We must move faster towards a low-carbon world if we are to limit global warming to 2 C this century, experts warn.
Changes in electricity, heat, buildings, industry and transport are needed rapidly and must happen all together, according to researchers at the universities of Sussex, Manchester and Oxford in a new study published in the journal Science.
To provide a reasonable (66%) chance of limiting global temperature increases to below 2 C, the International Energy Agency and International Renewable Energy Agency suggest that global energy-related carbon emissions must peak by 2020 and fall by more than 70% in the next 35 years. This implies:
- Generating 95% of electricity from low-carbon sources by 2050
- Shifting almost entirely towards electric cars.
- Tripling the annual rate of energy efficiency improvement
- Retrofitting the entire building stock
This elemental challenge necessitates “deep decarbonisation” of electricity, transport, heat, industrial, forestry and agricultural systems across the world. But despite the recent rapid growth in renewable electricity generation, the rate of progress towards this wider goal remains slow.
Professor Frank Geels from the University of Manchester, lead author of the study, explains: “Our ‘big picture’ socio-technical framework shows how interactions between various social groups can increase the momentum of low-carbon transitions.”
Professor Nick Eyre from the University of Oxford, End Use Energy Demand Champion for the UK Research Councils’ Energy Programme, adds: “Accelerating transitions is critical if we are to achieve the goals of decarbonizing and saving energy faster, further, and more flexibly. This international quality study shows the importance of whole systems thinking in energy demand research.”
Professor Benjamin K. Sovacool from the University of Sussex, a co-author on the study, says: “Current rates of change are simply not enough. We need to accelerate transitions, deepen their speed and broaden their reach. Otherwise there can be no hope of reaching a 2 degree target, let alone 1.5 degrees. This piece reveals that the acceleration of transitions across the sociotechnical systems of electricity, heat, buildings, manufacturing, and transport requires new conceptual approaches, analytical foci, and research methods.”
The Policy Forum provides four key lessons for how to accelerate sustainability transitions.
Lesson 1: Focus on socio-technical systems rather than individual elements
Rapid and deep decarbonization requires a transformation of ‘sociotechnical systems’ — the interlinked mix of technologies, infrastructures, organizations, markets, regulations and user practices that together deliver societal functions such as personal mobility. Previous systems have developed over many decades, and the alignment and co-evolution of their elements makes them resistant to change.
Accelerated low-carbon transitions therefore depend on both techno-economic improvements, and social, political and cultural processes, including the development of positive or negative discourses. Professor Steve Sorrell from the University of Sussex, a coauthor of the study, states: “In this policy forum we describe how transformational changes in energy and transport systems occur, and how they may be accelerated. Traditional policy approaches emphasizing a single technology will not be enough.”
Lesson 2: Align multiple innovations and systems
Socio-technical transitions gain momentum when multiple innovations are linked together, improving the functionality of each and acting in combination to reconfigure systems. The shale gas revolution, for instance, accelerated when seismic imaging, horizontal drilling, and hydraulic fracturing were combined. Likewise, accelerated low-carbon transitions in electricity depend not only on the momentum of renewable energy innovations like wind, solar-PV and bio-energy, but also on complementary innovations including energy storage and demand response. These need aligned and then linked so that innovations are harmonized.
Prof. Paul Ekins, Director of the UCL Institute for Sustainable Resources, University College London, and project leader to the EU INNOPATHS consortium researching low-carbon transitions for Europe, comments: “One of the great strengths of this study is the equal emphasis it accords to technological, social, business and policy innovation, in all of which governments as well as the private sector have a key role to play.
“European countries will become low-carbon societies not only when the required low-carbon technologies have been developed but when new business models and more sustainable consumer aspirations are driving their deployment at scale. Public policy has an enormous role to play at every step in the creation of these changed conditions.”
Lesson 3: Offer societal and business support
Public support is crucial for effective transition policies. Low-carbon transitions in mobility, agro-food, heat and buildings will also involve millions of citizens who need to modify their purchase decisions, user practices, beliefs, cultural conventions and skills. To motivate citizens, financial incentives and information about climate change threats need to be complemented by positive discourses about the economic, social and cultural benefits of low-carbon innovations.
Furthermore, business support is essential because the development and deployment of low-carbon innovations depends upon the technical skills, organizational capabilities and financial resources of the private sector. Green industries and supply chains can solidify political coalitions supporting ambitious climate policies and provide a counterweight to incumbents. Technological progress can drive climate policy by providing solutions or altering economic interests. Shale gas and solar-PV developments, for instance, altered the US and Chinese positions in the international climate negotiations.
Lesson 4: Phase out existing systems
Socio-technical transitions can be accelerated by actively phasing out existing technologies, supply chains, and systems that lock-in emissions for decades. Professor Sovacool comments that: “All too often, analysists and even policymakers focus on new incentives, on the phasing in of low-carbon technologies. This study reminds us that phasing out existing systems can be just as important as stimulating novel innovations.”
For instance, the UK transition to smokeless solid fuels and gas was accelerated by the 1956 Clean Air Act, which allowed cities to create smokeless zones where coal use was banned. Another example is the 2009 European Commission decision to phase-out incandescent light bulbs, which accelerated the shift to compact fluorescents and LEDs. French and UK governments have announced plans to phase-out petrol and diesel cars by 2040. Moreover, the UK intends to phase out unabated coal-fired power generation by 2025.
Phasing out existing systems accelerates transitions by creating space for niche-innovations and removing barriers to their diffusion. The phase-out of carbon-intensive systems is also essential to prevent the bulk of fossil fuel reserves from being burned, which would obliterate the 2C target. This phase-out will be challenging since it threatens the largest and most powerful global industries (e.g. oil, automobiles, electric utilities, agro-food, steel), which will fight to protect their vested economic and political interests.
Deep decarbonization requires complementing model-based analysis with socio-technical research. While the former analyzes technically feasible least-cost pathways, the latter addresses innovation processes, business strategies, social acceptance, cultural discourses and political struggles, which are difficult to model but crucial in real-world transitions. As Professor Geels notes, an enduring lesson is that “to accelerate low-carbon transitions, policymakers should not only stimulate techno-economic developments, but also build political coalitions, enhance business involvement, and engage civil society.”
Additionally, the research underscores the non-technical, or social, elements of transitions. Dr. Tim Schwanen from the University of Oxford, a coauthor, states that “the approach described in this Policy Forum demonstrates the importance of heeding insights from across the social sciences in thinking about low-carbon transitions.”
While full integration of both approaches is not possible, productive bridging strategies may enable policy strategies that are both cost-effective and socio-politically feasible.
Frank W. Geels, Benjamin K. Sovacool, Tim Schwanen, Steve Sorrell. Sociotechnical transitions for deep decarbonization. Science, 2017; 357 (6357): 1242 DOI: 10.1126/science.aao3760
University of Sussex. “We must accelerate transitions for sustainability and climate change, experts say: Rapid changes in electricity, heat, buildings, industry and mobility are needed.” ScienceDaily. ScienceDaily, 21 September 2017. <www.sciencedaily.com/releases/2017/09/170921141235.htm
New climate risk classification created to account for potential ‘existential’ threats: Researchers identify a one-in-20 chance of temperature increase causing catastrophic damage or worse by 2050
The risk assessment stems from the objective stated in the 2015 Paris Agreement regarding climate change that society keep average global temperatures “well below” a 2°C (3.6°F) increase from what they were before the Industrial Revolution. Even if that objective is met, a global temperature increase of 1.5°C (2.7°F) is still categorized as “dangerous,” meaning it could create substantial damage to human and natural systems.
A temperature increase greater than 3°C (5.4°F) could lead to what the researchers term “catastrophic” effects, and an increase greater than 5°C (9°F) could lead to “unknown” consequences which they describe as beyond catastrophic including potentially existential threats. The specter of existential threats is raised to reflect the grave risks to human health and species extinction from warming beyond 5° C, which has not been experienced for at least the past 20 million years.
The scientists term warming probability of five percent or less as a “low-probability high-impact” scenario and assess such scenarios in the analysis “Well Below 2°C: Mitigation strategies for avoiding dangerous to catastrophic climate changes,” which will appear in the journal Proceedings of the National Academy of Sciences on Sept. 14.
Ramanathan and Xu also describe three strategies for preventing the gravest threats from taking place.
“When we say five percent-probability high-impact event, people may dismiss it as small but it is equivalent to a one-in-20 chance the plane you are about to board will crash,” said Ramanathan. “We would never get on that plane with a one-in-20 chance of it coming down but we are willing to send our children and grandchildren on that plane.”
The researchers defined the risk categories based on guidelines established by the Intergovernmental Panel on Climate Change (IPCC) and previous independent studies. “Dangerous” global warming includes consequences such as increased risk of extreme weather and climate events ranging from more intense heat waves, hurricanes, and floods, to prolonged droughts. Planetary warming between 3°C and 5°C could trigger what scientists term “tipping points” such as the collapse of the West Antarctic Ice Sheet and subsequent global sea-level rise, and the dieback of the Amazon rainforest. In human systems, catastrophic climate change is marked by deadly heat waves becoming commonplace, exposing over 7 billion people to heat related mortalities and famine becoming widespread. Furthermore, the changes will be too rapid for most to adapt to, particularly the less affluent, said Ramanathan.
Risk assessments of global temperature rise greater than 5°C have not been undertaken by the IPCC. Ramanathan and Xu named this category “unknown??” with the question marks acknowledging the “subjective nature of our deduction.” The existential threats could include species extinctions and major threats to human water and food supplies in addition to the health risks posed by exposing over 7 billion people worldwide to deadly heat.
With these scenarios in mind, the researchers identified what measures can be taken to slow the rate of global warming to avoid the worst consequences, particularly the low-probability high-impact events. Aggressive measures to curtail the use of fossil fuels and emissions of so-called short-lived climate pollutants such as soot, methane and HFCs would need to be accompanied by active efforts to extract CO2 from the air and sequester it before it can be emitted. It would take all three efforts to meet the Paris Agreement goal to which countries agreed at a landmark United Nations climate conference in Nov 2015.
Xu and Ramanathan point out that the goal is attainable. Global CO2 emissions had grown at a rate of 2.9 percent per year between 2000 and 2011, but had slowed to a near-zero growth rate by 2015. They credited drops in CO2 emissions from the United States and China as the primary drivers of the trend. Increases in production of renewable energy, especially wind and solar power, have also bent the curve of emissions trends downward. Other studies have estimated that there was by 2015 enough renewable energy capacity to meet nearly 24 percent of global electricity demand.
Short-lived climate pollutants are so called because even though they warm the planet more efficiently than carbon dioxide, they only remain in the atmosphere for a period of weeks to roughly a decade whereas carbon dioxide molecules remain in the atmosphere for a century or more. The authors also note that most of the technologies needed to drastically curb emissions of short-lived climate pollutants already exist and are in use in much of the developed world. They range from cleaner diesel engines to methane-capture infrastructure.
“While these are encouraging signs, aggressive policies will still be required to achieve carbon neutrality and climate stability,” the authors wrote.
The release of the study coincides with the start of Climate Week NYC in New York, a summit of business and government leaders to highlight global climate action. Ramanathan and colleagues will deliver a complementary report detailing the “three-lever” mitigation strategy of emissions control and carbon sequestration on Sept. 18 at the United Nations. That report was produced by the Committee to Prevent Extreme Climate Change, chaired by Ramanathan, Nobel Prize winner Mario Molina of UC San Diego, and Durwood Zaelke, who leads an advocacy organization, the Institute for Governance and Sustainable Development, with 30 experts from around the world including China and India.
Yangyang Xu, Veerabhadran Ramanathan. Well below 2 °C: Mitigation strategies for avoiding dangerous to catastrophic climate changes. Proceedings of the National Academy of Sciences, 2017; 201618481 DOI: 10.1073/pnas.1618481114 firstname.lastname@example.org or email@example.com.
The historic Paris Agreement calls for limiting global temperature rise to “well below 2 °C.” Because of uncertainties in emission scenarios, climate, and carbon cycle feedback, we interpret the Paris Agreement in terms of three climate risk categories and bring in considerations of low-probability (5%) high-impact (LPHI) warming in addition to the central (∼50% probability) value. The current risk category of dangerous warming is extended to more categories, which are defined by us here as follows: >1.5 °C as dangerous; >3 °C as catastrophic; and >5 °C as unknown, implying beyond catastrophic, including existential threats. With unchecked emissions, the central warming can reach the dangerous level within three decades, with the LPHI warming becoming catastrophic by 2050. We outline a three-lever strategy to limit the central warming below the dangerous level and the LPHI below the catastrophic level, both in the near term (<2050) and in the long term (2100): the carbon neutral (CN) lever to achieve zero net emissions of CO2, the super pollutant (SP) lever to mitigate short-lived climate pollutants, and the carbon extraction and sequestration (CES) lever to thin the atmospheric CO2 blanket. Pulling on both CN and SP levers and bending the emissions curve by 2020 can keep the central warming below dangerous levels. To limit the LPHI warming below dangerous levels, the CES lever must be pulled as well to extract as much as 1 trillion tons of CO2 before 2100 to both limit the preindustrial to 2100 cumulative net CO2 emissions to 2.2 trillion tons and bend the warming curve to a cooling trend.
The Paris Agreement and its intended nationally determined contributions (INDCs) to reduce emissions (1) are unprecedented first steps for stabilizing global average warming to well below 2 °C (WB2C). It is generally acknowledged that the INDCs must be strengthened significantly to bend the climate emissions curve sufficiently and soon enough to limit the warming to WB2C (1⇓–3). The overall objectives of this perspective piece are threefold:
i) Assess the low-probability (5%) high-impact (LPHI) warming outcomes in the absence of a climate mitigation policy after accounting for major uncertainties in: (a) future emission trajectories; (b) physical climate feedback involving water vapor, clouds, and snow/ice albedo; (c) carbon cycle feedback involving biogeochemistry; and (d) aerosol radiative forcing. We ensure that the extreme outcomes projected in this study are consistent with published model parameters. The warming estimates in this study account for the well-known greenhouse gases (GHGs) and various aerosols (Box 1).
ii) Identify the constraints imposed by WB2C and the criteria for meeting WB2C, and thus sharpen the definition of WB2C.
iii) Explore the mitigation pathways that are still available to meet the WB2C goal.
Box 1. The Non-CO2 Climate Pollutants
The first category is SLCPs, which include GHGs such as methane, tropospheric ozone, HFCs, and aerosols such as BC, and coemitted OC. The lifetimes of these pollutants range from days (BC and OC), to months (tropospheric ozone), to about a decade (methane and HFCs), which explains the term “short-lived” in SLCPs. The second category of non-CO2 climate pollutants includes LLGHGs such as N2O and halocarbons other than HFCs (e.g., CFCs, HCFCs). Our model is forced by IPCC historical forcing of all non-CO2 gases and aerosols. The third category of non-CO2 climate pollutants is cooling aerosols (other than BC and OC) such as sulfates, nitrates, and dust. It should be noted that those cooling aerosols, along with the BC and OC aerosols included under SLCPs, are the major source of air pollution, leading to about 7 million deaths annually (31).
Box 1 Figure shows the individual contribution of CO2, SLCPs, and cooling aerosols (other than those included in SLCPs) to the transient warming during the 20th and 21st centuries. All of the warming trends are relative to preindustrial temperatures. By 2015, the warming due to CO2 is about 0.8 °C and that due to SLCPs is about 1.1 °C. The sum of the CO2 and SLCP warming is already close to the Paris Agreement limit of 2 °C by 2015. On the other hand, the aerosols have a cooling (“masking”) effect of about −0.9 °C. When we add the sum of the CO2, SLCPs, and aerosol effects to the warming due to non-CO2 LLGHGs, the estimated warming by 2015 is ∼1.1 °C (black curve in Box 1 Figure), which can be compared with the observed warming of about 1 °C (SI Appendix, Fig. S1). The main inference from Box 1 Figure is that CO2 and SLCPs have exerted comparable warming effects (0.8 °C and 1.1 °C) to the past, while the aerosol masking effect is also comparable in magnitude but of opposite sign, with a cooling of −0.9 °C.
Box 1 Figure. Simulated transient warming (°C) following the baseline-fast scenario, as a function of the cumulative emission of CO2 (x axis; black line). The decades at which each additional trillion tons of CO2was emitted and the corresponding CO2 concentration are shown at the top. The red, blue, and green lines illustrate transient simulated warming due to CO2, cooling aerosols, and SLCPs only, respectively.
This perspective article weaves in science perspectives with societal perspectives since the two are inextricably linked. For example, the mitigation pathways we choose are largely motivated by the magnitude and rapidity of societal as well as ecosystem impacts (4) (Box 2). We recognize that the metrics for fully comprehending the societal impacts need to extend beyond global average warming (5), but global warming is still a valuable and accepted metric for strategizing mitigation options (6).
Box 2. Risk Categorization of Climate Change to Society
The United Nations Framework Convention on Climate Change coined the phrase “dangerous anthropogenic interference” (DAI) with the climate system. The DAI phrase spurred quite a bit of research on what climate change means for society and the ecosystem (45). Subsequently, in 2001, the IPCC (46) came up with the burning embers diagram, in which it categorized climate risks under five reasons for concern (RFCs) that ranged from risks to natural systems, risks of extreme weather events, distribution of impacts between regions of the world, aggregate impacts, and risks of large-scale discontinuities. In the burning embers diagram, risks under each RFC were ranked based on the warming magnitude. For what follows, we adopt the most recent version of DAI analysis (47). At 2 °C, risks for two RFCs were designated as high, while at 4 °C, all RFCs were ranked as a high-risk category, with two of them ranked as very high. The burning embers diagram does not extend beyond 5 °C.
We are proposing the following extension to the DAI risk categorization: warming greater than 1.5 °C as “dangerous”; warming greater than 3 °C as “catastrophic?”; and warming in excess of 5 °C as “unknown??,” with the understanding that changes of this magnitude, not experienced in the last 20+ million years, pose existential threats to a majority of the population. The question mark denotes the subjective nature of our deduction and the fact that catastrophe can strike at even lower warming levels. The justifications for the proposed extension to risk categorization are given below.
From the IPCC burning embers diagram and from the language of the Paris Agreement, we infer that the DAI begins at warming greater than 1.5 °C. Our criteria for extending the risk category beyond DAI include the potential risks of climate change to the physical climate system, the ecosystem, human health, and species extinction. Let us first consider the category of catastrophic (3 to 5 °C warming). The first major concern is the issue of tipping points. Several studies (48, 49) have concluded that 3 to 5 °C global warming is likely to be the threshold for tipping points such as the collapse of the western Antarctic ice sheet, shutdown of deep water circulation in the North Atlantic, dieback of Amazon rainforests as well as boreal forests, and collapse of the West African monsoon, among others. While natural scientists refer to these as abrupt and irreversible climate changes, economists refer to them as catastrophic events (49).
Warming of such magnitudes also has catastrophic human health effects. Many recent studies (50, 51) have focused on the direct influence of extreme events such as heat waves on public health by evaluating exposure to heat stress and hyperthermia. It has been estimated that the likelihood of extreme events (defined as 3-sigma events), including heat waves, has increased 10-fold in the recent decades (52). Human beings are extremely sensitive to heat stress. For example, the 2013 European heat wave led to about 70,000 premature mortalities (53). The major finding of a recent study (51) is that, currently, about 13.6% of land area with a population of 30.6% is exposed to deadly heat. The authors of that study defined deadly heat as exceeding a threshold of temperature as well as humidity. The thresholds were determined from numerous heat wave events and data for mortalities attributed to heat waves. According to this study, a 2 °C warming would double the land area subject to deadly heat and expose 48% of the population. A 4 °C warming by 2100 would subject 47% of the land area and almost 74% of the world population to deadly heat, which could pose existential risks to humans and mammals alike unless massive adaptation measures are implemented, such as providing air conditioning to the entire population or a massive relocation of most of the population to safer climates.
Climate risks can vary markedly depending on the socioeconomic status and culture of the population, and so we must take up the question of “dangerous to whom?” (54). Our discussion in this study is focused more on people and not on the ecosystem, and even with this limited scope, there are multitudes of categories of people. We will focus on the poorest 3 billion people living mostly in tropical rural areas, who are still relying on 18th-century technologies for meeting basic needs such as cooking and heating. Their contribution to CO2 pollution is roughly 5% compared with the 50% contribution by the wealthiest 1 billion (55). This bottom 3 billion population comprises mostly subsistent farmers, whose livelihood will be severely impacted, if not destroyed, with a one- to five-year megadrought, heat waves, or heavy floods; for those among the bottom 3 billion of the world’s population who are living in coastal areas, a 1- to 2-m rise in sea level (likely with a warming in excess of 3 °C) poses existential threat if they do not relocate or migrate. It has been estimated that several hundred million people would be subject to famine with warming in excess of 4 °C (54). However, there has essentially been no discussion on warming beyond 5 °C.
Climate change-induced species extinction is one major concern with warming of such large magnitudes (>5 °C). The current rate of loss of species is ∼1,000-fold the historical rate, due largely to habitat destruction. At this rate, about 25% of species are in danger of extinction in the coming decades (56). Global warming of 6 °C or more (accompanied by increase in ocean acidity due to increased CO2) can act as a major force multiplier and expose as much as 90% of species to the dangers of extinction (57).
The bodily harms combined with climate change-forced species destruction, biodiversity loss, and threats to water and food security, as summarized recently (58), motivated us to categorize warming beyond 5 °C as unknown??, implying the possibility of existential threats. Fig. 2 displays these three risk categorizations (vertical dashed lines).
Projected Warming in the Absence of Climate Policies
A convenient place to start the discussion is the projected warming in the absence of climate policies. Determining this baseline warming sets the stage for exploring and justifying mitigation pathways. Published future CO2 emission scenarios (1, 4), along with historical emissions, are fed into a carbon cycle model to estimate the CO2 concentration during the 20th and 21st centuries (SI Appendix, Figs. S1 and S2). The calculated CO2 concentration is used to estimate its climate forcing (SI Appendix, Figs. S4 and S6). For climate forcing due to other atmospheric compositions, we adopted the Intergovernmental Panel on Climate Change (IPCC)-derived historical values (Box 1) and future projections (SI Appendix, Fig. S6). The temperature response is estimated with an energy balance climate model (SI Appendix, section 1). In a series of published studies (7⇓⇓–10), the carbon cycle model and the climate model simulations have been extensively validated by comparison with observations of atmospheric CO2 (7) (SI Appendix, Figs. S1 and S2), global mean temperature (8), ocean heat content (box 1 of ref. 7), and sea-level rise (figure 2 of ref. 9), as well as by comparison with published projections from 3D global climate model simulations (9, 10).
The projected warming (relative to 1900) at 2050 and 2100 is shown in Fig. 1, under the two baseline emission scenarios in the absence of climate policy during the 21st century. The temporal evolution of CO2emissions, CO2 concentrations, and global average temperatures are shown in SI Appendix, Fig. S1. To generate the probability distribution curves shown in Fig. 1 (as well as in SI Appendix, Figs. S3, S9, and S10), temperature under each scenario is simulated with 1,500 stochastic runs to cover the full range of climate sensitivity involving feedback related to water vapor, clouds, and snow/ice albedo (SI Appendix, section 1), hereafter referred to as climate physical-dynamical feedback. The projected warming is shown for two baseline emission scenarios proposed by the IPCC (4): baseline-fast and baseline-default. The baseline-fast scenario assumes an aggressive 80% reduction in the energy intensity of the economy (still using fossil fuels) compared with the 2010 energy intensity. The baseline-default scenario adopts the current rate of reduction in energy intensity until 2100, achieving a 50% reduction from the 2010 level. The two baseline emission trajectories, along with the corresponding 5–95% range within each scenario (shading in SI Appendix, Fig. S1), capture expert projections for a plausible range of future emissions in the absence of climate policies.
The warming probability distribution shown in Fig. 1 (and elsewhere in this paper) is due to the wide range of uncertainties in modeling the climate feedbacks (14). The upper range of warming projection, with a probability of less than 5% (Figs. 1 and 2), may appear unrealistically large, but this may not be the case. Here, we choose to use a high range of climate sensitivity because some studies have suggested that 3D climate models have underestimated three major positive climate feedbacks: positive ice albedo feedback from the retreat of Arctic sea ice (15), positive cloud albedo feedback from retreating storm track clouds in midlatitudes (16, 17), and positive albedo feedback by the mixed-phase (water and ice) clouds (18) (more discussion is provided in SI Appendix, section 5). The potential underestimation of these feedbacks, along with the positive carbon cycle feedback to be described below, persuaded us to show the warming distribution (Figs. 1 and 2 and SI Appendix) for low probabilities much less than 5%. Again, we caution that we do not use the projected warming with a probability less than 5% for rest of the mitigation analyses.
Thus far, the thick curves in Fig. 1 and SI Appendix, Fig. S1 capture the uncertainties in the emissions scenarios and in the model treatment of climate physical-dynamical feedback. There are two other major sources of uncertainties:
i) Aerosol radiative forcing uncertainties and their entanglement with climate sensitivity estimate (19). These uncertainties, when included in the probability distribution shown in Fig. 1B (blue dashed line), would slightly change the skewed distribution of the projected warming.
ii) Biogeochemical feedback between climate change and the carbon cycle. Different climate-carbon feedbacks, all of which amplify the warming, are considered below. The first feedback deals with the decrease in the oceanic and land uptake of carbon with warming, the second is the release of soil carbon due to the thawing of the permafrost, and the third is the increase in carbon emission (as CO2 and methane) from wetlands (20). Recent studies using 3D climate models coupled with a biogeochemistry component have systemically examined the carbon cycle response to future warming (21), which revealed the following: Modeling uncertainties introduce a −16% to 13% uncertainty range in the cumulative emission (as of 2100), climate-carbon feedback on the carbon uptake by land and oceans introduces a 6–27% increase in the cumulative emissions from 2010 to 2100, and the permafrost thawing can release soil carbon and increase cumulative emission by 3–13% from 2010 to 2100. Introducing the central value of the three processes above (21) effectively increases the baseline-fast carbon emissions by ∼20% and can enhance warming by less than 0.5 °C (red dashed lines in Fig. 1). Also, the warming has been projected to increase methane emissions from wetlands by 0–100% compared with present-day wetland methane emissions. A 50% increase in wetland methane emissions by 2100 in response to the 4.1–5 °C warming could add at least another 0.5 °C (50% probability) to the projected warming.
In summary, the aerosol forcing uncertainty, although large in the 20th century, had only a small impact (<0.2 °C) on the projected warming trends (Fig. 1B). The climate-carbon feedback can amplify warming by ∼0.5–1 °C. This amplification, albeit large, was not included in our discussion on LPHI (5% probability) projections, because although the CO2 emissions from permafrost thawing and methane emissions from warmer wetlands are certain to increase, there is little confidence in the magnitude of the increase. A more detailed discussion of the major uncertainties mentioned above and the skewed probability distribution due to climate sensitivity is given in the SI Appendix, sections 4 and 5. It should also be pointed out that the probabilistic approach for projecting climate has been recognized and adopted in earlier studies (22⇓⇓–25), and its application in assigning climate risk is an active area of research (Box 2).
Based on analyses of available studies and model projections presented here, we propose the following constraints and criteria as governing principles for mitigation strategies.
Assigning Climate Risks.
Following the societal risk characterization as defined in Box 2, the projected warming trends in Fig. 1 for the two scenarios without climate policies fall under the following risk categories.
Near term (<2050).
Within three decades, the warming has a 50% probability of reaching dangerous levels (>1.5 °C), with the LPHI warming reaching catastrophic levels (>3 °C).
Long term (>2050).
Within eight decades, the warming has a 50% probability of subjecting the global population to catastrophic (>3 °C) to unknown risks (>5 °C) and a 5% probability of being fully in the unknown risk category, which also includes existential threats for everyone.
Mitigation Criteria for Warming.
The meaning of the phrase “well below 2 °C” was not adequately defined in the Paris Agreement. A hint was given through the aspirational goal of limiting the warming to “below 1.5 °C.” Using the probability approach, we propose that mitigation measures attempting to limit the warming to WB2C must consider adopting the following criteria: (i) The warming should be limited to below dangerous levels with at least a 50% probability; (ii) in addition, the LPHI warming should be limited to below catastrophic levels; and (iii) instead of stabilizing at 1.5 °C or 2 °C, the warming must begin to decrease with time before the end of the 21st century. In other words, we must bend the warming curve by the end of the century. Why is this criterion for bending the warming curve important? The Eemian period of 130,000 years ago was an interglacial period similar to the present and was warmer by ∼1 °C. It was associated with a 6- to 9-m rise in sea level (26), which suggests that a warming of 1.5 °C or more sustained over centuries can cause a catastrophic sea level rise.
The near-term (<2050) risk of dangerous (50% probability) to catastrophic imposes severe constraints on the urgency of the mitigation measures. Bending of the emission curves must begin now. As shown in earlier studies (27), future emissions largely determine the future warming for CO2. The future net emission of CO2 must be brought to zero before the warming exceeds dangerous levels. If the CO2emission is abruptly brought to zero by 2020, the CO2 concentration will decrease soon after and the warming (due solely to CO2) will stabilize at 2020 levels or even decrease slightly (SI Appendix, Fig. S8 D–F).
Because of the inertia in the socioeconomic system, the emissions most likely cannot be brought to zero immediately. Even if a scalable renewable technology were invented today to zero out all of the CO2emissions, it would be likely to take between three and five decades to spread such technology to the whole world (28), assuming a globally binding policy for carbon neutrality had already been put into place. This delay is partly due to the locked-in infrastructure and the upfront capital cost of quickly replacing as opposed to distributing the cost over decades. This inference is also consistent with most scenario studies (29, 30) for carbon neutrality pathways. The opposite extreme of zeroing out CO2 emissions by 2020 is a more gradual reduction to near-zero emissions by 2100. For this case, SI Appendix, Fig. S8B shows simulated CO2concentrations increase by ∼20 ppm to peak levels by 2030 and stay flat post-2050 and CO2-induced warming increases by another 0.6 °C (SI Appendix, Fig. S8C).
The constraint posed by the near-term (next three decades) risk of dangerous (50% probability) to catastrophic (5% probability) warming is that emission of CO2 and short-lived climate pollutants (SLCPs) should peak immediately and bend downward by 2020. There are hopeful signs that this is not an unrealistic goal. Worldwide CO2 emissions grew at a rate of 2.9% per year from 2000 to 2011, slowed to 1.3% per year from 2012 to 2014, and further decreased to near-zero growth (−0.2% per year) for 2015. This near-zero growth rate continued into 2016 (2). The low to near-zero growth rate since 2014 is due to a combination of several factors: switching from coal to oil and natural gas; an increase in production of renewable energy such as nuclear (1.3%), hydro (1%), and wind and solar (15%); and a reduction in carbon intensity of the economy. The negative growth rate from the United States (−2.6%) and China (−0.7%) mostly contributed to the recent bending of the emissions curve. While these are encouraging signs, aggressive policies will still be required to achieve carbon neutrality and climate stability.
The other long-lived GHG (LLGHG) with nonnegligible forcing is nitrous oxide (N2O) (SI Appendix, Fig. S6). Its current forcing is ∼0.15 watts per square meter (Wm−2) and is projected to increase to 0.23 Wm−2 by 2100 (SI Appendix, Fig. S6). Its net contribution to the warming from 2010 to 2100 is only about 0.1 °C (50% probability). Given the small size of its warming from the present to 2100 and the fact that N2O emission is tied to agriculture and it is the greatest challenge in limiting N2O emissions by 2100 with a world population of 10 billion, we are not targeting N2O in the mitigation measures discussed here.
Super Pollutants (SLCPs).
As of now, the CO2 warming in conjunction with the larger than 1 °C future warming imposed by SLCPs (Box 1 Figure) makes it extremely difficult, if not impossible, to limit the near-term (next three decades) warming below the dangerous levels by reducing CO2 concentrations. The following discussion is restricted to the impact of reducing the atmospheric concentrations of SLCPs with the required measures discussed later. This distinction is important since measures to reduce CO2 would also indirectly reduce some of the SLCP emissions.
The four SLCPs [methane, tropospheric ozone, black carbon (BC), and hydrofluorocarbons (HFCs)] are referred to as super pollutants (SPs) since their global warming potential (on a 100-year time scale) ranges from 25 to 2,000. When the warming due to CO2 is added to the SP warming, the 2 °C threshold should have been crossed, but for the masking effect of the −0.9 °C (−0.5 to −1.5 °C) aerosol cooling due to sulfate, nitrate, and dust aerosols (Box 1 Figure). Since these cooling aerosols [along with BC and organic carbon (OC) aerosols] are also the major sources of air pollution, leading up to 7 million mortalities (31), they are being regulated independent of climate mitigation regulations. Phasing out the cooling aerosols completely within a few decades (e.g., by switching to renewable fuels) can lead to an additional warming of about 0.3 °C between 2020 and 2050 (SI Appendix, Fig. S7).
In summary, dangerous to catastrophic climate changes in the near term can be avoided only by reducing the concentrations of SPs substantially beginning in 2020. Technological measures to reduce SLCPs are mostly available and fall under two categories:
i) CO2-dedicated measures: Technology measures to curb CO2 emissions such as switching to renewables will also mitigate some of the emissions of SLCPs: methane (22% of methane emissions are due to production and consumption of fossil fuels), BC emitted by diesel vehicles, and emissions of ozone precursors such as carbon monoxide and NOx (nitrogen oxides) by fossil fuel consumption (32, 33).
ii) SLCP-dedicated measures. Technological measures independent of CO2-dedicated measures are already available (34) for reducing methane, ozone, and BC concentration, and deploying immediately to scale is feasible. For the halocarbons, including HFCs, the Kigali amendment to the Montreal Protocol, which was approved by 160 nations (35), will phase out high-GWP (global warming potential) HFCs by 2050. The Montreal Protocol was unanimously adopted by the United Nations in the 1980s to ban chlorofluorocarbons (CFCs) due to the negative impact on the ozone layer, but it was also effective in mitigating the super greenhouse effect of these halocarbons (36). The Kigali amendment specifically recognized the climate-warming effect of halocarbons and approved phasing down these powerful climate pollutants.
The mitigation of the coemitted SLCPs and cooling aerosols by CO2-dedicated measures requires special consideration (33). SLCP emissions are not entirely independent of CO2 emissions, and emission rates of SLCPs can decrease due to CO2 mitigation, and likewise CO2 emissions can decrease due to mitigation of SLCPs. The role of coemitted SLCPs that are dependent on CO2 is estimated in SI Appendix, Fig. S5. A fraction of CH4 (about 70%) and BC (about 30%) emissions can be mitigated through CO2-dedicated measures. While HFCs are not dependent on CO2 mitigation, CO2-dedicated mitigation measures can accomplish roughly 50% of the 0.6 °C mitigated warming by SLCPs by 2050 and 40% of the 1.2 °C mitigated warming by 2100. Another complexity of the coemission issue is that a major part of the cooling aerosols (mostly sulfates and nitrates) is also coemitted by CO2-dedicated measures. Hence, the CO2 measures implemented in 2020 will unmask some of the aerosol cooling (red lines in SI Appendix, Fig. S5) and offset the warming reduction by CO2 and SLCP mitigation. In the baseline scenarios of this study, the cooling aerosols are regulated gradually between 2020 and 2100 (SI Appendix, Fig. S6), whereas in the mitigation scenario examined here, CO2 mitigation is implemented starting from 2020 and CO2 emission is brought to net zero in about three decades (SI Appendix, Fig. S2B). As a result, the unmasking of coemitted aerosol cooling (a net warming effect) is more rapid in the decreasing CO2 emissions beginning in 2020 (CN2020) mitigation scenario (SI Appendix, Fig. S5B vs. S7).
Given the uncertainties in aerosol forcing; the socioeconomic and political challenges involved in CO2mitigation; climate feedback uncertainties; and, above all, the possibilities of catastrophic climate change (a 5% probability) within three decades, a no-regret policy will be to undertake both CO2-dedicated and SLCP-dedicated measures simultaneously. In 2016, California did exactly that by passing the SB-1383 bill (37), which targets mitigation of methane, HFCs, and BC beginning in 2020. Since California has already passed laws to reduce CO2 emissions by 80% before 2050, the SB-1383 bill is a demonstration that both CO2 and the SPs can be mitigated simultaneously using complementary technologies. Furthermore, in 2011, the United Nations Environment Program formed the Climate and Clean Air Coalition (www.ccacoalition.org/en) to mitigate all four SLCPs in coalition with many member nations.
We will now take up the mitigation strategy subject to the criteria and constraints identified above. We have to consider two time scales. First is the near term of three decades extending from now to midcentury, when the warming is likely to cross over to the dangerous threshold (baseline curves in Fig. 1A and Fig. 3). Next is the long term, extending from midcentury to 2100, when the baseline LPHI warming can reach beyond the catastrophic regime into the unknown domain (baseline curves in Fig. 1B and Fig. 2).
The carbon neutral (CN) lever is for mitigation of CO2 emissions. It has taken society nearly 220 years (from 1750 to 1970) to emit the first trillion tons of CO2 and only another 40 years (1970–2010) to emit the next trillion tons. The third trillion tons, under current emission trends, would be emitted by 2030 and the fourth trillion tons before 2050 (Box 1 and SI Appendix, Fig. S1A). Even if the INDCs are implemented rigorously and verifiably, the third trillion tons will be added by 2035 (SI Appendix, Fig. S2A). Earlier studies (30) have identified that cumulative CO2 emissions must be limited to less than 3.7 trillion tons (or 1 trillion tons of carbon) to have any chance of limiting the warming below 2 °C. These studies often focused on targeting the central value (50% probability) of the warming and less on the LPHI warming. The maximum warming reduction feasible by pulling on the CN lever can be inferred from Box 1, which shows the 2100 baseline-fast warming by CO2 alone to be 2.6 °C. Since the lifetime of CO2 ranges from decades (for the first 50%), to centuries, to millennia (for 20%) (38), not all of the 2.6 °C warming can be mitigated by 2100. Constrained by CO2 lifetime and the diffusion time of new technologies (decades), the scenarios considered here (SI Appendix, Fig. S2A) suggest that about half of the 2.6 °C CO2 warming in the baseline-fast scenario can be mitigated by 2100 and only 0.1–0.3 °C can be mitigated by 2050.
Had we followed the baseline-default trajectory, the CO2-alone warming would have been 3.5 °C instead of 2.6 °C as shown in Fig. 2. It is important to note both scenarios use fossil fuels. Since the baseline-default scenario reduces carbon intensity of the economy by only 50% from the 2010 values compared with an 80% reduction in the baseline-fast scenario, we infer that reducing the carbon intensity of the economy is a very potent mitigation measure since, by itself, it can reduce the 2100 CO2 warming by 0.9 °C from 3.5 to 2.6 °C (additional details are provided in SI Appendix, section 6).
The SP lever targets SLCPs. Reducing SLCP emissions thins the SP blanket within few decades, given the shorter lifetimes of SLCPs (weeks for BC to about 15 years for HFCs). The mitigation potential of the SP lever with a maximum deployment of current technologies (32) is about 0.6 °C by 2050 and 1.2 °C by 2100 (SI Appendix, Fig. S5B and Table S1).
Carbon Extraction and Sequestration Lever.
The third lever is the carbon extraction and sequestration (CES) lever, which will extract CO2 from the source (e.g., the coal power plant) or from the air and sequester it. While the CN and SP levers can help mitigate the 50% probability warming targets, they are inadequate to mitigate the LPHI warming. Ultimately, we must thin the CO2 greenhouse blanket by removing the CO2 that is already in the atmosphere.
Given the near-term risk of exceeding the dangerous to catastrophic thresholds, the timing for pulling these levers is a crucial issue. Ideally, these levers should be pulled immediately by 2020. We will now elaborate on three options to constrain the choices considered in earlier studies, starting with the least preferable option first.
This option involves following the INDCs until 2030 and bending the CO2 emissions downward by 2030, and bending the SP (SLCP) emissions downward by 2020 and reaching full potential by 2060. The CO2 part of this option is referred to as CN2030, while the SLCP part is referred to as SLCP2020 (Table 1). CN2030 will achieve carbon neutrality by 2060–2070, which will limit the cumulative CO2 emissions (since preindustrial) to 3.2 trillion tons (SI Appendix, Fig. S2A). We refer to this as the Target-2C option since it has been proposed by several earlier studies (3, 23). However, even when CN2030 is combined with SLCP2020, the Target-2C option will only be able to limit the 50% probability warming below 2 °C (Fig. 3) but will fail to meet the mitigation criteria of avoiding dangerous warming (50% probability of warming less than 1.5 °C) both in the near term and in the long term (SI Appendix, Figs. S9 and S10).
Scenarios of CO2 and SLCPs considered in the study
Instead of allowing CO2 emissions to increase until 2030, we should start bending the curve by 2020 [i.e., CN2020 and achieving a CN status by 2050 (SI Appendix, Fig. S2B)]. Since 2020 is just a few years away, this is a highly optimistic option. The 10-year head start in bending the CO2 curve, when combined with SLCP2020, was sufficient to bring down the probability of 1.5 °C warming (the threshold for dangerous warming) from more than 99% to less than 50% (blue dashed curve in SI Appendix, Figs. S9 and S10). Furthermore, advancing the CN lever by 10 years has reduced the probability of catastrophic warming (>3 °C) to below 5%. The main reason is because the CN2030 case allows additional emissions of 1.2 trillion tons between 2010 and 2050 (SI Appendix, Fig. S2A), whereas in the CN2020 case, the additional increase is only 0.5 trillion tons (SI Appendix, Fig. S2B). The inference is that to meet the criteria for avoiding dangerous warming (<1.5 °C warming with 50% probability) as well as catastrophic warming (<3 °C warming with 95% probability), the cumulative emissions from preindustrial to 2100 must be less than 2.5 trillion tons of CO2. This option, compared with the Target-2C option, illustrates the large impact of a 10-year delay in bending the CO2 emissions curve on increasing the risks of climate change.
This case involves pulling all three levers (CN, SP, and CES levers) with the CN2030 and the SLCP2020 options. This case is shown in Figs. 2 and 3 (green curves in both). The model simulations suggest that CES needs to be deployed by 2030 and to sequester 16 billion tons (Gt) of CO2 per year (SI Appendix, Fig. S2C) for several decades into the late 21st century to limit the cumulative CO2 emissions to 2.2 trillion tons (or 0.6 trillion tons of carbon). The CES of 16 Gt of CO2 per year will extract one-third of the 3.2 trillion tons of CO2 (CES1t) that would have been added by human activities since the industrial era. To get a perspective on the enormity of this extraction, the 2010 fossil fuel CO2 emission is 32 Gt of CO2 per year. This case meets all three criteria with a small exception. First, the option meets the criteria of limiting the long-term warming below the dangerous level (<50% probability of exceeding 1.5 °C) and below the catastrophic level (<5% probability of exceeding 3 °C). Next, the end-of-century temperature curve is trending downward, providing great relief for the expected sea level rise during centuries beyond 2100. The one exception is that this case does not limit the near-term warming below the dangerous level (with an “overshoot” at 2050) (6).
Basically, for a safe climate, all three levers (CN, SP, and CES) must be deployed as soon as possible. The CN and SP levers must be deployed by 2030 and 2020, respectively; the cumulative CO2 emissions from preindustrial must be limited to 2.2 trillion tons of CO2 (or 0.6 trillion tons of carbon); and the CES lever should extract and sequester as much as 1 trillion tons of CO2 (CES1t), depending on when the CN lever is deployed. If the CN lever is deployed as early as 2020, the required CES is much less than 1 trillion tons.
We propose that mitigation goals be set in terms of climate risk category instead of a temperature threshold. In this paper, we offer three broad risk categories, but it is likely that a more granular set of categories is required. The temperature threshold has served policy very well; however, given the imminence of dangerous warming within decades, the focus must broaden to include extreme climate changes. Precipitation, flooding, fire, and drought will all become serious sources of concern. The temperature will still occupy our attention because of the heat stress phenomenon and the likelihood of approximately half of the population exposed to deadly heat by 2050 (Box 2).
We conclude with a commentary on the feasibility of the mitigation options considered thus far. Over 24 technological measures to reduce SLCPs have been detailed previously (39) (details are provided in SI Appendix). These measures include providing clean cook stoves to the poorest three billion of the world’s total population and installing particulate filters in all diesel vehicles to reduce global BC emissions by nearly 80% and also reduce air pollution-related mortalities by ∼2 million; routine maintenance of gas pipes and banning gas flaring to reduce methane leaks; recovering methane from landfills, water sewage treatment plants, and farm manure; replacing HFCs with other available refrigerants that have negligible greenhouse effects; and installing catalytic converters in vehicles to reduce emissions of ozone precursors.
CN levers require switching from fossil fuels to renewables such as wind, solar, geothermal and nuclear sources, among others. Also, CO2 emissions from industrial processes should be eliminated. This requires electrification of all end uses and production of electricity from renewables (40). Since many renewables (solar and wind) are intermittent, storage is a crucial issue. Batteries, hydrogen production by renewables, and pumped hydropower are all possible options for storage. While about 50% of reductions are possible with scaling up of existing technologies, innovations are required for achieving carbon neutrality in a cost-effective manner (40). Achievement of carbon neutrality also requires societal transformation, governance, and market mechanisms such as cap and trade and carbon pricing (40). The encouraging sign is that 52 cities, 65 businesses, and numerous universities have already embarked on the CN pathway (41). Some of these living laboratories, like California and Stockholm, have shown that the gross domestic product (GDP) can be decoupled from carbon emissions. Their carbon emission per GDP has decreased by 20% while bending the carbon emissions curve. The technology development and innovations from these living laboratories should be scaled to the world to greatly accelerate efforts to achieve CN within decades.
Of the three levers recommended here, the third lever dealing with CES is the most challenging and formidable due to lack of scalable technologies. However, many technologies are being explored, including capturing CO2 in bioenergy power plants (42), biochar production by pyrolysis and storage in soils (43), restoration of soil organic pools (44), chemical weathering of rocks, mineral sequestration, reforestation, and urban forestry, among others. The availability of land and conflict with food production is another important constraint in some of the CES solutions. Major breakthroughs are needed urgently, and in the meantime, the best option is to start on the CN goal by 2020 and mitigate the SPs as soon as possible, since cost-effective technologies are already present to immediately start bending the emission curves.
Author contributions: V.R. designed research; Y.X. and V.R. performed research; Y.X. and V.R. analyzed data; and Y.X. and V.R. wrote the paper.