetric reservoirs have one well,
then finally 247 MMScf/d in Mahogany (Mb-13). These with the exception of a single (low-permeability) large
increases were in part, due to an understanding of the reservoir reservoir (>1 tcf). Only 4 of the gas reservoirs have been
quality and size – historically, many of the wells have developed with three or more wells and each of these had
produced for long periods at rates constrained by tubing large reserves (three close to 300 bcf and one over 1 tcf) and
erosion limits and not by the reservoir. This understanding led low permeabilities (7 md to 32 md). Clearly, the smaller-sized
to the use of bigger tubing sizes and a desire to (1) reduce the reservoirs require fewer wells to recover their reserves. As
well count per reservoir, (2) maximize the reserves and (3) Figures 2 – 6 show, the Trinidad gas reservoirs exist in highly
meet market requirements. compartmentalized and complex structures, and as discussed
previously, the majority of the reservoirs are water-drive.
120 These reservoirs have been developed with one or two wells.
Average Well Rate

100
8 1200
(MMScf/d)

80
7
1000

Number of Reservoirs
60 6

Reserves (bcf)
40 800
5
20 4 600
0 3
400
nt

2
y
l le

tia
k

ia

an
a

ya
s

te

rs
Te

as

og

200
bo

or

he

1
C

ah
m
am

m
Im

A
Fl

0 0

0-50

50-100

500-1000
100-150

150-200

200-250

250-300

300-350

350-400

400-450

450-500

> 1000
Figure 23: Average well rate in each field

Reservoir Size (bcf)

300
Maximum Well Rate

250 Figure 25: Reservoir Size – Current gas reservoirs, exclusive of

gas reservoirs with thin oil columns
(MMScf/d)

200
150
100 4
Number of Reservoirs
50 18 13 1 2 1
Total Reserves (tcf)

0 3
Tb-12 Ca-01 Fa-01 Ma-02, A-06 Mb-13
(1974) (1983) (1993) Mb-02 (2001) (2003) 2
(1999)
Well (Year start production)
1

Figure 24: Maximum well rate history 0

1 2 3 4 5
A study by Kromah et al (6) also demonstrates that the Trinidad Number of wells per Reservoir
gas well rates can increase in an attempt to achieve high
deliverability with fewer wells and consequently, reduce the
Figure 26: Total Reserves per number of wells per reservoirs
overall amount of development capital.

Reservoir Size and Well Count BpTT is currently evaluating the potential of 4D seismic to
The gas reservoirs developed to date in the bpTT gas portfolio identify by-passed reserves in gas reservoirs. A clearer
have reserves that vary from about 20 bcf up to 1.1 tcf. The understanding of the well count per reservoir may be obtained,
average reservoir size is 190 bcf, and the average should this technology prove to be applicable to our
reserves/well has averaged 132 bcf. The reserves distribution environment.
per reservoir is shown in Figure 25.
Conclusions
Most of the reservoirs have been developed using only one or a) The Trinidad gas reservoirs have shown four different
two wells. Figure 26 shows the total reserves and number of production mechanisms and they are (1) volumetric, (2)
reservoirs versus the number of wells that were used to volumetric with compartmentalization (3) waterdrive with
develop each reservoir. The number of wells required to weak to strong aquifer support and (4) strong waterdrive
develop the gas reservoirs is a function of (1) reservoir size (2) with compartmentalization.
SPE 81010 9

b) The volumetric reservoirs have a gas recovery factor that

varies from 0.61 to 0.74 and they show a relationship with
initial reservoir pressure and depth. These reservoirs
displayed an average condensate recovery factor of 0.34
with a range of 0.29 to 0.40.

c) The waterdrive reservoirs have a gas recovery factor that

varies from 0.53 to 0.83.

d) The waterdrive reservoirs in Flamboyant, Cassia and

Mahogany displayed an average condensate recovery
factor of 0.50 with a range of 0.31 to 0.68. The values for
these waterdrive reservoirs were higher than the
volumetric reservoirs because of the aquifer pressure
support that meant the reservoirs were abandoned at a
higher pressure.

e) The Teak waterdrive reservoirs displayed an average

condensate recovery factor of 0.26 with a range of 0.16 to
0.42. Teak, even with the aquifer pressure support, had
lower condensate recovery values than the volumetric
reservoirs and this was because of the different fluid
characteristics.

f) The reservoirs have varied in size, having reserves of 20

bcf to 1.1 tcf, with an average of 190 bcf.

g) To date, close to 89% of the gas reservoirs and 66% of the

gas reserves have been developed with 1 or two wells per
reservoir.

Acknowledgements
The authors would like to thank bpTT for permission to publish the
paper.

References
1. Lumsden, P.: “The Kapok Field, Trinidad: A Step Change for
Trinidad Gas Developments” paper SPE 75670 presented at SPE
Gas Technology Symposium in Calgary, Alberta, Canada Apr. 30
– May 2 2002.
2. Bayley-Haynes, E. & Shen, E.: “Thin Oil Development in the
Amherstia / Immortelle Fields, Offshore Trinidad” paper SPE
81088 presented at 2003 SPE Latin American and Caribbean
Petroleum Engineering Conference in Trinidad, Apr. 27-30.
3. Pucknell, J., Holder, G. & Seesahai, T.: “Amherstia Field,
Trinidad: Breaking Barriers in the New Millennium” paper SPE
81012 presented at 2003 SPE Latin American and Caribbean
Petroleum Engineering Conference in Trinidad, Apr. 27-30.
4. Ali-Nandalal, J., Staines, M., Bally, Y.K., Finneran, J., “Optimal
Locations and Performance Prediction of Horizontal Oil Wells in
the Oil Rim at Mahogany Field, Offshore Trinidad” SPE 56814
presented at 1999 SPE Annual Technical Conference in Houston,
Texas, October 3-6.
5. Jemmott, S., Hallam, R. & Maharaj, S.: “Condensate
Performance Trends in Trinidad Gas Reservoirs” paper SPE
81011 presented at 2003 SPE Latin American and Caribbean
Petroleum Engineering Conference in Trinidad, Apr. 27-30
6. Kromah, M.J., Lumsden, P.J., Hennington, E.R. and Brayshaw,
A.C., “Trinidad’s First 500 MMscfd well: Fact or Fiction?”
paper SPE 81045 presented at 2003 SPE Latin American and
Caribbean Petroleum Engineering Conference in Trinidad, Apr.
27-30
10 SPE 81010

Table 1 : Trinidad Reservoir Property Data

Field Reservoir Ref Depth Res. Pressure Permeability Initial Yield (Yi) Drive
ft TVDSS Psi md b/MMscf Mechanism
Cassia 22 8150 3712 - 8 waterdrive
Cassia 23 9200 4370 5 25 Volumetric
Cassia 25 11150 5150 6 9 Vol. Comp
Cassia 30 12200 7420 13 36 Vol. Comp
Teak 2 8600 3700 85 20 waterdrive
Teak 4 9554 4420 32 25 waterdrive
Teak 6 10908 5020 70 30 waterdrive
Teak 7 12415 5732 7 30 waterdrive
Teak 7al 13000 6525 7 50 waterdrive
Immortelle 13 5927 2638 970 1 waterdrive ?
Immortelle 15 6930 3084 145 4 waterdrive ?
Immortelle 18 7806 3473 180 10 waterdrive
Immortelle 23 9764 4345 165 13 ?
Immortelle 25N 10951 5085 254 22 Vol., comp
Immortelle 30 12100 5385 6 25 Vol., comp
Flamboyant G60 10440 6000 500 16 waterdrive
Flamboyant MP50 12730 7389 1025 24 volumetric
Amherstia 13UT 5200 2342 300 1 ?
Amherstia 13DT 5200 2308 480 1 ?
Amherstia 23L 23 U 9250 4350 38 9.5 ?
Amherstia 24M 24 U 9490 4500 20 9 ?
Mahogany 17 FBII 7690 3477 747 6 waterdrive
Mahogany 18 FBII 8315 3845 853 8.4 waterdrive
Mahogany 18a FBII 8500 3915 138 9.5 waterdrive
Mahogany 18a FBIII 8315 3522 131 10 waterdrive
Mahogany 19 FBII 8850 4132 560 6.3 waterdrive
Mahogany 20 FBIV 9685 4758 539 15.5 ?
Mahogany 20 FBV 9685 4731 418 15.8 ?
Mahogany 23 FBV 11022 5875 188 22 waterdrive
Mahogany 24 FBVI 11000 5983 253 27 waterdrive
Mahogany 24a FB V 12519 7352 6 31 volumetric*
Mahogany 24a FB VI 11600 7328 5 35 waterdrive
Mahogany 25 FBVI 12573 6008 83 20 waterdrive

Footnotes:
? = unknown as yet
volumetric * = well in small compartment - main reservoir may be waterdrive
vol, comp = Volumetric with Compartmentalization
waterdrive ? = expected to be waterdrive but too early to have definitive data
SPE 81010 11

Table 2 : Volumetric Reservoir Metrics

Field Reservoir Ref Depth Res. Pressure Gas Recover Condensate Drive
ft TVDSS Psi Rec. Factor Rec. Factor Mechanism
Cassia 23 9200 4370 0.61 0.29 Volumetric
Cassia 25 11150 5150 0.61 0.34 Vol. Comp
Cassia 30 12200 7420 0.74 0.33 Vol. Comp
Immortelle 25N 10951 5085 0.67 0.40 Vol., comp
Immortelle 30 12100 5385 0.67 - Vol., comp
Flamboyant MP50 12730 7389 0.70 0.35 volumetric

Table 3 : Waterdrive Reservoir Metrics

Field Reservoir Ref Depth Res. Pressure Aban. Pressure Gas Condensate Sweep Efficiency Drive
ft TVDSS psi psi Rec. Factor Rec. Factor (Back Calculated) Mechanism
Cassia 22 8150 3712 2714 0.67 0.48 0.86 waterdrive
Teak 2 8600 4074 2500 0.74 0.42 0.89 waterdrive
Teak 4 9554 4420 1950 0.73 0.22 0.68 waterdrive
Teak 6 10908 5020 1450 0.74 0.24 0.35 waterdrive
Teak 7 12415 5732 2600 0.78 0.25 0.89 waterdrive
Teak 7al 13000 6525 1700 0.83 0.18 0.68 waterdrive
Flamboyant G60 10440 6000 4500 0.64 0.57 0.86 waterdrive, comp
Mahogany 17 FBII 7690 3477 3050 0.59 0.49 0.59 waterdrive
Mahogany 18 FBII 8315 3845 3343 0.60 0.37 0.81 waterdrive
Mahogany 18a FBIII 8315 3522 2200 0.53 * 0.31 0.40 waterdrive
Mahogany 19 fb II 8850 4132 3600 0.70 0.57 0.92 waterdrive
Mahogany 23 fb V 11022 5875 1850 0.80 0.55 0.67 waterdrive

* recovery factor low due to premature failure of the well (sand production)