Users' Guide for GLIMPSE:
a Tool for Integrated Air-
Climate-Energy Planning

Version 1.0

Office of Research and Development

EPA600/B-23/125 | June 2023 | www.epa.gov/research


-------
» United States
tMli Environmental Protection
^#^1 #* Agency

June 2023 | www.epa.gov

Users' Guide for GLIMPSE:
a Tool for Integrated Air-
Climate-Energy Planning

Version 1.0

Office of Research and Development

1

EPA/600/B-23/125


-------
ACKNOWLEDGMENTS

GLIMPSE is an acronym for the "GCAM Long-Term Interactive Multi-Pollutant Scenario
Evaluator", a decision support system developed at the U.S. EPA Office of Research and
Development. GCAM is the "Global Change Analysis Model", an open-source Human-Earth
Systems model. GCAM development is led by Pacific Northwest National Laboratory (PNNL).

The core EPA GLIMPSE Team consists of EPA's Dan Loughlin, Chris Nolte, Carol Lenox, Joyce
Kim, and Fahim Sidi, as well as ORAU student services contractor Sarah Simms. Additional EPA
contributors and collaborators have included Ozge Kaplan, Michael Shell, Jeff Cole, Julien
Isnard, Chris Ramig, Colby Tucker, Shutsu Wong, and Shannon Koplitz. Former team members
include Tai Wu (EPA) and Samaneh Babaee, Wenjing Shi, Paelina DeStephano, Fanqi Jia, Farid
Alborzi, Michael Wu, and Yang Ou (ORISE). We also acknowledge the important contributions
provided by the EPA managers and staff who assisted with quality assurance, communications,
data management, computing, and review.

We appreciate the review of this documentation by Joyce Kim and Julien Isnard, as well as by
EPA's Heidi Paulsen and Paula Gomez of the State of Connecticut's Department of Energy and
Environmental Protection. We also appreciate the external reviewers who served as peer
reviewers and beta testers for this release.

For the GLIMPSE project, GCAM development, data processing, and support for policy
implementations has been led by Dr. Steven J. Smith of PNNL, via Interagency Agreements 89-
92423101 and 89-92549601. Contributors from PNNL include Maridee Weber, Catherine Ledna,
Gokul Iyer, Page Kyle, Marshall Wise, Matthew Binstead, and Pralit Patel.

This work has benefited from many interactions with our EPA Program Office and Regional
Office colleagues, as well as via conversations with state air quality, climate, and energy staff.
We also acknowledge the valuable testing performed by students within the Integrated
Assessment Modeling course within Duke University's Nicholas School of the Environment.

DISCLAIMER

The views and information presented in this document represent those of the authors and do
not necessarily reflect those of the Agency or other parties. The mention of specific software or
institutions does not convey endorsement. No warranty is provided regarding the functionality
or results provided by GLIMPSE and GCAM.

2


-------
Citation: Loughlin, D.H. and C.G. Nolte. "Users' Guide for GLIMPSE: a Tool for Integrated Air-
Climate-Energy Planning" EPA/600/B-23/125, U.S. Environmental Protection Agency,
Washington, DC, USA. June 2023.

3


-------
CONTENTS

ACKNOWLEDGMENTS	2

DISCLAIMER	2

INSTALLATION INSTRUCTIONS	8

System requirements	8

Installation steps	8

Step 1: Download "GLIMPSE-vl.zip" from the link that is provided by EPA	8

Step 2: Unzip "GLIMPSE-vl.zip" to install the GLIMPSE package	8

Step 3: Test the GLIMPSE software	9

Next steps	12

CHAPTER 1. OVERVIEW	14

1.1	Background	14

1.2	Motivation	16

1.3	GCAM-USA	21

1.4	Components of GLIMPSE	24

1.5	Design philosophy	24

1.6	Computer and software requirements	29

1.7	Organization of this Users' Guide	29

1.8	GLIMPSE Version	29

1.9	A note on units	30

1.10	Interpreting GCAM results	30

1.11	Known bugs, limitations, and other considerations	32

1.12	Where to learn more about GCAM and GLIMPSE	34

1.13	How to get started using GLIMPSE	35

1.14	Where to get assistance or provide feedback	36

CHAPTER 2. GLIMPSE REFERENCE SCENARIO	36

2.1	Introduction	36

2.2	The GCAM-USA 5.4 Reference Scenario	37

2.3	The GLIMPSE Reference Scenario, GUMPSEvl-Reference	38

4


-------
2.4	Selected results for the GLIMPSE Reference Scenario	45

2.5	Evaluation of GUMPSEvl-Reference	84

2.6	Discussion and Summary	94

CHAPTER 3. HOW DOES GCAM WORK?	96

3.1	General solution process	96

3.2	Markets in GCAM	96

3.3	Determining market share	98

3.4	Calibration	99

3.5	Relaxing shareweights over time	99

3.6	Illustrating the operation of the "Car" market	100

3.7	Markets, Logits, Shareweights throughout GCAM	109

3.8	The logit function and market share constraints	Ill

CHAPTER 4. HOW DO I....?	112

4.1	How do I change options in the GLIMPSE options files?	112

4.2	How do I update the JAVA_HOME environmental variable?	113

4.3	How do I create a new database?	113

4.4	How do I import and export scenarios from the Mode/Interface?	115

4.5	How do I manage database size?	117

4.6	How do I archive scenarios in GLIMPSE?	117

4.7	How do I access and interpret the mainjog file?	118

4.8	How do I edit a scenario's configuration file?	118

4.9	How do I import Scenarios into GLIMPSE?	119

4.10	How do I import files into the Component Library?	119

4.11	How do I recover deleted Scenario Components and Scenarios?	119

4.12	How do I save files associated with each run?	120

4.13	How do I clean up saved files?	120

4.14	How do I use the CSVtoXML utility?	120

4.15	How do I know if my computer's resources are running low?	121

4.16 How do I determine why a GCAM run did not complete and GLIMPSE reports "DNF"? 122

5


-------
4.17 How do I access the GCAM data system and find the original sources of data?	123

CHAPTER 5. ADVANCED TOPICS	125

5.1	What is happening behind the scenes in GLIMPSE...?	125

5.2	GLIMPSE folder structure	126

5.3	Interpreting and debugging unsolved market information in the main_log.txt file	128

5.4	Biomass flows and CO2 accounting in GCAM-USA	131

5.5	Considerations when modeling a deep decarbonization scenario	132

CHAPTER 6. REFERENCE	136

6.1	Buttons and menu options on the Scenario Builder	136

6.2	Scenario Components	144

6.3	GLIMPSE queries in the Mode/Interface	161

6.4	Troubleshooting	169

6.5	Glossary	175

APPENDIX: TUTORIALS	1

TUTORIAL 1: RUNNING GCAM THROUGH GLIMPSE	2

Tl.l Overview	2

T1.2 Opening the GLIMPSE software	2

T1.3 Executing the GLIMPSE Reference Scenario	7

T1.4 Examining information saved with each run	11

TUTORIAL 2: EXAMINING MODEL RESULTS	13

T2.1 Overview	13

T2.2 Viewing model results with the Model I interface	13

T2.3 Analyzing model results outside of the Model I interface	17

T2.4 Using the Mode/Interface to visualize model results	18

T2.4 Comparing results across scenarios	22

T2.5 Additional suggestions for exploration	28

TUTORIAL 3: MODELING AND EVALUATING A CARBON TAX	30

T3.1 Overview	30

T3.2 Constructing Carbon tax Scenario Component	30

6


-------
T3.3 Exploring the response to the tax	39

TUTORIAL 4: ADDITIONAL TOOLS FOR COMPARING SCENARIOS	47

T4.1 Overview	47

T4.2 Using the Diff Query	47

T4.3 Additional analysis suggestions	51

TUTORIAL 5: MODELING AN EV MARKET SHARE TARGET	52

T5.1 Overview	52

T5.2 Constructing the EV sales target components	52

T5.3 Verifying the performance of the policy	57

T5.4 Exploring the response to the EV market share target	59

7


-------
INSTALLATION INSTRUCTIONS

System requirements

Currently GLIMPSE runs only on Windows computers. At least 14 GB of RAM are required for
GCAM-USA, and 16 GB or more are recommended. The hard drive should have at least 80 GB
of free space. Typical runtime is 30 minutes to 5 hours, depending on computational power,
memory, and the complexity of the scenarios being simulated.

Installation steps

Please read these instructions in full before installing.

Step 1: Download "GLIMPSE-vl.zip" from the link that is provided by EPA.

The GLIMPSE package is approximately 1 GB in its zipped form.

Step 2: Unzip "GLIMPSE-vl.zip" to install the GLIMPSE package.

Unzipped GLIMPSE-vl.zip will be approximately 7 GB in size.

When selecting a location, it is suggested that you choose a hard disk with fast read/write
speeds and at least several hundred free GB of storage space. Furthermore, since GCAM
generates 1 to 2 GB of output with every execution, it is advisable to select a location that is not
automatically backed up. For example, some users have experienced difficulties when installing
GLIMPSE on OneDrive. Finally, please select a location without spaces in the names of folders
since these can result in errors in filename parsing within GLIMPSE.

One option is to create the following folders and install GLIMPSE there:

"C:\Users\INSERT_USERNAME\Documents\local_folder\GLIMPSE-vl"

For the purpose of these instructions, GLIMPSE is being installed in the following location:

"E :\P rojects\G LI M PSE-vl"

This folder should now include the contents shown in the image below.

8


-------
i I

B

P 9

c ^ I GLIMPSE-v1



~

X ,

File



Home

Share View





v 0

<-

->

v ^

P Data (E:) > Projects > GLIMPSE-v1

v O

P Search GLIMPSE-v1



P amazon-coireto-8.372.07.1-windows-x64-jre	P Contrib

P Docs	P GCAM-Model

P GLIMPSE-Data	P GLIMPSE-GUI

P ORDModellnterface	B options_GCAM-USA-5p4.txt
Q run_GLIMPSE_GCAM-USA-5p4.bat

9 items I	I 3

Figure Al. The GLIMPSE folder.

When unzipping, please note that various unzip programs treat the root folder of the zipped file
differently, and a common installation issue is the nesting of folders:

"E:\Projects\GLIMPSE-vl\GLIMPSE-vl"

If this is the case in your installation, please move the contents of the nested GLIMPSE-vl folder
(E:\Projects\GLIMPSE-vl\GLIMPSE-vl) to the parent folder (E:\Projects\GLIMPSE-vl). You can
then delete the nested folder (E:\Projects\GLIMPSE-vl\GLIMPSE-vl).

Step 3: Test the GLIMPSE software.

To test your installation, start GLIMPSE by double-clicking on "run_GLIMPSE_GCAM-
USA_5p4.bat". A black system window (also known as the Cmd.exe or DOS window) will
appear, and text printed to that window will indicate that options are being loaded from the
"options_GCAM-USA-5p4.txt" file. Note: additional diagnostic information will be output to this
window during your GLIMPSE session. You can generally ignore this information; however, if
you have issues with GLIMPSE or GCAM, it may be useful for debugging. Closing the Cmd.exe
window will terminate your GLIMPSE session.

9


-------
Several seconds after the Cmd.exe window appears, the GLIMPSE Scenario Builder wili appear.
If the GLIMPSE window does not appear, a common cause is that Windows is preventing
execution of the "run_GLIMPSE_GCAM-USA-5p4.bat" file since execution of ".bat" files can be a
security issue. To signal to Windows that it is OK to allow this file to be run, right-click on the
file, choose "properties", and then check the box next to "unblock execution". Then try double-
clicking on the file again.

Once GLIMPSE appears, the Component Library in the top left pane should include a number of
scenario components, and there should be at least one scenario listed in the Scenario Library
table at the bottom. If there are no scenario components, this indicates that the installation
was not successful, and that GLIMPSE cannot find the correct folder.

Q GLIMPSE Scenario Builder
File Tools View Help

Component Library Search:

-1- ^ b x loi

Component Name

Created



Calib-biomass_constraints.txt

2022-07-18:19:19

A

Calib-coal_egu_2020.csv

2022-07-18:19:19



Calib-coal egu_2020.txt

2022-07-18:19:19



Calib-coal_egu_2020b.csv

2022-07-18:19:19



Calib-H DV-BEV-SW-1 X50.CSV

2022-07-22: 20:25



Calib-LDV-EV-AE02020.csv

2022-07-18:19:19



< j

>



Create Scenario

Component Name

No content in table

©

Scenario Library

Scenario Name

GCAM5p4-Ref-Orig
GLIMPSE-Reference

Search:

7





>

X

~~~ mil

V



!



El O

Created

2022-07-19: 08:05

2022-07-25:10:55

Completed

2022-07-19:10:26

Status	ProbMkts	Runtime

Success	3441.29 s

Figure A2. The Scenario Builder.

Next, click on "Tools->Check Installation" from the main menu bar of the Scenario Builder.
When you do this, GLIMPSE analyzes the options that were loaded from the "options_GCAM-

10


-------
USA-5p4.txt" file, noting when critical options have not been specified or where folders
specified in the file do not exist.

Then, GLIMPSE checks for a frequent problem when installing updates to GLIMPSE that leads to
an incorrect folder structure. Finally, GLIMPSE checks the JAVA_HOME environmental variable
that was specified in "run_GLIMPSE_GCAM-USA_5p4.bat" to ensure that this version of Java is
on the system. The results of these checks are reported to a popup window. If all checks are
successful, the text will include: "Installation appears to be correct."

JO Analysis of GLIMPSE setup	— ~ X

	Analysis of GLIMPSE setup	

No problems found with parameters or folders.

No problem was found with nested GLIMPSE folders.

Your JAVA_HOME folder, E:\Projects\GLIMPSE-v1Wamazon-correto-8.372.07.1-windows-x64-jre, was successfully found.
Installation at location E:\Projects\GLIMPSE v1 appears to be succesful.

Check to verify that key files exist as specified
XML header file: E:\Projects\GLIMPSE-v1\GLIMPSE-GUI\templates\glimpseXMLHeaders.txt - true
Tech Bound file: E:\Projects\GLIMPSE v1\GLIMPSE-GUI\templates\tech_bnd_list_usa_5p4.txt true
Configuration template file: E:\Projects\GLIMPSE-v1\GLIMPSE-GUI\templates\configuration_usa_5p4_template.xml - true
Query file: E:\Projects\GLIMPSE v1\GLIMPSE Modellnterface\exe\Main_queries_GLIMPSE 5p4.xml true
GCAM executable: E:\Projects\GLIMPSE-v1\GCAM-Model\gcam-v5.4\exe\gcam.exe - true
Model Interface executable: E:\Projects\GLIMPSE v1\GLIMPSEModellnterface\exe\ORDModellnterface.jar true

	Computer Information	

— Memory analysis —

Total physical memory: 31.9 GB

Free physical memory: 20.8 GB

-- Disk space analysis —

Total space: 4657.4 GB
Free space: 3001.7 GB
Total swap space: 63.9 GB
Free swap space: 47.5 GB

-- Processor analysis —

Available processor cores: 8
Current usage: 10.2%

As a last step, click on the "results" button, ADD , that is located on the button bar in the
Scenario Library area. After several seconds, the GLIMPSE Mode/Interface should appear. The
"Scenarios" pane at the top left should include at least one scenario.

Close

Figure A3. Output from the Check Installation option.

11


-------
l::\Projerts\GLIMPSE-5p4\GCAM-Model\gcam-v5.4\output\database] - Model Interface — ~
File Edrt Table Help	

Scenario

Regions





Queries

GC AM5p4-R ef-Orig 2022-1



USA a







A





Africa Eastern



&

GLIMPSE





Africa Northern



0

Primary and final energy





Africa Southern







. Primary energy consumption by region (direct equivalent)





Africa Western







~ « Final energy consumption by region





Australia NZ







* Final energy consumption by aggregate sector





Brazil







Final energy consumption by aggregate sector and fuel





Canada v







ai Final energy consumption by sector and Fuel v

<



<



RunQuer> Dtff Query Q Total Collapse Update Single Queries Create Remove Edir

Figure A4. The GLIMPSE Mode/Interface.

Step 4 (Optional): Configure GLIMPSE to use specific text and XML editors

By default, GLIMPSE is configured to use Window's Notepad application to open text and XML
files. Alternatively, you can specify that GLIMPSE use different applications to open these files.
For example, Notepad++ is an open-source text editor that will automatically color-code and
format XML. files. To change the specified editors, open the "options_GCAM-USA-5p4.txt" file
and find the lines starting with "textEditor" and "xm I Editor". Change these to refer to your
preferred applications. If you have installed Notepad++ on your computer, you can also remove
the comment symbol, #, from "#textEditor" and "#xmlEditor" and add it to the start of the prior
lines.

This change will not take effect until you either restart GLIMPSE or choose "File->Reload
Options" from the main pulldown menu of the Scenario Builder.

12


-------
Next steps

At this point, GLIMPSE should be set up correctly on your computer. We recommend reading
"Chapter 1: Overview" and "Chapter 2: GLIMPSE Reference Scenario", then going through the
Tutorials that are provided in the Appendix.

If GLIMPSE does not start or you run into additional issues that are not described in this
Installation Guide, please see the Troubleshooting Section of the Users' Guide. If your problem
cannot be resolved, please contact Dan Loughlin at Loughlin.Dan@epa.gov with a detailed
description of your problem and a summary of the approaches you have taken to solve it.

13


-------
CHAPTER 1. OVERVIEW

GLIMPSE is a decision support tool being developed at the U.S. EPA to assist the EPA, states,
and others with long-term environmental and energy planning. GLIMPSE is an acronym for
the "GCAM Long-Term Interactive Multi-Pollutant Scenario Evaluator", where GCAM is
the "Global Change Analysis Model." GCAM, a human-Earth systems model developed by
Pacific Northwest National Laboratory (PNNL), simulates the co-evolution of the economy,
energy system, land use, and climate systems, including how this co-evolution is shaped by
policy and other external factors. GLIMPSE acts as a graphical user interface for GCAM.

Using GLIMPSE, decisionmakers and analysts at the national, regional, and state levels can
examine potential policies, investigate the impacts of emerging technologies, develop cost-
effective strategies for meeting air pollutant and greenhouse gas (GHG) mitigation targets, and
explore the tradeoffs at the nexus of energy, water, and land use. Additionally, GLIMPSE can be
used in a classroom setting, providing students with the ability to use a state-of-the-art human-
Earth systems model to investigate alternative scenarios of the future.

1.1 Background

GLIMPSE can be used with GCAM or with variants of GCAM that have additional spatial
resolution, such as GCAM-USA, which represents the U.S. energy system at the state level.

Among the attributes of GCAM (and GCAM-USA) that led to its inclusion in GLIMPSE are:

•	spatial coverage and resolution (global, with available state-level resolution for the U.S.),
allowing examination of national and state actions in a global context,

•	temporal range and resolution (2010-2100 in 5-year increments by default), supporting
long-term air-climate-energy planning,

•	runtime of one to several hours, depending on which policies are included in the simulation,
facilitating sensitivity and scenario analyses,

•	input and output formats that are amenable to integration with a user interface,

•	emission outputs including both greenhouse gases (GHGs) (CO2, CH4, N2O, HFCs and CFCs)
and traditional air pollutants (NOx, SO2, CO, PM2.5, VOCs, and NH3), covering major
pollutants of concern in the US,

•	characterization of water supply and demand across sectors, allowing investigation of
water-energy-land-agricultural dynamics in the context of a changing climate,

•	no requirement for specialized hardware or proprietary software, thus lowering the barriers
for adoption,

14


-------
•	the GCAM source code and data, both of which are regularly updated, existing in the public
domain and freely available, promoting transparency, and,

•	the model's primary developers, the Joint Global Change Research Institute (JGCRI) of PNNL,
provide model documentation, have helped cultivate a broad user community, and hold an
annual modelers' workshop.

JGCRI makes versions of GCAM and GCAM-USA available via GitHub

(https://github.com/JGCRI/gcam-core). Documentation (http://igcri.github.io/gcam-doc/) is
also available, and tutorials and training are provided at annual GCAM Community Modeling
Meetings (https://gcims.pnnl.gov/community).

We have adopted a publicly available version of GCAM-USA 5.4 for use in GLIMPSE and
have worked with PNNL researchers to modify that version by adding air pollutant emission
factors, updating technology attributes in the transportation and power sectors,
and incorporating key U.S. air quality and energy policies.

Scenario
assumptions

GCAM

Population
growth

Economic
growth

Energy

Agriculture

Economy

Simulates the co-evolution of these systems through time

Outputs

Energy

Technology penetrations

Fuel lss and prices

Economic

Cost of energy services

Land and food prices

	Climate	

GHG emissions
Global mean temperature

Environmental

Air pollutant emissions

Health impacts

Figure 1.1 GCAM inputs, outputs, and major components. GCAM includes representations of
energy, water, land use, agricultural, and climate systems, simulating their co-evolution.

15


-------
1.2 Motivation

1.2.1 The energy system

Understanding energy system concepts and terminology is important for understanding how

GLIMPSE and GCAM-USA can be used to explore GHG mitigation strategies, air pollution control

strategies, and strategies for meeting climate and air quality goals simultaneously.

In the context of GLIMPSE, the term "energy system" refers to all processes and fuels that

extend from:

•	importing or extracting raw forms of energy (e.g., crude oil, coal, natural gas, uranium, and
wind),

•	converting (e.g., in refineries and power plants) those raw forms of energy into useful forms
of energy (e.g., gasoline, diesel, and electricity), and

•	applying useful energy to meet end-use energy service demands (e.g., passenger and freight
travel, space conditioning, water heating, and lighting).

The figure below shows a depiction of these components:

Primary Processing and conversion of energy carriers Final energy
energy

Wind, Solar,
Hydro

Direct Electricity
	Generation	

Industry

Figure 1.2 Schematic of the energy system. The energy system extends from the import or
extraction of primary energy, through its processing and conversion into useful forms,
through its use in meeting final end-use energy demands.

16


-------
Several important terms used when describing the energy system are defined below:

•	Primary, secondary, and tertiary energy - Primary energy is in the raw form in which it is
first accounted for in a statistical energy balance before any transformation to secondary or
tertiary forms occurs. For example, coal can be converted to synthetic gas, which can be
converted to electricity; coal is primary energy, synthetic gas is secondary energy, and
electricity is tertiary energy. (Source: EIA)

•	Final energy - The energy that is consumed by end-users, including for transportation,
residences, commercial buildings, and industry. Examples include electricity, gasoline, and
natural gas.

•	Useful energy - The portion of final energy that is used to meet energy services. This
portion does not include energy that is wasted due to factors such as line loss, plug loss,
leakage, and waste heat.

•	Rejected energy - Energy that is lost through inefficiencies such as line loss, plug loss,
leakage, and waste heat.

•	Energy services - Activities that require energy, including space conditioning, water heating,
lighting, and passenger and freight transportation. Energy services can be expressed in units
of energy (e.g., exajoules), but are also often expressed in physical units, such as lumens in
lighting and passenger-km for travel.

As energy is transformed through the energy system, there are inherent inefficiencies and

losses. In the U.S., the quantity of useful energy is less than the quantity of rejected energy, as

shown in the Sankey diagram below.

17


-------
Estimated U.S. Energy Consumption in 2021: 97.3 Quads

Lawrence Livermore
National Laboratory







Energy
Services
31.8

Figure 1.3 Energy flows and consumption in the U.S. in 2021. This Sankey diagram tracks the
flow of energy through the U.S. energy system, including useful and wasted energy. Source:
LLNL

1.2.2 Energy and the environment

The energy system has many intersections with the environment. For example, based on the
2020 EPA Inventory of Greenhouse Gas Sources and Sinks, energy supply and use are
responsible for more than 96% of U.S. anthropogenic CO2 emissions and 82% of overall U.S.
anthropogenic GHGs. As a result, the energy system is a major focus of climate action at the
state and federal levels, including in the climate action plans enacted by more than half of
states (https://www.c2es.org/document/climate-action-plans/) and in federal regulations such
as the Corporate Average Fleet Efficiency standards for onroad vehicles
(https://www.nhtsa.gov/laws-regulations/corporate-average-fuel-economy).

Energy is also a source of air pollutants. According to the EPA's National Emission Inventory
(NEI), the energy system contributes 91% of U.S. anthropogenic NOx emissions, 75% of SO2,
74% of CO, 45% of VOCs, and 22% of directly emitted fine PM in 2021.
(https://www.epa.gov/air-emissions-inventories/air-pollutant-emissions-trends-data)
Combustion of fossil fuels is the main contributor to most of these emissions, although natural
gas leakage, evaporative processes, cement and fertilizer manufacturing contribute as well.

Despite significant improvements in air quality over the past decades, more than 100 million
people in the U.S. are estimated to live within areas that exceed one or more National Ambient

18


-------
Air Quality Standards (NAAQSH https://www3.epa.gov/airquality/greenbook/popexp.html).
Thus, reducing the contribution of the energy system to air pollutant emissions is a priority for
agencies such as the U.S. EPA.

Counties Designated "Nonattain merit"

• err?--

GU	PR

Legend "

L_| County Designated Nonaflammenl for 6 KAAQS Polutants
\ | County Designated Nonat1»nment fo« 5&AAQS Polutant*

County Designated NonafiaBfimwH r« 4 UAAQS Polufcant*

County Designated Nortattanment tor 3 NAAOS Po« utants

¦ County DcsKJnated NonaOanmenl tor 2 HAAQS PeMants
County Designated Nonadiienment for 1JSAAQ5 Pdutant

¦ The National Amt>ent Air OuaWy Standards . and SuHw OtowJt (1971 and 2010)

" Included in the counts arc counCtcs designated for NAAQS arwj reused NAAQS pollutants
Revoked 1-hour (1979) and 6-hour Ozone (1997) we e*cfruded Pamial counties, those with part
d the county denigrated nonaflamment and part attainment, are shown as Ml counties on the map

Figure 1.4 Map of U.S. non-attainment areas. Shaded counties did not attain one or more
National Ambient Air Quality Standards as of Nov. 30th, 2021. Source: EPA Greenbook

(https://www3.eoa.aov/airaualitv/areenbook/oooexD.html)

Environmental impacts associated with energy are not limited to climate and air quality. In the
U.S., freshwater withdrawals for thermoelectric power plant operations in 2015 were nearly as
great as those of agriculture (https://pubs.usgs.gov/fs/2018/3035/fs20183035.pdf). 41% vs
42%. As a result, cooling water requirements make the energy system susceptible to droughts.
Furthermore, used cooling water is typically discharged into rivers or lakes. Drought and high
ambient temperatures can limit the ability of these bodies to absorb additional heat without
damaging sensitive aquatic ecosystems. These conditions can result in the temporary shutdown

19


-------
of thermal generating capacity if the Total Maximum Daily Load (TMDL) limits for a discharge
body are exceeded (https://www.epa.eov/tmdl/overview-total-maximum-daily-loads-tmdls).
Energy is also a major source of solid and liquid waste. Electricity production, for example, was
responsible for 450,000 tons of solid waste in 2019

(https://www.epa.eov/trinationalanalvsis/electric-utilities-waste-management-trend).
1.2.3 Environmental management

There are many options available for addressing the environmental impacts of energy
production and use. The traditional approach is to use control devices to capture pollutants
from exhaust gases and liquid waste streams. While pollution controls have been used
successfully and are responsible for much of the environmental progress that has been made to
date, controls can result in important tradeoffs. For example, scrubbers that remove SO2 from
exhaust gases generate a liquid waste stream that must be treated. Also, carbon capture
devices that remove CO2 from an exhaust stream are energy intensive; by requiring more fuel,
pollution associated with coal and gas production may increase.

Non-traditional control options include reducing demands for energy through energy efficiency,
switching to fuels with lower emissions intensity, and the combination of electrifying end-use
technologies in transportation and buildings while simultaneously shifting electrification to
clean sources. An important aspect of many of these management strategies is that they have
the potential to benefit climate and environmental endpoints simultaneously.

However, some non-traditional measures can result in changes in the types and locations of
pollutants that are emitted. For example, while considered to be a low-carbon fuel, biomass
can have high air pollutant emissions, and there are emissions associated with the
manufacturing of batteries, solar panels, and wind turbines. There can also be cross-sector
interactions. For example, depending on how electricity demands are met, electrification can
increase electric sector emissions, as well as the emissions associated with natural gas
production and transportation. Shifts in energy supply and demand can also result in price-
induced fuel switching in other sectors.

In this complex landscape, it is important for policymakers to be able to:

•	quantify the benefits and co-benefits associated with various management options,

•	understand cross-sector interactions, tradeoffs, and potential unintended consequences,

•	evaluate how uncertainty in future conditions may present opportunities or challenges for
climate action and environmental management, and

•	identify management options that are cost-effective and robust through approaches such as
scenario analysis.

20


-------
Addressing these needs is the objective of the GLIMPSE project at the U.S. EPA.

1.3 GCAM-USA

The computational engine for GLIMPSE is GCAM. Specifically, we are using GCAM-USA, a
version of GCAM that includes state-level resolution of the U.S. energy system. GCAM-USA
version 5.4 covers a time horizon of 2015 through 2100, simulated at the annual resolution in 5-
year time steps.

In GCAM-USA, the 50 U.S. states plus the District of Columbia are explicit regions that operate
within the global GCAM model. Energy transformation (electricity generation and refined
liquids production) and end-use demands (buildings, transportation, and industry) are modeled
at the state resolution. Interstate trade of all energy goods is simulated, with state-specific
consumer price mark-ups assigned for coal, natural gas, and refined liquids based on price data
from EIA 2017b.

Note that several aspects of the energy system are not disaggregated to the state level. Most
notably, this applies to primary production of fossil resources including oil, gas, and coal. The
supply of biomass energy feedstocks, which include residues and dedicated energy crops, is
modeled at the level of 22 water basins in the United States (Calvin et al. 2019, Calvin et al.
2014. Wise etal.2014).

21


-------
32 geopolitical regions

50 states + D.C. in the U.S.

Figure 1.5 GCAM-USA regions. GCAM Global Regions and GCAM-USA States Map. Source:
PNNL.

Attributes of GCAM-USA are summarized in the table below:

Table 1.1 Summary of GCAM-USA attributes and links to documentation and source code. See
the appendix for definition of abbreviations.

Attribute

Description

Type

Technology-rich, market-based human-Earth systems model

Solution approach

Partial equilibrium (GDP is determined exogenously)

Nonlinear programming to identify market clearing prices for all
markets

Myopic (no foresight to future time periods)

Dynamic recursive (each time step starts with the prior time step's
solution)

22


-------
Economic choice

Market shares of competing technologies are determined using a
logit function, which considers the relative costs of technology
options, technology-specific shareweights (representing bias), and a
logit exponent (reflecting the degree to which cost-differences
impact choice)

Temporal coverage

2015-2100 (although typically run to 2050 in GLIMPSE applications)

Temporal resolution

5-year timesteps (although alternative values, including annual are
possible)

Spatial coverage

Global

Spatial resolution

Energy and economic: 32 global regions, with the U.S. disaggregated
by state

U.S. electric grid: 15 regions, similar to the NERC regions, but
following state boundaries

Water and land use: 235 regions, based on water basins

Sectoral coverage

Resources (extraction and mining), refining (oil and biomass),
electricity production (fossil and renewables), industry
(manufacturing, non-manufacturing, and nonroad), commercial,
residential, passenger travel (onroad, air, rail), freight travel (truck,
rail, marine), and agriculture (livestock, poultry; biomass for energy,
food, feed)

Pollutant coverage

GHGs and SLCPs: C02, CH4, N20, BC, OC

Air pollutants: NOx, SO2, CO, NH3, VOC, PM10, PM2.5

Policy

representations

Emission taxes and caps, technology and fuel taxes and subsidies,
technology capacity and market share targets, clean energy and
renewable portfolio standards

Implementation

Model coding: Object-oriented C++

Data system: CSV tables that are converted to XML input files using R

Platforms

Windows, Apple, Linux (including high performance clusters)

Note: To date, GLIMPSE has only been tested on the Windows
platform.

23


-------
Computing
requirements

At least 12 GB of RAM and 80 GB of free disk space are needed for
core GCAM. 16 GB or more of RAM is recommended for GCAM-USA,
and We have found that 20 GB or more is preferred state-level
renewable portfolio standards are included in scenarios.

Runtime

30 minutes to 4 hours on a typical desktop Windows computer,
depending on policies and other modifications to the reference
setup.

Source code and data

https://github.com/JGCRI/gcam-core (Open source)

Documentation

GCAM 5.4 documentation: http://igcri.github.io/gcam-
doc/vS.4/toc.html

GCAM Developer's Guide: http://igcri.github.io/gcam-doc/dev-
guide.html

1.4Components of GLIMPSE

GLIMPSE serves as a graphical interface to GCAM and GCAM-USA. We refer to both versions of
the model as "GCAM" from here on unless clarification is necessary. The graphical interface
consists of two primary components, the Scenario Builder and the Mode/Interface.

The Scenario Builder allows users to alter input assumptions and construct policy scenarios, as
well as to manage the execution of GCAM. Through its New Scenario Component Creator, users
can construct GHG targets, sectoral emission caps, technology subsidies, clean energy
standards, electric vehicle market penetration targets, and different assumptions about the
characteristics of technologies. The GLIMPSE Model Interface builds upon the Model Interface
that PNNL distributes with GCAM, providing additional capabilities for filtering, visualizing,
analyzing, comparing, and exporting model results. In this Users' Guide, we refer to the
GLIMPSE Mode/Interface as the Mode/Interface.

These components are described and demonstrated in the Tutorial section of the Users' Guide.
Attendance of GLIMPSE hands-on training sessions is highly recommended for those who are
considering using GLIMPSE for their applications.

1.5 Design philosophy

GLIMPSE has been developed to meet the needs of both experienced GCAM users and those
who are new to GCAM. For experienced users, the Scenario Builder will enhance their typical

24


-------
GCAM workflow by organizing a library of scenarios and scenario components, managing single
and batch execution, providing quick access to logs, and archiving the files that are specific to a
scenario. GLIMPSE also automates some activities that would be tedious even for an
experienced user, such as developing policy "add-on" files that implement an emissions cap or
clean energy standard over a group of states.

Those who are new to GCAM will benefit by being able to rapidly set up, execute, and examine
the results of scenarios. Furthermore, conversations with prospective GCAM users have helped
us identify and implement the scenario "levers" that address many of their modeling needs,
and these levers have been added to GLIMPSE. While GCAM still has a substantial learning
curve, GLIMPSE can address many of the barriers that new users face.

When GCAM execution of a scenario begins, the model reads a configuration file which
specifies many aspects of the run, such as the number of time periods to simulate, the name of
the output database into which results will be placed, and which scenario components will be
included. Scenario components are extensible Markup Language (XML)-formatted files that
provide the data used by the model, including parameterizations of the electricity production,
refining, industrial, commercial, residential, and transportation sectors. Scenario components
can also include representations of policies or alternative assumptions about technologies,
population, and GDP growth.

To simulate an alternative scenario, references to different or additional scenario components
can be included in the configuration file. The order in which these "add-on" scenario
components are listed is important: if a parameter value occurs in several scenario
components, the last value overrides prior values. Thus, policies or alternative assumptions
about technologies typically can be specified in "add-on" files that are listed at the bottom of
the scenario component list.

GLIMPSE supports this workflow. GLIMPSE includes a template configuration file. Through the
Scenario Builder the user can easily create a scenario based on this template, modified to
reflect the user's choices of output database, years to simulate, etc. Furthermore, the user can
add new scenario components by selecting them from a Component Library. The resulting
scenario can then be saved to a Scenario Library. Scenarios in the library can be executed
individually or in batches, and the status of each is displayed.

25


-------
Q GLIMPSE Scenario Builder
File Tools View Help

~

Component Library Search:

Component Name

Calib-2025lndCoal-GHGPIanStates.csv

Calib-biomass_constraints.txt

Cali b-coa l_egu_2020.txt

Calib-HDV-BEV-SW-1x50.csv

Calib-LDV-EV-AE02020.csv

Calib-Lower lncome_Elasticity_Tran.txt

+ b x O

C alih-l

< (	

alt Ki^f.iol C\A/»-

Created

2023-02-08: 08:45
2023-02-08: 08:45
2023-02-08: 08:45
2023-02-08: 08:45
2023-02-08: 08:45
2023-02-08: 08:45

Create Scenario

Component Name

No content in table

Scenario Library	Search:

Scenario Name

GCAM5p4-Ref-Orig
GLIMPSEv! -Reference

^ E

Created

2023-04-24: 12:20
2023-04-27: 11:52

~ ~~

Completed

2023-05-01:08:48

= H ID i) H O

Status

Success

Runtime

0 hr 51 min

Resources... CPU: 29% | RAM: 31.9GB Free:70% | Disk: 2,761.7GB available | Swap: 63.9GB Free: 73% // Database: database Size: 2.1GB Used: 5.4%

Figure 1.6 The Scenario Builder. The Scenario Builder facilitates development and execution of
scenarios.

GLIMPSE also supports exploration of model results through the Mode/Interface. With this tool,
users cars extract, filter, rapidly visualize, and compare many outputs across regions or
scenarios.

The Mode/Interface uses queries to extract data from the GCAM output database. We have
organized the query list such that those that are anticipated to be of greatest utility to GLIMPSE
users are grouped at the top. Hovering the mouse over these queries will produce a "tooltip"
that includes a brief description of the query. As GLIMPSE develops, we plan to have the option
of viewing results in units more meaningful for users in the energy and air quality management
fields (e.g., GWh instead of EJ; short tons instead of metric tonnes). We also plan to support
additional types of graphics, including maps and Sankey diagrams.

While some graphical capabilities are provided by the Mode/Interface, users can also readily
export data for analysis using other tools. For example, open query results can be saved to
comma-separated-value (CSV) files. Alternatively, users can drag a table into Excel by dragging
the tab associated with the table into an open Excel workbook.

26


-------
ft GLIMPSE Modellnterface [E:\Projects\GLIMPSE-5p4\GCAM-Model\gcam-v5.4\output\database]
| File Edit Table Help	

~

Scenario

GC AM5p4-R ef-Oi ig 2023-1-5T0:
GLIMPSEv 1 -R eFerenc e 2023-1-5

GLIMPSEv 1 -Tax- 100C-5pct 202,

Africa_Eastern

Afi ica_Northern

Africa_Southern

Africa_Western

Australia_NZ

Brazil

Canada

Central America and Caribbean

Central Asia

China

EU-12

EU-15

Europe_Eastern
Europe_Non_EU

European Free Trade Association

India

Indonesia

Japan	

Queries

- Energy inputs to refining activities (GCAM-U5A)

Technologies

a Electricity generation by region (no cogen)(GCAM-USA)

. Electricity generation by gen and cooling tech (incl cogen)(GCAM-USA)
Electricity generation by aggregated subsector (GCAM-USA)	

Electr icity generation by aggregated subsector rnw detail(GCAM-USA)

a Electricity generation by subsector (GCAM-USA)

-	Electricity generation by tech and cooling (incl cogen)(GCAM-USA)

-	Electricity generation by tech and cooling (incl cogen)(GCAM-USA)(new capacity)

-	Electricity generation by cogen only (GCAM-USA)

¦ Electricity generation input by subsector (GCAM-USA)

. Electricity generation by region (incl cogen)(core)
ss Electricity generation by subsector (core)

-	Electricity generation by gen tech (core)

-	Building final energy by tech

. Building service output by tech
. Industry final energy by tech and fuel

Run Query Diff Query £7] Total

Collapse

Update Single Queries

Q Electricity generation by aggregated subsector rnw detail(GCAM-USA)

GCA,,
GCA..

Filter

Graph

region
[Total

sub... 2015
biomass|0.240
coal 5.30

GCA...
GCA,..
GCA...
GCA...
GCA...
GCA...
GCA...
GCA...
GCA,..
GCA...
GCA...
GUM..
GUM..
GLIM..

Total

[Total

[Total
[Total
i Total

qi?

geo

2030
0.205

2040 2045
0.229 10.248

hydro ,0.910
hydr... 0.00

11.06
0.00...

refin... 0.139 0.0645 0.0598

[Total
[Total
[Total

solar PVjQ.116
solar... 0.00

Total

[Total
Total

jTotal
[Total
Total

0.730 1.04
|o.0268 |0.0557

11.06
0.00. ¦¦
0.0514
1.62
0.103

0.143 0.112 0.120

1.06
10.00...

1.06 1.05
0.00190 0.00...

0.00 0.0189 0.0730

nuclear
solar ...

2^99	

0.0128

biomass

coal

loas

0.240
5.30
4.99

0,0467
2.17
0.117
0.164

1.05

[0.00433
0.0436 10.0426 10.0429

0.0600 0.0851

2.68
jO. 128

2.25

0.207 10.170 0.166

2.69
6-52

2.64
16.35

2050
|0.276

Units
EJ

11.7

¦71 3.25
0.124 0.132

j3.43
10.160

0.286 |0.448 0.621 EJ

3.36 3.51 3.78

1.79 1.36
0.224 0.306

0.188
1.64

0.868 EJ
10.347 EJ

1.42
8.45

1.13
9.45

GCAM5p4-Ref Orig
region: Total

GLIMPSEvI-Reference
region: Total

#

o 10

§

o 10

o o o

h h 8

















1

¦

¦

|

¦

ii







1

1

|

1

¦

¦



1

¦

¦

¦

¦

m







N

o





K

N



Ol

y

Ol

s





&

a

Figure 1.7 The GLIMPSE Modellnterface. The Modellnterface supports exploratory
investigation ofGCAM results, including visualization and examining the differences in model
outputs from one scenario to another.

GLIMPSE simplifies the process of creating add-on scenario components. GCAM modelers
typically create add-on files in one of several ways. Simple scenario components are often
created by a user in a text or XML editor. However, constructing complex scenario component
files in this manner can be very tedious. Instead, GCAM modelers often use a two-step
approach. First, a table of data is saved as a CSV file. A header file defines how the tabular data
in the CSV file is to be converted to XML by the CSVtoXML.jar Java program that is integrated
into the Modellnterface. Even this process can be time consuming if there are many
technologies or regions that are being affected. For example, a state-level renewable portfolio
standard CSV file may require hundreds of thousands of rows.

The Scenario Builder includes a New Scenario Component Creator that provides an alternative
way to construct add-on files. Using this feature, users can implement a variety of policies and
introduce alternative assumptions about technologies and fuels. These modifications can easily
be applied to a single region or to a group of regions. Among the options available include:

27


-------
•	pollutant taxes or caps, introduced for a single sector or economy-wide,

•	technology market share constraints as a fraction of new sales or of total stock,

•	technology availability, including first and last year available,

•	technology-specific taxes and subsidies,

•	alternative values for technology costs and efficiencies,

•	adjustments to consumer preferences via the share weight parameter, and

•	fuel price adjustments.

GLIMPSE automates the process of generating the appropriate CSV file, selecting the matching
header, then executing CSVtoXML.jar to produce the corresponding XML.

* P»r«i 'kmirc L smpcncrrf C»iafpr

XMi UU F^jTjhI Tw/Clp ' ffcfi Avail j M«ffcs|Swi	T*~ f*m T>Mf Anooi! .cr.
SMn3pftE»Tt#rr,

lun^t.itodJU

Afftt*rt

Pw
-------
1.6Computer and software requirements

The GLIMPSE software is written in the Java programming language and requires the 64-bit
version of the Java Runtime Environment (JRE), version 8, which is also referred to as JRE 1.8.
We include Amazon's Correto version of the JRE as part of the GLIMPSE package
(https://docs.aws.amazon.com/corretto/latest/corretto-8-ue/downloads-list.html). although
users are able to configure GLIMPSE to user other versions of JRE 1.8. Note that GLIMPSE will
not work with OpenJDK since it does not include the JavaFX library.

While future versions of GLIMPSE are expected to be able to operate on PCs, Macs, and Linux,
the software currently is available on Windows PCs only.

The GCAM model itself is computationally and memory intensive. At least 12 GB of RAM and 80
GB of free disk space are needed for GCAM. 16 GB or more of RAM is recommended for GCAM-
USA. We have found that 20 GB or more is preferred state-level renewable portfolio standards
are included in scenarios.

Each simulation generates up to 3 GB of results. Hard disk space is also used as virtual memory
by GCAM. We recommend that your computer has at least 80 GB of free hard disk space, but
more is preferable.

1.7	Organization of this Users' Guide

This Users' Guide includes installation instructions, a multi-part tutorial, a description of the
GLIMPSE Reference Scenario, and key results. The Users' Guide also includes a brief overview of
how GCAM works, instructions on performing common tasks, additional information for
advanced users, and descriptions of the components of the graphical user interface. A trouble-
shooting section helps with common problems, and a glossary defines key terms and acronyms.

1.8	GLIMPSE Version

This Users' Guide has been developed specifically for GLIMPSE vl. If you are using a more
recent version of GLIMPSE, some of the information provided here may no longer be accurate.
Furthermore, this User's Guide will evolve as users report their experiences.

GLIMPSE currently incorporates GCAM 5.4, which was released in the summer of 2021. GCAM
releases typically occur at least once per year. See https://github.com/JGCRl/gcam-
core/releases for information on each GCAM release, including new features. We expect to
update GLIMPSE to incorporate GCAM 7.0 during the second half of 2023.

29


-------
Future versions of the Users' Guide will include information on how to update GLIMPSE to work
with other versions of GCAM. In general, updating can be straightforward, but changes to the
model or technology names can require non-trivial changes to GLIMPSE setup files and code.

1.9A note on units

As a result of GCAM's origins in GHG emission projections and climate analyses, the units in
GLIMPSE and GCAM may be different than those typical for applications in the air quality and
energy fields.

•	Metric units are used in all instances. In this documentation, we refer to metric tons as
"tonnes". However, in the GCAM outputs, "tons" is used.

•	Energy values are typically provided in Exajoules, which are joules x 10A18. For reference,
one EJ is 277,778 Gigawatt hours (GWh) or 277.778 Terawatt hours (TWh). For reference,
the US power sector produced approximately 16 EJ of electricity in 2020, and the onroad
transportation sector used approximately 21 EJ of gasoline, diesel, and ethanol in the same
year.

•	Where CO2 outputs are presented in the Mode/Interface, these are in units of million metric
tonnes of Carbon, or MTC. To convert MTC to MTC02, multiply by the ratio 44/12, which is
based on the relative molecular weights of C02 (44) and C (12).

•	Air pollutant emissions are reported in tera-grams (Tg), or grams x 10A12. A Tg is equivalent
to a million metric tonnes, or a MT. One MT is equivalent to 1,102,310 US short tons.

•	The $-years for monetary values are different than today's $s, resulting in the need to
adjust for inflation when interpreting results.

o Prices and costs typically are reported in 1990$s.

o Technology taxes and subsidies are represented in 1975$s per unit of output
o Based on the Consumer Price Index,

¦	$1 in 1975$s is equivalent to $2.40 in 1990$s

¦	$1 in 1975$s is equivalent to $5.66 in 2023$s

¦	$1 in 1990$s is equivalent to $2.34 in 2023$s.

•	Travel demands are represented in million passenger-km or million tonne-km.

1.10 Interpreting GCAM results

While GLIMPSE simplifies tasks such as developing policy scenarios, executing GCAM, and
analyzing results, GLIMPSE users should keep in mind that neither it, nor the underlying GCAM

30


-------
model, are commercial products. Documentation is available, but support is limited. As of May
2023, GLIMPSE has been tested by users within the U.S. EPA and has been used in four
semesters of a graduate-level course on Integrated Assessment Modeling. Please consider
reporting back your experiences, feature requests, comments, and suggestions to Dan Loughlin
(Loughlin.Daii@epa.gov) such that these can be considered as GLIMPSE development
continues.

Furthermore, please note that GCAM is a complicated model. Applying it for research and
policy analyses requires experience and skill that takes time to develop. For example, when
assessing model results, it is recommended that users look beyond the high level results (e.g.,
whether emissions went up or down) and examine the detailed technology-level outputs,
asking questions such as "Do these responses make sense?", "Is this result reasonable, or did I
uncover a response that arose because the scenario pushed the model in a new direction?",
"How did limitations in the model formulation affect this result?", and "How were the results
impacted by calibration to historical data or by assumptions about future conditions?"

Users should not expect GLIMPSE to provide "turnkey" answers. Analyses with tools such as
GLIMPSE and GCAM are iterative processes, with each iteration providing additional
information about the problem, potential solutions, and the representation of these in the
model. It is common to make refinements to scenario assumptions and policy implementations
as part of this iterative process.

When presenting model results, it is important to avoid statements such as: "If policy X is
implemented, Y will happen." Results are contingent on many factors, including the
assumptions about population growth and migration, economic growth, technology costs and
efficiencies, climate change, and human behavior and choices. Predicting these factors into the
future is inherently uncertainty. Additionally, while the logit algorithm that predicts market
shares has been calibrated to past decisions, real-world human decision-making involves many
considerations beyond the relative costs of competing options.

Thus, to properly present and caveat findings, it is important for analysts to understand the
operation of the model, key assumptions, and limitations. For any given analysis, users are
encouraged to explore how a scenario impacts technology market shares, market prices, and
fuel use. These will provide insights regarding the underlying pathways and mechanisms that
led to observed results. Sensitivity analysis is encouraged, which will indicate how the model's
results change in response to incremental perturbations in key input parameters.

To begin the process of learning how GCAM operates, we highly recommend that users read
the "How does GCAM work?" portion of this Users' Guide, which includes information about
the operation of markets, the logit function, shareweights, and model calibration.

31


-------
When applying GCAM to particularly challenging policy scenarios (e.g., a Net Zero CO2 target or
to specific GHG emissions targets), users should think holistically about the scenario and how
Reference Case assumptions may shift under such a target. For example, users may want to
modify shareweights for electric vehicles to reflect conditions that are not simulated by GCAM,
such as the investment and build-out of a charging infrastructure. The process of developing
Deep Decarbonization scenarios, and others that diverge significantly from historic and
Reference Case conditions, may involve many such considerations.

1.11 Known bugs, limitations, and other considerations

GLIMPSE users should take note of the following.

•	Unexpected termination - GCAM will occasionally terminate unexpectedly, reporting "DNF"
to the Scenario Builder. There are several causes, including the computer's available
resources being exhausted (e.g., RAM or disk space) and conflicts in the names of markets
or policies. While we have built tools into the Scenario Builder to avoid and help deduce the
cause of these problems, some level of "debugging" by users is required. See Section 4.16
for more information.

•	Numerical issues - With GCAM's global coverage and representation of thousands of
markets of diverse sizes, the model's solver must deal with numbers that vary greatly in
magnitude. At the same time, the model takes advantage of the multi-threading capabilities
of modern computers to parallelize computations, reducing runtime. However, parallelizing
computations introduces small numerical errors that can sometimes lead to the problem of
"Unsolved Markets." This Users' Guide includes information about how unsolved market
messages can be interpreted and how to avoid numerical issues by turning off parallel
computation or adjusting the solver parameterization. See Section 5.3 for more
information.

•	Transportation constraints - For legacy reasons, GCAM's units used in internal calculations
for the transportation sector are million British Thermal Units, or MMBtu. One MMBtu is
equivalent to 1.054e-9 EJ. The conversions required result in transportation policies and
constraints being particularly prone to numerical issues. While we have attempted to
address these through scaling during new scenario component creation and through
modifications to the solver settings, some numerical issues may arise, particularly in
transportation market share scenarios.

•	Output database size limitations - GCAM uses the BaseX database software for storing
model results. We have found the BaseX software to be unable to open databases that

32


-------
exceed approximately 40 GB in size. For GCAM-USA, this is the equivalent of 15 to 20
results. Once the database can no longer be opened, the data inside is lost. We recommend
that users pay particular attention to database sizes and to information in Section 4.5 of this
Users' Guide about how to monitor database size, export and import results from a
database, and create new databases.

•	Model I interface freezes - There are instances when the Mode/Interface freezes. When this
occurs, users can terminate the Model I interface task without terminating GLIMPSE by using
the Windows Task Manager. See Section 6.4.2 for instructions on terminating the

Model I interface when it is frozen.

•	Boundaries on system representations - In a complex model such as GCAM, decisions must
be made by the developers regarding where to draw the boundaries for various systems
represented in the model. These decisions are often impacted by factors such as data
availability and computation challenges. Knowing where these system boundaries are is
important in understanding the responses of the model to policies and other perturbations.

•	Feedbacks - In the real world, there are feedbacks that may impact the co-evolution of the
economic, energy, agricultural, land use, water, and climate sectors. For example, climate
impacts may change the availability of water, which would impact agricultural and energy
choices. These would have ramifications for the economy. While some feedbacks are
incorporated in research versions of GCAM, they were not been included in the publicly
released version of GCAM-USA 5.4.

•	Myopic solution process - GCAM is a dynamic recursive simulation model. This means that
the model steps through time, using the solution from the last modeled time period as the
starting point for the current time period. Markets are solved in the current period by
considering conditions in that period, but not taking into consideration future conditions. As
a result, GCAM's solution process cannot anticipate conditions in future time periods, such
as the planned tightening of a carbon cap over time. As a result, the model may make
decisions in the short term that are not optimal from a long-term perspective.

•	Calibration - GCAM uses a logit function to assign market share to competing technologies.
The parameters used in the logit function for many technologies and markets are
determined based upon conditions in the calibration year, which is 2015 in GCAM-USA 5.4.
For new and emerging technologies, these parameters are based upon assumptions about
the speed at which barriers and biases will be addressed. As a result, choice function
parameterization may not reflect behavior in the future, particularly for deep
decarbonization scenarios that differ significantly from historical decisions and Reference

33


-------
Case assumptions. Please see Section 5.5 for a discussion of adjustments that are
recommended for consideration when modeling deep decarbonization scenarios.

•	Query results for the USA region - In GCAM-USA 5.4, most US sectors have been
disaggregated to the state level. Thus, state-level queries are used to obtain information
about technology shares, fuel use, and emissions in buildings, industry, and transportation.
In contrast, the USA region predominantly includes categories that are represented at the
water basin level, such as water supply, agriculture, and other land uses. The USA region
also includes several industries that have not yet been disaggregated to the state level,
including coal mining, oil and gas production, and hydrogen production. Users should note
that queries of the USA region will not produce national totals. To obtain results for the
U.S., select all states and the USA region. To produce national totals, the "Total" check box
will sum across all selected regions.

•	This work was conducted under an approved Quality Assurance Project Plan (QAPP) that
was developed in accordance with guidance provided in EPA ORD's quality assurance
requirements for modeling and software development. The QAPP ("GCAM-based Long-term
Interactive Multi-Pollutant Scenario Evaluator (GLIMPSE)", J-AESMD-0033472-QP-1-0) was
approved by EPA prior to the initiation of GLIMPSE development.

1.12 Where to learn more about GCAM and GLIMPSE

If you would like to know more about GCAM, please see the model's on-line documentation.
Important links include:

General GCAM documentation, including for the latest public release:

•	http://iecri.github.io/ecam-doc/

GCAM documentation for version 5.4, which is included in this GLIMPSE release:

•	https://eithub.eom/JGCRI/ecam-core/releases/tae/ecam-v5.4
GCAM-USA documentation:

•	http://iecri.eithub.io/ecam-doc/ecam-usa.html
GCAM Users' Guide:

•	http://iecri.eithub.io/ecam-doc/user-euide.html
GCAM Video Tutorials:

•	https://ecims.pnnl.eov/community

34


-------
List of GCAM publications by PNNL authors:

•	https://www.pnnl.eov/publications-reports?keywords=GCAM&fieid document type=3

For more information about the GLIMPSE software and the Environmental Protection Agency's
GLIMPSE project, please see the following links:

GLIMPSE website (subject to change):

•	https://www.epa.gov/air~research/glimpse~computationai~framework~supporting~state~

level-environmental-and-energy

In addition, we have several EPA publications that involve using GCAM for various applications.
These include:

•	Ou, Yang, Noah Kittner, Samaneh Babaee, Steven J. Smith, Christopher G. Nolte, and Daniel
H. Loughlin. "Evaluating long-term emission impacts of large-scale electric vehicle
deployment in the US using a human-Earth systems model." Applied Energy 300
(2021):117364. doi:10.1016/j.apenergy.2021.117364.

•	Ou, Yang, J. Jason West, Steven J. Smith, Christopher G. Nolte, and Daniel H. Loughlin. "Air
pollution control strategies directly limiting future health damages in the US." Nature
Communications 11 (2020): 957, doi:10.1038/s41467-020-14783-2.

•	Babaee, Samaneh, Daniel H. Loughlin, and P. Ozge Kaplan. "Incorporating upstream
emissions into electric sector nitrogen oxide reduction targets." Cleaner Engineering and
Technology 1 (2020): doi:10.1016/j.clet.2020.100017.

•	Ou, Yang, Steven J. Smith, J. Jason West, Christopher G. Nolte, and Daniel H. Loughlin.
"State-level drivers of future fine particulate mortality in the United States." Environmental
Research Letters 14 (2019): 124071, doi:10.1088/1748-9326/ab59cb.

•	Ou, Yang, Wenjing Shi, Steven J. Smith, Catherine M. Ledna, J. Jason West, Christopher G.
Nolte, and Daniel H. Loughlin. "Estimating environmental co-benefits of U.S. GHG reduction
pathways using the GCAM-USA Integrated Assessment Model." Applied Energy 216 (2018):
482-493, doi:10.1016.j.apenergy.2018.02.122.

•	Shi, Wenjing, Yang Ou, Steven J. Smith, Catherine M. Ledna, Christopher G. Nolte, and
Daniel H. Loughlin. "Projecting state-level air pollutant emissions using an integrated
assessment model: GCAM-USA." Applied Energy 208 (2017): 511-521,
doi:10.1016/j.apenergy.2017.09.122.

1.13 How to get started using GLIMPSE

To install GLIMPSE, please follow the instructions described in the GLIMPSE Installation Guide.

A good place to start in understanding GLIMPSE'S underlying assumptions and capabilities is to
read Chapter 2 of this Guide, which describes the GLIMPSE Reference Case. While key results

35


-------
are provided at the national level, similar results can be generated at the state level or for other
countries.

For those who would like to use GLIMPSE themselves, the GLIMPSE tutorials are the next step.
These will take you through important steps of setting up, running, and analyzing scenarios. The
tutorials are in the Appendix of the Users' Guide.

Once you have completed the tutorials, we recommend that you read Chapter 3, "How does
GCAM work?", in the Users' Guide. This will provide an overview of how GLIMPSE simulates
markets in determining the market shares of technologies and fuels, as well as how markets are
impacted by policies such as taxes and caps.

Next, try to develop and evaluate your own policy scenarios, using the information in Chapter 4,
"How do I?", and in Chapter 6.4, "Troubleshooting", as needed.

To develop a more detailed understanding of GCAM, please see the PNNL GCAM
documentation: http://iecri.eithub.io/ecam-doc/user-euide.html.

Demonstrations and trainings are periodically available. Please inquire with Dan Loughlin at

Louehlin.Dari@epa.eov.

1.14 Where to get assistance or provide feedback

If you have difficulties installing or running GLIMPSE, please see the "Troubleshooting" section
of this Users' Guide to see if your problem can be readily addressed. If the information there
does not help solve your problem, please contact the GLIMPSE development team by emailing
Dan Loughlin at Louehlin.Dan@epa.eov. Include detailed information about your problem, as
well as what steps you have already taken to address it. Thank you!

As this is the first edition of GLIMPSE and of its Users' Guide, there are undoubtedly many ways
that both could be improved. Please send bug reports or feature suggestions to
Louehlin.Dan@epa.eov for consideration in future updates.

CHAPTER 2. GLIMPSE REFERENCE SCENARIO
2.1 Introduction

There are many pathways by which the U.S. energy system could evolve over the coming
decades. GCAM-USA can be used to construct internally consistent scenarios that describe
potential pathways. Users are also able to construct their own scenarios, introducing
alternative assumptions about inputs such as: population growth and migration, economic

36


-------
growth and transformation, technology change and adoption, climate change, human behavior
and choice, and policies, including those targeting climate, environmental, or energy endpoints.
The model's projections reflect the underlying assumptions about the present and future, as
well as the formulation of the GCAM-USA itself, including the aspects of the system the
developers chose to include, how various phenomena are represented, and how those
formulations are parameterized. Thus, understanding the underlying assumptions, formulation,
and parameterization is important in interpreting the results. We encourage GCAM-USA users
to become familiar with the GCAM documentation (http://jgcri.github.io/gcam~
doc/v5.4/toc.html). and, in particular, the section on economic choice
(https://iecri.eith ub.io/gcam-doc/v5.4/choice. html).

We include a GLIMPSE Reference Scenario with our GLIMPSE distribution. This Reference
Scenario (GLIMPSEvl-Reference in the Scenario Library) is provided for convenience, and no
endorsement of the Reference Scenario by the U.S. EPA is implied. There is no specific
likelihood assigned to the Reference Scenario or to any other scenario since there is
considerable uncertainty inherent in making long-term energy system projections.

Furthermore, there are differences between the GLIMPSE Reference Scenario and those used in
regulatory modeling activities by the Agency. This Reference Scenario is not intended to
supplant the scenarios used in those activities. GLIMPSE users may choose to tailor the
Reference Scenario for their specific application.

A primary goal of this chapter is to describe GLIMPSEvl-Reference, including how it differs from
the reference scenario included by PNNL with GCAM-USA 5.4. We provide national-level results
to convey fuel, technology, and emission trends, overall and for key energy sectors. Finally, we
compare several energy-related outputs with the U.S. Department of Energy's Annual Energy
Outlook 2022 and air pollutant projections with those of the U.S. EPA's 2016v2 modeling
platform.

2.2 The GCAM-USA 5.4 Reference Scenario

The GLIMPSE Reference Scenario, GLIMPSEvl-Reference, is built upon the GCAM-USA 5.4
Reference Scenario that was distributed by PNNL with GCAM-USA 5.4 but includes several
updates and additional policies. We first describe the GCAM-USA 5.4 Reference Scenario, then
discuss how the GLIMPSE Reference Scenario differs from it.

GLIMPSE users are encouraged to read the "Story-line for the GCAM-USA Reference Scenario"
in the GCAM-USA 5.4 documentation (http://igcri.github.io/gcam~doc/v5.4/gcam~usa.html).
That document provides a narrative, including key assumptions about socioeconomics and

37


-------
energy demands, electricity demand, electricity supply, inter-state electricity trade, refining,
water supply and demand, and carbon sequestration. Primary data sources are provided, and
there is guidance regarding interpreting results. We also include the GCAM-USA 5.4 Reference
Scenario in the GLIMPSE Scenario Library as "GCAM5p4-Ref-Orig".

In general, the GCAM-USA 5.4 Reference Scenario does not explicitly include representations of
air, climate, and energy policies. Policies implemented through 2015 are assumed to be
reflected in the calibration-year emission factors that are derived from EPA's National
Emissions Inventory (https://www.epa.gov/air~emissions~inventories/national~emissions~
inventory-nei). Exceptions include the Tier 3 mobile vehicle and fuel emission standards, as well
as the production and investment tax credits for renewables:

•	Tier 3 standards - Tier 3 standards limit air pollutant emissions from onroad vehicles.
GCAM-USA 5.4 includes onroad mobile emission factors developed using EPA's MObile
Vehicle Emissions System (MOVES), version 2014.

•	Production Tax Credit (PTC) - These credits subsidize renewable generation by providing a
l-to-2 cent tax credit per kWh of electricity generated. The PTC is represented as a
reduction in the cost of wind, geothermal, and biomass technologies in 2020. The policy
reduces the levelized cost of energy by $4.81 (2020 dollars) per gigajoule (GJ) for wind,
$8.71 for geothermal, and $4.62 for biomass energy.

•	Investment Tax Credit (ITC) - The ITC is an alternative to the PTC that provides a credit for
12-to-30% of the capital costs of a renewable energy project. The ITC is represented as a
cost reduction in solar in 2020. The reduction is made to the fixed charge rate, which has
the effect of reducing the portion of overnight capital paid each year.

Please note that we have made changes to the solver settings in GLIMPSE to support state-level
policies, such as electric vehicle sales targets. These changes include modifying the tolerances
and solution floors to support additional decimal places in market share calculations. As a
result, the GCAM-USA 5.4 Reference Scenario results may be slightly different than in PNNL's
distribution.

2.3 The GLIMPSE Reference Scenario, GLIMPSEvl-Reference
2.3.1 Overview

Broadly, GLIMPSEvl-Reference is constructed with an underlying story-line that assumes
that historical trends continue in the near term (e.g., through 2030). However, over the
longer term (e.g., 2030-2050), parameters that are used to calibrate to historic trends are

38


-------
relaxed, and outcomes are increasingly driven by economic forces and policies. We assume
a continuation of current, "on-the-books" national and regional policies. This type of scenario is
sometimes referred to as a "current legislation" scenario. When a policy reaches its "sunset
date" (e.g., the final year for which a policy target has been specified), we hold the policy target
constant through the rest of the time horizon. For example, the Regional Greenhouse Gas
Initiative (RGGI) electric power sector CO2 cap has been specified through 2030. We hold that
2030 cap through our final modeled time period, 2050.

In the real world, policies tend to become more stringent over time. GCAM-USA allows
exploration of the impact of alternative assumptions about how policy targets are extrapolated,
and the GLIMPSE Reference Scenario can be compared to alternatives, such as removing policy
targets after they sunset, continuing the historic downward trajectory of targets through
2050, or exploring more stringent options.

Other modifications to the GCAM-USA 5.4 Reference Scenario that are introduced in GLIMPSE-
Ref include the following:

•	Updated technology characterizations - incorporates PNNL's recent updates to the
transportation sector, which include hybrid, fuel cell, and electric versions of most onroad,
air, rail, and marine technologies,

•	Modifications to onroad transportation technology shareweights - adopts shareweight
trajectories for new and emerging onroad transportation technologies that are intended to
reflect current level of commercial viability and barriers to adoption,

•	Updated electric sector costs - incorporates updated costs from the National Renewable
Energy Laboratory (NREL) 2021 National Annual Technology Baseline,

•	Air pollutant emissions - incorporates air pollutant emissions developed by PNNLthat were
not ready in time for the GCAM-USA 5.4 public release,

•	Regulations and standards

o incorporates transportation emissions factors for onroad vehicles that were
developed using MOVES 3.0, a successor to MOVES 2014, that includes the
impacts of new emissions regulations as well updated emission estimation
methods,

o reflects federal rules and standards on emissions from marine and rail sources,

o includes lower-bound market share constraints for light-duty electric vehicles
based upon the recently finalized Near-Term Light-Duty GHG Standards; and,

o represents New Source Performance Standards for the electric sector.

•	Inclusion of several multi-state policies - adopts representations of RGGI
(https://www.rggi.org/). the Section 177 Zero-Emission Vehicle targets

39


-------
(https://ww2.arb.ca.gov/resources/docurnents/states~have~adopted~californias~vehicle~

standards-under-section-l??-federal).

•	Adoption of limits ofbiomass availability- limits US consumption of bioenergy to a
trajectory that reflects the Department of Energy's "2016 Billion Ton Study"

(https://www.enerev.eov/eere/bioenerev/2016~billion~ton~report), and,

•	Calibrations and other constraints - include constraints intended to represent current or
expected conditions, including:

o adjusts onroad vehicle lifetimes and retirement to be more consistent with
assumptions in EPA analyses,

o reflects witnessed coal plant retirements through 2020 by limiting state-level
generation from existing coal plants to 2021 levels,

o requires offshore wind in the US Northeast to be equal to or greater than current
procurement contracts, including planned nuclear plant shutdowns in New
England,

o limits the market share of light-duty natural gas vehicles and cellulosic ethanol
production to reflect that the commercialization of these technologies has been
slow,

o does not allow new industrial coal capacity in states with explicit GHG reduction
targets, and

o does not allow new coal capacity (including gasified coal and coal with CCS) in
states that will have eliminated coal from the electric sector by 2025.

GLIMPSEvl-Reference currently does not include:

•	Provisions of the 2022 Inflation Reduction Act and 2021 Infrastructure Bill - This legislation
is expected to be incorporated into future versions of GCAM and will be available in
GLIMPSE subsequently.

•	State-specific Renewable Portfolio Standards (RPSs)
(https://www.ncsl.org/research/energy/renewable~portfolio~standards.aspx) and
representations of state GHG reduction targets - Both can be simulated in GLIMPSE and are
incorporated in the Component Library. However, we have found the RPSs to be highly
memory intensive, limiting GCAM execution to computers with 20 GB of RAM or more. For
state GHG targets, many of the targets can be challenging to meet with GCAM and may
require additional assumptions that are not reflected GLIMPSEvl-Reference.

40


-------
•	The 45Q subsidy for Carbon Capture Utilization and Storage - This policy requires a
modification to the GCAM-USA source code that was not available in publicly available
version of GCAM-USA 5.4.

•	Other state and regional policies, including cap-and-trade, energy efficiency standards.

•	Explicit representations of location-specific air pollution control requirements associated
with attainment status of the National Ambient Air Quality Standards.

Below, additional information about components of the GLIMPSEvl-Reference scenario are
provided.

2.3.2 Additional detail

Onroad Transportation technology shareweights

Shareweights are used in the logit function to represent a wide range of otherwise unmodeled
factors that affect the relative sales of competing technologies within a market. For existing
technologies, 2015 technology-specific shareweights are calculated as a function of real-world
market share and the relative costs of competing technologies. The calculated shareweight is
"relaxed" over time, transitioning to a value of 1.0 in a future year, typically linearly. A value of

1.0	signifies that there is no bias for or against the technology, and that its market share is
determined solely on differences in price and on the logit parameter, which specifies how much
differences in price matter. The year in which this transition is completed is specified at the
market level, reflecting external assumptions about technological progress and the breaking
down of barriers.

For new and emerging technologies, the shareweight in the calibration year is zero because the
market share was negligible. The trajectory of how these technology-specific shareweights
transition to 1.0 is thus important, because it is indicative of the rate at which barriers against
technology adoption, whether psychological or technological, are likely to be addressed. Setting
shareweight trajectories involves modelers' judgement. The shareweight trajectories in Table

2.1	reflect that judgement but can also be modified by the use via the Scenario Builder.

For the light duty vehicle (LDV) fleet, we assume that all barriers against the purchase of hybrid
vehicles will be gone by 2025, and in that year, the vehicles will compete based upon cost
alone. Battery-electric vehicles (BEVs) reach this point of perfect competitiveness by 2035,
indicating that issues such as range anxiety and lack of charging infrastructure will no longer be
a barrier to buyers. Since hydrogen-powered fuel cell electric vehicles (FCEVs) are not yet
widely available, we assume that their path to a shareweight of 1.0 is 5 years longer.

41


-------
Heavy duty freight vehicle (HDV) technology shareweights are assumed to follow similar, linear
trajectories; however, the year by which the shareweights reach 1.0 is delayed relative to light
duty vehicles. Shareweights for other technologies can be obtained via the Mode/Interface
component of GLIMPSE. For convenience, many shareweights are available via queries in the
"GLIMSPE->Assumptions" query group.

Table 2.1 Shareweights for selected onroad vehicle technologies

Sector

Technology

2020

2025

2030

2035

2040

2045

2050

Transport-LDV (passenger)

Liquids

1.0

1.0

1.0

1.0

1.0

1.0

1.0

Transport-LDV (passenger)

Hybrid liquids

0.50

1.0

1.0

1.0

1.0

1.0

1.0

Transport-LDV (passenger)

BEV

0.25

0.50

0.75

1.0

1.0

1.0

1.0

Transport-LDV (passenger)

FCEV

0.17

0.33

0.50

0.67

1.0

1.0

1.0

Transport-HDV (freight)

Liquids

1.0

1.0

1.0

1.0

1.0

1.0

1.0

Transport-HDV (freight)

Hybrid liquids

0.25

0.50

0.75

1.0

1.0

1.0

1.0

Transport-HDV (freight)

BEV

0.00

0.17

0.33

0.50

0.67

1.0

1.0

Transport-HDV (freight)

FCEV

0.00

0.14

0.29

0.43

0.57

0.71

0.86

Air pollutant emissions

Most air pollutant emission factors (EFs) for GCAM-USA are derived from the 2016 EPA National
Emission Inventory (NEI). For most source categories, activity based EFs are calculated using the
following methodology. First, emissions from the NEI are aggregated by pollutant species, fuel,
and GCAM energy category. These values are divided by the fuel used in each category,
producing input based EFs (e.g., Tg of emissions per EJ of fuel used). For other sectors that are
expected to be impacted by regulations, output based EFs may be used. For example, onroad
emission factors currently are obtained from MOVES 3.0. These emission factors represent
regulations such as the Tier III mobile vehicle engine and fuel emission limits.

Policy representations

To characterize RGGI, we constrain the electric sector CO2 emissions of the member states
following the program's stated budgets through 2030. The cap is held at the 2030 level through
2050. The states included in our representation are Connecticut, Delaware, Maine, Maryland,

42


-------
Massachusetts, New Hampshire, New Jersey, New York, Rhode Island, Vermont, and Virginia.
New Jersey and Virginia fall under the cap after 2020. Cap values are provided below. The cap
increases in 2025 to reflect the addition of NJ and VA.

Table 2.2 RGGI targets and membership

Year

Cap (MTC)

Assumed membership

2020

19.34

CT, DE, ME, MD, MA, NH, NY, Rl, VT

2025

26.01

CT, DE, ME, MD, MA, NH, NY, Rl,
VT, NJ, VA

2030-2050

21.50

CT, DE, ME, MD, MA, NH, NY, Rl,
VT, NJ, VA

The Section-177 policy representation involves a minimum fraction of passenger vehicle sales
that must be electric. This Zero-Emission Vehicle (ZEV) target was adopted by California and
signed on to by the following states: Connecticut, Delaware, Massachusetts, Maryland, Maine,
Minnesota, New Jersey, New Mexico, New York, Oregon, Rhode Island, Vermont, and
Washington. While Minnesota plans to join the program in 2025, we have made the simplifying
assumption that the sales target is applicable from 2020. We also assume that the target is
applicable to the percentage of passenger-kilometer (pass-km) demand met by new sales as
opposed to vehicle count. For 2020, 6% of pass-km met by new sales across the member states
must be electric vehicles.

For 2025 through 2050, we shift to representing lower bounds on the combined sales shares of
EVs and fuel cell vehicles, based upon estimates from the Near-Term Light-Duty GHG standards

(https://www.epa.gov/regulations~emissions~vehicles~and~engines/final~rule~revise~existing~
national~ghg~emissions). These targets, which are applied nationally, are shown in the table
below.

Table 2.3 Sum of electric and fuel cell vehicle sales share

Year

Passenger cars

Passenger and light
commercial trucks

2025

12.9%

7.1%

2030

17.0%

10.6%

43


-------
2035

18.5%

11.0%

2040

19.7%

11.5%

2045

21.3%

12.0%

2050

22.1%

12.6%

Constraints on biomass availability

While the biomass constraint will not come into play in many scenarios, it is included in our
GLIMPSE Reference Scenario in general so that it will be in place if the user is simulating CO2 or
GHG reduction targets. In those instances, and without any additional limits on biomass, GCAM
tends to turn to biofuels and bioenergy as a primary means of mitigation. However, GCAM does
not represent important aspects of the US and global energy markets. For example, biomass
used for bioenergy is generally expensive to transport because of its bulk and water content.
This can limit the geographical size of markets. Also, in a scenario in which the U.S. has a
mitigation target in place, it is reasonable to assume that other countries would be competing
for available biomass, limiting US imports. To represent this limit, we apply the constraints in
the following table, which were developed by PNNL and are intended to represent values from
the U.S. Department of Energy's 2016 Billion-Ton Report
(https://www.energy.gov/eere/bioenergv/2016~billion~ton~report).

Table 2.4 Biomass limits in the US and globally. Global values do not include the US region.

Year

US biomass ceiling
(EJ)

Global biomass ceiling
(EJ)

2015

4.7

195.3

2020

5.8

194.2

2025

7.3

192.7

2030

8.8

191.2

2035

10.5

189.5

2040

12.3

187.7

2045

14.8

185.2

2050

17.4

182.6

44


-------
2.4 Selected results for the GLIMPSE Reference Scenario

Below, national-level GLIMPSE Reference Scenario results are presented for several key
outputs. The results are divided into three sections to reflect the three portions of the figure
below: (1) primary energy, (2) processing and conversion of energy carriers, and (3) final
energy. In addition, we provide results indicating the market shares for various technologies in
selected end-use sectors, as well as emissions of CO2 and air pollutants.

Primary Processing and conversion of energy carriers Final energy
energy	^	

Fossil Fuels

Refining & Processing

Biomass

Gasification Combustion-Based
Electricity Generation



Uranium

Conversion &
Enrichment

Nuclear Power









Wind, Solar,
Hydro

H, Generation

Carbon
Sequestration

Transportation

Residential

Commercial

Direct Electricity
	Generation	

Industry

Figure 2.1 Energy system schematic. The energy system extends from the import or extraction
of primary energy, through its processing and conversion into useful forms, through its use in
meeting final end-use energy demands.

To obtain national totals, all states, DC, and the USA region were selected. The USA region
includes a few sectors that have not yet been disaggregated to the state level, such as
agriculture, hydrogen production, and oil, natural gas, and coal operations.

All data were extracted from the output database using queries available in the "GLIMPSE"
section of the Mode/Interface query list. Except for the emissions graphics, all images shown

45


-------
were generated using the Mode/Interface's graphing tools. Emissions data were combined from
two queries in a spreadsheet.

46


-------
2.4.1 Primary energy

Fig. 2.2 shows the GLIMPSEvl-Reference projection of the consumption of primary energy (e.g.,
in its "raw" form) through 2050. Renewables are shown as direct equivalent, indicating that the
quantity shown is the energy produced from wind, solar, geothermal, and hydro.

Overall, primary energy consumption is stable from 2015 on, although consumption of energy
increases steadily from 2030 through 2050, returning to 2015 levels in 2040. Imported oil
decreases over time but imported natural gas increases. The reduction in primary energy use in
2020 reflects factors such as improved vehicle efficiencies and the transition from coal-
powered electricity production to higher efficiency gas combined-cycle turbines. The impacts of
the COVID-19 pandemic are not included in this version of the model.

Primary energy consumption by region (direct equivalent)
GLIMPSEul -Reference
region: Total



95



90



85



80



75



70



65



60

—J

55

UJ

50

"5

Cl

45

—

40



35



30



25



20



15



10



5



o -

i traded oil
traded natural gas
I traded coal
I oil

natural gas
I coal
i biomass
i nuclear
wind
solar
i hydro
I geothermal

Figure 2.2. Reference Scenario primary energy consumption by region (direct equivalent).

47


-------
2.4.2 Processing and conversion of energy carriers

In this section, we examine electricity production, refining, and hydrogen production by
aggregated technology category.

Electricity production

Figure 2.3 shows GLIMSEvl-Reference electricity generation by aggregate subsector. Categories
are fuel-based, but differentiate between onshore and offshore wind, as well as by type of solar
power. PV stands for photovoltaic technologies, while CSP indicates concentrated solar power
technologies. National electricity production grows by more than 60% from 2015 to 2050.
Generation from coal and nuclear power diminish over the modeling horizon as existing plants
retire and are replaced with increased generation from wind, solar, and natural gas. By 2050,
renewables constitute more than 50% of generation. This result is impacted by the use of
updated electric sector technology costs that reflect reduced capital costs for wind and solar.

Electricity generation by aggregated subsector rnw detail(GCAM-USA)
GLIMPSEvI -Reference
region: Total

26
24
22
20
18
16
14
12
10

wind onshore
wind offshore
solar rooftop PV
solar PV
¦ solar CSP
i refined liquids
i nuclear
i hydrogen
I hydro
I geo
gas
I coal
i biomass

Figure 2.3 Reference Scenario electricity generation by aggregate subsector.

48


-------
Refined liquids production

In GCAM and GCAM-USA, liquid fuel production technologies include those shown in Fig 2.4.
The fuels that they produce are then lumped together and referred to as "refined liquids".
Refined liquids can be used by any sector. FT biofuels are those produced by the Fischer-
Tropsch process. In GLIMPSEvl-Reference, domestic liquid fuel production decreases by 21%
from 2015 to 2050, driven by vehicle efficiency improvements, fuel switching, and
electrification of the transportation sector. The primary source of liquid fuels is oil refineries.
Corn ethanol production is small in comparison and is steady over the modeling horizon until
2040, when it begins to grow. Gas-to-liquid, cellulosic ethanol, and FT biofuels have small
production shares.

Refined liquids production by tech (GCAM-USA)

GLIMPSEvI -Reference
region: Total

38
36
34
32
30
28
26
24
Uj 22

I 20

•Zl

"3 18
o

10
14
12
10
8
0
4
2

0 *

Figure 2.4 Reference Scenario liquid fuel production by technology.

corn ethanol
cellulosic ethanol
biodiesel
I FT biofuels
oil refining
gas to liquids

49


-------
Hydrogen production

In GCAM-USA 5.4, H2 production is represented at the USA level only. Technologies in
GLIMPSEvl-Reference that produce H2 include: natural gas steam reforming, thermal splitting
(which uses the heat produced by nuclear power), and biomass technologies, with and without
CCS. While production grows dramatically from 2020 in the GLIMPSE Reference Scenario, the
overall quantity of H2 production in 2050 is still less than a tenth of refined liquids from
conventional refineries. The majority of H2 is produced by steam reforming.

Hydrogen production by tech
GLIMPSEul -Reference
region: Total

2.4

2.2

Figure 2.5 Reference Scenario hydrogen production by technology.

50


-------
2.4.3 Final energy

This section, we show the total final energy consumption by sector over time, independent of
fuel. We then examine this information further by inspecting the use of specific fuels by sector.

Total energy consumption by sector

Final energy is the energy used in various sectors to meet end-use demands, such as space
heating, water heating, lighting, and transportation. Here, transport-LDV refers to passenger
cars and trucks. Transport-HDV refers to trucks used for delivering freight, and these vehicles
are further categorized as being small, medium, or large. Transport-ALM refers to the
combination of air, locomotive (rail), and marine vehicles. In 2015, industry, buildings (the
residential and commercial sectors), and transportation each are responsible for approximately
one-third of final energy use. The industrial final energy usage grows, while most other sectors
stay relatively constant. Energy use by transport-LDV decreases as those vehicles become more
efficient. Fuel use in transport-HDV and in transport-ALM is steady over time. Overall, final
energy is relatively constant through 2035. From 2035, however, the final energy consumption
overtakes efficiency improvements, resulting in steady growth in consumption.

80
75
70
65
60
55
50
45

UJ

~ 40

Cl

.E 35
30
25
20
15
10
5
0

Figure 2.6 Reference Scenario final energy consumption by sector.

Final energy consumption by aggregate sector
GLIMPSEvI -Reference
region: Total

I transport-LDV
I transport-HDV
I transport-ALM
i residential
industry
buildings

51


-------
Electricity use by sector

The residential and commercial sectors are responsible for more than two-thirds of use
electricity use in 2015. By 2050, this share is closer to half as industry and transport grow in
electricity use over the time horizon. Fuel production refers to fuel extraction and refining
activities, which use very little electricity.

Electricity use by aggregate sector
GLIMPSEvI -Reference
region: Total

I transport-LDV
l transport-HDV
i transport-ALM
residential
industry

i fuel production
commercial

Figure 2.7 Reference Scenario electricity use by aggregate sector.

52


-------
Coal use by sector

Coal use declines significantly over the modeling horizon. The largest decreases are prior to
2020, when coal loses market share to natural gas and, to a lesser extent, wind. Post 2025, coal
declines accelerate, driven by retirements. Industrial coal use is steady over time, although
there is a slight increase. Only a negligible amount of coal is used in the buildings sectors.
Exported coal is not shown.

Coal use by aggregate sector
GLIMPSEvI -Reference
region: Total

17
16
15
14
13
12
11

^ 10

a B

E 8

7
6
5
4
3
2
1
0

Figure 2.8 Reference Scenario coal use by aggregate sector.

53


-------
Natural gas use by sector

Natural gas is used across many sectors of the economy. Use in electricity production grows
significantly from 2015 to 2020. Use in industry, transportation, and fuel production increases
over time as well. In contrast, commercial natural gas use is steady, while residential natural gas
use decreases as more residential end-use services (e.g., space and water heating) are met with
electricity.

Natural gas use l>y aggregate sector
GLIMPSEul -Reference
region: Total

¦	transport-HDV
residential
industry

¦	fuel production

¦	electricity
commercial

Figure 2.9 Reference Scenario natural gas use by aggregate sector.

54


-------
Refined liquids use by sector

GCAM does not differentiate whether the liquid fuels are gasoline, diesel, or biofuels. The
transportation sector used more than three-quarters of liquid fuels in 2015. Use in light duty
transportation declines over time as vehicles become more efficient and alternative fueled
vehicles achieve greater market share. Use of liquid fuels in industry grows slowly but steadily.
Much of this industrial refined liquid use is in construction, agriculture, and mining.

Refined liquids use by aggregate sector
GLIMPSEvI -Reference
region: Total

2

OOOOOOOO

-<¦	ro	ro	cj	to	ii	-t.	en

ChOChOChOOio

Figure 2.10 Reference Scenario refined liquids use by aggregate sector.

¦	transport-LDV

¦	transport-HDV

¦	transport-ALM

¦	residential
industry

¦	electricity
commercial

55


-------
Biomass use by sector

The industrial sector is the greatest user of biomass in the early years. However, use in liquid
fuel production grows significantly from 2040 as advanced biofuel technologies become
available in the model.

Biomass use by aggregate sector
GLIMPSEvI -Reference
region: Total

Figure 2.11 Reference Scenario biomass use by sector.

56


-------
H2 use by sector

The transportation sector is the greatest user of H2, and more than three-quarters of total H2
use occurs in onroad transportation. Please note that the H2 representation in GCAM-USA is
being revised, and versions after 5.4 may reflect additional uses, such as in industry and power
generation.

Hydrogen use by aggregate sector
GLIMPSEvI -Reference
region: Total

2.4
2.2
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2

0.0 1-

I transport-LDV
I transport-HDV
I transport-ALM
industry

Figure 2.12 Reference Scenario hydrogen use by sector.

57


-------
Queries are also available for examining specific sectors and their fuel use, as is shown in the
following graphics.

Energy use in the industrial sector

Overall, industrial energy use increases approximately 34% from 2015 to 2050. Use of all fuels
increases over that time horizon, but the increase is dominated by electricity (2.2 EJ), refined
liquids (2.2 EJ), and natural gas (2.0 EJ).

End-use energy consumption in industry
GLIMPSEvI -Reference
region: Total

OOOOOOOO

-<¦	ro	ro	cj	to	ii	-t.	en

chooiochooio

Figure 2.13 Reference Scenario industrial sector energy use by fuel.

¦	refined liquids

¦	hydrogen

¦	gas

¦	electricity

¦	coal
bioniass

58


-------
Energy use in residential and commercial buildings

Overall, energy use in buildings increases approximately 7% from 2015 to 2050. Of the fuels
used in buildings, only electricity use increases over that time horizon.

End-use energy consumption in buildings
GLIMPSEvI -Reference
region: Total

I refined liquids
gas

electricity
I coal
biomass

Figure 2.14 Reference Scenario buildings energy use by fuel.

59


-------
Energy use in in the onroad light-duty transportation sector

The on road light-duty vehicle sector includes the "Car" and "Large Car and Truck" passenger
vehicle categories. Overall, light-duty vehicle energy use decreases by 31% from 2015 to 2050.
Use of refined liquids decreases by more than 50%. By 2050, electricity and hydrogen account
for 29% of fuel use.

End-use energy consumption in transportation (detail)
GLIMPSEirt -Reference
trn_|)ass_road_LDV_4W
region: Total

I refined liquids
I hydrogen
electricity

Figure 2.15 Reference Scenario light-duty transportation energy use by fuel.

60


-------
Energy use in the onroad freight transportation sector

The onroad heavy-duty vehicle (HDV) sector includes the "Small", "Medium", and "Large"
freight truck categories. Overall, HDV energy use decreases by 12% from 2015 to 2050. Use of
refined liquids decreases by more than 31%. By 2050, electricity and hydrogen account for 21%
of fuel use.

End-use energy consumption in transportation (detail)
GLIMPSEirt -Reference
trn_freiglit_roatl
region: Total

I refined liquids
i hydrogen
gas

electricity

£ £

Figure 2.16 Reference Scenario heavy-duty transportation energy use by fuel.

61


-------
Energy use in across air, locomotive, and marine transportation sectors

The "ALM" sector includes air, locomotive, and marine passenger and freight vehicles. Overall,
energy use in this sector increases by 26% from 2015 to 2050. Use of refined liquids increases
by 12%, or 0.67 EJ. By 2050, electricity and hydrogen account for approximately 12% of this
combined sector's fuel use.

Final energy consumption by aggregate sector and fuel
GLIMPSEvI -Reference
transport-ALM
region: Total

I refined liquids
I hydrogen
electricity

£ £

Figure 2.17 Reference Scenario energy use of fuel for the transport-ALM sector.

62


-------
2.4.4 Technology market shares for specific end-uses

In this section, we examine the market shares (in terms of service demands met) of competing
technologies for a selection of end-use markets. Here, we focus on the commercial, residential,
and transportation sectors. In GCAM-USA 5.4, the industrial sector has limited technological
detail, reflecting "industrial energy use" by fuel, as well as fuel used in the cement and fertilizer
industries. We do not present additional results for the industrial sector, but GLIMPSE users are
encouraged to use the "Industry final energy by tech and fuel" to explore industrial fuel use
further. Future versions of GCAM-USA are expected to have additional detail in the industrial
sector.

Commercial space cooling

Overall, commercial space cooling demands increase by 37% from 2015 to 2050. The greatest
increase is in high-efficiency air conditioning (0.71 EJ), output of which nearly triples from 2015
to 2050.

Building service output by tech
GLIMPSEvI -Reference
comm cooling
region: Total

3.2 1

3.0

2.8

2.6

2.4

2.2

2.0

Figure 2.18 Reference Scenario service output for commercial space cooling technologies.

63


-------
Commercial space heating

Overall, commercial space heating demands increase by 29% from 2015 to 2050. The greatest
increases are from high-efficiency gas furnaces (0.72 EJ) and from electric heat pumps (0.43 EJ).

Building service output hytech
GLIMPSEvI -Reference
comm heating
region: Total

I wood furnace
gas furnace hi-eflf
gas furnace
I fuel furnace
electric heat pump
electric furnace
I coal furnace

Figure 2.19 Reference Scenario service output for commercial space heating technologies.

64


-------
Commercial water heating

Overall, commercial water heating demands increase by 43% from 2015 to 2050. The greatest
increase is from high-efficiency natural gas water heaters (0.35 EJ).

Building service output by tech
GLIMPSEvI -Reference
comm hot water
region: Total

0.30

OOOOOOOO
-j-roroojcoiJCkCh
ChOUiOOiOOiO

gas water heater hi-eff ¦ gas water heater ¦ fuel water heater electric resistance water heater
electric heat pump water heater

Figure 2.20 Reference Scenario service output for commercial water heating technologies.

65


-------
Commercial lighting

Solid state technologies include Light-Emitting Diodes (LED) lighting. Fluorescent includes
compact (CFL) and linear (LFL) bulbs. Overall, commercial lighting demands increase by 49%
from 2015 to 2050. While some market share remains through 2050 for fluorescent and
incandescent lighting, the only category that grows over that period is solid state lighting (12.4
petalumens-hours).

Building service output hy tech
GLIMPSEvI -Reference
comm lighting
region: Total

solid state
i incandescent
fluorescent

Figure 2.21 Reference Scenario service output for commercial lighting technologies.

66


-------
Residential space cooling

Overall, residential cooling demands increases by 43% from 2015 to 2050. The high-efficiency
version of electric air conditioning achieves only a very small market share.

Building service output l>y tech
GLIMPSEvI -Reference
resid cooling
region: Total

5.5
5.0
4.5

3.5
B 3.0

1"

o

2.0
1.5
1.0
0.5
0.0

air conditioning hi-eff
air conditioning

Figure 2.22 Reference Scenario service output for residential space cooling technologies.

67


-------
Residential space heating

Overall, residential heating demands increase by 25% from 2015 to 2050. Output from electric
heat pumps increases by 2.8 EJ. Heating from high-efficiency natural gas furnaces grows in
output as well, but only by 0.36 EJ. Overall, heating provided by natural gas decreases by 45%
from 2015 to 2050.

Building service output hy tech
GLIMPSEvI -Reference
resid heating
region: Total

I wood furnace
gas furnace hi-eff
gas furnace
i fuel furnace hi-eff
i fuel furnace
electric heat pump
electric furnace
I coal furnace

Figure 2.23 Reference Scenario service output for residential space heating technologies.

68


-------
Residential water heating

Overall, residential water heating demand increases by 35% from 2015 to 2050. Much of the
growth over time is met by high-efficiency electric water heaters. However, starting in 2035,
output from electric heat pump water heaters begins to grow.

Building service output liy tech
GLIMPSEvI -Reference
resicl hot water
region: Total

o.e

0.4
0.2
~ .0 -

gas water heater hi-eff ¦ gas water heater ¦ fuel water heater hi-eff ¦ fuel water heater

electric resistance water heater hi-eff electric resistance water heater electric heat pump water heater

Figure 2.24 Reference Scenario service output for residential water heating technologies.

69


-------
Residential lighting

Solid state technologies include Light-Emitting Diodes (LED) lighting. Fluorescent includes
compact (CFL) and linear (LFL) bulbs. Overall, residential lighting demand increases by 47% from
2015 to 2050. Output from fluorescent bulbs is relatively constant over time. Incandescent
bulbs have been significantly displaced by solid state lighting by 2025.

Building service output hy tech
GLIMPSEvI -Reference
resid lighting
region: Total

5.0

kj	ro	ro	ro	ro	ro	ro	ro

oooooooo

-J-NiKJCilCdiiOl

oiaoiooiooia

Figure 2.25 Reference Scenario service output for residential lighting technologies.

70


-------
Onroad light-duty technologies

Conventional vehicles that operate on gasoline and diesel dominate the market through 2020,
from which hybrid and electric vehicle market share increases steadily. Hydrogen fuel cell
vehicles begin to penetrate the market in 2030, achieving a market share (based on pass-km) of
approximately 11% in 2050.

Passenger car and truck service output hy tech
GLIMPSEvI -Reference
trn_|)ass_road_LDV_4W
region: Total

Figure 2.26 Reference Scenario service output for light-duty passenger vehicle technologies.

71


-------
Onroad freight truck technologies

This graphic includes onroad freight demand met by technology across three size categories:
"Light" and "Medium" and "Heavy". The load factors for these three sizes differ, and are 0.27,
2.07, and 4.16 tonnes per vehicle. Adoption of alternative fuel vehicles across these size classes
will differ. The "Transport service output by tech" query can be used to explore market shares
for each class separately. Conventional vehicles that operate on gasoline and diesel dominate
the market in 2015. From 2020, however, hybrid vehicle market share grows, eventually
representing more than a third of overall ton-km. Electric and fuel cells begin to appear in 2020.
By 2050, they have reached 17% and 14% of overall onroad ton-km, respectively.

3,400.000
3,200.000
3,000.000
2,800,000
2,600,000
_ 2,400,000
5 2,200,000
§ 2,000,000
o 1,800,000
^ 1,600,000

Freight truck service output by tech (no bus)
GLIMPSEvI -Reference
trn_freiglit_roacl
region: Total

Figure 2.27 Reference Scenario service output for heavy-duty truck technologies.

72


-------
Bus technologies

Overall demand for bus travel peaks in 2025, then declines slowly. Alternative fuels play an
increasing role in this sector. By 2050, hybrids, electric, and fuel cell buses each represent
approximately 25% of the category's service demand.

1,200,000

1,100,000

1,000,000

900,000

^ 800,~~~
to

ro 700,~~~

•Zl

c.

S 600,000
~ 500.000 |

=3
•Zl

"3 400,000 I
o

300,000 I
200,000 I
100,000 I

0

Transport service output by tech
GLIMPSEvI -Reference
Bus
region: Total

Figure 2.28 Reference Scenario service output by bus technologies.

73


-------
Domestic aviation technologies

Overall demand for domestic aviation more than doubles from 2015 to 2050. Refined liquid
technologies dominate the market, although there is significant growth in hydrogen and electric
plants, after 2035. Note that biofuels are not represented as a separate end-use fuel, but that
biofuels represent a portion of refined liquids.

Transport service output by tech
GLIMPSEvI -Reference
Domestic Aviation
region: Total

Figure 2.29 Reference Scenario service output by domestic aviation technologies.

74


-------
International aviation technologies

Overall demand for international passenger aviation nearly doubles from 2015 to 2050.
Hydrogen and electric planes do not achieve a meaningful share of the market by 2050.

Transport service output by tech
GLIMPSEvI -Reference
International Aviation
region: Total

Figure 2.30 Reference Scenario service output by international aviation technologies.

75


-------
Freight rail technologies

Freight rail service demand increases by 60% from 2015 to 2050. Hybrid and alternative fuel
technologies begin to appear in the 2025 model year. By 2050, fuel cells, electric, hybrid, and
convention rail all have similar market shares.

3,800,000
3,600,000
3,400,000
3,200,000
3,000,000
2,800,000

-g 2,600,000
2,400.000

£Z

3 2,200,000

| 2,000,000
1,800,000

^ 1,000,000

3- 1,400,000

=3

° 1,200,000
1,000,000
800,000
000,000
400,000
200,000
0 *

Transport service output by tech
GLIMPSEvI -Reference
Freight Rail
region: Total

Figure 2.31 Reference Scenario service output by freight rail technologies.

76


-------
Domestic marine technologies

Demand for domestic marine freight doubles between 2015 and 2050. From 2020, hybrid
technologies begin to appear and by 2035 have gained more than half the market. Hybrids
remain dominant, but electric and fuel cell ships begin to achieve a growing market share in
2040.

Transport service output by tech
GLIMPSEvI -Reference
Domestic Ship
region: Total



2,200,000



2,000,000



1,800,000



1,600,000

E







£Z
O

1,400,000

£=



O

1,200,000

I

1,000,000





Q.



~=3

800,000

O



600,000



400,000



200,000

Figure 2.32 Reference Scenario service output by domestic marine shipping technologies.

77


-------
International marine technologies

Demand for international shipping grows at a smaller rate than domestic shipping, increasing
42% from 2015 to 2050. By 2035, hybrid technologies will provide more than half of the ton-
km. By 2050, that has grown to two-thirds.

Transport service output by tech
GLIMPSEvI -Reference
International Ship
region: Total

20,000,000

Figure 2.33 Reference Scenario service output by international marine shipping technologies.

78


-------
2.4.5 Emissions of CO2 and air pollutants

CO2 emissions

CO2 emissions are produced at the state level, as well as from the "USA" region that includes
several source categories that have not been disaggregated to the state level (e.g., oil and gas
production, coal mining, and H2 production).

US CO2 emissions are estimated to decline steadily from 2015 to 2050, ultimately falling 26%.
Note that emissions of CO2 are presented in units of Megatonnes of Carbon (MTC). To convert
to MTCO2, multiply these values by 44/12.

C02 emissions by region
GLIMPSEvI -Reference

1,500
1,400
1,300
1,200
1,100
1,000

O
I—

IE 900

CO

£= 800
o

$ 700
aj 600
500
400
300
200
100
Q *

Total

Figure 2.34 Reference Scenario CO2 emissions.

79


-------
In Figure 2.35, CO2 emissions are aggregated by sector. Negative emissions from biomass
growth reflect the CO2 that is removed from the atmosphere when growing the biomass used
for bioenergy. This graphic was created using the results of two queries: "All emissions by
aggregate sector" and "All emissions by resource production." The latter query returns
emissions from coal, natural gas, and crude oil operations, which are not reported by the "All
emissions by aggregate sector" query.

In the scenario, CO2 emissions decline over time, driven by reductions from light duty
transportation (transport-LDV) and the electric sector. Economic growth drives an increase in
industrial CO2 emissions. Other sectors remain relatively constant despite increasing energy
service demands.

1600

1400

1200

1000

j? 800
2

CO

C 600
o

'to

E 400

LU

200
0

-200
-400

All emissions by aggregate sector AND All emissions by resource production
GLIMPSEvl-Refrence
C02 emissions
region: Total

¦lllllll

biomass growth
1 commercial
1 electricity
1 fuel production

industry
1 residential
I transport-ALM
1 transport-HDV
1 transport-LDV

2015 2020 2025 2030 2035 2040 2045 2050

Figure 2.35 Reference Scenario CO2 emissions by sector.

80


-------
NOx emissions

NOx emissions decline dramatically through 2035. Onroad vehicles (transport-LDV and
transport-HDV) and the electric sector are primarily responsible for this decline. For onroad
vehicles, the Tier 3 engine and mobile vehicle fuel standards drive this trend. After 2035,
additional reductions come from vehicle electrification. In the electric sector, fuel switching
from coal to natural gas and renewables is driving emissions lower. Industry and air-locomotive-
marine (transport-ALM) together emitted nearly 50% of total anthropogenic NOx in 2015. This
percentage grows to approximately 70% in 2050 as emissions from other sectors decline.

All emissions by aggregate sector AND All emissions by resource production

GLIMPSEvl-Reference
NOx emissions
region: Total

¦	electricity

¦	fuel production
industrial processes

¦	industry

¦	commercial

¦	residential

¦	urban processes

¦	transport-ALM

¦	transport-HDV

¦	transport-LDV

Figure 2.36 Reference Scenario NOx emissions by sector.

81


-------
SO2 emissions

Coal-fired power generation dominates SO2 emissions in 2015. However, emission limits
combined with the retirement of coal plants reduces the power sector's contribution by
approximately two-thirds by 2020. Industrial SO2 grows slowly over the time horizon, as does
transport-ALM, which is driven by international shipping.

All emissions by aggregate sector AND All emissions by resource production

GLIMPSE-Reference
S02 emissions
region: Total

3.5

¦

2.5

GO

to

.2 2

j)



£

LU

1.5

0.5



2015 2020 2025 2030 2035 2040 2045 2050

I electricity
I fuel production
industrial processes
industry
i commercial
I residential
I transport-ALM
Itransport-HDV
itransport-LDV

Figure 2.37 Reference Scenario SO2 emissions by sector.

82


-------
PM2.5 emissions

The urban processes category represents approximately half of PM2.5 emissions. This category
includes waste incineration, landfills, and wastewater treatment. Of the remaining half,
industrial sources, including combustion and processes, are the primary PM sources. Note that
this category includes construction, mining, and agriculture. Residential wood combustion also
accounts for a large share of PM2.5 emissions. Emissions from this category decrease over time,
driven both by cleaner wood-burning technologies and by a transition to greater use of electric
heating technologies.

All emissiosn by aggregate sector AND All emissions by resource production

GLIMPSE-Reference
PM2.5 emissions
region: Total

1.6

1.4

1.2

o 0.8

0.6

0.4

0.2

¦

¦iiiiii

2015 2020 2025 2030 2035 2040 2045 2050

i electricity
I fuel production
industrial processes
industry
I commercial
i residential
urban processes
Itransport-ALM
Itransport-HDV
Itransport-LDV

Figure 2.38 Reference Scenario PM2.5 emissions by sector.

83


-------
2.5 Evaluation of GLIMPSEvl-Reference

2.5.1 Comparison with the Annual Energy Outlook 2023

In this section, national-scale results from GLIMPSEvl-Reference are compared to the U.S.
Energy Information Administration's 2023 Annual Energy Outlook, or AEO 2023
(https://www.eia.gov/outlooks/aeo/tables ref.php).

Electricity production

Electricity production in GLIMPSEvl-Reference is greater than that in AEO2023. Much of the
increased demand is being met by electricity production from natural gas technologies.
AEO2023 includes aspects of the Inflation Reduction Act, which may be driving some of the
differences.

25

Electricity production by fuel category (EJ)
AEO2023

2025 2030 2035 2040 2045 2050

Electricity production by fuel category (EJ)
GLIMPSEvl-Reference

25

2030 2035 2040 2045

Difference in electricity production by fuel category
(EJ) GLIMPSEvl-Reference minus AEO2023

Legend

¦	Coal

¦	Nuclear Power

I Petroleum
Renewable Sources

I Natural Gas
] Other

2025 2030 2035 2040 2045 2050

Figure 2.39 Comparison of electricity production by fuel category with AEO2023.

84


-------
Electricity use by sector

Electricity use in GLIMPSEvl-Reference is greater than in AEO2023. A significant portion of the
difference is a result of transportation sector electricity demands. GLIMPSEvl-Reference also
has greater electricity demands in the residential, commercial, and industrial sectors.

Electricity use by sector (EJ)
AEO2023

mil!

2025 2030 2035 2040 2045 2050

Difference in electricity use by sector (EJ)
GLIMPSEvl-Reference minus AEO2022

-12

2025 2030 2035 2040 2045 2050

Electricity use by sector (EJ)
GLIMPSEvl-Reference

22

16

2025 2030 2035 2040 2045 2050

Legend

¦ commercial ¦ industry ¦ residential ¦ transport

Figure 2.40 Comparison of electricity use by sector with AEO2023.

85


-------
Energy use in buildings

Energy use in residential and commercial buildings is lower in GLIMPSEvl-Reference than in
AEO2023. GLIMPSEvl-Reference predicts greater electrification and lower demand for natural
gas and refined liquids.

10

Energy use in buildings by fuel (EJ)
AEO2023

iiiiii

Energy use in buildings by fuel (EJ)
GLIMPSEvl-Reference

HUH

10

Difference in energy use in buildings by fuel (EJ)
GLIMPSEvl-Reference minus AEO2023

Legend

¦ biomass acoal ¦ electricity Hgas ¦ refined liquids

2025	2030	2035	2040	2045	2050

Figure 2.41 Comparison of energy use in buildings by fuel with AEO2023.

86


-------
Energy use in industry

GLIMPSEvl-Reference projections for industrial energy use are very similar in magnitude to
AEO2023, although there are some differences in which fuels are being used.

Industrial energy use by fuel (EJ)
AEO2023

-III!1

Industrial energy use by fuel (EJ)
GLIMPSEvl-Reference

miir

2025	2030	2035	2040	2045	2050

Difference in industrial energy use (EJ)
GLIMPSEvl-Reference minus AEO2023

Legend

ibiomass acoal ¦ electricity Bgas ¦ hydrogen ¦ refined liquids

Figure 2.42 Comparison of industrial energy use by fuel with AEO2023.

87


-------
Energy use in transportation

Transportation sector fuel use is less in GLIMPSEvl-Reference than in AEO2023. Contributing to
this difference is greater market shares for electric and hydrogen fuel cell vehicles, which are
more efficient than conventional internal combustion engines.

Transportation energy use by fuel (EJ)
AEO2023

20

Transportation energy use by fuel (EJ)
GLIMPSEvl-Reference

30

2025 2030 2035 2040 2045 2050

Difference in transportation energy use (EJ)
GLIMPSEvl-Reference minus AEO2023

Legend

¦ electricity agas ¦ hydrogen ¦ refined liquids

Figure 2.43 Comparison of transportation energy use by fuel with AEO2023.

88


-------
Onroad light-duty demand and vehicle technologies

Demand for onroad passenger travel is slightly higher in GLIMPSEvl-Reference than in
AEO2023. The market shares of hybrid (hybrid liquids), electric (BEV), and fuel cell (FCEV)
vehicles are much greater, however.

Onroad service demand (mil-pass-km)
AEO2023

12,000,000

4,000,000

2,000,000

2025 2030 2035 2040 2045 2050

Difference in onroad service demand (mil-pass-km)
GLIMPSEvl-Reference minus AEO2023

4,000,000

-4,000,000

2025 2030 2035 2040 2045 2050

12,000,000

10,000,000

8,000,000

6,000,000

4,000,000

Onroad service demand (mil-pass-km)
GLIMPSEvl-Reference

2025 2030 2035 2040 2045 2050

Legend

¦ BEV ¦ FCEV ¦ Hybrid Liquids ¦ Liquids ¦ Other

Figure 2.44 Comparison of passenger car and truck service output by fuel with AEO2023.

89


-------
2.5.2 Comparison of air pollution outputs with the EPA's 2016v2 emissions modeling platform

In this section, we evaluate GLIMPSEvl-Reference national and sectoral emission projections by
comparing them with the U.S. EPA 2016v2 emission platform (http://epa.gov/air-emissions-
modeling/2016v2-platform). EPA developed the 2016v2 platform to support regulatory air
quality modeling for the 2022 Good Neighbor Rule (http://www.epa.gov/csapr/good-neighbor-
plan-2015-ozone-naaqs). The platform includes a 2016 air pollutant emission inventory based
upon the EPA's National Emissions Inventory (NEI), as well as projections to 2023, 2026, and
2032.

National totals

In general, GLIMPSEvl-Reference emission projections are similar to those of the 2016v2
platform, both in magnitude and trends through time. However, SO2 emissions tend to be
higher than the platform, which PM2.5, VOC, CO, and NH3 are lower.

NOx - Total

12,000,000
10,000,000
8,000,000
6,000,000
4,000,000
2,000,000

S02 - Total

5,000,000
4,000,000
3,000,000
2,000,000
1,000,000

O o

PM2.5-Total

2,000,000
1,500,000
1,000,000
500,000

VOC - Total

CO - Total

NH3-Total

12,000,000
10,000,000
8,000,000
6,000,000
4,000,000
2,000,000

V

40,000,000
35,000,000
30,000,000
25,000,000
20,000,000
15,000,000
10,000,000
5,000,000

400,000
350,000
300,000
250,000
200,000
150,000
100,000
50,000

Figure 2.45 Comparison of GLIMPSEvl-Reference emission projections (lines) with those of the
EPA 2016v2 emission platform (dots).

90


-------
Next, we compare sectoral emissions by species. For NOx, most sectors match the platform
well, although NOx from industry and buildings are low compared to the platform. The air, rail,
and marine categories initially match well, but diverge over time. The reason for this divergence
is that the emission standards for vehicles are specified per unit of fuel consumed. As advanced
technologies achieve greater market share in the GLIMPSE results, emissions from these sectors
decrease.

NOx total and by sector

NOx - Industry

12,000,000
10,000,000
8,000,000
6,000,000
4,000,000
2,000,000

4,000,000
3,500,000
3,000,000
2,500,000
2,000,000
1,500,000
1,000,000
500,000

1,600,000
1,400,000
1,200,000
1,000,000
800,000
600,000
400,000
200,000

200,000
150,000
100,000
50,000

3,500,000
3,000,000
2,500,000
2,000,000
1,500,000
1,000,000
500,000

700,000
600,000
500,000
400,000
300,000
200,000
100,000

NOx - Buildings

600,000
500,000
400,000
300,000
200,000
100,000

350,000
300,000
250,000
200,000
150,000
100,000
50,000

Figure 2.46 Comparison of GLIMPSEvl-Reference NOx emission projections, in total and by
sector (lines), with those of the EPA 2016v2 emission platform (dots).

91


-------
S02 total and by sector

Evaluation of GLIMPSEvl-Reference SO2 emissions against the 2016v2 platform provides similar
results for most categories. Exceptions are SO2 from the industrial and buildings sectors. Many
refineries are transitioning to ultra-low sulfur fuel oil, a transition that is not yet fully captured
in the GLIMPSEvl-Reference scenario.

Figure 2.47 Comparison of GLIMPSEvl-Reference SO2 emission projections, in total and by
sector (lines), with those of the EPA 2016v2 emission platform (dots).

92


-------
Other pollutant totals and by sector

The comparisons for other pollutants are similar, although there may be specific reasons for
differences. For example, we expect that the GLIMPSEvl-Reference scenario's VOC emissions
from buildings are below the 2016v2 emission platform because of incomplete coverage of
evaporative emissions from paints. Reducing the remaining discrepancies will be a goal of
future updates.

Figure 2.48 Comparison of GUMPSEvl-Reference PM2.5 emission projections, in total and by
sector (lines), with those of the EPA 2016v2 emission platform (dots).

Figure 2.49 Comparison of GUMPSEvl-Reference VOC emission projections, in total and by
sector (lines), with those of the EPA 2016v2 emission platform (dots).

93


-------
40,000,000
35,000,000
30,000,000
25,000,000
20,000,000
15,000,000
10,000,000
5,000,000

2010 2020 2030 2040

700,000
600,000
500,000
400,000
300,000
200,000
100,000

CO - Industry

20,000,000

15,000,000 o O O °
10,000,000
5,000,000

2010

2020 2030

CO - Buildings

3,500,000
3,000,000
2,500,000
2,000,000
1,500,000
1,000,000
500,000

2010 2020

25,000,000
20,000,000
15,000,000
10,000,000
5,000,000

700,000
600,000
500,000
400,000
300,000
200,000
100,000

140,000
120,000
100,000
80,000
60,000
40,000
20,000

50,000
40,000
30,000
20,000
10,000

Figure 2.50 Comparison of GUMPSEvl-Reference CO emission projections, in total and by
sector (lines), with those of the EPA 2016v2 emission platform (dots).

2.6 Discussion and Summary

The GLIMPSE Reference Scenario is intended to simulate a cohesive and plausible evolution of
energy demands, fuels, and technology market shares through 2050. Thus, it represents a
baseline against which scenarios with alternative assumptions about policies, technology
advances, and various socio-economic drivers can be compared and evaluated.

Overall, GUMPSEvl-Reference results are generally consistent with AEO2023, with differences
largely being explainable. For example, AEO's inclusion of Inflation Reduction Act of 2022 (IRA)
provisions in the electric sector result in less coal and natural gas than projected by GLIMPSE. In
other sectors, GUMPSEvl-Reference tends to transition from fossil fuels to electricity more
quickly than AEO, particularly in the transportation sector.

Note that there are real-world barriers and constraints that are not reflected in GCAM-USA, and
thus are not captured in GUMPSEvl-Reference. Examples include bottlenecks in the
transmission grid and limitations in the availability of funding for companies to build new
electricity production and refining capacity or for households to purchase more expensive
technologies.

There also are limitations to how GCAM-USA apportions market share to competing
technologies. For example, the logit will apportion at least some market share to each
competing technology. For light-duty transportation, this means that hydrogen fuel cell vehicles
will achieve a market share in the first year that they are available to the model. However,
GCAM-USA is not considering the potential barriers associated with introducing a refueling
infrastructure for only a small number of vehicles. Market share apportionment also considers
technology-specific shareweights, which are intended to represent factors such as bias and
unmodeled issues, as well as market-specific logit exponents, which reflect the sensitivity of the

94


-------
market to price differences between technologies. Values chosen for these parameters result in
technology shares that simulate historical choices reasonably. However, there is uncertainty in
determining the "best" values for each parameter, including how they transition over time.

As a result of these limitations, GCAM-USA results, including GLIMPSEvl-Reference, should be
considered in the context of the inclusions and omissions in model coverage, as well as
limitations associated with model formulation and parameterization. Considering these
limitations, some users may choose to develop their own Reference Scenario that involves
parameterizations, assumptions, or policies that are different from those used here. That is one
of the strengths of the GLIMPSE framework - users can explore alternative scenarios and
develop their own Reference Scenario and alternative scenarios.

When comparing GCAM-USA outputs with those of other models and data sources, there
undoubtedly will be differences. There are many potential reasons for these discrepancies,
including the inherent uncertainty in projecting energy system transformation for decades into
the future, and various projections use different assumptions, models, methods, and
formulations to make their projections. Nonetheless, understanding similarities and differences
can provide insights into the key factors that may drive future trends, as well as how a
particular projection falls within the range of possibilities.

95


-------
CHAPTER 3. HOW DOES GCAM WORK?

GCAM is a complex model and understanding how GCAM operates is beneficial to GLIMPSE
users in constructing scenarios and interpreting results. In this Chapter, we describe the general
solution process, including how GCAM steps through time, its market-based approach, how
market shares are determined via a logit equation, and how logit parameters such as
shareweights affect technology choice. This Chapter is a work in progress. Interested readers
also are encouraged to read PNNL's GCAM documentation, including the section on technology
choice (https://igcri.github.io/gcam-doc/choice.html).

3.1	General solution process

GCAM is a dynamic-recursive, market-based simulation model. This description refers to the
process by which GCAM steps through time. In each time period, GCAM solves for the vector of
prices at which quantities supplied equals quantities demanded for all modeled markets. When
all markets have been solved, GCAM transitions to the next time period, starting with the
solution from the prior period.

3.2	Markets in GCAM

To illustrate how markets work in GCAM, we will focus on one sector, passenger travel demand.
The figure below illustrates the hierarchical structure of this sector. Horizontal boxes represent
various subsectors and modes, while the vertical boxes represent technologies.

96


-------
Passenger travel demand

/(populationrcost)

Walking



Cycling



Road

Figure 3.1 Hierarchical representation of passenger travel in GCAM. Passenger travel is
subdivided into various subsectors, modes, and technologies. Options at each level compete
for market share.

The quantity of passenger travel demand, represented in units of passenger-km, is a function of
the size of the population and the cost of travel. As population increases, the quantity of
demand for passenger travel increases. However, if travel costs increase, demand for passenger
travel per person decreases; and, if travel costs decrease, per capita demand increases.

The "market" for passenger travel is apportioned across a range of modes, including "Walking",
"Cycling", "Road", "Domestic Air", "International Air", "High Speed Rail (HSR)", and "Passenger
Rail". The relative costs of these options influence the mixture used to fulfill overall passenger
travel demand; decreasing the travel costs in any of these modes will increase its share
compared to competing modes.

The "Road" mode is subdivided further, with "Bus" competing with "Light-duty". Within "Light-
duty", "2&3W" (2 or 3-wheeled vehicles, such as motorcycles) compete with "4W" (passenger
vehicles with 4 wheels, such as cars, SUVs, and trucks). For convenience, "4W" vehicles are
lumped into the categories "Car" and "Large Car and Truck".

At the bottom of the hierarchy are the technologies that are competing against each other. For
example, for the "Car" category, "Liquids" (conventional internal combustion engines) are
competing with "Hybrid", "CNG", "Fuel Cell", and "EV". This is the most granular level in
GCAM's market-based structure. The mix of technologies selected at this level determines the
levelized cost ($/passenger-km) of meeting travel demand by "Car". How this cost compares to
that of "Large Car and Truck" determines how "4W" demand is apportioned as well as the
overall cost of "4W". The relative costs of "4W" and "2&3W" determines apportionment of

97


-------
"Light-duty" across those two categories. In this manner, costs cascade up the hierarchy,
determining how passenger travel demand is allocated to each branch of the tree.

3.3 Determining market share

In determining the relative market share of new purchases of competing technologies, GCAM
uses a logit function. For example, the share of new "Car" purchases assigned to technology /' is
calculated using the modified logit formula:

atipj

where: N is the total number of technologies competing for market share; s, is the market share
of technology /'; p, and py- are the prices of technologies /' and j, respectively; a/and a/are the
shareweights of technologies /' and j; and y is the logit exponent.

The price of each technology is its levelized cost, which is determined from the capital,
operation and maintenance, and fuel costs and represented in units of $/pass-km.

The technology-specific shareweight is a means of representing non-modeled issues, such as
logistical barriers for the adoption of a technology or bias against a new technology. By
convention, a value of 1 indicates that the technology is perfectly competitive, meaning there is
no bias and only relative costs come into play when determining the technology's market share.
In contrast, a value of 0 indicates that the technology will not be purchased regardless of cost.

If a technology is assumed to be purchased at a lower rate than its cost would suggest, then its
shareweight typically is a value between 0 and 1. For example, in the "Car" category, a 0.2
shareweight for the "Fuel Cell" technology could be used to reflect implicitly how the limited
availability of hydrogen would dampen market share. Similarly, a reduced shareweight for "EV"
could be used to reflect factors such as limited charging network and range anxiety.

The logit exponent controls the degree to which cost differences among technologies
determine their relative market shares. A value of 0 means that the logit function will ignore
cost differences. Assuming every technology has the same shareweight, they would be assigned
equal shares. Typical logit exponent values are -4, -6, and -8. The more negative the value, the
greater the weight that is given to cost differences. For the "Car" market, the logit exponent is
-8, indicating that consumer decisions on which technology to purchase are highly influenced
by cost. In contrast, the logit exponent that is used to allocate passenger travel between
"2&3W" and "4W" is -4, indicating that this market is less sensitive to price.

98


-------
3.4	Calibration

The shareweights for most technologies in GCAM are estimated during the calibration process.
For GCAM 5.4, the calibration year is 2015. In the calibration year, GCAM starts with estimates
of real-world market share for each technology. The technology with the greatest market share
in the calibration year is assigned a shareweight of 1. Thus, in the "Car" market, "Liquids" would
have a shareweight of 1.

Since the market shares and costs for competing technologies (e.g., "Hybrid", "CNG", ...) are
known for the calibration year, and since a logit exponent has been assumed, GCAM can solve
for the shareweights of each technology.

For technologies that did not exist in the calibration year, shareweights are set to 0. In the "Car"
market, the calculated "Hybrid" shareweight would be near 0, and the "Fuel Cell" and "EV"
shareweight would effectively be 0.

Note that the calculation of shareweight values is based upon overall stock of each technology
in the market in the calibration year, not sales in that year. As a result, the shareweights can be
biased toward historic factors as opposed to any new technology or developments that may
have impacted more recently.

3.5	Relaxing shareweights over time

To allow more flexibility over time, the shareweights for many technologies are set to follow a
linear path from their calibrated value to 1 over time, implying that logistical barriers and biases
in the calibration year would diminish over time. In many markets, technology shareweights are
set to 1 in 2050, although for some technologies 2100 is used to indicate a very long technology
development timeline is assumed until all barriers have been addressed. These decisions have
been made by the GCAM developers using modelers' judgement but can be overridden by
GCAM users.

The best trajectory to use for a technology's shareweight may be technology- and scenario-
dependent. For "Hybrid" cars, for example, there is no range anxiety or fueling infrastructure
barriers. When introduced, there was some hesitancy to adopt hybrid vehicles because they
represented a new technology. Arguably, this barrier is greatly diminished or no longer exists. It
would be reasonable to have the "Hybrid" shareweight transition to 1 by 2020 or 2025. In
GLIMPSE, we use "2025", "2035", and "2040" as the years at which "Hybrid", "EV", and "Fuel
Cell" technologies, respectively, achieve a shareweight of 1 for the "Car" and "Large Car and
Truck" categories.

99


-------
Shareweight trajectories may also be modified to reflect a particular scenario. For example, in a
scenario involving deep decarbonization, there could be additional public investment in
charging infrastructure and research and development in improving battery cost and range. In
such a scenario, it could be reasonable to adjust the shareweight trajectory for EVs to reach 1
sooner.

3.6 Illustrating the operation of the "Car" market

Next, we provide simplified description of the "Car" market to illustrate how the logit factors
into technology choice in GCAM.

The "Car" market in GCAM is depicted in the figure below. In the calibration year, 2015, "Car"
travel demand is met by the initial stock of vehicles. This stock is assumed to retire over time,
with that retirement curve represented by the red line and defined by the lifetime (e.g., 30
years), half-life (e.g., 7 years), and slope. Thus, the travel demand that is met by the initial stock
declines over time while the overall demand for "Car" travel increases.

[30 years)

Figure 3.2 Initial vehicle stock, retirement, and demand. The initial stock retires following an
s-curve that is defined by its half-life, slope, and vehicle lifetime.

At each modeled time period (typically 5 years), GCAM must purchase "Car" capacity to bridge
the gap between demand and what can be met by the "Car" fleet remaining from the previous
period.

100


-------
Service demand met by Cars

pass-km





	.	——"



1 X

<2020 vintage

¦^025 vintage

For each model period
GCAM purchases additional
capacity to meet demand









2015	2020	2025	2030	2035	2040	2045

(30 years)

Figure 3.3 Vintaging and demand for new vehicles. Each vintage retires following an s-curve.
The difference between demand and remaining stock is made up through the purchase of new
stock in each modeled year.

The logit function helps determine the technology composition and cost of "Car" travel for each
vintage.

To do this, GCAM takes an iterative approach to determine the "Cost" of car travel, An initial
guess is made. At this price, GCAM estimates the amount of demand that would be allocated to
"Car" in the passenger transportation hierarchy, as well as the supply of vehicles that would be
brought to market at that price. This supply is informed by the logit function.

In the illustration below, the guess for the initial price is low, resulting in demand for "Car"
outstripping supply.

101


-------
Market for new cars

E

Demand Q

VI

n
C.

~>-
—>

—¦

c

I nit. guess

Price of travel ($/pass-km)

Figure 3.4 An initial guess for a market-clearing price. At the initial guess, supply and demand
values are determined.

Next, a second guess is made. At this higher price, GCAM finds that supply would be greater
than demand.

Supply $

Low

High

102


-------
Market for new cars

VI
Mn

TO

a

c

TO

3

a

Demand Q

Supply

0 Demand

Supply

Low	Z Z High

2nd guess
Price of travel ($/pass-km)

Figure 3.5 A second guess for a market-clearing price. For the new price, quantities of supply
and demand are again estimated.

The guessing process continues via the bisection method until the difference between supply
and demand is within a specified tolerance, such as 0.001. Once this criterion is met, the market
is considered to have been solved.

103


-------
Market for new cars

VI
0

ra
a

c
ro
~
a

Demand ^

o

CO

o

Supply

Low

TF

High

Tolerance threshold: 0.01

| Supply - Demand | =0.007

Verdict: Stop

This market has been solved

5th guess
Price of travel ($/pass-km)

Figure 3.6 Identification of a market clearing price that solves this market. Guessing continues
until the difference between supply and demand is below the solution tolerance, indicating
that the market is cleared.

We can visualize how the logit apportions market share to each technology if we think of
technology costs as distributions.

104


-------
Supply market

Market shares are determined by
the relative areas under the distributions

Liquids Hybrid EV

Purchases:
Liquids — 90S
Hybrid - 10%
EV-0%

-Q
05
-Q
O

Q

>

M

Low

High

In it. guess

Price of travel ($/pass-km)

Figure 3.7 Conceptual diagram indicating how the logit allocates market share. The price at
which each technology can be brought to market is represented with a mean value and with a
distribution.

For example, at our initial guess for price, the composition of sales is determined based on the
relative areas under the curves. In this illustrative example, at the initial price, 90% of sales are
apportioned to "Liquids", 10% to "Hybrids". Because sales are represented as distributions, the
amount assigned to "EV" would be non-0, but very small.

The solution algorithm continues to search for a market clearing price, and, in this example,
would result in sales shares of 50% for "Liquids", 38% for "Hybrid", and 12% for "EV".

105


-------
Supply market

Market shares are determined by
the relative areas under the distributions

Liquids Hybrid EV	p

Purchases:
Liquids - 50K
Hybrid - 3B%
EV —12%

Market
cleared!

High

5th guess
Price of travel ($/pass-km)

Figure 3.8 Market shares of sales at the market-clearing price. Shares are determined based
upon the relative areas under the three distributions.

This visualization also allows us to depict how reducing a technology's shareweight would
impact apportionment of market share. For example, in the next figure, we illustrate market
shares in which the shareweight for "EV" has been reduced.

106


-------
>

-Q

fj

O
Q.

Figure 3.9 Impact of a reduced shareweight for EVs on market share allocation. The market
share for EVs is reduced as a result of its lower shareweight.

A policy such as a carbon tax directly influences market share decisions. By increasing the
operating costs for fossil fuels, such a policy would modify the levelized costs of each
technology. This would shift the distributions and result in a change in the sales share
apportioned to each technology.

Purchases:
Liquids — 56W
Hybfid - 42%
EV-2M

Supply market

Reduced share weight for EVs

Liquids Hybrid

1>

Price of travel ($/pass-km)

107


-------
Supply market

Price on carbon

Hybrid

Purchases:
Liquids - 2Q&
Hybrid - SOW
EV—20%

Liquids



:

/\ A EV "Vl
fl	EV

Low

High

Price of travel ($/pa55-km)

Figure 3.10 Impact of a price on carbon on market share allocation. The distributions shift
along the x-axis, reflecting changes in the costs associated with each technology that result
from the price of carbon.

The costs for technologies can also be shifted higher or lower based on user-specified
technology-specific taxes or subsidies. A tax or subsidy would directly affect the technology
allocations in the logit. Alternatively, GCAM offers the option of specifying a market share
constraint, and GCAM then solves to determine the subsidy necessary to achieve this share.

For example, in this illustration, GCAM solves for the subsidy that would shift the EV cost
distribution lower such that a 45% market share is achieved.

108


-------
Supply market

Achieving a 45% EV share

Figure 3.11 Impact of a technology subsidy on market share allocation. The subsidy shifts the
distribution of the subsidized technology, increasing its market share allocation.

3.7 Markets, Logits, Shareweights throughout GCAM

While this example has focused on the "Car" market, the logit functions are used throughout
GCAM in the solution of thousands of markets in every modeled time period. In passenger
transportation, logit functions are used to apportion market share within each horizontal box in
the figure below.

109


-------
Passenger travel demand
/!populatiofi,cost)

r=-2



a







" a

a \



a



7a—-w

a



Walking



Cycling



Road

Y = —4

y = -6

Dom. Air



Intl. Air

y = -3

HSR



Pass Rail

y=-fi

Figure 3.12 Shareweights and logit exponents across the passenger transportation sector.

Each sector, subsector, and technology has a shareweight, and each market is assigned a logit
exponent.

These mechanisms govern shares in other components of the energy system, as well as in
GCAM's representation of choice in allocating water and land use.

Residential heating

f{ population,cost)

Biomass



Coal



Electricity



Gas



S 5"

Si

CD ^

0) *5;
•=> u

2 c

Water

Water demands

Land use

Land use categories compete against each other

hnp://]gcri^ithu*).«o/grjni-(Joc/jiglu.hlml

Figure 3.13 Shareweights are present across GCAM's systems and sectors. The logit
determines market share allocations throughout GCAM.

110


-------
3.8 The logit function and market share constraints

The logit-based estimation process will attempt to assign at least some market shares to each
technology option, with some exceptions (e.g., technologies with integrated carbon capture
and storage are not available in scenarios where there is no price on carbon). As a result, it may
be difficult for GCAM to achieve market share targets in which a technology or set of
technologies achieves 100% (or near 100%) market share. Users can implement high market
share targets by complementing the policy with low or zero shareweights for competing
technologies.

For example, California has recently specified a target that 100% of light duty vehicle sales be
electric by 2035. One way to implement this policy would be via the following combination:

•	Using a market share constraint to increase the EV sales share to 33% in 2025, 66% in 2030,
and 100% in 2035

•	Setting the shareweights for non-EVs to 0 from 2035 through the final model year

Modifying technology and sectoral shareweights is a complicated process in GCAM-USA.
Shareweights can be specified for specific years and regions. However, these values are
overridden if an interpolation rule is in place, which can result in unexpected behavior. To assist
with this challenge, GLIMPSE'S "Tech Param" tab allows the user to specify new sector- or
technology-specific shareweights. When creating the resulting policy file, GLIMPSE attempts to
automatically delete relevant interpolation rules. Users are encouraged to view the affected
technology's shareweights via the Model I interface to verify that the values used within the
model reflect what was expected.

A more detailed description of where shareweight rules are specified and how they are
interpreted will be included in future versions of this Users' Guide.

Ill


-------
CHAPTER 4. HOW DO I....?

In this Chapter, we describe how to accomplish important activities in the Scenario Builder.

4.1 How do I change options in the GLIMPSE options files?

The "options_GCAM-USA-5p4.txt" file resides in the GLIMPSE-5p4 folder. This file lists many
GLIMPSE options that are loaded when the program starts. Important settings include:

•	the location of the GLIMPSE folder (glimpseDir),

•	the location of your GCAM installation (gCamHomeDir),

•	the name of the output database (gCamOutputDatabase),

•	the number of model periods to execute (stop-period),

•	the region to use as the debug region (debugRegion),

•	whether to create the debug file (debugCreate),

•	which programs to use as text and xml file editors, and

•	which query file to use with the ORDModellnterface.

There are a few things to note in the options file:

•	Any line that starts with "#" is interpreted as a comment.

•	Where "#glimpseDir#" and "#gCamGuiDir#" appear to the right of an equal sign, the
values of these parameters that are defined above are inserted. These are the only two
parameters that are replaced in subsequent lines in this manner. For example, if you
include "#scenarioDir#" to the right side of equals, it will not be replaced with the value
of "scenarioDir".

Any changes that are made to the options file do not take effect until: (1) you restart GLIMPSE,
or (2) you choose "File->Reload Options".

You also can access the options file from the Scenario Builder via "File->Edit Options", which will
display the options file in a text editor. Any changes you make and save will not take effect until
you choose "File->Reload Options" from the main menu bar or restart GLIMPSE.

If you want to see the contents of the options file with the #glimpseDir# and #gCamGuiDir#
parameters resolved, choose "File->Show Options", which will pop up a non-editable dialog
window.

112


-------
Please note that changes you make to options such as the "stop-period" or the
"gCamOutputDatabase" are not automatically reflected in scenarios that have already been
created. These settings are reflected in scenarios that are created via the "Create Scenario"

button, © , after the options file has been edited and re-loaded.

4.2	How do I update the JAVA_HOME environmental variable?

Periodically, your system administrators may update the version of Java on your computer.
Depending on the setup, this may result in GCAM-USA no longer running. Typically, the
command window will appear, then disappear, within a second. In addition, the Model I interface
may not run when you press the Results button. If you are experiencing these symptoms, the
next step is to check your JAVA_HOME setting, which lets GLIMPSE know where to find Java.
You can do this using the following step:

•	Open your run_GCAM-USA-5p4.bat file by right-clicking on the file and choose to edit
the file.

•	Identify the JAVA_HOME setting.

•	Identify the location of the JRE folder on your computer. It will be similar to the
JAVA_HOME setting, but may have a different number appended.

•	Update the JAVA_HOME setting in the run_GCAM-USA-5p4.bat file to reflect the folder
number.

•	Save the run_GCAM-USA-5p4.bat file.

•	Re-start GLIMPSE by double-clicking on the run_GCAM-USA-5p4.bat.

If this does not solve your problem, please see the Troubleshooting section of this Users' Guide.

4.3	How do I create a new database?

GLIMPSE is shipped with a BaseX database named "database".

There are several ways to create a new database, which you may want to do if "database" is
getting close to 40 GB in size.

One way to create a new database is to modify the database setting in the options file to
provide a new database name (e.g., database_v2). Re-load the options file or restart GLIMPSE.
Then press the "Results" button.

113


-------
If the database does not exist, the following message will appear:

Open DB Error

X

i

Could not open the database. Attempt to create a new one?
— WARNING doing so will delete all files in the directory,

Yes

No

Figure 4.1 Warning message associated with creating a new database.

In general, you can answer "Yes". The new database will be created and placed in "GLIMPSE-
5p4\GCAM-Model\gcam-v5.4\output". You will still need to update the database specified in
the options file, then reload the options file, if you want this new database to be used in
subsequent runs.

Alternatively, there is a "Create new GCAM output database" option in the Scenario Builder's
tools menu. Select this option and the following dialog window will appear:

lij Create new GCAM output database	— ~ X

If specified database does not already exist it will be created in the output folder.
Name:

Create	Cancel

Figure 4.2 Dialog window for opening or creating a new database.

Enter the name of a database. If a database with that name exists, it will be opened in
Model I interface. The dialog window shown in Figure 4.1 will appear. Click "Yes".

Please note that each scenario's configuration file includes a specification of the database to
use for its results. The database name is inserted at the time the scenario and its configuration

file are created (e.g., when you press © , then click "OK" in the "Creating Scenario" dialog).
Providing a new database name in the options file does not change the database specified in

114


-------
any existing configuration files. You will need to either edit existing configuration files or re-
create the scenarios if you would like them to use the new database setting.

4.4 How do I import and export scenarios from the Mode/Interface?

Users may have reasons to export scenario results from the Model I interface, such as:

•	There are limits on the size of databases that the Modellnterface can open. In general,
the limit is approximately 20 scenarios (or approximately 40 GB in total size). After the
database reaches the limit, it can no longer be opened, and the data are inaccessible.
Users thus may wish to maintain a smaller database that includes only results of
particular interest.

•	Users may wish to archive model results for specific scenarios that have been run.

•	Users may wish to share scenario results with other GLIMPSE or GCAM users.

The Modellnterface includes features for importing and exporting scenarios that can address
these needs. To export one or more scenarios from the Modellnterface, choose "File->Manage
DB". The "Manage Database" window appears:

115


-------
"'I—
~~~

Manage Database	X

Scenarios in Database:

GCAM 5p4-Ref-Orig 20 22-3 -ST 19:01:44-04:00

Add

Done

Figure 4.3 The Manage Database dialog.

If you select a scenario from the list, you have the option to "Remove" it from the list,
"Rename" it, or to "Export" the scenario.

It is particularly important to note that "Remove" currently removes the scenario name from
the scenario list; however, it does *not* decrease the size of the database. Thus, this option
does not solve the problem of a database getting too large.

Exporting a scenario writes the scenario's contents to an XML file. This XML file is generally too
large to open in an editor. However, you can import the scenario into another database, thus
facilitating sharing or keeping a smaller database of import model results.

To import a scenario's XML file into a database via the Mode/Interface, choose "File->Manage
DB", then click on "Add". Note: Do not try to open a scenario's XML file via "File->Open->XML".
This will not work, and your computer's hourglass will spin indefinitely.

116


-------
4.5	How do I manage database size?

As mentioned previously, when the GCAM database reaches a size of approximately 40GB, the
Model I interface is no longer able to open the database and access its contents. As a result, your
data are effectively no longer available.

To avoid this option, we suggest that you switch to a new database when yours includes the
results of approximately 20 scenarios. To create a new database, follow the steps outlined
below:

•	Use "Tools->Create or open output database". A dialog will appear in which you can
specify a name for the new database.

•	Select "Create" to create the new database. Scenarios will not automatically output to
this database.

•	Update the database setting in your options file to refer to the new database.

•	Restart GLIMPSE or choose "Reload Options" for the change to take effect.

Note that the new database will only be reflected in scenarios created after this action. Pre-
existing scenarios in the Scenario Library are not modified.

It also may be advantageous to export scenarios that you want to save for future use. These can
be archived and imported into a database when needed.

4.6	How do I archive scenarios in GLIMPSE?

When conducting studies with a model like GLIMPSE, repeatability is important. An archive
feature has been added to GLIMPSE to support repeatability. If you click on a scenario name in
the Scenario Library list, then choose "Tools->Archive Scenario" from the main menu bar, the
following steps are carried out:

•	A dialog appears to confirm whether you would like to create an archive.

•	If "Yes", the scenario's configuration file is parsed, and all of the scenario components
listed in the file are identified (this includes from lines that are "commented out" via
"
-------
•	A copy of the scenario's configuration file is created but modified to point to the scenario
components within the "archive" folder instead of their original location.

•	The archive folder is zipped (although the unzipped folder is also currently maintained).

•	This archive folder can also be found in the scenario's folder.

When the user selects a scenario name in the Scenario Library table and presses play, GLIMPSE
first checks to see if an archive version of the configuration exists. If so, it asks whether the user
would like to run the scenario from the archive.

4.7 How do I access and interpret the main Jog file?

Most of the diagnostic information that is printed to the command terminal during GCAM
execution is also saved to the main_log.txt file that is in the "GCAM-Model\gcam-
v5.4\exe\logs" folder. This information can be extremely useful in diagnosing issues with the
run, including identifying when during execution that problems occurred and in identifying
which time periods had unsolved markets. For a run that is ongoing, you can access the
"main_log.txt" file easily in one of two ways. First, you can select the menu item "View-
current Main Log". The file will be displayed using the text editor that is specified in the

By default, GLIMPSE saves the "main_log.txt" file that is associated with each run after that run
terminates. You can find a scenario's "main_log.txt" file in the folder associated with the
scenario via GLIMPSE'S Scenario Library. To access the folder, click on a scenario, then click the

Library, then click the "Log" button,

4.8 How do I edit a scenario's configuration file?

You can edit a scenario's configuration file by double-clicking on the scenario name in the
Scenario Library pane. This will display the configuration file in a text editor. Settings such as the
output database or stop period can be edited and will be reflected when the scenario is
executed.

options file. Alternatively, press the "EXE Log" button,

"Browse" button, . Alternative, you can click on the scenario's name in the Scenario

118


-------
4.9 How do I import Scenarios into GLIMPSE?

Configuration files developed outside of GLIMPSE can be added as scenarios within the Scenario
Library. To use this function, use "File->lmport Scenario". Select the configuration file using the
File Browser that appears. GLIMPSE will then create a new Scenario entry in your Scenario
Library and will make a copy of the configuration file in that location. The scenario can be
executed and archived as if it had been created in GLIMPSE. However, you will not be able to
edit its components.

4.10	How do I import files into the Component Library?

You can include XML add-on files in GLIMPSE in one of several ways.

•	Using the New Scenario Component Creator, you can use the XML List feature to "point"
to one or more external XMLs. The resulting file list will be saved to the Component
Library. Adding this scenario component to a scenario will result in the files in the file list
being inserted into the scenario's configuration file when it is created.

•	You can place XML files directly into the Component Library folder. The XMLs will appear
in your Component Library and can be added to scenarios as you would add any other
scenario component.

4.11	How do I recover deleted Scenario Components and Scenarios?

You can view deleted scenario components and Scenarios by selecting "View->Browse Folders-
>GLIMPSE Trash Folder" via the main menu of the Scenario Builder. Deleted scenario
component files are located within this trash folder, while deleted scenarios are subfolders
within the trash folder.

If you have not emptied the GLIMPSE trash folder, you can restore deleted scenario
components by placing them back in the Component Library folder. Access that folder by

clicking on the Open Folder button, , in the Component Library section of the Scenario
Builder.

Similarly, you can restore deleted scenarios by moving the sub-folder with the scenarios name
from the trash folder to the "GLIMPSE Scenarios Folder". You can access the "GLIMPSE
Scenarios Folder" via "View->Browse Folders->GLIMPSE Scenario Folder" in the main menu of
the Scenario Builder.

119


-------
4.12	How do I save files associated with each run?

In the GCAM-USA output file, options_GCAM-USA-5p4.txt, is an option "gCamOutputToSave",
which is followed by a list of files to save for each run. Upon the completion of each run, these
files are saved to the scenario folder.

Modify the gCamOutputToSave parameter to change the files to be saved. After the values
have been modified, select "File->Update Options" to update the settings in GLIMPSE.

4.13	How do I clean up saved files?

By default, after every run, GLIMPSE will save a number of log files, including main_log.txt,
solverjog.csv, calibration_log.txt, as well as debug.xml. These logs can be large. For example,
the solverjog.csv file may be several hundred MBs. The main_log.txt file should be saved.
However, the other files are most useful for debugging, and it is not necessary to save them for
future use. Selecting "Tools->Advanced->Cleanup Saved Files" will delete the unneeded log
files.

4.14 How do I use the CSVtoXML utility?

There are several approaches one can use to develop input files for GCAM. One approach
involves integrating the new data and processing into the GCAM data system. This approach
requires significant understanding of the data system structure and operation. Please see
PNNL's github site for information about the datasystem, including an online manual
(https://eithub.com/JGCRI/gcamdata). See the next section for links to additional information
on the data system.

Most GLIMPSE users may not need to use the data system. Instead, these users can create
"add-on" files that are formatted in XML. These add-on files could be created by hand, but the
process can be tedious for complicated files. Alternatively, the Model I interface (and
Modellnterface) provide a "CSVtoXML" utility that can convert CSV-formatted files into XML
files using rules provided in a header file. For more information on using the "CSVtoXML" utility,
please see the instructions that are available via the GLIMPSE help menu option accessible from
the Scenario Builder. Choose "Help->GLIMPSE Documents", then navigate to the subfolder
"CreatingXMLs" and open the file "Using the Model Interface to create XML files from CSV
files.pdf".

120


-------
4.15 How do I know if my computer's resources are running low?

When GLIMPSE starts up, several computing system properties are evaluated. Results are
displayed to the Command window, including warnings if specific thresholds are not met. This
analysis can be executed at any time by selecting "Tools->Check Installation" from the Scenario
Builder's main menu. Thresholds include:

Table 4.1 Thresholds used for evaluating computer resources. Resources are compared to
these thresholds at GLIMPSE startup and when "Tools->Check Installation" is selected.

Parameter

Threshold

Total physical memory (RAM) for GCAM

12 GB

Total physical memory (RAM) for GCAM-USA

16 GB

Free hard disk space

100 GB

Swap space size

25 GB

In addition, GLIMPSE checks computer resource status approximately every 20 seconds and
updates the information in the status bar at the bottom of the Scenario Builder. Status shown
includes CPU usage, total RAM, free RAM, total swap space, free swap space, the name of the
current database, and its size and fraction of the maximum recommended size that has been
used.

These values are also compared against runtime thresholds:

Table 4.2 Runtime thresholds used for evaluating computer resources. Resources are
compared to these thresholds approximately every 20 seconds.

Parameter

Threshold

Free physical memory (RAM)

5%

Free swap space

5%

Free hard disk space

40 GB

Free database size as a percentage of the max
size listed in the options file

80%

When specific thresholds have been exceeded, the status message is concatenated with "111"
and asterisks are placed next to the values in exceedance. Additionally, this information is

121


-------
saved to a log file that can be accessed by "View->Resource Logs->Current Session". If GLIMPSE
crashes and needs to be restarted, you will find information from the prior session at "View-
resource Logs->Prior Session". For each log entry there is a time stamp and the name of the
currently executing scenario. This information can be useful in debugging execution problems
that are caused by computer resource limitations.

4.16 How do I determine why a GCAM run did not complete and GLIMPSE reports
"DNF"?

There are several reasons why a GCAM run would not complete successfully. These include the
following:

•	JAVA_HOME is set incorrectly. GCAM relies on the Java virtual machine (JVM) for execution.
The location of the JVM is defined by the JAVA_HOME environmental variable in the
run_GLIMPSE_GCAM-USA-5p4.bat. If JAVA_HOME is set incorrectly, a gcam.exe window will
appear then disappear nearly instantly. If you are using a computer that is managed by an
administrator, one potential cause is that a new version of Java may have been installed on
your computer, and the previous setting for JAVA_HOME is no longer correct. Please follow
the instructions in section 4.2, "How do I update the JAVA_HOME environmental variable?"

•	There are errors in the formatting of the configuration file. Configuration files for GCAM are
formatted in XML. If there are errors in the formatting of this file, such as the use of
incorrect syntax, GCAM's XML parser will not be able to read the configuration file. Similar
to when JAVA_HOME is set incorrectly, the gcam.exe window will appear then immediately
disappear. For the scenario with the problem, open the scenario's configuration file and
check the formatting. Some text editors can color-code the content of XML files and can
check the structure and format of the file, helping you identify if formatting is the problem.

•	There are errors in a scenario component file. When GCAM starts, the parser will indicate in
the gcam.exe window when the process of reading in each scenario component begins. If
the parser tries to load an incorrectly formatted scenario component file, or if the file does
not exist, GCAM will terminate and the gcam.exe window will disappear. When this occurs,
you can often examine the scenario's main_log.txt file to find the last file that the parser
attempted to load. This file typically is the one that caused the run failure. Make sure the
referenced scenario component exists in the location at which it is referenced. If it does,
check its format for errors.

•	Computer resources have been exceeded. Running out of available RAM, swap space, or
disk space will result in your GCAM run being terminated abruptly, with no message or sign

122


-------
in the main_log.txt file indicating the reason for termination. If this appears to be the case,
please use the "View->Resource Logs->Current Session" option in the Scenario Builder main
menu to see if there are any resource warnings that were reported around the time the
problem occurred. If the information in the log indicates that disk space or swap space were
limited, consider taking actions to free up hard disk space. If the warning indicates that RAM
was the issue, then you may need to revisit the policies represented in your scenario. For
example, market share constraints such as Renewable Portfolio Standards can be
particularly memory-intensive. You may want to consider alternative approaches for
simulating these types of policies, such as via the "Tech Bounds" or "Tech Avail" scenario
components, or by adjusting shareweights.

•	There are naming conflicts for markets or policies. GCAM users have considerable flexibility
in naming markets and policies. However, if the same name is used for more than one
market or for more than one policy, this can create conflicts that may cause the model to
terminate abruptly during the model run. GLIMPSE attempts to address this by assigning
unique names to markets and policies within scenario components that are created within
the system. However, if you have imported or referenced scenario components created
outside of GLIMPSE, there is the possibility that naming conflicts exist. Please check any
add-on files constructed outside of GLIMPSE to determine if there are conflicts.

•	The scenario's gcam.exe file was closed during execution. Closing the gcam.exe window that
is displaying text from the GCAM run will terminate the run.

If the GCAM run completes but GLIMPSE is reporting problem markets, please see "Section 5.3
Interpreting and debugging unsolved markets".

4.17 How do I access the GCAM data system and find the original sources of data?

The GCAM data system includes the underlying data used to generate the model's XML-
formatted input files. You can access the data system at GCAM-Model/gcam-
v5.4/input/gcamdata. This folder includes an xml folder that contains the model's input files, as
well as many other folders that contain data and code that were used to create the inputs.

Much of the input data can be found in the following subfolder: "GCAM-Model/gcam-
v5.4/input/gcamdata/inst/extdata". Within this folder are subfolders that include data for the
various human-Earth systems represented in GCAM. These include:

•	aglu (agriculture and land use files)

•	emissions

123


-------
•	energy

•	socioeconomics

•	water

In addition, there is a subfolder named "gcam-usa" for GCAM-USA-specific data.

124


-------
CHAPTER 5. ADVANCED TOPICS

In this Chapter, we provide information about advanced concepts and operations, such as
understanding how GLIMPSE interacts with GCAM-USA, debugging model solution errors, and
interpreting biomass flows.

5.1 What is happening behind the scenes in GLIMPSE...?

In this section, we describe actions taken by GLIMPSE when you perform various functions
using GLIMPSE.

5.1.1... when you create a new Scenario Component?

When you press "Save" in the New Scenario Component Creator window, what happens next
depends on what type of scenario component you are creating. If you are creating an XML list
(e.g., a text file that points to one or more XML files), GLIMPSE prompts you for the file name,
then saves the resulting file to the Component Library or another location of your choice. For
GLIMPSE to "see" the new Scenario Component, you will need to place it in the Component
Library or a subfolder of that library.

If you are creating one of the other types of Scenario Components, the first step that occurs is a
quality assurance step. GLIMPSE checks to see if all of the necessary options have been
selected. If not, a warning message is provided. Otherwise, you are prompted for a filename. At
this point, GLIMPSE scans all of the policy names and market names for the CSV-formatted
Scenario Components. It chooses unique names when writing the current file to disk to avoid
naming conflicts with existing policy representations. This action will produce a CSV file,
formatted to match one or more headers in the GLIMPSE header file: "glimpseXMLheaders.txt".
Note that later, when the scenario is created, its CSV files will be converted to XML files
automatically using these headers.

5.1.2... when you construct a new Scenario?

When you press the Create Scenario button, , GLIMPSE first checks to see if this scenario
already exists. It does so by checking if there is a subfolder in the Scenario Library folder with
the same name. If there is, the user is prompted to determine if they wish to continue. Next,
GLIMPSE checks the size of the database in which the scenario's outputs are to be saved. If the
database exceeds a specific size (default is 40 GB, although alternatives can be specified in the
options file), a warning is presented, and, again, the user has the option of continuing or
canceling. Next, a dialog appears. It lists the scenario name, the database to which the results
will be stored, and the size of that database. In addition, the user is given the options of
whether to generate a debug file (recommended) and which region to use for the debug region
The option is available to change the final year of the run, and a comment area is provided for
the user to leave any text that describes the scenario.

125


-------
When you click "OK", the scenario's folder is created. Next, the Scenario Template file is
accessed (typically, configuration_usa_5p4_template.xml). Metadata is written to the top of
the file, then the database, stop-year, and debug information are updated to reflect what the
user has selected.

Next, GCAM goes through the Scenario Components that have been selected for inclusion.
When it reads an XML list file, it places references to each of the indicated XML files at the
bottom of the Scenario Components block within the configuration files. When it reads an XML
file, it places a reference to that file into the same location. When it reads a CSV file, it identifies
the matching headers in the glimpseXMLheaders.txt file, then uses the CSV2XML.jar utility to
convert the CSV file into an XML. The resulting XML is automatically placed in the new scenario
folder, and a reference to it is placed in the configuration file. Finally, the edited configuration
file is renamed to reflect the scenario's name and placed in the scenario's folder.

5.1.3... when you run a scenario?

Selecting one or more scenarios in the Scenario Library, then clicking "Play", , adds those
scenarios to the run queue. GCAM runs are made sequentially from the queue, on a first-in
first-out basis. When running each scenario, GLIMPSE calls the GCAM executable identified in
the options file, using the "-C" argument to use that scenarios configuration file. When a run
has finished, whether successful or not, the "main_log.txt" file, several other log files, and the
debug file are copied to the scenario's folder.

Every 20 seconds, GLIMPSE updates the information about each scenario in the Scenario
Library. It does this by traversing the GCAM-Data/GCAM-USA/ScenarioFolders folder. For each
scenario's folder within, it searches for the main_log.txt file. If that file exists, then it is searched
for the text "Model completed successfully", "Model periods not solved", and for the runtime.
If no main_log.txt file exists, GLIMPSE checks to see which run is currently executing by
checking the current main_log.txt file in the exe folder. Based on the results, GLIMPSE assigns
status as "Running", "Success", "PblmMrkts", "DNF", or blank. PrbmMrkts indicates that there
were market errors in one or more time periods. DNF stands for Did Not Finish.

5.2 GLIMPSE folder structure

In this section, we describe some of the most important folders in GLIMPSE.

•	GLIMPSE - The root folder for GLIMPSE. This may be different, depending on your
installation.

The following folders contain the GCAM model and data.

•	GUMPSE/GCAM-Model/gcam-v5.4 - The root folder for a particular version of GCAM 5.4.

•	GUMPSE/GCAM-Model/gcam-v5.4/exe - Includes the GCAM executable (gcam.exe) and
example configuration files (e.g., configuration_ref.xml and configuration_usa.xml).

126


-------
•	GLIMPSE/GCAM-Model/gcam-v5.4/exe/logs - Includes log files that are created during
GCAM calibration and execution, such as main_log.txt, calibration_log.txt, and
solverjog.csv.

•	GUMPSE/GCAM-Model/gcam-v5.4/input - Includes folders with input data used by
GCAM or the GCAM data system.

•	GUMPSE/GCAM-Model/gcam-v5.4/input/policy - Includes example policy files for GCAM
(but not for GCAM-USA), including files that impose carbon taxes and that implement
various global warming targets.

•	GUMPSE/GCAM-Model/gcam-v5.4/input/gcamdata - Includes the gcam data system, an
R-based system that creates XML-formatted input files for GCAM from CSV-formatted
source data.

•	GUMPSE/GCAM-Model/gcam-v5.4/input/gcamdata/xml - Includes the xml-formatted
input files that are produced by the GCAM data system and read in by GCAM as scenario
components.

•	GLIMPSE/GCAM-Model/gcam-v5.4/input/gcamdata/inst/extdata - Includes the CSV-
formatted source data. The CSV files have metadata identifying the original source of data
tables.

•	GUMPSE/GCAM-Model/gcam-v5.4/output - Includes GCAM's output databases. Each
database is a subfolder that includes seven files with the ".basex" extension. These are
binary files in which GCAM's XML-formatted outputs are stored.

GLIMPSE folders:

•	GLIMPSE/Docs - Includes GLIMPSE documentation, such as this Users' Guide.

•	GUMPSE/GLIMPSE-Data/GCAM-USA/ScenarioComponents - Includes Scenario
Components available to GLIMPSE. These can be XML add-on files, text files that point to
one or more external XML files, or CSV files that are formatted specifically for use by
GLIMPSE.

•	GLIMPSE/GLIMPSE-Data/GCAM-USA/ScenarioFolders - Includes folders for each scenario
in the GLIMPSE Scenario Library. These folders are where the scenario's log files are
stored after each run, as well as where any scenario-specific XML files are stored. If the
scenario is archived, the copies of all scenario components will be placed in a subfolder
named "archive".

•	GLIMPSE/GLIMPSE-Data/GCAM-USA/trash - Includes deleted Scenario Component files
and Scenario folders. To restore from trash, move deleted components and scenarios
back into "ScenarioComponent" and "ScenarioFolders", respectively.

•	GUMPSE/GUMPSE-GUl - Includes all of the files associated with the GLIMPSE graphical
user interface (GUI)

•	GLIMPSE/GLIMPSE-GUI/exe - Includes ScenarioBuilder.jar, the GLIMPSE GUI executable
file.

127


-------
G UMPSE/G LIMPS E-G Ul/templates - Includes files that are used as templates by GLIMPSE,
such as the template configuration file used when constructing new scenarios.

GUMPSE/GLIMPSE-GUI/resources - Includes ancillary files used by GLIMPSE, such as
button images and the technology listing that is used the New Scenario Component
Creator window.

GLIMPSE/GLIMPSE-Modellnterface - Includes a modified version of PNNL's
Mode/Interface, which adds features for filtering, graphing, and exporting data.

GUMPSE/GLIMPSE-Modellnterface/exe - Includes ORDModellnterface.jar, the executable
used to run the program, as well as the modellnterface.properties file that has start-up
settings and the Main_queries_GLIMPSE.xml file that includes the queries used by
GLIMPSE.

5.3 Interpreting and debugging unsolved market information in the main_log.txt file.

The GCAM solver works by attempting to identify the price in each market at which supply
equals demand. When GCAM experiences an unsolved market, information about that market
is reported to the "main_log.txt" file. Here is an example of such a report:

./restart/restart.6..



Writing restart file:

Period 7: 2030

Number of accive scare values: 2846621

Starting Solution. Solving for 1667 markets.

ERROR:Model did not solve period 7 within set iteration S124

ERROR:Currently Unsolved Markets;

ERROR:Unsolved Part 1: Solvable Markets

ERROR:X,
ERROR:1.04764
ERROR:0.0126816
ERROR:0.0015292
ERR0RS7.12739
ERR0R:2.07346
ERROR:0.0543962
ERROR:0.142831
ERROR:10.3277
ERROR:0.43S26
ERROR:0.036S611
ERROR;0.0441924

XL,	XR,	ED.	EDL,	EDR,	RED,	brk. Supply,

,	7.S5702 ,	3.21381 , -0.00245122, -0.00133156, -0.00153139, 0.000902816,

,	0.0123835 ,	0.0123838 , -8.38918e-05, 0.000263064, -0.000728867, 0.000187506,

,	0.00148264,	0.0014826S, -6,5974Se-05, 0.000561821, -0.000387099, 0.000249359,

,	6.48911 ,	6.48912 , -0.697474 , 0.00275568, -0-0152208, 0.108473 , 1

,	2.03844 ,	2.03844 , -0.00146691, 0.000212768, -0.000286213, 0.000707972,

,	0.0544458 ,	0.0544458 , 0.0325314 , 0.00482003, 0.00482003, 0.000439323, 0

,	0.140889 ,	0.140889 , -6.12616e-05, 1.25307e-05, 1.25307e-05, 0.000429094,

,	10.258 ,	10.258 , 0.0123906 , 0.00084659, 0.00084659, 0.00119831, 0

,	0.405354 ,	0.405354 , -0.00122034, -1.12272e-05, -1.12272e-05, 0.00279229,

,	0.0365907 ,	0.0365907 , 0.0312499 , -0.000672724, -0.000672724, 0.000873203,

.	0.0442336 ,	0.0442336 , 0.0331784 , -0.000886108, -0.000886108, 0.000868204,

Demand, Mxk Type, Market,

0	, 2.71753 , 2.71508 , RES	, Biofuel_All_Reg_MktBiofuel_All_Reg ,

1	, 0.447492 , 0.447408 , Normal , TXonshore wind resource	,
1 , 0.264643 , 0.264577 , Normal , TXPV_resource

, 7.12739 , 6.42992 , Price , DSArefining	,

1 , 2.07346 , 2.07199 , Demand , Texas gridelectricity domestic supplyDemand_lnt,
, 74.0164 , 74.0439 , Normal , OSAWheat

0 , 0.142831 , 0.142769 , Trial-Value, Texas gridsolar-trial-supply	,

, 10.3277 , 10.3401 , Price , Hawaii gridelectricity domestic supply,

0 , 0.43826 , 0.437039 , Trial-Value, Hawaii gridsolar-trial-supply ,
0 , 35.7564 , 35.7876 , Normal , Europe_EasternCorn ,
0 ,	, Hzrrnal , Arqentir.aCorr.	,	

Figure 5.1 Example of the table used to report market solution errors.

Parameters of relevance are defined in the table below.

Table 5.1 Parameters relevant to solver operations

Parameter

Description

X

The current guess for the price that will solve the market.

XL

XL and XR represent the previous two guesses, with XL being the
lower of the two values.

128


-------
XR

XR is the upper value of the prior two guesses.

ED

At price X, ED represents the excess demand, or Demand minus
Supply.

EDL

At price XL, EDL represents the excess demand, or Demand minus
Supply.

EDR

At price XLR, EDR represents the excess demand, or Demand minus
Supply.

RED

Relative excess demand (RED) is the absolute value of ED divided by
demand.

Supply

The amount of Supply at price X.

Demand

The amount of Demand at price X.

max-model-calcs

The maximum number of solver calculations that will be conducted in
a model period, specified in the solver configuration file.

solution-tolerance

A threshold against which the RED is compared to determine whether
the market has been solved, specified in the solver configuration file.

solution-floor

A threshold against which the absolute value of ED is compared to
determine whether the market has been solved, specified in the
solver configuration file.

The process GCAM's solver uses to determine the "market clearing" price is similar to the
familiar "pick a number between 1 and 100" game. An initial set of guesses for X are made, with
XL being the lower of the two values and XR being the higher. Excess demands are calculated
for both XL and XR, providing EDL and EDR. These values are used to determine the next guess
for X. This value of X then displaces either XL or XR, new values for EDL and EDR are calculated,
etc.

The search for the best value for X ends when one of three stopping conditions is met:

•	The calculation limit, max-model-calcs, has been exceeded

•	The value of RED at X falls below the specified solution-tolerance

•	The value of | ED | at X falls below the specified solution-floor

The second and third criteria are similar, but the normalization that occurs when evaluating
RED is useful evaluating convergence in small markets. The solution-floor is typically smaller

129


-------
than the solution-tolerance and can also be interpreted as the difference between supply and
demand being so small that it can be interpreted as being zero.

If the calculation limit is reached, but not all markets have been solved, the markets that have
not been solved are reported to the "main_log.txt" file.

For the example report shown in Figure 1, there are 11 market failures reported in time period
7 of the model, 2030. The columns are not aligned well, but the fields in each row are separated
by commas, and the proper column for each value can be readily deduced. See Table 1 in this
section for a brief description of the parameters in the table.

Here, the supply and demand for each market are quite similar and are often different at the 4th
or 5th decimal point. This is evident in the RED column, where many of the values are small. The
solution-tolerance for this run was 0.0001, so many of these markets were close to having been
solved. One approach to address very small market errors is to relax the tolerance in the
solver's configuration. Alternatively, if these markets would have met the relaxed tolerance, it
may be possible to assume they have been solved. If this latter approach is taken, it is
suggested that the modeler carefully inspect model results for that time period for
discontinuities or other unexpected behavior.

In contrast, in the error table, the RED value for the "USArefining" market indicates that supply
and demand were different by 11%. That is a relatively large difference in an important market.
Additional examination is necessary, including evaluating whether any constraints in the model
make the solution infeasible.

There are other causes of unsolved markets. In some instances, a particular policy or
technology may have introduced complexities that are resulting in the need for more
computations to determine the market-clearing prices. Increasing the max-model-calcs in the
solver configuration file allows the search to perform additional iterations. Modifying solution-
tolerance and solution-floor may also provide some benefit, although changing these values
impacts the solution process and may not lead to the desired results.

Some markets may also be unsolvable or difficult to solve because of numerical issues involving
very low demands or extremely high prices, producing conditions that are incompatible with
the scaling approach used by the solver. State-level constraints on transportation technologies
are particularly problematic because of the unit conversions used specifically for that sector
within the model. In the solver configuration included in GLIMPSE, we have lowered the
solution-tolerance and solution-floor from their default values, which has addressed many of
these instances.

130


-------
For some of these markets that suffer from numerical issues, the multi-threading capability of
GCAM may exacerbate these errors by introducing additional small numerical errors. For
scenarios experiencing errors in small markets, turning off multi-threading may solve the
problem at the expense of a longer runtime. This option is available in the dialog window that

appears when you click on the "Create Scenario" button, .

At this time, our goal is to provide some insights that may help a GLIMPSE user understand and
recover from some market solution problems. As we seek to improve the solver settings and
configuration to improve its robustness, this section wiil evolve.

5.4 Biomass flows and CO2 accounting in GCAM-USA

As alluded to in the Tutorial, biomass has the potential to play a role in GHG mitigation
strategies, including taxes and caps, renewable portfolio standards, and clean energy standards.
In this subsection, we investigate CO2 accounting associated with biomass as weil as its use in
the energy system. The information provided here was developed by analyzing the results of
the "inputs by tech" and "output by tech" queries, which are highly recommended for assessing
material flows in GCAM.

There are several pathways by which biomass can enter a solution, as shown in the following
figure.

131


-------
Figure 5.2 Biomass and biofuel activities in GCAM-USA.

Biomass (e.g., fiber, grass, tree, root/tuber, sugar, and corn) for use in bioenergy and biofuels is
"grown" in GCAM and GCAM-USA at the land-use resolution (e.g., water basins). Corn and oil
crop can be used to create corn ethanol and biodiesel. Alternatively, these crops and other
forms of biomass can enter states as "regional biomass", which, in turn, can feed applications in
the electric sector and advanced biofuel production. In addition, regional biomass can be
converted into "delivered biomass", which is used in the residential, commercial, and industrial
sectors.

From a CO2 accounting standpoint in GCAM, combustion of biomass results in the release of
CO2 emissions (just as it does in the real world). To counter these releases, GCAM represents
the CO2 uptake from the atmosphere that is associated with biomass grown for bioenergy or
biofuels. This uptake is accounted for at the state level (see "Up" on the biomass flow diagram),
allowing bioenergy use and biofuel production to play a role in meeting an economy-wide CO2
cap or for reducing costs under an economy-wide CO2 tax.

There are several limitations to this method of accounting. One limitation is that a state would
not receive "credit" for reducing transportation emissions by increasing the biofuel content at
the pump. Instead, the credit for that ethanol would occur in the state where the ethanol was
produced. Furthermore, a cap on CO2 emissions from the electric sector would not increase
biomass use in that sector since the credit would be occurring in the state's "regional biomass"
sector instead.

5.5 Considerations when modeling a deep decarbonization scenario

A deep decarbonization scenario is one that achieves significant reductions in CO2 emissions,
often targeting an 80% or even 100% reduction by 2050, relative to a historical year such as
1990 or 2005. "Net-zero" strategies typically achieve zero CO2 emissions through a combination
of reducing emissions from the energy system, applying carbon capture to fossil- and biomass-
fueled combustion activities, removing CO2 from the atmosphere via direct air capture, and
through changes in land use and farming practices.

GLIMPSE can be used to explore pathways for achieving deep decarbonization scenarios. One
approach is to introduce a declining CO2 or GHG cap, allowing GCAM to select a technology
pathway through time for achieving the target. Alternatively, GLIMPSE can be used to simulate
sectoral mitigation strategies, including introducing market share targets for vehicle
electrification and renewable portfolio standards for electricity production.

132


-------
For either approach, GLIMPSE users should be aware of how GCAM's myopic solution process,
logit-based market share allocation, and shareweight assumptions affect the results (described
in "Chapter 2. How does GCAM work?"). Depending on the scenarios being modeled, some
adjustments may be desired or necessary, as discussed in this section.

5.5.1 Potential adjustments for deep decarbonization scenarios

When simulating deep decarbonization strategies, users may want to consider making the
following adjustments:

•	Including direct air capture (DAC) technologies. A representation of DAC was not
included in the original GCAM-USA 5.4 release but was later developed by PNNL. We
have included DAC.txt Scenario Component that can be used to integrate DAC into
decarbonization scenarios.

•	Modifying the shareweight trajectories of decarbonization technologies. In the
Reference Case assumptions, many of the electric and hybrid transportation
technologies have shareweights that begin at 0 and transition to 1 by 2050, 2070, or
even 2100. A value of 1 indicates that all biases have been addressed such that the
technology competes for market share based upon its cost (and the impact of any
policies on technology costs). Under a deep decarbonization scenario, it is reasonable to
assume that many of the barriers to hybrid, electric, and fuel cell technologies will be
addressed sooner than in the Reference Case. For example, investments likely will be
made to build out charging and hydrogen fuel infrastructure. Thus, it may be reasonable
to modify the shareweight trajectories for these technologies to reach 1 sooner. We
have created a DeepDecarbAssumptions.txt scenario component that includes revised
shareweight trajectories for advanced transportation technologies; however, GLIMPSE
users may opt to introduce their own assumptions instead.

•	Reducing the elasticities on end-use service demands. In GCAM, end-use service
demands have price elasticities such that increases in cost lead to a reduction in
demand. These elasticities have been developed with input based on past human
behavior. However, a deep decarbonization scenario would diverge considerably from
past conditions, and thus consumer choices and behaviors could be different as well.
Using the default elasticities, common response in GCAM and GCAM-USA to deep
decarbonization is significant reductions in service demands for historically difficuIt-to-
decarbonize sectors, such as international aviation and cement production. When
modeling decarbonization scenarios, we recommend that users examine the response

133


-------
of end-use demands and consider modifying the elasticities such that this is less demand
response.

•	Modifying the electric sector's coal plant retirement trajectory. GCAM and GCAM-USA
include a "profit-shutdown-decider" function that drives retirement as operational costs
increase. We have found that the parameterization of this function results in the model
being unable to retire existing capacity quickly enough to meet some short-term
decarbonization targets. As a result, users may choose to drive retirements via adjusting
power plant lifetimes or by using the "Tech Bound" option in GLIMPSE.

•	Eliminating conventional technologies when using high market share constraints. Some
deep decarbonization scenarios may involve representations of sector specific policies,
such as electric vehicle sales reaching 100% or an RPS reaching 100% in a particular
modeled year. When modeling such scenarios, users are advised to keep in mind how
the logit function makes market share decisions in GCAM. The logit will assign market
share based on relative costs, adjusted by technology-specific shareweights. In most
instances, GCAM will assign a non-zero market share to all technologies that are
competing in the market. As a result, market share constraints above 90% can be very
challenging, often resulting in unsolved markets. This problem can be addressed by
adjusting the shareweights of conventional technologies. For example, to meet a 100%
EV target, in the years that target must be met, the shareweights for other technologies
in that market can be set to zero, effectively removing them from the market. In
GLIMPSE, the "Tech Avail" feature can be used for this purpose, and shareweight can
also be modified via the "Tech Param" option. See Tutorial 5 for an example.

5.5.2 Additional considerations

There are several additional aspects of deep decarbonization that should be considered. When
interpreting the resulting mitigation strategy, it is important that users keep in mind the myopic
nature of GCAM's solution process. GCAM solves each time step by optimizing technology and
fuel choices based upon conditions within that time period only - but with no knowledge of
conditions in the future. Thus, GCAM will not make short-term decisions with long-term policy
targets in mind, and this will impact the decarbonization pathway that is produced.

Uncertainty is another important consideration. Assumptions about the cost and performance
of decarbonization technologies typically are derived from a combination of peer-reviewed
literature and government reports. Nonetheless, future technology cost and performance are
very uncertain, particularly for technologies that have not yet been deployed commercially.
Sensitivity analysis to explore alternative pathways can provide valuable insights into the roles

134


-------
that technologies may play and into the amount of flexibility available in meeting
decarbonization targets.

135


-------
CHAPTER 6. REFERENCE

6.1 Buttons and menu options on the Scenario Builder

In this section, the function of each button and menu option of the Scenario Builder is
described.

6.1.1 Buttons

Q : r.IPSE Scenario Builder
File Tools View Help

~

Component Library Search:

Component Name

Calib-2025lndCoal-GHGPIanStates.csv

Calib-biomass_constraints.txt

Calib-coal_egu_2020.txt

Calib-HDV-BEV-SW-1x50.csv

Calib-LDV-EV-AE02020.csv

Calib-Lower_lncome_Elasticity_Tran.txt

CaliK_l r\ alt Kja+iioI C\A/ rei/

+ B O

Created

2023-02-08: 08:45
2023-02-08: 08:45
2023-02-08: 08:45
2023-02-08: 08:45
2023-02-08: 08:45
2023-02-08: 08:45
-)m-x.ro.r\«•

Create Scenario

Component Name

No content in table

©

Scenario Library

Scenario Name

GCAM5p4-Ref-Orig
GLIMPSEv! -Reference



MlE
~~~

Created

2023-04-24:12:20
2023-04-27: 11:52

Completed

2023-05-01:08:48

Status

Success

Hi R Hi Pi Ei |Q

ProbMkts

Runtime

0 hr 51 min

Resources... CPU: 29% | RAM: 31.9GB Free:70% | Disk: 2,761.7GB available | Swap: 63.9GB Free: 73% // Database: database Size: 2.1GB Used: 5.4%

Figure 6.1 Scenario Builder window showing buttons.

See the table for descriptions of the buttons associated with various parts of the Scenario
Builder.

Table 6.1 Buttons in the Component Library.

Icon

Description





"New" - Open dialog to create a new scenario component.

136


-------


"Edit" - Open dialog with the attributes of the selected scenario component loaded
for editing.

&

"Browse" - Opens the folder where the scenario components are stored.

X

"Delete" - Removes the selected scenario component(s). Deleted components are
sent to the "Trash" folder. GLIMPSE currently does not support restoring from
trash.

O

"Refresh" - Re-scans the scenario component folder in GLIMPSE and refreshes the
list of available components.

Table 6.2 Buttons between the Component Library and "Create Scenario" area.

Icon

Description

~

"Include" - Copies the selected scenario component(s) from the Component Library
into the scenario being created in the "Create Scenario" area.



"Remove" - Removes the selected scenario component(s) in the "Create Scenario"
area, returning them to the Component Library.



"Remove all" - Removes all scenario components in the "Create Scenario" area,
returning them to the Component Library.

Table 6.3 Buttons in the "Create Scenarios" area.

Icon

Description

©

"Create" - Create a new scenario by converting scenario components to XML files,
as needed, modifying the template configuration file, creating a folder for the
scenario, and adding the scenario to the Scenario Library table.

w

"Move down" - Move the order of the selected scenario component down in the
"Create Scenario" list.



"Move up" - Move the order of the selected scenario component down in the
"Create Scenario" list.

137


-------
Table 6.4 Buttons in the Scenario Library.

Icon

Description



"Modify" - copies the selected scenario from the Scenario Library area to the
"Create Scenario" area where it can be modified.



"Edit" - opens the configuration files for the scenarios that are selected in the
Scenario Library area. The editor specified in the options file is used.

[



"Browse" - opens the scenario folder for the selected scenario. The scenario folder
includes the configuration file and XML files that were created with the "Create"
button in the "Create Scenario" area is pressed. After a run has been completed,
the main_log.txt file, as well as other files (specified in the options file) associated
with the run, are moved to the scenario folder.



>

"Play" - adds the selected scenarios to the queue of scenarios to be run. If the
queue is empty, the scenario begins to run. If multiple scenarios are selected, they
are added to the que in order of where they appear in the table, from top-to-
bottom.

~ ~~

"Results" - starts the Mode/Interface and automatically opens the database that is
specified in the options file.

J

3

"Results-Selected" - starts the Model Interface and automatically opens the
database associated with the selected scenario.



"Diff" - compares the configuration files of two selected scenarios. Lines that differ
between the two files are identified and those differences are listed.





"Queue" - shows a list of the runs that have been added to the queue since this
session of GLIMPSE was started. Note that completed runs are not removed from
this list. That feature may be added later.



3|

log|

"Log" - displays the "main_log.txt" file that is in GCAM's "exe/log" folder. This
button allows you to monitor the status of the current GCAM run while it is
executing, including viewing any warnings or error messages.



log|

"Log-Selected" - displays the "main_log.txt" file that is associated with the selected
scenario(s).



1 rn m 1

[iSpJ

"Errors" - examines the "main_log.txt" file in the "exe/log" folder, extracting lines
that include the word "Error". The lines are displayed in a popup dialog window.

138


-------
E^l

"Errors-Selected" examines the "main_log.txt" file for the selected scenario and
extracts the lines from it that include the word "Error". The lines are displayed in a
popup dialog window.

run)

Irpti

"Run report"- scans the "main_log.txt" files associated with the runs in the
Scenario Library table and generates a report of the status of each run, including
when the scenario was created, when it was executed, how long the run took,
whether it was completed successfully, the number of errors reported, and in
which model periods did not solve correctly.

o

"Refresh" - updates the completion status of the scenarios in the Scenario Library
table. Note that status is not currently updated automatically.

6.1.2 Menu Options

The menu bar for the Scenario Builder includes the following menus: "File", "Tools", "View",
and "Help". The contents of each of these menus is described below.

Table 6.5 Contents of the "File" menu.

Menu Item

Description

Show Options

Displays a non-editable dialog that shows the contents of the options
file. Paths that use #glimpseDir# and #gCamGuiDir# are resolved.

Edit Options

Opens the options file in a text editor. Users can edit the settings in
this file, but any changes do not take effect until either GLIMPSE is
restarted or "File->Reload Options" is chosen.

Reload Options

Reloads the contents of the options file, updating settings in GLIMPSE.

Import Scenario

Allows users to identify an existing configuration file. GLIMPSE reads
the scenario name from the file and creates a new scenario with that
name in the Working Scenarios table. The scenario can then be
executed within GLIMPSE.

Exit

Exits GLIMPSE and terminates related processes. Ongoing GCAM runs
will continue, however.

139


-------
Table 6.6 Contents of the "Tools" menu.

Menu Item

Description

Create or Open
Output Database

Opens the Model Interface to the specified database. If the database
does not exist, the user is given the opportunity to create it.

Check Installation

Parses the options file to search for errors, including if required
parameters are not defined and if folders specified in the file do not
exist. Also examines the folder structure and reports if the main
GLIMPSE folder is nested within the main GLIMPSE folder, indicating
an installation problem. Finally, reads the "JAVA_HOME"
environmental variable to determine which version of java is
specified, then checks to see if that version is on the system.

Check Current DB
Size

Analyzes the current database (usually the one specified in the
options file) to determine its size. Reports the database size in a
popup dialog, including a warning if the database is near its size limit.

Archive Scenario

Supports repeatability to analyzing the configuration file associated
with the selected scenario, then copying all referenced scenario
components to a subfolder. Users can then execute the model using
the archived files if desired.

Fix Lost Handle

If GLIMPSE is terminated while a GCAM run is ongoing, GLMPSE is no
longer able to move the main_log.txt file to the current scenario's
folder. Since GLIMPSE relies on this file to determine the status of a
run, it is no longer able to update the scenario's status in the Scenario
Library. The "Fix Lost Handle" searches the exe/log folder for a
main_log.txt file, reads the scenario name in that file, then moves the
main_log.txt file into the scenario folder with the same name.

Browse Trash

When scenarios and scenario components are deleted in GLIMPSE,
they are placed in a trash folder. This menu item displays the contents
of the trash folder in a file explorer. Note that there is currently no
mechanism to undelete items from the trash folder.

Empty Trash

Deletes all items in the trash folder.

Advanced

Provides access to features for advanced GLIMPSE users.

CSV-to-XML

Advanced users may wish to create their own xml files using the
CSVtoXML function. This function requires the user to specify a data

140


-------


file (a CSV file with a specific format) and an XML header file. The
header file includes instructions on how to convert the CSV to XML.
For more information, please see the "CreatingXMLs" folder,
accessible via "Help->GLIMPSE docs".

Cleanup saved files

Removes ancillary files that are saved when a GCAM run completes
from the various Scenario folders. These files are specified in the
options file, but often include debug files and solver logs, both of
which can be very large.

Table 6.7 Contents of the "View" menu:

Menu Item

Description

Current Main Log

Opens the current "main_log.txt" file from GCAM's exe folder. This file
can be useful for debugging many errors and warnings.

Errors in Main Log

Parses the "main_log.txt" from GCAM's exe folder and extracts lines
starting with the text "Error". These lines are then displayed to a
dialog window, along with an analysis of the severity of market
solution errors.

Current Solver Log

Displays the solver log, which provides very detailed outputs
generated during the solution process. Experienced users may find
this log helpful in identifying problematic markets.

Current Worst
Market Log

Opens the current "worst_market_log.txt" file from GCAM's exe
folder. This file can be useful in understanding solution bottlenecks.

Current Calibration
Log

Opens the current "calibration_log.txt" file from GCAM's exe folder.
This file can be useful in debugging calibration issues.

Debug File

Opens the current "debug.xml" file from GCAM's exe folder. For the
debug region, this file provides detailed output that can be useful in
debugging.

Resource Logs

Provides access to the GLIMPSE log that records computer resource
levels at the initiation of the session and whenever specific thresholds
are exceeded.

Current Session

Resource log for the current session.

141


-------
Prior Session

Resource log for the prior session. The prior log may be most useful if
GLIMPSE or GCAM terminate unexpectedly, and you are checking
whether limited resources may be the underlying cause.

Browse Folder

Opens a file explorer window in the specified location

GLIMPSE Folder

The GLIMPSE folder (GLIMPSE-5p4)

Scenario Folder

The parent folder to the scenario folders

Scenario

Component Folder

The folder that contains the scenario components

Contrib Folder

The folder that contains XML files that are not part of the GCAM data
system

Trash Folder

The folder that contains deleted scenario components and scenario
folders

GCAM exe folder

The folder where the gcam.exe file is located

GCAM log folder

The folder where log files are stored

GCAM input folder

The folder that contains gcam inputs files and the gcam data system

GCAM output folder

The folder that contains the output databases

Advanced

This menu provides access to files that should only be modified by
advanced GLIMPSE users.

Scenario Template

Opens the current "worst_market_log.txt file from GCAM's exe folder.
This file can be useful in understanding solution bottlenecks.

Tech List for Bounds

The "Tech Bounds List" file is used to populate many of the pulldown
menus and technology lists in the "New Scenario Component" dialog.
Only advanced users should modify this file.

Present Regions File

The New Scenario Component Creator's region tree allows users to
select the regions to which a policy will apply. Below the tree is a list
of multi-region groupings that can be used to automatically check
groups of states (e.g., New England) or regions (e.g., North America).

XML Header File

Includes the headers that are used to convert GLIMPSE-developed
scenario components into XML files. Only advanced users should
modify this file.

142


-------
Solver Config File

The "Solver Config File" is used to modify parameters associated with
the GCAM solver components.

Log Config File

This "Log Config File" can be modified to change the types and
verbosity of messages written to the log files and to the screen during
GCAM execution.

Query File

The "Query File" lists all the queries available to the Model Interface.
The name of this file is obtained from the options file and that is listed
in the "modelinterface.properties" file that is in the "GLIMPSE-
Modellnterface/exe" folder.

Table 6.8 Contents of the "Help" menu.

Menu Item

Description

GCAM Docs

Opens a web browser to the GCAM documentation on PNNL's GCAM
GitHub site.

GCAM-USA Docs

Opens a web browser to the GCAM-USA documentation on PNNL's
GCAM GitHub site.

GLIMPSE
Webpage

Opens a web browser to the U.S. EPA's GLIMPSE webpage.

GLIMPSE
Documents

Opens a file browser to a folder with the "GLIMPSE Users' Guide" and
other files that may be of interest.

About

Dates are provided listing when the Scenario Builder, Modelinterface,
and GCAM were last compiled. Clicking on "About" produces an
information dialog with additional information about GLIMPSE.

143


-------
6.2 Scenario Components

GLIMPSE comes with a number of scenario components included in the Component Library. In
this section, brief descriptions are provided for each of the scenario components included with
GLIMPSE. In addition, users can develop their own scenario components using the New
Scenario Component Creator dialog. We describe each of the types of scenario components
that can be created.

6.2.1 Scenario Components included in the Component Ubrary

We include example policies and other scenario components in the GLIMPSE Component
Library. These are listed and described below. Scenario components with an asterisk are part of
the GLIMPSEvl-Reference scenario.

Table 6.9 Scenario components included with GLIMPSE.

Scenario Component

Description

GCAM-Updates-postRelease.txt*

Includes several files that were produced by PNNL
but that were made available after the GCAM-USA
5.4 release. These include non-CCh GHG emission
factors, air pollutant emission factors, and updates
to transportation technologies, including the
addition of electric and hydrogen options across
transportation subsectors.

Calib-biomass_constraints.txt*

Limits the quantity of biomass available in the U.S.

Calib-coal-egu_2020.txt*

Constrains state-level output from coal-fired power
plants in the 2020 model year to be no more than
2021 real-world data. The purpose of this constraint
is to reflect coal plant retirements that have
occurred in recent years. 2020 was not used as the
source of the state-level values since data from that
year were skewed because of COVID-19-related
lockdowns and their impacts on energy demands.

Calib-LDV-EV-AE02020.csv

Limits onroad passenger vehicle electrification (e.g.,
"Cars" and "Large Cars and Trucks") to levels
projected in the Energy Information

144


-------


Administration's Annual Energy Outlook 2020. AEO
projections are relatively low, with the sales share
growing slowly to reach about 10% in 2050.

Calib-Lo_alt_biofuel_SW.csv*

Applies low shareweights to advanced biofuel
production technologies. The purpose of these
constraints is to reflect the slow development of
these technologies in recent years. In aggressive
decarbonization scenarios, modelers may wish to
remove or modify this file.

Calib-NE_fixed_nuke_output.csv

Forces electricity production from nuclear power
plants in New England to follow a retirement
schedule that reflects projections provided by
Connecticut's Department of Energy and
Environmental Resources.

Calib-

NE_fixed_nuke_output_Thru_2020.csv

A version of Calib-NE_fixed_nuke_output that adds
constraints only through 2020, allowing the model
to make its own decisions in subsequent years.

Ca lib-Offshore Windjower-
Bound_EastCoast_thru2035.csv*

Forces generation from offshore wind on the East
Coast of the US to be at least equal to current
procurements.

Calib-onroad_veh_lifetimes.txt*

Modifies onroad vehicle lifetimes to be similar to
assumptions used in the U.S. EPA's MObile Vehicle
Emissions Simulator (MOVES).

Calib-OnroadTrn-SW-Ref-Updates.txt*

Updates shareweights for alternative vehicles. For
light-duty vehicles, hybrid shareweights become 1
by 2025, electric vehicles by 2035, and fuel cell
vehicles by 2050. For heavy-duty vehicles, hybrid
shareweights become 1 by 2035, electric by 2050,
and fuel cell by 2055. In aggressive decarbonization
scenarios, modelers may wish to modify these
assumptions.

Calib-

Lower_lncome_Elasticity_Tran.txt

Some scenarios that result in cost increases to
transportation can result in significant declines in

145


-------


service activity (e.g., international aviation demand
declining by 50%). This file adjusts the elasticity of
transportation sectors to cost, addressing this issue.

Ca 1 ib-20251 ndCoa 1-
GHGPIanStates.csv*

Does not allow investment in new industrial coal
capacity in states that have GHG reduction targets.

Calib-

NoNewCoalEGUslnOCoalStates.csv*

Does not allow new coal capacity, including IGCC
and coal with CCS, to be added in states that are
projected to have eliminated coal in the electric
sector by 2025.

Calib-Transport-ALM-EF-Adj.txt*

GCAM emission factor estimates for air and marine
sources are higher than EPA estimates since GCAM
also includes emissions that could occur outside of
US borders. This file scales emission factors such
that GCAM's national emission projections for these
sectors are similar to EPA estimates.

Policy-C02-Cap-100x50_USA.csv

Specifies a declining cap on CO2 emissions. The cap
starts at roughly 2020 levels in 2020, declining
linearly to zero in 2050. This cap is not intended to
represent any particular policy, but instead to
capture some of the dynamics necessary to meet
deep decarbonization targets. Achieving this target
may require aggressive use of direct air capture
(DAC) and carbon capture and storage (CCS). DAC is
not included in GLIMPSEvl-Reference but can be
added via the "tech-DAC.txt" scenario component.

Policy-C02-Cap-80x50_USA.csv

Specifies a declining cap on CO2 emissions. The cap
starts at roughly 2020 levels in 2020, declining
linearly to a value that is 20% of the starting value.
This cap is not intended to represent any particular
policy, but instead to capture some of the dynamics
necessary to meet stringent CO2 reduction targets.

Policy-C02-Cap-netZero-after-land-
sink.csv

Specifies a declining cap on CO2 emissions. The cap
starts at roughly 2020 levels in 2020, declining
linearly to zero in 2050 - after accounting for land

146


-------


sink opportunities. Thus, the cap does not go to
zero, and instead is more similar to the 80x50 cap.
This cap is not intended to represent any particular
policy, but instead to capture some of the dynamics
necessary to meet deep decarbonization targets.

Policy-HDV-MOU-Heavy-EV_New.csv,

Policy-HDV-MOU-Medium-
EV_New.csv

Policy-HDV-MOU-NoConv-in-2050.csv

Implements medium-and heavy-duty electrification
targets that have been outlined in a multi-state
memorandum of understanding

(https://afdc.energv.gov/Iaws/12460). In this
implementation, we model sales targets that
increase from 15% in 2025 to 100% in 2050.

Policy-RGGI-hold.txt*

Caps electric sector CO2 emissions for a set of states
in the Northeast US through 2030 based upon the
targets outlined by the Regional Greenhouse Gas
Initiative. Holds the targeted values constant from
2030 through 2050.

Policy-RGGI-Thru20200nly.txt

Caps electric sector CO2 emissions for a set of states
in the Northeast US through 2020 based upon the
targets outlined by the Regional Greenhouse Gas
Initiative. No caps are included after 2020. This
version should be used instead of "Policy-RGGI-
hold.txt" for deep decarbonization scenarios in
which the targets post-2020 would not have been
binding.

Policy-RPS_80x50_Natl.csv

Sets an increasing target on the fraction of
electricity that must be produced from renewable
sources. The target increases linearly to a value of
80% in 2050.

Policy-Secl77-ZEV-FreightLtTruck.csv,

Policy-Secl77-ZEV-PassCars.csv,

Policy-Secl77-ZEV-LgCarAndTruck

Implements light-duty onroad electric vehicle sales
targets of 6% in 2020 and 15% in 2025, reflecting
California's ZEV targets. These targets are extended
to all states that had signed on to California's goals,
allowable via Section 177 of the Clean Air Act.

147


-------
Policy-Secl77-ZEV-FreightLtTruck-
20200nly.csv*,

Policy-Secl77-ZEV-PassCars-
20200nly.csv*,

Policy-Secl77-ZEV-LgCarAndTruck-
20200nly.csv*

Implements light-duty onroad electric vehicle sales
targets of 6% in 2020. This target is extended to all
states that had signed on to California's goals,
allowable via Section 177 of the Clean Air Act. The
target ends in 2020 since these constraints would
otherwise interfere with the constraints used to
implement the Light-Duty Near-Term GHG Rule.

Policy-LD-NTR-BEV-Sales-Car.csv*,

Policy-LD-NTR-BEV-Sales-LDTruck.csv*,

Policy-LD-NTR-BEV-Sales-
LDComTruck.csv*

Implements national electric vehicle sales estimates
through 2030 for light-duty onroad vehicles based
on projections from the light-duty, near-term GHG
rule that was finalized in late 2021.

Policy-Secl77-ZEV-LDFreight-
NewCATargets-from2025.csv

Policy-Secl77-ZEV- NewCATargets-
from2025.csv

Policy-Secl77-ZEV-NewCATargets-
from2025-NoConv2030_v2.csv

Policy-Secl77-ZEV-LDFreight-

NewCATargets-from2025-

NoConv2030.csv

Policy-Secl77-ZEV-FreightLtTruck-
20200nly.csv,

Policy-Secl77-ZEV-PassCars-
20200nly.csv,

Policy-Secl77-ZEV-LgCarAndTruck-
20200nly.csv

A set of files that together are intended to
represent the new CA light duty electrification
targets, extended to all of the Section 177 states.

Policy-State_RPS_targets.txt

Reflects state-level Renewable Portfolio Standards
(RPS) and Clean Energy Standards (CES) that had
been adopted by late 2021. These constraints are
mapped to the grid region level, and so each state
may not meet its target, but collectively the targets
will be met. This RPS representation is very memory

148


-------


intensive. While it leads to some additional
renewables, the difference is small after our
renewable cost updates. Based on these factors, we
have chosen not to include the RPS targets in
GLIMPSEvl-Reference.

Policy-NoBiomassEGUinMA.csv*

Massachusetts has eliminated biomass from its
Renewable Portfolio Standard. This file does not
allow new biomass capacity to be built in the state.

Policy-StateCChRdxTargets-
asOf2021.txt

Reflects state-level, economy-wide CO2 reduction
targets that had been adopted by 2021. Many of
these targets are aggressive (e.g., net-zero or 80%
reduction by 2050).

Policy-ROW-CTax-25dpt5pct.csv

US decarbonization results in GCAM tend to use
significant quantities of biomass, including imported
biomass. In deep decarbonization scenarios, other
countries would also have decarbonization targets
and would be competing for this biomass. This tax is
used to simulate that competition.

Tech-DAC.txt

Adds a representation of Direct Air Capture
technologies (DAC). These technologies, which have
not yet been commercialized, are able to remove
CO2 from the atmosphere.

Tech-HighEffTechs-bldgs.csv

Forces high efficiency technology adoption in
buildings by not allowing the purchase of non-high
efficiency technology options after 2020.

Tech-NoCCSinUSA.csv

Does not allow carbon capture and sequestration
(CCS) to be used.

Tech-NoLDVCNGs.csv,
Tech-NoLDVFCEVs.csv,
Tech-NoOn road EVs.csv

Do not allow light-duty natural gas vehicles, fuel cell
vehicles, and electric vehicles, respectively. These
are useful for sensitivity analyses.

149


-------
DeepDeca rbAssumptions.txt

A set of add-on XMLs that may be helpful when



running deep decarbonization targets. These



include:



• a price elasticity of zero for international



aviation, which avoids the problem of



international aviation demand diminishing



drastically when faced with deep



decarbonization targets



• adjustments to the shareweights of advanced



transportation technologies such that any non-



cost biases (e.g., range anxiety) and technical



issues (e.g., weight of batteries for aviation) are



addressed earlier, allowing these technologies



to play a greater role in mitigation.

6.2.2 Developing new scenario components using the New Scenario Component Creator

GLIMPSE supports the creation of several types of scenario components, which are summarized
in the following table. Additional information regarding the options for each are provided in the
next subsection. Please see the Tutorial, Part 3, for an example of how to construct a new
Scenario Component. Here, we provide supplemental information to explain the differences
between the available options.

Table 6.10 Tabs and types of scenario components supported by the New Scenario Component
Creator

Tab name

Type

Description

XML List

Various

Construct a scenario component that consists of a
user-defined list of existing XML add-on files.

Pollutant Tax/Cap

Pollutant taxes
or caps

Specify a tax (1990$/tonne) on CO2 or on GHGs,
or specify system-wide caps on annual emissions
for a region or across regions.

Tech Avail

Technology
availability

Specify whether and when specific technologies
are available for purchase.

150


-------
Market Share

Minimum or
fixed market
share

constraints

Specify the minimum or fixed ratio (as a %)
between the sum of output from one set of
technologies and the sum of output from another
set. This feature has pre-sections for policies such
as RPSs, CESs, LED, heat pump, and vehicle
electrification targets.

Tech Bound

Technology

output

constraint

Specify an upper, lower, or fixed constraint on the
output of a technology.

Tech Tax/Subsidy

Technology
taxes and
subsidies

Modify the levelized cost of a technology by a
specific amount to reflect a tax or subsidy.

Tech Param

Technology
parameters

Provide updated values for the attributes of a
technology, including capacity factor, capital cost,
efficiency, fixed output, shareweight, subsector
shareweight, lifetime, half-life, or emission factor.

Fuel Price Adj

Modifies fuel
prices

Allows prices for coal, crude oil, unconventional
oil, natural gas, and biomass to be adjusted.

Each of these options is provided on a separate tab of the New Scenario Component Creator
dialog window:

¦ New Scenario Component Creator

| XML List | Pollutant Tax/Cap Tech Avail Market Share Tech Bound Tech Tax/Subsidy Tech Pa ram Fuel Price Adj

Figure 6.3 Tabs reflecting the main categories of Scenario Components that can be
constructed via the New Scenario Component Creator.

Together, these tools will allow you to create a wide range of scenario components that reflect
alternative policies and assumptions. In the next subsection, we go through each tab,
highlighting key features and limitations.

151


-------
XML list

Experienced GCAM users likely will have created XML-formatted "add-on" files previously.
These may represent alternative technology costs or efficiencies, or may represent specific
policies. An "XML List" component allows a GLIMPSE user to "point" to one or more existing
XML add-on files. When GLIMPSE creates a scenario and reaches an XML List scenario
component, it will add the files in the XML List to the scenario template.

Clicking "Add" will display a file browser that will allow you to select one or more XML files.
Since the order of input files is important in GCAM (with later files overriding values from
earlier files if the same parameters are in both), you can use the "Move Up" and "Move Down"
buttons to change the order of the files in this list. Click on the "Save" button to create the
XML-list-type scenario component. You will have the opportunity to provide a unique name
when saving.

In the example below, the XML list points to three files. Including this Scenario Component in a
Scenario would result in these files being added to the scenario's configuration file.

New Scenario Component Creator

| XML List
XML Files:

Pollutant Tax/Cap Tech Avail Market Share Tech Bound Tech Tax/Subsidy Tech Param Fuel Price Adj

Add

Move Up

Move Down

Delete

Clear

XML Filename

../../../Contrib/Calibration/bioenergy_constraint\bio_ceiIing_constraint_state.xml
../../../Contrib/Calibration/bioenergy_constraint\bio_ceiling_constraint_ROW.xml
../../../Contrib/Calibration/bioenergy_constraint\bio_ceiling_mkt_global.xml

< I

Save	Close

152


-------
Figure 6.4 "XML list" components point to existing xml-formatted add-on files. In this
example, the XML list points to three files. Including this Scenario Component in a Scenario
would result in these files being added to the scenario's configuration file.

Pollutant Tax/Cap

Pollutant taxes and caps can be created on this tab. "Measure" allows the user to indicate
which of these two options is being created. The tax or cap can be applied system-wide or, for
CO2, can be applied to a specific sector. Taxes or caps can be applied to the pollutants listed in
the "Pollutant" choice box. Note the units in parentheses. For CO2, you have the ability to
specify your cap or tax in units of MTC or MTCO2. If you chose "GHG (MT CO2E)", the cap or tax
is specified in units of CO2 equivalent and is applicable across GHG species, considering their
global warming potentials.

3 • New Scenario Component Creator



XML List

Pollutant Tax/Cap

Tech Avail

Market Share

1

Tec

I Specification:	Va

e Measure:	Emission Tax ($/t)

Pollutant:

Select One

:i

Sector:

Select One

/N. |



C02 (MT C)



Names:

C02 (MT C02)



Policy:

GHG (MT COZE)



Market

NOx (Tg)
S02 (Tg)



Populate:

PM2.5 (Tg)



Type:

CO (Tg)



Start Yean

NH3 (Tg)

—J

End Year:

CH4 (Tg)



Initial Val:

1

Figure 6.5 Pollutants to which a cap or tax can be applied.

153


-------
The figure below shows an example of how the New Scenario Component Creator could be used
to construct a tax on CO2 emissions from the electric sector that starts at $100/tCO2, increasing
at 5% per year. Here, the "USA" region is selected, indicating that the tax is applied to all states.

New Scenario Component Creator

XML List Pollutant Tax/Cap Tech Avail Market Share Tech Bound Tech Tax/Subsidy Tech Pa ram Fuel Price Adj

Specification:

Measure:

Emission Tax (S/t)

•W

Pollutant:

C02 (MT C02)

-

Sector:

EGU

-







Names:
Policy:

yj Auto?



Market

Tax_EGUC02_Reg_Mkt



Populate:





Type:

Initial w/% Growth/yr

-

Start Yean

2020



End Year:

2050



Initial Val:

100

Growth (%}:

S

Values:
Populate

Delete

Clear

Year

Value

2020

100

2025

128

2030

163

2035

208

2040

265

2045

339

2050

432





< I

	I >

Add

'	"	'!	 I

Select region(s):

~ — world
~ @ USA
Canada

Central America and Caribbean
J Mexico
Brazil

South America_Northern
South America_Southern
! Argentina
| Colombia

~	EU-15

~	EU"12

European Free Trade Associatio
Europe_Eastern
I Europe_Non_EU
Africa Eastern

Africa Northerr

Presets:

Select (optional)

Save

Close

Figure 6.6 The "Pollutant Tax/Cap" tab.

Tech Avail

Controlling the availability of technologies is a powerful tool for conducting sensitivity analysis.
For example, by not allowing conventional vehicles to be purchased after 2030, one could
explore a scenario where 100% of vehicle purchases post-2030 are electric vehicles. Similarly,
this option can be used to examine the benefits of buildings electrification or of only allowing
high-efficiency end-use technologies.

The "Tech Avail" tab allows the user to specify over what time period and in which regions a
particular technology or set of technologies is available for purchase.

154


-------
Clicking on "Never" next to a technology will mean that the technology can never be purchased.
This option can be used to specify scenarios that, for example, do not include nuclear capacity
additions or CCS. Alternatively, you can make a technology available only within a specific range
of years.

The "Filter" options allow the user to view just a subset of the technologies. For example,
typing "resid" would show residential technologies. The "Never" and "Range" buttons will
automatically toggle all visible check boxes in the "Never?" and "Range?" columns, respectively.
You can edit the years for which each technology is available by double-clicking on a cell, typing
a new value, then pressing "Enter". Alternatively, you can set the values in the "First" and
"Last" columns for all visible cells simultaneously using the "First yr:" and "Last yr:" fields at the
bottom, then clicking "Set Years".

In the following example, we have specified that "Car", "Large Car and Truck", and "Mini Car"
vehicles powered by natural gas can only be purchased from 1975 to 2015 in NC. This
essentially eliminates their sale in the state after 2015.

l!_

New Scenario Component Creator



X

XML List Pollutant Tax/Cap Tech Avail Market Share Tech Bound Tech Tax/Subsidy Tech Pa ram

Fuel Price Adj

Select technologies and specify all, first or last years to constrain new purchases:



Select region(s)r

Filter by Sector: trn_pass_road_LDV_4W » ^e>ct:



~	Ml

~	MN



Never? Range? First Last Sector: Subsector: Technology: Units



~ MO



© 1975 2015 trn_pass_road_LDV_4W : Car : NG : million pass-km



~ MS



IV 1975 2015 trn_pass_road_LDV_4W : Large Car and Truck : NG : million pass-km





Q mt



0 1975 2015 trn_pass_road_LDV_4W : Mini Car : NG : million pass-km





0 NC









~ ND









~ NE









~ nh







~ NJ









~ nm









~ NV









~ NY









~ OH

















~ OK









~ 0R









~ PA











< C	J

>



Presets: Select (optional) ~

Select 1 Never I 1 Range 1 F,rst^ [ 1975 | "^ty* 2015 f Set Year;





Save Close

155


-------
Figure 6.7 The "Tech Avail" tab.

Market Share

The "Market Share" tab can be used to implement market share targets for one or more
technologies. The user specifies a ratio (expressed as a % of either sales or activity) between
the outputs of one set of technologies and those of another. The technologies that make up the
"numerator" in the ratio are specified via the "Subset:" selector, while the technologies in the
"denominator" are specified via "Superset:". For example, if you want to specify that 50% of
"Car" sales be electric, the "Subset" would include BEV (battery-electric-vehicles) Cars, while
the Superset would include all Car technologies.

For convenience, there are several pre-specified options, including a Renewable Portfolio
Standard (RPS), Clean Energy Standard (CES), light duty electrification targets for cars and trucks
etc. Selecting one of these options will pre-populate the "Subset" and "Superset" selections. A
message will pop up indicating that the user should check the selections to determine if they
meet their objectives. For example, an RPS may include hydropower in one state, but not in
another. Hydro is automatically checked when the user selects "RPS".

Select One

A 1

Renewable Portfolio Standard (RPS)



Clean Energy Standard (CES)



EV passenger cars and trucks



EV passenger cars trucks and MCs



EV freight light truck



EV freight medium truck



EV freight heavy truck



EV freight all trucks



LED lights



Initial and Final % ~

Figure 6.8 Preset options for specifying market share constraints.

The options next to "Constraint:" allow you to specify whether this is a "Lower Bound" or a
"Fixed Bound". If "Lower Bound", percentages must be equivalent to the constraint or higher. If

156


-------
"Fixed Bound", the constraint must be met exactly. The "Lower Bound" tends to be easier for
the solver to meet, and we advise that it be used instead of the "Fixed Bound" where possible.

The "Applied to:" pulldown menu allows the market share constraint to apply to either the
entire stock or just to new purchases. This distinction is important since some policies (e.g., EV
targets) typically target a sales percentage, while others (e.g., RPS or CES) are applied to the
output of both existing and new stock.

The "Treatment:" option allows the constraint to be applied to "Each Selected Region" or
"Across Selected Regions".

Together, these options provide a wide range of constraint designs. Below, we have specified
an RPS that increases from 20% in 2020 to 100% in 2050. The target is applicable to overall
electricity production as opposed to sales and is a national total.

£3 New Scenario Component Creator

XML List Pollutant Tax/Cap Tech Avail Market Share Tech Bound Tech Tax/Subsidy Tech Param Fuel Price Adj

Specification:

Type?

Subset:

Superset:

Constraint:

Applied to:

Treatment:

Names:
Policy:
Market

Populate:
Type:

Start Year:
End Year:
Initial (%):
Final (%):

Renewable Portfolio Standard (R...	~
base load generation : biomass=...

base load generation : biomass=...	~

Lower	~

All Stock	~
Across Selected Regions

\>/\ Auto?

Renewable Portfolio Standard RPS All

Initial and Final %

2020
2050
20
100

Values:
Populate
Year

Delete

Clear
Value

2020

0.200

2025

0.333

2030

0.467

2035

0.600

2040

0.733

2045

0.867

2050

1.00

Select region(s):

~ — world
~ @ USA

Canada

Central America and Caribbean

Mexico

Brazil

South America_Northern

~	South America .Southern
Argentina

Colombia

~	EU-15

~	EU 12

European Free Trade Associatio
Europe_Eastern

~	Europe_Non_EU
Africa_Eastern

Africa Northern	\

Presets:

Select (optional)

Add

Save	Close

Figure 6.9 The "Market Share" tab.

157


-------
Tech Bound

While "Market Share" allows the user to specify targets as percentages, the "Tech Bound"
option allows a constraint to be placed on the output of one or more technologies. The
constraint can represent an upper, lower, or fixed bound, and is specified in the output units of
the technology.

Similar to the "Market Share" option, the constraint can be applied for each selected region or
across regions.

Below is an example in which a "Tech Bound" constraint is used to force offshore wind in the
Northeast U.S. to be equal to or higher than current planned procurements.

* New Scenario Component Creator

XML List Pollutant Tax/Cap Tech Avail Market Share Tech Bound Tech Tax/Subsidy Tech Pa ram Fuel Price Adj
Specification:	Values:

Select region(s):

Populate:
Type:

Start Yean
End Year:
Initial Val:
Final Val:

Sector:
Tech(s):
Constraint:
Treatment:

All



Populate Delete

Clear



base load generation : wi...

¦w

Year

Value



Lower



2025
2030



0.0639
0.213
0.213

Across Selected

-







2035

Names:

\/ Auto?



2040



0.213

Policy:





2045



0.213

Market:





2050



0.213

Initial and Final

2020
2050

< I

Add

Presets:

Select (optional)

Save

Close

Figure 6.10 The "Tech Bound" tab.

158


-------
Tech Tax/Subsidy

The "Tax/Subsidy" tab allows the user to adjust the price of a technology. This price adjustment
is in million 1975$s per unit of output. Thus, a subsidy for electric vehicles would be in units of
million 1975$s per million pass-km, or 1975$s/pass-km. The user is responsible for any
conversions necessary to convert units. Thus, a subsidy of $1000 for an electric car in today's $s
would need to be modified to consider the load (passengers per vehicle), usage (km per year
per vehicle), the amortization of the car's capital cost (which by default assumes a 5% discount
rate and a 10-year payment period), and conversion to 1975$s.

Alternatively, one can think of the subsidy as reducing the levelized cost of the car. In the
example below, we are reducing the cost of EVs by $0.10/pass-km (in 1975$s),

* New Scenario Component Creator

j XML List | Pollutant Ta^/Cap || Tech Avail | Market Share |j Tech Bound || Tech Tax/Subsidy || Tech Param || Fuel Price Adj

Specification:

Type:

Names:

Policy:

Market:

Values:

Sector:

trn_passj"oad_LDV_4W

¦w

Populate

Delete

Clear

Tech(s):

trn_pass_road_LDV_4W: C...

~

Year

Value

Subsidy

•y Auto?

T echC A_MJct

Populate:



Type:

Initial w/% Growth/yr ~

Start Yean

2020

End Year:

2050

Initial Val:

,10

Growth (%}:

0

2020
2025
2030
2035
2040
2045
2050

< i

0,100
0,100
0.100
0.100
0.100
0.100
0.100

Add

Select region(s):
~ — world
~ H USA

~	AK

~	AL

~	ar

~	AZ
0 CA

~	co

~	CT

~	DC

~	de

~	fl

~	ga

~	HI

~	IA

~	ID

n it

Presets:

Select (optional)

Save

Close

Figure 6.11 The "Tech Tax/Subsidy" tab.

Note that both taxes and subsidies are expressed as positive values. Taxes add to the cost of a
technology while subsidies subtract from the cost.

159


-------
Tech Param

The "Tech Param" tab allows users to modify a subset of technology attributes, including their
shareweight, lifetime, and half-life. Additional attributes options are being tested and may be
added in the future.

!¦_! New Scenario Component Creator	X

1

XML List Pollutant Tax/Cap Tech Avail Market Share Tech Bound Tech Tax/Subsidy

Tech Param

:uel Price Adj

Specification:



Values:





Select region(s):

Sector:

H2 central production ~











Populate Delete

Clear



v |—| world















Tech(s):

H2 central production : bL. ~



Year

Value





~ |Vj USA

Parameter:

Shareweight ~



2020



0.0500



Canada

Central America and Caribbean
Mexico



Fixed Output



2025



0.208



Input:

Shareweight



2030



0.367



' Brazil

Output:

Lifetime
Ha If life



2035
2040



0.525
0.683



South America_Northern
South America_Southern
Argentina
| Colombia

Populate:



2045



0.842



Type:

Initial and Final ~



2050



1.00



~ EU-15

Start Year

2020











~ EU-12

European Free Trade Associatio
Europe_Eastern

End Year:

2050











Initial Val:

.05











Europe_Non_EU

Final Val:

1











Africa_Eastern
I Africa Northern v

























< I _ " l >_|





< I



I >



Presets: Select (optional) ~





Add















Save	Close

Figure 6.12 The "Tech Param" tab.

Fuel Price Adj

"Fuel Price Adj" allows users to develop alternative scenarios about the prices of coal, natural
gas, oil, and biofuels into the future, One or more fuels can be selected for adjustment. The
adjustments must be provided in units of million 1990$s per EJ. In the example below, a
$10Q/EJ increase in cost is applied globally from 2020 through 2030.

160


-------
¦ New Scenario Component Creator

XML List Pollutant Tax/Cap Tech Avail Market Share

Specification:



Fuel:

Crude Oil, Unconv Oil

Units:

million 1990$s per EJ

Names:

V Auto?

Policy:

FuelPriceAdj-mult-Reg

Market:

FuelPriceAdj-mult- Reg_Mkt

Populate:



Type:

Initial w/% Growth/yr ~

Start Yean

2020

End Year:

2030

Initial Val:

100

Growth (%):

0

Values:

Populate

1 	

Year

Delete

Clear
Value

2020

100

2025

100

2030

100

< I

Add

~ 0 world

A

~ 0

USA



¦

Canada



0

Central America and Caribbean



0

Mexico



0

Brazil



0

South America_Northem



0

South America_Southern



0

Argentina



0

Colombia



0

EU-15



0

EU-12



0

European Free Trade Associatio



0

Europe_Eastern



0

Europe_Nort_EU



0

Africa_Eastern



>/ Africa Northern

V

< l

	J > _

Presets:

Select (optional)

Save	Close

Figure 6.13 The "Fuel Price Adj" tab.

6.3 GLIMPSE queries in the Mode/Interface

When a GCAM run is completed, the results are saved to a database. The database is typically in
a subfolder of the GCAM-Model\gcam-v5.4\output, such as "database". The contents of the
database folder include seven files that end in ".basex". These are binary files that cannot be
opened directly. Instead, data are extracted from the database using queries that are written in
the XPath query language. These queries can be executed via R scripts or via the
Mode/Interface. Queries are located in a query file, which is specified in "GLIMPSE-
Modellnterface/exe/model_interface.properties". GCAM's query file is typically
"standard_queries.xml". In GLIMPSE, a file named "Main_queries_GLIMPSE-5p4.xml" is used
instead. This file includes all the queries in the standard_queries.xml file but adds several
groups of "GLIMPSE" queries to the top of the list. These include the queries that we have
found to be of most use to GLIMPSE users.

161


-------
When choosing which queries to use, it is important to be familiar with the terminology in their
names.

Primary energy refers to energy (e.g., natural gas, crude oil, coal, or biomass) that has not been
converted or transformed to another form (e.g., electricity, gasoline, or biofuels). Wind and
solar energy are sometimes reported as direct equivalents (the amount of electricity produced)
or as fossil equivalent (the amount of electricity produced, divided by the average fossil
efficiency factor, often 33%).

Final energy is the energy used in meeting end-use demands. Therefore, the natural gas going
into power generation is not final energy, but the electricity being used in residences is.

Cogen and combined heat and power (CHP) are used synonymously. Both represent industrial
technologies that produce both electricity and heat. Electricity production from these
technologies is typically used within the industry's fence line, meaning that it is not sold onto
the electric grid.

Many of the queries refer to "sectors". This is a GCAM model construct that is somewhat
analogous to economic sectors. However, the sector names are not always intuitive. For
example, instead of reporting "electric sector" outputs, results are reported as "baseload
generation", "intermediate generation", "peak generation", and "subpeak generation".

The GLIMPSE queries include additional aggregation into groupings that may be more intuitive.
Aggregated sectors include "electricity", "industry", "fuel production", "commercial",
"residential", "transport-LDV", "transport-HDV", and "transport-ALM". In these aggregations,
"industry" includes all industrial processes except those related to the fuel chain. "LDV" is
"Light-Duty Vehicles" and represents passenger cars and trucks. "HDV" is "Heavy-Duty
Vehicles". This includes Heavy, Medium, and Light trucks used for moving freight. "Fuel
production" includes oil and gas production, coal mining, refining, and pipelines.

Below, we list the queries that fall under the GLIMPSE categories, providing a brief description
of each. Where a query is followed by "(GCAM-USA)" or "(core)", users must use the query
corresponding to the version of the model they are using, where "core" indicated GCAM.

Queries that end in "(Total)" will sum results across all selected regions. The "(Total)" queries
are not shown in the tables below.

162


-------
Table 6.11 Query group: Primary and final energy

Query

Description and use

Primary energy consumption by
region (direct equivalent)

Primary energy consumption without converting wind
and solar to fossil equivalents. Useful in examining how
resource consumption changes over time.

Final energy consumption by
region

Regional totals of final energy, summing all fuels and
energy services. Useful in examining how demands for
energy change over time.

Final energy consumption by
aggregate sector

Final energy, summed across fuels, but does so at the
aggregated sector level. Useful in understanding which
sectors consume the most energy and how their energy
demands change over time.

Final energy consumption by
aggregate sector(v2)

Similar to the prior query but combines the residential
and commercial sectors into a buildings category and
combines light- and heavy-duty onroad vehicles into the
Transport-onroad category.

Final energy consumption by
aggregate sector and fuel

Final energy consumption by fuel, at the aggregated
sector level. Useful for understanding how much fuel is
used in each sector.

Final energy consumption by
sector and fuel

Final energy by fuel but does so at the GCAM sector level.

Table 6.12 Query group: Energy inputs by sector

Query

Description and use

Electricity use by aggregate
sector

Electricity use at the aggregated sector level. Results
include both final energy and the electricity consumed in
conversion technologies (e.g., refineries). Useful in
understanding in which sectors electricity is being used
and how that use changes over time.

Coal use by aggregate sector

Coal use at the aggregated sector level. Useful in
understanding in which sectors coal is being used and
how that use changes over time.

163


-------
Natural gas use by aggregate
sector

Natural gas use at the aggregated sector level. Useful in
understanding in which sectors natural gas is being used
and how that use changes over time.

Refined liquids use by aggregate
sector

Refined liquids (e.g., gasoline, diesel, and biofuels) use at
the aggregated sector level. Useful in understanding in
which sectors liquid fuels are being used and how that
use changes over time.

Biomass use by aggregate sector

Biomass use at the aggregated sector level.

Hydrogen use by aggregate
sector

Hydrogen use at the aggregated sector level.

End-use energy consumption in
buildings

Energy used across the buildings sector by fuel.

End-use energy consumption in
transportation

Energy used across the transportation sector by fuel.

End-use energy consumption in
industry

Energy used across the industrial sector by fuel.

Energy inputs in electricity
production (GCAM-USA)

Quantities of various fuels used by each electricity
production subsector. Useful in quantifying fuel use in the
electric sector.

Energy inputs to refining
activities (GCAM-USA)

Fuels used in producing refined liquids, including
petroleum-based fuels and alternative fuels.

Table 6.13 Query group: Technologies

Query

Description

Electricity generation by region
(no cogen)

Total electricity produced in each region. Industrial
generation via combined-heat-and-power and co-gen are
not included in this total. Useful in understanding how
much electricity is produced in each region and how that
quantity changes over time.

Electricity generation by gen
and cooling tech (incl cogen)

Generation of electricity by each combination of
generation and cooling technology. This is the most

164


-------


detailed electric sector output. Useful for performing your
own aggregations.

Electricity generation by
aggregated subsector

Generation of electricity by major fuel category (such as
coal, natural gas, wind, and solar), with electricity
production with CCS reported separately.

Electricity generation by
aggregated subsector rnw detail

Generation of electricity by major fuel category (such as
coal, natural gas, wind, and solar), with electricity
production with CCS and various solar and wind
technologies reported separately.

Electricity generation by
subsector

Electricity generation at the subsector level. Useful when
you are interested in broad shifts over time (e.g., changing
market shares among coal, gas, wind, and solar), as
opposed to detailed information about the specific coal or
gas technologies.

Electricity generation by cogen
only

Electricity produced by cogen and combined-heat-and-
power technologies. Useful when examining cogen and
CHP at the technology level.

Electricity generation input

Fuel used in electricity production.

Building final energy by tech

Final energy used in buildings by technology. Useful in
understanding the quantity of fuel used by various
technologies in meeting service demands.

Building service output by tech

Service outputs in the residential and commercial sectors,
reported by uses such as space heating, water heating,
and lighting. Useful in understanding the market share of
various technologies.

Industry final energy by tech
and fuel

Final energy used in industry by technology. Useful in
understanding how and where fuels are used to meet
industrial demands.

Refined liquids production by
tech

Quantity (EJ) of liquid fuels produced by various
technologies, such as oil refining, corn and cellulosic
ethanol production, biodiesel, and flow-through gasoline
fuels. Useful in understanding the liquid fuel mix, including
the fraction from alternative fuel sources.

165


-------
Hydrogen production by tech

Hydrogen production by technology.

Transport final energy by tech
and fuel

Final energy used by technology in meeting end use
service demands. Useful in understanding sectoral fuel use
and the impacts of changes in vehicle efficiency.

Transport service output by
tech

Transportation service outputs met by various
technologies and fuels. Useful in understanding the
market penetrations of alternative fuel vehicles.

Transport service output by
tech (new capacity)

Vintage-specific transportation service outputs met by
new sales (in units of pass-km).

Passenger car and truck service
output by tech

Onroad light-duty passenger travel demands (pass-km)
met by various technologies across vehicle sizes.

Freight truck service output by
tech (no bus)

Onroad freight travel demand (tonne-km) met by various
technologies across vehicle sizes.

Table 6.14 Query group: Emissions

Query

Description

CO2 emissions by region

Total CO2 emissions (in units of MTC) per region

CO2 emissions by aggregate
sector

CO2 emissions (in units of MTC) by aggregate sector.
Combustion emissions include all CO2 in the exhaust gas.
For biofuels and bioenergy, the negative CO2 associated
with biomass growth is captured in the biomass category.

CO2 emissions by sector

CO2 emissions (in units of MTC) per GCAM sector

CO2 emissions by sector (no bio)

CO2 emissions (in units of MTC) per GCAM sector, not
including emissions from the combustion of bioenergy.
Biofuels-related emissions are handled as usual, however.

CO2 emissions by resource
production

CO2 emissions (in units of MTC) from resource production
activities, including mining and oil and gas operations.
These emissions are not reported in the emissions-by-
technology queries.

CO2 emissions by tech

CO2 emissions (in units of MTC) by technology. This query
does not include the results of "CO2 emissions by resource
production".

166


-------
NOx, SO2, PM2.5 by region

Total emissions of several air pollutants (in Tg, which is
equivalent to million metric tonnes)

NOx, SO2, PM2.5 by aggregate
sector

Sectoral emissions of several air pollutants (in Tg, which is
equivalent to million metric tonnes)

NOx, SO2, PM2.5 by sector

Sectoral emissions of several air pollutants (in Tg, which is
equivalent to million metric tonnes)

NOx, SO2, PM2.5 by tech

Emissions of several air pollutants by technology

All emissions by region

All tracked emission species by region

All emissions by aggregate
sector

All tracked emission species by aggregate sector

All emissions by sector

All tracked emission species byGCAM sector

All emissions by tech

All tracked emission species by technology

All emissions by resource
production

All reported emissions from resource production
activities, including mining and oil and gas operations.
These emissions are not reported in the emissions-by-
technology queries.

Table 6.15 Query group: Impacts

Query

Description

CO2 concentration

Global average CO2 concentration in parts-per-million
(PPM)

Forcing total

Climate forcing in watts per square meter

Global mean temperature

Global mean temperature in degrees C

Table 6.16 Query group: Markets, prices, and costs

Query

Description

CO2 prices

Market price for CO2 reduction in 1990$s/tC

Prices of all markets

Market prices in 1990$s

Supply of all markets

Supplies at the equilibrium prices

Demand of all markets

Demands at equilibrium prices

167


-------
Final energy prices

Prices for energy commodities in meeting end use
demands in 1990$s

Costs by tech and input

Costs of technologies and fuels in 1990$s

Building service costs

Costs of meeting energy services in 1990$s

Costs of transport services

Costs of providing transportation services in 1990$s

Elec prices by sector

Electricity prices in 1990$/EJ

Table 6.17 Inputs and outputs

Query

Description

Inputs by tech

Inputs into each technology. Useful for tracking fuel and
material flows.

Outputs by tech

Outputs from each technology. Useful for tracking fuel and
material flows.

Table 6.18 Query group: Assumptions

Query

Description

Population by region

Exogenously determined population by region
(thousands). Note: USA population is reported both at the
state and national level. Do not sum over population USA
and states to avoid double-counting.

GDP per capita PPP by region

Gross domestic product per person, with GDP converted
to 1990$s US using purchasing power parity conversions
for each country.

Elec shareweights by subsector

Shareweights for electricity production subsectors. A
value of 1 reflects no bias relative to other subsectors.

Elec shareweights by gen tech

Shareweights for electric sector technologies. A value of 1
reflects no bias relative to other technologies.

Building floorspace

Building floorspace in billion square meters. Calculated as
a function of population.

Building shareweights by
subsector

Shareweights for buildings subsectors. A value of 1
reflects no bias relative to other subsectors.

168


-------
Building shareweights by tech

Shareweights for buildings technologies. A value of 1
reflects no bias relative to other technologies.

Industry shareweights by
subsector

Shareweights for industrial subsectors. A value of 1
reflects no bias relative to other subsectors.

Refining shareweights by tech

Shareweights for refining technologies. A value of 1
reflects no bias relative to other technologies.

Refining and oil shareweights by
subsector

Shareweights for refining subsectors. A value of 1 reflects
no bias relative to other subsectors.

Transport subsector
shareweights

Shareweights for transportation subsectors. A value of 1
reflects no bias relative to other subsectors.

Transport tech shareweights

Shareweights for transportation technologies. A value of
1 reflects no bias relative to other technologies.

Table 6.19 Query group: Water demands

Query

Description

Water withdrawals by state,
sector, basin (include desal)

Water withdrawals in units of km cubed

Water consumption by state,
sector, basin (includes desal)

Water consumed in units of km cubed. Does not include
quantity replaced into reservoir, river, etc.

6.4 Troubleshooting

6.4.1 Common Problems and Errors Using the Scenario Builder

Table 6.20 Scenario Builder problems and solutions

Problem

Potential Cause

Solution

After clicking "Run", ^
, the Cmd window
appears, then
immediately
disappears.

The location of Java on your
computer has changed as
the result of an update. For
example, Java may have
been located at "c:\Program
Files\Java\jrel.8.0_321"

Check the latest version of Java on
the computer, then update the
JAVA_HOME setting in the batch
file (run_GLIMPSE_ORD-GCAM-
USA-5p4.bat) in the GLIMPSE

169


-------


prior to the update but may
have been moved to
"c:\Program
Files\Java\jrel.8.0_333"
after the update, reflecting
the new version number.

folder. Restart GLIMPSE for the
change to take effect.

The configuration file is not
a properly formatted XML
file.

Inspect the configuration file for
errors. Some text editors have
features for checking the
formatting of XML files.

After clicking "Run",
only some components
are loaded before the
Cmd window
disappears.

One of the scenario
components does not exist
or the path name is
incorrect.

Check the last loaded scenario
component and the next scenario
component listed in the
main_log.txt file. Ensure that they
exist and are properly formatted.

After clicking "Run", all
components are
loaded, but the Cmd
window disappears
before the first period
completes.

A technology is referenced
in one of the add-on files,
but that technology does
not exist in the specific
region.

Check the main_log.txt file for an
error related to a DISCRETE market
not having been set up for a
technology. If the technology name
is misspelled, correct the spelling in
the relevant add-on file. If the
technology should not exist in some
markets, modify the add-on XML to
include the NOCREATE=l tag.

The model appears to
be running correctly;
however, the gcam.exe
window disappears
while the Scenario
Library. The
"main_log.txt" file in
the "exe" folder does
not include any details

The computer may have run
physical memory or disk
space (either on the drive
where your model is being
run or on the drive where
your "swap" file is stored).

A log is kept that records entries
indicating that key thresholds have
been exceeded. You can check
these resources via "View-
resource Logs->Current Log".
Please see "How do 1 ... determine
what caused GCAM to fail?"

170


-------
about why the run
terminated.





In the scenario's
components, there may be
conflicting market or policy
names that can cause the
model to terminate
unexpectedly. This typically
will not be a problem with
components created wholly
in GLIMPSE since GLIMPSE
generates unique market
and policy names.

Examine the scenario's market and
policy names for conflicts.
Iteratively executing the model
while removing one or more
scenario components may help
identify which files include
problematic naming conflicts.

The model seems stuck
and is not producing
new output lines in the
Cmd window.

The model is having
difficulty solving one or
more markets and the
solver is requiring an
unusual number of
iterations.

Within GLIMPSE, choose "View-
>Worst Market Log" from the menu
bar. Wait several minutes and
repeat. Compare the end of the
two files. Are they the same? If not,
then the model is still executing.
You may want to check the GCAM
website information on the solver
and its settings for guidance:
https://igcri.github.io/gcam-
doc/solver.html

GCAM has paused execution
for some other reason. A
common reason is that the
user has selected text in the
Cmd window with the
mouse, which will pause the
model execution.

Sometimes clicking in the Cmd
window, then pressing "Enter" will
result in GCAM execution
continuing.

After revising the
option file in the

The options file has not
been reloaded.

Restart GLIMPSE or choose "File-
reload Options" from the Scenario
Builder menu bar.

171


-------
notepad, the options
have not taken effect.





Modifications to
options file parameters
such as

"gCamOutputDatabase"
and "stop-period" are
not reflected in existing
scenarios.

The scenario was created
before the changes to the
options file were
introduced.

Click on the scenario in the Working
Scenarios table, then click on the
arrow button, . In the "Create
Scenario" pane, click on the

"Create" button . This will
update the scenario with the new
settings.

GCAM outputs
warnings and errors to
the Cmd window and to
the main_log.txt file.

Many such errors and
warnings are not serious
and do not affect the model
results. Most serious errors
will cause GCAM to
terminate.

Examine the run's main_log.txt file.
Does it include the text "Model
completed successfully?" If not,
then there may be more serious
issues.

The Scenario Builder
indicates that there
were solution problems
and identifies specific
time periods with
market failures.

The solver had difficulty
solving one or more markets
before the allowable
number of solver iterations.

Please see the section of this
manual in the Advanced Concepts
and Operations that discusses
market solution errors. GCAM
documentation for guidance on
changing solver settings may also
be useful:

https://igcri.github.io/gcam-
doc/solver.html

Some scenario
component XMLs are
not being created when

the "Create" button ®
is pressed.

GLIMPSE may be having
difficulty interpreting the
path names to files. This can
be the case if you have
scenario components or
scenario names that have
spaces or non-standard
characters.

Do not use spaces in path or
filenames when installing GLIMPSE
or constructing new scenarios or
scenario components.

172


-------
When you double-click
on a scenario name in
the Scenario Library or
do other actions that
should display a file, no
text or XML editor is
displayed.

The XML and text editor
options in the options file
may be referring to editors
that you do not have on
your computer.

Check and revise the text and XML
editor specifications, if necessary.
Reload the options file.

The status of a scenario
in the Scenario Library
is "Lost Handle".

Stopping the Scenario
Builder will not terminate
any ongoing GCAM runs.
However, GLIMPSE will not
be able to copy the
main_log.txt file to the
scenario's folder when the
run terminates. However,
when determining status,
GLIMPSE uses the
main_log.txt file, so status
cannot be updated.

In the Scenario Builder, select
"Tools->Fix Lost Handle". This will
find the main_log.txt file in the
exe/logs folder, parse it to
determine the scenario with which
that main_log.txt file is associated,
then place a copy in the scenario's
folder. This should resolve the lost
handle message.

In the Scenario Builder
or New Scenario
Component Creator,
some options are not
visible.

Windows includes options
for scaling font sizes to be
larger or smaller. Depending
on these settings, it could be
that the text no longer fits
on the application window.

In the options file, there is an
option for setting the font size.
Adjust the size up or down as
needed, restarting GLIMPSE to test
whether this addresses the issue.

6.4.2 Common Problems and Errors Using the Modellnterface

Table 6.21 Modellnterface problems and solutions

Problem

Potential Cause

Solution

Using GCAM-USA,
queries of the USA
region produce only

InGCAM-USA, the "USA"
region outputs include only
those categories that have

Select the "USA" region and all states
to obtain U.S. national totals.

173


-------
a small number of
outputs.

not been assigned to states,
such as: agriculture; oil, gas,
and coal extraction; and
hydrogen production.



Using GCAM-USA,
electric sector
queries produce
results for only a few
technologies.

GCAM-USA uses a different
set of queries for many
electric sector outputs.

If GCAM-USA-specific queries exist
for an output, use those.

The Model Interface
freezes and is non-
responsive

This happens occasionally for
non-obvious reasons.

There may be some incompatibilities
with certain versions of the Java
virtual machine. For example, we
have found that the Model Interface
occasionally will freeze when one
tries to open a File Chooser, for
example, when exporting a result
from the database. Terminating the
Scenario Builder allows the
Mode/Interface's File Chooser to
appear. It is not clear what is causing
this behavior. However, we have not
had this issue with jrel.8.0_333, so
upgrading your version of Java may
be a solution.

You may need to terminate, then re-
start the Model Interface. Press cntl-
alt-del and select the Task Manager.
Click on "More details". Find the
option "Java™ Platform SE binary"
and expand the "carrot" to the left. If
you see an icon for the
Mode/Interface, select that process.
If you do not, see if the
Model Interface icon is below
another Java process. Click on it.

174


-------




Press "End Task". You may need to





do this twice to kill the process.





Restart by pressing the "Results"





button on the Scenario Builder.



The database size has

Unfortunately, this means that the



become too large and the

data in the database is lost. Use the



basex software used by the

"Check Current DB Size" option to



Model Interface can no longer

determine the database size. If it is



access the database's files.

larger than 40 or 50 GB, size is the





issue. You will need to create a new





database. Please see Database Size





Management in the "How do 1...?





section for best practices.

6.5 Glossary

In this section, key terms and acronyms used in GLIMPSE and in the query list are defined.

Table 6.22 Glossary of acronyms and key terms.

Acronym or term

Description

2W

2-wheel powered vehicles, such as motorcycles

2W&3W

2- and 3-wheel powered vehicles, such as motorcycles

4W

Onroad passenger vehicles with 4 wheels, including cars, pickup
trucks, SUVs, and vans. To simplify data requirements, these are
categorized as "Car", "Large Car and Truck", and "Mini-Car" in
GCAM 5.4.

Adv

Advanced

Aggregate sector

Aggregation of GCAM sectors into categories such as "Electricity",
"Industry", "Commercial", "Fuel production", "Transportation-
LDV", "Transportation-HDV", and "Transportation-ALM".

175


-------
Base

Baseload component of a load duration curve

BC

Black Carbon, a climate pollutant

BEV

Battery-Electric Vehicle, also known as electric vehicle

CC

Combined-Cycle, a type of high efficiency turbine

CCS

Carbon Capture and Sequestration

CES

Clean Energy Standard

ch4

Methane, a climate pollutant

CHP

Combined heat and power technologies that produce both
electricity and heat

CO

Carbon Monoxide, an air pollutant

Cogen

Cogeneration technology that produces both heat and electricity

Comm

Commercial sector

CSP

Concentrated Solar-Thermal Power

Conv

Conventional

CT

Combustion Turbine

DAC

Direct Air Capture - technologies capable of removing CO2 from
ambient air

Delivered

Fuel that has been delivered to meet end-use demands and thus
includes costs associated with transportation and distribution

Direct equivalent

When comparing primary energy usage, renewable energy is often
expressed in terms of fossil equivalent or direct equivalent. For
example, with solar power, "direct" is equal to the actual electricity
produced by the solar device. If "fossil", then the electricity is
divided by the average fossil plant efficiency, often 0.33 to 0.4.

EJ

Exajoules, equivalent to lel8 Joules or 2.78e5 gigawatt-hours

Elec_td_bld

Electricity used in buildings

Elec_td_ind

Electricity used in industry

Elec_td_trn

Electricity used in transportation

176


-------
Enduse

Energy uses in buildings, industry, and transportation.

FCEV

Fuel Cell Electric Vehicle, a technology that uses electricity
produced by a fuel cell to power an electric vehicle

Final energy

Energy used in meeting end-use energy service demands such as
passenger travel and lighting

FT

Fischer-Tropsch process that converts carbon monoxide and
hydrogen or water into liquid hydrocarbons

Gas

Natural gas

Gen_ll_LWR

Conventional nuclear plants utilizing a Light Water Reactor

GenJII

Advanced nuclear power technologies

H2, H2

Hyd rogen

HDV

Heavy-duty onroad vehicle

Hi-eff

High-efficiency version of a technology

HSR

High-speed electric rail

Hybrid liquids

Hybrid technologies that are fueled by refined liquids

Hydro

Hyd ropower

IGCC

Integrated Gasification Combined-Cycle advanced combustion
technologies

Int

Intermediate portion of a load duration curve

Kg

Kilogram

km2, kmA2

Square kilometer

km3, kmA3

Cubic kilometer

LD-NTR

Light-Duty, Near-Term Greenhouse Gas Rule

LDV

Light duty onroad vehicles

Liquids

Often refers to technologies fueled by "refined liquids"

m3, mA3

Cubic meters

Mt

Mega-tonnes, or million metric tonnes

177


-------
MTC

Mega-tonnes of Carbon, where one MT is equivalent to 1.102xl0A6
short tons

N20, N20

Nitrous oxide - A climate pollutant

N

Nitrogen

NG

Natural gas or Compressed Natural Gas (CNG)

NH3, NH3

Ammonia - an air and water pollutant, precursor to tropospheric
ozone and secondary particulate matter

NMVOC

Non-methane Volatile Organic Carbon - air pollutants and
precursors to tropospheric ozone

NOx

Nitrogen oxides - an air pollutant and precursors to tropospheric
ozone and secondary particulate matter

OC

Organic Carbon

Pass-km

Passenger-kilometer

Peak

Portion of the load duration curve that represents the hours with
the highest electricity demand

Petalumen-hours

Units used to express demand for lighting

PM10, PM10

Particulate matter with a diameter less than 10 microns

PM2.5, PM2.5

Particulate matter with a diameter less than 2.5 microns which is
also referred to as fine PM

Pul

Pulverized, referring to coal

Primary energy

Raw forms of energy. Examples include crude oil, natural gas, and
coal.

PV

Photovoltaic solar power technology

Refined liquids

Liquid fuels. GCAM does not differentiate these fuels, which
include gasoline, diesel, liquid petroleum gas, kerosene, and fuel oil

RPS

Renewable Portfolio Standard

Resid

Residential sector

S02, SO2

Sulfur dioxide, and air pollutant

Solid state

Lighting technology also known as LED or light-emitting diode

178


-------
Subpeak

Portion of the load duration curve that represents the hours
between peak and intermediate electricity demand

Tech

Technology

Tg

Teragram, or le9 grams

Thous

Thousand

Ton-km

Tonne-kilometer

Transport-ALM

Transportation subsectors that include air, locomotive (aka rail),
and marine (aka shipping)

Transport-LDV

Light-duty transportation subsectors used foronroad passenger
travel, including "2W&3W", "Car", and "Large Car and Truck"

Transport-HDV

Heavy-duty transportation subsectors used foronroad freight,
including "Heavy", "Medium", and "Light" trucks

Transport-Onroad

All onroad transportation, including passenger and freight vehicles

Trn

Transportation

VOC

Volatile Organic Carbon - air pollutants and precursors to
tropospheric ozone

179


-------
APPENDIX: TUTORIALS

T-l


-------
TUTORIAL 1: RUNNING GCAM THROUGH GLIMPSE

Tl.l Overview

The purpose of this tutorial is to walk new GLIMPSE users through the steps of starting
GLIMPSE, executing GCAM-USA for the GLIMPSEvl-Reference scenario, and monitoring
progress. In Tutorial 2, users will evaluate the results of the simulation.

The reference material in Chapter 6 of this Users' Guide may be useful as you carrying out the
steps of the Tutorial. Chapter 6 includes descriptions of GLIMPSE'S buttons and menu options,
brief descriptions of the types of Scenario Components that can be created and the options for
each, descriptions of the scenario components that are included in the Component Library,
descriptions of commonly-used queries, information for troubleshooting, and a Glossary.

T1.2 Opening the GLIMPSE software

Open Windows Explorer. In the "GLIMPSE-5p4" folder, double click "run_GLIMPSE_ORD-GCAM-
USA-5p4.bat".

P 1 G P 9 C ~ | GLIMPSE-vl

File	Home Share View	V ?

<- -> v /f' R Data (E:) > Projects > GLIMPSE-v1

P amazon-correto-8.372.07.1-windows-x64-jre	P Contrib

P Docs	P GCAM-Model

P GLIMPSE-Data	P GLIMPSE-GUI

P GLIMPSE-Modellnterface	B options_GCAM-USA-5p4.txt
El run_GLIMPSE_GCAM-USA-5p4.bat

9 items I

Figure Tl.l Contents of the main GLIMPSE folder, GLIMPSE-vl.

T-2


-------
This will start the Scenario Builder, using the options specified in the file "options_GCAM-USA-
5p4.txt".

A Cmd.exe window will open for your GLIMPSE session. Diagnostic information about your
GLIMPSE setup and computer's resources are output to this window.

fcT] C:\WINDOWS\system32\cmd.exe	— ~ X

E:\Projects\GLIMPSE-vl>java -jar .\GLIMPSE-GUI\exe\ScenarioBuilder.jar -options options_GCAM-USA-5p4.txt
Loading settings and initializing.

Loading options from options_GCAM-USA-5p4.txt
No match for gcameexecutableargs

	 Analysis of GLIMPSE setup 	

No problems found with parameters or folders.

No problem was found with nested GLIMPSE folders.

Your 3AVA_HOME folder, E:\Projects\GLIMPSE-vl\\amazon-correto-8.372.07.1-windows-x64-jre, was successfully found.
Installation at location E:\Projects\GLIMPSE-vl appears to be succesful.

	 Check to verify that key files exist as specified 	

XML header file: E:\Projects\GLIMPSE-vl\GLIMPSE-GUI\templates\glimpseXMLHeaders.txt - true
Tech Bound file: E:\Projects\GLIMPSE-vl\GLIMPSE-GUI\templates\tech_bnd_list_usa_5p4.txt - true

Configuration template file: E:\Projects\GLIMPSE-vl\GLIMPSE-GUI\templates\configuration_usa_5p4_template.xml - true
Query file: E:\Projects\GLIMPSE-vl\GLIMPSE-ModelInterface\exe\Main_queries_GLIMPSE-5p4.xml - true
GCAM executable: E:\Projects\GLIMPSE-vl\GCAM-Model\gcam-v5.4\exe\gcam.exe - true

Modellnterface executable: E:\Projects\GLIMPSE-vl\GLIMPSE-ModelInterface\exe\ORDModelInterface.jar - true

	 Computer Information 	

-- Memory analysis —

Total physical memory: 31.9 GB

Free physical memory: 19.6 GB

-- Disk space analysis --
Total space: 4657.4 GB
Free space: 3001.7 GB
Total swap space: 63.9 GB
Free swap space: 45.6 GB

-- Processor analysis --
Available processor cores: 8
Current usage: 0.0%

Starting GLIMPSE Graphical User Interface...
time now=06/12/2023 13:40:44

Getting lastModifiedDate for E:\Projects\GLIMPSE-vl\GLIMPSE-GUI\exe\ScenarioBuilder.jar: 23.06.11

Getting lastModifiedDate for E:\Projects\GLIMPSE-vl\GLIMPSE-ModelInterface\exe\ORDModelInterface.jar: 23.06.01

Getting lastModifiedDate for E:\Projects\GLIMPSE-vl\GCAM-Model\gcam-v5.4\exe\gcam.exe: 23.02.08

v

Figure T1.2 Diagnostic information displayed to the GLIMPSE session's Cmd.exe window.

Several seconds later, the GLIMPSE Scenario Builder graphical user interface will appear.

T-3


-------
Q GLIMPSE Scenario Builder
File Tools View Help

~

Component Library Search:

Component Name

Calib-2025lndCoal-GHGPIanStates.csv

Calib-biomass_constraints.txt

Cali b-coa l_egu_2020.txt

Calib-HDV-BEV-SW-1x50.csv

Calib-LDV-EV-AE02020.csv

Calib-Lower lncome_Elasticity_Tran.txt

+ b x O

C alih-l

< (	

alt Ki^f.iol C\A/»-

Created

2023-02-08: 08:45
2023-02-08: 08:45
2023-02-08: 08:45
2023-02-08: 08:45
2023-02-08: 08:45
2023-02-08: 08:45

Create Scenario

Component Name

No content in table

Scenario Library	Search:

Scenario Name

GCAM5p4-Ref-Orig
GLIMPSEv! -Reference

^ E

Created

2023-04-24: 12:20
2023-04-27: 11:52

~ ~~

Completed

2023-05-01:08:48

= H ID i) H O

Status

Success

Runtime

0 hr 51 min

Resources... CPU: 29% | RAM: 31.9GB Free:70% | Disk: 2,761.7GB available | Swap: 63.9GB Free: 73% // Database: database Size: 2.1GB Used: 5.4%

Figure T1.3 The Scenario Builder.

The Scenario Builder consists of three panes. In the top-left pane is the scenario "Component
Library," which contains a list of scenario components that have been previously created. A
scenario can be created by adding one or more components from the library to the "Create
Scenario" pane on the top right. Scenarios that have been created are listed in the Scenario
Library table at the bottom.

The Scenario Library table lists the date and time that the scenario was created, the date and
time that its execution was completed, the status of the run, the modeled time periods that
experienced solution errors (if any), and the total runtime for any runs that have completed.
Options for status include blank (indicating the run has not been started), "In queue" to be run,
"Running", "Success", "DNF" (an abbreviation for "Did Not Finish"), and "Unsolved mkts".

The GLIMPSE-5p4 distribution includes two scenarios. The first is "GCAM5p4-Ref-Orig". This is
the reference scenario that is distributed with GCAM by PNNI. The Scenario Builder reports that
"GCAM5p4-Ref" was created on April 24th, 2023, execution was completed successfully later
that day, and that execution required 51 minutes.

T-4


-------
The second scenario included with this distribution is GUMPSEvl-Reference, which includes
several updates specific to GLIMPSE. Note that the GLIMPSE Reference Scenario may change
over time, including with new distributions. This scenario was created on April 27th, 2023, but
the empty "completed" field indicates that it has not yet been run.

You can see which components were included in each scenario by clicking on the scenario, then
hovering the mouse over the table. A "tooltip" appears, listing the scenario's components.
"GCAM5p4-Ref-Orig" includes one scenario component: "GCAM-Updates-postRelease.txt". This
component includes updates to transportation technologies, non-CCh GHGs, and air pollutants,
none of which were included in the public release of GCAM-USA 5.4.

Note if you select one scenario, then hover the mouse over another scenario, the components
that are shown will reflect the selected scenario.

A scenario's configuration file indicates options such as which input files are included, the name
of the database where outputs will be sent, how many model periods to execute, and whether
to generate a detailed "debug" file. To view the configuration file associated with a scenario,
double-click on the scenario's name in the Scenario Library. For example, double-clicking on
"GLIMPSE-Reference" displays the following file, using the text editor specified in your options

file. Alternatively, you can select a scenario, then click on the "Edit" button, , which will
open the configuration file(s) associated with the selected scenario(s).

T-5


-------
tl E:\Projects\GLIMPSE-v1\GLIMPSE-Data\GCAM-USA\ScenarioFolders\GLIMPSEv1-Reference\configuration_GLIMPSEv1-Reference.xml... — ~ X
File Edit Search View Encoding Language Settings Tools Macro Run Plugins Window ?	+ ~ X

olSS© -oi3l Jr %CiI 9 CI ft *b| -t	l PS El ui ®|S 0 IB H B|® ® * «' > * « *

i' i configuration_GLIMPSEv1-Reference.xml £3

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40



FjcConf iguration>

##################### Scenario Meta Data #####################

Scenario name: GLIMPSEvl-Reference

Database: database

Debug region: CT

Stop year:2050

Comments:

Components:

GCAM-Updates-postRelease.txt
Calib-biomass_constraints.txt
Calib-coal_egu_2020.txt
Calib-Lo_alt_biofuel_SW.csv
Calib-2025IndCoal-GHGPlanStates.csv
Calib-NoNewCoalEGUsInOCoalStates.csv
Calib-OnroadTrn-SW-Ref-Updates.txt
Calib-Onroad-LifetimeAdj-fromMOVES.txt
Calib-OffshoreWind_lower-Bound_EastCoast_thru2035.csv
Calib-Transport-ALM-EF-Adj.txt
Tech-EGU-Cost-Updates-NREL-ATB2021-Mod.txt
Policy-RGGI-hold.txt

Policy—Seel77-ZEV-FreightLtTruck-202OOnly.csv
Policy-Secl77-ZEV-PassCars-20200nly.csv
Policy-Secl77-ZEV-PassLgCarAndTruck-20200nly.csv
Policy-LD-NTR-BEV-Sales-Car.csv
Policy-LD-NTR-BEV-Sales-LDComTruck.csv
Policy-LD-NTR-BEV-Sales-LDTruck.csv
Policy-NoBiomassEGUinMA.csv
Policy-EGU-NSPS-EFs-N0xS02PM.txt

###############################################################

	>

F-]

../input/gcamdata/xml/modeltime.xml
batch_ag.xml

CValue name="policy-target-file">../input/policy/forcing_target_4p5.xml
../input/magicc/inputs/input_gases.emk

length: 19,144 lines : 242	Ln : 1 Col: 1 Pos: 1	Windows (CR LF) UTF-8	IN

Figure T1.4 The configuration file for the GCAM5p4-Ref scenario.

You may want to use an XML editor when viewing or editing XML files. For example, Notepad++
color codes the XML code, including marking comments as green, tags as blue, attributes in red,
and attribute values in purple. The programs to use for opening ".txt" and ".xml" files can be
specified in GLIMPSE'S options file.

If your configuration file did not appear, your XML editor's path may be specified incorrectly in
your options file. You can correct this setting in the options file, then choose "File->Reload
Options" in the Scenario Builder menu to adopt the change.

The GCAM User's Guide provides information about the contents and sections of the
configuration file: GCAM v5.4 Documentation: GCAM User Guide (igcri.github.ioi. For scenarios
that are constructed in GLIMPSE, meta-data is added the top of the configuration file,
surrounded by XML comment indicators and The meta-data indicates the scenario

T-6


-------
name, output database, end year, and the scenario components from the Component Library
that were included in the run.

T1.3 Executing the GLIMPSE Reference Scenario

To execute the GLIMPSE Reference Scenario, select it in the Scenario Library, then press the

"play" button, t . The scenarios status will immediately change to "In queue". If the queue is
empty, the status will change to "Running" within a few seconds.

A black gcam.exe window will be displayed. If this window appears, then immediately
disappears, this is usually an indication that JAVA_HOME is not configured properly in the
options file or that the XML formatting in the configuration file is incorrect. If the window does
not disappear immediately, but disappears while reading in input files, this is an indication that
the last input file that was read had an error. Please see the Troubleshooting section if you
encounter either of these instances or other problems as the model executes.

This status window displays diagnostic information for the run, including which input files have
been imported, as well as warnings and errors. By default, there are many warnings reported to
this window that can be ignored. Critical problems typically are reported as "Errors", "Critical
Errors".

3 E:\Projects\GLIMPSE-v1\GCAM-Model\gcam-v5.4\exe\gcam.exe	— ~ X

License version 2.0 (ECL 2.0). http://www.opensource.org/licenses/ecl2.php

For further details, see: http://www.globalchange.umd.edu/models/gcam/

Running GCAM model code base version 5.4 revision gcam-v5.4

Configuration file: E:\Projects\GLIMPSE-vl\GLIMPSE-Data\GCAM-USA\ScenarioFolders\GLIMPSEvl-Reference\configuration_GLIM
PSEvl-Reference.xml
Parsing input files —

parsing ../input/gcamdata/xml/no_climate_model.xml scenario component.

Parsing ../input/gcamdata/xml/socioeconomics_gSSP2.xml scenario component.

Parsing ../input/gcamdata/xml/resources.xml scenario component.

Parsing ../input/gcamdata/xml/en_supply.xml scenario component.

Parsing ../input/gcamdata/xml/en_transformation.xml scenario component.

Parsing ../input/gcamdata/xml/electricity_water.xml scenario component.

Parsing ../input/gcamdata/xml/heat.xml scenario component.

Parsing ../input/gcamdata/xml/hydrogen.xml scenario component.

Parsing ../input/gcamdata/xml/en_distribution.xml scenario component.

Parsing ../input/gcamdata/xml/industry.xml scenario component.

Parsing ../input/gcamdata/xml/industrv incelas gssp2.xml scenario component.

Parsing ../input/gcamdata/xml/cement.xml scenario component.

Parsing ../input/gcamdata/xml/cement_incelas_gssp2.xml scenario component.

Parsing ../input/gcamdata/xml/en_Fert.xml scenario component.

Parsing ../input/gcamdata/xml/HDDCDD_constdd_no_GCM.xml scenario component.

Parsing ../input/gcamdata/xml/building det.xml scenario component.

Parsing ../input/gcamdata/xml/transportation_UCD_CORE.xml scenario component.

Parsing ../input/gcamdata/xml/Ccoef.xml scenario component.

Parsing .,/input/gcamdata/xml/Cstorage.xml scenario component.

Parsing ../input/gcamdata/xml/ag For Past bio base IRRMGMT.xml scenario component.

Figure T1.5 Diagnostic information displayed while a scenario is executing.

T-7


-------
You may see many warnings or other messages printed to this window. In most instances, these
warnings can be ignored. For example, the "Market info object" messages reflect how water
demands are integrated into the model, but do not affect model performance.

[3 \Projects\GLIMPSE-5p4\GCAM-Model\gcam-v5.4\exe\gcam.exe

4

iterations:

465

total:

1881

solved:

1663

4

iterations:

467

total:

1883

solved:

1663

4

iterations:

494

total:

1910

solved:

1692

4

iterations:

576

total:

1991

solved:

1693

4

iterations:

578

total:

1993

solved:

1693

4

iterations:

586

total:

2001

solved:

1694

4

iterations:

586

total:

2001

solved:

1694

Model solved normally. Iterations period 4: 586. Total iterations: 2001

Period 5: 2020

Market info object cannot be returned	because market elect_td_ind in USA does not exist.

Market info object cannot be returned	because market water_td_ind_C in USA does not exist.

Market info object cannot be returned	because market water_td_ind_W in USA does not exist.

Market info object cannot be returned	because market elect_td_ind in USA does not exist.

Market info object cannot be returned	because market water_td_ind_C in USA does not exist.

Market info object cannot be returned	because market water_td_ind_W in USA does not exist.

Market info object cannot be returned	because market elect_td_ind in USA does not exist.

Market info object cannot be returned	because market elect_td_ind in USA does not exist.

5

iterations:

3

total

2004

solved: 184

5

iterations:

35

total

2036

solved: 630

5

iterations:

410

total

2411

solved: 423

5

iterations:

412

total

2413

solved: 424

5

iterations:

444

total

2445

solved: 593

5

iterations:

728

total

2729

solved: 861

5

iterations:

730

total

2731

solved: 866

5

iterations:

762

total

2763

solved: 934

5

iterations:

953

total

2954

solved: 796

5

iterations:

955

total

2956

solved: 800

5

iterations:

987

total

2988

solved: 838

- ~ X

/s



V

Figure T1.6 Example of warnings that can be ignored.

Additional diagnostic information is also printed to the gcam.exe window. For each model
period, these include the number of solver iterations within that period, the total solver
iterations, and the number of solved markets at that time.

T-S


-------
9 C:\PirojKCs\GyMP5E- 5p4\GCAM Modei\gcarn v5.4\cxc\gcarm.exe



-

~

X

Starting a run. Running period 11













¦lodel run beginning.













Period fi: 1975













todel solved with last period's prices.













'eriod 1: 199«













1 iterations: 1 total; 4

solved;

436









1 Iterations: 35 total; 36

solved:

K05









1 iterations; l$e total: 191

solved:

1582









1 iterations: 192 total; 193

solved:

1S81









1 iterations: 224 total; 225

solved:

1617









i iterations: 425 total; 426

solved:

1659









1 iterations: 427 total; 423

solved:

1659









1 iterations: 415 total; 436

Solved:

1660









1 iterations: 415 total: 436

solved:

166®









"lodel solved msmiilly. Iterations period 1: 415

Total

iterations: 436







'eriod 2: 2005













I iterations: 3 total; 439

solved:

592









2 iterations: 35 total: 471

solved;

1491









2 iterations: 229 total: 664

solved:

1653









2 iterations: 231 total: 566

solved;

1653









2 iterations: 251 total: 6B6

solved;

1660









2 iterations; 251 total: 686

solved;

1666









¦iodel solved normally, Iterations period 2: 251

total

iterations; 686







?eri«i J: 2&i©













1 iterations; 3 total: 699

solved;

601







¦

3 iterations: 35 total; 721

solved:

146S







Figure Tl. 7 Additional diagnostic information.

This information can be useful in determining that the model is actively running and in gaging
the progress that it is making. The solved market number tends to increase rapidly, then
progress tapers off. Occasionally the number of solved markets will decrease as the solver
adjusts its solution approach. The solver configuration file includes a parameter that specifies
the maximum number of iterations allowed in each model time period. In the default GLIMPSE
setup, this maximum is 8,000 for most time periods. Thus, the value following "iterations" can
provide some indication of the maximum number of remaining iterations in that period. When
the iteration limit is reached but unsolved markets remain, information about those markets is
written to the status window. Please see the Users' Guide for help with interpreting
information on unsolved markets.

At the run continues, the information in the Scenario Table is periodically updated. For
example, "Running (6)" indicates that GCAM is currently running and that the model is in the
sixth period. The status bar at the bottom of the Scenario Builder window shows current
computer resources and utilization. For many of these metrics there are thresholds beyond
which problems may occur. When specific thresholds have been exceeded, the status message
is concatenated with "I!!". Additionally, this information is saved to a log file that can be
accessed by "View->Resource Logs->Current Session". For each log entry there is a time stamp

T-9


-------
and name of the currently executing scenario. This information can be useful in debugging
execution problems that are caused by computer resource limitations.

See the chapter on "Advanced Concepts and Operations" for information about addressing
unsolved markets. Also, the GCAM documentation includes a detailed description of the solver
and its parameterization (https://iecri.eithub.io/gcam-doc/solver.html) as well as debugging
(https://iecri.eithub.io/gcam-doc/dev-euide/debue.html).

Most of the information written to the status window is also saved to the "main_log.txt" file,
which is in the "GLIMPSE-5p4\GCAM-Model\gcam-5p4\exe\logs" folder. You can access this log

file during the run by clicking on the "EXE LOG" button,

When the run completes, several sequential steps occur:

•	The results are written to the output database. Note that if the Mode/Interface is currently
open and viewing the output database, then this step cannot proceed. A message appears
in the gcam.exe window asking that the Model I interface be closed. Once it is closed, the
process of writing results to the database begins.

•	The gcam.exe status window disappears.

•	Several files are moved to the scenario's folder, including the main_log.txt file and other
files that are specified via the "gCamOutputToSave" option in the options file. The saved
files may include the following:

o calibration_log.txt - diagnostic information from GCAM's calibration process in
which the model determines shareweights based on real-world, calibration-year
data

o debug.xml - detailed information provided for a single region or state

o solverjog.csv - information on market prices and solution status at each iteration of
GCAM's solver

•	The main_log.txt file is parsed by GLIMPSE to ascertain whether the run completed
successfully and whether there were any solution periods with unsolved markets.

•	The scenario's status is updated in the Scenario Library to "Success", "DNF" (Did Not
Finish), or "Unsolved mkts" (Unsolved Markets).

•	If there were periods with unsolved markets, those are listed in the "ProbMkts" (Problem
Markets) column.

•	The scenario's runtime is listed in the table.

T-10


-------
B GLIMPSE Scenario Builder
File Tools View Help

~

Component Library Search:

~4~

o

Component Name

Created



Calib-2025lndCoal-GHGPIanStates.csv

2023-02-08: 08:45



Calib-biomass_constraints.txt

2023-02-08: 08:45



Ca li b-coa l_egu_2020.txt

2023-02-08: 08:45



Calib-HDV-BEV-SW-1x50.csv

2023-02-08: 08:45



Calib-LDV-EV-AE02020.csv

2023-02-08: 08:45



Calib-Lower_lncome_Elasticity_Tran.txt

2023-02-08: 08:45



r3|;h_l r>. afrJaioftn*! CW

nRvK



< 	



>

Create Scenario

Component Name

No content in table

Scenario Library Search:

7 ^



[ D* j X am ^

= H ^ H

lib Esi lO

Scenario Name

Created



Completed

Status ProbMkts

Runtime

GCAM5p4-Ref-Orig

2023-04-24: 12:20



2023-05-01: 08:48

Success

0 hr 51 min

GLIMPSEvI -Reference

2023-05-01:09:20



2023-05-01:11:04

Success

1 hr41 min



Resources... CPU: 29% | RAM: 31.9GB Free:50% | Disk: 2,758.6GB available | Swap: 63.9GB Free: 61% // Database: database Size: 4.3GB Used: 10.8%

Figure T1.8 Scenario Builder indicating successful completion of GUMPSE-Reference.

Runtime for GUMPSE-Reference can vary significantly from one computer to another,
depending on processing speed, available RAM, disk speed, and the types of policies included in
the run. You should expect this run to require 1 to 5 hours. For the mid-range desktop used in
constructing this tutorial (vintage 2018 with 32 GB of RAM), execution of GLIMPSEvl-Reference
required 1 hour 41 minutes.

T1.4 Examining information saved with each run

To access the folder associated with a scenario, click on the scenario's name in the Scenario

Library, then press the open folder button above the table, ! s In addition to the saved files
listed in the previous section, the scenario folder also includes the scenario's configuration file
and input files that were developed in the creation of the scenario.

For GUMPSE-Reference, the contents of the scenario folder are shown below.

T-ll


-------
^ 1 I I C* » | GLIMPSEvI-Reference





~ X

File Home Share View





v %

<- -> v P GCAM-USA > ScenarioFolders > GLIMPSEvI-Reference

v O

P Search GLIMPSEvl-Reference

Name

Date modified

Type

Size

I Calib-2025lndCoal-GHGPIanStates.xml

4/24/2023 4:32 PM

XML Document

705 KB

¦ Calib-LoaltbiofuelJSW.xml

4/24/2023 4:32 PM

XML Document

3,291 KB

I Calib-NoNewCoalEGUslnOCoalStates.xml

4/24/2023 4:32 PM

XML Document

336 KB

I Calib-NoOnroadCNGFreight.xml

4/24/2023 4:32 PM

XML Document

1,075 KB

I Calib-OffshoreWindJower-Bound EastCoast th...

4/24/2023 4:32 PM

XML Document

27 KB

I calibration_log.txt

4/24/2023 6:15 PM

Text Document

52,860 KB

B configuration_GLIMPSEv1 -Reference.xml

4/24/2023 4:32 PM

XML Document

19 KB

I debug.xml

4/24/2023 6:22 PM

XML Document

70,359 KB

I mainJog.txt

4/24/2023 6:27 PM

Text Document

21,201 KB

¦ Policy-LD-NTR-BEV-Sales-Car.xml

4/24/2023 4:32 PM

XML Document

726 KB

H Policy-LD-NTR-BEV-Sales-LDComTruck.xml

4/24/2023 4:32 PM

XML Document

944 KB

¦ Policy-LD-NTR-BEV-Sales-LDTruck.xml

4/24/2023 4:32 PM

XML Document

724 KB

| Policy-NoBiomassEGUinMA.xml

4/24/2023 4:32 PM

XML Document

212 KB

¦ Policy-Sec177-ZEV-FreightLtTruck-20200nly.xml

4/24/2023 4:32 PM

XML Document

48 KB

I Policy-Sec177-ZEV-PassCars-20200nly.xml

4/24/2023 4:32 PM

XML Document

48 KB

¦ Policy-Sec177-ZEV-PassLgCarAndTruck-20200n...

4/24/2023 4:32 PM

XML Document

45 KB

I solverjog.csv

4/24/2023 6:21 PM

Microsoft Excel Com..

373,394 KB

17 items I





8 a

Figure T1.9 Files saved to the scenario's folder.

As a shortcut to access a scenario's mainJog.txt file, you can click on the scenario's name in the

Scenario Library, then click on the LOG button, 1 . The arrow on the button indicates that the
main_log.txt file(s) for the selected scenario(s) will be opened in a text editor.

Users generally will not need to visit a scenario's folder. However, the "debug.xml" and
"solverjog.csv" files may be useful to advanced users in diagnosing problems with GCAM
execution. Users can modify which outputs are saved with each run by changing the
"gCamOutputToSave" settings in the GLIMPSE options file. Alternatively, when a scenario is

created via the create scenario button, © , the user has the opportunity to override these
settings.

Scenario results are saved in an output database, which is accessed via the Mode I Interface.

Part 2 of the Tutorial discusses how to examine model results.

T-12


-------
TUTORIAL 2: EXAMINING MODEL RESULTS

T2.1 Overview

The Model I interface is used to access results by performing queries on the GCAM-USA database.
Results are provided in a table, and additional tools are available for filtering, sorting, and
visualizing the data.

T2.2 Viewing model results with the Modellnterface

To view GCAM results, you will need to open the output database via the Modellnterface. There
are two Results buttons available. The "Results" button,

Ion

, will open the Modellnterface,

initialized to the output database specified in the options file. Alternatively, press ^ to
initialize the Modellnterface to the database referenced in the selected scenario.

For this tutorial, select the scenario "GLIMPSE-Reference" in the Scenario Builder's Scenario
Library. Then click the results button with the arrow, ^ .

After several seconds, the Modellnterface window will appear. You should see two scenarios in
the "Scenarios" pane, "GCAM5p4-Ref-Orig", which was included with the GLIMPSE installation,
and the new scenario, GLIMPSEvl-Reference. The name for each is followed by the date and
time that it was loaded into the output database.

T-13


-------
ddd GLIMPSE Model Interface [E:\Projects\GLIMPSE-5p4\GCAM-Model\gcam-v5.4\output\database]
File Edit Table Help

~

X

Scenario

Regions

GCAM5p4-Ref-Orig 2023-1

USA 1

IdllilMliUliimMAlikJi

Africa_Eastern
Africa_Northern





Africa_Southern



Africa_Western



Australia_NZ



Brazil



Canada



Central America and Cari



Central Asia



China



EU-12



EU-15 v

< >

< >

Queries

~ GLIMPSE
B Q Primary and final energy

•J~ - Primary energy consumption by region (direct eguivalent)
i - Final energy consumption by region

Final energy consumption by aggregate sector
••••"" - Final energy consumption by aggregate sector and fuel
~ - Final energy consumption by aggregate sector and fuel(v2)
L~- Final energy consumption by sector and fuel
Energy inputs by sector
~ - Electricity use by aggregate sector
r - Coal use by aggregate sector

- Natural gas use by aggregate sector
~ - Refined liquids use by aggregate sector
r - Biomass use by aggregate sector
r - Hydrogen use by aggregate sector
'"-End-use energy consumption in buildings
~ - End-use energy consumption in buildings (detail)

El-

Run Query Diff Query | [Total Collapse Update Single Queries Create Remove Edit

Figure T2.1 The Modellnterface.

In the "Regions" pane, all the socio-economic regions included in GCAM-USA are listed, as well
as the model's electricity transmission grid regions within the U.S.

Note that most of the energy system-related activities within the U.S. have been disaggregated
to the state level. Thus, the "USA" region does not provide national totals. Instead, "USA"
includes several sectors that have not been apportioned to states, including agriculture, coal
mining, oil and gas operations, and hydrogen production.

The "Queries" pane lists the various outputs that can be extracted from the output database.
Currently, several hundred queries are available. The queries are organized in a tree structure.
The "GLIMPSE" set of queries includes those that we anticipate will be of particular interest to
many GLIMPSE users. These queries are grouped into sub-categories, including "Primary and
final energy", "Technologies", "Emissions", "Impacts", "Markets, prices, and costs", "Inputs and
outputs", "Assumptions", and "Water supply and demand". For the queries under "GLIMPSE",
hovering the mouse over a query will pop up a tooltip with a description of the query. Several
queries are specific to either GCAM or GCAM-USA, as indicated in their title. Different versions

T-14


-------
of these queries are necessary because of the naming and structural differences between the
two versions of the model.

Additional queries follow, grouped into the category "Standard 5p4 queries." These are the
queries that are distributed with GCAM by PNNL.

For this tutorial, we will use several commonly examined outputs, including:

•	"CO2 emissions by region"

•	"Electricity generation by aggregate subsector rnw detail (GCAM-USA)", where "rnw
detail" indicates that this query returns more detailed information about renewables
than "Electricity generation by aggregate subsector (GCAM-USA)"

In the "Scenario" pane, select "GLIMPSE-Reference".

Next, in the "Regions" pane, we want to select all the states and the "USA" region. An easy way
to do this is with the following steps:

•	Scroll down to "WY" and click select it.

•	Scroll up to "AK" and shift-click select it.

•	Scroll up to the top and control-click on "USA".

In the "Queries" list, scroll down and select the query: Queries->GLIMPSE->C02 emissions by
region

Next, click on the "Run Query" button.

A tab will appear at the bottom of the Mode/Interface, showing the message "Waiting for query
to complete. Close to terminate." After a short period of time, the tab will be populated by a
table that shows the query results for each selected scenario and region.

T-15


-------
nm GLIMPSE Modellnterface [E:\Projects\GLIMPSE-5p4\GCAM-Model\gcam-v5.4\output\database]	~ ~ X

File Edit Table Help

Scenario

Regions

Queries

GCAM5p4-Ref-Orig 2023-lj

South Korea a

h ^ Refined liquids production by tech (core) A
- Hydroqen production by tech
Transport final energy by tech and fuel
Jransport service output by tech
^Transport service output by tech (new capacity)

*	Passenqer car and truck service output by tech

•	Freiqht truck service output by tech (no bus)

GLIMPSEvl-Reference 2023

Southeast Asia



Taiwan

Argentina

Colombia



AK
AL
AR

1

CO
CT

| B Emissions

-	C02 emissions by aggregate sector
• C02 emissions by sector

-	C02 by resource production

-	C02 emissions by sector (no bio)

-	C02 emissions by tech
N0x,S02,PM2.5 by region
N0x,S02,PM2.5 by aggregate sector
N0x,S02,PM2.5 by sector

!••••" - N0x,S02,PM2.5 by tech v

All • r L. .



DC

< >

1 < >

Run Query Diff Query [	| Total Collapse II Update Single Queries Create Remove Edit

O C02 emissions by region
Filter Graph Format

scenario

region

2015

2020

2025

2030

2035

2040

2045

2050

Units

GLIMPSEvl-Reference...

AK

8.89

7.91

7.59

7.47

7.39

7.46

7.46

7.43

MTC

GLIMPSEvl-Reference...

AL

32.3

27.9

26.7

25.9

25.0

23.7

21.9

21.9

MTC

GLIMPSEvl-Reference...

AR

15.8

13.2

12.6

12.2

11.9

11.4

10.8

10.2

MTC

GLIMPSEvl-Reference...

AZ

24.9

20.2

19.7

19.4

19.0

18.3

17.3

17.6

MTC

GLIMPSEvl-Reference...

CA

102

100

96.5

92.3

89.0

87.5

£6.4

84.5

MTC

GLIMPSEvl-Reference...

CO

24.1

20.5

20.0

19.1

18.6

18.2

17.5

16.9

MTC

GLIMPSEvl-Reference...

CT

9.01

8.20

7.36

6.64

6.20

5.98

5.76

5.75

MTC

GLIMPSEvl-Reference...

DC

0.726

0.697

0.652

0.618

0.593

0.588

0.578

0.587

MTC

GLIMPSEvl-Reference...

DE

3.37

3.07

2.78

2.53

2.37

2.26

2.15

2.05

MTC

GLIMPSEvl-Reference...

FL

66.7

62.4

60.4

58.0

56.4

55.8

54.9

54.3

MTC

GLIM PSEv 1 - Ref eren ce...

GA

37.5

33.1

31.4

29.8

28.6

28.2

27.7

26.7

MTC

GLIMPSEvl-Reference...

HI

5.33

4.64

4.17

3.79

3.53

3.39

3.25

3.08

MTC

GLIMPSEvl-Reference...

IA

13.5

9.64

9.41

9.36

9.82

8.51

6.60

4.89

MTC

GLIMPSEvl-Reference...

ID

4.57

4.71

4.58

4.36

4.11

3.95

3.79

3.68

MTC

GLIMPSEvl-Reference...

IL

55.8

48.0

45.7

37.6

35.3

33.6

32.9

33.7

MTC

GLIMPSEvl-Reference...

IN

48.5

39.5

38.2

31.0

25.6

25.2

24.6

24.7

MTC

GLIMPSEvl-Reference...

KS

16.5

13.3

12.9

12.5

12.2

11.7

11.1

10.0

MTC

GLIMPSEvl-Reference...

KY

35.5

27.0

25.9

25.1

23.7

22.9

21.6

21.2

MTC

GLIM PSEv i - Reference...

LA

54.9

51.2

49.3

48.8

48.5

49.0

49.0

48.9

MTC

GLIMPSEvl-Reference...

MA

16.4

14.9

13.8

12.6

11.8

11.4

11.0

10.8

MTC

GLIMPSEvl-Reference...

MD

16.1

13.2

12.3

11.4

10.8

10.5

10.2

9.94

MTC

GLIMPSEvl-Reference...

ME

4.10

3.78

3.47

3.17

2.92

2.75

2.59

2.51

MTC

Figure T2.2 Viewing the results of a query at the regional level.

The Modellnterface also makes it very easy to develop regional totals. To do this, click on the
"Total" button in the center for the graphical interface, then re-run the query by clicking on
"Run Query". The values that are reported in the table now represent the sum across all
selected regions.

T-16


-------
nm GLIMPSE Modellnterface [E:\Projects\GLIMPSE-5p4\GCAM-Model\gcam-v5.4\output\database]	~ ~ X

File Edit Table Help

Scenario

Regions

Queries

GCAM5p4-Ref-Orig 2023-lj

South Korea a



Refined liquids production by tech (core) A

-	Hydroqen production by tech
Transport final energy by tech and fuel
Jransport service output by tech

Transport service output by tech (new capacity)

-	Passenqer car and truck service output by tech
• Freiqht truck service output by tech (no bus)

GLIMPSEvl-Reference 2023

Southeast Asia





Taiwan

Argentina

Colombia





AK
AL
AR

1

CO
CT



EE} Emissions

-	C02 emissions by aggregate sector

-	C02 emissions by sector

-	C02 by resource production
C02 emissions by sector (no bio)

-	C02 emissions by tech
N0x,S02,PM2.5 by region
N0x,S02,PM2.5 by aggregate sector
N0x,S02,PM2.5 by sector

-	N0x,S02,PM2.5 by tech v



°C



< >

[<

Run Query Diff Query 0 Total Collapse Update Single Queries Create Remove Edit

~ C02 emissions by region O C02 emissions by region
Filter Graph Format

scenario region 2015 2020 2025 2030 2035 2040 2045 2050 Units
GUMPSEvl-... [Total	11420	11240	11190	11120	11080	11070	11040	11030	IMTC

Figure T2.3 Viewing the results of a query, totaled across selected regions.

T2.3 Analyzing model results outside of the Modellnterface

There are several ways to analyze the results of the queries. One is to move the data to a
spreadsheet. If you have Excel open, you can drag the label of the tab associated with a query
over to an Excel worksheet and drop the data there. This approach moves the table headings
along with the data.

For large datasets (e.g., with several hundred rows or more), this approach can be very overly
memory intensive. An alternative is to select rows in your table (multi-selecting with shift-click
or control-click), then pressing ctrl-c to copy the data. You can then paste the data into the
spreadsheet by choosing a location and pressing ctrl-v. This second approach only pastes the
contents of the table; it does not paste the headings.

The Modellnterface offers a third mechanism for exporting query results. If you have already
executed the queries of interest, one or more tables of data will be shown on tabs. Choose
"File->Export tabs as CSVs" from the main Modellnterface menu bar. A file browser will appear,

T-17


-------
and you can select the folder for where you would like to place the query results. The data on
each tab, including the column headers, are then written to that folder as CSV files, with the
filenames base on the tab names.

An issue can arise with these approaches if you have the "Text to Columns" feature in your
spreadsheet. The scenario name has a comma embedded in the cell, and this cell can
sometimes be displayed over two cells. Turning off "Text to Columns" can alleviate this
problem.

For practice, use each of the approaches described above to transfer the query results to an
Excel workbook.

T2.4 Using the Mode/Interface to visualize model results

The Model I interface includes graphical capabilities that are intended to support interactive,
exploratory data analysis - allowing the user to quickly visualize and iteratively explore the
model results. In this section, we use these tools to understand how electricity production
differs between the GCAM5p4-Ref-Orig scenario and GLIMPSEvl-Reference.

Run the query "Electricity generation by aggregate subsector rnw detail (GCAM-USA)". When
the query is complete, your Model I interface should report resulting electricity production
values.

T-18


-------
nm GLIMPSE Modellnterface [E:\Projects\GLIMPSE-5p4\GCAM-Model\gcam-v5.4\output\database]	~ ~ X

File Edit Table Help

Scenario

Regions

Queries

GCAM5p4-Ref-Orig 2023-lj

South Korea a





- End-use energy consumption in industry (detail) A
Energy inputs to electricity production (GCAM-USA)

Energy inputs to electricity production (GCAM-USA) (detail)

Energy inputs to refining activities (GCAM-USA)

Technologies

* Electricity generation by region (no cogen)(GCAM-USA)

Electricity generation by gen and cooling tech (incl cogen)(GCAM-USA)

GLIMPSEvl-Reference 2023

Southeast Asia







Taiwan

Argentina

Colombia



i 9-i



AK
AL
AR

1

CO
CT





Electricity generation by aggregated subsector (GCAM-USA)
j i i N >3.1 tHI 1 1 IjeiW ^

-	Electricity generation by subsector (GCAM-USA)

Electricity generation by tech and cooling (incl cogen)(GCAM-USA)

Electricity generation by tech and cooling (incl cogen)(GCAM-USA)(new capacity)

Electricity generation by cogen only (GCAM-USA)

Electricity generation input by subsector (GCAM-USA)

Electricity generation by region (incl cogen)(core)

-	Electricity generation by subsector (core)

Electricity generation by gen tech (core)

« Building final energy by tech v



DC





< >

1 < >

Run Query Diff Query 0 Total Collapse Update Single Queries Create Remove Edit

~ C02 emissions by region O C02 emissions by region O Electricity generation by aggregated subsector rnw detail(GCAM-USA)
Filter Graph Format

scenario

region

subsector

2015

2020

2025

2030

2035

2040

2045

2050

Units

GLIMPSEvl...

Total

biomass

0.240

0.207

0.170

0.166

0.178

0.188

0.207

0.231

EJ

GUMPSEvl...

Total

coal

5.30

2.69

2.64

2.08

1.79

1.64

1.42

1.13

~

GLIMPSEvl...

Total

gas

4.99

6.52

6.35

6.76

7.27

7.79

8.45

9.45

EJ

GLIMPSEvl...

Total

geo

0.0674

0.0907

0.117

0.132

0.145

0.161

0.130

0.136

EJ

GLIMPSEvl...

Total

hydro

0.910

1.06

1.06

1.06

1.06

1.06

1.05

1.05

EJ

GLIMPSEvl...

Total

hvdroqen

0.00

0.000389

0.000199

0.000421

0.000641

0.00113

0.00160

0.00255

EJ

GLIMPSEvl...

Total

refined liqu...

0.139

0.0666

0.0417

0.0359

0.0341

0.0330

0.0322

0.0317

EJ

GLIMPSEvl...

Total

solar PV

0.116

0.752

1.67

2.44

2.99

3.48

4.02

4.30

EJ

GLIMPSEvl...

Total

solar roofto...

0.00

0.0268

0.0871

0.159

0.169

0.208

0.237

0.287

~

GLIMPSEvl...

Total

wind offshore

0.00

0.00

0.171

0.377

0.747

0.871

1.03

1.09

EJ

GLIMPSEvl...

Total

wind onshore

0.695

1.57

2.37

3.08

3.68

4.25

4.58

5.07

EJ

GLIMPSEvl...

Total

nuclear

2.99

2.92

2.79

2.71

2.29

1.83

1.44

1.05

EJ

GLIMPSEvl...

T otal

solar CSP

0.0128

0.0628

0.160

0.277

0.390

0.535

0.758

0.927

EJ



Figure T2.4 Examining electricity production by subsector.

Click the "Graph" button above the table. A simple line chart will appear to the right of the
table.

T-19


-------
nm GLIMPSE Modellnterface [E:\Projects\GLIMPSE-5p4\GCAM-Model\gcam-v5.4\output\database]	~ ~ X

File Edit Table Help

Scenario

Regions

Queries

GCAM5p4-Ref-Orig 2023-lj

South Korea

A







- End-use energy consumption in industry (detail) A
Energy inputs to electricity production (GCAM-USA)

Energy inputs to electricity production (GCAM-USA) (detail)

Energy inputs to refining activities (GCAM-USA)

Technologies

* Electricity generation by region (no cogen)(GCAM-USA)

Electricity generation by gen and cooling tech (incl cogen)(GCAM-USA)

GUMPSEvl-Reference 2023

Southeast Asia











Taiwan

Argentina

Colombia





i

3



AK
AL
AR
AZ
CA
CO
CT









Electricity generation by aggregated subsector (GCAM-USA)
j i i N >3.1 tHI 1 1 IjeiW ^

-	Electricity generation by subsector (GCAM-USA)

Electricity generation by tech and cooling (incl cogen)(GCAM-USA)

Electricity generation by tech and cooling (incl cogen)(GCAM-USA)(new capacity)

Electricity generation by cogen only (GCAM-USA)

Electricity generation input by subsector (GCAM-USA)

Electricity generation by region (incl cogen)(core)

-	Electricity generation by subsector (core)

Electricity generation by gen tech (core)

« Building final energy by tech v



DC

V







< >

1 < >



Run Query Diff Query 0 Total Collapse Update Single Queries Create Remove Edit

~ C02 emissions by region O C02 emissions by region O Electricity generation by aggregated subsector rnw detail(GCAM-USA)

Filter Graph Format



More Display 1 J LineChart V I Is





scenario
GUMP...

region
Total

subse...
biomass

2015
0.240

2020
0.207

2025
0.170

2030
0.166

2035
0.178

2040
0.188

2045
0.207

2050
0.231

Units
~

GLIMPSE v1 Reference
region: Total

10
0
8

7

a a
i:

< >

GUMP...

Total

coal

5.30

2.69

2.64

2.08

1.79

1.64

1.42

1.13

~

GUMP...

Total

gas

4.99

6.52

6.35

6.76

7.27

7.79

8.45

9.45

~

GLIMP...

Total

geo

0.0674

0.0907

0.117

0.132

0.145

0.161

0.130

0.136

~

GUMP...

Total

hydro

0.910

1.06

1.06

1.06

1.06

1.06

1.05

1.05

~

GLIMP...

Total

hvdroq...

0.00

0.0003...

0.0001...

0.0004...

0.0006...

0.00113

0.00160

0.00255

~

GUMP...

Total

refined...

0.139

0.0666

0.0417

0.0359

0.0341

0.0330

0.0322

0.0317

~

GUMP...

Total

solar PV

0.116

0.752

1.67

2.44

2.99

3.48

4.02

4.30

~

GLIMP...

Total

solar r...

0.00

0.0268

0.0871

0.159

0.169

0.208

0.237

0.287

~

GUMP...

Total

wind o...

0.00

0.00

0.171

0.377

0.747

0.871

1.03

1.09

~

GUMP...

Total

wind o...

0.695

1.57

2.37

3.08

3.68

4.25

4.58

5.07

~

GLIMP...

Total

nuclear

2.99

2.92

2.79

2.71

2.29

1.83

1.44

1.05

~

GUMP...

Total

solar C...

0.0128

0.0628

0.160

0.277

0.390

0.535

0.758

0.927

~



Figure T2.5. Generating a thumbnail graphic.

This image is a thumbnail. Clicking on it will display a larger version with a legend. Right-clicking
on this larger image and choosing "Copy" will add a copy of the image to your computer's
clipboard, allowing it to be pasted into other documents. Note that this option may be

problematic on Corretto's versions of Java and may cause the Modellnterface to freeze.
Alternatively, you can right-click and choose "Save As..." to save the image as a PNG file or use
Windows screen capture tools to copy the image.

T-20


-------
^0_ Electricitygenerationbyaggregatedsubsectorrnwdetail(GCAM-USA)	X

Electricity generation hy aggregated subsector rnw detail(GCAMUSA)
GLIMPSEvI -Reference
region: Total

10
9



output (EJ)

O-^NJCJ-tlUtCCi-JCO





© solar CSP

¦	nuclear
wind onshore
wind offshore
solar rooftop PV
solar PV

¦	refined liquids

¦	hydrogen

¦	hydro

¦	geo

¦	gas

¦	coal

¦	biomass

fO

o

Ul

2050
2045
2040
2035
. 2030
j 2025
f 2020



ChartOptions



Year
biomass

2015

0.24

2020

0.207

2025

0.17

2030

0.166

2035

0.178

2040

0.188

2045

0.207

2050

0.231



A
V

coal

5.3

2.69

2.64

2.08

1.79

1.64

1.42

1.13



gas

4.99

6.52

6.35

6.76

7.27

7.79

8.45

9.45



geo

0.0S7

0.091

0.117

0.132

0.145

0.161

0.13

0.136



hydro

0.91

1.06

1.06

1.06

1.06

1.06

1.05

1.05



hydrogen

0.0

0.0

0.0

0.0

0.001

0.001

0.002

0.003



refined li...

0.139

0.067

0.042

0.036

0.034

0.033

0.032

0.032



solar PV

0.116

0.752

1.67

2.44

2.99

3.48

4.02

4.3



solar roo...

0.0

0.027

0.087

0.159

0.169

0.208

0.237

0.287



wind offs...

0.0

0.0

0.171

0.377

0.747

0.871

1.03

1.09



wind ons...

0.695

1.57

2.37

3.08

3.68

4.25

4.58

5,07



nuclear

2.99

2.92

2.79

2.71

2.29

1.83

1.44

1.05



solar CSP

0.013

0.063

0.16

0.277

0.39

0.535

0.758

0.927























Figure T2.6 Clicking on the thumbnail produces a larger version with legend.

There are a variety of options for modifying the figure, including changing the texture and
shading of the lines. These changes can be saved and are applied automatically the next time a
similar figure is plotted.

Other types of graphs can also be displayed. Above the thumbnail, change the selection in the
pulldown menu from "LineChart" to "StackedBarChart". The thumbnail is updated to reflect the
change.

T-21


-------
nm GLIMPSE Modellnterface [E:\Projects\GLIMPSE-5p4\GCAM-Model\gcam-v5.4\output\database]	~ ~ X

File Edit Table Help

Scenario

Regions

Queries

GCAM5p4-Ref-Orig 2023-lj

South Korea

A

«End-use energy consumption in industry (detail) A
- Enerqy inputs to electricity production (GCAM-USA)

Energy inputs to electricity production (GCAM-USA) (detail)

Energy inputs to refining activities (GCAM-USA)
f 9-E3 Technologies

f--~ * Electricity generation by region (no cogen)(GCAM-USA)

• Electricity generation by gen and cooling tech (incl cogen)(GCAM-USA)

GUMPSEvl-Reference 2023

Southeast Asia





Taiwan

Argentina

Colombia





AK
AL
AR
AZ
CA
CO
CT



Electricity generation by aggregated subsector (GCAM-USA)

*	Electricity generation by subsector (GCAM-USA)

Electricity generation by tech and cooling (incl cogen)(GCAM-USA)

-	Electricity generation by tech and cooling (incl cogen)(GCAM-USA)(new capacity)

-	Electricity generation by cogen only (GCAM-USA)

-	Electricity generation input by subsector (GCAM-USA)

-	Electricity generation by region (incl cogen)(core)

*	Electricity generation by subsector (core)

-	Electricity generation by gen tech (core)

j { | ^ * Bujjding final energy by tech _ v



DC

V

< >

1 < >



Run Query Diff Query M Total Collapse Update Single Queries Create Remove Edit

~ C02 emissions by region O C02 emissions by region O Electricity generation by aggregated subsector rnw detail(GCAM-USA)

Filter

Graph

Format

















scenario

region

subse...

2015

2020

2025

2030

2035

2040

2045

2050

Units



GUMP...

Total

biomass

0.240

0.207

0.170

0.166

0.178

0.188

0.207

0.231

~



GUMP...

Total

coal

5.30

2.69

2.64

2.08

1.79

1.64

1.42

1.13

~



GUMP...

Total

gas

4.99

6.52

6.35

6.76

7.27

7.79

8.45

9.45

~



GLIMP...

Total

geo

0.0674

0.0907

0.117

0.132

0.145

0.161

0.130

0.136

~



GUMP...

Total

hydro

0.910

1.06

1.06

1.06

1.06

1.06

1.05

1.05

~



GUMP...

Total

hvdroq...

0.00

0.0003...

0.0001...

0.0004...

0.0006...

0.00113

0.00160

0.00255

~



GUMP...

Total

refined...

0.139

0.0666

0.0417

0.0359

0.0341

0.0330

0.0322

0.0317

~



GUMP...

Total

solar PV

0.116

0.752

1.67

2.44

2.99

3.48

4.02

4.30

~



GLIMP...

Total

solar r...

0.00

0.0268

0.0871

0.159

0.169

0.208

0.237

0.287

~



GUMP-

Total

wind 0...

0.00

0.00

0.171

0.377

0.747

0.871

1.03

1.09

~



GUMP...

Total

wind 0...

0.695

1.57

2.37

3.08

3.68

4.25

4.58

5.07

~



GLIMP...

Total

nuclear

2.99

2.92

2.79

2.71

2.29

1.83

1.44

1.05

~



GUMP...

Total

solar C...

0.0128

0.0628

0.160

0.277

0.390

0.535

0.758

0.927

~







More Displayl1

~ s

GLIMPSEvI Reference
region: Total











til









M1

¦

1









¦

























—



l

¦

¦

¦

¦

.



I



°

8

°

s

§

&

a

Figure T2.7 Viewing query results as a stacked bar chart.

Try out some of the other graphing options.

T2.4 Comparing results across scenarios

Next, we will walk through some ways you can use the Mode/Interface to compare the results
of two scenarios.

In the Scenario pane at the top left of the Modellnterface, keep GLIMPSEvl-Reference selected,
but add "GCAM-5p4-Ref-Orig" to the selection by control-clicking on it.

in the Query pane on the top right, change your selection to "CO2 emissions by aggregate
sector".

Make sure the "Total" checkbox is checked, then click on the "Run Query" button. After several
seconds, the table will be populated with CO2 emissions data.

T-22


-------
ft GLIMPSE Modellnterface [E:\Projects\GLIMPSE-5p4\GCAM-Model\gcam-v5.4\output\database]
| File Edit Table Help	

~

Scenario

GC AM5p4-R ef-Orig 2023-1-5T0:
GLIMPSEv 1 -R eferenc e 2023-1-5

GLIMPSEv 1 -Tax- 100C-5pct 202,

Africa_Eastern

Africa_Northern

Africa_Southern

Africa_Western

Australia_NZ

Brazil

Canada

Central America and Caribbean

Central Asia

China

EU-12

EU-15

Eui ope_Eastern
Europe_Non_EU

European Free Trade Association

India

Indonesia

Japan	

Queries

- Energy inputs to refining activities (GCAM-U5A)

Technologies

a Electricity generation by region (no cogen)(GCAM-USA)

. Electricity generation by gen and cooling tech (incl cogen)(GCAM-USA)
Electricity generation by aggregated subsector (GCAM-USA)	

Electr icity generation by aggregated subsector rnw detail(GCAM-USA)

a Electricity generation by subsector (GCAM-USA)

-	Electricity generation by tech and cooling (incl cogen)(GCAM-USA)

-	Electricity generation by tech and cooling (incl cogen)(GCAM-USA)(new capacity)

-	Electricity generation by cogen only (GCAM-USA)

¦ Electricity generation input by subsector (GCAM-USA)

. Electricity generation by region (incl cogen)(core)
ss Electricity generation by subsector (core)

-	Electricity generation by gen tech (core)

-	Building final energy by tech

. Building service output by tech
. Industry final energy by tech and fuel

Run Query Diff Query £7] Total

Collapse

Update Single Queries

Q Electricity generation by aggregated subsector rnw detail(GCAM-USA)

Filter Graph Format

scenario

legion

subsector

2015

2020

2025

2030

2035

2040

2045

2050

Units

GCAM5p4-Ref,..

Total

biomass

0.240

0.202

0.197

0.205

0.218

0.229

0.248

0.276

EJ

A

GCAM5p4-Ref...

Total

coal

5.30

2.89

2.83

2.71

2.54

2.31

1.92

1.38

EJ



GCAM5p4-Ref...

Total

gas

4.99

6.31

6.63

7.28

8.07

8.93

10.1

11.7

EJ



GCAM5p4-Ref...

Total

geo

0.0674

0.0885

0.105

0.116

0.129

0.143

0.112

0.120

EJ



GCAM5p4-Ref...

Total

hydro

0.910

1.06

1.06

1.06

1.06

1.06

1.05

1.05

EJ



GCAM5p4-Ref...

Total

hydrogen

0.00

0.000389

0.000493

0.000839

0.00125

0.00190

0.00276

0.00433

EJ



GCAM5p4-Ref,..

Total

refined liquids

0.139

0.0645

0.0598

0.0514

0.0467

0.0436

0.0426

0.0429

EJ



GCAM5p4-Ref...

Total

solar PV

0.116

0.730

1.04

1.62

2.17

2.71

3.25

3.43

EJ



GCAM5p4-Ref...

Total

solar rooftop PV

0.00

0.0268

0.0557

0.103

0.117

0.124

0.132

0.160

EJ



GCAM5p4-Ref...

Total

wind offshore

0.00

0.00

0.0189

0.0730

0.164

0.286

0.448

0.621

EJ



GCAM5p4-Ref...

Total

wind onshore

0.695

1.50

1.87

2.32

2.85

3.36

3.51

3.78

EJ



GCAM5p4-Ref...

Total

nuclear

12.99

2.92

2.79

2.68

2.25

1.79

1.36

0.868

EJ



GCAM5p4-Ref...

Total

solar CSP

0.0128

0.0600

0.0851

0.128

0.173

0.224

0.306

0.347

EJ



GLIMPSEvl-Re,..

Total

biomass

0.240

0.207

0.170

0.166

0.178

0.188

0.207

0.231

EJ



GLIMPSEvl-Re...

Total

coal

5.30

2.69

2.64

2.08

1.79

1.64

1.42

1.13

EJ



GLIMPSEvl-P.e,,.

Total

oas	

4,99

6.52

6.35

6.76

7.27

7.79

8,45

9.45

EJ

V

Figure T2.8 Examining the results of two scenarios simultaneously.

Click on the "Graph" button to generate thumbnails. You may need to reposition the dividers to
see both thumbnails fully. Click on the checkbox next to "Same Scale". This will ensure that the
Y axis is shown at the same scale for all thumbnails.

T-23


-------
ft GLIMPSE Modellnterface [E:\Projects\GLIMPSE-5p4\GCAM-Model\gcam-v5.4\output\database]
| File Edit Table Help	

~

Scenario

GC AM5p4-R ef-Orig 2023-1-5T0:
GLIMPSEv 1 -R eFerenc e 2023-1-5

GLIMPSEv 1 -Tax- 100C-5pct 202,

Africa_Eastern

Africa_Northern

Africa_Southern

Africa_Western

Australia_NZ

Brazil

Canada

Central America and Caribbean

Central Asia

China

EU-12

EU-15

Europe_Eastern
Europe_Non_EU

European Free Trade Association

India

Indonesia

Japan	

Queries

- Energy inputs to refining activities (GCAM-U5A)

Technologies

a Electricity generation by region (no cogen)(GCAM-USA)

. Electricity generation by gen and cooling tech (incl cogen)(GCAM-USA)
Electricity generation by aggregated subsector (GCAM-USA)	

Electr icity generation by aggregated subsector rnw detail(GCAM-USA)

a Electricity generation by subsector (GCAM-USA)

-	Electricity generation by tech and cooling (incl cogen)(GCAM-USA)

-	Electricity generation by tech and cooling (incl cogen)(GCAM-USA)(new capacity)

-	Electricity generation by cogen only (GCAM-USA)

¦ Electricity generation input by subsector (GCAM-USA)

. Electricity generation by region (incl cogen)(core)
ss Electricity generation by subsector (core)

-	Electricity generation by gen tech (core)

-	Building final energy by tech

. Building service output by tech
. Industry final energy by tech and fuel

Run Query Diff Query £7] Total

Collapse

Update Single Queries

Q Electricity generation by aggregated subsector rnw detail(GCAM-USA)

Filter

GCA...
GCA...

Graph

sub... 2015
biomass]0.240
coal 5.30

GCA...
GCA,..
GCA...
GCA...
GCA...
GCA...
GCA...
GCA...
GCA,..
GCA...
GCA...
GUM..
GUM..
GLIM..

Total

[Total

[Total
[Total
i Total

qi?

geo

2030
0.205

2040 2045
0.229 10.248

hydro ,0.910
hydr... 0.00

11.06
0.00...

refin... 0.139 0.0645 0.0598

[Total
[Total
[Total

solar PV'0.116
solar... |0.00

(Total

[Total
Total

jTotal
[Total
Total

0.730 1.04
|o.0268 |0.0557

11.06
0.00. ¦¦
0.0514
1.62
0.103

0.143 0.112 0.120

1.06
'0.00...

1.06 1.05
0.00190 0.00...

0.00 0.0189 0.0730

nuclear
solar ...

2^99	

0.0128

biomass

coal

loas

0.240
5.30
4.99

0,0467
2.17
0.117
0.164

0.0436 0.0426

0.0600 0.0851

2.68
jO. 128

2.25

0.207 10.170 0.166

2.69
6-52

2.64
16.35

2050
|0.276

Units
EJ

11.7

1.05

0.00433

0.0429

¦71 3.25
0.124 0.132

0.286 0.448

1.79 1.36
0.224 0.306

0.188
1.64

3.43
0.160

0.868
0.347

1.42
8.45

1.13
9.45

GCAM5p4-Ref Orig
region: Total

GLIMPSEvI-Reference
region: Total

o o o

6 & S

Figure T2.9 Comparing two results scenarios graphically.

As before, you can click on each thumbnail to view a larger version. However, additional
features facilitate comparison.

Click on "More" in the green area above the thumbnails and click "Transpose".

A popup window of thumbnails appears, but this time each series is given its own plot and the
results of the two scenarios are compared.

Here, we have changed the "Display" option to "4", which changes the number of charts shown
on a row.

T-24


-------
!:&! Transpose ThumbnailsJEIectricitygenerationbyaggregatedsubsectorrnwdetaii(GCAM-USA)

biomass

coal

0.30
0.25
qj 0.20
"5 0.15

Cl

I 0.10

0.05
0 .00 '

WWWMMWWK)

oooooooo

—» K) t J U U L L Ol

oiooioaoao

lOWUUUWUW

oooooooo

-i M IO CJ U L ii Ol
OlOOiOOiOOiO

gas

geo

12
10
8
6
4
2

0 *



- ooooooo
r j to (>) u i i ui

O Ol O Ol O Ol O

LU

r 0.10

° 0.05

/

0.00

MfOMWWIOMK)

00000000
-> ro m u u •

01	O Ol O Ch

6 & :

HI

1.0
0.8

3 06

o.

| 0.4
0.2
0.0

hydro

\/~

OOOOOOOO

w r o to co Ji t oi

OlOOlOOlOiSO

liydrogen

0.004

w 0.003
"3

3- 0.002

=3
O

0.001
0.000

refined liquids

solar PV

0.15

0.05

WWWKIWWWW

oooooooo
r j fo (o co ^ it oi
OlOOiOOiOOiO

Ol O Ol O Ol

MWMMMMNW
OOOOOOOO
-itOKJUUAiOl
OlOOiOOiOOiO

wind offshore

wind onshore

nuclear

OOOOOOOO
-iK)K)UUlLOl
OlOOiOOiOOiO



3.0



2.5

—>



UJ

2.0

Q.

1.5

3

1.0

O



0.5



0.0

W M M K) M to K)

OOOOOOO
M W U (O Ji L Ol
O Ol O Ol O Ol 0

MNMMWMWW
OOOOOOOO
ho M (O CO L 1. Ol
OlOOiOOiOOiO

solar CSP

1.0
^ 0.8

T
LU

— 0.0
|o.4

O

0.2

0.0

OOOOOOOO
-LrOhJC0(O££Ol
OlOOiOOiOOiO

Figure T2.10 Using the transpose option to compare the scenario results by series.

T-25


-------
If we click on the "electricity" thumbnail, we can see that the electric production from coal is
less in GLIMPSEvl-Reference (blue) relative to GCAM5p4-Ref-Orig (red).

IM1. Electricitygenerationbyaggregatedsubsectorrnwdetail(GCAM-USA)

5.5
5.0
4.5
4.0

— 3.5

—)

UJ

S 3 0

Q.

"3

o 2.5
2.0
1.5
1.0
0.5
0.0

¦ Total GLIMPSEv1-Reference_date=2023-1-5T09:23:06-04:00 ¦ Total GCAM5p4-Ref-Orig_date=2023-1-5T07:57:08-04:00

ChartOptions

Year

Total GCA...

2015
5.3

2020
2.89

2025
2.83

2030
2.71

2035
2.54

2040
2.31

2045
1.92

2050
1.38



Total GLI...

5.3

2.69

2.64

2.08

1.79

1.64

1.42

1.13























Figure T2.ll Comparing output from coal plants in the electric sector from one scenario across
scenarios.

Close these windows when you are ready.

Next, we will generate a "difference" plot.

First, change the chart type to "StackedBarChart" and click on "Same Scale" to synchronize the
magnitude of the Y-axis.

T-26


-------
ft GLIMPSE Modellnterface [E:\Projects\GLIMPSE-5p4\GCAM-Model\gcam-v5.4\output\database]
| File Edit Table Help	

~

Scenario

GC AM5p4-R ef-Orig 2023-1-5T0:
GLIMPSEv 1 -R eferenc e 2023-1-5

GLIMPSEv 1 -Tax- 100C-5pct 202,

Africa_Eastern

Africa_Northern

Africa_Southern

Africa_Western

Australia_NZ

Brazil

Canada

Central America and Caribbean

Central Asia

China

EU-12

EU-15

Eui ope_Eastern
Europe_Non_EU

European Free Trade Association

India

Indonesia

Japan	

Queries

- Energy inputs to refining activities (GCAM-U5A)

Technologies

a Electricity generation by region (no cogen)(GCAM-USA)

. Electricity generation by gen and cooling tech (incl cogen)(GCAM-USA)
Electricity generation by aggregated subsector (GCAM-USA)	

Electr icity generation by aggregated subsector rnw detail(GCAM-USA)

a Electricity generation by subsector (GCAM-USA)

-	Electricity generation by tech and cooling (incl cogen)(GCAM-USA)

-	Electricity generation by tech and cooling (incl cogen)(GCAM-USA)(new capacity)

-	Electricity generation by cogen only (GCAM-USA)

¦ Electricity generation input by subsector (GCAM-USA)

. Electricity generation by region (incl cogen)(core)
ss Electricity generation by subsector (core)

-	Electricity generation by gen tech (core)

-	Building final energy by tech

. Building service output by tech
. Industry final energy by tech and fuel

Run Query Diff Query £7] Total

Collapse

Update Single Queries

Q Electricity generation by aggregated subsector rnw detail(GCAM-USA)

Filte

Graph

Format





































see...

region

sub...

2015

2020

2025

2030

2035

2040

2045

2050

Units



More

Display











StackedBarCh...

v

|vj Same Scale



Refresh















GCA,..

[Total

biomass 10.240

0.202

0.197

0.205

0.218

0.229

0.248

0.276

EJ

/s



















GCA...

Total

coal

5.30

2.89

2.83

2.71

2.54

2.31

1.92

1.38

EJ









region: Total











region: Total







GCA...

Total

gas

4.99

6.31

6.63

7.28

8.07

8.93

10.1

11.7

EJ











































GCA...

ITotal

geo

0.0674

0.0885

i0.105

0.116

0.129

0.143

0.112

0.120

EJ





25



















25















a



GCA...

iTotal

hydro

0.910

1.06

1.06

1.06

1.06

1.06

1.05

1.05

EJ













































GCA...

Total

hydr...

0.00

0.00...

0.00...

0.00...

0.00...

0.00190 [0.00...

0.00433

EJ





20













m







20









¦







GCA...

Total

refin...

0.139

0.0645 10.0598

0.0514

0.0467

0.0436

0.0426

0.0429

EJ



HA





























¦













GCA...

ITotal

solar PV (0.116

0.730

|l .04

1.62

2.17

2.71

3.25

3.43

EJ



15

I



















~ 15

¦

¦













GCA...

Total

solar...

0.00

0.0268

10.0557

0.103

0.117

0.124

0.132

0.160

EJ



a



-







-



mm







a











-





GCA...

Total

wind...

0.00

0.00

10.0189

0.0730

0.164

0.286

0.448

0.621

EJ



o

10

¦¦













o 10



¦¦



mm

¦¦

™1





GCA,..

|Total

wind.,.

0.695

1.50

1.87

2.32

2.85

3.36

3.51

3.78

EJ















































GCA...

Total

nuclear

2.99

2.92

b.79

2.68

2.25

1.79

1.36

0.868

EJ





5

¦



















5

¦

















GCA...

Total

solar...

0.0128

0.0600

0.0851

0.128

0.173

0.224

0.306

0.347

EJ







|

¦

|

1

¦

¦

¦



-





|

¦

¦

¦

¦

a

ft

a

ft

-



GLIM,..

ITotal

biomass

0.240

0.207

0.170

0.166

0.178

0.188

0.207

0.231

EJ





0















*



0



3
H





GLIM...

Total

coal

5.30

2.69

2.64

2.08

1.79

1.64

1.42

1.13

EJ









o

Ol





ft

8









8

a

a

GLIM,,.





4.99

6.5">

16.35

6jZ6	

7.77

7.79

8.45

9,45

EJ

V































Figure T2.12 Viewing the thumbnails as stacked bar charts.

Click on "More" in the green area above the thumbnails and click "Difference". A popup
appears with the names of all the scenarios in the table. First, choose "Total GLIMPSEvl-
Reference", then press "OK".

£*"i Select first chart to compare

[Total GCAIM5p4-Ref-Orig_date= 2023- 1-5T07:57:08-04:00/0

Total GLIMPSEvl-Reference_date=2023-l-5T09:23:06-04:00,l

Ok

Figure T2.13 Specifying which scenario to subtract from the other. Typically, the Reference
Scenario is the second selected.

1-27


-------
Then select "GCAM5p4-Ref-Orig" and press "OK".

A difference plot appears. Items that are greater in GUMPSEvl-Reference are shown above the
line, while items that are less are shown below the line.

Um 0_Chart1 -ChartO_Electridtygenerationbyaggregatedsubsectorrnwdetail(GCAM-USA)

UJ 1.0
S- 0.5

Year
biomass

hydro
hydrogen

GLIMPSEvI-Reference Total GCAM5p4-Ref-Orig Total



* solar CSP
> nuclear
wind onshore
wind offshore
solar rooftop PV
solar PV
I refined liquids

¦	hydrogen
I hydro

I geo
a gas

¦	coal

I biomass

2015

o.o

2020
0.005

2025
-0.027

2030
-0.039

2035
-0.04

2040
-0.041

2045
-0.041

2050
-0.045

Figure T2.14 Difference plot where items above the axes represent increases and items below
the axis represent decreases.

The updates incorporated in GLIMPSEvl-Reference have resulted in increased electricity
production from onshore and offshore wind, and utility-scale solar and rooftop photovoltaics.
GLIMPSEvl-Reference also has less output from coal and natural gas.

Stacked bar plots are particularly useful for showing changes from one scenario to another.
T2.5 Additional suggestions for exploration

Use the Mode/Interface's graphical capabilities to examine other outputs of interest, and how
these outputs differ between GLIMPSEvl-Reference and GCAM5p4-Ref-Orig. Suggestions
include:

• "Final energy consumption by aggregate sector"

T-28


-------
•	"Building service output by tech (GCAM-USA)"

•	"Refined liquids production by tech (GCAM-USA)"

•	"Passenger car and truck service output by tech"

T-29


-------
TUTORIAL 3: MODELING AND EVALUATING A CARBON TAX

T3.1 Overview

In this portion of the tutorial, we will develop a new scenario that incorporates a hypothetical
economy-wide tax on Carbon. The tax will start at a value of $100/tC in 2025 (in 1990$s) and
will increase at 5% per year. Please note that this tax is being applied per metric tonne of
Carbon. Converting this its $/tC02 equivalent involves multiplying it by 12/44ths, the ratio of
the molecular weight of C to the molecular weight of CO2. By default, GCAM uses 1990$s. $1 in
1990 is the equivalent of $2.27 in 2022. Considering both conversions, our starting tax is
approximately $62/tC02 in 2022$s.

T3.2 Constructing Carbon tax Scenario Component

Our first step is to create a new Scenario Component that represents our hypothetical tax. First,

click on the "New" button in the Component Library, . This will pop up the New Scenario
Component Creator. Tabs across the top represent several types of policies or alternative
assumptions that can be created. Select "Pollutant Tax/Cap".

T-30


-------
* New Scenario Component Creator

XML List | Pollutant Tax/Cap | Tech Avail Market Share Tech Bound Tech Tax/Subsidy Tech Pa ram Fuel Price Adj

Specification:
Measure:
Pollutant:
Sector:

Names:

Policy:

Market:

Populate:
Type:

Start Yean
End Year:
Initial Val:
Growth (%):

Select One

Values:
Populate
Year

Delete

Clear
Value

jV* Auto?

No content in table

Initial w/% Growth/yr

2020
2050

Add

T world

A

~

USA
Canada





Central America and Caribbe...



~

Mexico



~

Brazil



~

South America_Northern



~

South America_Southern



l_l

Argentina



~

Colombia



~

EU-15



~

EU-12



~

European Free Trade Associa...



~

Europe_Eastern



~

Europe_Non_EU





Africa_Eastern
Africa_Northern



Presets: Select (optional)

Save

Close

Figure T3.1 The New Scenario Component Creator dialog.

The top-left portion is where we define what type of policy we will be creating. The middle
portion will contain the values associated with the policy. The tree at the right allows us to
specify to which regions or states the policy applies. The bottom-left portion can be used to
automate the process of populating the central table with data.

In the top-left portion, under the "Measure:" choice menu, you will find the two types of
policies that can be created on this tab. Select "Emission Tax ($/t).

Next, the "Pollutant:" choice menu can be used to select to which pollutant the tax will be
applied, choose "CO2 (MT C)".

We want this to be an economy-wide tax, so in the "Sector:" choice menu, choose "AN".

The tax will start at $100/tC, starting in 2025, so change the "Start Year:" value to "2025" and
add "100" and "5" in the "Initial Val:" and "Growth (%):" boxes, respectively.

Above the table, press the "Populate" button to populate the table with values that reflect
these options.

T-31


-------
The tax will be applied to all US states, so click the check box next to "USA" on the region table.
When you have completed these steps, the scenario component window should look like this:

New Scenario Component Creator

XML List

Pollutant TaVCap

Tech Avail Market Share Tech Bound Tech Tax/Subsidy Tech Pa ram Fuel Price Adj

Specification:

Measure:

Emission Tax ($/t)

-

Pollutant:

COZ (MT C)

-

Sector:

All

-







Names:

|

Select regson(s):

~ j—| world
~ 0 USA

Canada

Central America and Caribbean
Mexico
I	| Brazil

_j South America_Northern
South America_Southern
Argentina
| Colombia
Q EU-15

y EU-12

| European Free Trade Associatio
| Europe_Eastern
Europe_Non_EU
_ Africa_Eastern
I Africa Northern

Presets: Select (optional)

Add

Save

Close

Figure T3.2 Specifying the options for our carbon tax policy.

Next, press the "Save" button. The "Save As" dialog will appear, and a default name will be
provided. Here, the name has been changed to be more descriptive. When naming scenario
components and scenarios in GLIMPSE it is important only to use alpha-numeric characters and
either or Do not use spaces, or other non-alpha-numeric characters, such as "+", "/" or
"\". Do not use spaces in Scenario Component filenames.

T-32


-------
Save As

X

«- ¦* -

Organize -

I « Projects > GLIMPSE-5p4 > GLIMPSE-Data > GCAM-USA > ScenarioComponents

Search ScenarioComponents

5 This PC

Name

Date modified

Type

Size

0" 3D Objects

*1 C a 1 ib_C oalEG U_2020_AK. csv

7/18/2022 7:19 PM

Microsoft Excel C...

78 KB

XHR Desktop

*1 C a 1 ib_C oalEG U_2020_ALcsv

7/18/20227:19 PM

Microsoft Excel C...

78 KB

Qj Documents

*5 Calib-coal_egu_2020.csv

7/18/20227:19 PM

Microsoft Excel C„.

1,375 KB

Downloads

*J| Calib-coal_egu_2020b.csv

7/18/2022 7:19 PM

Microsoft Excel C...

1,375 KB

Music

*1 C a 1 i b- LDV- EV - AE02020,csv

7/18/20227:19 PM

Microsoft Excel C...

585 KB

13 Pictures
££ Videos

*1 Calib-Lo_alt_biofuel_SW.csv
Calib-NE_fixed_nuke_output.csv
Ca 1 ib-OffshoreWindJower-Bound_NE.csv

7/18/20227:19 PM
7/18/20227:19 PM
7/18/2022 7:19 PM

Microsoft Excel C...
Microsoft Excel C...
Microsoft Excel C...

173 KB
43 KB
10 KB

OSDisk (CO

xl LDV- EV-4W_N ew_Reg_Ea. csv

7/18/2022 7:19 PM

Microsoft Excel C...

530 KB

& DVD Drive (DO

xjj No-ccs-NE.csv

7/18/20227:19 PM

Microsoft Excel C...

42 KB

i—i New Volume (E v

xj P olicy- CES_1 OOx 50_USA_acr.csv

7/18/2022 7:19 PM

Microsoft Excel C...

79,084 KB

File name: | Tax_C_10Qdpt-5pct-2025.csvj
Save as type: I CSV files (*.csv)

^ Hide Folders

Figure T3.3 Saving our scenario with a unique name.

Note that the New Scenario Component Creator does not automatically disappear after you
have clicked "Save". This behavior is intentional and is intended to allow the user to modify the
constraint readily.

Once you have pressed "Save", the component will appear in the Component Library. By
default, the components are in alphabetical order. You can click on the appropriate column
name to sort the table. Here, the components are sorted by the date they were created:

T-33


-------
Q GLIMPSE Scenario Builder
File Tools View Help

Component Library

EE ^ B X o

Component Name

Created



~

T ax_C_ 10Od pt-5 pct-2025.csv

2023-04-24:

22:55



GCAM-Updates-postRelease.txt

2023-04-24:

12:49



Calib-OffshoreWind_lower-Bound_EastCoast_thru2035.csv

2023-04-24:

11:20



Emissions-AirPollutants-AIISectors.txt

2023-04-24:

10:47



Tech-EGU-Cost-Updates-NREL-ATB2021-Mod.txt

2023-04-24:

10:42



Calib-Transport-ALM-EF-Adj.txt

2023-04-24:

10:39



Policy-EGU-NSPS-EFs-NOxS02PM.txt

2023-04-24:

10:29

V

< 1



	i >



Create Scenario

Component Name

No content in table

Scenario Library

7





>

X











ftraj

i)





li

Scenario Name

Created

Completed

Status

ProbMkts

Runtime

GCAM5p4-Ref-Orig

2023-04-24: 12:20

2023-04-24:13:49

Success



0 hr47 min

GLIMPSEvI -Reference

2023-04-24:16:32

2023-04-24:18:27

Success



1 hr 54 min



Resources... CPU: 38% | RAM: 31.9GB Free:70% | Disk: 2,797.9GB available | Swap: 63.9GB Free: 71% // Database: database Size: 13.0GB Used: 32.5%

Figure T3.4 Selecting a scenario to modify.

The next step is to construct a scenario that includes this policy. We will do this by modifying an
existing scenario.

Start by clicking on GLIMPSEvl-Reference in the Scenario Library. Then, click on the button
to allow you to modify it.

You will see components associated with GLIMPSEvl-Reference displayed in the table at the
top-right. Change the name above the table to "GL!MPSEvl-Tax-C100dpt5pct" to describe your
new scenario.

Then, click on the new tax component in the Component Library and press the button to
add it to your scenario:

T-34


-------
f^\ jLIMPSE Scenario Builder	— ~

File Tools View Help

Component Library Search:

Component Name

Tax_C_ 10Od pt - 5 p ct-2025.csv

Policy-ROW-CTax-25dpt5pct.csv

Calib-OnroadTrn-SW-Ref-Updates.txt

DeepDecarbAssumptions.txt

Tech-NoLDVCNGs.csv

Policy-100pct-EV-NoConv2050.csv

Dolirw-1 nnr.H--P\/-Croinh»Tr>-U.I-loa>n. ww

4- # x Q

Created

2023-05-01:

2023-04-28
2023-04-27
2023-04-27
2023-04-25
2023-04-25

12:14

06:56
11:50
11:43
16:31
13:35



~
«
4

Create Scenario

GLIMPSEvI -Tax-100C-5pct

Component Name

Policy-LD-NTR-BEV-Sales-LDComTruck.csv

Policy-LD-NTR-BEV-Sales-LDTruck.csv

Policy-EGU-NSPS-EFs-NOxS02PM.txt

Policy-NoBiomassEGUinMA.csv

Tax_C_.100dpt-5pct-2025.csv

©

Scenario Library

Search:







>

X

™ itaii



C-=-

!

Pi Pi

| iog | |t»i|

Hi



Scenario Name

Created

Completed

Status

ProbMkts

Runtime

GCAM5p4-Ref-Orig

2023-04-24:12:20

2023-05-01: 08:48

Success



0 hr 51 min

GLIMPSEvI-Reference

2023-05-01:09:20

2023-05-01:11:04

Success



1 hr41 min

Resources... CPU: 0% | RAM: 31.9GB Free:61% | Disk: 2,758.6GB available | Swap: 63.9GB Free: 66% // Database: database Size: 4.3GB Used: 10.8

Figure T3.5 Adding the new scenario component to the scenario.

To create the new scenario, press the button. A dialog will appear that indicates the name
of the scenario, the database to which the results will be sent, the final year to include in the
simulation, which region to use as the debug region, whether to create the debug file, and
check boxes allowing you to choose which files to save when the scenario run is complete.

Many of these settings are available in the GLIMPSE options file, but this dialog gives the ability
to override those settings. The debug file includes detailed outputs for a specific region. Only
one region or state can be selected since the debug file is very large.

The "Use all available processors?" option allows you to indicate that GCAM should distribute
the computational load across all available processing cores. In general, this option should be
selected. However, there may be instances where unsolved markets occur as a result of the
pre-emptive calculations that occur during parallel computations. For scenarios that have
unsolved markets for which the cause is unclear, repeating the model run, but without the "Use
all available processors?" box checked, may help pinpoint or eliminate errors introduced by
parallel calculations as the cause.

There is a box to add comments or other meta-data that describes the scenario.

T-35


-------
El Creating Scenario	— ~ X

Scenario name:	GLIMPSEv1-Tax-100C-5pct

Database:	database (4 GB)

Final model year.	2050 ~r

\/ Create debug file? CT	i

Use all available processors?

Save files in scenario folder: (global setting)

V	Main log	V Debug file

V	Calibration log \V Solver log
Comments:

Scenario based on GLIMPSEvI-Reference that incorporates a
carbon tax of $100/tC (1990$s), starting in 2025 and growing
at 5 percent per year

OK	Cancel

Figure T3.6 The Creating Scenario dialog.

Note that the database size is important. When the database reaches approximately 40 GB, it
can no longer be opened in Windows. If you attempt to create a new scenario and the database
size is already 36 GB, a warning message will occur. See the Users' Guide for information on
managing database size.

Once you press "OK", the new scenario will be added to the Scenario Library. Double-clicking on
the new scenario's name in the Scenario Library will display its configuration file.

T-36


-------
Ifaif E:\Projects\GLIMPSE-5p4\GLIMPSE-Data\GCAM-USA\ScenarioFolders\GLIMPSEv1-Tax-100C-5pct\configuration_GLIMPSEv1-Tax-100C-5pc... —
File Edit Search View Encoding Language Settings Tools Macro Run Plugins Window ?

o'oa	i ^igaais?ii psanaaa«>i®ii'iHjiBi® a * < >

9 configuration GLIMPSEvI -Reference.xml £3 SI options Gi

.H configuration_GLIMPSEv1-Tax-100C-5pct.xml EZ3

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41



3

3

P]

CValue name="xmlInputFileName">../input/gcamdata/xml/modeltime.xml

"D 4-^.V.TT-; 1 cVo



length : 19,249 lines : 245

Ln : 1 Col: 1 Pos : 1

Windows (CR LF) UTF-8

Figure T3.7 The new scenario's configuration file.

Opening the scenario folder by pressing - will show the XML-formatted files that were
generated from scenario components during the scenario's creation.

T-37


-------
i ! 1 P 9 C1 - 1 GLIMPSEvI -Tax-100C-5pct





— H X

File Home Share View





v £

-> v 1" P ScenarioFolders > GLIMPSEv1-Tax-100C-5pct

<
c

Search GLIMPSEv1-Tax-100...

A

Name

Date modified

Type

Size

¦ Calib-2025lndCoal-GHGPIanStates.xml

5/1/2023 12:20 PM

XML Document

705 KB

I Calib-Lo_alt_biofuel_SW.xml

5/1/2023 12:20 PM

XML Document

3,291 KB

B Calib-NoNewCoalEGUslnOCoalStates.xml

5/1/2023 12:20 PM

XML Document

336 KB

B Calib-OffshoreWind_lower-Bound_EastCoast_th...

5/1/2023 12:20 PM

XML Document

27 KB

I configuration_GLIMPSEv1-Tax-100C-5pct.xml

5/1/2023 12:20 PM

XML Document

19 KB

¦ Policy-LD-NTR-BEV-Sales-Car.xml

5/1/2023 12:20 PM

XML Document

726 KB

I Policy-LD-NTR-BEV-Sales-LDComTruck.xml

5/1/2023 12:20 PM

XML Document

944 KB

¦ Policy-LD-NTR-BEV-Sales-LDTruck.xml

5/1/2023 12:20 PM

XML Document

724 KB

| Policy-NoBiomassEGUinMA.xml

5/1/2023 12:20 PM

XML Document

212 KB

I Policy-Sec177-ZEV-FreightltTruck-20200nly.xml

5/1/2023 12:20 PM

XML Document

48 KB

B Policy-Sec177-ZEV-PassCars 20200nly.xrnl

5/1/2023 12:20 PM

XML Document

48 KB

| Policy-Sec177-ZEV-PassLgCarAndTruck-20200n...

5/1/2023 12:20 PM

XML Document

45 KB

B Tax_C_100dpt-5pct-2025.xml

5/1/2023 12:20 PM

XML Document

8 KB

13 items I





I a

Figure T3.8 Contents of the scenario's folder.

Close the text editor and scenario folder.

Next, start running the scenario by clicking on "GLIMPSEvl-Tax-C100dpt5pct" and pressing play,

>

A black window will appear, providing diagnostic information about the run.

Once the run has completed after 1 to 3 hours, you will see its status updated in the Scenario
Library.

T-38


-------
Q GLIMPSE Scenario Builder
File Tools View Help

~

Component Library Search:

Component Name

Tax_C_ 10Od pt - 5 p ct-2025.csv

Policy-ROW-CTax-25dpt5pct.csv

Calib-OnroadTrn-SW-Ref-Updates.txt

DeepDecarbAssumptions.txt

Tech-NoLDVCNGs.csv

Policy-100pct-EV-NoConv2050.csv

Oolirw-1 nnr.H--P\/-Croir.h»TrrU-MaiaWt/ /-co

4- # x Q

Created

2023-05-01:

2023-04-28
2023-04-27
2023-04-27
2023-04-25
2023-04-25

12:14

06:56
11:50
11:43
16:31
13:35



~
«
4

Create Scenario

GLIMPSEvI -Tax-100C-5pct

Component Name

Policy-LD-NTR-BEV-Sales-LDComTruck.csv

Policy-LD-NTR-BEV-Sales-LDTruck.csv

Policy-EGU-NSPS-EFs-NOxS02PM.txt

Policy-NoBiomassEGUinMA.csv

Tax_C_.100dpt-5pct-2025.csv

©

Scenario Library

Search:







>

X

™ itaii



C-=-

!

Pi Pi

liQ&l

Hi



Scenario Name

Created

Completed

Status

ProbMkts

Runtime

GCAM5p4-Ref-Orig

2023-04-24:12:20

2023-05-01: 08:48

Success



0 hr 51 min

GLIMPSEvI -Reference

2023-05-01:09:20

2023-05-01:11:04

Success



1 hr41 min

GLIMPSEvI-Tax-100C-5pct

2023-05-01:12:20

2023-05-01:13:55

Success



1 hr 34 min



Resources... CPU: 8% | RAM: 31.9GB Free:66% | Dislc 2,756.0GB available | Swap: 63.9GB Free: 68% // Database: database Size: 6.5GB Used: 16.3%

Figure T3.10 After the new scenario run has been completed, its status is updated
T3.3 Exploring the response to the tax

The Mode/Interface provides a variety of features for examining the differences between
scenarios. Several examples were provided in Tutorial Part 2, including transposing plots so to
compare data by series and creating "difference" graphs. In this section, we use similar
approaches to examine the impacts of the carbon tax policy.

First, start up the Mode/Interface by clicking on "GLIMPSEvl-Tax-100C-5pct" in the Scenario

T^i

Library, then clicking the Results button with the arrow, POD. GLIMPSE will then access the
scenario's configuration file, read the name of the database to which its results were stored,
then open that database via the Mode/Interface.

For this tutorial, we will be exploring national-scale impacts of the tax policy. Select "GLIMPSE-
vl-Reference" and "GL!MPSEvl-Tax-100C-5pct" in the Scenario pane. Next, in the Regions
pane, select all states and the USA region. Click the check box next to "Total" to obtain national
totals.

T-39


-------
Select "CO2 emissions by region", then click the "Run Query" button. The results will populate
the table after several seconds.

~nj GLIMPSE Modellnterface [E:\Projects\GLIMPSE-5p4\GCAM-Model\gcam-v5.4\output\database]
File Edit Table Help

~

X

Scenario

Regions

Queries

GCAM5p4-Ref-Orig 2023-1-5T07:57:0

GUMPSEvl-Reference 2023-l-5T09:23
GLIMPSEvl-Tax-100C-5pct 2023-1-5TI

Africa_Southem
Africa_Western
Australia_NZ
Brazil

Canada

Central America and Caribbean

Central Asia

China

EU-12

EU-15

Europe_Eastern
Europe_Non_EU

European Free Trade Association
India

H

¦ Transport service output by tech
Transport service output by tech (new capacity)
Passenger car and truck service output by tech
Freight truck service output by tech (no bus)
Emissions

C02 emissions bv reqion

«C02 emissions by aggregate sector
- C02 emissions by sector
C02 by resource production
C02 emissions by sector (no bio)
C02 emissions by tech
N0x,S02,PM2.5 by region
N0x,S02,PM2.5 by aggregate sector
N0x,S02,PM2.5 by sector
* N0x,S02,PM2.5 by tech
All emission'; bv reninn

Run Query Diff Query 0 Total Collapse Update Single 1

~ C02 emissions by region

Filter Graph Format













scenario region 2015

2020

2025

2030 2035 2040

2045 2050

Units



GLIMPSEvl. JTotal 11420

1240

1190

1120 11080 11070

1040 11030

MTC



GUMPSEvl. JTotal 1420

1240

1120

1020 935 853

724 572

MTC







Figure T3.ll CO2 emissions for the two scenarios

Under the policy, CO2 emissions decrease to 572 MTC by 2050. This is 460 MTC less than
GUMPSEvl-Reference in 2050 and a reduction of nearly 55% from 2020 levels. Pressing the
"Graph" button shows each scenario's CO2 trajectory on a separate thumbnail. However, using
the "More->Transpose" option generates the following graphic.

T-40


-------
^0_ C02emissionsbyregion

X

¦ GLIMPSEv1-Tax-100C-5pct_date=2023-1-5T12:21:28-04:00 ¦ GLIMPSEv1-Reference_date=2023-1-5T09:23:06-04:00

ChartOptions

Year

GLIMPSEvl-Reference date...

2015

1420.0

2020

1240.0

2025

1190.0

2030

1120.0

2035

1080.0

2040

1070.0

2045

1040.0

2050

1030.0



GLIMPSEvl-Tax- 100C-5pct_...

1420.0

1240.0

1120.0

1020.0

935.0

853.0

724.0

572.0























Figure T3.12 Visualizing the CO2 trajectories for the two scenarios

Next, we will explore from which sectors GCAM is obtaining CO2 reductions.

Select and execute the "CO2 emissions by aggregate sector" query. Graph the results, then
change the graph formatted to stacked bar charts.

Next, using "More->Difference", generate a difference plot. When the popup appears, select
one scenario then the next, first click on "GLIMPSEvl-Tax-100C-5pct", click "OK", then click
GLIMPSEvl-Reference and click "OK".

This ordering - policy case, then reference case - is typical because values that go up as a result
of the policy will be shown in the positive direction and values that go down will be negative.

T-41


-------
Sk Chart!-ChartO_C02emissionsbyaggregatesector

GLIMPSEv1-Tax-100C-5pct Total GLIMPSEvI-Reference Total

o

-50

s


-------
jjjj GLIMPSE Modellnterface [E:\Projects\GLIMPSE-5p4\GCAM-ModeI\gcam-v5.4\output\database3]
File Edit Table Help	

p:

|GCAM5p4-Ref-Orig 2023-1-5T

GLIMPSEvl-Reference 2023-1
GLIMPSEv 1 -Tax-100C-5pct 2C

Africa_Eastern

Afi ica_Northern

Afi ica_Southei n

Afriea_Westein

AustraliaJMZ

Brazil

Canada

Central America and Cai ibbe-

Central Asia

China

EU-12

EU-15

Europe_Eastern
Europe_Non_EU
European Free Trade Associe s

< >

Queries

. End-use energy consumption in industry (detail)

« Energy inputs to electricity production (GCAM-USA)

-	Energy inputs to electricity production (GCAM-USA) (detail)

. Energy inputs to refining activities (GCAM-USA)

Technologies

Electricity generation by region (no cogen)(GCAM-USA)

-	Electricity generation by gen and cooling tech (ind cogen)(GCAM-USA)
Electricity generation by aggregated subsector (GCAM-USA)

Electricity generation by aggregated subsector rnw detail(GCAM-USA)

-	Electricity generation by subsector (GCAM-USA)

Electricity generation by tech and cooling (ind cogen)(GCAM-USA)

Electricity generation by tech and cooling (ind cogen)(GCAM-USA)(new capacity)
« Electricity generation by cogen only (GCAM-USA)

. Electricity generation input by subsector (GCAM-USA)
o Electricity generation by region (ind cogen)(core)

-	Electricity generation by subsector (core)

Run Query

Diff Query 0 Total Collapse Update Single Queries Create Remove

Edit

~ Diff: Outputs by tech Q Diff: Prices of all markets Q Diff: Building service costs ~ Electricity generation by aggregated subsector rnw detail(GCAM-USA)

Filtei



Graph

Format













see...
GLIM...

reg.
Total

. sub...
|biom...

2015 2020
0.240 |0.207

2025
0.170

2030
0.166

2035 2040
0.178 10.188

2045
0.207

2050
0.231

Units
EJ

A

GLIM...

Total

{coal

5.30 2.69

2.64

2.08

1.79 1.64

1.42

1.13

EJ



GLIM...

Total

gas

4.99 6.52

6.35

6.76

7.27 |7.79

8.45

9.45

EJ



GLIM...

Total

|geo

0.0674 0.0907

0.117

0.132

0.145 0.161

0.130

0.136

EJ



GLIM...

Total

[hydro

0.910 1.06

1.06

1.06

1.06 11.06

1.05

1.05

EJ



GLIM...

Total

|hydr...

0.00 0.00...

0.00...

0.00...

0.00... 0.00...

0.00...

0.00...

EJ



GLIM...

Total

jrefin...

0.139 0.0666

0.0417

0.0359

0.0341 0.0330

0.0322

0.0317

EJ



GLIM...

Total

Isolar...

0.116 0.752

1.67

2.44

2.99 3.48

4.02

4.30

EJ



GLIM...

Total

solar...

0.00 0.0268

0.0871

0.159

0.169 !0.208

0.237

0.287

EJ



GLIM...

Total

[wind...

0.00 |0.00

0.171

0.377

0.747 ,0.871

1.03

1.09

EJ



GLIM...

Total

wind...

0.695 1.57

2.37

3.08

3.68 4.25

4.58

5.07

EJ



GLIM...

Total

Inuclear

2.99 12.92

2.79

2.71

2.29 1.83

1.44

1.05

EJ



GLIM...

Total

solar...

0.012810.0628

0.160

0.277

0.390 '0.535 0.758

0.927

EJ



GLIM...

Total

biom...

0.240 0.207

0.189

0.191

0.211 >0.230 ;0.254

0.255

~

V

GLIMPSEvl-Reference
region: Total

GLIMPSEvl -Tax-100C-5pct
region: Total





o o o o

—« K) N) GJ

N> K) ro ro k>

k & 8

a a

M\

linn

io	ro ro	to

0	o o	o

CO	£ JL	Ol

01	O oi	O

Figure T3.14 Electricity production by technology category for the Reference and Tax
scenarios.

Next, use "More->Difference" to generate show the changes in electricity production under the
task. The results suggest that the tax may result in reductions in electricity production from
conventional coal and gas, offset by renewables, fossil production with CCS, biomass with CCS,
and nuclear power. Overall, the quantity of electricity produced does not change substantially
since the height of the stacked bars above and below 0 for each time period are roughly the
same.

T-43


-------
I jfa| 0_Chart1-Chart0_Electricitygenerationbyaggregatedsubsectorrnwdetail(GCAM-USA)

X

GLIMPSEv1-Tax-100C 5pct Total GLIMPSEvI Reference Total	a

9
7
e
5
4
3
2

—>

111

3

a

3 0
o

-1
-2
-3
-4
•5
¦e

-7

2015	2020	2025	2030	2035	2040	2046	2050

Year

2015

2020

2025

2030

2035

2040

2045

2050

biomass

0.0

0.0

0.019

0.025

0.033

0.042

0.047

0.024



A

biomass w/CCS

0.0

0.0

0.009

0.018

0.037

0.07

0.174

0.478





coal

0.0

0.0

-0.34

-0.54

-0.851

-1.044

-1.106

-0.988





coal w/CCS

0.0

0.0

0.143

0.259

0.408

0.585

0.831

1.16





gas

0.0

0.0

-0.94

-1.43

-2.07

-2.85

-4.06

-5.73





gas w/CCS

0.0

0.0

0.192

0.353

0.57

0.814

1.15

1.59





geo

0.0

0.0

0.017

0.024

0.034

0.043

0.056

0.067





hydro

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0





hydrogen

0.0

0.0

-0.0

-0.0

-0.0

-0.0

-0.0

-0.001





refined liquids

0.0

0.0

-0.002

-0.002

-0.002

-0.004

-0.007

-0.008





solar PV

0.0

0.0

0.24

0.27

0.33

0.42

0.62

0.96





solar rooftop PV

0.0

0.0

0.025

0.06

0.107

0.178

0.238

0.265





wind offshore

0.0

0.0

0.021

0.054

0.097

0.126

0.18

0.22





wind onshore

0.0

0.0

0.29

0.46

0.65

0.83

1.04

1.19





nuclear

0.0

0.0

-0.01

0.08

0.19

0.29

0.46

0.66





solar CSP

0.0

0.0

0.059

0.113

0.179

0.256

0.362

0.503





refined liquids w/CCS

0.0

0.0

0.0

0.0

0.0

0.0

0.0

0.0



V

Figure T3.15 Difference graph showing the electric sector response to the tax.

In this result, biomass use is increasing in the electric sector. Next, we will explore the degree to
which biomass use is increasing in other sectors as well.

Run the query "Biomass use by aggregate sector", graph the results, and display them as
stacked bar charts. The results suggest that the greatest increase in biomass use is in the fuel
production sector.



¦	refined liquids w/CCS
solar CSP

¦	nuclear
wind onshore
wind offshore
solar rooftop PV
solar PV

¦	refined liquids

¦	hydrogen

¦	hydro

¦	geo

¦	gas w/CCS

¦	gas

»coal w/CCS

¦	coal

¦	biomass w/CCS

¦	biomass

T-44


-------
jjjj GLIMPSE Modellnterface [E:\Projects\GLIMPSE-5p4\GCAM-Model\gcam-v5.4\output\database3]
File Edit Table Help	

p:

|GCAM5p4-Ref-Orig 2023-1-5T

GLIMPSEvl-Reference 2023-1
GLIMPSEv 1 -Tax-100C-5pct 2C

Africa_Eastern

Africa_Northern

Africa_Southern

Africa_Western

AustraliaJMZ

Brazil

Canada

Central America and Caribbe-

Central Asia

China

EU-12

EU-15

Europe_Eastern
Europe_Non_EU

Queries

Primary energy consumption by region (direct equivalent)
« Final energy consumption by region
Final energy consumption by aggregate sector
. Final energy consumption by aggregate sector and fuel
Final energy consumption by aggregate sector and fuel(v2)
Final energy consumption by sector and fuel
Energy inputs by sector

Electricity use by aggregate sector
Coal use by aggregate sector
« Natural gas use by aggregate sector
Refined liquids use by aggregate sector

Biomass use by aggregate sector

- Hydrogen use by aggregate sector
. End-use energy consumption in buildings
. End-use energy consumption in buildings (detail)

Run Query Diff Query 0 Total Collapse Update Single Queries Create Remove

El Diff: Outputs by tech

Q Electricity generation by aggregated subsector

~ Diff: Prices of
rnw detail(GCAM-

all markets
USA)

Diff: Buildina service costs
Q Biomass use by aggregate sector

Graph

see...

reg...

sec...

2015

2020

2025

2030

2035

0

s

ri

2045

2050

Units

GLI...

Total

com...

0.103

0.109

0.114

0.118

0.124

0.128

0.132

0.134

EJ

GLI...

Total

elect..,

0.560

0.344

0.182

0.156

0.187

0.210

0.258

0.321

EJ

GLI...

Total

indu...

1.47

1.60

1.62

1.67

1.71

1.77

1.80

1.81

EJ

GLI...

Total

resid...

0.419

0.405

0.388

0.364

0.337

0.310

0.287

0.268

EJ

GLI...

Total

fuel...

1.31

1.45

1.39

1.30

1.19

1.80

2.77

3.54

EJ

GLI...

Total

com...

0.103

0.109

0.117

0.131

0.149

0.169

0.198

0.199

EJ

GLI...

Total

elect...

0.560

0.344

0.295

0.314

0.437

0.589

0.972

1.91

EJ

GLI...

Total

indu...

1.47

1.60

1.63

1.70

1.77

1.87

1.94

1.96

EJ

GLI...

Total

resid...

0.419

0.405

0.393

0.377

0.364

0.355

0.350

0.326

EJ

GLI...

Total

Fuel...

1.31

1.45

1.39

1.33

1.28

2.85

6.35

10.7

EJ



Moie Display * '

StackedBarC... v 0 Same Scale

GLIMPSEvl Reference
region: Total

S 10

i:

Refresh

II

-ll

II II

GLIMPSEvl Tax 100C-5pct
region: Total

LJ

II

¦

M KJ K> M

o o o o

a s ft 8

Figure T3.16 Difference graph showing the sectoral CO2 response to the tax.

The query "Refined liquids production by tech(GCAM-USA)" provides insight into how this
biomass is being used.

Run that query, then generate a difference plot indicating how refined liquids production
technologies are responding to the tax. Here, we see a reduction in oil refining, which is being
offset by an increase in bio-refining, some of which integrates CCS.

T-45


-------
|&| 0_Chart1 -ChartO_Refinedliquidsproductionbytech(GCAM-USA)

X

GLIMPSEv1-Tax-100C-5pct Total GLIMPSEvI Reference Total

3

CL

5
4
3
2
1
0
-1
-2
-3
-4
-5
-6

2015

2020

2025

2030

2035

2040

2045

2050

corn ethanol ¦ cellulosic ethanol CCS level 2 ¦ cellulosic ethanol CCS level 1 cellulosic ethanol ¦ biodiesel
FT biofuels CCS level 2 FT biofuels CCS level 1 ¦ FT biofuels ¦ oil refining gas to liquids
i coal to liquids CCS level 2 ¦ coal to liquids CCS level 1

ChartOptions

Year	2015 2020 2025 2030 2035 2040 2045 2050

coal to liquids CCS level 1	0.0	0.0	0.0	0.0	0.0	0.017 0.049 0.078

coal to liquids CCS level 2
gas to liquids

0.0

0.0

0.0

0.0

0.0

0.015

0.046

0.074

0.0

0.0

0.0

0.0

0.0

-0.059

-0.26

-0.542

oil refining

0.0

0.0

¦0.4

-0.4

-0.4

-1.3

-3.1

-5.4

FT biofuels

0.0

0.0

0.0

0.0

0.0

0.01

0.042

0.085

FT biofuels CCS level 1

0.0

0.0

0.0

0.0

0.0

0.034

0.154

0.361

FT biofuels CCS level 2

0.0

0.0

0.0

0.0

0.0

0.031

0.15

0.364

biodiesel

0.0

0.0

¦0.001

0.0

0.003

0.058

0.11

0.155

cellulosic ethanol

0.0

0.0

¦0.0

0.0

0.0

0.032

0.136

0.27

cellulosic ethanol CCS level 1

0.0

0.0

0.0

0.0

0.0

0.068

0.273

0.544

cellulosic ethanol CCS level 2

0.0

0.0

0.0

0.0

0.0

0.08

0.395

0.996

corn ethanol

0.0

0.0

0.0

0.0

0.02

0.4

0.97

1.49

Figure T3.17 Difference graph showing the sectoral CO2 response to the tax.

In the next part of the tutorial, we explore additional ways to identify the impacts of policy
across scenarios.

T-46


-------
TUTORIAL 4: ADDITIONAL TOOLS FOR COMPARING SCENARIOS

T4.1 Overview

The Model I interface provides a variety of features for examining the differences from one
scenario to another. Several examples were provided in Tutorial Part 2, including transposing
plots so to compare data by series and creating "difference" graphs. These options are
particularly useful if you know ahead of time which results you would like to examine. However,
in some instances, it may be difficult to determine where to get started. Helping identify major
changes between scenarios is a strength of the "Diff Query" feature, which is discussed here.

T4.2 Using the Diff Query

Using the "Diff Query" is similar to executing a query using the "Run Query" button.

First, select the scenarios and regions of interest. Here, we have selected GLIMPSEvl-Reference
and our new carbon tax scenario from Tutorial Part 3. For this demonstration, we will start with
examining national totals, so select all of the states and the USA region, then check the box
next to "Total".

The "Diff Query" option can be used for any query, but it is often most useful for queries that
return results across many sectors, including "Inputs by Tech", "Outputs by Tech", and "Prices
for all markets". For this tutorial, we have chosen "Outputs by tech".

T-47


-------
jjfgg GLIMPSE Modellnterface [E:\Projects\GLIMPSE-5p4\GCAM-Model\gcam-v5.4\output\database3]
File Edit Table Help

Scenario	

GCAM5p4-Ref-Ortg 2023-l-5lj |

Regions

GLIMPSEv 1 -R ef erence 2023-1
GUMPSEvl-Tax-100C -5pct 2G

Africa_Eastern

Africa_Northern

Africa_Southern

Africa_Western

Australia_NZ

Brazil

Canada

Central America and Caribbe.

Central Asia

China

EU-12

EU-15

Europe_Eastern

Europe_Non_EU

European Free Trade Associc

India

Indonesia

Queries

Total climate forcing
* Global mean temperature
Markets, prices, and costs
m C02 prices
t Prices of all markets
• Supply of all markets
. Demand of all markets
a Final energy prices
. Costs by tech and input

Building service costs
- Costs of transport services
. Elec prices by sector (GCAM-USA)
Elec prices by sector (core)

Inputs and outputs
Inputs by tech

Outputs by tech

. elec production (ind CHP)
. electricity use

Run Query Diff Query 0 Total Collapse Update Single Queries Create Remove

Edit

Figure T4.2 Selecting the "outputs by tech" query.

Next, press the "Diff Query" button. After several seconds, a dialog will appear that provides
options for how you would like differences to be displayed.

The top area is where you select the from which scenario differences will be calculated. Below
that area, you can select the smallest difference value (Minimum value) and smallest percent
difference (Minimum percent) that you are willing to display. Differences that do not meet
these criteria are not displayed.

The final option allows you to display the query results as values or percent differences.

T-48


-------
jag Difference Options

Choose Reference Scenario for Calculations

Minimum value:

Minimum percent:

Show differences as:

OK	cancel

Figure T4.3 Selecting Diff Query options, including the Reference, difference criteria, and
whether to report differences as percents.

Here, parameters that differ at least by "0.00001" and "30%" at some point over the modeled
time horizon are shown in the table as percent changes from GLIMPSEvl-Reference. Note that
instances involving divide-by-zero are reported as "1000" in the table.

0.00001 v
30	v

Percent v

T-49


-------
TTnh GLIMPSE Modellnterface [E:\Projects\GLIMPSE-5p4\GCAM-Model\gcam-v5.4\output\database3]	— ~ X

File Edit Table Help

Scenario

Regions

Queries

GCAM5p4-Pef-Orig 2023-1-5T!

	H









• « Total climate forcing a

GLIMPSEv 1 -R ef erence 2023-1



Africa_Eastern









* Global mean temperature

GLIMPSE v 1 -T ax-100C -5pct 2C



Africa_Northern







Markets, prices, and costs





Africa_Southern









m C02 prices





Africa_Western









« Prices of all markets





Australia_NZ









« Supply of all markets





Brazil









. Demand of all markets





Canada









m Final energy prices





Central America and Caribbe.









» Costs by tech and input





Central Asia









* Building service costs





China









« Costs of transport services





EU-12









¦ " « Elec prices by sector (GCAM-USA)





EU-15









. Elec prices by sector (core)





Europe_Eastern







Inputs and outputs





Europe_Non_EU









Inputs by tech





European Free Trade Assoeic v











< >



< >



Run Query Diff Query 0 Total Collapse Update Single Queries Create Remove Edit

Q Diff: Outputs by tech

Filter Graph Format

scenario
GLIMPSE...

re...
Total

sector

comm heating

subsector
biomass

output

comm heating

technology 2015
wood furnace lO.OO

2020

0.0000705

2025
3.40

2030
10.4

2035
20.3

2040
31.7

2045
50.1

2050
48.6

Units
pet (EJ)

GLIMPSE...

Total

comm heating

electricity

comm heating

electric heat pumpi0.00

0.0000313

3.79

11.1

19.0

27.5

39.8

48.6

pet (EJ)

GLIMPSE...

Total

comm heating

gas

comm heating

gas furnace jO.OO

-0.0000...

-1.90

-4.23

-8.37

-14.2

-28.7

-33.8

pet (EJ)

GLIMPSE...

Total

comm hot water

electricity

comm hot water

electric heat pu... 1000

0.000103

16.7

29.8

38.6

47.3

59.1

71.1

pet (EJ)

GLIMPSE...

Total

gas to liquids

gas to liquids

gas to liquids

gas to liquids 1000

1000

1000

1000

1000

-38.8

-47.6

-54.3

pet (EJ)

GLIMPSE...

Total

regional biomass

regional biomass

regional biomass

regional biomass 0.00

0.000228

9.46

13.2

21.8

51.1

127

231

pet (EJ)

GLIMPSE...

Total

resid cooling

electricity

resid cooling

aii conditioning ...

1000

0.000141

8.02

15.3

23.1

30.7

37.3

41.2

pet (EJ)

GLIMPSE...

Total

resid hot water

electricity

resid hot water

electric heat pu...

1000

0.0000722

26.8

37.4

46.6

52.7

59.6

66.3

pet (EJ)

GLIMPSE...

Total

tin aviation intl

International Av...

trn_aviation_intl

BEV

1000

1000

1000

1000

8.45

14.5

22.7

32.3

pet (millio...

GLIMPSE...

Total

trnjreight

Domestic Ship

trnjreight

Liquids

0.00

-0.000139

-14.5

-18.0

-21.8

-29.0

-37.2

-46.9

pet (millio...

GLIMPSE...

Total

water td elec C

PacArctic

water_td_elec_C

PacArctic

0.00

0.00209

-18.2

-15.5

-9.42

0.293

21.4

42.6

pet (km^3)

GLIMPSE...

Total

water_td_elec...

PacArctic

water td elec W

PacArctic

0.00

0.00449

-56.5

-62.9

-63.3

-67.5

-70.4

-68.6

pet (kmA3)

GLIMPSE...

Total

biomass liquids

biomass liquids

biomass liquids

FT biofuels

1000

0.00322

-0.424

0.131

1.80

108

158

194

pet(EJ)

GLIMPSE...

Total

biomass liquids

biomass liquids

biomass liquids

biodiesel

0.00

0.000239

-0.599

-0.0233

1.65

20.7

31.6

39.8

pet(EJ)

GLIMPSE...

Total

biomass liquids

biomass liquids

biomass liquids

cellulosic ethanol

1000

0.00323

-0.320

0.225

1.89

111

163

199

pet (EJ)

GLIMPSE...

Total

biomass liquids

biomass liquids

biomass liquids

corn ethanol

0.00

0.000296

-0.393

0.169

1.83

24.9

41.0

51.2

pet (EJ)

Figure T4.4 Diff Query results shown as percents.

The items and values in the table provide valuable insights into the model's response to the
carbon tax. For example, In the commercial sector, by 2050, space heating by heat pumps has
increased nearly 49%, while heating from natural gas furnaces has declined 34%. Other sectors
where major changes are occurring are residential water heating, domestic and international
aviation, and cement.

Applying the "Diff Query" to prices can also provide valuable insights. Prices are not additive (if
a good is $l/unit in state A, and $4/unit in state B, the sum of those values would not be
$5/unit), however, so it is recommended that the "Diff Query" be applied to prices for a single
state or region at a time.

Here, we are examining the change in "Building service costs" in North Carolina, using a
minimum percent difference of 15%. These results could be useful from a policy design
standpoint, particularly in determining how carbon tax proceeds could be allocated to reduce
impacts.

T-50


-------
jj[fj iLIMPSE Modellnterface [E:\Projects\GLIMPSE-5p4\GCAM-Model\gcam-v5.4\output\database3]	— ~

File Edit Table Help

Scenario

Regions

Queries

GC AM5p4-R ef-Oi ig 2023-l-5l|

MA a









—AN emissions by resource production a

GLIMPSEvl-Reference 2023-1

MD





B-L

Impacts

GLIMPSE v 1 -Tax -100C-5pct 2C

ME









* C02 concentrations



MI









« Total climate forcing



MN









« Global mean temperature



MO





b-T

Markets, prices, and costs



MS









4 C02 prices



MT









« Prices of all markets



NC









- Supply of all markets



ND









¦ • " « Demand of all markets



NE









—Final energy prices



NH









. Costs by tech and input



NJ













NM









* Costs of transport services



NV









« Elec prices by sector (GCAM-USA)



NY v









~ * Elec prices by sector (core) v

< >

< >



Run Query Diff Query 0 Total Collapse Update Single Queries Create Remove Edit

Q Diff: Outputs by tech ~ Diff: Prices of all markets ~ Diff: Building service costs

Filter Graph Format

scenario

region

sector

2015

2020

2025

2030

2035

2040

2045

2050

Units

GLIMPSEvl-...

Total

comm cooking

0.00

0.00

12.2

14.8

19.3

24.6

30.2

34.9

|pct (1975$/GJ)

GLIMPSEvl-...

Total

comm heating

0.00

0.00

10.2

12.4

15.9

19.8

23.4

26.5

pet(1975$/GJ)

GLIMPSEvl-...

Total

comm hot water

0.00

0.00

13.2

16.1

21.3

27.7

34.8

41.8

pet(1975$/GJ)

GLIMPSEvl-...

Total

comm non-building

0.00

0.00

14.1

15.3

19.6

21.8

24.7

25.4

pet(1975$/GJ)

GLIMPSEvl-...

Total

esid cooking

0.00

0.00

7.38

8.21

10.4

12.4

14.5

15.6

pet(1975$/GJ)

GLIMPSEvl-...

Total

resid heating

0.00

0.00

8.56

8.95

11.4

13.0

14.8

15.2

pet(1975$/GJ)

GLIMPSEvl-...

Total

1 esid hot water

0.00

0.00

8.53

9.41

12.0

13.8

15.8

16.6

pet(1975$/GJ)



Figure T4.5 Differences in energy service costs in buildings.

T4.3 Additional analysis suggestions

Use "Diff Query" to uncover to examine technology changes in the residential, commercial, and
transportation sectors.

If you apply the "Diff Query" to emissions by sector or emissions by technology, what do the
results tell you about potential low-hanging fruit for mitigating CO2 emissions?

Try graphing "value" and "percent" results of the "Diff Query".

T-51


-------
TUTORIAL 5: MODELING AN EV MARKET SHARE TARGET

T5.1 Overview

In this part of the tutorial, we use a combination of GLIMPSE'S "Market Share" and "Tech Avail"
features to simulate a scenario in which the EV sales share for onroad passenger cars and trucks
increases to 100% nationally by 2050.

T5.2 Constructing the EV sales target components

Implementing the electrification target requires two steps: representing the market share
constraint and adding a complementary that addresses numerical.

T5.2.1 Introducing a market share constraint

We will begin by creating EV market share targets for passenger cars and trucks.

In the Scenario Builder, click on " " button to open the New Scenario Component Creator
dialog, then click on the "Market Share" tab.

Select the pulldown menu next to "Type?". The menu shows several types of market share
constraints, including a Renewable Portfolio Standard (RPS), Clean Energy Standard (CES), and
EV targets for various vehicles classes and combinations, as well as options for light-emitting
diodes (LEDs) and heat pumps.

T-52


-------
Select One

A

Renewable Portfolio Standard (RPS)



Clean Energy Standard (CES)



EV passenger cars and trucks



EV passenger cars trucks and MCs



EV freight light truck



EV freight medium truck



EV freight heavy truck



EV freight all trucks



LED lights



Initial and Final % *

Figure T5.1 Options available in the "Type?" pulldown menu of the "Market Share" tab.

Select the "EV passenger cars and trucks" category.

This selection results in the "Subset" and "Superset" pulldown menus being populated. Check
the settings in those pulldown menus.

The "Subset" menu shows all "Car", "Large Car and Truck" and "Mini-Car" technologies, but
only the battery electric vehicle (BEV) options have check marks next to them.

The "Superset" menu shows the same set of technologies, but all have checked boxes.

The selected items on these menus specify numerator (subset) and denominator (superset) for
the market share constraint. The settings were chosen automatically based on the "Type"
selection. However, the specific technology selections can be customized. For example, if you
also want fuel-cell electric vehicles to be eligible to meet the market share constraint, you can
click the box next to the FCEV technologies in the subset menu.

Next, click on the "Constraint" pulldown menu. The options listed are "Lower" and "Fixed".
Choosing "Lower" indicates that you would like the constraint to be a lower bound, meaning
that GCAM can opt to exceed that percentage. "Fixed" requires GCAM to hit the specified
market share exactly, which can be more computationally challenging. Select "Lower".

Next, click on the "Applied to" pulldown menu. Here you have the option of applying the
constraint to "All Stock" or just to "New Purchases". Select "New Purchases" since we are
setting a sales target.

Then click on the "Treatment" pulldown menu. This allows the constraint to be applied "To
Each Region" or "Across Selected Regions". The latter provides GCAM with more flexibility since

T-53


-------
it may choose to have some states be above and others below the constraint as it seeks to hit
the target at the lowest cost. Choose "Across Selected Regions".

Next, we will select the regions that will be constrained. On the tree at the right side of the
dialog, click the triangle next to the "USA" region. This will expand the region to show the 50
states and DC. The tree allows you to select one or more constraints. In addition, the "Presets"
option at the bottom will automatically check the boxes next to specific regions, such as "New
England", "North America", or "Europe". Click on the check box next to "USA", which will select
all the states.

The next step is to populate the data table in the center of the dialog. One approach is to enter
a year and value in the text fields next to "Add", then pressing the "Add" button. Alternatively,
you can use the "Populate" options at the bottom left to add data to the table. We use this
approach in the tutorial.

Enter "2025" in the text field next to "Start Year".

Next, for "Initial %", enter "15", and for "Final %", enter "100".

Press the "Populate" button, which will add your data to the table.

T-54


-------
11 New Scenario Component Creator

XML List Pollutant Tax/Cap Tech Avail Market Share Tech Bound Tech Tax/Subsidy Tech Param Fuel Price Adj

Specification:

Type?

Subset:

EV passenger cars and trucks	~

trn_pass_road_LDV_4W: Car: BE... ~

Superset

trn_pass^road_LDV_4W : Car: BE...

-

Constraint:

Lower

-

Applied to:

New Purchases

-

T reatment:

Across Selected Regions

-

Names:

>/ Auto?



Policy:





Market:





Populate:





Type:

Initial and Final %

*







Start Year:

2025



End Year

2050



Initial (%):

15



Final (%):

100



Values:
Populate

Year

2025
2030
2035
2040
2045
2050

Select region(s):

Clear
Value

0.150
0.320
0.490
0.660
0.830
1.000

T — world



~ 0 USA



0

AK



0

AL



0

AR



0

AZ



0

CA



0

CO



0

CT



0

DC



0

DE



0

FL



0

GA



0

HI



0

IA



0

ID



0

IL



Select (optional)

Save	Close

Figure T5.2 Options chosen for the passenger vehicle market share constraint

Next, press the "Save" button to save this new policy representation to the Component Library.
Name the new component "EV-passenger-cars-and-trucks_AII_Reg_100x50.csv".

T5.2.2 Addressing numerical issues using Tech A vail

A next step could be to create a new scenario based upon GLIMPSEvl-Reference that includes
the new policy component. However, that scenario would experience solution problems in
2050 because of how the logit function assigns market shares in GCAM (see Chapter 3 for a
description of how the logit works). In summary, is not possible for GCAM to find a subsidy that
would achieve a 100% market share since technology costs are represented in the logit function
as distributions with infinite tails. Instead, we must also introduce a complementary measure
that eliminates non-EV technologies in that year.

We can use a "Tech Avail" scenario component for this purpose. 'Tech Avail" allows the range
of years over which a technology is available to be specified. For the years outside this range,
the technology shareweight is set to zero by GLIMPSE, effectively eliminating those
technologies from their respective markets.

T-55


-------
We start by filtering the list to show just the technologies within the passenger car and truck
sector. From the "Filter by Sector" pulldown menu, choose "trn_pass_road_LDV_4W".

The table is updated to show only the technologies available to the "Car", "Large Car and
Truck", and "Mini Car" categories. Only rows with checks in the "Never?" or "Range?" columns
are being constrained, so currently all the boxes are unchecked.

Clicking on the "Never?" checkbox the technology in that row is not available in any year.
Clicking on "Range?" results in the technology only being available between the "First" and
"Last" years that are specified, inclusive of those years.

Start by clicking the "Range" box next to all of the non-BEV technologies. Next, enter "2045" in
the "Last yr" textfield near the bottom and press "Set Years". This will update the years for all
technologies visible in the table.

Next, click on the checkbox next to the "USA" region.

Your dialog should look like this.

IB New Scenario Component Creator	X

XML List Pollutant Tax/Cap Tech Avail Market Share Tech Bound Tech Tax/Subsidy Tech Param Fuel Price Adj
Select technologies and specify all, first, or last years to constrain new purchases:

Filter by Sector: trn_pass_road_LDV_4W	I Text:

Never?

Range?

First

Last



Sector : Subsector : Technology : Units

~



1975

2045

trn_pass_road_LDV_4W

Car : BEV : million pass-km

-



0

1975

2045

trn_pass_road _LDV_4W

Car : FCEV : million pass km





0

1975

2045

trn_pass_road_LDV_4W

Car : Hybrid Liquids : million pass-km





0

1975

2045

trn_pass_road_LDV_4W

Car : Liquids : million pass-km



~

0

1975

2045

trn_pass_road_LDV_4W

Car : NG : million pass km







1975

2045

trn_pass_ road_ LDV_ 4W

Large Car and Truck : BEV : million pass km





0

1975

2045

trn_pass_road _LDV_4W

Large Car and Truck : FCEV : million pass km





~

1975

2045

trn_pa ss_ road_LDV_4W

Large Car and Truck : Hybrid Liquids : million pass !





0

1975

2045

trn_pass_road_LDV_4W

Large Car and Truck : Liquids : million pass km





0

1975

2045

trn_pass_ road. LDV_ 4W

Large Car and Truck : NG : million pass-km







1975

2045

trn_pass_road..LDV_4W

Mini Car : BEV : million pass-km





0

1975

2045

trn_pass_road^LDV_4W

Mini Car : FCEV : million pass-km



~

0

1975

2045

trn_pass_ road _ LDV_4W

Mini Car : Hybrid Liquids : million pass km





0

1975

2045

trn_pass_road_LDV_4W

Mini Car : Liquids : million pass km





IT71

1Q 75

7045

trn na« rrvarj I HV 4W

Mini f~ar • Nlf^ • million na«-Vm



rr









"	;	j >

Select region(s):

~ — world
~ g] USA

Canada

Central America and Caribbean
J Mexico
I Brazil

South America_Northern
South America_Southern
Argentina
Colombia
Q EU-15
Q EU-12

European Free Trade Association

Europe_Eastern

Europe_Non_EU

Africa ^Eastern

Africa_Northern

Presets:

Select (optional)

Select:

Range

First yr: 1975

Last yr: 2045

Save	Close

Figure T5.3 "Tech Avail" setup to eliminate non-EVpassenger car and trucks in 2050.

T-56


-------
Finally, press "Save" to save this new component to the Component Library. Name this
component "EV-passenger-cars-and-trucks All Reg 100x50-part2.csv".

Add both "EV-passenger-cars-and-trucks_AII_Reg_100x50.csv" and "EV-passenger-cars-and-
trucks_AII_Reg_100x50-part2.csv" to GLIMPSEvl-Reference, then name the new scenario
"GLIMPSEvl-PassEV100x50".

Create the new scenario by pressing



When you are ready, add the scenario to the execution queue by pressing " After the
scenario execution completes successfully, move on to the next section of the tutorial.

T5.3 Verifying the performance of the policy

First, we will verify that the new policy files achieved their objective. Select the "GLIMSPEvl-
PassEV100x50" scenario, select all states and the "USA" region, select the "Transport service
output by tech (new capacity)", click on the check box next to "Total", then press "Run Query".

After several moments the table will be populated with data representing new sales for each
transportation subsector (in units of capacity, which are million pass-km for passenger vehicles
and million ton-km for freight).

We are interested in the "Car" and "Large Car and Truck" categories, so use the "Filter" option
to view just those subsectors. Note that the USA region of the model does not currently include
"Minicar", so there are no results for that subsector.

Click "Graph" to visualize the data and choose "StackedBarChart" as the format.

T-57


-------
OM GLIMPSE Model Interface [E:\Projects\GLIMPSE-5p4\GCAM-ModeI\gcam-v5.4\output\database3]	— ~

File Edit Table Help

Scenario

Regions	

Queries

GCAM5p4-Ref-Orig 2023-1-5T07:57:08-04:00



USA









. Electricity generation by subsector (GCAM-USA) a

GLIMPSEv 1 -R ef erence 2023-1-5T09:23:06-04:00



Africa_Eastern









- Electricity generation by tech and cooling (ind cog

GLIMPSEv 1 -Tax-100C-5pct 2023-l-5T12:21:28-04:00



Africa_Northern









- Electricity generation by tech and cooling (ind cog

GLIMPSEv 1 -PassEV100x50 2023-4-5T15:38:04-04:00



Africa_Southei n









« Electricity generation by cogen only (GCAM-USA)





Africa_Western









- Electricity generation input by subsector (GCAM-U





Australia_NZ









4 Electricity generation by region (ind cogen)(core)





Brazil







« Electricity generation by subsector (core)





Canada







* Electricity generation by gen tech (core)





Central America and Caribbean







- Building final energy by tech





Central Asia







« Building service output by tech





China







• Industry final energy by tech and fuel





EU-12







Refined liquids production by tech (GCAM-USA)





EU-15







- Refined liquids production by tech (core)





Europe_Eastern







. Hydrogen production by tech





Europe_Non_EU







- Transport final energy by tech and fuel





European Free Trade Association

India

Indonesia







Transport service output by tech
- Passenger car and truck service output by tech





Japan







. Freight truck service output by tech (no bus) v





Mew o
Middle East

Palrfehan



h	

>





V

| Run Query Diff Query 0 Total Collapse Update

Q Transport service output by tech (new capacity)

Filter Graph Format



1 More Display

1 A
7 V

StackedBarC... v I~1 Same Scale B Refresh

sc... re...

su... te... 2...



2... 2030 2... 2040 2... 2050 U...









GLI...

lotal

Car

BtV

0.00

8440U

28b...

622...

94b...

141...

184...



milli...

GLIMPSEv1-PassEV100x50
Car
region: Total

"p

GLIMPSEv1-PassEV100x50
Large Car and Truck
region: Total

E 3.500.000 _

(J> 3.000.000 1
m

2. 2.500.000 ¦
o 2.000.000 1

| 1.500,000 1 ¦¦¦¦.
~ 1.000.000 1
B 500.000 l|l

to n to to to to n to
OOOOOQOO

(Sooioiji^o

GLI...

Total

Car

FCEV

0.00

2470

14000

73300

68200

65900

38700

0.00

milli...

GLI...

Total

Car

Hyb...

7030

414...

689...

565...

382...

241...

97100

0.00

milli...

GLI...

Total

Car

Liqu...

417...

114...

873...

659...

439...

266...

104...

0.00

milli...

GLI...

Total

Lar...

BEV

0.00

67100

219...

497...

674...

854...

102...

199...

milli...

GLI...

Total

Lai...

FCEV

0.00

2060

12600

69100

68600

76600

62200

0.00

milli...

GLI...

Total

Lar...

Hyb...

0.00

322...

538...

445...

323...

235...

131...

0.00

milli...

-f 4.000.000 ¦
v>

s

2. 3.000.000 ¦

1

= 2.000.000 ¦

§ 1

= 1.000.000 ¦

f 1

0*

w
a

Ol



GLI...

Total

Lar...

Liqu...

347...

916...

729...

568...

409...

291...

159...

0.00

milli...



Ifl'll

M M M M U M H

o o o o o o o
o Ch o (Ji & & o

Figure T5.4 Sales (in units of million pass-km) for the "Car" and "Large Car and Truck"
transportation subsectors.

By examining the sales data, we can verify by that target was met.

Table T5.1 On road EV sales shares by subsector and total

Category

2020

2025

2030

2035

2040

2045

2050

Car

5%

15%

32%

52%

71%

88%

100%

Large Car and Truck

5%

15%

31%

46%

59%

74%

100%

Total

5%

15%

32%

49%

66%

83%

100%

T-58


-------
Note that while the overall sales target is met in each year, "Car" EV sales are greater than
"Large Car and Truck". The bars for 2015 are considerably higher than the others because 2015
is the model's calibration year and new capacity added that year is equivalent to the entire
stock.

Some interesting dynamics occurring in 2040 through 2050 in the "Large Car and Truck"
subsector. We applied our constraint across the "Car" and "Large Car and Truck" categories.
However, there are differences in the costs of electrifying vehicles in these two categories,
resulting in GCAM choosing to electrify the "Car" category at a level higher than the target and
the "Large Car and Truck" at a level below the target. Furthermore, our constraint was applied
based on capacity in units of "million pass-km", but cars are assumed to have an average
ridership of 1.58 people per vehicle, while the "Large Car and Truck" ridership is assumed to be
1.66 people per vehicle. Finally, note that our policy representation resulted in a small increase
in the cost of onroad light duty travel, leading to a small amount of mode switching to buses
and motorcycles, as well as a small decrease in demand for passenger travel overall.

T5.4 Exploring the response to the EV market share target

Based upon what you have learned through the course of these tutorials, explore the answers
to some of the following questions:

•	How is electricity demand changing under this scenario?

•	How is the additional electricity being produced?

•	How is demand for refined liquids changing across sectors? Natural gas?

•	How is refinery output changed? What is the impact on biofuel production?

•	What are the impacts on the price of electricity, refined liquids, and natural gas?

•	How are CO2 emissions impacted overall and by sector? How about air pollutants?
Remember that it does not make sense to sum prices across states or regions.

T-59


-------
v>EPA

United States
Environmental Protection
Agency

PRESORTED STANDARD
POSTAGE & FEES PAID
EPA

PERMIT NO. G-35

Office of Research and Development (8101R)
Washington, DC 20460

Official Business
Penalty for Private Use
$300

Recycled/Recyclable Printed on paper that contains a minimum of
50% postoonsumer fiber content processed chlorine free


-------