Abstract

The 2015 Paris Agreement calls for countries to pursue efforts to limit global-mean temperature rise to 1.5 °C. The transition pathways that can meet such a target have not, however, been extensively explored. Here we describe scenarios that limit end-of-century radiative forcing to 1.9 W m−2, and consequently restrict median warming in the year 2100 to below 1.5 °C. We use six integrated assessment models and a simple climate model, under different socio-economic, technological and resource assumptions from five Shared Socio-economic Pathways (SSPs). Some, but not all, SSPs are amenable to pathways to 1.5 °C. Successful 1.9 W m−2 scenarios are characterized by a rapid shift away from traditional fossil-fuel use towards large-scale low-carbon energy supplies, reduced energy use, and carbon-dioxide removal. However, 1.9 W m−2 scenarios could not be achieved in several models under SSPs with strong inequalities, high baseline fossil-fuel use, or scattered short-term climate policy. Further research can help policy-makers to understand the real-world implications of these scenarios.

Main

Scenarios of the energy–economy–land system can facilitate the integrated assessment of the impacts and mitigation of climate change. For the Fifth Assessment Report (AR5) of the Intergovernmental Panel on Climate Change (IPCC), four Representative Concentration Pathways1 (RCPs) have provided climate researchers with a set of consistent climate forcings2,3,4. More recently, the Shared Socio-economic Pathways (SSPs) have been developed5,6. SSPs provide a socio-economic dimension to the integrative work started by the RCPs7. This framework provides a basis of internally consistent socio-economic assumptions that represent development along five distinct storylines8: development under a green-growth paradigm9 (SSP1); a middle-of-the-road development along historical patterns10 (SSP2); a regionally heterogeneous development11 (SSP3); a development that results in both geographical and social inequalities12 (SSP4); and a development path that is dominated by high energy demand supplied by extensive fossil-fuel use13 (SSP5).

Prior to 2015, international climate policy under the United Nations Framework Convention on Climate Change focused on the goal of keeping the global-mean temperature increase below 2 °C relative to pre-industrial levels14. The Paris Agreement reset this long-term goal to holding the increase well below 2 °C and pursuing efforts to limit it to 1.5 °C15. In this study, we present a set of stringent climate change mitigation scenarios consistent with an increase of 1.5 °C in 2100. Six integrated assessment models were included in this study (AIM, the Asia–pacific Integrated Model11; GCAM4, the Global Change Assessment Model12; IMAGE, the Integrated Model to Assess the Global Environment9; MESSAGE-GLOBIOM, the Model for Energy Supply Strategy Alternatives and their GeneralEnvironmental Impact combined with the Global Biosphere Management Model10; REMIND-MAgPIE, the Regionalized Model of Investments and Development combined with the Model of Agricultural Production and its Impact on the Environment13; and WITCH-GLOBIOM, the World Induced Technical Change Hybrid model combined with GLOBIOM16), with which we attempted to model scenarios that limit end-of-century radiative forcing to 1.9 W m−2 under various SSPs (hereafter called ‘SSPx–1.9’ scenarios, with SSPx indicating the specific SSP assumed by the scenario and 1.9 the radiative forcing target in 2100, Methods). This scenario set allows the structured exploration of climate change at a level consistent with limiting the global-mean temperature increase in 2100 to 1.5 °C with approximately 66% probability (see Fig. 1 and results described below). Overall, all teams were able to produce 1.9 W m−2 scenarios in SSP1, and four teams were successful in SSP2. Of the three and four modelling frameworks that attempted to model 1.9 W m−2 scenarios in SSP4 and SSP5, one and two were successful, respectively (see Methods, Supplementary Table 1, Supplementary Fig. 1, Supplementary Text 2). From this set of 1.9 W m−2 scenarios, a further, stringent climate mitigation scenario has been selected for inclusion in the Scenario Model Intercomparison Project17 (ScenarioMIP) of the Sixth Phase of the Coupled Model Intercomparison Project18 (CMIP6), as well as other CMIP6 MIPs (for example, refs 19,20, Fig. 1a, Supplementary Text 1, Methods).

Fig. 1: Emission and temperature characteristics of 1.9 W m−2 scenarios under various SSPs.
Fig. 1

a, Global CO2 emissions of SSP scenarios with the selected CMIP6 ScenarioMIP subset highlighted. Historical emission from ref. 52. All other panels show 1.9 W m−2 scenario data only. b, Global Kyoto GHG emissions. Shaded areas show the range per SSP, solid lines the marker scenarios for each SSP and dashed lines single scenarios that are not markers. Single model detail is provided in Supplementary Fig. 2. c, Non-CO2 GHGs per scenario in 2100. d, Exceedance probability of various temperature limits for the 1.9 W m−2 scenarios with bars showing the full range over all available scenarios per SSP. Except for the first sub-panel, all other panels give the exceedance probability over the entire twenty-first century. e, Probability of peak warming versus 2030 GHG emissions in 1.9 W m−2 scenarios. f, Dependence of cumulative CO2 emissions on non-CO2 radiative forcing in 2100.

Emission and climate-related outcomes

CO2 and other greenhouse gas (GHG) emissions peak before 2030 and decline rapidly over the next two to three decades in SSPx-1.9 scenarios (Fig. 1, see Supplementary Figs. 26 for other emissions). By 2050, annual CO2 and GHG emissions are in the range of −9–6 and 1–13 billion tons of CO2-equivalent emissions (gigaton GtCO2e yr−1, Methods), respectively, across all available scenarios. Underlying these reductions is a phase-out of industry and energy-related CO2 production at a rate of 0.2–7.1% yr−1 (median: 3.0% yr−1, see Supplementary Tables 2, 3 for a complete overview), combined with rapid upscaling of carbon capture and storage (CCS) and carbon-dioxide removal (CDR, see section on system transformations below). Near-term emissions vary across the SSPs, because, in contrast to SSP1, the effectiveness of near-term climate policies is assumed to be limited in other SSPs (defined by so-called Shared Policy Assumptions5,21). In that case, global mitigation is regionally scattered and accelerates slower over the next few decades,requiring it to accelerate faster later on.

All scenarios presented here lead to 1.9 W m−2 radiative forcing in 2100 within rounding precision (Supplementary Fig. 7), but they differ in their likelihood of limiting warming below specific temperature levels. All scenarios keep warming to below 2 °C with more than 66% probability (Fig. 1d), and maximum (peak) median temperature estimates vary from 1.5 °C to 1.8 °C. Near-term mitigation has a determining role here: higher 2030 emissions come with a temperature penalty (Supplementary Fig. 8). The probability of limiting peak warming to below 1.5 °C relative to pre-industrial levels is approximately halved and peak temperature about 0.2 °C higher if emissions are at the high (>45 GtCO2e yr−1) instead of the low (<30 GtCO2e yr−1) end of the available range in 2030 (Fig. 1e). By 2100, this variation disappears and all scenarios limit warming below 1.5 °C with about 66% probability (Supplementary Figs. 8, 9). Whether these pathways provide an acceptable interpretation of the Paris Agreement long-term temperature goal is not a scientific but a political question22,23, which we do not address here.

Across all 13 available scenarios, net zero GHG emissions are reached around 2055–2075 (rounded to the nearest 5 years). Net zero CO2 emissions are reached earlier (Supplementary Table 2). The year of reaching net zero GHG emissions is inversely correlated with emissions in 2030. For example, scenarios with 2030 GHG emissions higher than 40 GtCO2e yr−1 reach global net zero GHG emissions before 2060 (Supplementary Fig. 10). Cumulative CO2 emissions over the 2016–2100 period range from −175 to 475 GtCO2 (SSP2 median: 250 GtCO2, rounded to the nearest 25 GtCO2). End-of-century non-CO2 radiative forcing strongly influences the variation across this range24 (Fig. 1f). These values are consistent with earlier published estimates2427 (Supplementary Text 3) and lead to 2100 atmospheric CO2 concentrations in the 350–390 p.p.m. range. Potential feedbacks that are currently not included, such as CO2 and CH4 release from permafrost thawing or changes in other natural sources, can reduce carbon budgets further28,29 and therefore alter the presented climate outcomes.

Even in these very stringent mitigation pathways, sizeable remaining CH4 and N2O emissions are projected by all models (Fig. 1c, Supplementary Fig. 6), and in 2100, respectively, 53–85% and 59–95% of these emissions originate from agriculture. The uncertainty in CH4 and N2O emissions is large with inter-model variations dominating inter-SSP variations. High and low estimates for 2100 differ by a factor of 2–3, mainly owing to uncertainties in how emissions from agriculture are treated and can be mitigated in different models30,31. Important uncertainties also remain in the CO2 mitigation contribution of the land-use sector31 (Supplementary Fig. 5). Here, emissions decline over the long term, but whether and to what degree the land-use sector becomes a global sink is very model-dependent (Supplementary Text 4).

System transformations

Achieving pronounced emission reductions requires a transformation of the global economy. Previous studies have discussed the implications of such a global transformation for the energy and land-use system32, highlighting the importance of limiting future energy demand32 to keep warming to below 1.5 °C and of changing consumption patterns33 combined with sustainable intensification of agriculture34. We here focus on confirming these characteristics and exploring the extent to which they vary across SSPs.

All 1.9 W m−2 scenarios in this study strongly limit energy demand growth (Fig. 2d, Supplementary Fig. 11), with energy intensity reduction rates of 2–4% yr−1 from 2020 to 2050 (Fig. 2d). In SSP2, final energy demand in 2050 is limited to 10–40% above 2010 levels (rounded to the nearest 5%). This compares to 10% below to 30% above, and 45–75% above 2010 levels in SSP1 and SSP5, respectively. Energy conservation is therefore a common strategy in stringent mitigation scenarios, but it also has limits.

Fig. 2: Overview of key decarbonization characteristics in 1.9 W m−2 scenarios.
Fig. 2

a, Primary energy from non-biomass renewables (wind, solar, hydro and geothermal energy). b, Primary energy from biomass with CCS (BECCS). c, Primary energy from coal without CCS. Shaded areas in ac show the range per SSP, solid lines the marker scenarios for each SSP and dashed lines single scenarios that are not markers. d, Three illustrations of global final energy demand in 1.9 W m−2 scenarios showing, from left to right, the average reduction from baseline over the 2020–2100 period, the change in 2050 compared to 2010 levels, and the annual rate of final energy intensity change. e, Global forest cover, and change relative to 2010 due to afforestation and reforestation in 2.6 and 1.9 W m−2 scenarios. f, Change in global cropland for agriculture in 2100 relative to 2010 in ‘Baseline’ scenarios in the absence of climate change mitigation, as well as in 2.6 and 1.9 W m−2 scenarios. Results are grouped per SSP (coloured lines with black symbols). rel., relative. Mha, million hectares.

Energy supply also has to be transformed to achieve reductions in deep emissions. This includes upscaling of bioenergy and renewable energy technologies, shifting away from freely emitting fossil-fuel use, and the deployment of CDR, such as Bioenergy with Carbon Capture and Sequestration (BECCS) or large-scale afforestation (see Supplementary Text 5 for a discussion of CDR in SSPx-1.9 scenarios). Non-biomass renewables (solar, wind, hydro and geothermal energy) scale up rapidly over the twenty-first century (Fig. 2a), reaching mid-century electricity shares of 60–80% and 32–79% in SSP1 and SSP2, respectively (Supplementary Fig. 12). In the marker SSP scenarios, these shares are 79%, 60% and 61% in SSP1, SSP2 and SSP5, respectively. Both solar and wind energy is projected to scale up consistently across the different SSPs (Supplementary Fig. 13). Particularly for wind energy, inter-model variations dominate over differences induced by different SSPs, a feature also present in less stringent mitigation pathways35 (Supplementary Table 4). SSP2 and SSP5 1.9 W m−2 scenarios see a strong upscaling of nuclear power, whereas in SSP1, and particularly its marker implementation, the contribution of nuclear energy use decreases compared to today’s levels (Supplementary Fig. 13).

Under all SSPs, 1.9 W m−2 scenarios show a clear shift away from unabated fossil fuels (that is, without CCS, Fig. 2c), and a phase-out of all fossil fuels. The marker implementations exhibit rapidly declining contributions of coal until 2040 (less than about 20% of its 2010 contribution in 2040), followed by a phase-out of oil until 2060 (Supplementary Figs. 14, 15). The potential contribution of natural gas to the primary energy mix is the most uncertain, with mid-century contributions ranging from 22 to 267 exajoules (EJ) yr−1 across all scenarios compared to about 100–110 EJ yr−1 in 2010. Differences in preferences for gas supply across models here dominate the variation in costs and availability assumptions owing to alternative socio-economic pathways (Supplementary Table 4, Supplementary Fig. 16).

Bioenergy is used in large amounts in all 1.9 W m−2 scenarios, and this can raise concerns for food security or biodiversity36,37,38. These concerns depend both on how and how much bioenergy is produced. Bioenergy demands can be met through dedicated energy crops or through residues. The latter option comes with fewer trade-offs than dedicated bioenergy crops38. Models, however, project very different shares for the use of residues (Supplementary Table 5), and further research clarifying its potential would be essential. For 2050, global technical bioenergy potentials (including energy crops and residues) were identified ranging from <50 to >500 EJ yr−1. High, medium and low agreement was attributed to potentials of 100, 300 and >300 EJ yr−1, respectively36. Bioenergy use is increased by 1–5% per year between 2020 and 2050 in 1.9 W m−2 scenarios. Total bioenergy use in 2050 is kept below about 300 EJ yr−1, and in most cases below 150 EJ yr−1 (Supplementary Fig. 17). In a green-growth SSP1 world, markedly lower bioenergy contributions are projected compared to an SSP2 world that continues the historical experience (34–112 EJ yr−1 lower in 2050). Putting this into context, scenarios project approximately 100 EJ yr−1 of bioenergy use (full range: 38–112, with important variations across SSPs) in baseline scenarios without any climate policy (Supplementary Fig. 17).

In 1.9 W m−2 scenarios, land for energy crops and forest area is generally projected to expand during the twenty-first century, with large variations across models, and this can impact land for agriculture and water availability39,40 (Fig. 2f, Supplementary Fig. 18). However, in SSP1 the decrease in agricultural land in 1.9 W m−2 scenarios is reasonably similar to what is projected in a no-climate-policy baseline merely owing to low demand for agricultural commodities and high agricultural intensification. Pasture is one of the activities most affected by expanding other land uses and declines robustly across models and SSPs (Supplementary Fig. 19). In the middle-of-the-road SSP2 world, pastures decreases by 1–20% in 2050 compared to 2010 levels, and in SSP1, pastures also decrease by 8–16%. In a fossil-fuel intensive SSP5 scenario, it declines by 15–25%. It is important to note that SSP1 baseline scenarios already project a pasture-land decrease of 1–11% due to shifts towards less meat-intensive diets, limited food waste and a return of the world population to 7 billion people by 21005,9,31. This reaffirms the important role that changes in food consumption in combination with sustainable intensification of agriculture have for stringent mitigation31,34,41.

Large-scale afforestation and reforestation can make an important contribution to the overall CDR effort. In the sustainable SSP1 world, pressure on land is relatively low, and the forest area in 2050 can therefore expand by 0–24% relative to 2010. However, in the middle-of-the-road SSP2 scenarios, results are mixed, with some models projecting forest area to decrease by 2% and others report an increase of up to 18%. SSP5 sees a change of 0–16% (Supplementary Table 6). Not all models explicitly include afforestation as a mitigation option and ranges therefore span results that are not fully comparable across models. However, in all 1.9 W m−2 scenarios climate policy leads to a net forest expansion compared to no-climate-policy baselines (Fig. 2e). Integrated policy packages are required that ensure food security is achieved together with climate change mitigation42.

BECCS contributes the largest part of CDR in 1.9 W m−2 scenarios (Supplementary Fig. 20). Between 150–1,200 GtCO2 (rounded to nearest 25 GtCO2), equivalent to about 4–30 years of current annual emissions, is removed from the atmosphere via BECCS during the twenty-first century, with important variation between models and across SSPs (Fig. 3a, d). SSP1 shows the lowest BECCS deployment over the twenty-first century (150–700 GtCO2) owing to its lower final energy demand and baseline emissions, compared to SSP2 (400–975 GtCO2) and SSP5 (950–1,200 GtCO2). None of the SSPx-1.9 scenarios explicitly attempted to limit the contribution from BECCS. The numbers reported here therefore represent projections of estimated cost-effective BECCS deployment in 1.9 W m−2 scenarios, but do not represent minimum BECCS requirements in a strict sense.

Fig. 3: BECCS, fossil–CCS and CCS across SSPs and across climate targets.
Fig. 3

a, Annual amount of CO2 stored by CCS in 1.9 W m−2 scenarios. Shaded areas show the range per SSP, solid lines the marker scenarios for each SSP and dashed lines single scenarios that are not markers. b, Variation per modelling framework and per SSP of cumulative CO2 stored by CCS during the twenty-first century when moving from a world in the absence of climate policy (baseline) to increasingly more stringent climate targets (6.0, 4.5, 3.4, 2.6 and 1.9 W m−2) c,d, As b but for fossil–CCS and BECCS, respectively. Note that axis limits vary across models.

Abated fossil fuels—that is, fossil fuels combined with CCS (fossil–CCS)—are often used in models as a bridging solution. However, fossil–CCS still results in residual CH4 emissions from coal mining or gas handling, and CO2 emissions due to imperfect capture and leakage. These emissions can become too substantial for very stringent mitigation transitions. Indeed, almost all 1.9 W m−2 scenarios deploy less cumulative fossil–CCS than weaker mitigation scenarios (Fig. 3c). Optimal 1.9 W m−2 strategies are therefore not merely ‘more of the same’. Overall, the BECCS share of total CCS increases (Supplementary Fig. 20). CDR is thus preferred over fossil–CCS in very stringent mitigation scenarios.

Differential mitigation

A previous study43 has identified characteristics of 1.5 °C pathways in comparison to 2 °C pathways. These characteristics were (i) greater mitigation efforts on the demand side; (ii) energy efficiency improvements; (iii) CO2 reductions beyond global net zero; (iv) additional GHG reductions mainly from CO2; (v) rapid and profound near-term decarbonization of energy supply; (vi) higher mitigation costs; and (vii) comprehensive emission reductions implemented in the coming decade. Using our 1.9 W m−2 and 2.6 W m−2 scenarios as proxies for 1.5 °C and 2 °C pathways, these characteristics still hold when assessed with four additional models and varying socio-economic assumptions (Fig. 4, Supplementary Text 6, and results above). None of the 1.9 W m−2 scenarios show a peak of emissions after 2020, and 82–98% of additional cumulative mitigation over the 2020–2100 period is achieved through CO2 reductions (Supplementary Fig. 21). Fig. 4 further illustrates the relatively stronger demand-side mitigation efforts in 1.9 W m−2 scenarios, particularly in the transport and building sectors (see also Supplementary Figs. 2224).

Fig. 4: Differential mitigation characteristics when moving from a 2.6 W m−2 SSP2 scenario to a 1.9 W m−2 scenario under three SSP assumptions (SSP1, SSP2, SSP5).
Fig. 4

Updated from ref. 43. Indicators are: long-term mitigation costs (2010–2100 aggregate consumption losses relative to baseline discounted at 5%); short-term mitigation costs (2010–2040 aggregate discounted at 5%); 2040 global emission-weighted equivalent carbon price level; electricity price in 2030; cumulative CDR between 2010 and 2100 including BECCS and CO2 removal by land use and land-use change; decarbonization pace (average linear 2010–2050 rate of reductions in energy-related CO2 emissions); reductions in CO2 emissions from electricity from baseline in 2050; reductions in CO2 emissions from industry from baseline in 2050; reductions in CO2 emission from transport from baseline in 2050; and reductions in CO2 emissions from buildings from baseline in 2050. Data are shown for the marker implementations of SSP1, SSP2 and SSP5. Ranges per SSP are provided in Supplementary Figs. 22–24. RF, radiative forcing. *Not available for all models.

Mitigation costs increase substantially between 1.9 and 2.6 W m−2 scenarios reflecting higher marginal abatement costs (Figs. 4,5). The relative carbon price increase is largest in SSP2 (Fig. 4) and also SSP1 sees large relative increases across all models (Supplementary Figs. 2224). However, in absolute terms, carbon prices (Fig. 5), consumption losses and energy supply mitigation investments (Supplementary Fig. 26) are highest when assuming the less favourable socio-economic conditions of SSP2, SSP4 and SSP5. For instance, the average discounted carbon prices (discounted to 2010 over the 2020–2100 period; Fig. 5) are estimated to be about 50–165 US$ per tCO2e in SSP2 (rounded to the nearest 5). They are approximately 35–65% lower in SSP1, and for the two reported SSP5 scenarios the change is −30% and +5%, respectively. The large range of carbon prices is mainly driven by model uncertainties, which were already identified for 2.6 W m−2 scenarios5, but are more pronounced here owing to the more stringent target.

Fig. 5: Variation of carbon prices over SSP and radiative forcing target space.
Fig. 5

Values are shown as average global average carbon prices over the 2020–2100 period discounted to 2010 with a 5% discount rate. Mitigation challenges are assumed to increase from left to right across the SSPs (that is, SSP1, SSP4, SSP2, SSP3, SSP5). Each box represents one model–SSP–radiative forcing target combination. A, AIM/CGE; G, GCAM4; I, IMAGE; M, MESSAGE-GLOBIOM; R, REMIND-MAgPIE; W, WITCH-GLOBIOM. All scenarios with a carbon price greater than 0 (that is, all but the baselines) have been designed to reach one of the radiative forcing targets on the vertical axis. Models for which no baseline data are indicated have baselines that result in an end-of-century radiative forcing between 6.0 and 8.5 W m−2.

Enabling and disabling factors

Our results show that some socio-economic developments and assumptions about policy effectiveness preclude achieving stringent mitigation futures (Fig. 5). Such failures were anticipated for SSP3, in which very heterogeneous regional development and debilitating policy assumptions already rendered limiting end-of-century radiative forcing to 2.6 W m−2 unachievable in the models5 (Supplementary Text 2). However, in SSP4 and SSP5 limiting radiative forcing to 1.9 W m−2 proved difficult too. In SSP4, a world that promotes both geographical and social inequalities, only one out of three models attempting a 1.9 W m−2 scenario was successful. Weak mitigation is achieved rather easily in SSP45,12. However, the lack of control over land-related emissions in developing countries and lower acceptability of CCS in developed countries in SSP4 make very low emissions pathways unachievable12. Also in SSP5, a world dominated by high economic growth and fossil-fuel development, challenges to mitigation are high13. Finally, under a middle-of-the-road development (SSP2) and under a green-growth paradigm (SSP1) four and six models, respectively, were able to produce a 1.9 W m−2 scenario (Supplementary Table 1).

Mitigation challenges for achieving a 1.9 W m−2 target thus differ strongly across the SSPs, as illustrated in Fig. 6. For example, the amount of CO2 emission that has to be avoided varies by a factor of two between SSP1 and SSP5 worlds in 1.9 W m−2 scenarios (Fig. 6a). The projected use of BECCS varies by a factor 2 to almost 3 between SSP1, and SSP2 and SSP5, respectively (Fig. 6c), and also land-use CO2 mitigation contributions vary massively yet less distinctly (Fig. 6b). Furthermore, the shift away from baseline development implied by the energy system transformation is also markedly smaller in SSP1 than in SSP2 or SSP5 (Fig. 6d–f), and therefore comes with potentially lower overall societal hurdles. Even when overcoming these differences in starting points, the difficulty or facility of achieving deep mitigation remains very diverse across SSPs. In particular, the lower level of final energy demand that can be achieved in SSP1 implies a smaller energy supply system5,35 (Fig. 6g) and thus also a smaller amount of investment needs to decarbonize it (Fig. 6h). Finally, also residual emissions from agriculture and the emission intensity of food production differ strongly between SSPs (Fig. 6i,j) highlighting that challenges have to be overcome in all sectors. Each of these dimensions identifies possibilities for potential policy intervention.

Fig. 6: Variation in mitigation challenges for limiting end-of-century radiative forcing to 1.9 W m−2 across the SSPs.
Fig. 6

aj, Various dimensions of climate change mitigation challenges are shown. A description of the ten indicators shown here is provided in Supplementary Table 7. Ranges show the minimum–maximum range across models per SSP. Symbols show single models. The yellow line indicates the marker implementation for each respective SSP. As not all modelling frameworks provide all necessary indicators, some panels show fewer models. No model was able to produce a 1.9 W m−2 scenario for SSP3. a, Cumulative CO2 emission reduction from baseline in the 2020–2100 period for the 1.9 W m–2 scenarios (GtCO2). b, Cumulative net land-use CO2 in the 2020–2100 period for the 1.9 W m–2 scenarios (GtCO2). c, Average annual CO2 storage from BECCS for the 1.9 W m2 scenarios for the 2020–2100 period (GtCO2 yr–1). d, Upscaling of low-carbon primary energy share for the 1.9 W m–2 scenarios in 2050 relative to baseline. e, Reduction in coal primary energy for the 1.9 W m–2 scenarios in 2050 relative to baseline (EJ yr–1). f, Reduction in carbon intensity of primary energy in 2050 for the 1.9 W m–2 scenarios relative to baseline (tCO2 TJ–1). g, Average final energy demand for the 2020–2100 period for the 1.9 W m–2 scenarios (EJ yr–1). h, Average annual energy system investment for the 2020–2100 period for the 1.9 W m–2 scenarios (trillion 2005 US$). i, Non-CO2 emissions from agriculture for the 1.9 W m–2 scenarios in 2050 (GtCO2e yr–1). j, Emission intensity of food production for the 1.9 W m–2 scenarios in 2050 (gCO2e kcal–1).

Interpretation and feasibility

Wh at can SSPx-1.9 scenarios teach us about the feasibility of limiting warming to 1.5 °C? Typically, feasibility refers to a multi-dimensional concept that considers aspects of geophysics, technology, economics, societal acceptance, institutions and politics, among other disciplines. In this context, integrated scenarios provide insights about the technological and economic assumptions under which a global climate goal could or could not be achieved. However, because models are stylized, imperfect representations of the world, feasible dynamics in a model might be infeasible in the real world, while vice versa infeasibility in a model might not mean that an outcome is infeasible in reality.

For example, modelled energy transition pathways assume broad social acceptance, convergence towards global cooperation, and limited political inertia or institutional barriers—conditions that are different in reality. At the same time, reality can also move faster than assumed in models44. Advanced and pervasive information technologies that dominate our lives today would not have been considered feasible half a century ago, and also recent real-world cost reductions for renewable energy technologies exceeded expectations even of the more optimistic scenarios from 20 years ago.

Earlier studies have highlighted the importance of deriving insights from scenarios that are able to reach the intended target, and scenarios that indicate under which conditions a target cannot be met45. This has led to the development of more sophisticated interpretations of structured scenario ensembles, which suggest that the proportion of successful scenario results can be used as an indicator of infeasibility risk46. In this context, our scenarios can illustrate that multiple technologically salient options are available for limiting warming increase to 1.5 °C, but that the risk of failure increases markedly in the high growth, unequal and/or energy-intensive worlds of SSP3, SSP4 and SSP5. Any interpretation of models that are unable to reach a certain target comes with caveats, because models, including IAMs, are coarse approximations of reality. Real-world feasibility of a particular scenario also depends on factors that are not covered by current IAMs (such as social support) or enabling factors (such as rapid technological developments). These might shift assessments of feasibility in either a more positive or negative direction.

The policy scenarios reported here thus inform certain aspects, but should not be considered as an absolute statement on feasibility32. Policy analysts and advisors still need to translate the insights of this and other related studies39,43,47,48,49,50,51 into a more complete assessment of feasibility, which accounts for the broader context of societal preferences, politics and recent real-world trends.

Going forward

This study aimed to develop a set of stringent integrated community scenarios that can facilitate the assessment of climate impacts, mitigation and adaptation challenges in the context of the Paris Agreement. However, continued research is needed. A stronger involvement of the social sciences that study how societies change and transform can provide valuable complementary insights. To facilitate such further analyses, data presented here are made available to the wider community. Finally, the SSP1-1.9 marker implementation will be included as a very low climate change scenario in CMIP6 ScenarioMIP (Supplementary Text 1), and detailed climate data for these scenarios will become available during the 2018–2020 time period17,18.

Methods

Methodological context

The IPCC AR5 assessed pathways that limited radiative forcing in 2100 to 2.6 W m−2, allowing a higher level during the century32. This level was deemed likely (>66% probability) to limit global-mean temperature rise to below 2 °C relative to pre-industrial levels by 210053. There are various motivations to explore even more stringent scenarios. For example, in several regions and particular subsystems, such as tropical coral reefs, the impacts for a global average temperature rise of 2 °C can already be considerably large54,55. Recent research has also reported discernible differences in climate impacts between a world that is 1.5 °C or 2 °C warmer56, and these future impacts depend on the evolution of both the climate and the socio-economic system. Our scenarios provide a quantification of these dimensions for worlds that are 1.5 °C warmer, and can serve as a starting point for further research by other communities, such as, for example, the adaptation, water or sustainable development communities. The scenarios presented here are an extension of efforts to provide scenarios for the integrated assessment of climate-change-related challenges5,6,57: the SSP scenario matrix framework7. Studies already use these narratives to explore the actions required to limit radiative forcing in 2100 to levels varying from 8.5 W m−2 down to 2.6 W m−2 (refs 5,9,10,11,12,13,16,31,35,47), and their detailed emissions and land-use developments5 serve as inputs for CMIP6 ScenarioMIP17, as well as other MIPs19,20.


Modelling protocol

Participating modelling teams were asked to provide scenarios that comply with specific modelling characteristics and that are derived with the same models, model versions and assumptions as used for the SSPs5 (see below). The modelling protocol consisted of a set of simulations in which total anthropogenic radiative forcing in 2100 is limited to 1.9 W m−2. The limit of 1.9 W m−2 is evaluated with the simple carbon-cycle and climate model MAGICC (Model for the Assessment of Greenhouse Gas Induced Climate Change)58 in a setup comparable to the initial setup used for the RCPs4. The 1.9 W m−2 limit was selected to result in at least 0.3 °C of global mean temperature increase difference with corresponding 2.6 W m−2 scenarios, which would be consistent with at least 50% of the global land surface experiencing statistically significant changes in temperatures59. The 1.9 W m−2 limit is achieved in the IAMs by adjusting the CO2-equivalent carbon price. This means that the radiative forcing target is achieved through reductions in GHG emissions and related co-emissions, but not through intentional increases in aerosol emissions or solar radiation management. Scenarios are run for all SSPs available in each respective modelling framework, and with their corresponding Shared Climate Policy Assumptions21, which influence the regional and sectorial application of CO2-equivalent carbon prices (see appendices in ref. 5). Scenarios are labelled with the forcing target identifier '1.9' in combination with the respective SSP identifier, for example, SSP1–1.9 for a 1.9 W m−2 scenario with SSP1 assumptions. For each SSP, a marker implementation was selected, which represents the characteristics of that SSP particularly well5. If appropriate, insights are drawn from a comparison of marker scenarios only. As was the case with RCP and SSP construction, no account of climate feedbacks to human activities and associated emissions is taken into account in the scenarios reported here.


Model participation

Six modelling frameworks were included in this study: AIM/CGE11, GCAM412, IMAGE9, MESSAGE-GLOBIOM10, REMIND-MAgPIE13 and WITCH-GLOBIOM16. To ensure consistency and comparability, the study was carried out with the same model versions and setup as used for the other SSP–RCP work5. Detailed descriptions of the SSP implementations in all participating frameworks are available as part of a special issue on the quantification of the SSPs9,10,11,12,13,16, with overview papers showing a comparison of results5 as well as a synthesis of key insights related to the energy system35 and land use31. An overview of model documentation, including the native regional resolution of the models and extensive references, is available in appendix D of ref. 5. Supplementary Table 1 provides a succinct overview of the modelling frameworks and key references.

Two modelling frameworks have slight updates to their model setups since earlier publication of the SSP-RCP work5: (i) GCAM: The implementation of near-term policy restrictions as dictated by the Shared Policy Assumptions5,21 has been modified for 'F2' (see ref. 5) by ensuring that a linear carbon–price trajectory is followed between 2020 and 2040. GCAM’s agricultural assumptions in 2020 have been adjusted to better align emissions with observations. In particular, agricultural productivity estimates from 2011 to 2020 have been reduced. (ii) WITCH: A recalibration in the supply cost curves of Storage and Transportation of CO2 has been carried out. On the basis of the regional storage cost curves of ref. 60, availability curves per region have been fitted to provide better cost estimates as the amount of stored CO2 increases markedly, and to ensure the estimated storage potential is in line with more recent publications.

Not all modelling teams attempted to model all SSPs, and many only implemented a subset, either because their model was not appropriate to represent the particularities of a specific SSP or because of time and resource constraints. No SSP3–1.9 scenarios have been reported as reaching a 2.6 W m−2 target was, under these assumptions, already not possible5 (Supplementary Table 1, Supplementary Fig. 1, Supplementary Text 2). Marker implementations are available for the 1.9 W m−2 scenarios for SSP1, SSP2 and SSP5.

The set of modelling frameworks participating in this study represents an ensemble of opportunity. However, it nevertheless represents a wide variety of modelling approaches and model behaviour. Several different model types are represented, including Computable General Equilibrium (CGE) models, partial equilibrium models and hybrid models that combine a systems dynamics or a systems engineering model with a CGE (see Supplementary Table 1). Three frameworks are intertemporal optimization frameworks, and the other three are recursive dynamic frameworks (see table 1 of ref. 5). The set of modelling frameworks spans the whole spectrum of model response classes as identified in ref. 61, that is, from low (WITCH) to high response (for example, REMIND, GCAM and MESSAGE). Considering these various dimensions, the ensemble of opportunities described by modelling frameworks participating in this study spans a wide diversity of the models that are available.

The scenarios presented here do not consider all potential CDR options (for example, they do not include direct air capture, enhanced weathering, biochar, soil organic carbon or ocean fertilization) and exclude solar radiation management. In these scenarios, CDR is thus mainly achieved by BECCS or afforestation.


Emission and temperature assessment

GHG emissions here always refer to the gases of the Kyoto basket (that is, CO2, CH4, N2O, HFCs, PFC and SF6 but excluding the recently added gas NF3)62, aggregated with 100-year Global Warming Potentials from the IPCC Fourth Assessment Report63. Global-mean temperature change is reported relative to the 1850–1900 base period, here referred to as the pre-industrial period. Exceedance probabilities are computed with a probabilistic setup of the MAGICC model64,65 similar to the setup used in the IPCC AR5 Working Group III contribution32. The distribution of equilibrium climate sensitivity assumed in this setup is derived from the climate sensitivity assessment of the IPCC Fourth Assessment Report and hence fully consistent with this Report65. Our setup has similar results when updated to the values of the IPCC’s most recent assessment (see ref. 66). The implied transient climate-response distribution has a median of 1.7 °C with a 5–95% range of 1.2–2.4 °C. The performance of this model setup is compared to the response of complex general circulation models in fig. 6.12 of ref. 32.


Data availability

Scenario data for all SSPx–1.9 scenarios will be made accessible online via the SSP Database portal: https://tntcat.iiasa.ac.at/SspDb/.

Additional information

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  1. 1.

    van Vuuren, D. et al. The representative concentration pathways: an overview. Climatic Change 109, 5–31 (2011).

  2. 2.

    Taylor, K. E., Stouffer, R. J. & Meehl, G. A. An overview of CMIP5 and the experiment Design. Bull. Am. Meteorol. Soc. 93, 485–498 (2011).

  3. 3.

    Warszawski, L. et al. The inter-sectoral impact model intercomparison project (ISI–MIP): project framework. Proc. Natl Acad. Sci. USA 111, 3228–3232 (2014).

  4. 4.

    Meinshausen, M. et al. The RCP greenhouse gas concentrations and their extensions from 1765 to 2300. Climatic Change 109, 213–241 (2011).

  5. 5.

    Riahi, K. et al. The Shared Socioeconomic Pathways and their energy, land use, and greenhouse gas emissions implications: an overview. Glob. Environ. Change 42, 153–168 (2017).

  6. 6.

    O’Neill, B. et al. A new scenario framework for climate change research: the concept of shared socioeconomic pathways. Climatic Change 122, 387–400 (2014).

  7. 7.

    van Vuuren, D. P. et al. A new scenario framework for climate change research: scenario matrix architecture. Climatic Change 122, 373–386 (2014).

  8. 8.

    O’Neill, B. C. et al. The roads ahead: narratives for shared socioeconomic pathways describing world futures in the 21st century. Glob. Environ. Change 42, 169–180 (2017).

  9. 9.

    van Vuuren, D. P. et al. Energy, land-use and greenhouse gas emissions trajectories under a green growth paradigm. Glob. Environ. Change 42, 237–250 (2017).

  10. 10.

    Fricko, O. et al. The marker quantification of the Shared Socioeconomic Pathway 2: a middle-of-the-road scenario for the 21st century. Glob. Environ. Change 42, 251–267 (2017).

  11. 11.

    Fujimori, S. et al. SSP3: AIM implementation of Shared Socioeconomic Pathways. Glob. Environ. Change 42, 268–283 (2017).

  12. 12.

    Calvin, K. et al. The SSP4: a world of deepening inequality. Glob. Environ. Change 42, 284–296 (2017).

  13. 13.

    Kriegler, E. et al. Fossil-fueled development (SSP5): an energy and resource intensive scenario for the 21st century. Glob. Environ. Change 42, 297–315 (2017).

  14. 14.

    Decision 1/CP.16 The Cancun Agreements: Outcome of the Work of the Ad Hoc Working Group on Long-term Cooperative Action Under the Convention (UNFCCC, 2010).

  15. 15.

    Paris Agreement (UNFCCC, 2015).

  16. 16.

    Emmerling, J. et al. The WITCH 2016 model — documentation and implementation of the Shared Socioeconomic Pathways. FEEM Working Paper 42.2016 (2016).

  17. 17.

    O’Neill, B. C. et al. The scenario model intercomparison project (ScenarioMIP) for CMIP6. Geosci. Model Dev. 9, 3461–3482 (2016).

  18. 18.

    Eyring, V. et al. Overview of the coupled model intercomparison project phase 6 (CMIP6) experimental design and organization. Geosci. Model Dev. 9, 1937–1958 (2016).

  19. 19.

    Jones, C. D. et al. C4MIP — the coupled climate–carbon cycle model intercomparison project: experimental protocol for CMIP6. Geosci. Model Dev. 9, 2853–2880 (2016).

  20. 20.

    Lawrence, D. M. et al. The land use model intercomparison project (LUMIP) contribution to CMIP6: rationale and experimental design. Geosci. Model Dev. 9, 2973–2998 (2016).

  21. 21.

    Kriegler, E. et al. A new scenario framework for climate change research: the concept of shared climate policy assumptions. Climatic Change 122, 401–414 (2014).

  22. 22.

    Schleussner, C.-F. et al. Science and policy characteristics of the Paris Agreement temperature goal. Nat. Clim. Change 6, 827–835 (2016).

  23. 23.

    Knutti, R., Rogelj, J., Sedlacek, J. & Fischer, E. M. A scientific critique of the two-degree climate change target. Nat. Geosci. 9, 13–18 (2016).

  24. 24.

    Rogelj, J. et al. Differences between carbon budget estimates unravelled. Nat. Clim. Change 6, 245–252 (2016).

  25. 25.

    IPCC Climate Change 2014: Synthesis Report (eds Core Writing Team, Pachauri, R. K. & Meyer L. A.) (IPCC, 2015).

  26. 26.

    Clarke, L. et al. in Climate Change 2014: Mitigation of Climate Change (eds Edenhofer, O. et al.) 413–510 (IPCC, Cambridge Univ. Press, 2014).

  27. 27.

    Rogelj, J. et al. Energy system transformations for limiting end-of-century warming to below 1.5°C. Nat. Clim. Change 5, 519–527 (2015).

  28. 28.

    MacDougall, A. H., Zickfeld, K., Knutti, R. & Matthews, H. D. Sensitivity of carbon budgets to permafrost carbon feedbacks and non-CO2 forcings. Environ. Res. Lett. 10, 125003 (2015).

  29. 29.

    Schneider von Deimling, T. et al. Estimating the near-surface permafrost–carbon feedback on global warming. Biogeosciences 9, 649–665 (2012).

  30. 30.

    Gernaat, D. E. H. J. et al. Understanding the contribution of non-carbon dioxide gases in deep mitigation scenarios. Glob. Environ. Change 33, 142–153 (2015).

  31. 31.

    Popp, A. et al. in Land-use futures in the shared socio-economic pathways. Glob. Environ. Change 42, 331–345 (2017).

  32. 32.

    Clarke, L. et al. Climate Change 2014: Mitigation of Climate Change (eds Edenhofer, O. et al.) Ch. 6, 413–510 (IPCC, Cambridge Univ. Press, 2014).

  33. 33.

    Popp, A., Lotze-Campen, H. & Bodirsky, B. Food consumption, diet shifts and associated non-CO2 greenhouse gases from agricultural production. Glob. Environ. Change 20, 451–462 (2010).

  34. 34.

    Havlík, P. et al. Climate change mitigation through livestock system transitions. Proc. Natl Acad. Sci. USA 111, 3709–3714 (2014).

  35. 35.

    Bauer, N. et al. Shared Socio-Economic Pathways of the energy sector — quantifying the narratives. Glob. Environ. Change 42, 316–330 (2017).

  36. 36.

    Creutzig, F. et al. Bioenergy and climate change mitigation: an assessment. Glob. Change Biol. Bioenergy 7, 916–944 (2015).

  37. 37.

    Bonsch, M. et al. Trade-offs between land and water requirements for large-scale bioenergy production. Glob. Change Biol. Bioenergy 8, 11–24 (2016).

  38. 38.

    Smith, P. et al. in Climate Change 2014: Mitigation of Climate Change (eds Edenhofer, O. et al.) Ch. 11, 811–922 (IPCC, Cambridge Univ. Press, 2014).

  39. 39.

    Smith, P. et al. Biophysical and economic limits to negative CO2 emissions. Nat. Clim. Change 6, 42–50 (2016).

  40. 40.

    Field, C. B. & Mach, K. J. Rightsizing carbon dioxide removal. Science 356, 706–707 (2017).

  41. 41.

    Smith, P. et al. How much land-based greenhouse gas mitigation can be achieved without compromising food security and environmental goals? Glob. Change Biol. 19, 2285–2302 (2013).

  42. 42.

    Valin, H. et al. Agricultural productivity and greenhouse gas emissions: trade-offs or synergies between mitigation and food security? Environ. Res. Lett. 8, 035019 (2013).

  43. 43.

    Rogelj, J. et al. Energy system transformations for limiting end-of-century warming to below 1.5 °C. Nat. Clim. Change 5, 519–527 (2015).

  44. 44.

    Creutzig, F. et al. The underestimated potential of solar energy to mitigate climate change. Nat. Energy 2, 17140 (2017).

  45. 45.

    Tavoni, M. & Tol, R. Counting only the hits? The risk of underestimating the costs of stringent climate policy. Climatic Change 100, 769–778 (2010).

  46. 46.

    Riahi, K. et al. Locked into Copenhagen pledges — implications of short-term emission targets for the cost and feasibility of long-term climate goals. Technol. Forecast. Soc. Change 90, 8–23 (2015).

  47. 47.

    Sanderson, B. M., O’Neill, B. C. & Tebaldi, C. What would it take to achieve the Paris temperature targets?. Geophys. Res. Lett. 43, 7133–7142 (2016).

  48. 48.

    Azar, C., Johansson, D. J. A. & Mattsson, N. Meeting global temperature targets—the role of bioenergy with carbon capture and storage. Environ. Res. Lett. 8, 034004 (2013).

  49. 49.

    Su, X. et al. Emission pathways to achieve 2.0 °C and 1.5 °C climate targets. Earths Future 5, 592–604 (2017).

  50. 50.

    Walsh, B. et al. Pathways for balancing CO2 emissions and sinks. Nat. Commun. 8, 14856 (2017).

  51. 51.

    Scott, V., Gilfillan, S., Markusson, N., Chalmers, H. & Haszeldine, R. S. Last chance for carbon capture and storage. Nat. Clim. Change 3, 105–111 2013).

  52. 52.

    Le Quéré, C. et al. Global Carbon Budget 2015. Earth Syst. Sci. Data 7, 349–396 (2015).

  53. 53.

    IPCC Climate Change 2014: Impacts, Adaptation, and Vulnerability (eds Field, C. B. et al.) 1–32 (IPCC, Cambridge Univ. Press, 2014).

  54. 54.

    Frieler, K. et al. Limiting global warming to 2 °C is unlikely to save most coral reefs. Nat. Clim. Change 3, 165–170 (2013).

  55. 55.

    Schleussner, C. F. et al. Differential climate impacts for policy-relevant limits to global warming: the case of 1.5 °C and 2 °C. Earth Syst. Dynam. 7, 327–351 (2016).

  56. 56.

    Moss, R. H. et al. The next generation of scenarios for climate change research and assessment. Nature 463, 747–756 (2010).

  57. 57.

    Meinshausen, M., Raper, S. C. B. & Wigley, T. M. L. Emulating coupled atmosphere–ocean and carbon cycle models with a simpler model, MAGICC6 — part 1: model description and calibration. Atmos. Chem. Phys. 11, 1417–1456 (2011).

  58. 58.

    Claudia, T., Brian, O. N. & Jean-François, L. Sensitivity of regional climate to global temperature and forcing. Environ. Res. Lett. 10, 074001 (2015).

  59. 59.

    Hendriks C., Graus W. & Van Bergen F. Global Carbon Dioxide Storage Potential and Costs Report No. EEP-02001 (Ecofys, 2004).

  60. 60.

    Kriegler, E. et al. Diagnostic indicators for integrated assessment models of climate policy. Technol. Forecast. Social. Change 90, 45–61 (2015).

  61. 61.

    Decision 24/CP.19. Revision of the UNFCCC Reporting Guidelines on Annual Inventories for Parties included in Annex I to the Convention 1–54 (UNFCCC, 2013).

  62. 62.

    IPCC Climate Change 2007: The Physical Science Basis (eds Solomon, S. et al.) (Cambridge Univ. Press, 2007).

  63. 63.

    Meinshausen, M. et al. Greenhouse-gas emission targets for limiting global warming to 2 °C. Nature 458, 1158–1162 (2009).

  64. 64.

    Rogelj, J., Meinshausen, M. & Knutti, R. Global warming under old and new scenarios using IPCC climate sensitivity range estimates. Nat. Clim. Change 2, 248–253 (2012).

  65. 65.

    Rogelj, J., Meinshausen, M., Sedláček, J. & Knutti, R. Implications of potentially lower climate sensitivity on climate projections and policy. Environ. Res. Lett. 9, 031003 (2014).

  66. 66.

    IPCC Climate Change 2014: Mitigation of Climate Change (eds Edenhofer, O. et al.) 1–33 (Cambridge Univ. Press, 2014).

Download references

Acknowledgements

We thank the International Institute for Applied Systems Analysis (IIASA) for hosting and maintaining the SSP Scenario Database of the Integrated Assessment Modelling Consortium (IAMC), and thank P. Kolp for his reliable support with the administration of and access to scenario data, and administration of the database infrastructure. J.R., O.F., V.K., K.R., G.L., E.K. and A.P. have received funding from the European Union’s Horizon 2020 Research and Innovation Programme under grant agreement no. 642147 (CD-LINKS), no. 641816 (CRESCENDO) and the Framework Programme 7 under grant agreement no. 308329 (ADVANCE). J.S. has received funding from the Deutsche Forschungsgemeinschaft (DFG) in the SPP ED 178/3-1 (CEMICS). S.F. and T.H. are supported by JSPS KAKENHI Grant Number JP16K18177, and the Global Environmental Research Fund 2–1702 of the Ministry of Environment of Japan. J.R. acknowledges the support of the Oxford Martin Visiting Fellowship Programme.

Author information

Affiliations

  1. International Institute for Applied Systems Analysis (IIASA), Laxenburg, Austria

    • Joeri Rogelj
    • , Shinichiro Fujimori
    • , Tomoko Hasegawa
    • , Volker Krey
    • , Keywan Riahi
    • , Oliver Fricko
    •  & Petr Havlík
  2. Institute for Atmospheric and Climate Science, ETH Zurich, Zurich, Switzerland

    • Joeri Rogelj
  3. Potsdam Institute for Climate Impact Research (PIK), Member of the Leibniz Association, Potsdam, Germany

    • Alexander Popp
    • , Gunnar Luderer
    • , Jessica Strefler
    • , Elmar Kriegler
    •  & Florian Humpenöder
  4. Joint Global Change Research Institute, Pacific Northwest National Laboratory, College Park, MD, USA

    • Katherine V. Calvin
    •  & Jae Edmonds
  5. Fondazione Eni Enrico Mattei, Milan, Italy

    • Johannes Emmerling
    • , Giacomo Marangoni
    • , Laurent Drouet
    •  & Massimo Tavoni
  6. Centro Euro-Mediterraneo sui Cambiamenti Climatici, Milan, Italy

    • Johannes Emmerling
    • , Giacomo Marangoni
    • , Laurent Drouet
    •  & Massimo Tavoni
  7. PBL Netherlands Environmental Assessment Agency, The Hague, The Netherlands

    • David Gernaat
    • , Detlef P. van Vuuren
    • , Jonathan Doelman
    • , Mathijs Harmsen
    •  & Elke Stehfest
  8. Copernicus Institute for Sustainable Development, Utrecht University, Utrecht, The Netherlands

    • David Gernaat
    • , Detlef P. van Vuuren
    •  & Mathijs Harmsen
  9. National Institute for Environmental Studies, Tsukuba, Japan

    • Shinichiro Fujimori
    •  & Tomoko Hasegawa
  10. Department of Management, Economics and Industrial Engineering, Politecnico di Milano, Milan, Italy

    • Massimo Tavoni

Authors

  1. Search for Joeri Rogelj in:

  2. Search for Alexander Popp in:

  3. Search for Katherine V. Calvin in:

  4. Search for Gunnar Luderer in:

  5. Search for Johannes Emmerling in:

  6. Search for David Gernaat in:

  7. Search for Shinichiro Fujimori in:

  8. Search for Jessica Strefler in:

  9. Search for Tomoko Hasegawa in:

  10. Search for Giacomo Marangoni in:

  11. Search for Volker Krey in:

  12. Search for Elmar Kriegler in:

  13. Search for Keywan Riahi in:

  14. Search for Detlef P. van Vuuren in:

  15. Search for Jonathan Doelman in:

  16. Search for Laurent Drouet in:

  17. Search for Jae Edmonds in:

  18. Search for Oliver Fricko in:

  19. Search for Mathijs Harmsen in:

  20. Search for Petr Havlík in:

  21. Search for Florian Humpenöder in:

  22. Search for Elke Stehfest in:

  23. Search for Massimo Tavoni in:

Contributions

J.R. coordinated the conception and writing of the paper, performed the scenario analysis and created the figures; J.R., K.V.C., A.P., G.L., J.Em., S.F., E.K., K.R. and D.P.v.V. designed the scenarios, which were developed and contributed by all modelling teams, with notable contributions from S.F., T.H. (AIM/CGE), K.V.C., J.Ed. (GCAM), D.G., E.S., J.D., M.H., D.P.v.V. (IMAGE), O.F., P.H., V.K., J.R., K.R. (MESSAGE-GLOBIOM), J.S., F.H., A.P., G.L., E.K. (REMIND-MAgPIE) and J.Em., G.M., L.D. and M.T. (WITCH-GLOBIOM); all authors provided feedback and contributed to writing the paper.

Competing interests

The authors declare no competing interests.

Corresponding author

Correspondence to Joeri Rogelj.

Supplementary information

  1. Supplementary Information

    Supplementary Text 1–6, Supplementary Figures 1–26, Supplementary Tables 1–7 and Supplementary References

About this article

Publication history

Received

Accepted

Published

DOI

https://doi.org/10.1038/s41558-018-0091-3

Rights and permissions

To obtain permission to re-use content from this article visit RightsLink.