SPE 99720

Oligocene Vicksburg Thin-Bed Production Optimization Derived From Oil-Based Mud

Imaging: A Case Study
D.L. Fairhurst, B.W. Reynolds, S. Indriati, and M.D. Morris, SPE, Schlumberger, and E.G. Hanson, Abaco Operating LLC

Copyright 2006, Society of Petroleum Engineers

This paper was prepared for presentation at the 2006 SPE Gas Technology Symposium held
in Calgary, Alberta, Canada, 1517 May 2006.
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 Oligocene Vicksburg formation in South Texas has been a
prolific play for many years with targets of thick and stacked
sand bodies. These thick sections have been primarily
exploited and produced. Still existing are many previously
considered uneconomical sequences. These marginal sections
consist of highly laminated sand shale sequences along with
disbursed clay in sand. Standard cutoffs from basic log
evaluation work correctly for the disbursed clay sections. But
the cutoffs are inadequate for the highly laminated sequences;
many thin, high-quality sands have been overlooked. These
sections can now be discerned using microresistivity
measurements in oil-based mud systems and new highresolution cutoffs can be employed.
A production prediction model is critical to enhance the
chance of success. The model used here employs a
petrophysically consistent high-resolution permeability
estimate, fracture geometry prediction, and formation
pressure. The methodology identified several sands as
commercial that have been bypassed in offsets with the old
cutoffs.
Over a two-year drilling program, data gathered from several
field example wells were analyzed. These are presented here
to illustrate how production data was utilized to continuously
adjust and calibrate the high-resolution petrophysical model.
The incremental revenue from the added pay exceeded the
cost of this new methodology and enhanced the economic
viability of the field.
This integrated process of measurement, analysis, prediction,
evaluation, and model adjustment enables the operator in
South Texas to make timely completion decisions as well as
set-pipe decisions. This process is becoming a useful tool for

further exploitation of the mature Oligocene Vicksburg

formation of South Texas.
Introduction
The Vicksburg formation in South Texas has been exploited
since the 1920s and is still a prolific producer with over 20
Bcf per year average rate (Fig. 1). The play has seen both
productivity increases and declines depending on gas prices
and technology drivers. Since the mid-1990s, however, the
trend has been ever-decreasing productivity and faster rate
declines. At the same time, only 12% of the estimated 3,860
Bcf ultimate recoverable designated tight gas in Vicksburg has
been produced1, leaving much to be recovered. Some of this
recovery can be enhanced with recently developed highresolution technology.
The decision on whether to set pipe or complete a particular
zone usually is made once the logging run is complete.
During the standard logging run, the analyst will view the
density porosity output and question the economics. What is
the porosity cutoff to make a well here? The answer is found
over years of experience and the school of hard knocks.
Typically a Rule of Thumb is used and a line is drawn (Fig.
2). Many South Texas partners make their decisions based on
these cutoffs and individual experience. Worthington gives a
comprehensive perspective on the use of these cutoffs2. The
cutoff number most often used in the Oligocene Vicksburg
trend of South Texas is 15-16% porosity (Fig. 2). More
recently there has been success at much lower porosity in the
range of 8-10%3. Obviously, if a 16% porosity cutoff was
applied routinely, then somewhere in the thousands of wells
drilled, some pay has been bypassed.
One solution that has been used primarily in water-based
systems has been laminated sand analysis. This type of
analysis has been applied since the early 1990s primarily in
turbidite plays4 and not verified with production. The analysis
used here verified with production data, provides a better
answer for the less obvious and often bypassed pay sands.
Thin Bed Production Optimization Technique
Overview:
The standard cutoff technique that has been described above
can be applied to high-resolution petrophysical analysis
utilizing enhanced stratigraphic imaging. The method used
here was to predict the wells initial production via high

resolution petrophysical analysis, measured or estimated initial

reservoir pressure and assumption of the fracture height,
length and conductivity. Production data was compared with
the prediction and the high resolution cutoffs were corrected to
make a match. This iterative process was applied over a series
of wells. The result is a working predictive model that can
ensure the operator the best possible economic decision for
completion. The main difference from the standard resolution
cutoffs is net height and to a lesser extent, the relative
permeability to gas value. With the high-resolution analysis, it
can be shown that the net height of producible pay can go
from 10 feet to 50 feet, or stay at 10 feet in the instance of
disbursed clay.
Petrophysical Analysis:
Enhancement of the petrophysical analysis using highresolution stratigraphic imaging has been applied since the
early 1990s in water-based muds. In the oil-based mud
system, a similar analysis has only been available since 1999.
The importance of this is that most of the current drilling for
South Texas Vicksburg wells is now done with oil-based mud
for cost and better wellbore integrity. Cheung5 and Tabanou6
defined a convolution process for oil-based mud imaging. The
original intention of this device was to provide the geologist
with a stratigraphic image. For thin bed analysis however, the
same data can be used to compute a resolution enhanced
petrophysical analysis. Standard resistivity resolution is about
18 inches for 100% of the signal (fig. 3). The enhanced image
from oil-based mud imaging has an effective resolution of 1.2
inches6. This ultra-fine resolution gives a much clearer picture
of the highly laminated sand sequences of the Vicksburg
formation. Many sands are on the order of several feet down
to a few inches in thickness. Imaging tools also differentiate
between disbursed clay and laminated sand shale sequences
(fig 4). Thinly laminated sands are masked by the resolution of
standard triple combo logging tools and appreciable pay can
be bypassed.
The use of microimaging tools along with triple combo data
opens the door to resolution enhanced interpretations. The first
step to this sharpened analysis is to use the image data to
determine a lithofacies model of sand shale sequences. Next,
this lithofacies model is used with a mathematical method to
correct standard resolution logs to the best representation of
what they would read if they had the vertical resolution of
microimaging tools. The mathematical model is part of an
optimization process that calculates a squared output that is in
turn convolved and compared to the original data. Multiple
iterations refine the sharpened log by minimizing the
difference between the convolved output and the original logs.
The convolution process is used to correct the resistivity,
density, neutron, and gamma ray (Fig 5). These sharpened
outputs can then be used to compute an enhanced
petrophysical analysis that yields corrected effective
porosities, water saturations and estimates of relative
permeability for water and gas. Ahmed, Crary and Coates
described the techniques used for log derived permeability
estimates7. Here, a similar process is applied with magnetic
resonance permeability when available. The water saturation
correction is reliable wherever the mud invasion is low. It is

SPE 99720

typical that there is very little mud invasion in the deep

Oligocene Vicksburg and it is easily verified with the
induction curves.
Stimulation model:
Stimulation practices have been described over time for the
Vicksburg. Tucker describes steep declines observed and
attributed to proppant embedment8. Abrams and Vinegar
found that the water block additives when laboratory tested on
Vicksburg core samples were not effective9. Brin found that
high strength proppants led to 15% increase in NPV even
though the cost of the proppant was 6 times higher than
sand10. Similar practices were applied in all the wells in this
study. A pseudo 3D fracture design model was used and from
25 to 50 different layers were characterized, utilizing the
calibrated sharpened analysis described in the previous section
(Fig. 6). Fracture direction and height were then synthesized
via log data and the fracture modeling. The fracture design
lengths and conductivities were optimized for each of the
Vicksburg multi-strings based on the net height, permeability
values and layers obtained from the sharpened process. The
resultant stimulations varied in size from 140,000 lbs. to
400,000 lbs. of bauxite (20/40). The number of stages per
well also varied from 2 stages to 6 stages.
Production Data Analysis for Petrophysical Model
Calibration:
The first step to calibrate the cutoff and the permeability to gas
estimate of the sharpened petrophysical model was to perform
production history matching. Next was to characterize the
reservoir and fracture properties layer by layer in the wellbore.
The methodology to calculate the production from individual
layer and to characterize the reservoir and fracture properties
layer by layer in a wellbore has been well documented in
several papers.11,12 Daily production and wellhead pressure
was collected for three to six months from several wells in the
field. These wells have commingled production from several
stimulated layers. The goal of the production history matching
was to characterize the reservoir properties layer by layer
rather than the average value. A computational model was
used to generate individual production and flowing bottomhole pressure history from the total commingled production
with the aid of production logging13. Once the individual layer
production contribution and pressure traverse components
were determined, a unique rate transient analysis was
performed on each layer as if it was the only zone being
produced from the well14. The analysis was performed
graphically by matching the production data slope with the
theoretical slope of each of the transient flow regime in a
fractured well. The slope from each of the flow regime yields
specific reservoir and fracture properties layer by layer in the
wellbore. The result from this graphical rate transient analysis
was then compared and refined with the result from a
comprehensive analytical production simulator15.
The second step was to compare the results obtained from
production data analysis to the initial estimate from the
petrophysical model and fracture modeling. When the initial
estimate of the reservoir properties from the petrophysical
model did not support the results obtained from the production

SPE 99720

data analysis, the petrophysical model was adjusted. Primary

adjustments were made to saturation and porosity cutoffs to
determine net pay as well as the permeability model to
estimate permeability to gas.
The third step in the process and probably the most critical
was to perform production forecasting utilizing the adjusted
petrophysical model in order to evaluate the economics of
each potential zones and frac stage. The analytical production
simulator described above was utilized to forecast the
production layer by layer along with the commingled
production of the entire wellbore. The newly adjusted net pay
calculation, porosity, saturation, permeability to gas and water,
initial reservoir pressure, drainage area and modeled fracture
properties are the primary inputs in the production forecasting
step. For comparison purposes the decline analysis was
performed using petrophysical outputs based on standard
resolution analysis as well as the calibrated sharpened
petrophysical analysis.
Surface daily production and wellhead pressure were then
monitored and compared with the production forecast using
the calibrated model. This iterative method was conducted
until the production forecast using the calibrated petrophysical
model and the actual production data matched, resulting in an
optimized calibration for the petrophysical model.
Case Study
Well A is representative of the initiation of the sharpened
analysis study. Production forecasting was not available for
this well at the time of completion. In this well, with standard
analysis, density, neutron with a field cutoff of 15% gave a net
pay of 20 feet in the lower zone. Conventionally, this lower
zone would have been viewed as non-economic and bypassed.
The induction log was sharpened with microimaging, a field
cutoff of 15% was still utilized and net sand count improved to
55 feet of pay in this process. Stimulation was performed for
this lower zone. Production testing showed initial production
of 2MMcf/D coming from this zone that conventionally would
have been bypassed. There were two fracture treatments
performed on this wellbore. After the production was
commingled, a production log was run to further confirm the
contribution of the lower sand. The production log verified
that this zone was still contributing as the lowermost of three
commingled stimulation treatments (Fig 7). Four months of
production and wellhead pressure data were collected. The
methodology to characterize the reservoir and fracture
properties described above was applied (Fig. 8). The
production data analysis resulted in a change of the porosity
cutoffs from 15% to 12 %.
Well B is a representative of the second group of wells that
utilized the first attempt of the calibrated high-resolution
analysis. The production prediction yielded an estimate that
initial production would be 12 MMcf/D. This is a combined
two stage completion. Each stage was predicted using the
production optimization technique described above. The result
was under-calling the production. Actual field well production
came in higher than the prediction by 2mmcfd (Fig. 9).
Production logging was not available for this well, thus the

analysis was done on commingled basis. The production

analysis indicated that further adjustment of the porosity and
water saturation cutoffs was needed to predict more net pay.
Porosity cutoff was changed from 12% to 9% and water
saturation cutoff changed from 65% to 70%. The result was a
close match to the actual field production.
Well C and Well D utilized the calibrated results from Well B
for production prediction. Well C is a combination of three
stages of completion. Production logging confirmed the
contribution from all three stages and the actual production
followed closely to the production prediction, within 10-15%
range (Fig. 10). Well D only has one stage completed. The
actual production is in the 10-15% range of the production
prediction (Fig. 11). The optimization process now appears to
be a working predictive model for this field. The result is an
aid for the operator to make timely and effective completion
decisions.
Some of the wells declined faster than predicted after 50-60
days. It is believed that the reason is due to proppant
embedment from the high closure stress and soft shales.
Furthermore, geopressured reservoirs tend to have some
permeability compaction through time and faster fracture
degradation.
Conclusions
Thin bed production optimization can greatly improve the
operators economical success in completing thin bed sands.
Intervals that could be bypassed using standard log resolution
are revealed with the new cutoffs applied (Fig. 12). The
application tested here can be applied in many other laminated
sand reservoirs. The process will work for both oil-based and
water-based drilling systems.
The enhanced thin bed cutoff process used in this study relied
on already proven techniques. The process however is a
summation utilizing these techniques along with comparison
over time and adjustments to the model. The adjustments were
made through actual production analysis. The same process
may result in different cutoffs for other geological plays and
add again to the bank of bypassed pay discoveries.
After several iterations and changes, the actual production
matched the prediction more closely (within 10-20%). A
working predictive model that can ensure the operator the best
possible economic decision for completion in this field has
been established. The model can be easily re-applied to
evaluate different completion scenarios.
Although some wells declined faster than predicted after 5060 days, it is assumed that this is approximately the point at
which proppant embedment occurs. However, this
phenomenon is beyond the scope of this study and further
effort is necessary to prove this.
Acknowledgements
We thank Schlumberger for the time to complete this study.
Also B.J. Drehr and Kathy Huddleston of Abaco for their help
in providing data.

Nomenclature
k = permeability, L2, md
h = height
References
1. Gas Technology Instiute, Tight Gas Resource Map of the
United States, GTI-01/0114 (2000)
2. Worthington, P.F., The Role of Cut-offs in Integrated
Reservoir Studies, SPE Paper 84387 (2003) 16 p.
3. Erskine, R.D., Out Front In Deep Gas The American Gas
Reporter, (Dec. 2001) p. 44-54.
4. Hansen S M, T Fett.2000.Identification and evaluation of
turbidite and other deepwater sands using open hole logs and
borehole images. In A H Bouma and C G Stone. Fine-grained
turbidite systems, AAPG Memoir 72/SEPM Special
Publication 68. U.K.: Geological Society Publishing House,
317338.
5. Cheung, P. et al.: Field Test Results of a New Oil-Base
Mud Formation Imager Tool, (2001), SPWLA 42nd Annual
Logging Symposium.
6. Tabanou, J.R. et al.: Thinly Laminated Reservoir
Evaluation In Oil-Base Mud: High Resolution Versus Bulk
Anisotropy Measurement-A Comprehensive Evaluation,
(2002), SPWLA 43th Annual Logging Symposium.
7. Ahmed, U., Crary, S.F., Coates, G.R., Permeability
Estimation:
The
Various
Sources
and
Their
Interrelationships, SPE Paper 19604 (1991)
8. Tucker, R.L., Practical Pressure Analysis in Evaluation of
Proppant Selection For The Low Permeability, Highly
Geopressured Reservoirs of the McAllen Ranch (Vicksburg)
Field, SPE Paper 7925 (1979) 9 p.
9. Abrams, A. and Vinegar, H.J., Impairment Mechanisms in
Vicksburg Tight Gas Sands, SPE Paper 13883 (1985) 12 p.
10. Brin, H.P., A Post-Audit of Fracture Stimulations in the
Vicksburg Formation of South Texas, SPE 15508 (1986) p 8.
11. England, K.W, Poe, B.D.Jr., Conger, J.G.,
Comprehensive Evaluation of Fractured Gas Wells Utilizing
Production Data, SPE Paper 60285 (2000).
12. Larkin, S.D, Pickrel, H.M, et.al,Analysis of Completion
and Stimulation Techniques in a South Texas Field Utilizing
Comprehensive Reservoir Evaluation, SPE Paper 93996
(2005) 10 p.
13. Poe B.D Jr., Villarreal, R., et.al. Production Optimization
Methodology for Multilayer Commingled Reservoirs Using
Commingled Reservoir Production Performance Data and
Production Logging Information, US Patent Pending Sept
(2000).
14. Poe, B.D. Jr., Conger, J.G., et al.: Advanced Fractured
Well Diagnostics For Production Data Analysis, SPE Paper
56750 (1999) 22 p.
15. Poe, B.D. Jr., Zheng-Poe, A., Boney, C.L.: Production
Data Analysis and Forecasting Using a Comprehensive
Analysis System, SPE Paper 52178 (1999) 8 p.

SPE 99720

SPE 99720

Number Of Wells

50,000

2500

45,000

2250

40,000

2000

35,000

1750

30,000

1500

25,000

1250

20,000

1000

15,000

750

10,000

500

5,000

250

0
1965

1970

1975

1980

1985

1990

1995

Time
Fig 1. Vicksburg production data, 1965 to 2005, gas rate compared to number of wells drilled.

Net Pay Using 15 PU

Cut-off = 20ft
Considered noneconomical

Fig 2. Rule of Thumb cut-off technique applied in the Vicksburg.

2000

2005

0
2010

Number of Wells

Gas Rate (MMCF

Gas Rate (MMSCF)

SPE 99720

Rt

AIT Correct

Rt

AIT Misses 50%

Rt

TCOM misses 100%

Fig 3. OBMI Resistivity compared with Array Induction, Density and Neutron.

Disbursed Clay
Induction

Neutron/Density

Fig 4. Disbursed clay compared with laminated sand/shale sequence.

Laminated Pay

SPE 99720

Sharp outputs

BFV

GR

RT

Fig 5. Sharpened outputs after 5 iterations.

ABACO OPERATING
6
Vicksburg 5
12-07-2004

FracCADE*
ACL Fracture Profile and Proppant Concentration
14100

Well Depth - ft

14150
< 0.0 lb/ft2
0.0 - 0.3 lb/ft2
0.3 - 0.5 lb/ft2
0.5 - 0.8 lb/ft2
0.8 - 1.0 lb/ft2
1.0 - 1.3 lb/ft2
1.3 - 1.6 lb/ft2
1.6 - 1.8 lb/ft2
1.8 - 2.1 lb/ft2
> 2.1 lb/ft2

14200

14250

14300

14350
13200

14000
Stress - psi

14800

-0.2

-0.1
0
0.1
ACL Width at Wellbore - in

0.2

*Mark of Schlumberger

Fig 6. Geometrical pseudo 3d stimulation model used per layer.

500

1000
Fracture Half-Length - ft

1500

SPE 99720

Fig 7. Well A production log confirmation of the contribution from the added zone.

Actual Gas Rate

Sharpened Analysis History Match

Forecast with Standard Analysis

14,000

Gas Rate (MSCF/D)

12,000
10,000
8,000
6,000
4,000
2,000
0
0

20

40

60

80

100

120

140

160

Time (Days)

Fig 8. Well A actual production rate, production history match with sharpened analysis, production forecast with standard
analysis for comparison purposes.

SPE 99720

Forecast with adjusted cutoffs

Actual Gas Rate

Forecast with conventional cutoffs

16,000.00
14,000.00

Gas Rate (mscf/d)

12,000.00
10,000.00
8,000.00
6,000.00
4,000.00
2,000.00
0.00
0

10

20

30

40

50

60

70

Time (Days)
Fig 9. Well B Field production is 2 MMcf/D higher than predicted.

Actual Gas Rate

Forecast with Sharpen Analysis

7,000
6,000

Gas Rate (MSCF/D)

5,000
4,000
3,000
2,000
1,000
0
0

10

20

30

40
Time (Days)

Fig 10. Well C Production matches prediction up to 30 days.

50

60

70

80

10

SPE 99720

Forecast with Sharpened Analysis

Actual Gas Rate

9,000
8,000

Gas Rate (MSCF/D)

7,000
6,000
5,000
4,000
3,000
2,000
1,000
0
0

10

20

30

40

50
Time (Days)

Fig 11. Well D Production matches prediction up to 50 days.

60

70

80

90

100

SPE 99720

11

Flow Prediction From Standard Log Analysis.

Clark Sain #12

Flow Prediction From Sharpened Log.

Both Standard and High Resolution

See The Thick Sand, Same Rates.

The Sharpened High Resolution

Predictions Show Pay That Would
Have Been Bypassed.
The Actual Well Production Matches
The High Resolution Prediction.

Fig 12. Standard resolution rate predictions compared with high resolution production predictions.