Heat Pump Lifecycle Emissions Explorer · BDA / The Transition Accelerator · v1.0 · May 2026

Refrigerant leaks do not erase the climate benefit of a heat pump

A common question in discussions around heating electrification is whether refrigerant leaks outweigh the emissions benefits of heat pumps. It is a legitimate concern: the current industry-standard refrigerant, R-410A, has a global warming potential of roughly 2,000 times that of CO₂, and field studies suggest cumulative leakage can reach about 1.5 times the equipment's nominal charge over its lifetime.

But the more important question is whether those leaks materially erode — or even cancel out — the climate benefit of switching to heat pumps. In almost all cases, they do not. Even at the extremes — whether switching from gas to a heat pump on a relatively emissions-intensive grid under worst-case leak assumptions, or adding a heat pump to an electric resistance heated home on a very low-emitting grid — heat pumps still deliver lifetime GHG savings.

And those conclusions are conservative: not accounting for forecasted declines of grid emissions intensities over the heat pump's lifetime; applying worst-case refrigerant leak rates while ignoring methane leakage in the gas system; and discounting both the current industry shift toward lower-GWP refrigerants, and the equivalent refrigerant leakage from the air conditioner many households would otherwise install instead of the heat pump.

This tool lets users test those assumptions directly, using input ranges informed by the literature. Try it yourself — under almost any reasonable combination, heat pumps still reduce lifecycle GHG emissions, even when refrigerant leaks are fully accounted for.

Jump to:
The verdict under current assumptions
Calculating…
GHG impact without considering refrigerants
—
operational emissions only
Net GHG impact (full lifecycle)
—
below baseline emissions
Refrigerant share of HP lifetime impact
—
of HP lifetime emissions
Lifecycle emissions, broken down
Left: baseline emissions (stacked). Middle: each addition or subtraction. Right: HP lifecycle total. All values in tonnes CO₂e over 16 years.
Fossil emissions (combustion + methane) Electricity (grid) Refrigerant impact Net total Additional end-use electrification potential

Fixed inputs (held constant)

Industry-typical values. Editable in the HTML source.
  • HP system size3-ton (~10.5 kW) ccASHP
  • HP refrigerant charge3.6 kg R-410A
  • AC refrigerant charge2.5 kg R-410A
  • Equipment life16 years
  • Annual cooling load5 GJ/${_yr} (when AC on)
  • Cooling COP / SEER3.4 (HP & AC, equal)
  • Methane GWP-100 / GWP-2029.8 / 82.5 (AR6 fossil)
  • AC lifetime leaktracks HP leak scenario
Step-by-step calculation for the current scenario

Updates live as you change inputs. All values shown correspond to the current scenario. Every step below uses only the constants documented in the Methodology section — a reviewer with a calculator should be able to reproduce every number.

Methodology, caveats, and sources

The framework

Standard Total Equivalent Warming Impact (TEWI) approach. For each scenario we compute lifecycle emissions = direct (refrigerant leaks × GWP) + indirect (energy × emission factor), comparing the gas/oil/electric baseline against the HP scenario over 16 years.

Why HP and AC use the same refrigerant

The tool compares two future installation choices: (a) the household installs a HP, or (b) the household installs an AC for cooling and keeps its existing furnace. Both options are future installations, so both follow current refrigerant regulations and the same refrigerant applies. The AC is smaller (2-ton vs. 3-ton HP), so its charge is roughly 70% of the HP charge for the same refrigerant.

Refrigerant loss scenarios

  • Low (50%): IPCC 2019 Refinement lower-end (~2–3% annual) + EU F-gas enforced end-of-life recovery (85–95%). Represents well-managed fleets.
  • Median (100%): Eunomia 2014 residential field study (for UK DECC) — 3.5% average annual loss × 16 years ≈ 56% operational, plus moderate end-of-life loss (~45%). Residential average is heavily skewed by catastrophic leaks in ~10% of installations.
  • High (150%): Derived from California PUC's Refrigerant Avoided Cost Calculator (2022), combining CARB's ~5% annual operational rate with ~80% end-of-life venting. Represents worst-case US fleet with unenforced recovery.

Annual operational and end-of-life losses are summed into a single refrigerant impact bar in the chart.

Refrigerant charges (3-ton residential HP)

  • R-410A: 3.6 kg (DNV / CARB 2022) · R-454B: 3.4 kg (OEM specs for 2025+ equipment) · R-32: 2.3 kg (Daikin / Chemours specs) · R-290: 0.5 kg (monoblock; charge limited by IEC 60335-2-40)

AC charges are scaled to ~70% of HP charge to reflect 2-ton vs 3-ton sizing.

GWP values used

All Global Warming Potential (GWP) values are sourced from IPCC AR6 Working Group I, Chapter 7 (2021), the most current scientific consensus. For refrigerant blends, blend-level GWPs are calculated as the mass-weighted average of constituent gases per AR6 (the IPCC reports per-molecule values; UNEP, ECCC, and CARB use the same constituent-weighting convention).

Refrigerant Composition GWP-100 GWP-20
R-410A50% R-32 / 50% R-1252,2564,715
R-454B68.9% R-32 / 31.1% R-1234yf5311,854
R-32100% HFC-327712,690
R-290100% propane0.020.07
Methane (CH₄), fossilCH₄29.882.5

Note on industry-standard values: Many regulatory and industry sources still use AR4-derived constituent averages — including ECCC's Federal Halocarbon Regulations 2022, UNEP RTOC 2022, and CARB reporting. Under AR4 conventions, R-410A GWP-100 is 2,088 (vs. 2,256 in AR6) and R-454B GWP-100 is 466 (vs. 531). For propane (R-290), AR6 reports a direct GWP-100 of 0.02 (Annex 7.A), correcting the longstanding "3" value commonly cited in industry — which was based on the simplistic assumption that each carbon atom yields one CO₂ molecule on combustion, and which AR4 reported as 3.3. The tool uses AR6 throughout to reflect current scientific consensus; users comparing against regulatory documents may see slightly different numbers as a result.

Grid emission factors

  • Québec: ~2 g CO₂e/kWh (ECCC NIR 1990–2024, Annex 7 — generation intensity, 2024 value: 2.10 g/kWh). Hydro-dominated grid with negligible combustion emissions. Forecast: ~2 g/kWh (essentially unchanged through 2026–2041; Hydro-Québec Plan stratégique 2035 continues adding clean generation).
  • Ontario: ~65 g CO₂e/kWh (ECCC NIR 1990–2024, Annex 7 — 2024 value: 64.9 g/kWh). Low-emission grid dominated by nuclear (~52%) and hydro (~24%), with gas-fired marginal generation. Forecast: ~95 g/kWh (16-yr average 2026–2041, Current Measures). Near-term rise driven by Pickering retirement (2024–2026) and gas growth before nuclear refurbishments come online; declines post-2035 under Clean Electricity Regulations.
  • Alberta: ~335 g CO₂e/kWh (ECCC NIR 1990–2024, Annex 7 — 2024 value: 334.8 g/kWh). Post-coal-phaseout fleet (all coal generation retired June 2024); now natural-gas-dominant with growing wind and solar share. Historical anchors: 2005 = 910, 2022 = 470, 2023 = 424 g/kWh. Forecast: ~245 g/kWh (16-yr average 2026–2041, Current Measures — continued wind/solar additions, federal Clean Electricity Regulations compliance from 2035).

The "current" values are taken directly from ECCC's National Inventory Report 1990–2024, Annex 7 (Electricity in Canada), 2024 generation-intensity table. Forecasts are derived from CER's Canada's Energy Future 2026 Current Measures scenario, which reflects only currently-legislated policies, calibrated to the NIR 2024 anchor for each province. Per-fuel emission factors used for scenario projection: Coal 950 g/kWh, Natural Gas 490, Oil 750, Renewables/Nuclear 15.

Average vs. marginal grid emissions

The tool uses average emission factors (AEF) for each provincial grid. A defensible line of analysis — developed for Ontario in TAF's emissions-factor guidelines — argues that marginal emission factors (MEF), capturing the emissions of the generator that responds to the next unit of load, are more theoretically appropriate for new electrified loads in fossil-leaning grids and retrofit fuel-switching contexts. In those grids, MEF can substantially exceed AEF in the near term.

The tool deliberately stays with AEF for three reasons:

  • Time horizon. Over a 16-year HP lifetime, the relevant signal is closer to the long-run / build margin than the near-term dispatch margin. Under the federal Clean Electricity Regulations (finalized December 2024), fossil generating units in NERC-connected grids become subject to annual emission limits from January 1, 2035, with grid net-zero targeted for 2050. The dispatch margin in 2026 is not the dispatch margin in 2034, let alone 2041.
  • The "Forecast" trajectory already operationalizes a soft long-run margin. Forecast values are 16-year averages of CER-projected generation mixes, already weighted toward the cleaner units that increasingly set the margin as legacy fossil capacity retires.
  • A single province-specific MEF over 16 years is not defensibly singular. Alberta is gas-marginal today but reshaped by CER from 2035. Ontario sits in a transition window dominated by Pickering retirement, Bruce/Darlington refurbishment timing, near-term gas growth, and post-2035 CER constraints. Québec has no consensus marginal value: hydro ramping, gas peaker hours, displaced exports, and marginal imports each yield very different numbers, and Hydro-Québec itself reports on an average basis. Asserting one MEF per province would publish a position rather than a calculation.

Users who want to stress-test the result against a fossil-leaning grid assumption can do so with the tool's existing controls: select Alberta with the "Today (static)" trajectory, R-410A refrigerant, and the highest leak rate. HPs typically still deliver lifecycle savings against a high-efficiency gas furnace under that combination — which is the more important takeaway than any single MEF estimate.

Systems-level displacement (freed kWh on a low-carbon grid)

When a household replaces electric resistance heating with a HP, the freed kWh has small direct value at near-zero grid intensity, but represents capacity that can substitute for fossil fuels in other sectors. The chart and callout illustrate three anchors of what those freed kWh could accomplish, calculated net of the freed kWh's own grid-emissions cost (i.e., the kWh's grid intensity is subtracted so the displayed values reflect the actual displacement gain that survives that cost):

  • Resistance electrification (low anchor — gross 0.189 kg CO₂/freed kWh): 1 kWh of electric resistance heat displaces 1 kWh-equivalent of gas combustion at 95% AFUE. Calculation: (1 / 0.95) × 49.9 kg/GJ × 0.0036 GJ/kWh.
  • HP electrification at the user's COP (mid anchor — gross varies): Same calculation but with the user-selected COP applied (same-technology comparison: "what could this kind of HP do elsewhere?"). Calculation: (COP / 0.95) × 49.9 × 0.0036. At COP 2.5: ≈ 0.473 kg/kWh gross.
  • Transport electrification (high anchor — gross 0.924 kg CO₂/freed kWh): A BEV at 0.20 kWh/km displaces gasoline at 8 L/100 km × 2.31 kg CO₂/L. Calculation: (1 / 0.20) × (8 / 100) × 2.31.

Each anchor's net value is then: net = freed kWh × (gross per-kWh factor) − freed kWh × (grid emission factor). Methane is omitted from all three anchors so anchor values stay stable across user methane selections and remain easy to reproduce. On near-zero-carbon grids (QC), net ≈ gross. On higher-carbon grids, the leveraged kWh's cost subtracts meaningfully — and on coal-leaning grids, the low anchor (1:1 resistance) can go net-negative (i.e., the freed kWh's grid emissions exceed the resistance-displacement gain). The chart's two-tone treatment for ON/AB makes this visible.

Why this isn't double-counting against utility plans: Hydro-Québec's Action Plan 2035 explicitly counts residential efficiency and HP conversions as the mechanism by which kWh are freed for transport and industrial electrification. The displacement shown here is the same value HQ is planning around — not a separate credit.

Cold-climate HP COP

Per NRCan / CanmetENERGY (Ferguson & Sager 2022, Appendix A heat-pump performance tables), 3-ton ccASHPs achieve measured instantaneous COPs of 1.38 at −25°C, 1.82 at −15°C, 2.99 at 0°C, 3.49 at +8.3°C, and 3.74 at +18°C (high-capacity operation). The paper does not publish a seasonal-average COP table directly. The full-electric options in this tool (2.0 / 2.5 / 3.0) are derived values consistent with those instantaneous COPs integrated against Canadian climate hour-counts — typically 2.0 in cold climates (Edmonton, Winnipeg), 2.5 in central Canada, 3.0 in mild climates (Vancouver, Victoria). All-electric configurations include electric-resistance backup at extreme cold (NRCan's "all-electric scenario," Sec. 3.6). Hybrid (dual-fuel) options (2.8 / 3.2 / 3.6) are higher because the HP operates only above the switchover temperature — avoiding the coldest, lowest-efficiency hours — with the backup furnace handling the rest. Hybrid values are also derived, calibrated to be consistent with Ferguson & Sager's reported 15–35% hybrid GHG savings vs. gas furnace.

Hybrid configurations

The HP covers a fraction of annual heating energy (50%, 65%, or 80%), while the backup furnace covers the remainder at the same baseline efficiency. The HP still handles 100% of cooling. The chart's "avoided baseline emissions" bar is sized to reflect the HP's coverage (only the avoided fraction is shown).

What this tool does NOT capture

  • Manufacturing embodied carbon of the equipment (small and roughly comparable between HP and furnace).
  • Hour-by-hour grid emissions, or marginal-vs-average attribution (discussed under "Average vs. marginal grid emissions" above).
  • Performance degradation over time (typically 0.5–1% per year for both HP and furnace).
  • Peak demand or grid-capacity implications (separate considerations).
  • Refrigerant manufacturing emissions (small, ~5% of in-use GWP impact).

Key references

Where multiple editions exist, the most recent applicable version is linked.

Refrigerants and lifecycle GHG analysis

Heat pump performance (Canadian context)

Refrigerant leak rates

Canadian regulatory and stewardship framework

Grid emissions data

GWP values

Methane leakage

Built for the Building Decarbonization Alliance / The Transition Accelerator. Outputs are illustrative of the shape of the comparison, not precise estimates for any specific installation.