SPE-169856-MS

A Probabilistic Approach to Reserves Booking and Evaluation in Early

Development Fields
Víctor A. Huerta Quiñones, PETROPERU S.A. - UNI

Copyright 2014, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Hydrocarbon Economics and Evaluation Symposium held in Houston, Texas, USA, 19–20 May 2014.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been
reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessar ily reflect any position of the Society of Petroleum Engineers, its
officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohi bited. Permission to
reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
Technical recoverable volumes estimation and reserves booking are two of the most significant and complex tasks in the
petroleum industry, since figures add value to company assets and/or the economy of a country. Both processes, which not
always are quite compatible because of project maturity or non-technical issues, usually embrace uncertainty due to lack of
information and different criteria of reserves auditors.

Conventional incremental deterministic approach ruled by guidelines of the Petroleum Resources Management System
(PRMS-SPE/WPC/AAPG/SPEE) and the Securities Exchange Commission framework is quite restrictive to book for proved
reserves of a field in an early development phase or few years of production history; in addition, 2P and 3P estimations are
closely linked to current geological interpretation, and are not flexible to account for the whole uncertainties inherent to
project at that level of maturity.

A new probabilistic approach has been proposed to provide consistency and confidence to conventional 1P, 2P and 3P
deterministic estimations in recent gas condensate and oil fields. The incorporation of strong descriptive statistics concepts in
traditional probabilistic volumetric models, such as central limit theorem, dependency between model inputs (Bayes’
theorem), aggregation of volumes, as well as, the use of the standard error of mean, have given consistency to volumetric
model outputs. These probabilistic calculations, adjusted to SEC definitions for proved reserves, and SPE standards, for 2P
and 3P reserves categories, as well as, contingent resources, allows dealing with uncertainty inherent to a new field on
development phase.

Introduction

The value of a petroleum field is determined by its size and the volume of hydrocarbons in place. The volume of existing
hydrocarbons can be classified and categorized according to PRMS1 in two main factors:

1
Petroleum Resources Management System 2007
2 SPE-169856-MS

Figure N°1 PRMS Classification Framework

1. Degree of technical uncertainty (X axis)
2. Opportunity of commerciality (Y axis)

The volume of hydrocarbons in place is the key factor when conceiving field development strategies and plan, taking into
consideration all investment decisions (current and future); eventually, the field development plan is expressed as a cash flow
asessement which profit indicators, as well as other external factors (e.g social and legal issues) will result in the
implementation of the project or not. Decisions involving capital investment deployments include:

• Environmental Studies
• Subsurface studies
• Drilling and workovers
• Pipe Sizing
• Design of facilities
• Sale contracts
• Field abandonment, etc.

In this work, all efforts have been focused on reducing technical uncertainty (factor No. 1) and have not taken into account
technological, economic, legal, environmental and social risk factors which are present in all projects.

Estimated in place and technical recoverable volumes of hydrocarbons in a field should consider heterogeneity sources of
rock properties (porosity, permeability, fluid saturation, net thickness), fluid properties (formation volume factors, solution
gas, viscosity), as well as, complexity in geological and geophysical features; then, success will strongly depend on the
understanding and management of the uncertainty.

Uncertainty may be addressed with deterministic (scenario analysis) and probabilistic methods. Both methodologies,
independently, have strengths and weaknesses, but deterministic method (which average fixed values and discrete scenarios),
is conventionally the most used and preferred. However, the probabilistic approach has the power to use the full range of
values that can reasonably occur for each unknown parameter and can generate a full range of possible outcomes (expected
volumes in this case), which provides a better understanding of the risk. The incorporation of strong principles of descriptive
statistics such as central limit theorem, the use of the standard error of mean (SEM), Bayes’ theorem and aggregation of
volumes has contributed to give consistency to probabilistic approach of estimation of resources and reserves, which
nowadays is in accordance to PRMS guidelines and is more accepted by SEC regulations.

The integration of deterministic and probabilistic approaches will take advantage of the strenghts of these methods, towards
proving a better assessment and management of subsurface uncertainties, as well as, a better understanding of risk, in oil and
gas projects.

Two field cases in early development phase were studied to test this new approach; a gas condensate field (Field 1)
throughout its whole development before production phase, and a deep oilfield in initial development in Peru (Situche
Central), in which some development strategies are currently under evaluation.

Before discussing this novel probabilistic approach, it is quite important to clarify the following concepts:
 Procedures to calculate reserves and resources
 Assessment approaches
 Estimation Methods
SPE-169856-MS 3

I. Resources and Reserves Calculation Procedures

They can be classified into 3 categories:

• Analogies
• Static: Volumetric Estimation
• Dynamics: Material Balance, Decline Curve and Simulation.

Figure 22 shows how, depending on the maturity of the project, a calculation procedure is preferred to be applied. The
confidence in the evaluation results usually increases when estimations are supported by more than one analytical
procedure1.

Figure 2. Production profile

1. Analogies.-
The analogies are widely used in the resource estimate, particularly in the exploration and development stages, when
information is limited direct measurements. The methodology is based on the assumption that the analog field is
comparable with the reservoir study in terms of reservoir properties that control fluid and ultimate hydrocarbon
recovery1.
Advantage It is useful in very early stages when there is not much technical information.
Disadvantage It is hard to find a similar field, preferably with sufficient production history.

2. Volumetric estimation.-
It is mainly used to estimate the original volume in place, the petroleum initially in place (PIIP) is calculated using
an average value for each parameter. Reserves @ EUR are estimated by multiplying PIIP and recovery factor with a
mathematical model based on the volumetric calculation shown below, to estimate reserves at any time, it is
necessary to discount cumulative production if any.

Advantage It is fast and easy

Deterministic values of these parameters do not take into account dynamics of the
Disadvantage
reservoir.
Reservoir geometry and limits
Key Distribution of porosity and permeability
Uncertainties1 Contacts locations
Fluid type and saturation distribution
Recovery factor can be estimated by analogy or other analytical procedures.

3. Material balance.-
It is the simplest method based on dynamic expansion of the fluids in the reservoir. This analytical method relates
the in-situ original volumes with production and pressure history.

2
Senturk, Y. (2011). Chapter 4. Guidelines for the application of PRMS
4 SPE-169856-MS

Takes less time than a numerical model, it is independent of the geological model and use
Advantage production history. It simple because the reservoir is considered as a tank and petrophysical
average properties are assigned.
When it is considered the reservoir as a tank and average properties are assigned, it does
Disadvantage
not take into account the areal and vertical heterogeneities.
Quality data provided:
• Rock Properties
Key • Functions of relative permeability and fluid properties.
Uncertainties • Production history and pressures
Complex situations such as: water invasion, compartmentalization, multiphase reservoir
behavior and reservoir with multiple horizons or low permeability

4. Production Performance Analysis:

The essential factor for the occurrence of production decrease capacity of a reservoir occurs as a consequence of an
internal pressure decrease, which leads to lower energy levels. They are known traditionally as decline curves which
are based on: "Those factors wich have affected production in the past will continue to do in the future", but there
are other forms of curves which can be used (GOR, WGR).

This method is often used when there is sufficient production history so that you can distinguish a trend which has
been described mathematically in three types of curves: exponential, hyperbolic and harmonic.

Advantage Provides valuable information about the performance of the field.

Disadvantage In the early stages, there may be significant uncertainty in the final performance profile.
Key
Comercial uncertainties may impact the abandonment rate.
Uncertainties

5. Reservoir Simulation. -
It is a very important method to use in complex fields (highly faulted structures, pinch-out, etc).
In reservoir simulation, the field is divided into a series of blocks or cells and average properties are assigned to each
one. They interact through Darcy's equation and the principle of mass conservation. Reservoir simulation has the
ability to integrate the static model with dynamic model including field performance through the production history.
Advantage The estimation is more accurate because static and dynamic models tend to represent more
accurately the complexity of the structures
Disadvantage To more detail in building the model, more time it takes to perform calculations.
Key
Population of properties in areas without information
Uncertainties

II. Assessment approaches

Regardless of the procedure used to calculate resources and reserves, the PRMS recognizes two equally valid
approaches that can be interpreted as two different ways of seeing the same problem2. They can be classified into:
• Cumulative (scenario)
• Incremental (based on risk)
Both are equally reliable and depending on the level of technical uncertainty, both results can be comparable.

1. Cumulative approach2. -
Under this approach, what it is done is to prepare three separate scenarios that support the uncertainty through
sensitivity analysis (values estimated using 3 sets of plausible parameters). These scenarios are designed to represent
the low, medium (most likely) and high case of PIIP and reserves associated.

2. Incremental approach2. -
In the incremental approach, experience and professional judgment are used to estimate the volumes for each
category (proved, probable and possible) as discrete volumes separately and is intimately associated to the increased
level of uncertainty expressed qualitatively by the "reasonable certainty concerning" and quantitatively by their
respective average level of confidence (or risk factor).
SPE-169856-MS 5

Cumulative vs Incremental

Case 1 - Cumulative Approach: Wells are distributed throughout the structure where the main uncertainty goes through
out the vertical variability where sceneario N°1 (1P rserves) is defined by the Lowest Known Hydrocarbon (LKH),
scenario N°2 (2P reserves) is defined by a contact placed at the midpoint of LKH and the spill point, and finally
scenario N°3 (3P reserves) is defined by the spill point (see figure 3).
Case 2 - Incremental Approach: Wells are located toward the left side of the structure. In this case, the main
uncertainty goes through out the horizontal variability. The areas where the wells are located and those surrounding
areas defined according to drainage area (1 spacing) will be categorized as proved, those areas located at 2 spacings of
distance around the original locations are classified as probable and those areas located at 3 spacings around original
locations are classified as possible (see figure 3).

Figure 3. Technical Uncertainties & Reserves Categories

The cases are useful to highlight the fact that, depending on the technical uncertainties present, we can use the most
appropriate approach.
2
III. Estimation Methods

We can divide them into two groups:

• Deterministic Method
• Probabilistic Method

Regardless of the procedure used for calculation, it is possible to prepare resource estimations using deterministic or
probabilistic methods.

A deterministic estimation is a discrete scenario within a range of results which might arise from a probabilistic
analysis. In the deterministic method, you select a discrete value or set of values for each parameter based on the
selection by the estimator of the values that are more appropriate for the corresponding resource category. It derives into
a simple result of the recoverable amounts for each increment or deterministic scenario.

For probabilistic method, the evaluator defines a distribution that represents the full range of possible values for each
input parameter. These distributions can be sampled randomly (typically using a Monte Carlo simulation software) to
calculate a complete range and distribution of potential results of recoverable quantities.

IV. Applied Descriptive Statistics

Descriptive statistics is the most suitable discipline to describe uncertainties and evaluate the risk. It is intended to
develop features of data or of probability distributions3. Four common types of graphs are used to relate model variables
to probability: Pareto plots, histograms, probability density functions (PDFs) and cumulative distributions (CDFs). In
addition, measures of central tendency, mean, median and mode, as well as, percentiles P90, P50, and P10, are frequently
used to understand these graphs and assess risk; no less important, measures of dispersion and symmetry, variance,
standard deviation, standard error of the mean (SEM) 4 and skewness are needed to describe how the data are dispersed or
spread out from the “center trend”, and indeed, how the measures of central tendency are representative of the entire
population of data5.

3
Murtha, J. Peterson, S. Adams, W. (2007). Petroleum Engineering Handbook. Volume VI, Emerging and Peripheral Technologies, Chapter 10, Risk and
Decision Analysis.
4
Murtha, J. Ross, J. (2009). Uncertainty and the Volumetric Equation. SPE-JPT September 2009.
5
Murtha, J. Peterson, S. Adams, W. (2007). Petroleum Engineering Handbook. Volume VI, Emerging and Peripheral Technologies, Chapter 10, Risk and
Decision Analysis.
6 SPE-169856-MS

Application of descriptive statistics would be useless without having representative samples of data. Probability density
functions should be “fit” to real data of each input variable to assure consistency of the models (the goodness of fit is
usually judge by the “chi-square” value). In case fitting data to distributions is not possible, expert opinion would be
necessary to decide which distribution to use to represent a given parameter. The following list provides guidelines for
using some common distributions6:

Normal distributions.- they are often used to represent variables, which themselves are sums (aggregations), of average
of other variables, according to central limit theorem7. The most popular applications are:

 Field production, which is a sum of various wells and reservoirs.

 Reserves for a business unit, which means aggregation of reserves from various fields.
 Total cost, which is the sum of line-item costs.
 Average porosity of a given structure
 Average saturation of a given structure
 Average net pay and formation volume factor for a given structure

Log-Normal distributions.- the log-normal distribution is the most used in the oil and gas industry, because it arises in
calculating resources and reserves. It follows from the central limit theorem 8 that products are approximately log-normal.
The most popular applications are:

 Volumes of reserves and/or resources of a single reservoir.

 Areas of structures in a play.
 Production rates (Darcy’s equation)
 Time to reach pseudosteady state.
 Effective permeability

Triangular distributions.- They are widely used by people who simply wants to describe a variable by its range and a
mode (minimum, maximum and most likely values); the minimum and maximum are almost unlikely to occur.

V. New Probabilistic approach to Reserves Evaluation in Early Development Fields

Probabilistic assessment of reserves is now a fancy technique which supports traditional deterministic calculations;
advantages offered by including multiple scenarios give reserves estimates a stronger technical certainty towards their
booking and reporting. Traditional view of deterministic method to address uncertainty by 1-average or 3 scenarios (1P,
2P and 3P outcomes) may be outgrown by Decision / Scenario Analysis which allows multiple evaluations that
provides a range of discrete possible outcomes.

The multiple scenario approach yields to a discrete probability function, which in turn, can be converted into a
cumulative distribution curve (see figure 4).
Net
Pore GIIP GIIP 1
Structure GWC GRV por*ntg Volume 10**9 m3 Probability

0.6 0.14 479 94 0.18 0.8

Cumulative Probability

0.6 -3190 3420

0.4 0.23 787 155 0.12
0.6
high
0.5 0.6 0.14 234 46 0.12
0.4 -3040 1670 0.4
0.4 0.23 384 75 0.08

0.2
0.6 0.14 350 69 0.18
0.6 -3190 2500
0.5 0.4 0.23 575 113 0.12 0
low 0 50 100 150 200
0.6 0.14 181 36 0.12 GIIP (10**9 m3)
0.4 -3040 1290
0.4 0.23 297 58 0.08
Probabilities in italics

Figure 4. Multiple realisation approach

6
Murtha, J. Peterson, S. Adams, W. (2007). Petroleum Engineering Handbook. Volume VI, Emerging and Peripheral Technologies, Chapter 10, Risk and
Decision Analysis.
7
Murtha, J. (2004). Risk Analysis for the Oil Industry. HART’s E&P Magazine.
8
Murtha, J. (2004). Risk Analysis for the Oil Industry. HART’s E&P Magazine.
SPE-169856-MS 7

On the other hand, the Monte Carlo Simulation9 (see figure 5) allows using the full range of values that could
reasonably occur for each reservoir parameter to generate a full range of possible outcomes for the resource volume. To
do this, first, each reservoir parameter of a volumetrics-base equation is described by a so-called probability density
function (PDF) to cope with urcertainty regarding geoscience and engineering data; then, using a stochastic sampling
procedure, it is randomly drawn a value for each parameter to calculate a hydrocarbon in place volume or a technical
recoverable estimate. By repeating this process a sufficient amount of iterations, a cumulative distribution function,
commonly known as expectation curve, is generated. It is usually a continuous inverse cumulative curve that measures
the full range of uncertainty of a model output. Cumulative functions help us to determine traditional P90, P50 and P10
values (or other percentiles), as well as, the chance that the model output will or will not exceed some value, or fall in
between two values (confidence interval).

p
p
p
p
p 1/Bo
Sh p

N/G
GRV
STOIIP
p>x

STOIIP
Figure 5. Monte Carlo approach to volumetrics

This paper proposes a simple but robbust integrated method to estimate both original hydrocarbon and technical
recoverable volumes. This method uses the volumetric equation as a base model, and some probabilistic density
functions for each reservoir parameter (model inputs), selected by expert opinion and some descriptive statistics
principles; truncantion ranges, central limit theorem, SEM, dependency, and aggregation principles are carefully
analized to assure representative and consistent results. Finally, in case of the expectation curve follows a log-normal
trend, an special linear plot (log-normal-probabilistic) may be used to compare high, medium and low deterministic
figures with corresponding probabilistic estimates (P90, P50 and P10) as QA/QC.

It is worth pointing out that the concept of aggregation of reserves should be carefully treated. If two or more reservoirs
are consider fully independent, corresponding proved reserves should not be arimethically added (expected curves
should), because we run the risk of underestimating proved recoverable volumes of the combined reservoirs. On the
other hand, when there is partial or total dependency between two or more oil zones, Bayes’ theorem should be strictly
applied; however, calculations can turn into complicated and unpractical for many oil zones produced in commingled. A
simplified arimethical aggregation of proved reserves may be appropriate10.

A generalized volumetric equation11 has been prepared for estimating technical recoverable volumes of hydrocarbons in
oil and gas fields.

1
Re servesHC  ( Bulk volumentg )S h ( )(1  shrinkage)( RF )
FVF

9
Swinkels, W. (2011). Chapter 5. Guidelines for the application of PRMS
10
Cronquist, Ch. (2001). Chapter 9. Estimation and Classification of Reserves of Crude Oil, Natural Gas and Condensate.
11
Huerta, V. (2012). Session III. "Deterministic and Probabilistic Methods." SPE-ATW 2012 on Petroleum Reserves & Resources Estimation.
8 SPE-169856-MS

Figure 6 depicts the main parameters for a volumetric equation with their corresponding PFDs selected by expert opinion.

HEES
Preferred
Parameter P90 Most Likely P10 Model
Distribution
Volumen (Acre-ft)* 429824 962941 662979 log-normal
ntg: 0.23 0.32 0.4 0.32 normal
Porosidad 0.2 0.215 0.23 0.22 normal
sw 0.18 0.23 0.27 0.23 normal
1/Bg 207 217 228 217.50 normal
1-shrinkage 0.9 0.94 0.96 0.93 normal
RF 0.78 0.8 0.82 0.80 log-normal
Condensate Yield 25 28 32 28.50 normal
NGL Yield 28 36 43 35.50 normal
EUR (BCF) 237.2
Condensate (MMSTB) 6.8
NGL (MMSTB) 8.4

EUR SPE (BCF) 237.2

Condensate SPE (MMSTB) 6.8
NGL SPE (MMSTB) 8.4

*Inputs P95 & P5 for Volume (Acre-ft)

Figure 6. Probabilistic Volumetric Model

This method is situable for fields in exploratory, appraisal or early development phase. Depending on the case, a proved
distribution function12 should be recommended for 1P volume to conform to regulatory guidelines, while an overall
distribution is applied to account for 2P and 3P estimates.

VI. Case Studies

Two (2) case studies in Peru regarding a gas condensate field and a deep oilfield are presented to illustrate the
methodology; it should be pointed out that estimations of the first case study are shown in equivalent units (UN) for the
purpose of this manuscript.

1. Case study 1: Reserves estimation and follow up of Field 113

Field 1 is finishing development phase. In addition to the wildcat well (which was flanked), 2 additional development
wells were initiall drilled in the structure; as they were placed to the upper part of the field, gas water contacts of main
reservoirs were not seen.

A deterministic analysis was done to account for technical recoverable volumes by category. Lowest known gas levels
were set for proved reserves (1P), and spill points for 3P estimations. A source of uncertainty was detected to calculate
2P reserves; available pressure data taken from RFTs of Field 1 and analogous fields were carefully analized to set a
tentative GWC for 2P estimate. In addition, porosity and gas saturations were calculated from log measurements of
Field 1 wells, while formation volume, recovery and shrinkage factors and were taken from analogous field reports.
Results of deterministic analysis are shown in table 1.

Category 1P 2P 3P
Field 1 0.58 1.00 1.35
Table 1 Deterministic Gross Recoverable Gas (UN)

Then, a probabilistic analysis was carried out to cope with the uncertainty of placing GWC levels of main reservoirs. A
Monte Carlo simulation was chosen to run a sensitivity analysis, using the same volumetric inputs as reference, but
considering the full range of values for each reservoir and surface parameter in Field 1. As it was suggested previously,
a distribution considering only petrophysical and fluid properties variations while keeping constant net rock volumes
was used to determine 1P estimate; The P50 value of the proved reserve distribution corresponds to 1P estimate. On the
other hand, the overall curve was used to calculate 2P and 3P recoverable volumes of gas (see figure 7 and table 2).

12
Acuña, H. Harrel, D. (2000). "Adapting Probabilistic Methods to Conform to Regulatory Guidelines." SPE 63202.
13
Huerta, V. (2012). Session III. "Deterministic and Probabilistic Methods." SPE-ATW 2012 on Petroleum Reserves & Resources Estimation.
SPE-169856-MS 9

Probabilistic Calculation
Field 1 Reserves Distribution (UN)
100%
0.56
90%
80%
70%
60%
0.93
50%
40%
30%
Proved Distribution
20% Total Distribution
1.29
10%
0%
0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80
UN

Figure 7. Field 1 recoverable gas distribution

Category 1P 2P 3P
Field 1 0.55 0.93 1.29
Table 2 Probabilistic Gross Recoverable Gas (UN)

Differences between deterministic and probabilistic calculations at this stage of development seems to be quite small
and tolerable, as it is shown in table 3; only 2P estimates are not completely agreed.

Category 1P 2P 3P
Field 1 0.7 6.7 3.3
Table 3 Differences between Deterministic and Probabilistic Gross Recoverable Gas (%)

A log-normal probabilistic plot was built for QA/QC purposes 14. As it seen in figure 8, 1P estimate is considered as the
pivot, 2P and 3P deterministic are adjusted to yield a log normal distribution (straight line)

Probabilistic vs. Deterministic Reserves

P99 Field 1 P1

2P deterministic
P10
calculations
P90
might be
switched to corresponding
probabilistic estimates, so
as P50 to, account for P50
uncertainty and find an
appropriate straight line for
the three categories.
P10 P90
------- Deterministic P50P10
____ Deterministic P90P10
____ Probabilistic

P1 . Deterministic Figures
P99
0 1 10

UN

Figure 8. Log-normal straight line plot of Field 1 recoverable gas

14
Araujo, H. Rattia, A. (2011). "Reserves follow-up using an integrated Deterministic-Probabilistic Approach." SPE 143843.
10 SPE-169856-MS

To the end of the history, operator of Field 1 decided to drill a new development well which was placed at the flank of
the structure; fortunately, this third development well allowed identifying GWC levels for all the reservoirs, as well as,
move volumes of P2 and P3 categories to 1P recoverable resources.

A new Monte Carlo simulation was run to update estimation of recoverable gas volumes; as GWC levels are clearly
defined, it was not necessary to describe a proved reserves distribution; an overall curve was used to recalculate 1P, 2P
and 3P recoverable gas volumes. Results are shown in table 4 and figure 9.

Category 1P 2P 3P
Field 1 1.09 1.18 1.27
Table 4 Updated probabilistic Gross Recoverable Gas (UN)

Probabilistic Calculation
Field 1: Reserves Distribution after Development
100%
1.09
90%
80%
70%
60%
1.18
50%
40%
30%
20%
1.27
10%
0%
0.80 0.90 1.00 1.10 1.20 1.30 1.40 1.50
UN

Figure 9. Field 1 updated recoverable gas distribution

A log-normal probabilistic plot was rebuilt for QA/QC purposes. As it seen in figure 10, the monitoring of the 1P, 2P
and 3P estimates of recoverable gas volumes shows that figures follow a steep straight that fix a log normal distribution.

Probabilistic vs. Deterministic Reserves

P99 Field 1: follow up after finishing development P1

P90 P10

New reserves figures

concur steeply towards 3P
original reserves.
P50
Uncertainty
P50 has been
mitigated with information
of the latest well located at
the flank
P10
of the field
P90
------- Deterministic P50P10
____ Deterministic P90P10
____ Probabilistic

P1 . Reserves after new well

P99

0 1 10

UN

Figure 10. Log-normal straight line plot of Field 1 recoverable gas

SPE-169856-MS 11

2. Case study 2: Field Development Plan and Reserves Estimation – Situche Central - Block 6415

Situche Central is a deep oilfield discovered by the end of 2005. 2 wells, SC2X and SC3X found and tested oil in
Lower Vivian reservoir, as well as, contingent resources in Cushabatay Fm. 2D and 3D seismic information is
available, as well as, data of well logs, cores, pressure tests, PVTs and DSTs. In addition, there is a plenty of
information from analogous fields of Peruvian Block 1AB which includes production and pressure history, rock and
fluid properties and drive mechanism performance. Even though, there are some subsurface uncertainties that should be
addressed towards optimizing development plan and maximizing hydrocarbon recovery.

A new probabilistic tool has been proposed to integrate technical recoverable volumes estimation and production
performance prediction in Situche Central Field.

Hydrocarbon recoverable volumes are estimated using a volumetric equation set up to run Monte Carlo simulation;
thousands of realizations are then captured by the model to build an expectation curve (see table 5 & figure 11).

Vivian Inferior
P90 Most Likely P10
Area (Acres) 5552.51
h (ft): 56 69 62.50
Porosidad 0.0978 0.10
sw 0.189 0.26
Bo 1.25
RF 34.00% 40.00% 46.00% 40.00%
EUR VIV Inf (MMSTB) 62.4
P(S) 1.0
Success.SC? 1.00
RISK EUR VIV inf (MMSTB) 62.4
Table 5 Probabilistic Model to Estimate Recoverable Hydrocarbons – Situche Central Field

Figure 11. Volumetric Recoverable Hydrocarbons of Situche Central Field

As a result, probabilistic figures are quite representative as compared to certified deterministic values (see table 6 and
Figure 12) but proved reserves when there is a disagreement with the auditor criteria on estimation of bulk volume.

Situche Central 1P 2P 3P
Deterministic Estimation 31.45 55.44 85.43
(MMSTB)16
Probabilistic Estimation (MMSTB) 36.7 58.3 90.4
 Estimation (%) 17% 5% 6%
Table 6 Deterministic vs. Probabilistic Reserves – Situche Central Field

15
GEYE-PETROPERU. (2013). Initial Development Plan – Situche Central Field – Block 64.
16
Ryder Scott Co. (2013). Estimated Future Hydrocarbon Reserves and Resources in the Situche Central Field – Block 64.
12 SPE-169856-MS

Probabilistic vs. Deterministic Reserves

Situche Central Field

P10 P90

P50 P50

P90 P10

------- Deterministic Est.

------- Probabilistic Est.
10 100

MMSTB

Figure 12. Deterministic vs. Probabilistic Reserves Estimations – Situche Central Field

Furthermore, some development strategies were evaluated considering different well trajectories (directional and/or
horizontal), as well as, a reasonable range of number of wells to test overall recovery factor (from 6 to 9 according to
previous analysis to analogous fields); after some executions of the conceptual simulation model, a combination of 6
vertical + directional wells were found as the most suitable plan to produce around 40% of the original oil in place of
Situche Central field.

In addition, type curves derived from analogous fields of Block 1AB and a conceptual simulation study of Situche
Central were incorporated to the model to predict production performance under a probabilistic approach. Initial oil rate
along with decline rates and exponents of hyperbolic-type curves were varied following log-normal trends according to
production history of analogous fields in Block 1AB. Then, a Monte Carlo simulation was run to build a probabilistic
performance analysis with 10000 realizations; all the output recoverable hydrocarbons and production profiles with
corresponding percentiles are plot in Figure 13 and Figure 14, respectively.

Figure 13. Technical Recovery Volumes of Situche Central with Decline Curve Analysis (DCA)

In general terms, both the probabilistic Volumetric and Decline Curve Analysis yield consistent figures of technical
recoverable volumes of hydrocarbons; thus, it gives support to the 6-wells development strategy of Situche Field
towards accomplishing a 40% of primary recovery factor.
SPE-169856-MS 13

Production Forecast - Situche Central Field - Block 64

30,000 P90
P85
P80
25,000
P75
P70
20,000 P65
BOPD P60
P55
15,000
P50
P45
10,000 P40
P35
P30
5,000
P25
P20
0 P15
0 5 10 15 20 P10
Years

Figure 14. Probabilistic Production Profiles of Situche Central Field

RESULTS & CONCLUSIONS

1. The approach and methodology to estimate reserves or resources of a petroleum project will strongly depend on its
maturity, as well as, the analysis of field uncertainties.

2. Comparison between results of deterministic and probabilistic methods is used for QA/QC to detect deviations in any
category, and eventually, to mitigate uncertainties in calculations and booking.

3. The use of a specific cumulative distribution function for proved reserves might address effectively the issue of booking
probabilistic reserves under SEC regulations.

4. Standard deviation of continuous distributions for full field volumetric analysis should be adjusted for each model input
to the standard error of the mean (SEM).

5. Log-normal probabilistic plots may be useful to detect inconsistency between probabilistic and deterministic
calculations when there is applicable (e.g: estimation of recoverable volumes of one-single reservoir).

6. Dependency between model inputs should be set in order to be precise with probabilistic reserves estimation; standard
deviation and mean increases when correlations are set up.

7. It should be noted that when there is independency between two or more oil zones, percentiles should not be additive.
We run the risk of underestimating proved recoverable volumes.

8. Follow up of production performance and recoverable hydrocarbons estimates is a recommended practice that can be
addressed by integrating probabilistic analysis to analytical deterministic methods.

9. Decline curve analysis can be set effectively as a performance tool to run a probabilistic estimation of possible ranges of
production forecasts.

10. Integration of probabilistic volumetric and decline curve analysis provides consistency to estimation of recoverable
hydrocarbons, as well as, a better understanding of risk in early development fields.
14 SPE-169856-MS

Acknowledgments
The authors would like to thank PETROPERU and UNI for their authorization to publish this paper.

Nomenclature
PRMS = Petroleum Resources Management System
SPE = Society of Petroleum Engineers
WPC = World Petroleum council
AAPG = American Association of Petroleum Geologist
SPEE = Society of Petroleum Evaluation Enginieers
EUR = Estimate ultimate recovery
FVF = Formation Volume Factor
Bg = Gas formation volume factor (ft3/SCF)
Bo = Oil formation volume factor (bbl/STB)
PIIP = Petroleum initially in place
A = Area
h = thickness
Φ = Effective porosity
Sw = Water saturation
Bhi = Initial Hydrocarbon formation volume factor
Re, RF = Recovery factor
GOR = Gas oil ratio
WGR = Water gas ratio
LKH = Lowest Known hydrocarbon
HKW = Highest known water
GWC = Gas water contact
GRV = Gross rock volume
GIIP = Initial gas in place
PDF = Probability density function
Sh = Hidrocarbon saturation
Sg = Gas saturation
So = Oil saturation
N/G = Net to gross
STOIIP = Stock tank oil in place
QA/QC = Quality assurance/quality control
PVT = pressure/volume/temperature
DST = Drill Stem Test
RFT = Repeat Formation Tester
UN = Equivalent Units
Shrinkage = Includes shrinkage at surface and plant conditions
DC = Decline Curve Analysis

References

1. Petroleum Resources Management System (PRMS) 2007. SPE.

http://www.spe.org/spe-site/spe/spe/industry/reserves/Petroleum_Resources_Management_System_2007.pdf
2. Guidelines for Application of the Petroleum Resources Management System. 2011. Houston, Texas: Society of
Petroleum Evaluation Engineers.
3. Murtha, J. Peterson, S. Adams, W. (2007). Petroleum Engineering Handbook. Volume VI, Emerging and Peripheral
Technologies, Chapter 10, Risk and Decision Analysis.
4. Murtha, J. Ross, J. (2009). Uncertainty and the Volumetric Equation. SPE-JPT September 2009.
5. Murtha, J. Peterson, S. Adams, W. (2007). Petroleum Engineering Handbook. Volume VI, Emerging and Peripheral
Technologies, Chapter 10, Risk and Decision Analysis.
6. Murtha, J. (2004). Risk Analysis for the Oil Industry. HART’s E&P Magazine.
7. Swinkels, W. (2011). Chapter 5. Guidelines for the application of PRMS
8. Cronquist, Ch. (2001). Chapter 9. Estimation and Classification of Reserves of Crude Oil, Natural Gas and
Condensate.
9. Huerta, V., Lanchimba, A. (2012). "A holistic risk analysis approach for reservoir modeling." SPE 153387
SPE-169856-MS 15

10. Huerta, V. (2012). Session III. "Deterministic and Probabilistic Methods." SPE-ATW 2012 on Petroleum Reserves
Resources Estimation.
11. Acuña, H. Harrel, D. (2000). "Adapting Probabilistic Methods to Conform to Regulatory Guidelines." SPE 63202.
12. Araujo, H. Rattia, A. (2011). "Reserves follow-up using an integrated Deterministic-Probabilistic Approach." SPE
143843.
13. GEYE-PETROPERU. (2013). Initial Development Plan – Situche Central Field – Block 64.
14. Ryder Scott Co. (2013). Estimated Future Hydrocarbon Reserves and Resources in the Situche Central Field – Block
64.