SPE 92036

Forty Years of Improved Oil Recovery: Lessons from Low-Permeability Turbidites of

the East Wilmington Field, California
Creties Jenkins, SPE, DeGolyer and MacNaughton; Shahin Al-Sharif, SPE, OXY Long Beach Inc.; Randy Harris, SPE,
THUMS Long Beach Company; John Weisgram, SPE, Bainbridge Petroleum Consulting; and Don Michel, SPE,
DeGolyer and MacNaughton

Copyright 2004, Society of Petroleum Engineers Inc.

This paper was prepared for presentation at the 2004 SPE International Petroleum Conference
in Mexico held in Puebla, Mexico, 89 November 2004.
This paper was selected for presentation by an SPE Program Committee following review of
information contained in a proposal 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 a proposal of not more than 300
words; illustrations may not be copied. The proposal 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 U-P Ford zone in the Long Beach Unit of the East
Wilmington Field consists of low-permeability (2-50
millidarcies) turbidites that have been waterflooded since field
start-up. Forty years of successes and failures have provided
valuable insights into how to best waterflood these reservoirs
given their thin-bedded nature, lateral and vertical changes in
reservoir quality, formation damage susceptibility, and sand
control problems. Multiple techniques and technologies have
been applied to describe their reservoir architecture, quantify
reservoir performance, and extract the oil. This work has
become more challenging as the waterflood has matured and
will require the close integration of all disciplines to identify
and exploit remaining opportunities.
Introduction
The Long Beach Unit (LBU) of the East Wilmington Field has
been producing oil from waterflooded turbidite reservoirs
since 1965. The productive horizons include four sandstone
intervals that combine to form a 4000-foot thick oil column.
(Figure 1). The reservoirs are produced primarily from four
artificial drilling islands in Long Beach Harbor. The deepest
of the four intervals, the Union Pacific (U-P) Ford zone, has
the poorest reservoir quality and has therefore been the most
challenging interval to develop and produce.
To date, 198 wells have been drilled into the U-P Ford zone
and there are currently 75 active producers and 35 active
injectors. The U-P Ford produces 3,500 barrels of oil per day
(BOPD) and 36,500 barrels of water per day (BWPD).
Current water injection is 45,000 BWPD. These rates must
increase through more efficient waterflooding if the full
potential of the U-P Ford is to be realized. The Long Beach

Unit is very mature and production from shallower intervals is

declining. As a result, much of the oil in the U-P Ford is at
risk because this interval alone cannot carry the production
costs of the Unit.
Reservoir Geology
Introduction. The East Wilmington Field is part of the
northwest-trending Wilmington Anticline (Figure 2). This
structure lies between the Palos Verdes and NewportInglewood fault zones, and is underlain by the THUMSHuntington Beach fault. This is a southwestward-directed
blind thrust that offsets the basement (Figure 3). The anticline
is approximately 11 miles long and 3 miles wide, with dips
averaging about 20 degrees on the north flank and 60 degrees
on the south flank (Figure 4). The anticline is broken into a
series of individual fault blocks that compartmentalize the
reservoirs. The bounding faults are normal faults, tear faults,
or a combination of the two. The faults dip at 40-80 degrees
and their displacements range from 50-450 feet1.
The gross thickness of U-P Ford zone is about 2100 feet, and
depths to the top of the zone range from about 4200 to 4800
feet. The U-P interval is about 900 feet thick and consists of
the AE, AF, AI, AK1, AL1, and AM subzones. The U-P has
an average net-to-gross ratio of 0.25 and contains many thin
sands separated by shales, as well as some thicker sands. The
Ford interval is about 1200 feet thick and underlies the U-P
interval. The Ford consists of the AO, AR, AU, AU2, AV,
AW, AX, AX1, AY, AT, and AZ subzones. It has an average
net-to-gross ratio of 0.45 and contains thicker sand units than
the U-P, although many of the sand-rich intervals are
frequently found in the aquifer. The major sandstone units
show an overall thickening and higher net-to-gross values
from west to east.
Depositional Setting and Rock Properties. The interbedded
sands and shales of the U-P Ford zone formed in a
progradational, submarine-fan system trending from northwest
to southeast2. The sands were deposited by sediment gravity
flows that were funneled basinward by submarine canyons.
These flows spread laterally across the surface of the fan to
form lobes that exhibit compensating geometries. This pattern
is characterized by the thick portion of an underlying lobe
coinciding with the thin portion of an overlying lobe. Between

these episodic flows, fine-grained clastic and biogenic

material settled out of the water column to form intervening
shales. Individual sandstone beds range from less than one
foot to more than 10 feet in thickness. The thicker beds
formed by the stacking (amalgamation) of thinner beds
representing individual flows (Figure 5).
First-order shales, which are typically tens of feet thick,
extend across the entire U-P Ford zone and can be used to
establish a correlation framework. These shales are fluid
barriers and represent periods of submarine fan abandonment
due to decreased sediment supply. The sand packages
between these first-order shales typically contain thinner,
second-order shales that were deposited atop abandoned lobes
as the locus of deposition shifted laterally (Figure 6) These
shales serve as baffles to fluid flow, but could also be fluid
barriers if they effectively encase an entire lobe. Second-order
shales typically show an off-lapping character and have
steeper dips than first-order shales.
The log character and core descriptions of the U-P Ford
sandstones indicate several types of stacking patterns (Figure
5). Some sandstones show a coarsening and thickening
upward character suggestive of the prograding outer fan
portion of a lobe. Other sandstones show a blocky or finingand-thinning upward character indicating a channel. These are
thought to be unleveed distributaries on the fan surface that
are likely filled with coarser and cleaner sandstones. They
will therefore be preferred paths for fluid movement. Some of
these features have been identified and correlated to water
breakthrough in wells3.
U-P Ford sandstones are lithic arkoses composed of quartz,
plagioclase, potassium feldspar, mica, and rock fragments.
They are fine to coarse-grained and very poorly to well-sorted.
Clays comprise 10-20% of the rock, predominantly as
structural shale. Clay minerals include smectite, illite,
kaolinite and chlorite. Plagioclase grains commonly show
dissolution and replacement by these clays. Authigenic
kaolinite and illite are present as cement, as are siderite and
calcite. The U-P Ford contains multiple sand beds with
pervasive carbonate cement and range from a few inches to
five feet in thickness4. These are excellent correlation
horizons and some beds can be traced across the entire Unit
using porosity logs.
Pore types in the U-P Ford include interparticle macropores,
moldic macropores, and appreciable amounts of interparticle
microporosity associated with clay matrix and cement5. As
this microporosity increases, permeability decreases and
capillarity increases, resulting in higher water saturations and
poorer reservoir quality. As a result, permeabilities are
relatively low (2 to 50 millidarcies). To make matters worse,
the clay minerals in the U-P Ford can damage the reservoir
through clay swelling (smectite), fines migration (kaolinite),
and the formation of iron-hydroxide precipitates from
acidizing (chlorite and Fe-illite).
History and Impact of Description Work. The thin-bedded,
heterogeneous nature of the U-P Ford sandstones has made it

SPE 92036

difficult to calculate oil-in-place, quantify reservoir properties,

and assess waterflood performance. As early as 1970, it was
recognized from core and log data that the U-P Ford was
characterized by an erratic vertical permeability profile,
laterally discontinuous strata, cross-flow between strata, areal
permeability variations, and compartmentalization by faults
and fractures6. These heterogeneities, coupled with a
peripheral waterflood, led to relatively poor recoveries in most
of the U-P Ford sands in the 1970s and 1980s.
In 1992, an optimized waterflood was implemented in the
Long Beach Unit that included pattern injection and hydraulic
fracturing in the U-P Ford zone. To help quantify remaining
potential, a 3D-seismic survey was conducted in 1995 to
delineate complex faulting and areas of better sand
development. However, due to acquisition problems caused
primarily by poor weather conditions, data quality was
particularly poor over crestal areas containing U-P Ford
sandstones, and bed thicknesses and fault offsets of less than
about 50-75 feet could not be resolved7. The 3D seismic has
been useful for clarifying the geometries of major faults and
identifying amplitude anomalies that may indicate thicker,
cleaner sands (Figure 7). However, seismic modeling shows
that the rock and fluid properties of the U-P Ford do not
generate sufficient contrast to resolve whether the entrained
fluid is oil or water. The seismic data is currently being
reprocessed to provide better resolution, and the acquisition of
additional 3D seismic data has been discussed.
Volumetric estimates of original oil in place (OOIP) have
varied considerably over time. The most significant upward
revision occurred in 1998 with an increase in OOIP that nearly
doubled the previous value. This revision was the result of
revised pay counts and oil-water contact estimates as well as
new porosity and saturation models from cores and logs.
Using this higher OOIP, rate extrapolations based on historical
performance yielded a very low ultimate recovery efficiency.
Taken at face value, this indicates the potential for significant
additional waterflood recovery through aggressive waterflood
management. Alternatively, the higher OOIP may be too
optimistic. A key issue is whether thinly-bedded sands were
properly accounted for in this study, and whether these sands
should be considered effective pay for waterflooding. To help
understand this, a new U-P Ford reservoir characterization
study was conducted in 2003-2004.
The geological portion of this characterization study has shed
light on the architecture of thinner intervals3. Each of the 17
major subzones of the U-P Ford (intervals between first-order
shales) have been divided into 1-5 cycles based on changes
in core and log character. The cycles correspond to different
sand lobes and the mapping of these cycles shows a
compensating geometry that also appears in shallower zones
of the Long Beach Unit8. The cycles are capped by shales
which may or may not form a laterally-continuous barrier to
fluid movement. This zonation is a key first step to
understanding how stratigraphy impacts the distribution of
flow units, oil-water contacts, and reservoir compartments. A
geocellular model has been constructed using this zonation
and reservoir properties have been distributed using the results

SPE 92036

of the revised petrophysical model described in the next

section.
Petrophysics
Previous Studies. In the 1970s and 1980s, petrophysical
studies in the Long Beach Unit were largely driven by equity
redeterminations, with different parties conducting analyses
tailored to support their positions. Much of the work revolved
around the determination of porosity and permeability values
from core plugs, which are poorly-consolidated and must be
compacted to approximate in-situ conditions. In the U-P Ford
zone, compaction studies on cores from well C-657 in 1986
showed that porosity decreases from an average value of
22.6% at 90 psi confining stress to 18.8% at an in-situ
confining stress of 2800 psi. Similarly, permeability decreases
by ten-fold from an average of 800 md at 90 psi to 88 md at
2800 psi. Capillary pressure measurements on these same
cores showed a wide-range of irreducible water saturation
values ranging from 30 to 80% depending on permeability
(Figure 8). This is expected given the heterogeneous
composition of the U-P Ford sandstones.
Beginning in the early 1990s, several studies were conducted
to develop log-based algorithms for computing porosity and
water saturation from a subset of U-P Ford wells9,10. A key
part of this work was the development and application of an
inversion tool to improve the vertical resolution of induction
logs for thin bed detection (Figure 9). The studies also
attempted to use old sonic and neutron count logs for
determining porosity in early U-P Ford wells, but ultimately a
decision was made to map Vshale and porosity values from
more recent wells and use these to compute water saturations
in old wells.
A third study11 attempted to reconcile oil-base core saturations
and capillary pressure saturations from wells cored in the late
1960s, and then use these to calibrate log-derived water
saturations. This work reduced the log-derived water
saturations by an average of 10-20 saturation units, but the
quality of the core measurements and the match between oilbase and capillary pressure saturations reduced the confidence
in these results.
Over the past decade, the petrophysical models generated for
the U-P Ford have not been used on a routine basis. Their
complexity, the time required to conduct the analysis, and the
uncertainty in the results makes the old practice of counting
pay using a resistivity cutoff attractive. While this technique
may work in thicker sands at or near original conditions,
quantifying the remaining potential in thin sands and/or
partially swept sands requires a more rigorous approach.
New Petrophysical Model. As part of the 2003-04 reservoir
characterization study, a new, more comprehensive
petrophysical model has been developed for the U-P Ford
zone to calculate shale volume (Vsh), porosity, permeability
and water saturation in wells with or without porosity logs.
The raw logs used to develop this model were acquired from
multiple logging tools and companies from the late 1960s to
the present. Most of these wells were logged with 6FF40-style

induction logs, but some used dual laterologs because of high

salinity mud. Other logs included gamma ray (GR), formation
density, spontaneous potential (SP), sonic, compensated
neutron, and GR-neutron logs. Recent wells have either
obtained logging-while-drilling (LWD) resistivity and GR
curves, or traditional array induction, density, neutron, and GR
logs on wireline.
For this new petrophysical model, poorer-quality logs,
especially the bulk density curves, had to be normalized to
generate a consistent dataset. Shale volumes were calculated
using a Stieber-type relationship applied to a boreholecorrected GR log. This type of equation is commonly used in
arkosic or lithic sands that are more radioactive than typical
quartz sandstones. Where no GR log was available, the
minimum of SP and resistivity indicators was used to calculate
shale volumes.
Total porosities calculated from the logs were calibrated to a
dataset of 3200 compaction-corrected core porosities (CPHIC)
from 21 wells. An examination of the CPHIC versus subsea
depth plot (Figure 10) reveals that in addition to a depthrelated compaction gradient, there appear to be some
variations related to stratigraphic changes. For example, most
porosity values for the AM subzone are below the regression
line, indicating that some combination of factors (poorer
sorting, more clay and/or more cement) results in lower
porosity and poorer sandstone quality. The final porosity
versus depth relationship used a constant slope with offsets
varied by subzone.
Density porosity was calculated using a constant grain density
based on core data and a constant fluid density selected to
provide a good match to the CPHIC data on a field-wide basis.
This technique was used in all intervals where valid density
measurements were available. In intervals where no density
log was present, a sonic log or the plot of CPHIC versus
subsea depth was used to estimate total porosity. Carbonatecemented sandstones were identified using a porosity limit
equal to 5 porosity units less than the porosity-depth trend. In
wells without porosity logs, these tight sands were defined
qualitatively using a variety of methods that depended on the
data available in each case.
Permeability curves were computed from two logarithmic
functions of effective porosity that were derived using core
plug air permeability measurements corrected to liquid
permeability at in-situ conditions. The core permeabilities in
the U-P Ford sandstones range from less than one to several
hundred millidarcies in a few samples. The arithmetic average
of these core permeabilities is 14.3 md and the geometric
average is 4.4 md.
The Juhsz equation12, a dual-water model, was selected to
calculate log-derived water saturations, and the input
parameters were chosen to generate reasonable agreement
with oil-based core and capillary pressure saturations in the UP Ford. In wells with 6FF40-style induction logs or dual
laterologs for resistivity, an inversion program was used to
calculate a true formation resistivity.

The results of the study show that the average porosity ranges
from the mid-teens to low twenties with an overall average of
about 19 percent. Water saturations approach 20 percent in the
best quality sands, but average 53 percent for all sands that
meet a 70 percent water saturation cut-off for pay. This limit
corresponds to a water cut of approximately 97 percent. Net
sands are defined as those non-tight sands with a shale volume
of less than 42 percent. This shale volume limit was chosen to
yield sand counts that were very similar to those estimated
from core sand counts.
This new petrophysical model will be applied to both existing
and future U-P Ford wells, and the results have been used to
distribute properties in the new geocellular model (Figure 11).
This model has been constructed for both numerical
simulation and well planning. The new petrophysical model
provides a robust and consistent technique to analyze the U-P
Ford core and log data, but there are still lingering
uncertainties with regard to the initial and current water
saturation values in thin sands, and whether these can be
effectively waterflooded.
Operations
Performance History. When development of the U-P Ford
zone began in 1965, the pressure in many flow units was subhydrostatic due to an estimated 15 million barrels (MMBBLS)
of off-lease drainage. The majority of this is attributed to the
Belmont Offshore Field which adjoins the Long Beach Unit
on the southeast and produced oil for 11 years prior to LBU
start-up. Since start-up, the focus of waterflooding has been
the area east of the Long Beach Unit Fault (U-P Ford East).
West of this fault (U-P Ford 98), oil production is confined to
a relatively small area (Figure 4). The sands in U-P Ford 98
are thinner, of lower quality, and have higher water saturations
than U-P Ford East.
Early production from U-P Ford East was dominated by
higher-quality sands in the lower part of the Ford, especially in
the AX1 subzone, but these sands watered-out in the 1970s
and completions moved up-hole into the upper Ford and U-P.
During the 1970s and 1980s, the U-P Ford was primarily a
peripheral waterflood. However, due to aquifer losses and the
cycling of injection water through down-dip producers with
poor vertical conformance, aquifer injection was unable to
support producers at the crest of the anticline, where pressures
in some wells fell to 40-50% of hydrostatic. This led to the
drilling of pattern injectors and the hydraulic fracturing of
producers in the 1990s to improve performance.
The U-P Ford reached a peak rate of about 30,000 BOPD in
1969, after which the rate declined at more than 20% per year
through the mid-1970s (Figure 12). In 1976, the oil rate
increased as a result of oilfield activitythe first of many
such increases caused by drilling, recompletions, new
technology, and/or higher oil prices. These drilling and
completion techniques have evolved dramatically since the
first U-P Ford wells. From 1965 through the 1980s, most UP Ford wells were drilled and completed by running slotted
liners in open hole. By the middle of the 1980s, completion
targets had progressed uphole into thinner, poorer-quality flow

SPE 92036

units but the completion methods remained largely unchanged.

Rate performance was poorer in these sandstones due to lower
permeabilities, inadequate injection support, sand control
problems, and damaging completion and workover techniques.
Production indices (PI) in the U-P Ford are low relative to the
shallower Ranger and Terminal reservoirs. The average PI for
a U-P Ford well ranges from 0.2 bbls/psi to 0.8 bbls/psi. A
map of PI values shows that these change areally over the
reservoir and are a function of reservoir heterogeneity,
pressure support, and drilling and completion techniques
(Figure 13). The choice of which techniques to apply can have
a huge impact on well productivity, and therefore are the focus
of constant improvement. Over the past 20 years, innovations
have included better drilling methods, less damaging
workover fluids, cleaner injection water, cased hole
completions, hydraulic fracture stimulation, horizontal wells,
and improvements in sand control.
Drilling. Nearly 2000 wells have been drilled in the Long
Beach Unit since 1965, with as many as 20 rigs running at one
time. The Unit has maintained a one-rig pace since late 1999,
drilling 26 wells per year. In mid-2004, a second rig was
activated which will result in a total of 37 wells being drilled
in 2004, of which 10 will be U-P Ford wells. Thirteen more
U-P Ford wells are planned for 2005. New injectors cost $0.71.0 million depending on depth, and new cased-hole fracturestimulated producers cost $1.2-1.5 million. The time it takes to
drill a well has decreased substantially over the past 20 years
for many reasons including 1) improvements to steerable
bottom-hole assemblies, (2) logging-while-drilling, (3) the use
production rigs to complete the wells, and (4) the use of
premium milltooth bits. Minimizing these expenses is critical
because higher costs demand higher oil rates to justify
continued development of the U-P Ford.
Workover Fluids. Many different fluid compositions and
techniques have been tried in the U-P Ford zone to minimize
formation damage associated with swelling clays and fines
migration. Several studies were conducted during the early
1990s to identify the damage mechanisms and methods to
both remediate the damage and prevent it from occurring. This
damage is particularly important to avoid in workover
operations where productivity losses are very difficult to
restore without redrilling the well. Fluids used have included
potassium chloride (KCl) water and surfactant mixtures, sized
salt with KCl water, lease crude oil, diesel oil, and polymers.
Studies have shown that the greatest damage is caused by
sized salt with KCl water and lease crudeprobably due to
the way the crude is handled at the surface. These studies and
field experience have also shown that the least damaging kill
fluid is 4% KCl with a surfactant.
Large kill fluid volumes were found to be a primary
formation damaging factor during workovers. These volumes
sometimes approached three times the wellbore volume. This
large volume combined with low reservoir pressure ensured
that significant amounts of kill fluid invaded the formation,
which changed the wettability, caused clays to swell, and
mobilized fines. A limited kill procedure was therefore

SPE 92036

developed in which a reduced volume of kill fluid is pumped

into the well to prevent possible surges and gassing. This
procedure avoids replacing the entire volume of the well and
has become the preferred means for reducing damage in U-P
Ford wells. In a similar fashion, bailers are used to cleanout
fill from wells instead of circulating fluids and exposing the
formation to potential damage.
Cleaner Injection Water. The U-P Ford zone receives about
5% of the water injected into the Long Beach Unit. Because
water injectors are not permitted to inject above parting
pressure, it is critical that the water is clean so that
contaminants are not introduced that can reduce injectivity.
These include solids that can bridge pore throats and oxygen
that promotes the growth of bacteria. In addition, cleaner
water reduces the need for acid stimulations. Recent
improvements in water quality have been made as a result of
better monitoring, improved backwashing of filter media,
better chemical treatments and modifications to skimmer
units.
Cased Hole Completions. Selective perforated cased-hole
completions were adopted in the early 1990s as an alternative
to traditional slotted liner completions. These cased-hole
completions are perforated in selective oil-bearing sands
whereas slotted liners are open across a long interval that may
include high water-cut sands. In some instances, the casedhole completions result in well production rates that are
insufficient to cool the electrical submersible pump, resulting
in pump failures. These low rates are primarily the result of
having less interval open to the wellbore, but there are other
factors that contribute to this problem. These include
formation damage (skin factors as high as 4) associated with
drilling mud, cementing operations and perforating practices,
and the inability to resolve and perforate those thin oil sands
that would be captured in a slotted liner completion. As a
result, cased-hole completions in the upper Ford and U-P
intervals average 208 BPD of gross fluid whereas open-hole
completions average 335 BPD. Both completion types are
used depending on the purpose of the well and the
interpretation of pay distribution from the logs.
Hydraulic Fracture Stimulations. The first hydraulic
fracture stimulations of the U-P Ford Zone in the East
Wilmington Field were pumped from 1963-1965. The chosen
wells were located in Belmont Offshore which adjoins the
Long Beach Unit on the southeast (Figure 2). Five Belmont
wells along the leaseline were stimulated by pumping 10,000
to 50,000 pounds of 40-60 sand resulting in an average rate
increase of 50%. The first Long Beach Unit U-P Ford frac job
followed in 1992, but U-P Ford wells were not fracture
stimulated on a regular basis until 1996.
Poorly-consolidated sands of the U-P Ford pose special
problems for hydraulic fracture stimulation. First, the stress
contrast between sands and shales is very high, limiting the
vertical growth of fractures and requiring multiple stimulation
stages. Second, because the U-P Ford typically has
permeabilities in the tens of millidarcies, high fracture
conductivities (7,000 to 10,000 millidarcy feet) are needed to

double or triple the well rate. Third, because of the poorlyconsolidated nature of the U-P Ford, the proppant will become
embedded in the sandface and reduce the fracture width unless
resin-coated sand or a similar product is used to create a
web of proppant in the fracture.
Despite these complications, frac jobs were successfully
designed and pumped in 50 wells from April, 1996 to
December, 1998 (Figure 14). The typical job was designed
for a tip screen-out and pumped 150,000 to 200,000 pounds of
16-30 resin coated sand in a borate cross-linked gel followed
by a persulfate breaker. Skin factors as low -4 were observed
in these wells and the incremental oil rate from these jobs
peaked at 5000 BOPD in January, 1998. The success of this
program in new wells led to an aggressive campaign to
increase rates by fracturing older U-P Ford wells. Most of the
fracture stimulation workovers of older wells merely required
adding perforations and then fracturing in multiple stages. On
occasion, more complex practices were conducted, such as
dumping cement down an existing slotted liner, drilling it out,
perforating, and pumping the frac job.
In 1999, the fortunes of the hydraulic fracturing program
changed with decreases in initial well rates and long-term
performance. Contributing factors included insufficient
pressure support for the stimulated wells and changes in the
fracture stimulation design including shorter fracture lengths,
different perforating strategies, and modified pumping rates.
In addition, the use of ceramic proppants instead of resincoated sand contributed to proppant flow-back problems in 8
of the 10 wells in which ceramic proppant was used. As a
result, the frac jobs pumped from 2001 through 2003
significantly underperformed (Figure 15).
Figure 15 also shows that recent fracture stimulated well
performance has significantly improved due to several
changes in job parameters including perforation strategies,
proppant type, and well selection. One well in particular
stands out for its success the first U-P Ford frac-and-pack
was completed in April, 2004 producing 400 BOPD and 300
BWPD. Perhaps this completion technique is the key to
another surge of U-P Ford oil production.
Sand Control. Most U-P Ford wells have not required sand
control, but as the shallow U-P sands have been exploited,
more sanding problems have been encountered. In addition,
geography plays a role as demonstrated by greater sanding in
several wells in Fault Block 8 (Figure 4). This may be due to
a more mature waterflood in this areamore sand is often
produced after water breakthrough because there is less sticky
oil to hold the sand grains together. Sanding issues cause a
number of operational problems, particularly since all U-P
Ford wells are lifted using electrical submersible pumps
(ESP). Desanders are installed on most ESPs and electrical
submersible progressive cavity pumps have been used in
severe sanding cases.
Inner liner gravel packs have been attempted in wells
completed with open-hole slotted liners and in cased-hole
wells that have been fracture stimulated. In both cases, this

technique has generally resulted in a loss of production,

perhaps due to formation damage. In addition, wells with
inner liner gravel packs have higher decline rates possibly due
to fines plugging of the pack. Frac-and-packs hold out the
greatest promise for improved sand control, especially if these
completions consistently result in higher well rates than
traditional frac jobs.
Horizontal wells. Three horizontal wells have been drilled
and completed in the U-P Ford zone. All three wells were
drilled along faults to capture pooled reserves. Well C441 was
completed in 1996 into sands of the AO subzone (Ford
interval). The well has a 1092-foot horizontal section and was
completed with a wire-wrapped screen in open hole. This well
initially produced 230 BOPD and 90 BWPD, but declined
rapidly and was idled as uneconomic after producing 80
thousand barrels of oil (MBO) in just over 3 years.
Well B438A was drilled into the AF subzone (U-P interval) in
1995. The well has a 1284-foot horizontal section and was
completed with a wire-wrapped screen in open hole. B438A
had initial rates of 80 BOPD and 20 BWPD but declined
rapidly. Formation damage from a sized-salt drilling mud was
suspected, so the well was redrilled as B438B with 1958 feet
of horizontal section. This time the well was cased,
perforated, and hydraulically fractured in three stages. Initial
rates for this well were only 40 BOPD and 5 BWPD, so an
additional 570 feet of perforations were added in the
horizontal section and stimulated with acid. Unfortunately,
this only increased production by 10 BOPD.
Although the results have been disappointing, horizontal wells
are still being evaluated as a way to exploit selected sands.
Successful horizontal wells are common in shallower sands of
the Long Beach Unit where the sands are thick (20-50) with
good permeabilities (tens to hundreds of millidarcies),
homogeneous properties, and pressure support.
Waterflood Surveillance
Performance modeling of wells in the U-P Ford East area uses
a combination of decline curve analysis, log WOR versus
cumulative oil extrapolations, material balance, and numerical
simulation. Waterflood recovery can be reasonably well
projected using decline curves and plots of log WOR versus
cumulative oil, but these techniques assume current conditions
will persist into the future. This is not a good assumption in
the U-P Ford where drilling campaigns, new technologies, and
oil price fluctuations cause changes in the slope of the curves
(Figures 12 and 16). Fractional flow concepts also need to be
included in these rate extrapolations.13 For example, a plot of
specific productivity index (SPI) versus water cut shows a
sharp increase in SPI at water cuts above 90% using U-P Ford
relative permeability and well performance data (Figure 17).
This demonstrates that gross fluid rates will increase
dramatically at high water cuts, requiring more water handling
and re-injection.
The projected ultimate recovery (using extrapolations of
existing performance to an economic limit) can be compared
with the OOIP volume and with estimates of ultimate recovery

SPE 92036

efficiency to establish the remaining potential of U-P Ford

East. Based on fractional flow theory,14,15 and assuming a
volumetric efficiency of 0.70,16,17 it is estimated that 38% of
the OOIP can be recovered at a limiting WOR of 30 (97%
water cut) by injecting 1.8 hydrocarbon pore volumes of
water. Combining this with the current estimate of OOIP
results in a recovery factor for U-P Ford East of 29.4% of
OOIP. By comparison, the recovery factor using the 1998
OOIP is 17.0%. This indicates a very large uncertainty in
remaining potential and demonstrates how important it is to
accurately determine OOIP.
To more accurately determine the remaining waterflood
potential in the U-P Ford East, material balance-based pattern
analysis (pattern surveillance) can be used. Historically,
similar techniques have been applied to the Ranger and
Terminal zones for waterflood management.18,19 This
technique divides the U-P Ford East areally into pattern
elements bounded by injection wells and faults. These
elements are then vertically segregated into subzones and
injected volumes are assigned to each subzone over time using
periodic spinner surveys from each injector. These surveys,
combined with the vertical isolation of subzones by first-order
shales, allows injection throughput to be vertically allocated
with good confidence. Although fluid entry surveys are rarely
obtained in producing wells, the subzone production of oil and
water can be estimated using fractional flow behavior based
on relative permeability data.
By combining the production and injection data with
fractional flow concepts, pressure histories and an estimate of
volumetric OOIP, a variety of dimensionless plots can be
generated to assess pattern performance. Figure 18 shows five
pattern elements located in cut-recovery blocks (CRB) 44 and
45 of U-P Ford East. Figures 16, 19 and 20 are examples of
composite dimensionless plots from these CRBs. Figure 16
is a plot of water-oil ratio (WOR) versus recovery factor
showing that 34.5% of the OOIP will be recovered at a
limiting WOR of 30. This is less than the target recovery
factor of 38% and demonstrates the need to inject water into
unswept sands and/or reduce water cycling. Figure 19 is a
plot of WOR versus hydrocarbon pore volumes injected
(HPVI) showing that the limiting WOR of 30 will be reached
at 1.75 HPVI instead of the target value of 1.8 HPVI, again
demonstrating the need to redirect injection. Figure 20 is a plot
of recovery factor versus HPVI, again showing that 34.5% of
OOIP will be recovered at 1.75 HPVI, which is short of the
target value of 38% of OOIP.
After using these plots to identify underperforming patterns
and ranking those with the greatest potential, new wells,
recompletions, and profile modifications can be planned to
increase recovery efficiency. In the case of CRBs 44 and 45,
the pattern surveillance work showed that the best way to
optimize waterflood recovery was to create a series of
injection rows by drilling new injectors, restoring old
injectors, and converting existing producers to injection
(Figure 21). As part of this work, it is critical to determine the
pressures in each subzone and manage injection at this level,
or perhaps at a finer level if individual sand cycles show
significant pressure differences.

SPE 92036

Pattern surveillance in U-P Ford East has proven to be very

useful for validating reservoir parameters, defining improved
recovery potential, and determining well requirements over a
given control volume. However it falls short of quantifying
the benefits of site-specific projects such as optimizing the
location of producers and injectors, which requires numerical
simulation.
Numerical Simulation
Background. Numerical simulation can be used to
supplement analytical techniques to address a variety of
waterflood management issues including identifying infill
locations, quantifying incremental reserves, optimizing
injector placement, and assessing the benefits of reservoir
pressure on throughput and well productivity.
Numerical simulation has been used at the Long Beach Unit
since the mid-1980s when coarse, full-field simulation
models were built to understand aquifer behavior, off-lease oil
migration, and the degree of communication across faults in
the Ranger and Terminal zones. These models gave way to
smaller mechanistic (strip) models and front tracking models
in the 1990s, which were used to assess the benefits of pattern
injection, determine optimum injector/producer spacing, and
assess breakthrough timing. Because the Ranger and Terminal
zones held the greatest potential, earlier modeling efforts were
focused here. However, as these zones have matured,
attention has turned to numerical modeling of the more
complex U-P Ford East as a means to optimize its waterflood.
Two generations of reservoir modeling of the U-P Ford East
have been conducted using a front tracking simulator. This
type of simulator is especially well suited to modeling twophase fluid systems with limited compressibility like
waterfloods. Finely gridded models can be built and run much
more efficiently than conventional finite difference simulators.
The U-P Ford East interval generally meets the criteria for
modeling with a two-phase front tracking simulator, although
the modeler must be mindful that parts of the reservoir were
depleted below saturation pressure, evolving free gas that was
later displaced from the reservoir through waterflooding.
First Generation Model. The first generation of modeling
covered the more prolific eastern portion of the U-P Ford East
with 15 layers representing all of the major U-P Ford East
subzones. The model grid used 100x100 ft cells for a total of
170,000 active cells. This reservoir description was based on
the higher OOIP value from the 1998 description study.
Injection was specified to each layer explicitly to ensure that
historical injection data were honored. However, time
constraints only allowed for global history matching at the
major fault block level of reservoir pressure and water cut.
Significant modifications to relative permeability were
required to match historical water and oil rates. In addition,
model permeabilities were not closely calibrated to available
core and well test data. As such, the model was found useful
for qualitatively predicting sweep, but quantitative analysis of
incremental oil rate and volumes from infill drilling or other
well work was of more limited value.

The large increases in water relative permeability (Krwo)

needed to match history suggested that the effective volume of
oil being waterflooded was likely less than modeled (Figure
22). Modeling sensitivities done with OOIP reduced to the
level estimated prior to the 1998 study (40% reduction) were
able to achieve a similar quality of global history match with
significantly less manipulation of the relative permeability
curves. Detailed comparisons of net pay counts from the 1998
study to core oil sand counts and petrophysical analysis of one
area of the field also suggested that net pay and OOIP might
be lower than estimated in that study. These differences
prompted the most recent reservoir description study3
discussed above and eventually led to construction of a new
reservoir simulation model of the eastern area.
Current Modeling. The new reservoir simulation model is a
work in progress at the time of this writing, with model
construction completed and history matching just beginning.
A finer vertical description has been incorporated into the new
model, based on the 62 stratigraphic cycles mapped in the new
description study. Guided by historical injection profiles,
permeability differences between cycles, and the continuity of
the individual cycle net sand maps, the 62 cycles were
condensed to 26 layers for modeling purposes. Twenty model
layers cover the interval of primary focus (AE-AU subzones)
for the remainder of field life as opposed to eight layers in the
old model for the same interval (Figure 23). Areal grid size
remains similar to the old model resulting in a total of about
320,000 active cells (Figure 24).
Net sand maps from the description work were imported
directly into a geocellular model. Porosity and water saturation
obtained from the recent petrophysical study were used to
populate the model, which was subsequently scaled-up to the
reservoir simulation model. Permeability is based on the
porosity/permeability relationships developed in the
petrophysical model. Relative permeability is taken from a
limited suite of native state relative permeability experiments
and PVT properties are also tied to available experimental
data.
The productive limits in the model area are defined by the
sealing Long Beach Unit Fault to the west, and downdip oil
water contacts to the north and south. The north and south
aquifers are generally of limited activity based on previous
modeling and material balance. The model also includes the
in-place volumes and well production for the adjoining
Belmont Offshore Field that began production in 1954 causing
drainage and pressure depletion.
The finer vertical description should capture the key vertical
sweep characteristics better than the previous model. The
updated OOIP, which also incorporates offset lease volumes
and drainage, should result in a history match with fewer
adjustments to saturation functions or other description
parameters. To history match the new model, injection
volumes will be specified by layer using separate pseudoinjectors for each layer as was done in the first generation

SPE 92036

modeling. Total fluid production by well is specified and

producing water cuts and pressures will be history matched.

compacted core sand counts, core porosities, oil base core

and capillary pressure data.

A combination of static and dynamic well pressure data is

being used to calibrate individual layer permeabilities as a key
history match parameter. Dynamic pressure data includes
wellhead injection pressures and producing bottom hole
pressures. Static pressure data includes periodic static
measurements from a number of pressure monitoring wells
across the field, along with RFT data taken beginning in the
early 1990s. Layer permeabilities will be adjusted to match
actual bottom hole injection pressures with model injection
pressures while honoring producing bottom hole pressures at
the producers as well as static pressure levels. Positive skin
values representing formation damage on injectors as well as
both positive skins (poor completion/workover fluids) and
negative skins (from fracture stimulations) for producers will
be taken into account to the extent feasible.

Successful hydraulic fracture stimulation depends upon

combining the right strategy and execution, as
demonstrated by early success from 1996-1998,
underperformance from 1999-2003, and a recent
resurgence marked by the refinement of basic fracturing
practices and the first frac-and-pack job.

Limited horizontal drilling has not been very successful

due to lower permeabilities, more heterogeneity, greater
formation damage susceptibility, and less pressure
support than in shallower zones of the Long Beach Unit.

Clean injection water, non-damaging workover fluids, and

effective sand controls are keys to maintaining both
productivity and injectivity in active wells.

Although global matching on a fault block level is the initial

goal, the history matching approach requires considerable
review of individual well performance and pressures. Final
adjustment to match producing water cuts will no doubt
involve some adjustment of saturation functions as well. The
ultimate goal of the modeling is to provide forecasts of
waterflood performance that will allow poorly swept areas and
layers to be identified and the benefits of new producers,
injectors, and other well work to be quantified. The model is
also expected to provide insights regarding the performance of
low pressure areas that have resulted from the aggressive
drilling and stimulation program conducted over the past
dozen years, and the impact of increasing pressure on
throughput and well productivity in these areas.

Pattern injection rows are a much more efficient means to

achieve through-put targets than peripheral waterflooding
which fails to provide adequate pressure support to updip
producers.

Improved oil recoveries depend upon increasing injection

into low-pressure sands, re-directing injection into
unswept sands, and aligning patterns to take advantage of
preferred reservoir flow paths.

Waterflood monitoring needs to include a periodic review

of injection profiles and dimensionless performance plots
coupled with pressure data gathering by flow unit
(subzone or sand cycle) and the monitoring of offset well
response.

Numerical simulation using front tracking models can

capture detailed descriptions in full-field models with
reasonable turn-around time.

Lessons learned from a first generation of modeling have

resulted in a more detailed vertical description and more
thorough evaluation of in-place fluids in a secondgeneration simulation model to better capture vertical
profile variations of the thinly-bedded turbidite system.

The improved model description along with the matching

of historical dynamic and static pressures to calibrate
layer permeability-thickness (Kh) are the keys to
quantifying improved conformance and rate profiles from
infill drilling and other waterflood optimization programs.

Conclusions
Low permeability (2-50 millidarcies) turbidites of the U-P
Ford zone have been waterflooded in the Long Beach Unit for
nearly forty years. Improving oil recovery has become more
challenging with time as thinner bedded, more heterogeneous
sandstones are targeted for their remaining oil. Multiple
techniques and technologies have been applied to describe the
reservoir architecture, understand well performance, and
efficiently extract the oil. The successes and failures have
provided valuable lessons into how to best analyze and
optimize waterflooding of the U-P Ford:

Although off-lease drainage, poor data quality, and nonwaterfloodable pay have made it difficult to accurately
estimate OOIP, this number must be accurately quantified
to understand the remaining waterflood potential.

Preferred flow paths in the reservoir can be recognized by

mapping cycles whose flow characteristics are
controlled by stacking patterns, lobe geometries, and the
location of distributaries on the fan surface.

Petrophysical modeling needs to include density log

normalization, resistivity enhancement, and calibration to

SPE 92036

Acknowledgements
The authors would like to thank the following people for
providing their expertise: Geology: Jack Klotz, Scott Prior,
Tom Eggert, and Melissa Morse; Petrophysics: Kevin Crook
and Eric Prasse; Reservoir Engineering: Evan Rael, Scott
Weimer, and Keith Lynch; Operations: Eric Upchurch, Greg
Colazas, and Carl Montgomery; Drilling: Keith Turnage and
Mike McCarter. We would also like to thank Jeff Johnston,
John Bolling, Scott Prior and Don Clarke for their critical
review of the manuscript. Finally, the authors would like to
thank Occidental Petroleum, THUMS Long Beach Company
and the Department of Oil Properties in the City of Long
Beach for permission to publish this paper.
References
1.

D. D. Clarke, Structure of the Wilmington Oil Field, in:

D.D. Clarke and C.P. Henderson, eds., Geologic Field
Guide to the Long Beach Area, Pacific Section of
American Assoc. of Petroleum Geologists, Guidebook 58,
1987, 43-55.
2. T. Redin, Oil and Gas Production from Submarine Fans of
the Los Angeles Basin, Active Basin Margins, AAPG
Memoir 52, 1991, 239-259.
3. J. Klotz, Reservoir Geology of the Miocene Union Pacific
and Ford Reservoirs, Long Beach Unit, Wilmington Field,
California, Occidental Oil and Gas Corporation,
Production Geoscience internal report, 2004, 28 p.
4. B. H. Berman, The Origin and Habitat of Sedimentary
Carbonates from the Miocene Puente and Pliocene Repetto
Formations in the Long Beach Unit, Wilmington Field,
California, THUMS Long Beach Company internal report,
1995, 60 p.
5. C. L. Vavra, Petrographic Evaluation of Long Beach Unit
Samples, ARCO Exploration and Production Geology,
Reservoir Engineering Analysis Group, internal
correspondence, 1992.
6. O. A. Alpay and E. R. Green, Description of Reservoir
Characteristics, Union Pacific-Ford Zone, Long Beach
Unit, THUMS Long Beach Company Internal Report, July
24, 1970.
7. G.E. Otott, Long Beach Unit 3D Survey, Update of
Activity, February 1996 through September 1999,
THUMS Long Beach Company internal report.
8. P. Lowry, C. D. Jenkins, and D. J. Phelps, Reservoir Scale
Sandbody Architecture of Pliocene Turbidite Sequences,
Long Beach Unit, Wilmington Field, California, SPE
26440, Presented at the 1993 Annual Technical Conference
and Exhibition, Houston, Texas (October 3-6, 1993).
9. T. Dang, Long Beach Unit Log Analysis: Porosity and
Water Saturation Models, ARCO Exploration and
Production Geology Technical Service Report TSR-910066, 1991.
10. B. G. Gobran and T. Dang, Long Beach Unit Original Oil
in Place Determination, ARCO Exploration and
Production Geology Technical Service Report TSR-940036, 1994.
11. B. G. Gobran and C. D. Jenkins, Calibration of the U-P
Ford Petrophysical Model to Oil-base Core and Capillary
Pressure Data, ARCO Exploration and Production Geology
Technical Service Report TSR-95-0016, 1995.
12. I. Juhsz, Normalized Qv - The Key to Shaly Sand
Evaluation using Waxman-Smits Equation in the Absence
of Core Data, Paper Z, presented at the Society of
Professional Well Log Analysts 22nd Annual Logging
Symposium, June, 1981.

13. S. Al-Sharif and E. Rael, Rate Forecasting of a Mature

Waterflood Using Fractional Flow Theory and Hyperbolic
Decline Curve Analysis, East Wilmington, Long Beach,
California, SPE 83502, Presented at the 2003 SPE
Western Regional Meeting, Long Beach, California (May
19-24, 2003).
14. S.E. Buckley and M.C. Leverett, Mechanism of Fluid
Displacement in Sands, Transcript AIME (1942), 146,
107
15. F.F. Craig, The Reservoir Engineering Aspects of
Waterflooding, Monograph 3, Society of Petroleum
Engineers, Dallas, TX, 1971
16. C.R. Smith, Mechanics of Secondary Oil Recovery.
Huntington, NY: Robert E. Krieger Publishing, 1966
17. G.P Willhite, Waterflooding. Dallas. TX: Society of
Petroleum Engineers, 1986.
18. J.A. Robertson, J.A. Blesener, and S. Soo Hoo, Subzone
Redevelopment of the Long Beach Unit, Wilmington Oil
Field: A Case Study,
SPE Journal of Petroleum
Technology, October 1987, p 1229.
19. G.S. Woodling, P.J. Taylor, H.H. Sun, Q.N. Nguyen, and
T.R. Brix, Layered Waterflood Surveillance in a Mature
Field: The Long Beach Unit, SPE 26082, Presented at the
1993 SPE Western Regional Meeting, Anchorage, Alaska
(May 26-28, 1993).
20. S. Franks, D. Medwedeff, S. Raeuchle, and B. VerWest,
Additional Petroleum Potential and Hydrocarbon Phase
Analysis of the Long Beach Unit and Adjacent Area,
ARCO Exploration and Production Geology Technical
Service Report TSR-95-0056, 1995.
21. S. Prior and G.E. Otott, Long Beach Unit Seismic
Prospects, Plays, Leads and Concepts, THUMS Long
Beach Company Internal Report, September 8, 1999.

10

SPE 92036

NE

City of Long Beach

-48
00
-46
00
-44
00

SW

1 MILE

0
20
-4

Fault
Block 8

LBU Leaseline

UP-Ford 98
-4
20
0

-4400

-480
0
-5000

Ranger

- 4000

-5 -54
20 00
0

0
00
-5 00
8
-4 600
-4

-4600

Tar

Wells
0
40
-4

- 2000

-4400

Terminal

- 6000

Drilling Island
Long Beach
Unit Fault

UP/Ford

237 Shale

- 8000

-4600
-4800

Belmont

C-608 Fault

Offshore
-520
0
Field

1 Mile

Figure 1: Schematic NE-SW cross-section through one of the drilling

islands in the Long Beach Unit showing turbidite sandstone reservoirs
(Tar, Ranger, Terminal, U-P Ford zones) and a fractured shale
reservoir (237 Shale) above the basement.

Figure 4: Structure map on the top of the U-P Ford zone (top of AE
subzone) showing the southeastward-plunging Wilmington Anticline
and the major faults that compartmentalize it. The waterflooded U-P
Ford East area is located east of the Long Beach Unit Fault.

GR Curve
6400
Coarsening and
Thickening Upward

Palos
Verdes
Fault
Zone

Newport-Inglewood
Fault Zone
Wilmington
Field

Los Angeles Basin

LBU

THUMS-Huntington
Beach Fault Zone

Oil-Stained
Sandstone

Coarsening and
Thickening Upward
Fining and
Thinning Upward

Huntington
Beach Field

Carbonate
Cemented
Sandstone

Siltstone
6450

Blocky

Fining and
Thinning Upward

Belmont
Offshore

Amalgamated
Sandstones
6500

Blocky

Figure 2: Location map of the Long Beach Unit (LBU) in the East
Wilmington Field. Belmont Offshore Field adjoins the LBU to the
southeast.
SW

Rev
ers
e Fa
ult

C60
8 Fa
ult

Well GR curve

Approx.
UP-Ford
STILL WAITING ON THIS FIGURE
Interval
BA

First Order Shale

t
aul
9F
C53

LBU
Fau
lt

NE

Figure 5: Core description plot from the C-518 well showing the
lithologies, stacking patterns, and GR curve for portions of the AR and
AU subzones (top of the AU subzone = 6450 feet)

Second Order Shales

BA

BS
BS

THUM
S
Beach -Huntingto
n
Fault

Figure 3: Interpreted seismic line showing the southwestward directed

THUMS-Huntington Beach Thrust Fault. The Wilmington Anticline is
carried in the hanging wall of this fault and contains several faults that
20
compartmentalize the U-P Ford reservoirs. BA = top of the 237 Shale,
BS = top of basement schist.

First Order Shale

Figure 6: Cross-section showing the geometry of first and second

order shales in turbidite sand lobes. First-order shales form thick,
laterally-extensive barriers to vertical fluid movement. Second-order
(off-lapping) shales are thinner, more discontinuous, and form baffles
to fluid movement (from Phil Lowry, personal communication).

11

16
4400

4600

4800

5000

30

5200

5400

5600

5800

6000

HAFWL
DEPTH
ft
1000
FT
0.5

-80

SPBL
mV

20

GRCE
api

150

SN
OHMM

50 400

0.5

ILD
ohmm

50 1

0.5

ILDRT
RtMod

50 1.7

NEUT
API
VEL
f/ms

1400

SSD
FT

Capillary Pressure (psi)

4900

15

5000

CSROC
0 v/v 1

CPHIC
0.4 v/v 0

CSRWC

BVWT

16

1
4600

4700

4800

4900

5100

10

5000

5200

5100

5300
5200

0
50

60

70

80

90

100

7200

AY
13.5

7400

SW
v/v

VSHLIMIT
v/v

VSH
dec

SHUT
1 0 - 6

VSH
dec

PHIT

0.4

PERF
0 - 6

0 ERF
P

AE

AF

5400

Water Saturation (%)

Figure 8: Graph of capillary pressure versus water saturation for core

samples from well C-657. Samples are from the AU-AV subzones of
the U-P Ford. Permeability values are uncompacted.

7000

CKLC

4.4 1.7 0.9 1.4 md

20

40

6800

mD 0.1 1 v/v 0 0.4

1000

RHOB_NRM
g/c3
2.7

4800

30

6600

KL_PHIE
1000 mD 0.1

25

13

6400

Figure 10: Graph of compaction-corrected core porosity vs. TVDSS

depth for core samples from U-P Ford wells. Values from the shallow
AE subzone are on the left, the deep AY subzone is on the right.

4700

469 705 493 101 102 359 177 83 70

6200

True vertical subsea depth (feet)

35

AW
11
AM
6
AK1
3.5
AE
1

Y=0.2599-X*0.000013

4200

Base of UP/Ford Zone

Figure 7: Interpreted 3D seismic line showing a narrow horst block

between the Long Beach Unit Fault and the D-630 Fault. Potentially
significant amplitude anomalies appear in the AR, AU and AU2
21
subzones of the U-P Ford.

AU
8.5

0.3
0.25
0.2
0.15
0.1
0.05

AU2

B-339I
B-501
C-323
C-336
C-424
C-518
C-539
C-602
C-603
C-618
C-657
C-736S
D-101
D-102
D-104
D-113
D-308
D-600
D-703
D-825A

t
ul
Fa

Compaction-corrected core porosity (fraction)

ro
pe
ni
Ju

AU
AO

0.35

SPE 92036

AI

AK1
AK1-4

AL1

5300

Figure 11: Example plot showing raw log curves, core data, and
calculated log curves from a U-P Ford well analyzed using the new
petrophysical model.

Sw from
ILD curve
10,000

Shale
Volume

BPD
effective
from logs

1,
000
1,000

oil rate
injection rate

gross rate

1970

Recomplete High oil Low oil In-zone Hydraulic Low oil

into Upper prices prices injection fracture
prices
stimulation
added
Ford

1975

1980

1985

1990

1995

2000

Sw from
inverted ILD
curve

Figure 9: Log plot showing the results of applying an inversion

algorithm to improve the vertical resolution of the induction log for thin
9
bed detection in the U-P Ford.

Figure 12: Production history of the U-P Ford zone showing a peak
rate of 30,000 BOPD in 1969 and events since then that have affected
the production rate. The current rate is about 3,500 BOPD and 36,500
BWPD.

12

SPE 92036

100

.7
5

1.00

.25

1.5
0

5
.2

.50

.25

Long Beach
Unit Fault

.50

.50

0
1.0

1.

0
1.5

.25

5
.2

.50

00

.50

Expected
Ultimate
UtEr~0.345@30WOR
Recovery = 34.5%
at WOR = 30

.25

0
1.5

.50

.25

.25

1.0
0

.50

1.01.50
0

WOR

1.00

.75

.25

10

1.00

.25
.50

.75

1
5
.2

.50

.50

.25

.5
0

.2

.75

.25
5
.2

.2
5

.5
0

C-608 Fault

0.1
-

0.05

0.10

0.15

0.20

Belmont
Offshore

Long Beach Unit

0.25

0.30

0.35

0.40

0.45

0.50

Recovery factor

Figure 13: Contour map of production index (PI) values in U-P Ford
wells.

Figure 16: Semi-log plot of WOR versus recovery factor for 5 pattern
elements in cut-recovery blocks 44 and 45 of the U-P Ford East
waterflood.

0.007

5000

+
SPI, B/D/FT-PSI

0.006

BOPD

4000

3000

0.005
0.004

Oil + Water

0.003

+
0.002

2000
29 existing wells: oil rate increased from 42 to 153
BOPD with a 10% decrease in water cut

1000

++ + ++
+ + +
+

21 new wells: average initial production = 158 BOPD at

75% water cut

10

20

0
Jan 98

Jan 97

Apr 96

Jan 99

Figure 14: Plot of cumulative oil rate added through hydraulic

fracturing versus time showing the results of 100 stages pumped in 21
new wells and 29 existing wells of the U-P Ford zone.

80

90

100
D565

P1

D587A

D202A
D806
D734

D630Z
D104A
D212A

C539A

C418A

D627A
C418A
C608A
C533A

70

Figure 17: Plot of specific productivity index versus water cut derived
from relative permeability data (lines) and compared to actual well
performance in the U-P Ford zone.

Injector

P2

First UP-Ford
Frac and Pack

C539A

D707

40
50
60
Water Cut, %

P3

C440

D108A

D738

A415

C432A

C403A

C530B

BOPD
500

30

D566 C403 C446 D202A C648 D314A C603B C626A C646 + D819

1,500

1,000

Oil

Water

0.001

P4

2001

2002

2003

P5

2004

Producer

Figure 15: Plot of cumulative oil rate added by U-P Ford zone
hydraulic fracture stimulations (stippled area), risked expected
cumulative oil rate for the period 2001-2004 (solid line), and the names
of stimulated wells.

Figure 18: Pattern elements in cut-recovery blocks (CRBs) 44 and 45

of U-P Ford East showing the injection wells that have historically
provided waterflood support.

SPE 92036

13

45

0.9
Kr
Relative Permeability

40
35

WOR

30
25
20
Forecast

15
10
History

Model pseudo Krwo

0.8
0.7
0.6
0.5

Rock Krwo

0.4
0.3
0.2
0.1

0
0

0.0
0

0.4

0.6

Sw norm
Normalized

HPVI

Figure 19: Plot of WOR vs. HPVI for five pattern elements in cutrecovery blocks 44 and 45 of the U-P Ford East waterflood showing
that a limiting WOR value of 30 will be reached at 1.75 HPVI.

0.2

0.8

1.0

Sw

Figure 22: Plot of relative permeability versus normalized Sw for the

first generation U-P Ford East simulation model showing the large
increase in pseudo Krwo required to obtain a history match.

-4000

0.40
0.35

-5000

Forecast

AE

Recovery Factor

0.30
History

-6000

0.25

AU
-7000

0.20
0.15

-4000

0.10
-5000

0.05

AE
-6000

0.00
0

HPVI

-7000

Figure 20: Plot of recovery factor versus HPVI for five pattern
elements in cut-recovery blocks 44 and 45 of the U-P Ford East
waterflood showing that 34.5% of OOIP will be recovered.

-5
0
00

-47

50

AU

Figure 23: Comparison of model layering in the first generation model

(top) and the 2004 model (bottom).

Maintenance Well Work

Restore Jobs
Convert to Injection

0
55
-5

0
70
-5

New injector (drilling)

-52
50

-4750

-5000
-5250
-5500
-5750

-5250

-6000

Figure 21: Structure map on the top AK1 subzone showing the
planned injectors and work requirements in cut-recovery blocks 44 and
45 of the U-P Ford East waterflood.

Figure 24: Second generation simulation model for U-P Ford East with
26 layers and 320,000 active cells.
.