MESSAGEix-Canada covers all energy sectors in the economy, including oil & gas, power, buildings and transportation, and is designed to investigate long-term energy pathways and comprehensively assess energy policy packages. With its whole-energy system coverage, the model is well suited to answer policy and research questions that require feedbacks between downstream and upstream sectors to be accounted for (e.g., demand response).

MESSAGEix-Canada is built off of the MESSAGEix framework, a linear programming modelling framework developed by the International Institute for Applied Systems Analysis in Austria.

Key features of MESSAGEix-Canada:

• Provincial / territorial resolution: All provinces and territories are represented as separate nodes in the model, allowing for policy and pathway assessments at both the federal and provincial / territorial levels.
• Least-cost optimization: Since the model seeks a minimal cost solution to the policy and technology constraints provided to it, MESSAGEix-Canada can complement the use of models like CIMS, which account for behaviour, but do not give insight to the user on the most efficient pathway that satisfies the policies / other constraints provided by the user.
• Detailed representation of hydrogen: significant work has gone into developing and calibrating the model’s representation of hydrogen production, transmission & distribution and use. The model currently includes 20+ hydrogen technologies, with additional work on improving the representation of end use technologies planned.

Policy / research questions MESSAGEix-Canada is designed to explore:

1. How does a policy package being proposed by the federal government impact emissions?
2. What policies and technologies will most be needed in the 2030s and 2040s to ensure Canada reaches net-zero?
3. Is there a credible net-zero scenario for Canada by 2050 that involves significant oil and gas production increases? What would be the preconditions for this?
4. If Canada adds a production tax credit to hydrogen production, such as $​0.5-3 USD​ per kg of H2 to match the US policy, what will be the impact on H2 production and emissions? Where would the H2 be produced, and where would it be used?

CIMS is an integrated energy-economy-emissions model designed to simulate the interaction of energy supply-demand and estimate the effects of policy and market dynamics on Canada’s energy system. To do so, CIMS represents detailed energy-related technologies and the realistic preferences of consumers and firms across supply sectors such as oil & gas, as well as key energy demand end-uses in residential, commercial/institutional, transportation, and various industrial sectors.

The model is being open-sourced by the Energy & Materials Research Group at Simon Fraser University, in collaboration with the Open Insights project at the University of Victoria.

Key features of CIMS

• Whole energy system coverage: CIMS covers all energy-intensive industries, including personal / freight transportation, residential / commercial buildings, coal mining, natural gas and crude oil production, electricity generation, various light industrial sectors, petroleum refining, industrial minerals, mining, pulp & paper, and iron & steel.
• Behaviourally realistic: Instead of basing its simulation of technology choices only on financial information, CIMS differentiates itself from conventional bottom-up analysis by including non-financial costs, which reflect real-world consumer and business preferences.
• Bottom-up & technology-rich: Through a ‘stock-flow’ model, CIMS simulates new purchases, retrofits, and retirements over time for many technologies. The level of detail can be as granular as the type of lightbulb used in residential buildings.

Policy / research questions CIMS is designed to explore:

1. How might prospective market dynamics or energy and climate policies impact energy supply-demand, GHG emissions, and techno-economic outcomes?
2. What policies are required to meet economy-wide or sector GHG emissions targets?
3. What are the costs of different energy sources in 2050 under net-zero scenarios?

The M3 EnergyABM is an agent-based model designed to simulate the strategic investment decisions of energy companies, their profits or losses, and how they respond to policies and shocks. For the oil and gas sectors, the model represents individual firms in Canada, with foreign producers aggregated up to country-level cost curves. For the electricity sector, the heterogeneity of the various provincial electricity markets is represented by vertically integrated firms representing government-owned public utilities in most provinces and competitive markets in Ontario and Alberta. Electricity trade with the USA is represented by a statistical model, which can assess the trade impacts of changes in total USA demand and variable renewable penetration on the USA grid.

The model is being developed by Macrocosm Group, a company formed by researchers at the University of Oxford, in collaboration with the Open Insights project at the University of Victoria.

Figure 1: Schematic of the M3 EnergyABM. The model consists of two types of agents: firms and plants. Firms own and operate plants, make forecasts and invest into new assets. Plants produce extract/produce electricity, bid/sell in energy markets and face evolving costs.

Policy / research questions the M3 macromodel is designed to explore:

1. What impact will different production/generation tax credits have on capacity expansion in different Canadian energy markets?
2. How will oil and gas companies in Canada respond to various climate policies such as an increasing price of carbon?
3. How competitive (profits, cost structure) are Canadian oil & gas firms relative to global firms over the transition period to net-zero?

Why build an agent-based model of the energy sector?
Agent-based models (ABM) have several properties that make them particularly strong at analyzing the impacts of the energy transition:

• ABMs account for heterogeneity of economic agents and capture nonlinear and dynamic interactions between agents
• ABMs realistically incorporate institutional structures and thus model the economy “as is” (rather than “as if”), avoiding strong assumptions of equilibrium, market clearing and perfect rationality.
• Performance can be easily tested against historical economic data.
• ABMs track balance sheets over time, allowing for granular as well as aggregate insights.

Timeline:

A functioning version of the model is already developed but is missing several key features. These features will be added over the coming months, after which the model will be open-sourced.

The M3 macromodel is an agent-based model designed to simulate the economic growth, trade, household and labour impacts of provincial and federal climate policies. The model directly represents all major economic agents and their actions, 50+ sectors, all 10 provinces, and interprovincial and international tradeFigure 1. The number of firms and households the model runs with is defined by the user, making it straightforward to account for the diversity of economic agents. A dedicated calibration pipeline has been developed alongside the model, meaning that the model can be quickly recalibrated and tested against historical data when new data or features are added.

The model is being developed by Macrocosm Group, a company formed by researchers at the University of Oxford, in collaboration with the Open Insights project at the University of Victoria.

Policy / research questions the M3 macromodel is designed to explore:

1. How does an increase in the carbon price affect Canadian economic indicators such as household budgets, jobs, and trade?
2. How could the global transition impact Canadian jobs?
3. What is the economic impact of the existing tariff on Chinese EV imports?

Why is the M3 macromodel an agent-based model?
Agent-based models (ABM) have several properties that make them particularly strong at analyzing the impacts of the energy transition:

• ABMs account for heterogeneity of economic agents and capture nonlinear and dynamic interactions between agents
• ABMs realistically incorporate institutional structures and thus model the economy “as is” (rather than “as if”), avoiding strong assumptions of equilibrium, market clearing and perfect rationality.
• Performance can be easily tested against historical economic data.
• ABMs track balance sheets over time, allowing for granular as well as aggregate insights.

Timeline:

A functioning version of the model is already developed but is missing several key features. These features will be added over the coming months, after which the model will be open-sourced. If you are interested in obtaining early access and/or collaboration, please reach out to [email protected].

The M3 LabourABM is an agent-based model designed to understand the response of labour markets to the energy transition. The model estimates the unemployment rate and vacancy rate by occupation and sector, accounting for occupational and regional mobility of workers. Crucially, the model has been extensively calibrated, meaning that the costs to a worker of switching occupations and/or moving provinces has been carefully aligned with provincial historical data.

The model is being developed by Macrocosm Group, a company formed by researchers at the University of Oxford, in collaboration with the Open Insights project at the University of Victoria. The model is designed to be used in conjunction with the M3 Macromodel. The Macromodel generates employment shocks which can be fed into the LabourABM, where impacts on the unemployment and vacancy rate are calculated, accounting for more labour dynamics than the Macromodel is capable of on its own.

Figure 1: Network of occupations in the LabourABM and exposure to external employment shock (in this figure the shock is due to automation).

Policy / research questions the M3 LabourABM is designed to explore:

1. What are the job / wage impacts of across regions and occupations of different climate policies?
2. What impact will different job training programs have on reducing frictions of job transitions from fossil fuel to renewable industries?
3. How could the global transition impact Canadian jobs (requires linkage to the Macromodel)?

Why is the M3 LabourABM an agent-based model?
Agent-based models (ABM) have several properties that make them particularly strong at analyzing the impacts of the energy transition:

• ABMs account for heterogeneity of economic agents and capture nonlinear and dynamic interactions between agents
• ABMs realistically incorporate institutional structures and thus model the economy “as is” (rather than “as if”), avoiding strong assumptions of equilibrium, market clearing and perfect rationality.
• Performance can be easily tested against historical economic data.

Timeline:

A functioning version of the model is already developed but is missing several key features. These features will be added over the coming months, after which the model will be open-sourced. If you are interested in obtaining early access and/or collaboration, please reach out to [email protected].

M3’s technology cost models are a collection of data-driven methods designed to forecast the future costs of different energy technologies under certain deployment and investment scenarios. Models have been developed for all major clean energy technologies for which there is sufficient historical data, including: wind, solar, geothermal, small-modular reactors, electrolyzers, electric vehicles, bioelectricity and heat pumps.

These models have been developed by Macrocosm Group, a company formed by researchers at the University of Oxford, in collaboration with the Open Insights project at the University of Victoria. These models are intended to be used as inputs into power system and energy system models, enabling modelers to explore the sensitivity of their models to realistic alternative technology cost trajectories.

Figure 1: Example forecasts for Solar PV and Wind. As illustrated, the models account for the inherent uncertainty in these projections.

Key features of the technology cost models:

• Calibrated on historical data: Recent studies have shown that many energy-economy models historically underestimate deployment rates of clean technologies and overestimate costs. In contrast, the methods used here perform better at forecasting technology cost evolution1.
• Captures uncertainty: These methods are designed to easily produce many possible forecasts, the range of which is directly based off the model’s success in reproducing the historical cost evolution.
• Open-source and easily extendable: All the methods for calibrating these models will be released open-source, meaning that models for new technologies can be easily calibrated once the required data to do so is compiled.

Policy / research questions the M3 technology cost models are designed to explore:

1. How will evolving technology costs impact net-zero energy system mixes of the future? (requires linkage to an energy system model, such as CIMS or MESSAGEix-Canada)
2. What is the impact on deployment of cleantech of increasing the carbon price at different paces to 2050? (requires linkage to an energy system model)
3. What are the costs of different energy sources in 2050 under different net-zero scenarios, and what drives the differences? (requires linkage to an energy system model)

Timeline:

These models are complete and are in the process of being open-sourced. If you are interested in obtaining early access and/or collaboration, please reach out to [email protected].