Early decarbonisation of the European energy system pays off

The study addresses how different CO₂ reduction trajectories, given the same multi-decade carbon budget for Europe, affect the cost, technology deployment, and feasibility of decarbonising the energy system. Against the context of EU climate neutrality by 2050 and uncertain progress towards 2030 targets, the paper allocates a European carbon budget based on an equal per-capita share of the global budget, yielding about 48 GtCO₂ overall and 21 GtCO₂ for electricity and heating (33 GtCO₂ including transport). Prior work shows wind, solar, and hydro can supply hourly electricity demand with appropriate balancing and that sector coupling can provide flexibility. However, many models either ignore existing assets (greenfield) or lack temporal resolution and updated costs, limiting insights on transition pathways. This work investigates two paths—Early and Steady vs Late and Rapid—using an hourly resolved, sector-coupled, networked European model with myopic 5-year investment steps from 2020 to 2050 to determine the cost-effectiveness and operational implications of early versus delayed action under the same carbon budget.

The paper reviews evidence that cost declines in wind and solar PV enable high-renewable power systems if balancing is provided via reinforced interconnections or storage. Sector coupling (electricity, heating, transport) can supply additional flexibility, reducing costs and transmission needs, and deferring large storage investments. Greenfield optimisations inform fully decarbonised end-states but not transition dynamics. Myopic optimisation captures short-sighted decisions and tends to increase cumulative costs due to stranded assets compared to perfect foresight. Integrated Assessment Models often have low temporal resolution and outdated renewable cost assumptions, potentially underestimating wind and solar roles. Nordic experiences in heating decarbonisation via district heating, fuel shifts, and heat pumps illustrate effective sector-coupling strategies.

The system is modelled with hourly resolution and myopic optimisation in 5-year steps from 2020 to 2050, minimising annualised system cost under a cumulative CO₂ cap consistent with the chosen transition path. Two sector scopes are considered: (1) electricity and heating with a 21 GtCO₂ budget and (2) including road and rail transport with a 33 GtCO₂ budget. At each time step, capacities for generation, storage, and interconnections can be expanded if cost-effective, assuming perfect competition, perfect foresight within the step, and long-term market equilibrium. Constraints include nodal demand-supply balance, capacity limits, and a global CO₂ cap. The CO₂ price emerges endogenously as the KKT multiplier of the emissions constraint. Spatial resolution: one node per country for 30 countries (EU-28 as of 2018 excluding Malta and Cyprus, plus Norway, Switzerland, Bosnia-Herzegovina, and Serbia). Countries are interconnected by HVDC links; in the baseline, transmission follows ENTSO-E TYNDP expansions to 2030 and is fixed thereafter (variants allow further optimisation). Technologies in electricity: onshore/offshore wind, solar PV, hydro (reservoir, run-of-river), OCGT, CCGT, coal, lignite, nuclear, and CHP (gas/coal/biomass). Storage: pumped hydro storage, stationary batteries, hydrogen (electrolysers, H₂ storage, fuel cells). Power-to-gas: synthetic methane via combining DAC CO₂ with electrolytic H₂ (Sabatier). Heating: urban (district heating) served by large heat pumps, resistive heaters, gas boilers, solar collectors, CHP, and thermal energy storage (TES, e.g., hot water pits); rural served by individual heat pumps, electric boilers, gas boilers, and TES. Heat demand is split into urban and rural. Baseline assumptions: district heating penetration fixed at 2015 values; annual heat demand constant; transmission capacities expand per TYNDP to 2030 then fixed. Additional analyses vary DH expansion, building renovations reducing space heat demand (2% per year), allow transmission expansion beyond 2030, and include transport. Transport module: road and rail electrification levels exogenously follow the electricity+heating emissions reduction trajectory; BEV and battery costs are excluded; road/rail modelled as lumped demand by country. At each step, 50% of passenger BEVs enable smart charging (DSM) and 25% provide V2G. Hydrogen use in transport is not considered. Cost and finance: Technology costs vary over time (learning reflected exogenously) but not endogenously with cumulative installations. Annualisation discount rate: 7% for centralised assets and 4% for decentralised (e.g., rooftop PV, small water tanks). Cumulative path cost sums annualised costs using a social discount rate of 2%. Existing capacities (2020 and surviving to future steps) are exogenous; their annualised and operating costs are included. For fossil units constrained out by CO₂ caps before end-of-life, annualised costs are still accounted until technical lifetime ends (contributing to stranded costs).

• Cost advantage of early action: The Early and Steady path yields a cumulative system cost of 7875 B€, versus 8238 B€ for the Late and Rapid path (difference 363 B€), remaining cheaper for social discount rates below 15%.
• 2050 system: Wind (onshore, offshore) and solar PV dominate electricity supply, complemented by hydro and minor biomass. Cost per unit delivered energy in 2050 is ~59 €/MWh.
• Conventional capacity: No new lignite, coal, or nuclear is built along either path; newly built conventional assets by 2050 are limited to gas-fuelled power plants, CHP, and boilers.
• Sectoral decarbonisation: The power sector decarbonises faster than heating in both paths; differences are more pronounced in conventional heating capacity additions.
• Timing effects: In Late and Rapid, early budget depletion forces zero emissions by 2040, when renewables and balancing are costlier, increasing total cost. In 2040, up to 220 TWh/a of synthetic methane is produced to keep CCGT and gas boilers operating under high CO₂ prices, raising cumulative costs.
• Stranded assets: Expenditures not recovered via market revenues are similar in both paths and represent ~12% of cumulative system cost. Coal, lignite, CCGT (from early 2000s build-out), and gas boilers are most affected.
• Build rates: Achieving the Early and Steady path requires PV and wind installation rates comparable to historical peaks scaled to EU level (e.g., 50–100 GW/a for PV extrapolating Germany/Italy peaks). Early and Steady supports smoother build trajectories, avoiding boom-bust cycles and facilitating industrial and social adaptation.
• CO₂ price trajectory: The required CO₂ price is zero in 2020 (constraint non-binding) and rises with tightening caps, reaching levels well above historical ETS prices by 2050. Early and Steady yields a smoother CO₂ price evolution. Co-benefits of air quality and agriculture are estimated at 125–425 €/tCO₂, comparable to required late-stage prices.
• Balancing linkages: High PV shares pair with large battery capacities (diurnal balancing); high wind shares pair with hydrogen storage and reinforced interconnections (synoptic/seasonal balancing). Long-term storage (H₂, TES) is essential for seasonal variation and cold spells.
• Country mixes: Optimal national renewable mixes depend on local resources and existing capacities, but near-optimal alternatives exist with small cost penalties.
• Robustness and scenario variants (Table 1): • Expanding district heating to supply all urban heat by 2050 reduces cumulative cost by 2.4% (Early and Steady). • Building renovations cutting space heat demand by 2%/yr reduce cumulative cost by 11.3%. • Allowing transmission expansion after 2030 lowers cost by 1.3% and triples optimal interconnection volume by 2050, reducing H₂ storage needs. • Including road and rail transport increases cumulative cost by 5.4% (Early and Steady) due to +1102 TWh/a electricity demand when fully electrified; BEV flexibility reduces stationary battery needs and favours higher PV penetration. 25% V2G captures most cost reductions.
• Nuclear sensitivity: New nuclear appears only if nuclear costs are 15% lower than reference and no post-2030 transmission expansion is allowed; otherwise, no new nuclear is installed.

The findings directly address the core question: among pathways with identical carbon budgets, earlier and steadier reductions minimise cumulative system costs and smooth deployment and CO₂ price trajectories. High-resolution, sector-coupled modelling with updated technology costs reveals a cost-effective, fully decarbonised system primarily based on wind and solar, challenging prior IAM results that elevate nuclear, biomass, or CCS due to coarse time resolution or outdated cost assumptions. Policy relevance is high: the results support raising 2030 EU emissions reduction ambition, aligning with analyses of the EU ETS suggesting faster cap reductions are cost-effective. Early action mitigates stranded asset risks, enables smoother industrial scaling (avoiding boom-bust cycles), and aligns required CO₂ prices with quantified co-benefits for health and agriculture. Operationally, the hourly model highlights complementary roles of storage (batteries for PV; hydrogen for wind) and transmission for variability management, and confirms the importance of long-term thermal and hydrogen storage for seasonal balancing and cold spells. Compared to Clean Planet and ENTSO-E scenarios that rely heavily on nuclear or biomass/BECCS, the cost-optimal pathways here avoid associated concerns while remaining feasible when broader socio-technical factors are considered.

Early, steady decarbonisation of Europe’s energy system is consistently ~350 B€ cheaper than delayed, rapid action under the same CO₂ budget. With current and projected costs and hourly-resolved balancing, wind and solar (with hydro and minor biomass) can anchor a fully decarbonised sector-coupled system by 2050. Required renewable build rates match historical peaks, making the transition challenging but feasible, especially with stable policies that enable smooth deployment and CO₂ price trajectories. The analysis supports increasing 2030 ambition and highlights the benefits of district heating expansion, building efficiency, and strategic transmission reinforcement. Future research should assess inter-annual weather variability and climate change impacts with multi-year datasets, explore social acceptance and institutional constraints, consider hydrogen transport infrastructures, and evaluate broader interregional interconnections beyond Europe.

• Temporal and climatic data: Only one weather year is used; although costs appear robust across years for highly decarbonised power systems, sector-coupled impacts need further multi-year analysis. Climate change may alter wind correlation lengths (reducing transmission smoothing efficacy), hydro seasonality, and cooling/heating demands.
• Social and siting constraints: Limited social acceptance for onshore wind and utility-scale PV can constrain deployment; reduced onshore potential likely shifts capacity to offshore wind with limited cost increases.
• Hydrogen system scope: H₂ is assumed produced and used within the same country; hydrogen transport infrastructure is not modelled.
• Technology scope: Retrofitting of existing nuclear or deployment of coal with CCS is excluded. Nuclear only emerges in a narrow cost/transmission-constrained sensitivity.
• Stability timescales: The model resolves hourly operations; sub-hourly stability and ancillary services are not explicitly modelled but literature and field experience indicate feasible solutions (synchronous compensators, grid-forming inverters, demand response, curtailment).
• Geographic scope: Focus on Europe; interconnections and dependencies with neighbouring regions are neglected.
• Behavioural and adoption dynamics: Transport electrification shares are exogenous; BEV costs are excluded; potential hydrogen use in transport is omitted.