The Paris Agreement aims to limit global warming to well below 2 °C, preferably 1.5 °C, above pre-industrial levels. Achieving this necessitates net-zero greenhouse gas (GHG) emissions around 2050. Numerous countries, cities, and companies have set net-zero targets, sparking debate about achieving these goals and reliance on CDR technologies. Scenario studies suggest that net-zero emissions are more easily achieved by allowing residual emissions in certain sectors, offset by CDR in others. However, the IPCC's AR6 shows that most net-zero CO₂ scenarios allow 11 GtCO₂ of residual emissions annually, compensated by CDR. These residual GHG emissions mainly stem from energy-intensive industries (steel, cement, chemicals), international transport (air and shipping), and non-CO₂ emissions from agriculture (livestock, fertilization, rice cultivation). Transformations in these hard-to-abate (HtA) sectors are challenging due to technical limitations, complex value chains, and societal factors. Major CDR measures include BECCS, afforestation/reforestation, and direct air carbon capture and storage (DACCS), each involving environmental, social, and economic risks, particularly land use. This study explores demand and technological options to reduce HtA sector emissions and lessen BECCS reliance.

A comprehensive literature review (Supplementary Information) examined challenges in mitigating emissions in HtA sectors and potential demand and technology-oriented measures and policies. The review covered sustainable fuels, electrification, efficiency improvements, and operational advancements. For agriculture, buildings, and transport, demand-oriented measures focused on lifestyle changes like dietary shifts. Policy instruments included economic (subsidies, taxation), regulatory (standards), and other (information, infrastructure, agreements, technology transfer). The review informed the development of new scenarios.

The study employed the IMAGE integrated assessment model, chosen for its detailed representation of HtA sectors and ability to evaluate targeted measures and their impact on energy and land interactions. Three new scenarios were created: Reference 1.5 °C (default 1.5 °C scenario with a uniform global carbon price), Demand 1.5 °C (additional demand-side measures), and Technology 1.5 °C (additional technology-oriented measures). A 'Combined 1.5 °C' scenario combined both. Scenarios were run under SSP1, SSP2, and SSP3 socioeconomic assumptions. Demand-side measures targeted consumption, while technology measures focused on service provision and innovation. The model tracks costs, allowing assessment of cost-effectiveness. The IMAGE model incorporates TIMER (energy and industry) and MAGNET (agricultural demand, production, and trade), along with LPJmL (carbon, crops, vegetation, water) and MAGICC (climate). The model operates at regional and grid levels, offering detailed geographical representation. Specific measures within each sector (transport, buildings, industry, agriculture) are detailed in Table 1 and the Extended Data Tables (1-5). For example, the Demand scenario involved dietary shifts, reduced travel, material demand reductions, and decreased energy waste in buildings, while the Technology scenario emphasized technology adoption such as efficient aircrafts, heat pumps, and artificial meat.

The Reference 1.5 °C scenario showed slow emission reductions in HtA sectors compared to other sectors. The Demand 1.5 °C, Technology 1.5 °C, and Combined 1.5 °C scenarios significantly reduced residual emissions from HtA sectors to 5.6–7.1 GtCO₂e by 2060 under SSP2 (4.7–5.6 GtCO₂e in SSP1 and 5.8–7.8 GtCO₂e in SSP3), compared with 8.3 GtCO₂e in the Reference scenario. Stringent industrial measures effectively reduced emissions in both Demand and Technology scenarios. The Technology scenario prioritized immediate zero-carbon technology adoption, while the Demand scenario emphasized demand reduction. In transport, demand measures (airline tax, high-speed rail) reduced emissions by 43–64% by 2060. In agriculture, dietary shifts and reduced food waste strongly reduced non-CO₂ emissions and indirectly affected land use. The Combined scenario's impact was less than the sum of individual scenarios due to systemic interactions (e.g., efficiency improvements reducing the impact of demand reduction). The additional measures significantly reduced BECCS use. In the Demand 1.5 °C scenario, BECCS was reduced by 20%, 74%, and 94% under SSP3, SSP2, and SSP1, respectively, in the net-zero year. Peak annual BECCS use was limited to 0.5–2.2 GtCO₂e/yr in the Demand scenario and 1.9–7.0 GtCO₂e/yr in the Technology scenario (compared to 10 GtCO₂e/yr in the Reference scenario).

The study highlights the potential for significant emission reductions in HtA sectors through ambitious demand and technological interventions. The findings demonstrate that limiting BECCS use to sustainable levels is feasible while still achieving the 1.5 °C target. Demand-side measures, especially dietary changes, offer substantial mitigation potential and reduce the need for large-scale land-use changes associated with BECCS. Agriculture remains a critical sector, with substantial non-CO₂ emissions despite reductions. The study acknowledges limitations in representing transformative changes in integrated assessment models, particularly social dynamics and individual preferences. However, despite these simplifications, the results provide valuable insights into the critical relationship between HtA sectors and CDR.

This study demonstrates that significant reductions in hard-to-abate emissions are achievable, substantially reducing reliance on carbon dioxide removal technologies. Demand-side measures, particularly dietary shifts, show great promise in limiting BECCS deployment and achieving climate goals. While challenges remain in implementing these measures, the results offer alternative pathways toward net-zero emissions that may be more sustainable than those heavily reliant on CDR. Future research could focus on detailed cost-benefit analyses and assessments of policy feasibility for different regions.

The study acknowledges limitations of using global integrated assessment models to fully represent transformative change. The scenarios assume full effectiveness of demand-side measures, without accounting for all social dynamics and individual preferences. The model simplifies complex interactions within the HtA sectors and does not fully capture all potential environmental and social trade-offs associated with various mitigation strategies. The optimistic assumptions regarding technological advancements may also affect the generalizability of results.