Energy and environmental impacts of air-to-air heat pumps in a mid-latitude city

The study addresses how transitioning from conventional heating (electric resistive and gas) to air-to-air heat pumps (AAHPs) affects building energy consumption, electricity demand, urban microclimate, and CO₂ emissions in a mid-latitude European city. Residential space heating is a significant fraction of final energy use and CO₂ emissions in the EU and globally. Heat pumps provide a high-efficiency, electrified alternative, but their benefits depend on system type, operation, and the carbon intensity of electricity. The work focuses on Toulouse, France, where low-carbon electricity (notably nuclear) suggests strong potential for emission reductions. The research aims to quantify energy savings, grid implications via COP variability, and urban climate effects using an integrated modeling framework.

Prior studies report that HPs generally reduce carbon intensity versus fossil heating across most regions under current grid mixes (e.g., 53 of 59 regions), with particularly large benefits in low-carbon electricity contexts like France. Conversely, in regions with fossil-dominant grids (e.g., parts of the United States), electrification via HPs can increase costs and CO₂ emissions. Empirical work from Arizona found no notable electricity savings in summer and increased winter demand following HP adoption. A UK retrofit with air-to-water HPs saw a modest CO₂ reduction but higher operational costs. Collectively, the literature emphasizes that HP impacts are region- and grid-dependent, and that realized energy savings can be offset by behavioral and contextual factors. AAHPs are the most common HP type globally, warranting focused investigation of their performance under real-world indoor/outdoor conditions and their transient efficiency (COP) behavior.

An integrated modeling approach couples heating, ventilation, and air conditioning (HVAC) models with building energy and urban canopy models, and with a mesoscale atmospheric model for microclimate impacts. Core tools: SURFEX-TEB (v8.2) computes urban surface energy budgets and building energy consumption at district scale, using meteorological forcing and urban morphology/constructive parameters. MinimalDX (v0.3.0), a simplified single-speed direct-expansion coil model (derived from EnergyPlus), represents AAHP/AC performance via quadratic fits of capacity and electric input ratio (EIR=1/COP) against indoor wet-bulb and outdoor dry-bulb temperatures, using performance curves from prior work. Five AAHP scenarios are simulated with rated COP (RC) of 2.5, 3.0, 3.5, 4.0, 4.5, allowing COP to vary dynamically with conditions. Internal fan energy is neglected; defrost is modeled as resistive at 20% of rated capacity for outdoor air temperatures below 0 °C. Baseline configuration reproduces 2004–2005 Toulouse heating systems and usage as in referenced prior studies. Offline SURFEX-TEB simulations (forced with observed meteorology) cover March 2004–February 2005 for energy impacts; during summer, AAHPs are repurposed as ACs, with design setpoints adjusted to represent wider AC adoption to quantify rebound. Heating design temperatures vary by building type; baseline capacity limitations (3.23 GW) restrict setpoint maintenance during cold spells, while AAHP scenarios increase capacity (6.64 GW) and remove this limitation, reflecting adequate sizing and reduced occupant curtailment (rebound). Non-HVAC internal gains vary seasonally with outdoor temperature. Domain: 15×15 km at 250 m over Toulouse; urban morphology and construction parameters sourced from prior datasets; heating mixes by administrative area (electric resistive 52.3%, gas 41%, wood 4.4%, oil 2.3%), applied uniformly within each area. Electric energy consumption (EEC) is computed via an energy disaggregation approach and evaluated in the city center where data are available. Microclimate impacts during a cold spell (22–30 Jan 2005) are simulated using Meso-NH (v5.3) coupled to SURFEX-TEB-MinimalDX, nested down to 250 m resolution, with ECMWF operational analyses for initial/lateral conditions and standard physical parameterizations. Anthropogenic heat flux analyses exclude traffic and industry to isolate building contributions.

- Energy use: Baseline annual building energy consumption (BEC) is 186 GWd for heating and 4.1 GWd for cooling (total 190 GWd). With AAHPs, heating BEC reductions range from 61% (RC2.5) to 78% (RC4.5); for the median RC3.5, heating BEC drops 72% to 52 GWd. Cooling BEC increases due to rebound, by 54% to 6.3 GWd (RC3.5), but total BEC still falls 69% to 59 GWd (RC3.5). - Electricity use (center of Toulouse): Electric heating energy drops by 24–57% across scenarios; for RC3.5, EEC for heating decreases 45% from 2.4 to 1.3 GWd. Cooling EEC increases by 17% to 0.36 GWd (RC3.5). COP dynamically varies and can fall below 2.5 during the coldest days even in the RC3.5 scenario. - Electricity use (entire domain): With 52% baseline heating by electric resistive heaters, RC3.5 yields a 36% decrease in electric heating energy to 52 GWd and a 54% increase in cooling electricity to 6.3 GWd. Heating electricity savings range from 12% (RC2.5) to 50% (RC4.5). Total EEC (heating + cooling) falls by 32% in the RC3.5 scenario. - Surface energy balance: Anthropogenic heat flux from buildings peaks near 80 W m⁻² in the center; AAHPs reduce anthropogenic heat by ~50–80%. Winter sensible heat flux reaches ~100 W m⁻² in urban core under baseline and is reduced by ~50 W m⁻² with AAHPs in dense areas. - Microclimate: During a cold spell (22–30 Jan 2005), AAHP deployment lowers 2 m air temperature by up to ~0.45–0.5 °C, with the largest reductions in the urban core; average reductions across stations are under 0.5 °C. - CO₂ emissions: Baseline local heating-related CO₂ emissions reach up to 40 kg CO₂ m⁻² in the center and up to ~7000 t CO₂ day⁻¹ during winter cold spells at the agglomeration scale. Transition to AAHPs under low-carbon electricity implies near-zero local heating emissions; if electricity were fossil-based, indirect off-site emissions could exceed local combustion emissions. - Refrigerants: Assuming 1% annual refrigerant loss, estimated leakage equals ~11,205 t CO₂e year⁻¹ (~31 t CO₂e day⁻¹) for the modeled AAHP stock; end-of-life non-recovery for 20% of units would release ~224,100 t CO₂e—both below the baseline annual emissions estimate of ~577,740 t CO₂ (Mar 2004–Feb 2005).

The findings demonstrate substantial reductions in building energy use and electricity for heating when replacing resistive and gas systems with AAHPs, especially in a low-carbon electricity context like France. However, benefits are contingent on the regional energy mix; in fossil-dominant grids, electrification may raise emissions and costs. AAHP efficiency (COP) declines at low outdoor temperatures, introducing complex, temperature-dependent electricity demand patterns that challenge grid planning and peak management. While cooling electricity increases due to rebound and broadened AC adoption, net energy and electricity use still decline in Toulouse’s context. AAHPs also reduce anthropogenic and sensible heat fluxes, slightly cooling the urban near-surface air (≤0.5 °C) during cold spells, with minimal feedback on HP performance expected. CO₂ benefits are significant locally, but overall climate gains depend on low-carbon electricity supply. Refrigerant-related climate impacts, though smaller than baseline combustion emissions, must be managed via low-GWP refrigerants, leak minimization, and end-of-life recovery. The results highlight the need to align heating electrification with clean power, adapt grid operations to COP variability, and account for behavioral rebound.

Deploying AAHPs in Toulouse yields large reductions in heating energy consumption and electricity demand, modest increases in cooling electricity due to rebound, and near-elimination of local heating-related CO₂ emissions under a low-carbon grid. Urban climate impacts include reduced anthropogenic and sensible heat fluxes and slight coolings of near-surface air during cold spells. Effective implementation requires considering local heating mixes, building stock, behavior, and grid carbon intensity. Future research should explore broader city types and climates, quantify peak-demand and grid stability implications of COP variability, assess long-term rebound under warming climates with expanding AC use, and evaluate refrigerant alternatives and lifecycle management to minimize indirect emissions.

- Context dependence: Results hinge on Toulouse’s heating mix and low-carbon electricity; findings may not generalize to fossil-heavy grids. - Modeling assumptions: Anthropogenic heat flux excludes traffic and industry; heating system capacities differ between baseline and AAHP scenarios; internal fan energy is neglected; a simplified resistive defrost strategy is assumed; uniform heating system shares are applied within administrative areas, ignoring building-type correlations. - Temporal and spatial scope: Analysis focuses on one year (2004–2005) and a single city/domain. - Behavioral and rebound effects: Increased AC adoption and comfort setpoints may erode savings; future climates could amplify cooling demand beyond the estimates. - COP variability: Efficiency declines in cold conditions introduce uncertainty in peak loads and grid impacts. - Refrigerant impacts: Leakage and end-of-life losses add CO₂e; estimates rely on generalized rates and average charges. - Validation constraints: EEC evaluation is strongest in the city center where data exist; broader-domain disaggregation relies on model-based methods.