Replacing gas boilers with heat pumps is the fastest way to cut German gas consumption

To avoid exceeding 1.5 °C global warming above pre-industrial levels, rapid de-fossilisation of energy systems is required. In electricity, the cost-effective pathway involves replacing fossil fuels with photovoltaics (PV), wind power, and renewable balancing. Heating is another major emissions source; mitigation options include building efficiency, solar thermal, district heating using waste heat, geothermal, storage, and replacing gas and oil boilers with heat pumps powered by renewable electricity. Expansion of wood, pellets, and biomass for heating is limited and often unsustainable. Renewable hydrogen is unsuitable for domestic heating due to high system costs and energy inefficiency relative to heat pumps. The Russo-Ukrainian War has disrupted fossil gas supply security, prompting urgent decisions that risk locking in long-term contracts inconsistent with the Paris Agreement. The study hypothesizes that installing heat pumps, which are modular and require limited planning similar to PV, is among the fastest ways to reduce gas use. However, heat pumps compete for renewable electricity that could also displace gas-fired power generation, and in winter may rely on electricity partly generated from gas. The research question is how much gas can be substituted through rapid heat pump deployment alongside renewable expansion, especially when building insulation is modest. Using Germany as a case study, the authors model hourly gas consumption for buildings and industry and electricity generation, prioritizing maximal replacement of fossil gas with renewable electricity, and explore practical installation bottlenecks via bottom-up narratives and expert interviews to define acceleration scenarios.

The paper situates its work within established pathways for decarbonizing electricity through PV and wind with renewable balancing and highlights heating as a key sector for emissions reductions. Prior studies recognize heat pumps as a major component for de-fossilising heat, while noting limits and sustainability concerns for biomass-based heating. The authors reference analyses indicating renewable hydrogen is economically and energetically unsuitable for domestic heating compared with heat pumps. Integrated assessment models present diverse pathways for heat generation relative to electricity. The paper also draws on empirical datasets for heating load profiles, heat pump performance (COP) under varying conditions, and building energy classes and heating circuit temperatures, as well as sectoral gas consumption statistics for Germany to inform its modeling assumptions.

The study employs an hourly-resolved, bottom-up deterministic model to estimate substitution of fossil gas by renewable electricity via heat pumps in Germany, integrating both demand (heat) and supply (electricity) sides for years 2022–2030, with 2020 as the reference year. Key elements:

• Gas demand modeling: Residential sector gas in 2020 is split into 273 TWh for space heating and 12 TWh for cooking. Industrial low-temperature (<100 °C) heat is modeled for chemical (75.8 TWh), food processing (38.6 TWh), and paper (25.4 TWh) industries, including an estimated three-quarters of industrial heat losses below 100 °C. Trade and services are excluded due to diversity. Hourly load profiles depend on daily average temperature, with distinctions between weekdays and weekends. Industry hourly load curves are taken by season; heat pump demand is modeled as a function of outdoor temperature with data accounting for frosting/defrosting. Location-specific temperature inputs use DWD data (Hanover as representative), with error bars reflecting regional variability.
• Electricity system modeling: The baseline uses 2020 hourly generation from all sources, including installed wind (56 GW onshore, 3.6 GW offshore) and PV (54 GW), plus storage capacities. Near-future generation is projected using a data-driven approach (Ministry of Economics and Climate Protection), incorporating planned additions of wind and PV through 2030 and assuming sufficient grid expansions.
• Substitution logic: The model prioritizes maximal gas displacement either by (a) reducing load hours of gas-fired power plants with new renewable electricity, or (b) using renewable electricity to power heat pumps that replace gas boilers. The comparison hinges on heat pump COP versus gas plant efficiency. Gas CCGTs in 2020 exhibited ~55% efficiency (electricity/gas input), reduced to 50% to account for grid losses.
• Heat pump performance: Industrial heat pumps are modeled with COP=2 to reflect high output temperature needs but frequent availability of higher source temperatures via residual heat. Residential COPs are modeled hourly by building energy class (A–H) reflecting insulation levels and by heating circuit type. Only buildings using radiators with heat pumps are considered (conservative). Supply water temperature requirements are derived from a dataset of hundreds of German buildings with radiator heating; linear fits by energy class determine supply temperature versus outdoor air temperature, with a conservative +10 °C margin in poorer classes. Buildings in energy class H are excluded from heat pump retrofits without insulation upgrades. COP curves are taken from a commercially available air-to-water heat pump data sheet; a conservative quadratic fit is used. Only air-to-water systems are modeled, though alternatives may yield higher COPs; sensitivity is discussed in supplements.
• Scenario building and constraints: Semi-structured interviews with two plumbing/heating business owners (one non-adopter, one partial adopter) and an expert inform micro-level constraints: workforce shortages, limited/practical training quality, perceived technical maturity, integration challenges (domestic hot water), and the importance of high-output-temperature units for old buildings. These narratives, along with macro-level supply (global production ~10 million heat pumps/year; domestic production scaling), inform four deployment scenarios, including an installers’ roadmap baseline and accelerated S-curve adoption triggered by economic and policy shifts (e.g., gas price increases, electricity price caps). A “very fast” scenario installs up to 4 million heat pumps per year, assuming 4–6 person-days per install (including radiator upsizing and DHW integration), no net addition of workers, team-based installation with training for every second tradesperson (3–4 days), and a cap on heat pump supply temperature at 60 °C (target room temps up to 20 °C). Industrial applications below 100 °C deploy larger heat pumps; detailed incentives are out of scope.
• Mathematical treatment: Gas consumption per hour per sector is computed from temperature-dependent standard load profiles, normalized to match annual sectoral totals. Future substituted gas quantities are derived by applying scenario-specific green factors and heat pump adoption trajectories, with hourly resolution coupling to renewable generation profiles. Weather variability is addressed by comparing multiple historical years (2017–2019 vs. 2020), with standard deviations propagated to result error bars.
• Approximations: (1) Geographical distribution of new PV/wind simplified; assumes adequate grid expansion. (2) Non-gas power plant generation kept constant across years; market merit order implies coal displacement has minor effect on gas substitution estimates. (3) Interannual renewable variability considered via historical analogs; dark doldrums reduce savings only to some extent.
• Limitations acknowledged include neglect of transmission congestion, localized generation changes, and detailed grid dispatch; focus on newly added renewables and heat pumps; exclusion of trade/services sector; and deterministic single-realization approach.

• Accelerated heat pump deployment combined with rapid renewable capacity additions can substantially reduce Germany’s fossil gas consumption within a few years.
• In the very fast scenario, residual gas savings of about 30% by 2025 relative to total 2020 gas imports (971 TWh) are achieved; this corresponds to saving about 60% of the gas imported from the Russian Federation by 2025 (Russia supplied ~50% of gas in 2020). The text also cites approximately 28 billion m³ saved by 2025 in this scenario and at least 180 Mt of GHG emissions avoided by 2025.
• Early in the transition (first two years), most newly added wind and PV reduce gas-fired power plant load hours automatically through market dispatch. From 2024–2025 onward, a substantial share of new renewable electricity is used by heat pumps, while enough electricity remains to continue coal phase-out and enable growth in e-mobility.
• Hourly modeling indicates relatively small interannual weather-driven variation in gas substitution; error bars for 2026 illustrate modest year-to-year effects.
• Practical constraints (workforce, training, integration in buildings with radiators) can be overcome with targeted efforts: team-based installation, modest radiator upsizing, conservative supply temperature limits (~60 °C), and focused training enable scaling to millions of installations per year without adding new workers in the near term.

The results support the hypothesis that rapidly installing heat pumps, alongside expanding renewable electricity, is one of the fastest viable pathways to reduce fossil gas consumption in Germany. Despite competition for renewable electricity between displacing gas-fired generation and powering heat pumps, the modeled system dynamics show a temporal complementarity: initial renewable additions primarily curtail gas plant load hours, while subsequent increments are effectively absorbed by a growing fleet of heat pumps. Even with conservative assumptions—radiators retained, limited supply temperatures, COPs based on air-to-water units, and partial reliance on gas-generated electricity during winter—substantial gas savings are achieved. The findings suggest that accelerated heat pump deployment can simultaneously reduce fuel supply risk, dampen price volatility, and align with Paris Agreement trajectories while preserving capacity to progress on coal phase-out and e-mobility. The bottom-up scenario analysis, grounded in installer narratives, underscores that micro-level bottlenecks (workforce, training quality, integration concerns) are surmountable through targeted policy and industry coordination, including improved training, incentivization, and grid enhancements.

Accelerated installation of heat pumps, powered by rapidly expanding renewable electricity, provides an effective and near-term strategy to significantly cut fossil gas consumption in Germany. With coordinated efforts across government, industry, and consumers—ensuring installer lucrativeness, training offensives, and strategic grid expansion—approximately 60% of the 2020 volume of gas imported from the Russian Federation can be substituted by 2025 (around 40% under more moderate assumptions). This approach enhances energy security, reduces greenhouse gas emissions, and remains compatible with ongoing coal phase-out and electrification of transport. Future research should refine grid congestion modeling, extend sectoral coverage (e.g., trade/services), incorporate stochastic weather and market dynamics, and evaluate policy instruments and industrial incentives to further accelerate low-temperature industrial heat electrification.

• Deterministic modeling with a single nominal realization of renewable generation; interannual variability handled only via historical analog comparisons and summarized error bars.
• Assumes sufficient grid expansion and does not model transmission constraints or congestion explicitly, which may be significant under high heat pump uptake (particularly in 2024–2025 when a large share of new renewables serves heat pumps).
• Keeps non-gas generation constant across years and focuses on newly added renewables and heat pumps; changes in local generation capacities and detailed market interactions are simplified.
• Excludes the diverse trade and services sector from heat pump uptake modeling; results are conservative for total potential.
• Heat pump COPs are based on air-to-water systems and conservative temperature assumptions; buildings in the worst insulation class (H) are excluded without upgrades; potential of higher-COP technologies (air-to-air, water-to-water) is not fully captured.
• Industrial modeling limited to heat below 100 °C; detailed policy incentives for industry not assessed.
• Workforce and training assumptions (no new workers added, team-based installs, limited training duration) may affect scalability and timelines in practice.
• Code availability is limited; detailed datasets (2017–2022) are not publicly accessible, constraining reproducibility.