Information about http://www.eia.doe.gov/emeu/steo/pub/document/overview.pdf
Regional Short-Term Energy Model (RSTEM) Overview Table of…
Tags: consistent time series, documentation overview, electric power sector, electricity demand, electricity supply, energy data collection, energy information administration, energy model, integrated information system, model calibration, motor gasoline, petroleum products, profile generator, regional electricity, residential propane, simulation engine, steo, term energy outlook, time series format, winter fuels,Pages: 13
Language: english
Created: Sun May 7 21:11:39 2006
Display cached document
Page 1
Page 2
Page 3
Page 4
Page 5
Page 6
Page 7
Page 8
Page 9
Page 10
Page 11
Page 12
Page 13
Regional Short-Term Energy Model (RSTEM) Overview
Table of Contents
1. Background
2. Petroleum Products Overview
A. Regional Gasoline Model
B. Regional Residential Heating Oil Model
C. Regional Residential Propane Model
3. Natural Gas
A. Regional Natural Gas Demand Model
B. Regional Natural Gas Supply and Pricing Model
4. Electricity
A. Regional Electricity Demand
B. Regional Electricity Supply and Prices
C. Hourly Load Profile Generator
D. Regional Electric Power Sector Fuel Cost Model
E. Electricity Dispatch Model
F. Regional Electricity End-Use Price Model
5. Macro Bridge
6. Appendices
A. Electricity Dispatch Model Calibration
RSTEM Documentation: Overview 1
1. Background
The Energy Information Administration's (EIA) Office of Energy Markets and End Use
(EMEU) is responsible for producing the monthly Short-Term Energy Outlook (STEO),
which forecasts U.S. supplies, demands, imports, stocks, and prices of various forms of
energy. In addition to the STEO, EIA also produces a Summer Motor Gasoline Outlook
(in April) May and a Winter Fuels Outlook (in October) as part of the short-term
forecasting program. EIA provides a public PC version of the model, PC-STEO, which
presents EIA's latest monthly national energy forecast in an Excel-like format. PC-
STEO allows includes a simulation engine to allow users to change assumptions and
generate alternative forecasts. A data query system is also available to search of data
series. Backing up the forecasting system is EIA's extensive energy data collection and
publication process.
STEO uses the Short-Term Integrated Forecasting (STIFS) model, which is an
integrated information system, bringing together energy quantities and prices from
various sources within EIA (and from elsewhere) in a consistent, time series format.
This energy information is coupled with other economic and non-economic information
to form a modeling database from which forecasting equations are estimated, saved
and later used to produce monthly projections and reports. Other models that run
outside the STIFS system are needed to generate some forecast information, such as
the macroeconomic forecasts.
STIFS consists of over 300 equations (excluding equations used to convert standard
units into energy equivalents such as British thermal units (Btu's)), of which just over
100 are estimated. The estimated regression equations form a system of interrelated
forecasting equations. The selection of functional form and the estimation technique is
generally done on an equation-by-equation basis. The general method of estimation is
ordinary least squares. Some equations incorporate a correction for autocorrelation of
the error term. STIFS model documentation is available on the STEO Website.
The STIFS model is almost entirely framed as a national-level model. While this feature
is adequate for portraying some of the more important near-term developments in major
fuel markets, it is limited in its ability to provide regional forecasts, which are of
particular interest to EIA customers. Consequently, EIA expanded the STIFS model to
include selected regional forecasts to provide greater geographic detail to the national
forecasts. The regional model allows for unique regional factors that affect energy
demands, supplies, and prices to be explicitly modeled: these factors are obscured in
an aggregated, national-level model. The new model, the Regional Short-Term Energy
Model (RSTEM), makes use of some of the structure contained in STIFS, but is more
complex due to the new regional detail.
One of the goals of the regionalization project is to provide a national forecast while
providing regional detail. The most detailed regional forecasts are in the natural gas
and electricity markets, partly because these markets tend to have strong regional
RSTEM Documentation: Overview 2
differences and have available regional data. However, considerable effort has been
made to provide regional forecasts for key petroleum products, as well. The only
regional consideration for coal demand is for the demand from the electric power sector,
although that is the bulk of the market. The same is true for renewables.
For example, the national outlook for heating fuel prices will be determined from
regional (largely Census Division) data and forecasts; users will be able to see how
regional prices and demand vary. Similarly, the national forecast for gasoline prices will
be determined from regional supplies and demand and users will be able access both
the national average price (and demand) along with regional prices and demands.
The Regional Short-Term Energy Model (RSTEM) utilizes estimated econometric
relationships for demand, inventories and prices to forecast energy market outcomes
across key sectors and selected regions throughout the United States. The structure of
the model is sufficiently detailed to allow for a richer and deeper treatment of the key
trends and forces in major domestic energy markets than was possible under STIFS.
The frequency of the model database remains monthly, and the forecast horizon
remains at 12 to 24 months ahead. Due to the limitations of data availability at this
frequency, the economic sectors covered for the energy demands and end-use prices
are somewhat limited.
Table 1 provides a summary of the coverage, by major fuel category, of the concepts in
RSTEM.
Table 1. STEM Model Regionalization Scheme
Regions Comments
Regionalized Components
Macroeconomic/Weather/Household
Macro Data/Projections 9 Census Divisions GI quarterly regional macro
model/U.S. bridge
Household Characteristics 9 Census Divisions RECS/Census/NEMS and
interpolations
Weather 9 Census Divisions CPC/NOAA
9 Census Divisions + NY, FLA, Based on EIA state-level sales and
Electricity Demand (Retail Sales) CA, TX and AK+HI revenue
Electricity Supply 9 Census Regions + NY, FLA, Generation and fuel consumption
CA, TX and AK+HI patterns will be modeled for all
Census Divisions, except that NY,
FLA, TX and CA will be treated
separately form the rest of their
respective Census Divisions.
Electricity imports will be determined
first at the national level, then
estimates shared to the supply
regions
RSTEM Documentation: Overview 3
Natural Gas Demand 9 Census Divisions Based on EIA state-level sales and
revenue
Natural Gas Supply National/Regional Hybrid National-level mechanism for
benchmark gas commodity price,
selected regions for basis
differential calculations. End-use
prices (including power sector
prices) will be at the Census
Division or power supply region
level. Storage will be handled at
the AGA regional level.
Coal Supply 3 Production Regions Provided Exogenously to RSTEM by
(Eastern, Interior, and Western) EIA's Office Coal, nuclear, Electric
and Alternate Fuels. Selected sub-
regions possible.
Petroleum Prices/Inventories 5 PADD Regions Gasoline, heating oil and propane.
Based on EIA'S PSM, PMM, and
price survey data and commercial
information on spot prices
National Components
Gasoline/Hwy Travel Demand
Jet Fuel Supply/Demand
Non-power Distillate Fuel Demand/Supply
Non-power Residual Fuel Demand/Supply
LPG Supply/Demand Balance
Other Petroleum Products Supply/Demand
Crude Oil Supply/Demand
Petroleum Products Import
Non-power Coal Demand
Non-retail Electricity Demand
Electricity Imports
Electricity Exports
Natural Gas Imports
Natural Gas Exports
Natural Gas Drilling/Production
2. Petroleum Products
A. Regional Gasoline Model
B. Regional Residential Heating Oil Model
Distillate fuel oil is consumed in several different sectors, including on-highway
transportation, residential, commercial, industrial, and agricultural (Figure 1). Other
RSTEM Documentation: Overview 4
applications include off-highway diesel, railroad, vessel bunkering, electric power, and
oil company use.
Distillate fuel oil is a general classification for one of the petroleum fractions produced in
crude oil distillation operations. First, distillate fuel is classified as No. 1, No. 2, or No. 4
fuel oil where the higher number denotes a heavier or more viscous liquid. Heating oil
used in the residential sector is primarily No. 2 distillate fuel oil (about 1.5 percent is No.
1 fuel oil). Second, distillate fuel is classified as diesel fuel or fuel oil. The most
significant distinction between diesel fuel and fuel oil is that diesel fuel has a maximum
allowed sulfur level of 500 parts per million or lower while fuel oil has a 5,000 parts per
million maximum. Consequently, diesel fuel can be used as fuel oil but fuel oil generally
cannot be used as diesel fuel. Heating oil used in the residential sector is fuel oil that
has the higher sulfur limit.
Heating oil ranks as the third most important source of residential energy in the Nation,
with nearly 8 percent of all households using heating oil as their primary space heating
fuel (Energy Information Administration, 2001 Residential Energy Consumption Survey,
Table HC1-9a). Heating oil is also used by households for water heating.
The objective of the regional short-term residential heating oil model is to generate
residential price forecasts for the four census districts: Northeast, South, Midwest, and
West (see Appendix A1 for map). Regional residential heating oil prices are estimated
as a function of the wholesale heating oil price, regional stocks, and regional demand
(Figure 2). Regional residential heating oil prices are then aggregated to the U.S. level
by weighting regional prices by estimated regional demands.
C. Regional Residential Propane Model
Propane is consumed in several different sectors, including residential, commercial,
petrochemical, industrial, and agricultural (Figure x). Other applications, which account
for the remainder of propane demand, include use as fuel in internal combustion
engines (generators, pumps, and fork lifts) and in gas utility peak-shaving. Propane
ranks as the fourth most important source of residential energy in the Nation, with nearly
5 percent of all households using propane as their primary space heating fuel (Energy
Information Administration, 2001 Residential Energy Consumption Survey, Table HC1-
9a). Propane is also used by households for water heating and cooking.
The objective of the regional short-term residential propane model is to generate
residential price forecasts for the four census districts: Northeast, South, Midwest, and
West. Regional residential propane prices are estimated as a function of the wholesale
propane price to the petrochemical sector, regional stocks, and regional demand
(Figure xx). Regional residential propane prices are then aggregated to the U.S. level
by weighting regional prices by estimated regional demands.
RSTEM Documentation: Overview 5
2. Natural Gas
A. Regional Natural Gas Demand
Natural gas demands and end-use prices are determined for 3 economic sectors
(residential, commercial and industrial), the other principal source of natural gas
consumption (i.e. for electric power) is determined in the Electricity Supply Model. The
natural gas portions of fuel demands for commercial and industrial electricity production
and cogeneration are implicitly included in the regional commercial and industrial
demand totals. Except for a small amount of liquefied natural gas (LNG) exported to
Japan from Alaska, natural gas exports are via pipeline to Mexico. In addition to current
domestic gas use and exports, demand for gas for domestic storage is another source
of demand for new gas supply in RSTEM. Regionality for natural gas storage is limited
to the three-region approach used in the weekly reporting of underground storage,
namely East consuming, West consuming, and West producing regions. Not all of the
important elements of the demand for natural gas storage are tracked in EIA data (or
any published aggregate data). Unlike liquids, natural gas can be compressed easily,
and some of the demand for gas in storage is satisfied by compressing natural gas in
the existing pipeline system, particularly as the period of peak demand (winter heating
season) begins. In the aggregate data, this phenomenon shows up as a sharp
widening in the excess of reported supply over reported demand (including net
underground storage builds) in the fourth quarter. Aside from actual pipeline additions,
which have to be filled in any case, it may be possible to avoid dealing directly with this
particular measurement problem by assuming that the extent of compression is
constant and assume that this particular source of demand may be captured in
seasonal factors.
Residential. The primary determinants of residential natural gas demand in a region
are: occupied housing units with natural gas hookups; share of occupied housing units
with natural gas hookups using natural gas as the primary energy source for heating;
heating degree-days; real delivered residential per-unit natural gas prices; real personal
income per household. Autonomous trends in household natural gas use intensity are
additional factors in residential demand related to residential building shell efficiencies
and penetration of more efficient natural gas appliances (furnaces, water heaters,
clothes dryers and ranges) in the household. In RSTEM, the net effect of any such
trends will be estimated econometrically.
RSTEM Documentation: Overview 6
Commercial. The primary determinants of commercial natural gas demand in a region
are: hours worked or real wage disbursements in commercial employment (as a proxy
for commercial sector output); heating degree-days; real commercial-sector per-unit
delivered natural gas prices; self-generating natural gas-fired capacity in the commercial
sector. Autonomous trends in commercial natural gas use intensity are additional
factors in commercial demand related to commercial building shell efficiencies, average
commercial floor-space per unit of output, and penetration of more efficient natural gas-
using equipment in commercial establishments. In RSTEM, the net effect of any such
trends will be estimated econometrically.
Industrial. The primary determinants of industrial natural gas demand in a region are:
hours worked or real wage disbursements in industrial employment or industrial output
as measured by the Federal Reserve Board; real industrial-sector per unit delivered
natural gas prices; heating degree-days; self-generating natural gas-fired capacity in the
industrial sector. Autonomous trends in industrial natural gas use intensity are
additional factors in industrial demand related to energy efficiency trends in industrial
processes and equipment, and shifts in regional industrial output patterns among
industry sectors of varying levels of natural gas use intensity per unit of output. In
RSTEM, the net effect of any such trends will be estimated econometrically.
Storage. The demand for natural gas storage (that is, net injections into underground
storage facilities) follows from the behavioral relationship describing the aggregate
market propensity to hold gas stocks under various seasonal and general market
conditions. In RSTEM, a system of equations determining the level of storage in three
general regions will be estimated jointly, and the resulting per-period addition to stocks
(positive during the injection season, negative during the withdrawal season) will be
added to the natural gas demand total for each time period. In general, natural gas
storage levels exhibit very regular seasonal patterns and have generally understood
seasonal shifts in variances.
B. Regional Natural Gas Supply
We assume that the U.S. natural gas market is sufficiently integrated geographically so
that a key market location (the Henry Hub in Louisiana) provides price information that
is sufficiently representative of the overall natural gas market supply/demand balance in
the United States that it may be used as the benchmark price from which all major
regional price levels and deviations may be inferred. Additionally, we assume that
regional price differentials relative to the benchmark price are generally stable and
reasonably predictable, given some region-specific market characteristics (weather,
excess pipeline capacity, etc.). This assumption allows us to calculate regional spot
prices for natural gas, and from these, region-specific wholesale and end-use prices for
inclusion in regional natural gas demand estimation.
Natural gas supply (domestic production of dry natural gas and imports of gas from
Canada plus LNG imports from various international sources plus supplemental
gaseous fuels) is first treated at the national level in a procedure that determines an
RSTEM Documentation: Overview 7
equilibrium price for a key benchmark natural gas spot price: the spot price at the Henry
Hub in Louisiana. The Henry Hub price is determined by equilibrating aggregate
demand for natural gas (sum of the regional sectoral demands plus natural gas used in
the electric power sector plus lease and plant fuel use plus pipeline gas plus gross
exports plus storage demand) with aggregate supply.
Then, the model determines regional natural gas spot price differentials (for how many
regions) relative to the Henry Hub price, taking into account systematic historical and
seasonal patterns in the differentials as well as any regional supply factors that may be
useful in improving the overall accuracy of the estimating procedure.
The resulting regional spot prices are used to generate city gate prices for the demand
regions (how many demand regions) or may be used directly as proxies for market
prices (such as in the industrial sector) and used as determinants of delivered prices of
natural gas to the electric power sector by electricity supply region. The regional
natural gas price determinations will feed back into the national level supply-price
calculations by affecting regional demands which will feed back to the aggregate supply
balance, and so on.
The basic natural gas market equilibrium condition in RSTEM requires that the Henry
Hub price be such that supply equals demand for the national market as a whole. The
general approach in RSTEM is to obtain initial estimates of the Henry Hub spot price
from an appropriately derived reduced form equation for the price, calculate the
supply/demand imbalance from the demand and supply components in RSTEM and
apply the implied net supply elasticity from the system to the percent imbalance in such
a way that the spot price is adjusted to eliminate the excess supply. The system is
generally not amenable to direct analytical solutions for the equilibrium price but
iterative application of the equilibrating procedure can be (and is) utilized in RSTEM.
Drilling, Well Completions and Productive Capacity. RSTEM characterizes industry
efforts at natural gas productive capacity development and utilization (wells drilled and
dry gas production) at the national level by utilizing estimating relationships for active
rigs drilling for natural gas (as a function of oil and gas revenues, spot natural gas
prices, and other factors), well completions (as a function of rig efficiency trends, lagged
relationship to rig activity, etc.), productive capacity (lagged relationship to well
completions, cumulative production, and other factors) and capacity utilization or
production (as a function of capacity, current and lagged spot prices, and other factors).
Imports. Natural gas imports are assumed to be a function of current and lagged spot
prices, import capacity (separately for LNG and Canadian (and possibly Mexican)
pipeline imports), and other factors.
Supplemental Natural Gas Supply. Supplemental gaseous fuels supply, also
included in the total supply for natural gas, is a relatively small component that consists
of synthetic gas, refinery gas and some other components. It is assumed that this
source of supply is generally stable except for some regular seasonal components.
RSTEM Documentation: Overview 8
4. Electricity
In RSTEM, the greatest overall level of detail is provided for the electricity market.
Electricity demands and end-use prices are determined for 4 economic sectors
(residential, commercial, industrial, and other2) for 14 regions (9 Census Divisions plus
a separate combined Hawaii and Alaska aggregation plus the four individual States of
California, Florida, New York and Texas. Electricity supply (net generation by fuel
source and energy consumption to produce electricity by fuel) is determined for the
same 14 regions indicated for demand. Industrial-sector and commercial-sector
production of electricity (and related fuel consumption) are determined for the 14
regions but are not expected to be a part of the main procedure for determining
electricity output for the electric power sector.
A. Regional Electricity Demand
Residential. The primary determinants of residential electricity demand in a region are:
occupied housing units or households; share of occupied housing units or households
using electricity as the primary energy source for heating; share of occupied housing
units or households with installed air conditioning; cooling degree-days; heating degree-
days; real delivered residential per-unit electricity charges; real personal income per
household. Autonomous trends in household electricity use intensity are additional
factors in residential demand related to residential building shell efficiencies and
penetration of electricity-using appliances in the household. In RSTEM, the net effect of
any such trends will be estimated econometrically.
Commercial. The primary determinants of commercial electricity demand in a region
are: hours worked or real wage disbursements in commercial employment (as a proxy
for commercial sector output); cooling degree-days; heating degree-days; real
commercial-sector per-unit delivered electricity charges; self-generating capacity in the
commercial sector. Autonomous trends in commercial electricity use intensity are
additional factors in commercial demand related to commercial building shell
efficiencies, average commercial floor-space per unit of output, and penetration of
electricity-using equipment in commercial establishments. In RSTEM, the net effect of
any such trends will be estimated econometrically.
Industrial. The primary determinants of industrial electricity demand in a region are:
hours worked or real wage disbursements in industrial employment or industrial output
2
Beginning with 2003 data, EIA discontinued the reporting of the "Other" electricity sales sector and required
respondents to allocate the information previously included in this category to the commercial, industrial and
transportation sectors. Information on transportation sector electricity sales has, since 2004, been collected on a
monthly basis. For electricity demand modeling in RSTEM, the old-basis sectors are maintained since the longer
history is available that way, although a restatement to the new basis information is included and the reported
information from RSTEM will normally be done on the new basis so as to match EIA's current publication
standards. Once enough history is available for the new-basis categories, the RSTEM electricity demand equations
will be restructured accordingly.
RSTEM Documentation: Overview 9
as measured by the Federal Reserve Board; real industrial-sector per unit delivered
electricity charges; cooling degree-days; heating degree-days; self-generating capacity
in the industrial sector. Autonomous trends in industrial electricity use intensity are
additional factors in industrial demand related to energy efficiency trends in industrial
processes and equipment, shifts in regional industrial output patterns among industry
sectors of varying levels of electricity use intensity per unit of output, and penetration of
general electricity-using equipment in industrial establishments. In RSTEM, the net
effect of any such trends will be estimated econometrically.
Other. Other electricity demand in a region, which consists of electricity sales not
designated as residential, commercial or industrial (such as municipal lighting and other
general service as well as transportation), is assumed to be determined by general
growth of economic activity in the region, as measured by gross regional product or
other aggregate activity measures.
The demands for electricity relate to retail sales, or sales distributed by local distribution
companies, either for own account or for the account of independent service providers.
Two other demand components are: electricity generated by and consumed by an entity
or facility, such as an industrial establishment (with either combined heat and power
(CHP) or electric-only generating capability) or a commercial facility such as a
University or other non-industrial facility; electricity generated by one entity and
delivered directly to another entity, by-passing retail distribution. The bulk of such non-
retail demand is in the industrial sector (approximately X percent in 2002). In RSTEM,
the non-retail component of electricity demand is treated separately and at the national
level.
B. Hourly Load Profile Generator
C. Regional Electricity Supply and Prices
In RSTEM, electricity supply (net generation) is calculated for 14 regions, based on the
9 U.S. Census Divisions plus four States, namely California, New York, Florida and
Texas. Nuclear and hydroelectric generation (by regions) are determined outside of
STIFS and taken as exogenous. Historical patterns and recent trends are used to
estimate non fossil fuel based generation sources other than hydroelectric and nuclear.
A detailed dispatch model will be used to determine fossil fuel-based generation (coal,
oil, natural gas) by region.
D. Electricity Dispatch Model
The dispatch module predicts the composition of electricity supply within regions in the
United States for the forecasting horizon (18-24 months). The dispatch model uses
information on all the power generators in the United States to calculate the variable
cost of operating for each power generator. The generators are then sorted, within a
given region, from least-cost to highest-cost and are dispatched in that order based on
the demand for electricity in any given period.
RSTEM Documentation: Overview 10
The model allows for trading between regions, as long as the regions are within the
same interconnection. This is done by, rather than sorting all generation facilities within
a given region, sorting all generating facilities within a given interconnection, or
essentially adding the regional supply curves within each interconnection. The demand
curves are also added, coming up with supply and demand curves for the whole
interconnection. Dispatching decisions are then made based on the variable costs of all
facilities within that interconnection. Each facility is identified with its fuel type and the
region in which it is located, so once dispatching decisions are made, the model
indicates how much and what types of generation were dispatched from each region.
Additional modifications to this scenario are made to acknowledge transmission,
environmental, and engineering constraints.
E. Regional Electricity End-Use Prices
End use prices for electricity may not have a particularly straightforward relationship to
wholesale or spot prices for the underlying commodity. Retail electric rates are still
heavily regulated at the State level (particularly for the residential sector), and average
costs of production and the particular requirements of State regulators are generally
more important in the determination of the prices that consumers see than marginal
supply costs, which may or may not be presumed to be reasonably well represented in
spot electricity prices. Thus, consumers in areas that have an abundance of low cost
hydroelectric output may see much lower electricity prices than other regions, even if
marginal supply costs are roughly equivalent. Thus, much of the analysis needed to
generate good estimating equations for the electricity prices for the electricity demand
regions will focus on the particular patterns of historical average price behavior from the
time series in those regions rather than on marginal supply costs in the relevant
electricity supply regions. Nevertheless, an attempt to relate marginal electricity
production costs to spot electricity price patterns should be made. In future versions of
RSTEM, regional spot electricity prices will be considered as factors, along with average
fuel costs in the supply regions and other variables as determinants of retail electricity
prices.
F. Regional Electric Power Sector Fuel Cost Module
The Regional Power Sector Fuel Cost Module (RSTNGSM) of the Regional Short-Term
Energy Model (RSTEM) provides a procedure for calculating short-term delivered
average fuel costs to the electric power sector (electric utilities and independent power
producers) by RSTEM electricity supply region. In RSTEM, there are 13 electricity
supply regions in the Lower-48 States plus a separate accounting for the combined
activities of Alaska and Hawaii. Fuel costs by region are calculated only for the Lower-
48. Four categories of average monthly fuel cost are tracked, including: coal; natural
gas; residual fuel oil; and distillate fuel. Regional and national average natural gas and
fuel oil spot and wholesale prices determined in other RSTEM modules are used as key
determinants of the regional power sector delivered costs
RSTEM Documentation: Overview 11
As with the rest of RSTEM, the RSTPSFCM is designed to generate monthly forecasts,
in this case of per-unit average costs for fossil fuels in dollars per million Btu delivered
by 13 electricity supply regions.
The power sector fossil fuel costs used in RSTEM are taken from FERC Form 423 and
EIA Form 423, "Cost and Quality of Fuels Survey(?)." Costs associated with deliveries
to independent power producers are proprietary, and would not be reportable for some
of the electricity supply regions for some periods. For RSTPSFCM, only average costs
for deliveries to electric utilities are used.
G. Regional Electricity Generating Capacity
For RSTEM, generating capacity by fuel by region will be predetermined, based on
capacity reported in EIA's Form EIA-860, with recent historical values and forecasts
provided by utilizing EIA tracking of recent additions and planned additions. Capacity
tracking and projections by fuel by region is an important part of the regional short-term
forecasting program, and will require coordination between the Office of Energy Markets
and End Use, the Office of Integrated Analysis and Forecasting and the Office of Coal,
Nuclear, Electric and Alternate Fuels to insure consistency with forecasts and
assumptions about capacity reported in all EIA Offices.
5. Macro Bridge
The RSTEM Macro Bridge is designed to address the problem of maintaining regional
macroeconomic forecasts (which are only available on a quarterly basis) consistent with
monthly national macroeconomic forecasts, the latter of which are to serve as the basis
for EIA's assumptions about growth in aggregate output, income and employment for its
monthly model simulations used in the Short-Term Energy Outlook. Both the national
and regional macroeconomic forecasts are supplied by Global Insight (GI). Once each
quarter, the baseline forecasts for both the regional and national macroeconomic
forecasts are expected to be entirely consistent. For interim monthly forecasts,
however, a procedure is required to adjust the quarterly regional forecasts so that they
reflect aggregate economic activity that is consistent with the monthly national
forecasts. The Macro Bridge program utilizes simple scaling routines that align/and
update GI regional macroeconomic data and forecasts with monthly macroeconomic
data and forecast updates from the GI quarterly model of the U.S. economy.
6. Appendices
A. Electricity Dispatch Model Calibration
The generic nature of the regional dispatching model allows it to utilize all the relevant
EIA electricity data and to capture historical dispatching patterns. This chapter analyzes
available data at generator level and put them in a framework that shed lights on
RSTEM Documentation: Overview 12
operators' dispatching behavior. A systematic calibration process then uses the
findings to modify the dispatching order of power plants in each region. The short-term
energy outlook makes 24 monthly projections from the date of publication. Both
forecasting period and time interval are short, therefore, the regional electricity model
uses only near term monthly data, 2002 and 2003 data for model calibration.
RSTEM Data Flow and Modeling Process
Regional Regional EIA Regional EIA Regional W orld Oil Other
Macro Data/Forecasts W eather Energy Quantity Energy Price/Cost Price Input
(G I) (NOAA) Data Data Assum ptions Data
Regional Data Prep Econometric Model
Estim ation Petroleum Dem and
Program
Petroleum Supply
Petroleum Prices: m otor gasoline,
heating oil and propane
Natural Gas Demand: regional end
use
Update Econometric
Natural Gas Prices: regional spot,
Model Inputs
citygate, end use
Electricity Demand: regional end
use
Electricity Electricity Prices: regional end use
Load Parameters Power Sector Fuel Cost: regional
Electricity coal, dist., residual, and natural
Generation by Source gas
Dispatch Coal Dem and
EIA Electric
M odel Econometric Coal S upply
Generating M odel Forecasts Other
Capacity ts
Data C os
u el
Lo d ,F
ad
,F L oa
ue
lC
os Convergence Check
ts
Save Final Outputs,
No Yes Generate Reports
RSTEM Documentation: Overview 13