GCAM: An overview and applications

Pralit Patel

3/17/2020

PNNL is operated by Battelle for the U.S. Department of Energy

Outline

• Broad overview
• Where does this style of modeling fit in overall community
• Sponsors and research areas
• Modeling details
• Some examples

GCAM explores the interactions between multiple systems

The Global Change Assessment Model (GCAM)

• GCAM is a global integrated assessment model
• GCAM links Economy, Energy, Land, Water, and Climate systems
• Typically used to examine the effect of technology and policy on the economy, energy system, agriculture and land-use, and climate
• Technology-rich model
• Emissions of 24 greenhouse gases and short-lived species: CO₂, CH₄, N₂O, halocarbons, carbonaceous aerosols, reactive gases, sulfur dioxide.
• Runs through 2100 in 5-year time-steps.
• Open source: https://github.com/jgcri/gcam-core
• Documentation available at: http://jgcri.github.io/gcam-doc/

Global Coverage

384 Land Regions

32 Energy & Economy Regions

235 Water Basins

Reduced-Form ESM (Hector)

GCAM Produces Conditional Forecasts

Scenario Assumptions

Model Equations, Relationships, and Parameters

Modeled Scenario

The GCAM ecosystem is a suite of models and tools

GCAM DS

Data Development

Dynamic Integration

Disaggregation

32 Energy Economy Regions

235 Water Basins

384 Land Regions

Province-Level Energy Economy Regions

Moirai

Single System

Hector

Xanthos

Demeter

Tethys

Statistical Emulators

fldgen

Persephone

GCAM-core

7

Background

Modeling Community

• One of a handful of “Integrated Assessment Models”
• Has been included in every IPCC assessment

8

Modeling Community

• One of a handful of “Integrated Assessment Models”
• Has been included in every IPCC assessment
• Produced the scenarios for “RCP 4.5”

9

Modeling Community

• One of a handful of “Integrated Assessment Models”
• Has been included in every IPCC assessment
• Produced the scenarios for “RCP 4.5”
• Involved in planning and producing “SSP” scenarios

10

Sponsors and research areas

• US Department of Energy – Office of Science
• Other US government agencies
• Non-government partners and supports
• World Bank, Exxon, European Commission, NIES

11

State Department

US EPA

US AID

Non-Governmental Partners and Sponsors

(World Bank, Exxon, Chevron, European Commission, NIES, ….)

Department of

Defense

Home Institute

• Joint Global Change Research Institute
• Pacific Northwest National Laboratory
• University Of Maryland
• Staff of over 50 people
• From Senior Scientist to Undergraduate Interns
• Pretty diverse group

12

An Expanding GCAM Community

• In 2010, PNNL converted GCAM into a community model
• We have held an annual community modeling meeting since 2010 with a continuous steady growth in attendance, including more than 150 participants at the 2019 meeting

13

14

Modeling Details

Market Equilibrium in GCAM

ENERGY DEMANDS

AGRICULTURAL DEMANDS

WATER DEMANDS

MARKETS

(& Transformation sectors and intermediate demands)

ENERGY SUPPLIES

LAND

WATER SUPPLIES

PRICES

QUANTITIES

PRICES

QUANTITIES

ECONOMIC ACTIVITY

CLIMATE

Under development

Under development

Energy System flow

16

Logit Choice Mechanism for Decision Making

• Calibrated logit approach assumes a distribution of realized costs due to heterogeneous conditions.

​

• Market share based on probability that a technology has the least cost for an application.
• Avoids a “winner take all” result.

​

• Historical calibration influences future competition through the “share-weight” (α)

17

Land Supply

Solution Approach

20

Future Research Directions

Mission statement

• Improve understanding of the complex interactions among energy, water, land, climate, socioeconomics, and other important human and natural systems
• Captures the co-evolution of the human – Earth system

Heat-wave

Drought

Wildfires

Sea Level Rise

Flooding

Permafrost

How can we get at these influences in GCAM

22

GCAM DS

Data Development

Disaggregation

32 Energy Economy Regions

235 Water Basins

384 Land Regions

Province-Level Energy Economy Regions

Moirai

Single System

Hector

Xanthos

Demeter

Tethys

Statistical Emulators

fldgen

Persephone

GCAM-core

Drought

Heat-wave

Yield

Wildfires

Flooding

Does AI or machine learning help?

• Train our emulators
• Predicting “rare” events
• Scenario discovery

​

23

Lamontagne et al. (2018), Large Ensemble Analytic Framework for Consequence‐Driven Discovery of Climate Change Scenarios, Earth’s Future.

​

Snyder et al.. (2019). fldgen v2.0 software description PloS one, 14(10).

​

Discussion

PNNL is operated by Battelle for the U.S. Department of Energy