SPE 91069
Oil and Gas Reserves Estimates: Recurring Mistakes and Errors
D. Ronald Harrell SPE, John E. Hodgin SPE, Thomas Wagenhofer SPE, Ryder Scott Company, L.P., Houston, Texas
Copyright 2004, Society of Petroleum Engineers Inc.
This paper was prepared for presentation at the SPE Annual Technical Conference and
Exhibition held in Houston, Texas, U.S.A., 26–29 September 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
Although most petroleum reserves estimates are prepared by
qualified reservoir engineers and geoscientists, there continue
to be recurring mistakes and errors in estimates of reserves.
This statement is not intended to be presumptuous in any
regard, but is based upon observations made by the authors
over many years as independent petroleum consultants for
both publicly-owned and private exploration and production
companies.
This paper presents and discusses the most common
mistakes and errors associated with reserves estimates. In
addition, guidelines are offered that, if followed, may help
reduce the chance of such errors occurring.
It is the intent of this paper to focus on mistakes made in
the preparation of reserves estimates rather than to discuss
potentially incorrect applications of petroleum reserves
definitions. It is clearly not the intent of this paper to interfere
with the necessary judgment the reserves evaluator must
employ in preparing reserves estimates.
It is inevitable that adherence to applicable reserves
definitions must be recognized in any discussion of
recommended practices regarding the estimation of petroleum
reserves. A generalized definition of reserves without regard
to categories or authorship is: Reserves are estimated
remaining quantities of hydrocarbons to be commercially
produced from a known accumulation as of a given date under
specified definitions and economic conditions.
Introduction
The concerns expressed in this paper have been developed
over time by the authors and have been subdivided into
geoscience and engineering categories. Without full mutual
understanding and competent integration of the work products
of these similar, yet unique, professions, any resulting estimate
of in-place hydrocarbons and ultimate reserves will be suspect.
Since the geoscience component forms the basis for
engineering estimates, these concerns are presented first.
Geoscience-based Recurring Mistakes and Errors
This section discusses some of the most common geosciencebased errors associated with reserves estimation as observed
and confirmed over time by the authors of this paper. In
summary, errors are most commonly associated with the
following issues (each topic will be discussed in detail below):
• Mapping surfaces – top and base of contributing
reservoir unit
• Downdip limits – vertically stratified reservoirs
• Isopachous maps
• Attic volumes
Mapping surfaces - top and base of contributing reservoir
unit. Structure maps on the top and base of a reservoir unit
are among the most common maps constructed by
geoscientists. Maps on these surfaces used in conjunction
with lateral limits from structural or stratigraphic barriers and
downdip fluid limits describe the productive area of a
reservoir.
The selection of correct mapping surfaces
corresponding to the top and base of the contributing reservoir
unit is critical in determining an accurate volumetric estimate.
Structure on top surface. Structure maps tied to markers
(subsurface or seismic) or the top of a formation which do not
represent the top of the contributing or effective reservoir unit
may result in overstating the productive area, associated
volume and reserves.
Figure 1 illustrates the potential overstatement of reserves
(crosshatched areas) from using a marker picked from well
data to represent the structural surface at the top of the
reservoir unit. The illustration shows a difference of 50 feet
in the elevation between the structure map using the
subsurface marker top at -7000 feet and the top of the
contributing or effective reservoir at -7050 feet.
This distance translates into an exaggeration of the areal
extent based on the projection of the downdip limit to the top
surface of the reservoir. The significance of the discrepancy is
related to the magnitude in the distance between mapping
points and the structural dip.
2
SPE 91069
The intersection of the fluid contacts on the top and basal
surface determines the productive area of the reservoir.
Figure 3 demonstrates the potential overstatement
(crosshatched area) of the productive gross rock volume
using a marker picked from well data to represent the
structural surface at the base of the reservoir unit. In this
illustration, the gross interval appears to thicken in the updip
direction. The discrepancy can become significant as the
magnitude in the distance between the mapping points
increases.
Figure 3. Potential overstatement due to selecting marker as base
of formation instead of base of effective pay.
Figure 1. Structure on top surface mapping error using marker
instead of top of effective pay.
Similarly, Figure 2 illustrates the selection of a mapping
top corresponding to the top of the formation rather than the
top of the contributing or effective reservoir section. As in the
prior example, the selection of a correlative mapping point
results in a similar exaggeration of the areal extent and
potentially overstates the reserves attributable to the reservoir.
Figure 4 illustrates the error that may be introduced in the
determination of the inner limit of water by inaccurately
selecting the base of the contributing or effective reservoir
unit. In this illustration, the volume within the wedge area is
overstated and the volume above the inner limit of water is
understated.
This example illustrates the potential
understatement of reserves. This discrepancy can become
significant as the magnitude in the distance between the
mapping points increases. A decrease in structural dip would
further compound this problem.
Figure 2. Structure on top surface mapping error using top of
formation instead of top of effective pay.
These errors are also replicated when the top of a seismic
event is not adjusted to tie with the top of the contributing or
effective reservoir unit from subsurface well data.
Structure on basal surface. Structure maps tied to
markers (subsurface or seismic) or the base of a formation
which does not represent the base of the contributing or
effective reservoir unit may result in overstating the gross rock
volume and inaccuracy in the determination of the inner limit
of fluids used in constructing net pay isopachous maps. Most
computer aided mapping relies on the calculation of the gross
rock volume generated by the difference between structural
surfaces or maps on the top and basal surface of the reservoir.
Figure 4. Potential error of “inner water limit” as the result of
incorrectly picking base of effective pay.
These errors are also replicated when the base of a seismic
event is not adjusted to tie with the base of the contributing or
effective reservoir unit from subsurface well data.
Position of faults relative to the structure on top surface.
Faults which have not been adjusted to tie the structure map
on the top surface of the contributing or effective reservoir
unit may result in either the potential overstatement or
understatement of the productive area, associated volume and
reserves.
SPE 91069
Figure 5 demonstrates the potential error in relating
position of the updip trapping fault to the top of structure
based on a marker rather than the top of the effective
reservoir.
3
The diagrammatic cross section shown in Figure 6
illustrates the difference between the productive volume
attributable to the three zones with a common downdip contact
(lowest known hydrocarbon or LKH) represented by the cross
hatched area as compared to the shaded areas in each zone
with a separate LKH. “Most reserves definitions would
require the assumption of three reservoirs with separate
downdip limits for the attribution of proved reserves absent
any confirming data to contradict this assumption.”1
Isopachous maps. The estimation of volumetric reserves
depends on three main types of isopachous maps: (i) a map of
the gross thickness of reservoir unit, (ii) a map of the net
effective thickness generally based on the application of a
minimum porosity cut-off value, and (iii) a map of the net
effective pay thickness generally based on the application of a
maximum saturation cut-off value.
The construction of accurate net pay isopachous maps,
whether drawn by hand or generated with the aid of mapping
software, must adhere to the same basic geometric concepts
introduced by Wharton2 and Tearpock and Bischke3. Not
adhering to these principles introduces various errors that may
result in potential miscalculation of the productive reservoir
volume.
Figure 5. Potential error in picking fault location due to incorrectly
selecting marker as top of structure.
Once again, factors such as the magnitude in the distance
between mapping points, the structural dip and thickness of
the reservoir unit and the dip on the fault plane determine the
significance of the error.
Figure 7 illustrates the two main regions of a net pay
isopach map; the wedge zone and the area of maximum fillup, according to the Wharton method.
Downdip limits - vertically stratified reservoirs. The downdip extent of a productive area is defined by fluid contacts and
lateral limits from structural or stratigraphic barriers. An
overestimation of in-place volumes is the likely result when
assuming vertical communication and a common downdip
contact in stratified or layered reservoirs without adequate
support from pressure data.
Figure 6 illustrates a log section marked to indicate three
porous intervals shown as A, B, and C. “These three zones,
all assumed to be capable of commercial production, may be
part of a single pressure-connected reservoir or may
collectively comprise three separate reservoirs.
This
uncertainty may not be resolvable with well data alone.”1
Figure 6. Potential error of assuming a common downdip contact
in stratified reservoirs.
Figure 7. Illustration of wedge zone and area of maximum fill-up.
Net pay isopach maps - downdip wedge zone. The wedge
zone as illustrated in Figure 7 is the interval between the
intersection of the fluid contact and the structure on the top
and base of the contributing or effective reservoir unit. The
correct placement of contours representing net pay thickness
in the wedge zone is governed by the rate of structural gain
above the elevation of the downdip fluid contact and the
vertical distribution of the net pay.
A common technique in both hand drawn and computer
aided mapping involves the use of a net-to-gross ratio to
represent the change in vertical net pay proportionate to the
change in elevation above the intersection of the fluid contact
and the structure on the top of the effective reservoir unit. A
4
SPE 91069
net-to-gross ratio based on the net pay thickness of the
reservoir unit to the gross thickness of the reservoir unit
represents an average distribution for the entire interval. In
zones where the vertical distribution of net pay is fairly
constant, the average net-to-gross ratio may be a fair
representation. However, the use of an average net-to-gross
ratio in reservoirs where the distribution of net pay varies over
the vertical interval is unlikely to be correct and may lead to
either overstating or understating the reservoir volume and
associated reserves.
The net-to-gross ratio for the well illustrated in Figure 8 is
0.50; however, most of the net pay occurs in the upper 20 feet
of the 80 foot gross interval.
Figure 10. Illustration of wedge zone based on correct application
of a vertical net-to-gross ratio.
A similar but inverse error would occur if the vertical net
pay distribution was inverted from that shown in Figure 8. In
this case, a map constructed using the average net-to-gross
ratio would overstate the productive reservoir volume.
Figure 8. Sample wedge calculation with most of net-pay at top of
reservoir.
Figure 9 illustrates a net pay isopach map constructed
using the average net-to-gross ratio of 0.50 in the wedge zone.
Figure 9. Illustration of wedge zone using average net-to-gross
approach.
Figure 10 illustrates a net pay isopach map constructed
using the actual relationship of net pay thickness to height
above the downdip fluid contact. In this example, the net pay
isopach volume in Figure 9 is 18 percent smaller than the
correctly drawn map from Figure 10.
In both examples, the “average” net-to-gross approach
would result in mechanically equal spaced thickness contours
in the wedge zone which would not be representative of the
actual vertical distribution of the net pay in the well.
Net pay isopach maps - area of maximum fill-up. The
area of maximum fill-up as illustrated in Figure 7 is the
interval above the intersection of the fluid contact and the
structure on the base of the effective reservoir unit. Above
this inner limit of fluid, the placement of net pay thickness
contours is governed by the lateral change in the net effective
reservoir thickness.
Most computer-aided mapping relies on the calculation of
the gross rock volume generated by the difference between
structural surfaces or maps on the top and basal surface of the
reservoir. The net pay thickness is generated by applying a
net-to-gross ratio to the gross rock volume. A few of the
potential inherent errors are as follows:
• Use of an arithmetic average of the net-to-gross ratio
from multiple well penetrations may not be
representative of the lateral variation from well to
well. A more rigorous approach is to represent the
lateral variation by contouring the net-to-gross ratio
from actual well data. The resulting interpolated
distribution of net pay thickness should either tie or
be adjusted to match the actual well data points.
Consideration should be given to the validity of
estimates of interpolated net-to-gross ratios greater
than the maximum value obtained from well data.
• As previously noted, errors in the selection of the top
and/or base of the contributing or effective reservoir
unit will result in overestimating the gross interval
thickness and gross rock volume. The interpolated
SPE 91069
5
lateral distribution of the gross reservoir thickness
should either tie to or be adjusted to match the actual
well data points. In those cases where the top and/or
base of the reservoir unit are based on seismic data,
consideration must be given to the quality and
resolution of the seismic data. Consideration should
also be given to the validity of estimates of
interpolated gross reservoir thickness greater than the
maximum value obtained from actual well data.
Similarly, consideration should be given to the
validity of lateral variations in interpolated net pay
thickness derived from uncalibrated seismic
amplitudes that result in values greater than indicated
by the actual well data points.
the following processes (each topic will be discussed in detail
below):
• Production decline curves
• Operating costs
• Inappropriate selection of analogs
• Simulation-derived estimates of proved reserves
• Failure to incorporate early life performance data into
volumetric estimates
• Impact of partial waterdrive and over-pressured
reservoirs on gas material balance
• Assignment of proved reserves to undrilled fault
blocks
• Incomplete understanding of commercial economics
projection software
Attic volumes. In many cases reserves are assigned to
volumes updip to the last well penetration point in a reservoir.
The level of confidence in the structural and stratigraphic
continuity of the reservoir and the recognition of the
appropriate drive mechanism are critical to the correct
attribution of reserves.
Production decline curves. Performance decline analysis is
probably the most common technique to estimate reserves in
mature assets where ample performance data is available for
both primary and secondary products. Besides the obvious
subjectivity in determining a decline trend, common errors are
associated with: (i) composite field decline curves, and (ii)
neglecting to apply a minimum hyperbolic decline rate.
•
Figure 11 illustrates a case where the net pay thickness is
projected in association with structural gain only and exceeds
the maximum net effective sand thickness updip to the wedge
zone from the downdip well penetration.
Composite field production decline curves. Quite often,
an engineer has available only production history for a multiwell lease, a production unit, a single reservoir, or sometimes
an entire field. Individual well production history may not be
available or can be compiled only through the use of
allocations relying upon less-than-perfect well tests. When
aggregate well production history is displayed as a graphical
presentation of monthly oil or gas production, the historical
trend may show a continual decline over time. Indeed, this
trend may be well-defined as an exponential or hyperbolic
decline that can be projected into the future with a reasonably
high degree of reliance based upon the mathematical “best fit”
of the historical data. This is illustrated as Figure 12.
Example Decline Curve
Field Aggregrate Forecast
Figure 11. Potential error in estimating attic volumes based on
projecting structural gain greater than maximum sand thickness.
Though not necessarily a geoscience issue, one must also
give consideration to the possibility of a gas-saturated attic
above a highest-known oil limit.
Oil Production Rate (STB/day)
The level of confidence in the position of faults and
stratigraphic conditions away from the existing subsurface
well control must be considered. Seismic fault placement
should be collaborated by subsurface well control.
Stratigraphic continuity verified in zones above and/or below
the interval in question increases the level of confidence for
the attribution of reserves.
10,000
1,000
100
10
1980
1985
1990
1995
2000
2005
2010
2015
2020
2025
2030
2035
2040
Figure 12. Field aggregate forecast based on apparent trend.
Frequent Reservoir Engineering-based Recurring
Mistakes and Errors
This section discusses the most common reservoir
engineering-based errors associated with reserves estimation
as observed and confirmed over time by the authors of this
paper. In summary, errors are most commonly associated with
This projection clearly presents an appealing case for using
the entire production history to obtain an estimate of proved
reserves as of the effective date of the estimate. Such a
decline projection may be acceptable, however, only if (i) the
well count is relatively stable, (ii) production conditions and
6
SPE 91069
Example Decline Curve
Field Aggregrate Forecast w/ Well Count
50
10,000
40
Example Decline Curve
Forecast After Removing Effects of Drilling and Workovers
50
10,000
45
40
35
1,000
30
Well Count
25
20
Forecast after accounting for drilling and
single-event workovers.
100
5
35
1,000
30
25
20
Well Count
100
15
10
Average Well Performance.
Well Count
Oil Production Rate (STB/day)
45
Figure 15 presents a final forecast in which both effects of
drilling and single-event workovers have been removed from
the field trend. The final projection may yet overstate
remaining reserves unless one can be assured that
opportunities for these types of production enhancement
remain (e.g. limited number of re-completions, stimulation
treatments).
Well Count
Figure 13 is the same as Figure 12 but contains additional
data about the number of producing wells over the productive
life of the field. This additional data is often overlooked but
has a significant impact on the previous interpretation of
remaining proved reserves. It should be clear that the forecast
shown on Figure 12 is not achievable without the continual
drilling of additional wells at the same frequency and with
similar results, a condition that is highly unlikely in most
cases. This erroneous approach has been used by many
estimators in the past to estimate proved producing reserves,
in some cases with the mistake compounded by the addition of
yet even more proved undeveloped reserves assigned to
discrete drilling locations.
The average well production (field production devided by
well count) may have been sustained by the continuing impact
of production from new wells over time and well-maintenance
work.
Oil Production Rate (STB/day)
methods are largely unchanged over the producing life, and
(iii) wellbore intervention and other remedial work can be
classified solely as maintenance. If these rather stringent
conditions are not met, reliance upon this projection to
estimate proved reserves may be inappropriate.
0
10
1980
1985
1990
1995
2000
2005
2010
2015
2020
2025
2030
2035
2040
Figure 15. Alternate forecast after removing effects of drilling and
single-event well-maintenance work.
15
The preferred approach is to rely upon the performance of
individual wells whenever possible. Any other approach may
lead to an optimistic estimate of future performance and
proved reserves.
10
5
0
10
1980
1985
1990
1995
2000
2005
2010
2015
2020
2025
2030
2035
2040
Figure 13. Field forecast based on apparent trend with well count.
Even with the preparation of a projection such as that
shown in Figure 14, which is a restatement of the data
displayed on Figures 12 and 13 based on average monthly
production per well, one should be cautious when using
“average well” projections.
Example Decline Curve
Alternate Field Aggregrate Forecast w/ Constant Well Count
50
10,000
40
35
1,000
30
Well Count
25
Forecast based on constant well count
and average well performance.
100
20
Well Count
Oil Production Rate (STB/day)
45
15
10
Average Well Performance.
5
0
10
1980
1985
1990
1995
2000
2005
2010
2015
2020
2025
2030
2035
2040
Figure 14. Alternate forecast based on constant well count and
average well performance.
Failure to specify minimum decline rates in hyperbolic
projections. Virtually all commercial software programs
designed to assist in the forecast of future production rates and
the related cash flow projections provide an option to use a
hyperbolic projection with a specified N-factor and the final
decline rate. This N-factor can also be “calculated” by using
the curve-fitting section of the economic software package.
A decision to allow the software to default to the use of an
unspecified final decline rate may have little effect on present
value of the entity being evaluated but the additional reserves
“added” through an often unreasonable and unsupportable
final decline rate can become a serious overestimation in many
cases. A review of depleted or nearly depleted area analogs
can often be found to provide guidance in the selection of an
appropriate final decline rate.
Other errors associated with decline curve analysis.
• Ultimate recovery not related to volumetric estimates.
Apparent decline trends combined with relatively flat
flowing tubing pressures can lead to optimistic
reserves estimates, particularly in gas reservoirs with
partial to strong waterdrives.
SPE 91069
•
•
Assumption of exponential decline in reservoirs that
have a tendency to exhibit hyperbolic decline trends
(source of underestimating reserves). These include
(i) tight gas reservoirs (enhanced if multiple layers),
(ii) naturally fractured reservoirs, and (iii) waterflood
reservoirs.
Conversely, the assumption of a hyperbolic decline
may lead to overstating reserves in cases where an
exponential decline would also fit performance.
Guidelines to reduce mistakes in decline curve analysis.
• Always attempt to estimate performance decline on a
well or completion level for best results.
• Include trends in secondary products (condensate
yields, gas-oil-ratios, water-cuts) in analysis.
• When projecting group or field level rates, make sure
to review the components of the field curve and
properly account for well work and associated cost
that is required to maintain the decline trend. If well
work cannot be sustained, the field curve needs to be
adjusted to fit the true decline of existing wells.
• Use analogous fields or more mature wells in the
field or area to establish typical decline behavior,
including minimum hyperbolic decline rates.
• In addition, gain understanding of reservoir
properties (e.g. porosity, permeability, lithology,
depositional environment) in order to exercise better
judgment in selecting exponential vs. hyperbolic
decline models.
• Attempt to combine various types of evaluation
techniques with decline curve analysis to assure
consistency in results.
Operating costs. Operating costs reflect expenses that are
attributable to the daily operations of the field and typically do
not include G&A expenses or other overhead costs. Operating
costs are used to capture expenses (affects reserves value), and
to estimate an economic limit (affects reserves volume). The
economic limit is defined as the rate and time when revenue of
production becomes less than the cost of operations.
Typical errors or mistakes associated with operating costs
include the following and are discussed in more detail below:
(i) use of forecasted or budgeted operating costs that are lower
than actual long-term historic costs, (ii) recurring well or
facility costs that are assumed to be single events and
therefore excluded from future estimates of cost, (iii)
assumption of per unit cost of primary product (dollars/barrel
for example) without the proper treatment of fixed cost or
costs of producing secondary products, and (iv) failure to
evaluate changes to costs due to the introduction of new
recovery mechanisms.
Projected operating costs are lower than historic average
costs. Occasionally, forecasted or budgeted operating costs
that are lower than average historic costs are used to estimate
reserves. This may be based on an assumption rather than
established fact. This approach will, in most cases, result in
overstating both income and reserves. In general, regulatory
7
bodies require that operating costs be closely tied to at least
one if not several years of observed costs. Any deviation will
require sufficient evidence describing circumstances and
events that will result in lower future operating costs.
Recurring well or facility expenses. Most reservoir
engineers rely upon historic facility, lease, and/or well
operating cost statements as the basis for calculating historic
operating costs, typically expressed as a monthly cost, for
mature properties. This may further be subdivided into fixed
and variable components when appropriate. Historical costs
frequently will include costs that are deemed to be “nonrecurring” and these costs are typically excluded from average
costs for use in production forecasts. This approach is
acceptable only if the “non-recurring” costs are indeed nonrecurring. Altogether too often, such items as tubing repairs
and/or replacement or periodic platform or facility
maintenance, are deducted as being non-recurring. The failure
to recognize the periodic frequency of such maintenance can
lead to an overstatement of reserves and future net income.
Assumption of per unit operating cost. Alternatively, and
perhaps more serious, some evaluators will elect to use a
future operating cost expressed as a unit cost per volume
(barrel, mcf, or cubic meter), without a proper fixed cost
portion or proper inclusion of secondary products, based upon
their estimate of some current or past analog. This approach is
virtually never acceptable as unit costs of production almost
universally increase over time with declining production even
if the total monthly or annual costs remain constant or slightly
decline.
This increase in unit costs of production is
exacerbated by increasing needs for compression and artificial
lift and a continuing growth in maintenance related to
corrosion, equipment repairs, water treatment, and disposal
and
ever-expanding
environmental
concerns.
An
understatement of operating costs will lead to an
overstatement of future net income and reserves.
All performance-derived estimates of reserves are limited
by a terminal rate, which is typically described as an economic
limit. A unit cost of oil or gas production never leads to an
economic limit as the cost will simply remain a fraction of
revenue, which illustrates the improper assumption of a
constant unit operating cost.
Changes in recovery process. Problems concerning
operating cost estimates can also occur if future production
involves new recovery mechanisms (e.g. start of waterflood).
In such cases, a careful review is necessary to properly
account for changes in costs as a result of additional
operational requirements.
Guidelines to reduce operating cost mistakes.
• Future operating costs need to be in close agreement
with observed historic costs. Incorporate leaseoperating-expenses (LOEs) from at least two to three
years back in the estimate of future costs.
• Attempt to separate costs into fixed and variable
components.
8
SPE 91069
•
•
•
•
•
Include recurring well or facility expenses in
operating cost.
Account for changes in costs due to new recovery
mechanism.
Avoid simplifying by estimating cost per unit volume
without fixed/variable split.
Include cost for handling of secondary products.
Apply proper escalation of costs if applicable
reserves definitions allow for such.
Inappropriate selection of analogs.
Engineers and
geologists have historically relied upon the use of analogies to
estimate a number of reservoir parameters and performance
expectations. An ideal analog would be a developed reservoir
with well documented physical parameters and an adequate
performance history to rely upon for future production and
performance expectations. Such a reservoir is thus an
excellent analog for predicting the qualities of a nearby
undeveloped reservoir in the same formation assuming the
same development plan and operating scenario.
However, given a number of potential analogs in a
particular area, it is inappropriate to select the best-performing
reservoir as the basis of comparison for a subject reservoir
under study. Indeed, it may be necessary to analyze several
potential area analogs to more fully understand the extent and
impact of the variations in performance before selecting a
particular reservoir or family of reservoirs as “the” analog.
The suitability of a reservoir to serve as an analog is
directly related to the purpose of the comparison. Estimations
of gross rock properties, for example, may be reliably obtained
by direct comparisons with nearby similar reservoirs within
the same formation but ultimate recovery may vary
considerably depending upon well spacing, completion
practices and numerous other operational details that can
affect recovery efficiency.
The estimation of reserves by analogy is typically relied
upon during the early development stages of a field when no
definitive performance and/or geologic data are available. The
analogy method is also frequently applied when new recovery
mechanisms or enhancements are introduced to an otherwise
mature field (e.g. waterflood, well stimulation, infill-drilling).
The process of applying the analogy method typically
involves three stages (all have to be appropriately accounted
for in order to establish a valid estimate by analogy):
• Establish proof of analogy to a mature reservoir and
recovery process.
• Study performance and operations of analogous
reservoir.
• Apply analogy performance with appropriate
adjustments to account for deviations.
Problems associated with properly selecting and applying
analogs can be found with each of the three stages listed
above, but the typical problems encountered are associated
with the first and third item – proof of analogy and applying
analogy to the target reservoir.
Problems with establishing proof of analogy. In most
cases, errors occur as the result of omitting or misinterpreting
the impact of key parameters that are crucial to establishing
proof of analogy. The proof of analogy requires establishing
(i) geologic and petrophysical similarity, (ii) reservoir
engineering similarity, and (iii) operational similarity.
Operational similarity can typically be designed for by
assuring that the target field is operated in a way similar to the
analogous field.
Table 1 illustrates typical parameters that need to be
reviewed to make the case for applying an analogy method. A
case for analogy should only be made if items in Table 1 (as
applicable) are similar or more favorable for the target
reservoir, particularly for a proved reserves classification. Not
all items listed in Table 1 necessarily apply to each case. The
key is to identify the main performance drivers that influence
the process intended to be used for the analogous treatment
and to determine if similarity can be established.
Table 1
Typical parameters to establish proof of analogy
Geoscience
Engineering
Operational
Structural configuration
Lithology and stratigraphy
Principal heterogeneities
Reservoir continuity
Average net thickness
Water saturation
Permeability
Porosity
Areal proximity
Pressure and temperature
Fluid properties
Recovery mechanism
Fluid mobilities
Fluid distribution
Reservoir maturity
Well productivity
EOR specifications
Areal proximity
Well spacing
Artificial lift methods
Pattern type and spacing
Injector to producer ratio
Annual injection volumes
Fluid handling capacity
Stimulation design
Areal proximity
The importance of areal proximity is emphasized in the
following quotation from “Standards Pertaining to The
Estimating and Auditing of Oil and Gas Reserve Information,”
approved by the SPE Board of Directors, June 2001: “If
performance trends have not been established with respect to
oil and gas production, future production rates and reserves
may be established by analogy to reservoirs in the same
geographic area (emphasis added) having similar
characteristics and established performance trends.”4
Mistakes made establishing an analogy will ultimately
result in incorrectly applying analogous performance to the
target field. Typical mistakes associated with establishing
proof of analogy include the following:
• Assuming similarity because of areal proximity and
same formation without proper evaluation of all
parameters.
• Field not located in same geographic area.
• No similarity in critical parameters that have been
overlooked in the analysis.
• Bias towards trying to force analogy if a few key
parameters match.
Problems applying analogy to target field. In cases where
an analogy can be established but there are slight differences
in key parameters, it may be possible to apply the analogy
SPE 91069
method if appropriate adjustments are made that account for
the differences. In most cases, problems are the result of
inappropriately applying analogous behavior due to:
• Not designing for operational similarity, particularly
well density.
• Not making appropriate adjustments to account for
operational differences including costs.
• Not making appropriate adjustment to account for
differences in geoscience and engineering
parameters that can be quantified (e.g. displacement
efficiency as a result of differences in fluid
properties or differences in stratification that may
affect vertical sweep).
Examples.
• When estimating future recovery from a planned
waterflood by analogy, it is not only necessary to
establish similarity between geoscience and
engineering parameters that would assure similar
displacement and sweep behavior but to also design
the target waterflood in such a way that is similar to
the analogous field in terms of well spacing, pattern
type, and annual injection volumes. The lack of
creating operational similarity is frequently the
cause for projections that may be either too
conservative or too aggressive.
• Similarly, differences in mobility may not
necessarily disqualify an analogy as long as the
proper adjustments are being made to account for
the change in displacement efficiency.
Guidelines to reduce mistakes using analogies.
• Give preference to analogies in areal proximity to
target field.
• Follow a strict process whereby key parameters that
need to be similar are tabulated and compared.
• Accept analogy only if a good match exists or if
adjustments can be quantified to account for
differences - qualitative or “instinct” adjustments
need to be weighed carefully and may be cause for
down-grading to a lower reserves classification.
• Review, and if necessary, design for operational
similarity (this will also allow capture of appropriate
cost).
Simulation-derived estimates of proved reserves. Almost
all significant oil and gas reservoirs worldwide are managed
through the use of detailed reservoir models. These models
provide an excellent tool for decisions related to development,
operations and reservoir management. Palke and Rietz5 have
described some of the concerns about using even the most
robust models for the estimation of proved reserves under any
of the more commonly recognized definitions of reserves.
They suggest that for immature reservoirs, simulation is
most useful to estimate recovery efficiencies and for testing
the ranges of other parameters including permeability and
aquifer support. They further recommend that models of
mature reservoirs should be used for proved reserves estimates
9
only when reasonable history matches of the reservoir and
wells have been obtained. These observations clearly are not
intended to suggest that using reservoir simulation as the basis
of reserves estimation is a mistake, but are intended as a
warning about the dangers of doing so without a detailed
review of the model to fully understand the assumptions,
limitations, and applicability associated with the model.
Failure to do so may result in a significant overstatement of
proved reserves.
Failure to incorporate early life performance data into
volumetric estimates. Early life production and pressure
decline trends may not be sufficiently definitive to provide the
sole basis for reserves estimation but should be continuously
reviewed in the “fine tuning” of a volumetric, analogy, or
simulation derived reserves estimate. Quite frequently this
early life data (including initial rate and pressure data as well
as any available trends) has not been utilized to calibrate
“static” estimates until well past the half-life of a reserve
estimate. This disregard of early performance data (and
potential warning signs) may lead to significant reserves
revisions, either positive or negative.
Commonly found errors include: (i) not revising the
reserves expectations for undeveloped locations based on
performance data of producing wells, (ii) not anticipating the
impact of unexpected increase in water or gas production, and
(iii) not accounting for effects of pressure depletion on behindpipe and infill locations over time.
Updating undeveloped locations based on performance
data. Reserves estimated for undeveloped locations at the
beginning of field development are typically based on
drainage area and recovery factor assignments, many times in
combination with analogies from nearby fields. Volumetric
calculations and recovery factor estimates need to be reviewed
and revised (calibrated) as performance data becomes
available. Deviations from the initial estimates may require
adjustments to recovery factors, rate projections, and numbers
and locations of future development wells. Some of the
largest errors often occur if existing wells are adjusted for
lower productivity but ultimate reserves are maintained by
extending field life. This situation creates two critical
problems: (i) lower initial rates may be indicative of lower
productivity, thinner pay, interference effects, and/or smaller
drainage areas; therefore per-well reserves and/or in-place
volumes may be overestimated, (ii) capital allocations may be
underestimated as more wells may be necessary to achieve the
previously estimated volumes and the resulting net present
value will therefore be overstated.
Early or unexpected water production or increases in
gas-oil-ratio. The unexpected increase in water production in
down-dip wells or gas-oil-ratios in up-dip wells may have an
effect on reserves booked in wells throughout the field.
Problems associated with unexpected changes in water or gas
production are typically the result of uncertain drivemechanisms. For example: (i) undeveloped locations may
have been booked up-dip of an existing location based on an
expected strong waterdrive - but existing wells are
10
SPE 91069
Effect of depletion on behind-pipe and infill locations.
Behind-pipe reserves or infill wells are established at a certain
point in time under existing pressure and depletion (or sweep)
conditions. Behind-pipe reserves and infill wells are often
“kept on the books” for several years or longer depending on
the allocation of capital spending or timing of other projects.
Over time, the volumes assigned to behind-pipe and infill
wells need to be re-evaluated as existing wells may have
drained some or essentially all of these volumes, even in low
permeability reservoirs. A recommended approach to avoid
carrying reserves that may have already been drained is to
compare produced volumes with the expected ultimate
recovery for the entire reservoir. Such an approach allows
timely adjustments to the remaining volumes for behind-pipe
or infill wells. Judgment should be exercised to ensure that
the remaining volumes can reasonably be expected to be
drained by the proposed behind-pipe completion or
undeveloped locations.
Other commonly found problems associated with
performance adjustments.
• Recovery factors based on optimistic but
unconfirmed drive mechanisms.
• Assumed well drainage areas or reservoir area (updip locations, seismic amplitude maps).
• Setting up offset locations without compelling
evidence of reservoir continuity.
Events that should trigger review of all reserves.
• New wells with unexpected changes in reservoir
thickness, fluid contacts, pressures, and/or
productivity.
• Early or unexpected water production or
unanticipated increases in gas-oil-ratio.
• Significant deviations from expected production or
pressure decline trends.
• Reserves for undeveloped and behind pipe locations
that have not been reviewed in several years.
Guidelines to reduce frequency of mistakes.
• Always review the potential field-wide implications
of new data.
• Do not assume that, by chance, only poor locations
are being drilled and the good ones are yet to come.
• Exercise caution placing undeveloped locations in
where drive mechanisms or efficiencies are uncertain.
Impact of partial waterdrive and over-pressured
reservoirs on gas material balance. The standard gas
material balance analysis (p/z analysis) is a common tool used
to determine both reservoir size and recovery for a given
abandonment pressure. Combined with volumetric estimates,
gas material balance is an effective tool to estimate reserves,
particularly in mature reservoirs. Many problems with gas
material balance are typically encountered earlier in the field
life when less than 25 percent of the expected volume has
been produced or reservoir pressures are still above the normal
pressure gradient. During this early period, factors that
influence the reservoir pressure behavior such as compaction
and partial waterdrive can be indistinguishable from a pure
depletion drive (Figure 16).
Although it is sometimes difficult to isolate reservoir
mechanisms that may affect the p/z trend, there are a few
guidelines that, if followed, can reduce the risk of
overestimating gas in-place and recovery early in the field life.
Full Water Drive (constant pressure)
p/z (psia)
experiencing increased gas-oil-ratios indicative of a secondary
gas cap or a smaller than anticipated reservoir. Conversely,
(ii) undeveloped reserves may have been set up on-strike with
existing wells that water-out prematurely due to expectations
of a depletion or weak aquifer drive.
Under such
circumstances, not only do the wells affected need to be reevaluated but any undeveloped or behind pipe reserves need to
be reviewed as well.
Pu
re
Partia
l Wate
r
De
ple
tio
n
Dr
Ab
n
ive
Drive
or
m
al
Pr
es
su
r
e
Ap p
ar e
nt G
as in
Plac
e tr
e nd
G
Gp (scf)
Figure 16. Conceptual gas material balance graph.
Conditions that should trigger caution when using gas
material balance.
• Over-pressured reservoirs with a gradient of 0.6 psi/ft
or higher.
• Small pressure change to original pressure (may be
indicative of water influx).
• Apparent gas in place significantly larger than
volumetric estimate.
• Areas prone for water influx or high pressure
gradients.
• Cumulative production less than 25 percent of
expected ultimate based on volumetric estimate.
• High withdrawal rates that may mask water influx in
early life.
Guidelines to reduce risk of overestimating gas in-place
and ultimate recovery using material balance.
• Never base an early life reserves estimate on material
balance alone – include volumetric data and/or
performance analysis if available.
• Review other, more mature fields in the area to look
for trends in p/z behavior and observed abandonment
conditions.
• Over-pressured reservoirs typically exhibit linear p/z
trends until a normal pressure gradient is reached.
Exercise caution and revert to volumetric analysis
SPE 91069
•
11
until a second trend materializes below the normal
pressure gradient.
Exercise caution assuming low abandonment
pressures if water-loading may become an issue
(include nodal analysis calculations).
Assignment of proved reserves to undrilled fault blocks.
Virtually all recognized proved reserves definitions refer to
“known reservoirs” or “known accumulations” as being a
necessary qualifier for the attribution of proved reserves. The
term “known” has generally been interpreted as requiring a
well penetration. Accordingly, it is deemed inappropriate to
classify un-drilled fault blocks or reservoir segments as being
proved reservoirs.
Area experience with seismic
interpretations may have advanced to the point where a 90
percent confidence factor can be attributed to an undrilled
reservoir but even such a high level of confidence is typically
not considered adequate to declare an un-drilled fault block as
being “a known reservoir”.
Incomplete understanding of commercial economics
projection software. Thompson and Wright6 described an
investigation into the application and results achieved through
use of Economics Software Programs developed, sold, and
maintained by 12 vendors. They developed 30 test cases with
straightforward assumptions about future production rates of
oil and gas, constant and variable gas-oil-ratios and
condensate yields as well as reversionary interests and
overriding royalties. Assumptions further included prices and
costs, both constant and escalated. Results were to include
future net income undiscounted and discounted at several
designated annual rates. Each of the 12 vendors was provided
a copy of these cases.
These cases were not unusually complex nor were they
created in such a way as to be easily misunderstood. The
vendors were given one month to complete their forecasts of
results. One of the simpler cases specified a drilling cost, an
initial monthly oil production rate, an effective annual decline
rate, exponential production decline, working and net revenue
interests (constant), taxes as a percent of revenue, and a
beginning oil price and monthly operating cost, both escalated
at 3 percent annually. The ranges in certain results are
tabulated below:
Undiscounted net PV
Net Present Value – 20%
Internal rate of return
High
$181,000
97,000
104.2%
Low
$124,000
3,000
20.7%
Expected
$160,000
52,000
34.3%
The “expected” case was prepared by the authors and was
essentially hand calculated over the 5-year project life. The
differences reported above arose from one simple case but
were magnified as example cases became more complex.
How could this happen? Different program assumptions were
made involving the number of days in a year, the timing of the
receipt of income, timing of expenses, differing discounting
and escalation calculations and the timing of payouts
triggering reversionary interests.
This topic is presented herein simply as a caution to those
who find it necessary to rely upon results from software with
which the user has not developed a high level of confidence
over time. This is neither an endorsement nor condemnation
of any vendor-developed and supported evaluation software
but simply a warning about reliability of results.
Summary and Conclusions
In any discussion about errors and mistakes, it is almost
inevitable that some measure of personal experience and
opinion becomes imbedded in some way. The authors have
diligently tried to confine the foregoing to factual matters but
have found it necessary to blend factual findings with certain
personal observations and opinions.
We respectfully
acknowledge that other evaluators may not agree with the
entire contents of this paper but trust that these observations
may be a small contribution to the art and science of reserves
estimation. One final admonition to fellow geoscientists and
engineers; please document and archive copies of all pertinent
work notes and use this accumulated history as a basis of
refining future estimates of reserves.
Acknowledgements
The authors would like to express their gratitude to fellow
Ryder Scott colleagues Larry Connor, Bob Wagner, Joe
Magoto, James Latham, Dean Rietz, Debby Schiro, Pamela
Nozza, and Vickie Sublett for their comments and help during
the preparation of this paper.
References
1. Harrell, D.R. and Cronquist, C.: Petroleum Engineering
Handbook, Chapter 19, SPE, Richardson, TX (pre-publication).
2. Wharton, J.B., Jr.:“Isopachous Maps of Sand Reservoirs,” Am.
Assoc. Petroleum Geologists Bull, v.32, No. 7, pg. 1331-1339,
1948.
3. Tearpock, D.J. and Bischke, R.F.: Applied Subsurface Geological
Mapping, Prentice-Hall Inc., Englewood Cliffs, NJ (1991).
4. SPE: "Standards Pertaining to The Estimating and Auditing of Oil
and Gas Reserve Information", approved by the SPE Board of
Directors, June 2001. Article 5.7 Estimating Reserves by Analogy
to Comparable Reservoirs.
5. Palke, M.R. and Rietz, D.C.: “The Adaptation of Reservoir
Simulation Models for Use in Reserves Certification Under
Regulatory Guidelines or Reserves Definitions,” paper SPE
71430 presented at the 2001 SPE Annual Technical Conference
and Exhibition, New Orleans, September 30 – October 3.
6. Thompson, R.S. and Wright, J.D.: “A Comparative Analysis of
12 Economic Software Programs”, paper SPE 68588 presented at
the 2001 SPE Hydrocarbon Economics and Evaluation
Symposium, Dallas, 2-3 April.