Non-OPEC Oil Supply Data and Algorithms for SAGE

1
Non-OPEC Oil Supply Data and Algorithms for SAGE
prepared for Energy Information
Administration April 13, 2006
2
EIA Objectives

• Provide detailed description of conventional oil
  resource base with ability to investigate
  alternative resource assumptions
• Develop bottom-up estimates for oil discovery,
  development and production costs and ability to
  alter these based on assumptions for future
  technology
• Create data and algorithms that could go into
  SAGE to allow price-responsive non-OPEC oil
  production forecasts
• Data and process should be understandable,
  traceable and verifiable and updates should be
  unburdensome.

3
Scope of EEA Work

• Adapt EEAs existing World Assessment Unit (WAU)
  costing model to provide the needed oil resource
  base and cost information
• Create a standalone (Excel) model of annual world
  oil production trends based on the WAU economics,
  historical production, current reserves, resource
  depletion patterns, oil prices and other factors
• Extract the underlying data and results of the
  models into data sets and equations that could be
  used in SAGE
• Write a short report documenting data,
  methodology and results
• Note Modifying SAGE is not part of EEAs work
  scope for this task

4
Project Schedule

• Initial version of World Oil Logistic Model
  created in February.
• Final version of World Oil Production Model
  produced at the end of March including oil well
  accounting.
• Draft final report delivered April 7.
• Final report next week.

5
WAU Costing Model

• Created by EEA in 2004
• Undiscovered oil and gas resource base for oil
  and gas starts with USGS 2000 World Resource
  Assessment
• Adjustments are made to small fields consistent
  with field size distributions created by EEA for
  2003 NPC gas study for U.S., Canada and Mexico
• WAU model resource costs are based on EEA costing
  algorithms and consider
• field size distribution
• location onshore, offshore water depth
• drilling depth
• finding rate
• financial parameters
• Results are computed by country and assessment
  unit and can be aggregated by country or region
  into gas or oil supply curves
• Output from WAU Model for this project will be
  country-level supply curve segments for crude oil
  resource base including new fields and growth to
  existing fields.

6
Annual Oil Production Model (WOLM)

• This is a country-level model with annual
  historical data from 1980 to present and
  forecasts to year 2100
• Spreadsheet built to easily accommodate annual
  updates to historical data each year
• Underlying concept of forecast is a resource
  depletion model (akin to Hubert or logistic
  curve) for which area under production curve is
  equal to resource base

7
Key Relationships for Hubbert or Logistic Curves

• Prod/Cum Intercept Slope(Cum)
• Prod Intercept(Cum) Slope(Cum2)
• RB -Intercept/Slope
• Slope (Prod/Cum) / (Cum-RB)
• Intercept -Slope(RB)
• where
• Prodannual production
• Cumcumulative production
• RBultimate resource base
• Note also Prod Intercept(Cum) -
  Intercept(Cum2)/RB

8
WOLM Differs from Simpler Depletion Models

• Resource base economics are explicit and vary
  based on oil prices and technology drivers
• Shape of production curve is endogenous (see
  figure)
• WOLM accounting includes annual reserves
  additions and wells drilled

9
Data Sources for WOLM

• To keep consistent with other EIA models and data
  series, historical crude oil and NGL production
  data is from EIA (however some adjustments to to
  historical nonconventional oil production will
  have to be made)
• All time cumulative production is from USGS
  (originally from PetroConsultants) with
  adjustments and estimates made by EEA for some of
  the smaller countries.
• Reserves are from EIA also (originally from Oil
  and Gas Journal. EEA adjustments removed
  nonconventional oil reserves.
• Historical producing wells and new oil wells
  drilled are from OGJ and World Oil magazines.
  Unfortunately, data for some countries is sparse.
• Resource base and cost curves come from WAU based
  on USGS resource assessment with costing
  algorithms from EEA. Model user can adjust
  resource base in Production Model to create
  alternative scenarios.

10
Information Processing and Flow
USGS Resource Assessment
EEA Cost Algorithms
Financial Assumptions
Scenario Inputs
Reserves
Production History
Well Data
WAU Cost Model
Country-level Crude Oil Supply Curves
Oil Production Model
Data and Parameters for Non-OPEC Regions
Annual Production Forecasts
SAGE
11
Representation of Supply Curves

• prc1 price at first point on curve (20)
• prc1 price at second point on curve (50)
• prc3 price at third point on curve (100)
• rbf1 portion resource base economic at first
  price point
• rbf2 portion of resource base economic at
  second price point
• rbf3 portion of resource base economic at third
  price point
• These three points on the supply curve are used
  to develop a continuous equation that gives the
  economic resource base (as a fraction) as a
  function of oil price. The coefficients of that
  equation are defined as
• a3pt rbf3 - rbf1
• b3pt LOG((rbf2-rbf1)/a3pt)/LOG((prc2-prc1)/(prc3
  -prc1))
• The supply curve function that finds the economic
  resource as a function of price (variable
  price) is
• SCF rbf1 a3pt((price-prc1)/(prc3-prc1))b3pt

12
Example of Supply Curve
13
Production as Function of Resource Base and
Cumulative Production

• P(t) (CPL1PL1) (PL1/(CPL1(CPL1-RBL1))((CPL1
  PL1) - RBL1))
• where
• PL1 prior period production
• CPL1 prior period cumulative production
• RBL1 prior period resource base

14
Production in One Equation (substitute supply
curve equation for RBL1)

• P(t) (CPL1PL1) (PL1/(CPL1(CPL1-(rbf1 a3pt
  ((PRICEL1-prc1)/(prc3-prc1))b3pt)
  RB))((CPL1PL1) - (rbf1
• a3pt((PRICEL1-prc1)/(prc3 - prc1))b3pt)
  RB))
• where
• PRICEL1 prior period price
• PL1 prior period production
• CPL1 prior period cumulative production

15
Production Acceleration (Shape Multiplier)

• SM(t) MAX(MIN(MAX(0,PRICE-Parm1)Parm2,SM(t-1)
  0.005),SM(t-1) -0.005)
• P(t) after shape adjustment P(t) (SM(t) -
  SM(t-1))
• Where
• Parm1 the threshold price above which the shape
  adjustment is applied (25)
• Parm2 shaping coefficient (0.002)

16
Example Results at Three Oil Price Levels
(Billion barrels per year, sum of country-level
data)
17
Issues for SAGE Implementation

• Representation of upstream technology advances
  (in supply curves or in SAGE)
• Meaning of production in SAGE period (e.g.
  average of all years vs. last year)
• New field resource base (e.g. Canada is 2,774
  MMbbl vs. NPC estimate of 11,745 MMbbl)
• Growth resource base (WAU world total is 390 ex.
  US/Can vs. 612 billion barrels for USGS)
• Definitions and base year calibration of
  production and reserves.
• How to deal with nonproducing areas (e.g.
  Greenland)