2 °C and SDGs: united they stand, divided they fall?

There is hope that 2015 will be remembered as a defining year for setting human development on a more sustainable pathway. Two important milestones were reached. On 25 September, a new development agenda was adopted in New York aimed at eradicating poverty and facilitating inclusive development within ever tighter planetary boundaries. Economic, social and environmental progress will be tracked across a set of agreed sustainable development goals (SDGs). The SDG framework is intended to manage trade-offs and maximize synergies across the 17 different goals and associated 169 targets (Griggs et al 2013).

On 12 December, countries agreed upon a new international climate treaty, the Paris Agreement, at the United Nations Framework Convention on Climate Change (UNFCCC) Conference of Parties (COP21) in Paris. It 'aims to strengthen the global response to the threat of climate change, in the context of sustainable development and efforts to eradicate poverty, including by holding the increase in the global average temperature to well below 2 °C above pre-industrial levels' (UNFCCC 2015a).

Both processes are highly interrelated: SD is an explicit part of the Paris Agreement, while avoiding dangerous climate change features as one of the SDGs (#13). In fact, failure in one process would undermine the success of the other. Stringent and sustained mitigation is a necessary condition for SD, because unabated climate change will exacerbate many of today's development issues and negate future improvements (see Fleurbaey et al 2014). However, it is an insufficient condition for SD, because some 2 °C pathways could, if not designed properly, undermine SD in non-climate dimensions. For example, pathways with a limited short-term ambition like the current INDCs may have higher SD risks than more ambitious ones. Such broader SD implications could delegitimize some 2 °C pathways or even the 2 °C target itself (Edenhofer and Kowarsch 2015). SD further hinges on the successful implementation of non-climate policies that complement or support climate policies in other dimensions. Thus, identifying socially acceptable 2 °C pathways requires framing climate policy in a broader SD context.

Assessments of alternative mitigation pathways so far have mainly focused on characterizing the underlying technological and economic challenges (Clarke et al 2014), but less is known about the wider social, economic and environmental implications. For example, many 2 °C pathways project large amounts of bioenergy demand in the second half of this century. It is highly debated in the literature whether these can be provided sustainably: food security, place-specific livelihoods, water availability and biodiversity are amongst the critical issues being discussed (Creutzig et al 2012, Smith et al 2014). At the same time, many 2 °C pathways project potential health gains and co-benefits for other sustainability objectives. The balance of these co-effects is poorly understood, particularly on the supply side, because risks of alternative 2 °C pathways for non-climate sustainability objectives have not yet been systematically analyzed (von Stechow et al 2015).

In this letter, we analyze the implications of alternative 2 °C pathways for SD risk dimensions by drawing on existing, publicly available inter-model comparison results from integrated energy-economy-climate models—henceforth referred to as integrated models (see SI section 1, available at stacks.iop.org/ERL/11/034022/mmedia). We demonstrate how broadening the analytical framework can allow both for a more informed public debate about alternative 2 °C pathways and how achieving the climate SDG may affect the prospects of meeting other energy-related SDGs. This is important both for critically discussing the relationship between the international climate policy and SDG agendas as well as for identifying stringent mitigation pathways that are socially acceptable.

Choosing appropriate climate policies is an exercise in risk management for which it is key to understand and evaluate relevant uncertainties (Kunreuther et al 2013). We focus on uncertainties related to different model structures and assumptions, i.e. 'model uncertainty' (Drouet et al 2015) and draw on results from a structured inter-comparison exercise of integrated energy-economy-climate models, AMPERE (Kriegler et al 2015, Riahi et al 2015). To complement existing literature, this data is used to assess relevant SD implications of alternative clusters of mitigation pathways that are consistent with the 2 °C target (see table S2) to initiate a public debate on their wider sustainability implications.

2.1. Choice of indicators for SD risks

2.2. Choice of scenario data

The analysis is divided into two parts: we assess co-benefits of alternative 2 °C pathways before turning to their mitigation risk profiles. In each part, we systematically analyze different clusters of 2 °C pathways to understand the implications for SD outcomes of variations in (i) short-term climate policy stringency, (ii) availability of mitigation technologies or (iii) a combination of the two. Analyzing these clusters is highly relevant, because the current and projected INDC emission trajectories are not consistent with optimal 2 °C pathways (UNFCCC 2015b) and the standard assumption of full technological flexibility is inhibited as significant upscaling of low-carbon technologies faces many different hurdles in practice⁸ . Our analysis here focuses on the first half of the 21st century in which the interaction of short-term climate policies and the long-term climate target is strongest (Kriegler et al 2013, Luderer et al 2013a, 2013b, Riahi et al 2015, Eom et al 2015, Bertram et al 2015a).

3.1. Synergies across mitigation and sustainable energy objectives

Figure 1 uses cumulative indicators for (i) CO₂ emissions (Zickfeld et al 2012), (ii) the co-emitted air pollutants black carbon (BC) and sulphur dioxide (SO₂) and (iii) global oil extraction and trade as well as transport sector reliance on oil to present reduced SD risks, i.e. co-benefits of mitigation scenarios compared to baseline developments. Figure 1 shows that co-benefits in terms of lower ocean acidification, health and oil security increase relative to optimal 2 °C pathways by limiting the availability of key mitigation technologies, though considerable differences exist for different technologies and different sustainable energy objectives. This is for three main reasons:

• (i)  
  The unavailability of low-carbon technologies limits long-term mitigation potential, resulting in greater near-term emissions reduction requirements to meet a particular long-term climate goal. This leads to a decrease in fossil fuel use in the medium term (with lower cumulative global oil trade, oil extraction as well as transport sector reliance on oil) and the associated CO₂ emissions and co-emitted air pollutants. Limiting technologies that play a smaller role in reaching the long-term goal results in less dramatic transition requirements and fewer additional co-benefits.
• (ii)  
  When relying less on bioenergy and/or CO₂ capture and storage (CCS) technologies, the models are forced to switch more rapidly from fossil fuels to solar, wind and nuclear energy, which have higher co-benefits for air quality and oil security (Bruckner et al 2014, Hertwich et al 2015).
• (iii)  
  Limiting the deployment of bioenergy or CCS technologies that are associated with co-emitted air pollutants themselves (see SI section 3.1.9) additionally reduces air pollutant emission levels—which is not the case for limiting the availability of non-combustible RE or new nuclear capacity.

Zoom In Zoom Out Reset image size

Admittedly, these results only cover a small subset of potential co-benefits from mitigation. However, the literature suggests that this finding may apply more broadly (see von Stechow et al 2015 for a review and synthesis): climate policy that leads to less fossil fuel use and energy demand growth in the near term drives a broad range of co-benefits beyond air quality and oil security, such as reduced water use and pollution, reduced ecosystem impacts, reduced health impacts (also due to more physical activity under changed mobility patterns and less fuel poverty in insulated housing) as well as more local employment opportunities.

Comparing optimal 2 °C pathways with scenarios assuming weak short-term climate policies confirms the positive effect of stringent mitigation in the near term on the magnitude of co-benefits (see figure S5 for the year 2030): weak short-term climate policies imply a reduction in co-benefits relative to those that could materialize in optimal 2 °C pathways. This effect is, however, not as obvious for cumulative 2050 values (see figure 1) because some of the additional mitigation efforts in the period 2030–2050 partially compensate for weak climate policies until 2030. Since the transport sector is characterized by faster capital turnover rates (at least with regard to the vehicle fleet) (Bertram et al 2015a), it can react more quickly to carbon price changes, compensating for higher emissions from sectors that are less flexible. This may lead, for example, to a higher fuel diversity in the transport sector in the year 2050 in delayed mitigation scenarios compared to optimal 2 °C pathways albeit at high uncertainty.

3.2. Trade-offs between mitigation and sustainable energy objectives

While constraining a particular mitigation technology may minimize the mitigation risks specific to that technology, it usually implies an increase in the deployment of other low-carbon technologies, which may incur other mitigation risks. Figure 2 shows that limiting the availability of specific technologies in 2 °C pathways with immediate global climate policies substantially increases the risk of not meeting other sustainable energy objectives. While the unavailability of CCS and limitation of bioenergy potential lead to the largest co-benefits (see figure 1), they also entail significantly higher SD risks. This can be explained by the promise of greater flexibility in near-term emission pathways that are still able to meet the long-term climate goal through the presence of carbon dioxide removal technologies, such as bioenergy with CCS (BECCS). Constraining BECCS deployment by limiting the global bioenergy potential or ruling out CCS deployment results in substantially higher deployment of other mitigation technologies in the medium term. The increase is much less pronounced for limiting the potential for solar and wind energy or assuming no new nuclear capacity (see figure S6).

Zoom In Zoom Out Reset image size

Due to the different nature of the mitigation risks, it is unclear how decreasing risks in one dimension (e.g. bioenergy expansion or environmental risks associated with CCS deployment), can be traded off with risk increases in others (e.g. transitional growth reduction, energy price growth, nuclear proliferation or the technological challenges of integrating high amounts of fluctuating RE into existing power grids in a very short time frame). For example, a 20%–30% increase in energy prices may have a much more immediate, adverse effect on the poor in many countries than a 4-7-fold increase in maximum decadal upscaling of variable renewable energy sources, which is primarily a technological and institutional challenge for infrastructure provision. Rather than aggregating effects across different risk dimensions, the purpose of this analysis is to make the trade-offs across alternative clusters of mitigation pathways transparent. Hence, the way the climate SDG is met can substantially alter the risks of not meeting other SDGs and sustainable energy objectives.

This is confirmed by figure 3: delaying stringent mitigation in the near term leads to a significant increase in mitigation risk levels in the medium term compared to optimal 2 °C pathways. With more GHG emissions before 2030, subsequent reductions are more expensive (Luderer et al 2013b) and need to be faster to stay below 2 °C (Eom et al 2015)—with implications for the grid integration of fluctuating RE (see SI section 3.1.6) and for stranded investments in coal capacity (Johnson et al 2015) and the associated job losses (Rozenberg et al 2014). The carbon lock-in effect hence manifests itself particularly in technological and economic risk dimensions. To a lesser degree, these effects can also be seen for delayed mitigation scenarios with more optimistic assumptions about short-term climate policies (see figure S7). Hence, delaying stringent mitigation implies forgoing potential paths with lower risks along multiple SD dimensions.

Zoom In Zoom Out Reset image size

In contrast, assuming lower energy demand growth entails mitigation risk reductions relative to optimal 2 °C pathways (see figure 3). As each unit of energy not produced is free of pervasive supply-side risks, reducing energy demand by promoting energy efficiency in end-use sectors (e.g., consumer appliances), lifestyle changes (e.g., people living in higher-density areas and eating less dairy and meat) and structural changes in the economy (e.g., shifting to more service-oriented economies) is an important strategy both for mitigation and other sustainable energy objectives (von Stechow et al 2015).

Note that these reductions in energy demand growth are assumed to happen in the baseline scenarios, i.e. independent of the mitigation efforts and hence without a cost mark-up; it is unclear how future energy demand levels would develop under real-world conditions where clean energy and energy efficiency projects may compete for limited funds (McCollum et al 2013b). Furthermore, the models do not simply prescribe lower energy supply at the expense of energy service supply, but alter assumptions on the average energy intensity improvement rates and, e.g., on the viability of more compact, public transit-friendly urban areas (Riahi et al 2015). This does not imply, however, that all integrated models project final energy supplies in mitigation scenarios that are consistent with minimum thresholds of energy consumption to satiate basic needs related to cooking, heating, health and other infrastructure (Steckel et al 2013, Lamb and Rao 2015). Hence, projections of energy demand from individual models need to be interpreted with care (see discussion in SI section 2).

3.3. Trade-offs between mitigation and sustainable energy objectives for combined model constraints

As current GHG emission trends keep tracking along business-as-usual (Edenhofer et al 2014a) and societal concerns grow with regard to upscaling of many low-carbon technologies (see footnote 8), 2 °C pathways with multiple constraints seem to mirror most closely developments observed in the real world. In fact, delaying stringent mitigation in combination with technological constraints risks no longer meeting the climate goal (Riahi et al 2015), substantially increases mitigation risks (see figure 4) and increasingly jeopardizes our ability to manage risk trade-offs. For CCS and bioenergy whose unavailability/limitations already show substantial risk trade-offs in immediate mitigation scenarios, most models can no longer find a solution (for CCS unavailability only DNE21+ and GCAM; for limited global bioenergy potential only GCAM, POLES, and REMIND) implying a high risk of not meeting the 2 °C target.

Zoom In Zoom Out Reset image size

Figure 4 draws on AMPERE scenarios with multiple constraints but shows results for more optimistic—albeit not optimal — short-term climate policies⁹ , with and without limited global bioenergy potential. As models work close to their feasibility frontier, the additional constraint results in large mitigation risk increases. Even for non-biomass RE and nuclear energy, whose limitation/phase-out has rather small effects in immediate 2 °C pathways, risk trade-offs increase strongly for delayed mitigation scenarios in some dimensions (see figures S7 and S8).

This letter presents a first attempt to shed light on the question of how alternative 2 °C pathways perform in non-climate SD dimensions and to draw conclusions about important interactions between stringent mitigation and other sustainable energy objectives. Figure 5 shows an overview of the different clusters of constrained 2 °C pathways relative to (each model's) optimal pathways (i.e., those with immediate mitigation, full technology portfolios, and conventional energy demand growth). We use 'optimal' scenarios as benchmarks because they show comparatively balanced risk profiles relative to baseline developments (see figures 2– 4) and because they are commonly used as reference point for policy analysis, e.g. in the WGIII AR5 (Edenhofer et al 2014a). This enables the comparison of the various SD implications of one cluster of 2 °C pathways to those of all others and therefore facilitates an informed public debate on socially acceptable SD risks and thus the interaction between the international climate policy and the broader SDG agendas.

Zoom In Zoom Out Reset image size

Note that 'optimal' pathways are not necessarily the most socially desirable because they may already involve unacceptable risks. Scientific analysis alone cannot judge whether a particular 2 °C pathway poses acceptable or unacceptable risks to society (Edenhofer and Minx 2014). Science can, however, explore alternative mitigation pathways and inform an enlightened public debate across SD risk dimensions in an iterative learning process (Edenhofer and Kowarsch 2015). For example, annual bioenergy supply is projected to reach up to 168 EJ (median: 158 EJ) in 2050 in optimal scenarios. These levels of biomass extraction may already be associated with fundamental challenges with respect to food security, place-specific livelihoods, water availability and biodiversity (Creutzig et al 2012, Smith et al 2014). These numbers further increase substantially over the second half of the century, reaching up to 862 EJ (median: 268 EJ) with growing requirements for removing CO₂ from the atmosphere via bioenergy with CCS (BECCS) technologies in many available scenarios (Clarke et al 2014). Many 'optimal' 2 °C pathways have therefore been challenged on these grounds (Fuss et al 2014, Smith et al 2016).

In a world which is increasingly unlikely to develop along 'optimal' scenario trajectories, an informed public debate about synergies and risk trade-offs implied by alternative clusters of constrained 2 °C pathways is key for identifying those which are socially acceptable. For example, current INDCs at best add up to emission trajectories similar to those 2 °C pathways with low short-term ambition ('LST' scenarios, see table 3)¹⁰ . According to figure 5, these pathways (presented as circles) not only lead to fewer co-benefits compared to optimal 2 °C pathways (except for cumulative BC emissions and transport sector oil reliance) but also to significantly higher mitigation risk levels, particularly in socioeconomic dimensions—with higher risks of not meeting those SDGs related to economic growth, energy access, job preservation, food security and resilient grid infrastructure (see also figure S7).

When a technology constraint is added, only the risks specific to that technology can be lowered (e.g. reduced nuclear proliferation risks for scenarios with no new nuclear capacity or fewer grid integration challenges for scenarios with limited potential for solar and wind energy, see also figures S8 and S9). The other risk levels are exacerbated, particularly for those SDGs that relate to economic growth, job preservation, resilient infrastructure, and ocean acidification. This is particularly obvious for scenarios with limited global potential of bioenergy in which the risks related to bioenergy expansion are lower (including environmental effects related to BECCS deployment) but the risks of not meeting socioeconomic SDGs are significantly higher (see green circles in figure 5). Limiting the global use of bioenergy to 100 EJ per year by 2050—widely believed to be more sustainable (Creutzig et al 2014)—hence introduces a trade-off with socioeconomic objectives for weak short-term climate policies (see green circles in figure 5).

While there are uncertainties around acceptable levels of bioenergy deployment, the development and deployment of CCS technology is lagging behind expectations (IEA 2009), despite its important role in keeping mitigation costs at relatively low levels (Edenhofer et al 2014a). Our results highlight two things: first, those models that are flexible enough to compensate for the unavailability of CCS can only do so with increased upscaling requirements for other low-carbon technologies and related SD risks (see pink circles in figure 5). This also implies high near-term mitigation requirements with associated co-benefits. Second, the absence of CCS seriously questions the achievability of the 2 °C target in a world with delayed climate action and therefore threatens the climate SDG itself—only two models can report results for the combination with weak short-term climate policies.

In contrast, 2 °C pathways with lower energy demand growth generally entail a substantial reduction in SD risk levels (blue shapes in figure 5). This confirms results from a bottom-up assessment of the wider SD implications of technology-specific studies from a cross-sectoral perspective (von Stechow et al 2015). While these scenarios typically do not feature many additional co-benefits due to lower supply-side transition requirements, achieving lower energy demand growth has considerable synergies with the SDG agenda related to economic growth, food security, resilient grid infrastructure as well as with the peaceful use of nuclear energy. Delaying mitigation in scenarios with low energy demand growth only entails moderate risk increases—although some co-benefits are reduced and more coal capacity is likely to be retired early. Pursuing aggressive energy efficiency improvements across all sectors and rethinking high-energy lifestyles therefore seems essential to increase synergies and keep the trade-offs across SDGs manageable in a world that is characterized by multiple constraints. Unfortunately, model inter-comparison projects have not yet analyzed the combination of technology constraints and low energy demand growth pathways, which is a promising research area to better understand synergies between SDGs. Future research should also ensure that mitigation scenarios are consistent with minimum thresholds of energy demand necessary to satiate basic human needs (see discussion in SI section 2).

This letter has analyzed the changes in SD risks across alternative 2 °C pathways. These effects depend to a great extent on the development context, i.e., assumptions about baseline developments (Moss et al 2010, O'Neill et al 2014). To circumvent this potential caveat, the analysis used AMPERE data that stands out in its comprehensive effort to harmonize future socio-economic drivers of SD across models in the baseline scenarios: e.g., regional-level gross domestic product (GDP), population, and energy demand growth. This makes the results more comparable across models but begs the question of how the results would have changed for alternative assumptions beyond changes in energy demand growth. Research can and should build on alternative baseline developments as expressed by the 'shared socioeconomic pathways' (O'Neill et al 2014) that will soon be published even though important, non-trivial discussions remain on how SDGs can be adequately built into these baselines (O'Neill et al 2015).

Indicators that were used to track the changes in SD risks are only rough and sometimes very rough approximations of individual SDGs. There is no doubt that individual models—particularly those coupled to a detailed agro-economic and land-use model—could already provide better indicators, such as for water availability and ecosystem impacts which are important concerns in stringent mitigation pathways (see SI section 3.1.1). However, these have not yet been analyzed in a multi-model study (von Stechow et al 2015). We believe that such inter-model comparison results are crucial for a meaningful public debate about SD risks.

Another important caveat of the analysis is that we focus on 2050 and the preceding decades when looking at the implications of alternative 2 °C pathways for SD risk dimensions. The risks of some 2 °C pathways, however, only unfold later in that century when some particularly risky negative emissions technologies, such as BECCS, are being deployed at large scale to compensate for lower mitigation efforts in the first decades and residual GHG emissions in other sectors (Fuss et al 2014, Smith et al 2016). For illustrative purposes, figures S10 and S11 show how mitigation risks change from 2050 to 2080 for scenarios with substantially different amounts of negative emission requirements. Since the AMPERE scenario specifications do not allow for a meaningful comparison across scenarios with low or high amounts of negative emissions, we use the amount of radiative forcing overshoot to cluster scenarios with respect to their dependence on negative emissions (also used in the WGIII AR5 scenario database, see Krey et al 2014). It shows that the magnitude of the mitigation risk levels can change substantially over time for those dimensions that are related to negative emission technologies such as CCS and bioenergy deployment.

Our analysis points to important future challenges: first, the chosen indicators do not represent all SDGs as some touch on socio-cultural and institutional aspects which are challenging—if not impossible—to represent in an economic model framework (see SI section 2). Second, the changes in the indicators across scenarios are merely indicative for the change in risks to meet the related SDGs and sustainable energy objectives because there are many more relevant drivers that cannot be analyzed based on the available scenario data. Third, many relevant issues play out at lower geographic and time scales which are difficult to represent adequately in global-scale integrated models. For example, food security is driven by many socioeconomic drivers both on global and local scales and bioenergy expansion represents but one of those (Tscharntke et al 2012). And according to Creutzig et al (2012), the models are not (yet) suitable for operationalizing important global SD dimensions of bioenergy supply such as the socioeconomic convergence across different countries. Nevertheless, we argue that the indicators used in this letter are relevant for evaluating additional pressure on the energy-economy-climate system from additional constraints as represented in the models. As such, they supply important information from internally consistent model frameworks taking into account inter-sectoral and inter-regional interactions (von Stechow et al 2015 and SI section 1).

We provide this early contribution to a public debate on the relationship between the international climate policy and the SDG agendas based on existing multi-model scenario data that was not specifically developed for this particular purpose. This stimulus seems important because results from model inter-comparisons that are tailored towards the SDG-climate nexus will not be published for some years. Only by working with the available data can we start discussing relevant (risk) trade-offs and synergies. Based on our analysis, we argue that SD considerations are central for determining socially acceptable climate policies and that the prospects of meeting other SDGs need not dwindle and can even be enhanced for some goals if appropriate climate policy choices are made. Moreover, experiences and caveats of this analysis can help guide future research efforts at a relevant moment in time when new model comparison exercises are being designed. For example, to remain policy-relevant, SDG-focused multi-model comparisons will need to address inequality, poverty, and basic human needs as major drivers of the policy process much more adequately. This requires a serious discussion, e.g., on how to deal with the coarse regional disaggregation in the integrated modelling frameworks. Equally, successful efforts to address SDG-relevant issues in one model, e.g., for the analysis of water availability or ecosystem impacts (see SI section 2), will need to be lifted into a multi-model context.

Until now, no multi-model study has been used to systematically analyze the changes in SD risks implied by stringent mitigation scenarios and evaluate them across a set of SDGs. This letter addresses this research gap by analyzing a comprehensive set of alternative clusters of 2 °C pathways consistently formulated across many integrated models from the AMPERE model inter-comparison study, drawing on publicly available scenario results to calculate indicators for global SD risks. We shed light on the implications of alternative clusters of 2 °C pathways for meeting a set of energy-related SDGs and other sustainable energy objectives and to inform the public debate about the synergies and trade-offs across the international climate policy and the SDG agendas.

Our analysis shows that the near-term choice of 2 °C pathways has implications for the extent of synergies and trade-offs across energy-related SDGs in the medium term. Given current trends in emissions and technology deployment, we argue that mitigation pathways are likely to be characterized by multiple constraints. But adding limits on the availability of specific mitigation technologies on top of weak short-term climate policies decreases synergies and locks in substantial trade-offs across environmental and socioeconomic objectives. From an SDG perspective, the challenges of meeting other sustainable energy objectives substantially change with the way the climate SDG will be met. In some cases, meeting the 2 °C target is even threatened itself. Achieving low-energy demand growth, e.g., through aggressive energy efficiency improvements, helps to manage these trade-offs and attain multiple energy-related SDGs together. We find the greater the constraints on flexibility in meeting the 2 °C target, the higher the risks of not meeting other SDGs and the flexibility to manage these risks. Governments at all levels need to be informed about such implications of their collective decision for the attainability of global SDGs. This could avoid additional pressures on the sustainability of each region's development pathway.

After COP21, decision makers need to rethink their commitment to the SDG agenda, given that the short-term ambition for mitigation action falls short of the mitigation efforts consistent with staying below 2 °C in a cost-effective way. According to our results, this is likely to decrease co-benefits and increase the risks for attaining energy-related SDGs and other sustainable energy objectives. Since many of these SD risks are best dealt with at the global level, however, they might be good entry points into additional incentives for international cooperation. We suggest that the review of INDCs should provide for an assessment of policies at all scales to monitor global risks for non-climate sustainability objectives that arise from specific global mitigation pathways. Monitoring these risks could avoid unintended consequences (which might even delegitimize the 2 °C target), finding new entry points for global cooperation and providing rationales for ramping up mitigation ambition in the short to medium term.

Future research should extend the current system boundaries and, based on a comprehensive review of model literature on the climate-SDG nexus, establish indicators that help evaluate integrated policies addressing multiple SDGs in a unified framework. This would be a prerequisite for model inter-comparison projects with a focus on the interactions across multiple SDGs that could result in meaningful and robust results for better decision making. Climate policy will not be successful unless it seriously considers other policy objectives and therefore wider SD implications. Dividing the huge effort of achieving more sustainable development pathways into isolated policy problems will fall short of reaping synergies and successfully managing trade-offs across the many SDGs.

We are grateful to Ad-Willem Dashorst, Michael Jakob, Jan C Steckel, Nils Johnson, and Felix Creutzig for helpful comments on earlier versions of this manuscript. We acknowledge the work by integrated model teams that contributed to the AMPERE scenario database and thank IIASA for hosting the AMPERE scenario database. The AMPERE scenario work received funding from the European Union's Seventh Framework Programme (FP7/2007-2014) under grant agreement No. 265 139 (AMPERE). The research leading to this publication was supported by the German Federal Environment Agency (UBA) under UFOPLAN FKZ 3714 411 670 as well as the European Union's Horizon 2020 research and innovation programme under grant agreement No. 642 147 (CD-LINKS).