Increased energy use for adaptation significantly impacts mitigation pathways

The study investigates how energy-intensive climate adaptation actions feed back into energy systems, emissions, and mitigation policy design—an interaction largely absent from most current integrated energy scenarios and Illustrative Mitigation Pathways. Rising temperatures and extremes increase electricity use for space cooling, refrigeration, and industrial processes, while potentially decreasing (but not eliminating) heating demand; extreme temperatures also affect productivity, prompting adjustments in energy use across sectors. Given that low-energy-demand pathways facilitate stringent mitigation, increased adaptation-driven energy demand may jeopardize decarbonization goals. The paper’s core question is: how do direct energy needs for adaptation alter global and regional energy demand, capacity investment, costs, emissions, air pollution, and the carbon prices required to meet climate targets?

Empirical and engineering studies have long documented weather sensitivity of energy demand, yet most energy scenarios omit adaptation-energy feedbacks. Prior global analyses often used econometric projections without price-induced substitution and income effects, while CGE studies suggest macroeconomies can absorb adaptation-related energy costs but lack detailed energy system implications. IAM work commonly relies on average temperatures or HDD/CDD metrics, which underrepresent nonlinear demand surges at extreme temperatures. Few IAM assessments have quantified macroeconomic impacts of adaptation-driven energy use or its implications for power capacity, system costs, and co-pollutants at global scale. This work extends literature by endogenizing adaptation-energy feedbacks across fuels, sectors, and regions within a process-detailed IAM and comparing mitigation pathways and SSPs.

The study extends the WITCH 5.0 integrated assessment model—linking economy, detailed energy system, climate (MAGICC), and an air pollution module (FASST(R))—to endogenize an adaptation-energy feedback. Three steps: (1) Estimate reduced-form links between annual mean temperature and extreme temperature indicators (ETIs): annual counts of days with average daily T > 27.5°C and T < 12.5°C. Countries are clustered to capture heterogeneous ETI–temperature relationships; statistical emulators project future ETIs from regional mean temperatures which are tied to global mean temperature. (2) Map ETI changes to sector- and fuel-specific energy demand using semi-elasticities (from De Cian & Sue Wing, 2019) for electricity, gas, and oil in residential, commercial, and industrial sectors, distinguishing temperate vs. tropical macro-regions. This approach captures nonlinear tails better than HDD/CDD-based methods. Sectoral semi-elasticities are aggregated with sectoral energy shares to derive regional-fuel-time semi-elasticities that shift demand relative to conditional means as ETI counts change. (3) Feed climate-induced demand shocks into WITCH’s production tree as changes in energy input productivity (technological retrogression when demand rises), allowing endogenous supply-side adjustments: generation mix, capacity expansion (including flexibility and capacity adequacy constraints), grid and fuel investments, and prices. Scenarios: A current policy case (continuation of implied policy ambition post-2020, no fixed carbon budget) and two cost-effective mitigation pathways with uniform global carbon prices and fixed global GHG budgets aligning to about 2.5°C and well-below 2°C by 2100. Non-CO2 gases priced using 100-year GWPs. Socioeconomics follow SSP2 main case, with SSP3 and SSP5 sensitivity. Outputs include energy demand by sector/fuel, additional capacity by technology (coal, gas, oil, wind, solar, storage), energy system costs (NPV of generation, grid, O&M, fuels), electricity prices, GHGs, and air pollutants (BC, NOx, CO, SO2, OC, VOC). Data and code are provided via Zenodo; WITCH is open source on GitHub.

• Adaptation-driven energy demand growth: By 2050 (2100), global electricity demand increases by 7% (18%) and fuels by 1% (2.5%) under current policies, relative to no-adaptation cases. In 2100 (SSP2, current policy), added energy for adaptation equals ~20% of 2019 global final energy demand. Under SSP5, added energy exceeds 100 EJ/yr by 2100; mitigation halves (2.5°C) or cuts >70% (well-below 2°C) adaptation energy needs. Industrial sector supplies ~40% of additional electricity needs. Liquids/gases increases are small globally, masking sectoral heterogeneity: building heating fuels decrease while industrial fuel use rises; distributed petroleum-fired generation and absorption cooling can elevate fossil use in developing regions.
• Regional heterogeneity: Africa and Middle East experience the largest relative increases in final energy for adaptation; in current policy 2100, regional electricity demand rises ~50% vs. no-adaptation. Exposure to hot days (>27.5°C) surges in Indonesia, SE Asia, Sub-Saharan Africa (>100 additional days by 2100 vs. 2005); stringent mitigation drastically reduces such exposure.
• Additional power capacity and lock-in: Under current policies, cumulative additional fossil capacity installed by 2050 due to adaptation is ~300 GW coal, 390 GW oil, and 960 GW gas (≈55 GW/yr average across 2020–2050). Ambitious mitigation cuts extra coal/oil capacity by 50–90% and gas additions to ~300–580 GW by 2050. After 2050, additional capacity shifts toward renewables and storage. Fossil share in power mix changes little with adaptation; however, power carbon intensity rises modestly in less ambitious cases (e.g., 2030 current policy 471 vs. 460 gCO2/kWh with vs. without adaptation).
• Energy system costs and prices: Adaptation raises global electricity supply costs by ~21% (NPV 2020–2100) and total energy system costs by ~4.5% under current policies; electricity prices rise 2–6% depending on year/region. Ambitious mitigation substantially reduces the adaptation-induced cost increase; when accounting for reduced adaptation needs, additional power system costs of stringent mitigation are much lower than previously estimated and can turn negative (net gains) in the well-below 2°C case relative to current policy. Average annual per-capita ESC increase is ~$105; increases are larger in USA, MENA, SE Asia and Indonesia, with higher GDP burden in middle-income regions.
• Emissions and co-pollutants: Under current policies, cumulative GHGs attributable to adaptation reach ~350 GtCO2eq by 2100 (~7% of 2020–2100 cumulative emissions). Additional emissions are larger in developing/tropical regions given slower energy transitions; SSA has the largest total increase but low per-capita additions, whereas SE Asia/Indonesia show higher per-capita increases. Air pollutants (NOx, CO, SO2) increase notably, with peak average annual rises of ~200 kton/yr in SSA, 157 kton/yr in SE Asia, and 145 kton/yr in MENA, implying heightened pollution exposure risks.
• Carbon price impacts: To meet fixed carbon budgets, global carbon prices must be higher when accounting for adaptation-energy feedbacks: up to +30% in 2.5°C scenarios (e.g., 2050: $16 → $21/tCO2eq; 2100: $44 → $57), and about +5% in well-below 2°C scenarios (e.g., 2050: $151 → $158; 2100: $422 → $443).

By endogenizing adaptation-driven energy demand, the study demonstrates that adaptation materially alters mitigation pathways through increased energy intensity, capacity needs, and costs, particularly in the near term where added capacity is often fossil-based. This directly affects GHG and pollutant trajectories and necessitates higher carbon prices to achieve the same climate targets. Critically, integrating adaptation reveals additional co-benefits of mitigation: stringent policies reduce extreme temperature exposure, curtail adaptation energy demand, lower energy system costs, and diminish lock-in and co-pollutant emissions. The results reframe cost-effective mitigation design by showing that scenarios ignoring adaptation-energy feedback overstate the incremental costs of mitigation and understate its benefits, especially for vulnerable regions where power systems face large cooling-driven peaks and reliability issues. Regionally differentiated effects underscore equity and development considerations, as similar per-capita cost increases impose heavier income burdens in middle-income countries.

The paper contributes a comprehensive integration of adaptation-energy feedbacks into an IAM, quantifying their impacts on energy demand, capacity expansion, costs, emissions, air pollution, and carbon prices across scenarios and SSPs. It shows that adaptation considerably increases electricity and, to a lesser extent, fuel use, leading to short- to medium-term fossil lock-ins, higher GHGs and air pollutants, and elevated energy system costs; consequently, carbon prices must rise to meet climate targets. Conversely, ambitious mitigation substantially reduces adaptation energy needs and can yield net power system cost savings relative to current policies. Future research should broaden coverage of adaptation channels (e.g., water supply/treatment, transport, cold chains), incorporate planned adaptation (passive cooling, reflective roofs, urban greening), improve representation of peak load responses to extremes, consider humidity and region-specific thresholds, explore supply-side climate vulnerabilities, and quantify health and welfare effects of co-pollutants and residual damages.

Key limitations include: reliance on two extreme temperature bins (T < 12.5°C, T > 27.5°C) and omission of moderate temperature effects, potentially overstating net impacts; semi-elasticities derived from existing empirical work with limited alternatives across fuels/sectors and climate zones; not explicitly modeling peak electricity demand sensitivity, which could raise future system costs; omission of additional adaptation channels (water, transport, cold chains) and of welfare/health impacts from increased co-pollutants; not accounting for humidity-driven thermal discomfort or region-specific ETI thresholds that could amplify demand in tropics; potential behavioral and technological adaptations (efficiency, digitalization, passive cooling, building performance) that might attenuate future energy needs; and uncertainties in technological change, policy, and power system resilience under climate stressors.