An assessment of energy system transformation pathways to … – Nature.com

The Swiss TIMES energy systems model (STEM)

The Swiss TIMES Energy Systems Model (STEM)²⁰ is based on The Integrated MARKAL-EFOM System (TIMES) framework developed by the International Energy Agency’s Energy System Technologies Analysis Program (IEA-ETSAP)⁵⁴. STEM is a partial equilibrium framework representing the entire energy system, from an engineering techno-economic perspective, with a full range of energy processes, including exploitation, conversion, transmission, distribution, storage, and energy end-use. As STEM is a bottom-up technology-rich model, it does not explicitly include broader socio-technical or political perspectives of the energy system and energy transition⁵⁵. The model consists of the following sectors of the energy system:

• Upstream sector, including the production and extraction of resources, energy imports and exports

• Conversion sector that includes electricity and heat production, hydrogen production, production of biofuels and e-fuels

• Final energy consumption sectors, including industry, services, residential, transport and agriculture

• CO₂ capture, sequestration and utilisation sectors

STEM aims to supply energy services at a minimum energy system cost by simultaneously making energy technology investments and operating decisions to reach a demand-supply equilibrium state. The equilibrium found by STEM has the property that the total surplus is maximised. STEM has a long-term horizon. This analysis focuses on the developments until 2050, with reporting years 2020, 2030, 2040 and 2050. We apply a 2.5% social discount rate and technology-specific interest rates of up to 5.5%, depending on the sector and technologies. In each year, 288 operating hours are considered, representing four seasons, three average days per season (viz. working day, Saturday and Sunday), with 24 h resolution.

For each energy service demand sector and energy use, the model considers hourly profiles, which are given exogenously²⁰, but the hourly energy consumption needed to meet the hourly energy service demand is fully endogenous⁵⁶. Based on construction year and size, different residential building classes are distinguished. The transport sector includes passenger and freight transport, with different modes and many types of vehicles for each mode. Stock-based material flow analysis approach for the evolution of building stock⁴⁵, discrete choice methods for appliances, Gompertz functions for car ownership, etc., are used to project energy service demands for the subsectors according to the scenarios’ macroeconomic and demographic developments. Climate change impacts on heating and cooling demands, as well as on energy resources (e.g., water), are explicitly accounted for based on the climate scenarios that are consistent with the energy scenarios⁴¹.

STEM includes options for producing synthetic fuels, and it explicitly represents the storage, transport and distribution required for the secondary energy carriers. A particular emphasis was given to the electricity sector to address systems with high shares of variable renewable energy. Four electricity grid transmission and distribution levels are represented, from very high to low voltage³³. STEM includes direct current (DC) power flow model modelling of the electricity transmission grid with 15 nodes and 319 bi-directional lines^57,58. Technical operating constraints for the hydrothermal power plants are approximated via a continuous relaxation of the unit commitment problem⁵⁹. They include ramping rates, minimum stable operating levels, part-load efficiency losses, minimum online and offline times and start-up costs. While it is unsuitable for analysing actual system operation, it is valuable for including short-term decisions in long-term planning. The energy supply and use assets can be retired before the end of their technical lifetime when they have higher fixed or operating costs than the investment cost of new alternatives⁶⁰.

While energy service demands are determined exogenously based on macroeconomic and demographic developments, the model endogenously considers energy savings measures for each end-use to reduce energy consumption. These include the switch to a more efficient technology or another energy carrier like electricity, building renovations (walls, windows, floors, roof), adopting waste heat recovery practices, optimisation of motor drive operations and other industrial equipment, investments in specific industrial processes and measures that reduce consumption, and many others. Supplementary Information Note 1 elaborates a list of the technical measures in the model that characterises the demand reduction in our modelling framework. The costs for deploying such measures are fully accounted for in the objective function when minimising the total energy system cost.

Though STEM includes many technical and policy measures, changes in consumer behaviour or lifestyles, such as lowering indoor temperatures, are not considered because of the lack of reliable costs associated with the shift in consumer behaviour (e.g., discomfort costs). STEM is a bottom-up techno-economic partial equilibrium modelling framework. In this regard, consumer behaviour and acceptance of new technologies (in the meaning of openness or willingness to invest) is approximated by side constraints regarding the deployment level of these technologies or the access to the remaining exploitable and sustainable renewable energy potential.

STEM endogenously accounts for the stochastic variation of renewable energy availability, demand, and the residual load curve⁶¹. Besides energy markets, the requirements for primary and secondary operating reserves are endogenously modelled with explicit representation of ancillary services markets⁶². Hence, STEM includes both market- and technical-based flexibility mechanisms. For example, STEM represents energy storage, demand-side management schemes in the end-use sectors, grid-to-vehicle and vehicle-to-grid services from electric vehicles and detailed Power-to-X pathways³¹. The energy storage options are at different time scales and sizes, ranging from daily storage options to weekly and seasonal storage options³¹. In addition, we distinguish between centralised and large-scale storage (e.g. pumped hydro and large stationary batteries) and distributed storage (e.g. batteries and heat storage). The demand-side management schemes enable the shifting of electricity loads related to heating, cooking, washing, etc., also based on discomfort costs, shifting windows, and maximum load constraints derived from Swiss consumer surveys^63,64,65 and international studies^66,67.

The cross-border electricity trade in STEM is modelled through exogenous hourly electricity import price profiles. These profiles are based on interfacing STEM with complicated electricity systems and grid simulation models for the transmission grid of Switzerland and Europe, assuming European developments align with the ENTSO-E TYNDP scenarios⁴¹. Additionally, trade patterns concerning the potential of Switzerland to export excess electricity to the neighbouring countries are considered, which result from transmission grid reliability constraints and other power plant dispatching constraints in Switzerland and its neighbours.

Regarding modelling the district heating infrastructure, we apply an approach that considers the aggregate of the corresponding network infrastructure, distinguishing between existing and newly built equipment. When a pipeline comes to the end of its life, it can be replaced with a new investment or refurbished. In this regard, STEM does not allow disinvestments in the district heating network. However, the model ensures that the overall annualised investment and operating, and maintenance costs of the district heating infrastructure are covered in each year of the projection period. The connections to district heating of existing buildings⁶⁸ are explicitly modelled in STEM. When existing buildings are demolished, their connections can be used by new buildings built in the same locations. The development of the building stock is based on a detailed dynamic stock model. The model was developed in cooperation with our partners in this project from the University of Geneva⁴⁵. New buildings can also require new district heating connections, depending on the overall energy system cost minimisation. When the new connections exceed the current capacity of the process, then a new investment is made. The overall district heating potential is limited in the model at around 18 TWh/yr.⁶⁹.

Particular emphasis has been given to STEM for modelling battery electric vehicles. The charging and discharging profiles of electric vehicles are endogenously determined in STEM. We distinguish four consumer segments with different driving profiles and, consequently, charging and discharging needs. For each consumer segment, there is also a set of constraints to guarantee that the electric vehicle cannot be used if there is not enough energy in the battery to perform the requested trip at a given time. With the consumer segmentation and the set of constraints for having enough energy in the battery to complete the car trips, we try to improve the realism of the resulting charging and discharging profiles. The modelling of electric vehicles in STEM allows each vehicle’s battery to provide electricity back to the grid (at a cost reflecting the necessary electronic equipment to achieve the reverse electricity flow). Thus, the estimation of the number of electric cars participating in vehicle-to-grid schemes is calculated from the model results. It is based on the driving, charging and discharging profiles of the different consumer segments, the amount of electricity these segments give to the grid, and the total number of cars in each segment—all these aspects are endogenously determined by STEM based on the mobility energy service demands³².

The emissions included in STEM are CO₂ from fuel combustion and CO₂ from the cement production process, which are generated by decomposing raw materials. The model does not include the CO₂ emissions from international aviation and navigation in the overall mitigation target since these are not included in the Swiss National Determined Contribution and are not part of the Paris Agreement of 2015. STEM also does not calculate other non-CO₂ Kyoto GHGs related to the following sectors: from agriculture related to livestock, manure management, agricultural soils, liming, and urea applications; from the waste sector on solid waste disposal, biological treatment of solid waste, incineration and open burning of waste (other than the waste consumption for cogeneration of electricity and heat) and shredding; from fire damages; from land use, land use change and forestry.

STEM lacks endogenous mitigation options and technologies for non-CO₂ GHG emissions. Marginal Abatement Cost Curves (MACCs) can approximate emissions reductions from N₂O, CH₄, SF₆, etc., but these would need to be exogenously calculated based on complementary modelling frameworks and studies. In the context of the present study, performed within the project Swiss Competence Centre for Energy Research Joint Scenario and Modelling (SCCER JASM), such complementary modelling competencies to provide STEM with these MACCs were not available. Hence, we focused on the transformation of energy systems to achieve net-zero CO₂ emissions. This implies that the remaining non-CO₂ emissions in 2050, around 5.7 Mt CO₂-eq if the targets for reducing the non-CO2 GHG stated in the Long-Term Climate Strategy of Switzerland⁴ are met, should be compensated abroad in our analysis, as noted in the Swiss National Determined Contribution²⁶.

Most of the 5.7 Mt CO₂-eq that remain in 2050 and compensated abroad are emitted from the agriculture sector. This sector does not have mitigation options in STEM other than those for emissions from fuel combustion. Agriculture accounted for about 14% of the total GHG emissions in Switzerland in 2019, and in the Long-Term Climate Strategy of Switzerland the sector accounts for more than 95% of the remaining GHG emissions in 2050. We include the CO₂ emissions from fuel combustion in agriculture (about 10% of the total emissions from agriculture) but not the non-CO₂ GHG that are predominately related to biological and biophysical processes and come from diffuse sources or sources that fluctuate in terms of time and space. The complexity of the Swiss agriculture sector, with many small farms and various business operations, challenges the development of generally applicable mitigation measures—an argument also noted in the Long-Term Climate Strategy of Switzerland⁴. Given that more data and research is needed for identifying mitigation options for the GHG emissions from agriculture, and given that the Swiss climate change mitigation policy relies on the use of Internationally Transferred Mitigation Options for offsetting these emissions, we opted to exclude non-CO₂ emissions in our analysis. However, in an ongoing project, we have established collaborations with researchers from the agriculture and forestry sectors to assess potential mitigation options for the sector.

Finally, CO₂ domestic and cross-border pipelines are also considered for transferring captured emissions. As stated above, additional details about the structure and features of STEM are provided in Supplementary Information Note 1.

Scenario design

The scenario analysis was performed within the Swiss Competence Centre for Energy Research (SCCER) Joint Activity Scenarios and Modelling (JASM)⁷⁰ by using the Swiss TIMES energy systems model (STEM) of Paul Scherrer Institute. The SCCER energy programme²⁵ is a collection of eight university-networked research activities in Switzerland from 2013 to 2020 on energy efficiency, energy grids, electricity supply, economy, society, energy behaviour, mobility and biomass resources and technologies. It was funded by InnoSuisse, aiming to provide solutions to the technical, social and political challenges of the Swiss energy transition. In JASM, modelling teams from the eight SCCERs worked together to analyse the technical feasibility of energy transformation pathways to net-zero. This approach is different from other efforts to define long-term energy strategy for Switzerland, as it is based on the combined output of the Swiss energy systems research community. In JASM, a framework that soft-links various energy systems and sectoral technical-economic models from the project partners was developed. All models used a consistent set of drivers and scenario definitions⁴¹, derived from the Swiss Administrative Federal Offices.

In the context of the SCCER JASM, STEM is enhanced via the collaboration with the other modelling teams regarding renewable potentials, import prices, the potential for domestic sequestration of CO₂, cross-border trade infrastructure, characterisation of electricity, hydrogen, biofuel and synthetic fuel production technologies, the definition of end-use technologies such as boilers, appliances and vehicles, climate change impacts on energy demand and energy resources (e.g. water), and side-constraints reflecting consumer energy behaviour and social acceptance. The links between STEM and the rest of the sectoral models and approaches employed in SCCER JASM are detailed in Supplementary Information Note 2.

STEM explores possible energy futures based on contrasted scenarios. Unlike forecasts, scenarios do not presuppose knowledge of the main drivers of the energy system. Instead, they consist of a set of coherent assumptions about the future trajectories of these drivers, leading to a cohesive organisation of the energy system under study. A scenario builder carefully tests the hypotheses for internal consistency. A scenario definition requires at least four inputs: energy service demands, resources and potentials, policy settings, description and availability of technologies.

In developing the SCCER JASM scenarios, we examined two key methodologies:⁷¹ (a) using input from knowledge carriers such as scientific experts or stakeholders; (b) using intuitive logics (IL) or scenario axes. We also considered using the newly developed Cross Impact Balance (CIB) approach⁷¹ that enables the traceability and reproducibility of scenarios’ construction processes. At the same time, CBI can lead to scenarios with better consistency and more complete assumption sets than the two traditional approaches mentioned before. However, CIB was evaluated as unattractive to our exercise because of the considerable time and human resources required (which were unavailable to us) and the quality risks associated with its use by non-experienced teams in this method.

Therefore, in the SCCER JASM study, we based the scenario design process on our expertise with the scenario development methods we followed for the World Energy Council (WEC) scenarios⁷². The WEC methodology combines developing scenarios using three axes, the so-called energy trilemma, which refers to a balance between affordability, sustainability and security in energy system transformation⁷³, and stakeholder expertise⁷³. In SCCER JASM, the scenario axes were based on research and policy questions in Switzerland within the active discourse from 2017 to 2020, which was the timeframe within which the current exercise was performed. These mostly referred to the technical feasibility of transforming the Swiss energy system to a carbon-free energy supply and use and its economic implications regarding energy system costs and required investments in energy infrastructure. The technical feasibility of the presented pathways does not directly imply social, economic, or political feasibility.

Thus, the scenarios aim at exploring technically plausible pathways for reaching carbon neutrality by 2050 by adopting socio-economic drivers from Swiss Administrative and Federal Offices⁴¹. It should be noted that the assessed pathways presented in this study are only a subset of all the possible future energy system developments, and they are neither mutually exclusive nor forecasts. The main scenario axes used in their construction are: (a) technology development; (b) market integration; (c) security of energy supply. The first axis relates to sustainability, as new and immature technologies are needed in a decarbonised energy system. The second axis relates to market integration and trade of energy carriers, and emissions permits can reduce the mitigation costs. The third axis relates to the security of energy supply, which applies to different time and spatial scales, i.e., from hourly to annual and from local to global.

The BAU scenario extrapolates the energy consumption trends and energy policies in place for the future. It tries to answer the research question, “where does Switzerland land in terms of CO₂ emissions in 2050 if we continue the business-as-usual practices?”. The scenario reflects the energy and climate policy instruments currently in force but not measures under discussion and not included in the energy and climate Swiss legislation until 2019. Table 3 shows the major energy and climate policies implemented in the Baseline scenario.

The CLI scenario assumes that the world gradually shifts to implement green economy strategies. Most governments agree on driving environmental sustainability through related policies and practices. Emissions trading markets are globally introduced, and many international governments support standards and protocols to improve energy efficiency and adopt circular economy practices. Society has become aware of the challenges related to climate change mitigation and embraces cleaner and smarter lifestyles than in the BAU scenario. Shifts in energy consumer behaviour towards energy efficiency gradually developed in CLI from 2020 onwards. Thus, energy demand is reduced in the end-use sectors as more energy-sufficient lifestyles are developed. In this scenario, due to the global cooperation and effort in mitigating climate change, Internationally Transferred Mitigation Options (ITMOs) are available for Switzerland at competitive prices, which help the country compensate for a part of the domestic emissions abroad. Besides, CO₂ grids are developed, connecting the European countries, and regulations enabling carbon emissions transportation for sequestration abroad are enacted after 2030.

Until 2030 the energy sector in Switzerland is regulated through taxes, feed-in tariffs, capital subsidies and state-funded building renovation programmes to encourage clean energy solutions. Switzerland implements climate change mitigation policies to achieve emissions cuts by 2030, many of which are described in the revised CO₂ Act⁷⁴. A key subset of these policies is provided in Table 4.

Besides the core net-zero scenario, i.e. CLI, four additional scenarios are defined that are aimed at providing insights into the following questions related to energy transition towards net-zero CO₂ emissions: (a) what if the renewable expansion is not realised due to low population mobilisation (scenario ANTI); (b) what if the current bilateral energy agreements between Switzerland and the EU fail (ANTI and SECUR); (c) can Switzerland achieve the net-zero ambition by merely relying on domestic resources (SECUR); (d) what are the benefits for the energy transition cost, if increased integration between Swiss and international markets is achieved (MARKETS); (e) how the speed of the transition and its cost is affected by high innovation in low-carbon and clean technologies (INNOV); (f) which sectoral policies are incompatible with a least-cost trajectory to net-zero (LC). The design of these scenarios ensures comparability with CLI, and to some extent, among themselves.

The ANTI scenario assumes an international context with low cooperation in mitigating climate change. Due to the fragmented climate policies worldwide, R&D expenditures in low-carbon technologies are stagnated. There is limited technological progress, and, as a result, both consumers and utilities are faced with high upfront capital costs when adopting low-carbon solutions for energy supply and use. The integration of the Swiss and international energy markets is weaker in ANTI than in CLI, which hinders the availability of imported zero-carbon fuels and energy carriers, including electricity. However, environmental sensitivity still develops at local scales, and climate change mitigation goals are not abandoned. Communities seek to build a more sustainable future locally, and consumers rely on best-fit local solutions for energy supply and demand that maximise local benefits. They increasingly desire products and energy services customised for local contexts, which implies a high willingness to pay for these products. Individualistic lifestyles emerge with uneven energy efficiency and equipment replacement efforts between different consumer groups. A decentralised energy supply supports the development of local energy networks, accelerated building renovations to minimise import dependency at the community level, and widespread e-mobility and intelligent mobility. Local self-sufficiency becomes a priority in the ANTI scenario, which could often conflict with nationally-wide energy and climate targets. There is limited tolerance from society regarding landscape changes for implementing renewable and other low-carbon projects that do not directly benefit the local communities. In this context, we argue that the focus in ANTI is more on adaptation than mitigation, as capital-intensive projects for reducing the carbon footprint of the Swiss energy system are postponed to the future. Regarding implemented sectoral climate policies, the ANTI scenario extrapolates the BAU policies.

The SECUR scenario assumes that the world gravitates toward a multi-polar order, weakening the international governance systems. As a result, economic volatility increases and international trade agreements are not fully implemented due to rising nationalistic policies that make export-oriented growth less critical as a financial strategy. Most countries use whatever capabilities they have to achieve national-wide security in energy supply. Due to weakened international structures, access to energy resources becomes complicated, and import duties increase energy import prices. The shift in priority from environmental sustainability to energy supply security triggers conflicting objectives, but the overall global climate change mitigation targets are not abandoned. Switzerland relies upon its domestic energy resources, including hydropower and other renewable energy sources. The Swiss society is keener in the SECUR scenario than in CLI in fully exploiting sustainable domestic renewable resources to reduce the carbon intensity of the energy system. In SECUR, there is a higher social acceptance than in CLI in implementing projects based on new renewable energy forms such as bioenergy, geothermal and solar. Priority for consumers and utilities in SECUR is the secure and reliable operation of the Swiss energy system, given the weak integration of Swiss and international energy markets. The reduction of the overall import dependency in future is a priority. Domestic grid reinforcement is highly acceptable in SECUR and desirable to mitigate congestion for a secure and reliable grid operation. In terms of implemented sectoral policies, the SECUR scenario includes the ones of CLI. The SECUR scenario relaxes the domestic sustainable exploitable renewable energy potentials to their maximum values in the literature. It additionally imposes a constraint on annual net imports of all energy carriers to be as close as possible to zero without deviating from the net-zero CO₂ emissions target in 2050. The right-hand side of the constraint is found by continuously tightening the constraint and keeping the feasibility of the climate target.

The MARKETS scenario is a variant of CLI, which assumes higher global cooperation and integration of the Swiss and international energy markets beyond the levels considered in the core scenarios. Therefore, the MARKETS variant allows increased availability of imported biofuels, synthetic e-fuels, hydrogen and electricity. Priority in MARKETS is decarbonisation through affordable energy, which is also achieved via the development of local energy markets and intelligent prosumage schemes in coordination with national energy markets. As a result, technologies enabling sector coupling, prosumers and storages, demand-side management and vehicle-to-grid enjoy economies of scale, and their costs are lower than in the core scenarios. In supporting domestic and national markets integration, grid congestion is eliminated to a large extent by reinforcing domestic grids. With the focus on new business models based on renewable energy and P2X pathways, the social acceptance for using domestic renewable energy resources is higher than in the core scenarios and similar to the levels in SECUR. All policies included in the CLI scenario are also implemented in MARKETS.

The INNOV scenario is a variant of MARKETS. It assumes the same developments as MARKETS regarding integrating local, national and international energy markets and that a global effort exists to mitigate climate change and achieve the Paris Agreements. Closer international coordination is assumed concerning implementing climate change policies worldwide and reducing the costs of low-carbon technologies in energy supply and demand via increased R&D expenditures worldwide. Circular economy policies and schemes emerge globally, and there is a high level of material efficiency that results in lower energy conservation and renovation costs compared to the core scenarios. Moreover, in the context of a global joint effort to reduce GHG emissions, consumers also have high social acceptance for new technologies. Thus, the costs of clean and low-carbon technologies, such as renewable technologies, alternative vehicles and CCS, and energy supply and demand efficiency measures are lower in this variant than in the core scenarios. As INNOV builds on MARKETS, grid reinforcement and high availability of imported zero-carbon fuels and energy carriers are also assumed in this variant.

The LC scenario is another variant of CLI in which we sought the cost-optimal emissions reduction trajectory by excluding the sectoral policies in CLI (see Table 3 and “Table 4”). The LC scenario is regarded as the Least Cost solution and used as a benchmark to identify sectoral policies in CLI that could potentially entail high costs for energy consumers. The only constraints given to the LC scenario are the resource potentials, which are set to the levels of CLI, and the overarching CO₂ emissions reduction targets for 2030 and 2050.

The scenario design and quantification are based on systematic and informed research regarding alternative pathways and their costs. The derived paths are realistic and technically feasible under the evolving uncertainty of technical progress, resources, societal structures, national and international energy policies and markets.

All scenario assumptions used as input to the model are given in Supplementary Information Note 3 and Supplementary Data 1 with the key scenario parameters accompanying this manuscript. Further comparisons of the scenario results are provided in Supplementary Information Note 7. Detailed results are given in Supplementary Data 2 with energy balance tables for each scenario.

It should be noted that quantifying the scenarios by STEM assumes rational agents. In this regard, many of the policy interventions included in the scenarios are of financial nature. Moreover, aspects regarding lifestyle changes and sustainable and circular economy can only be given exogenously to our framework as part of a coherent storyline. An endogenous assessment of these aspects is beyond the scope of this paper as STEM needs further development to capture them. It is a limitation of our work, which we aim to address in future research with stronger collaboration with scholars from the social sciences and humanities fields.

Finally, the assessed scenarios did not include aggressive energy demand reductions in their hypotheses. While aspects of efficient lifestyles and circular economy are included in the scenarios, including ambitious energy savings due to lifestyle changes and beyond the levels achieved with technical measures needs extensive assessment of the energy saving potential that can be achieved with such changes. Such estimates are not yet available for Switzerland in a format that quantitative energy systems models like STEM can use. Thus, to avoid creating a paper exercise by arbitrarily changing the energy demand levels due to lifestyle changes, we opted for the scenario definitions to follow a conservative approach. Still, such a conservative approach is not necessarily out of context, given the uncertainty surrounding the energy demand development and associated with the future lifestyle changes affecting energy consumption behaviour and patterns. We aim to collaborate closely with the social sciences community in Switzerland to identify potential energy demand reductions associated with changes in energy consumption behaviour for the future years. Nevertheless, we should note that the model result regarding the need for accelerating renewable energy and introducing negative emissions technologies to meet the net-zero CO₂ emissions target is robust, as a recent study that assumed ambitious demand reductions due to lifestyle changes at global scales shows⁷⁵.

The meaning of the “policy cost” in a partial equilibrium framework like STEM

In a partial equilibrium framework, the term “policy cost” is defined as the difference in the total energy system cost with and without a specific policy/target, i.e., between the BAU and the suite of the net-zero CO₂ emissions scenarios assessed. Given that we do not account for feedback on the energy system to the rest of the economic sectors, we cannot claim that this is the overall cost associated with a specific policy measure. As we also do not account for the distributional impacts of the increased costs on different income classes, our results on the costs of the energy transition have limited information about the affordability of the energy transition.

The total energy system cost comprises: (a) capital costs for investing into and dismantling technologies for energy conversion, transmission and use, renovations and energy savings; (b) fixed and variable operation and maintenance costs; (c) costs for imports and domestic resource extraction and production; (d) revenues from exports; (e) delivery costs for energy carriers; (f) taxes and subsidies of energy carriers and technologies; (g) revenues from recuperation of embedded commodities accrued when a technology’s dismantling releases some valuable commodities and; (h) salvage value of technologies and embedded commodities at the end of the planning horizon. Moreover, the cost structure of the model is vintage-dependent. Thus, the costs for renewing the existing energy system infrastructure, independently of its decarbonisation, are included in the analysis.

The model’s objective is the discounted system cost from 2020 to 2050. For the discounting, a social discount rate of 2.5% is used. It should be noted that sectoral interest rates are applied for investments in energy technologies in the different sectors, which are higher than the social discount rate. In the end-use sectors, the interest rate is set to 5.5%, which implies a 3% risk premium (or hurdle rate) in the investment decision. In the supply sectors, the interest rate for investing in large-scale energy supply technologies is 2.5%, while the interest rate for investing in medium and small-sized energy technologies is 5.5%.

Further, we calculate costs for the baseline scenario (BAU), which does not achieve decarbonisation targets but extrapolates the current energy supply and consumption trends and the stated energy and climate policies in the Swiss legislation.

Our analysis does not include external costs associated with environmental damage, resource depletion, or human health. Moreover, our analysis does not include the distributional impacts of the energy transition on households of different income classes. Thus, it does not address equity concerns and affordability of the energy transition.