Strengthening and implementing the global response

Greater scale, speed and change in investment patterns

There is agreement in the literature reviewed by Chapter 2 that staying below 1.5°C would entail significantly greater transformation in terms of energy systems, lifestyles and investments patterns compared to 2°C-consistent pathways. Yet there is limited evidence and low agreement regarding the magnitudes and costs of the investments (Sections 2.5.1, 2.5.2 and 4.4.5). Based on the IAM literature reviewed in Chapter 2, climate policies in line with limiting warming to 1.5°C would require a marked upscaling of supply-side energy system investments between now and mid-century, reaching levels of between 1.6–3.8 trillion USD yr⁻¹ globally with an average of about 3.5 trillion USD yr⁻¹ over 2016–2050 (see Figure 2.27). This can be compared to an average of about 3.0 trillion USD yr⁻¹ over the same period for 2°C-consistent pathways (also in Figure 2.27).

Not only the level of investment but also the type and speed of sectoral transformation would be impacted by the transitions associated with 1.5°C-consistent pathways. IAM literature projects that investments in low-emission energy would overtake fossil fuel investments globally by 2025 in 1.5°C-consistent pathways (Chapter 2, Section 2.5.2). The projected low-emission investments in electricity generation allocations over the period 2016–2050 are: solar (0.09–1.0 trillion USD yr⁻¹), wind (0.1–0.35 trillion USD yr⁻¹), nuclear (0.1–0.25 trillion USD yr⁻¹), and transmission, distribution, and storage (0.3–1.3 trillion USD yr⁻¹). In contrast, investments in fossil fuel extraction and unabated fossil electricity generation along a 1.5°C-consistent pathway are projected to drop by 0.3–0.85 trillion USD yr⁻¹ over the period 2016–2050, with investments in unabated coal generation projected to halt by 2030 in most 1.5°C-consistent pathways (Chapter 2, Section 2.5.2). Estimates of investments in other infrastructure are currently unavailable, but they could be considerably larger in volume than solely those in the energy sector (Section 4.4.5).

Greater policy design and decision-making implications

The 1.5°C-consistent pathways raise multiple challenges for effective policy design and responses to address the scale, speed, and pace of mitigation technology, finance and capacity building needs. These policies and responses would also need to deal with their distributional implications while addressing adaptation to residual climate impacts (see Chapter 5). The available literature indicates that 1.5°C-consistent pathways would require robust, stringent and urgent transformative policy interventions targeting the decarbonization of energy supply, electrification, fuel switching, energy efficiency, land-use change, and lifestyles (Chapter 2, Section 2.5, 4.4.2, 4.4.3). Examples of effective approaches to integrate mitigation with adaptation in the context of sustainable development and to deal with distributional implications proposed in the literature include the utilization of dynamic adaptive policy pathways (Haasnoot et al., 2013; Mathy et al., 2016)²⁹ and transdisciplinary knowledge systems (Bendito and Barrios, 2016)³⁰.

Yet, even with good policy design and effective implementation, 1.5°C-consistent pathways would incur higher costs. Projections of the magnitudes of global economic costs associated with 1.5°C-consistent pathways and their sectoral and regional distributions from the currently assessed literature are scant, yet suggestive. For example, IAM simulations assessed in Chapter 2 project (with a probability greater than 50%) that marginal abatement costs, typically represented in IAMs through a carbon price, would increase by about 3–4 times by 2050 under a 1.5°C-consistent pathway compared to a 2°C-consistent pathway (Chapter 2, Section 2.5.2, Figure 2.26). Managing these costs and distributional effects would require an approach that takes account of unintended cross-sector, cross-nation, and cross-policy trade-offs during the transition (Droste et al., 2016; Stiglitz et al., Pollitt, 2018; 2017; Sands, 2018; Siegmeier et al., 2018)³¹.

Greater sustainable development implications

Few studies address the relations between the Shared Socio-Economic Pathways (SSPs) and the Sustainable Developments Goals (SDGs) (O’Neill et al., 2015; Riahi et al., 2017)³². Nonetheless, literature on potential synergies and trade-offs between 1.5°C-consistent mitigation pathways and sustainable development dimensions is emerging (Chapter 2, Section 2.5.3, Chapter 5, Section 5.4). Areas of potential trade-offs include reduction in final energy demand in relation to SDG 7 (the universal clean energy access goal) and increase of biomass production in relation to land use, water resources, food production, biodiversity and air quality (Chapter 2, Sections 2.4.3, 2.5.3). Strengthening the institutional and policy responses to deal with these challenges is discussed in Section 4.4 together with the linkage between disruptive changes in the energy sector and structural changes in other infrastructure (transport, building, water and telecommunication) sectors. A more in-depth assessment of the complexity and interfaces between 1.5°C-consistent pathways and sustainable development is presented in Chapter 5.

Climate variability and uncertainties in the underlying assumptions in Chapter 2’s IAMs as well as in model comparisons complicate discerning the implications for climate impacts, adaptation options and avoided adaptation investments at the global level of 2°C compared to 1.5°C warming (James et al., 2017; Mitchell et al., 2017)³³.

Incremental warming from 1.5°C to 2°C would lead to significant increases in temperature and precipitation extremes in many regions (Chapter 3, Section 3.3.2, 3.3.3). Those projected changes in climate extremes under both warming levels, however, depend on the emissions pathways, as they have different greenhouse gas (GHG)/aerosol forcing ratios. Impacts are sector-, system- and region-specific, as described in Chapter 3. For example, precipitation-related impacts reveal distinct regional differences (Chapter 3, Sections 3.3.3, 3.3.4, 3.3.5, 3.4.2). Similarly, regional reduction in water availability and the lengthening of regional dry spells have negative implications for agricultural yields depending on crop types and world regions (see for example Chapter 3, Sections 3.3.4, 3.4.2, 3.4.6).

Adaptation helps reduce impacts and risks. However, adaptation has limits. Not all systems can adapt, and not all impacts can be reversed (Cross-Chapter Box 12 in Chapter 5). For example, tropical coral reefs are projected to be at risk of severe degradation due to temperature-induced bleaching (Chapter 3, Box 3.4).

Realizing 1.5°C-consistent pathways would require rapid and systemic changes on unprecedented scales (see Chapter 2 and Section 4.2.1). This section examines whether the needed rates of change have historical precedents and are underway.

Some studies conduct a de-facto validation of IAM projections. For CO₂ emission intensity over 1990–2010, this resulted in the IAMs projecting declining emission intensities while actual observations showed an increase. For individual technologies (in particular solar energy), IAM projections have been conservative regarding deployment rates and cost reductions (Creutzig et al., 2017)³⁶, suggesting that IAMs do not always impute actual rates of technological change resulting from influence of shocks, broader changes and mutually reinforcing factors in society and politics (Geels and Schot, 2007; Daron et al., 2015; Sovacool, 2016; Battiston et al., 2017)³⁷.

Other studies extrapolate historical trends into the future (Höök et al., 2011; Fouquet, 2016)³⁸, or contrast the rates of change associated with specific temperature limits in IAMs (such as those in Chapter 2) with historical trends to investigate plausibility of emission pathways and associated temperature limits (Wilson et al., 2013; Gambhir et al., 2017; Napp et al., 2017)³⁹. When metrics are normalized to gross domestic product (GDP; as opposed to other normalization metrics such as primary energy), low-emission technology deployment rates used by IAMs over the course of the coming century are shown to be broadly consistent with past trends, but rates of change in emission intensity are typically overestimated (Wilson et al., 2013; Loftus et al., 2014; van Sluisveld et al., 2015)⁴⁰. This bias is consistent with the findings from the ‘validation’ studies cited above, suggesting that IAMs may under-report the potential for supply-side technological change assumed in 1.5°-consistent pathways, but may be more optimistic about the systemic ability to realize incremental changes in reduction of emission intensity as a consequence of favourable energy efficiency payback times (Wilson et al., 2013)⁴¹. This finding suggests that barriers and enablers other than costs and climate limits play a role in technological change, as also found in the innovation literature (Hekkert et al., 2007; Bergek et al., 2008; Geels et al., 2016b)⁴².

One barrier to a greater rate of change in energy systems is that economic growth in the past has been coupled to the use of fossil fuels. Disruptive innovation and socio-technical changes could enable the decoupling of economic growth from a range of environmental drivers, including the consumption of fossil fuels, as represented by 1.5°C-consistent pathways (UNEP, 2014; Newman, 2017)⁴³. This may be relative decoupling due to rebound effects that see financial savings generated by renewable energy used in the consumption of new products and services (Jackson and Senker, 2011; Gillingham et al., 2013)⁴⁴, but in 2015 and 2016 total global GHG emissions have decoupled absolutely from economic growth (IEA, 2017g; Peters et al., 2017)⁴⁵. A longer data trend would be needed before stable decoupling can be established. The observed decoupling in 2015 and 2016 was driven by absolute declines in both coal and oil use since the early 2000s in Europe, in the past seven years in the United States and Australia, and more recently in China (Newman, 2017)⁴⁶. In 2017, decoupling in China reversed by 2% due to a drought and subsequent replacement of hydropower with coal-fired power (Tollefson, 2017)⁴⁷, but this reversal is expected to be temporary (IEA, 2017c)⁴⁸. Oil consumption in China is still rising slowly, but absolute decoupling is ongoing in megacities like Beijing (Gao and Newman, 2018)⁴⁹ (see Box 4.9).

In some regions and places, incremental adaptation would not be sufficient to mitigate the impacts of climate change on social-ecological systems (see Chapter 3). Transformational adaptation would then be required (Bahadur and Tanner, 2014; Pant et al., 2015; Gillard, 2016; Gillard et al., 2016; Colloff et al., 2017; Termeer et al., 2017)⁵⁰. Transformational adaptation refers to actions aiming at adapting to climate change resulting in significant changes in structure or function that go beyond adjusting existing practices (Dowd et al., 2014; IPCC, 2014a; Few et al., 2017)⁵¹, including approaches that enable new ways of decision-making on adaptation (Colloff et al., 2017)⁵². Few studies have assessed the potentially transformative character of adaptation options (Pelling et al., 2015; Rippke et al., 2016; Solecki et al., 2017)⁵³, especially in the context of warming of 1.5°C.

Transformational adaptation can be adopted at a large scale, can lead to new strategies in a region or resource system, transform places and potentially shift locations (Kates et al., 2012)⁵⁴. Some systems might require transformational adaptation at 1.5°C. Implementing adaptation policies in anticipation of 1.5°C would require transformation and flexible planning of adaptation (sometimes called adaptation pathways) (Rothman et al., 2014; Smucker et al., 2015; Holland, 2017; Gajjar et al., 2018)⁵⁵, an understanding of the varied stakeholders involved and their motives, and knowledge of less visible aspects of vulnerability based on social, cultural, political, and economic factors (Holland, 2017)⁵⁶. Transformational adaptation would seek deep and long-term societal changes that influence sustainable development (Chung Tiam Fook, 2017; Few et al., 2017)⁵⁷.

Adaptation requires multidisciplinary approaches integrating scientific, technological and social dimensions. For example, a framework for transformational adaptation and the integration of mitigation and adaptation pathways can transform rural indigenous communities to address risks of climate change and other stressors (Thornton and Comberti, 2017)⁵⁸. In villages in rural Nepal, transformational adaptation has taken place, with villagers changing their agricultural and pastoralist livelihood strategies after years of lost crops due to changing rain patterns and degradation of natural resources (Thornton and Comberti, 2017)⁵⁹. Instead, they are now opening stores, hotels, and tea shops. In another case, the arrival of an oil pipeline altered traditional Alaskan communities’ livelihoods. With growth of oil production, investments were made for rural development. A later drop in oil production decreased these investments. Alaskan indigenous populations are also dealing with impacts of climate change, such as sea level rise, which is altering their livelihood sources. Transformational adaptation is taking place by changing the energy matrix to renewable energy, in which indigenous people apply their knowledge to achieve environmental, economic, and social benefits (Thornton and Comberti, 2017)⁶⁰.

Demand-driven disruptive innovations that emerge as the product of political and social changes across multiple scales can be transformative (Seba, 2014; Christensen et al., 2015; Green and Newman, 2017a)⁶¹. Such innovations would lead to simultaneous, profound changes in behaviour, economies and societies (Seba, 2014; Christensen et al. 2015), but are difficult to predict in supply-focused economic models (Geels et al., 2016a; Pindyck, 2017)⁶². Rapid socio-technical change has been observed in the solar industry (Creutzig et al. (2017)⁶³. Similar changes to socio-ecological systems can stimulate adaptation and mitigation options that lead to more climate-resilient systems (Adger et al., 2005; Ostrom, 2009; Gillard et al., 2016)⁶⁴ (see the Alaska and Nepal examples in Section 4.2.2.2). The increase in roof-top solar and energy storage technology as well as the increase in passive housing and net zero-emissions buildings are further examples of such disruptions (Green and Newman, 2017b)⁶⁵. Both roof-top solar and energy storage have benefitted from countries’ economic growth strategies and associated price declines in photovoltaic technologies, particularly in China (Shrivastava and Persson, 2018)⁶⁶, as well as from new information and communication technologies (Koomey et al., 2013)⁶⁷, rising demand for electricity in urban areas, and global concern regarding greenhouse gas emissions (Azeiteiro and Leal Filho, 2017; Lutz and Muttarak, 2017; Wamsler, 2017)⁶⁸.

System co-benefits can create the potential for mutually enforcing and demand-driven climate responses (Jordan et al., 2015; Hallegatte and Mach, 2016; Pelling et al., 2018)⁶⁹, and for rapid and transformational change (Cole, 2015; Geels et al., 2016b; Hallegatte and Mach, 2016)⁷⁰. Examples of co-benefits include gender equality, agricultural productivity (Nyantakyi-Frimpong and Bezner-Kerr, 2015)⁷¹, reduced indoor air pollution (Satterthwaite and Bartlett, 2017)⁷², flood buffering (Colenbrander et al., 2017)⁷³, livelihood support (Shaw et al., 2014; Ürge-Vorsatz et al., 2014)⁷⁴, economic growth (GCEC, 2014; Stiglitz et al., 2017)⁷⁵, social progress (Steg et al., 2015; Hallegatte and Mach, 2016)⁷⁶ and social justice (Ziervogel et al., 2017; Patterson et al., 2018)⁷⁷.

Innovations that disrupt entire systems may leave firms and utilities with stranded assets, as the transition can happen very quickly (IPCC, 2014b; Kossoy et al., 2015)⁷⁸. This may have consequences for fossil fuels that are rendered ‘unburnable’ (McGlade and Ekins, 2015)⁷⁹ and fossil fuel-fired power and industry assets that would become obsolete (Caldecott, 2017; Farfan and Breyer, 2017)⁸⁰. The presence of multiple barriers and enablers operating in a system implies that rapid change, whether the product of many small changes (Termeer et al., 2017)⁸¹ or large-scale disruptions, is seldom an insular or discrete process (Sterling et al., 2017)⁸². This finding informs the multidimensional nature of feasibility in Cross-Chapter Box 3 in Chapter 1 which is applied in Section 4.5. Climate responses that are aligned with multiple feasibility dimensions and combine adaptation and mitigation interventions with non-climate benefits can accelerate change and reduce risks and costs (Fazey et al., 2018)⁸³. Also political, social and technological influences on energy transitions, for example, can accelerate them faster than narrow techno-economic analysis suggests is possible (Kern and Rogge, 2016)⁸⁴, but could also introduce new constraints and risks (Geels et al., 2016b; Sovacool, 2016; Eyre et al., 2018)⁸⁵.

Disruptive innovation and technological change may play a role in mitigation and in adaptation. The next section assesses mitigation and adaptation options in energy, land and ecosystem, urban and infrastructure and industrial systems.

All renewable energy options have seen considerable advances over the years since AR5, but solar energy and both onshore and offshore wind energy have had dramatic growth trajectories. They appear well underway to contribute to 1.5°C-consistent pathways (IEA, 2017c; IRENA, 2017b; REN21, 2017)⁸⁶.

The largest growth driver for renewable energy since AR5 has been the dramatic reduction in the cost of solar photovoltaics (PV) (REN21, 2017)⁸⁷. This has made rooftop solar competitive in sunny areas between 45° north and south latitude (Green and Newman, 2017b)⁸⁸, though IRENA (2018)⁸⁹ suggests it is cost effective in many other places too. Solar PV with batteries has been cost effective in many rural and developing areas (Pueyo and Hanna, 2015; Szabó et al., 2016; Jimenez, 2017)⁹⁰, for example 19 million people in Bangladesh now have solar-battery electricity in remote villages and are reporting positive experiences on safety and ease of use (Kabir et al., 2017)⁹¹. Small-scale distributed energy projects are being implemented in developed and developing cities where residential and commercial rooftops offer potential for consumers becoming producers (called prosumers) (ACOLA, 2017; Kotilainen and Saari, 2018)⁹². Such prosumers could contribute significantly to electricity generation in sun-rich areas like California (Kurdgelashvili et al., 2016)⁹³ or sub-Saharan Africa in combination with micro-grids and mini-grids (Bertheau et al., 2017)⁹⁴. It could also contribute to universal energy access (SDG 7) as shown by (IEA, 2017c)⁹⁵.

The feasibility of renewable energy options depends to a large extent on geophysical characteristics of the area where the option is implemented. However, technological advances and policy instruments make renewable energy options increasingly attractive in other areas. For example, solar PV is deployed commercially in areas with low solar insolation, like northwest Europe (Nyholm et al., 2017)⁹⁶. Feasibility also depends on grid adaptations (e.g., storage, see below) as renewables grow (IEA, 2017c)⁹⁷. For regions with high energy needs, such as industrial areas (see Section 4.3.4), high-voltage DC transmission across long distances would be needed (MacDonald et al., 2016)⁹⁸.

Another important factor affecting feasibility is public acceptance, in particular for wind energy and other large-scale renewable facilities (Yenneti and Day, 2016; Rand and Hoen, 2017; Gorayeb et al., 2018)⁹⁹ that raise landscape management (Nadaï and Labussière, 2017)¹⁰⁰ and distributional justice (Yenneti and Day, 2016)¹⁰¹ challenges. Research indicates that financial participation and community engagement can be effective in mitigating resistance (Brunes and Ohlhorst, 2011; Rand and Hoen, 2017)¹⁰² (see Section 4.4.3).

Bottom-up studies estimating the use of renewable energy in the future, either at the global or at the national level, are plentiful, especially in the grey literature. It is hotly debated whether a fully renewable energy or electricity system, with or without biomass, is possible (Jacobson et al., 2015, 2017)¹⁰³ or not (Clack et al., 2017; Heard et al., 2017)¹⁰⁴, and by what year. Scale-up estimates vary with assumptions about costs and technological maturity, as well as local geographical circumstances and the extent of storage used (Ghorbani et al., 2017; REN21, 2017)¹⁰⁵. Several countries have adopted targets of 100% renewable electricity (IEA, 2017c)¹⁰⁶ as this meets multiple social, economic and environmental goals and contributes to mitigation of climate change (REN21, 2017)¹⁰⁷.

Bioenergy is renewable energy from biomass. Biofuel is biomass-based energy used in transport. Chapter 2 suggests that pathways limiting warming to 1.5°C would enable supply of 67–310 (median 150) EJ yr⁻¹ (see Table 2.8) from biomass. Most scenarios find that bioenergy is combined with carbon dioxide capture and storage (CCS, BECCS) if it is available but also find robust deployment of bioenergy independent of the availability of CCS (see Chapter 2, Section 2.3.4.2 and Section 4.3.7 for a discussion of BECCS). Detailed assessments indicate that deployment is similar for pathways limiting global warming to below 2°C (Chum et al., 2011; P. Smith et al., 2014; Creutzig et al., 2015b)¹⁰⁸. There is however high agreement that the sustainable bioenergy potential in 2050 would be restricted to around 100 EJ yr⁻¹ (Slade et al., 2014; Creutzig et al., 2015b)¹⁰⁹. Sustainable deployment at such or higher levels envisioned by 1.5°C-consistent pathways may put significant pressure on available land, food production and prices (Popp et al., 2014b; Persson, 2015; Kline et al., 2017; Searchinger et al., 2017)¹¹⁰, preservation of ecosystems and biodiversity (Creutzig et al., 2015b; Holland et al., 2015; Santangeli et al., 2016)¹¹¹, and potential water and nutrient constraints (Gerbens-Leenes et al., 2009; Gheewala et al., 2011; Bows and Smith, 2012; Smith and Torn, 2013; Bonsch et al., 2016; Lampert et al., 2016; Mouratiadou et al., 2016; Smith et al., 2016b; Wei et al., 2016; Mathioudakis et al., 2017)¹¹²; but there is still low agreement on these interactions (Robledo-Abad et al., 2017)¹¹³. Some of the disagreement on the sustainable capacity for bioenergy stems from global versus local assessments. Global assessments may mask local dynamics that exacerbate negative impacts and shortages while at the same time niche contexts for deployment may avoid trade-offs and exploit co-benefits more effectively. In some regions of the world (e.g., the case of Brazilian ethanol, see Box 4.7, where land may be less of a constraint, the use of bioenergy is mature and the industry is well developed), land transitions could be balanced with food production and biodiversity to enable a global impact on CO₂ emissions (Jaiswal et al., 2017)¹¹⁴.

The carbon intensity of bioenergy, key for both bioenergy as an emission-neutral energy option and BECCS as a CDR measure, is still a matter of debate (Buchholz et al., 2016; Liu et al., 2018)¹¹⁵ and depends on management (Pyörälä et al., 2014; Torssonen et al., 2016; Baul et al., 2017; Kilpeläinen et al., 2017)¹¹⁶; direct and indirect land-use change emissions (Plevin et al., 2010; Schulze et al., 2012; Harris et al., 2015; Repo et al., 2015; DeCicco et al., 2016; Qin et al., 2016)¹¹⁷²; the feedstock considered; and time frame (Zanchi et al., 2012; Daioglou et al., 2017; Booth, 2018; Sterman et al., 2018)¹¹⁸, as well as the availability of coordinated policies and management to minimize negative side effects and trade-offs, particularly those around food security (Stevanović et al., 2017)¹¹⁹ and livelihood and equity considerations (Creutzig et al., 2013; Calvin et al., 2014)¹²⁰ .

Biofuels are a part of the transport sector in some cities and countries, and may be deployed as a mitigation option for aviation, shipping and freight transport (see Section 4.3.3.5) as well as industrial decarbonization (IEA, 2017g)¹²¹ (Section 4.3.4), though only Brazil has mainstreamed ethanol as a substantial, commercial option. Lower emissions and reduced urban air pollution have been achieved there by use of ethanol and biodiesel as fuels (Hill et al., 2006; Salvo et al., 2017)¹²² (see Box 4.7).

Many scenarios in Chapter 2 and in AR5 (Bruckner et al., 2014)¹²³ project an increase in the use of nuclear power, while others project a decrease. The increase can be realized through existing mature nuclear technologies or new options (generation III/IV reactors, breeder reactors, new uranium and thorium fuel cycles, small reactors or nuclear cogeneration).

Even though scalability and speed of scaling of nuclear plants have historically been high in many nations, such rates are currently not achieved anymore. In the 1960s and 1970s, France implemented a programme to rapidly get 80% of its power from nuclear in about 25 years (IAEA, 2018)¹²⁴, but the current time lag between the decision date and the commissioning of plants is observed to be 10-19 years (Lovins et al., 2018)¹²⁵. The current deployment pace of nuclear energy is constrained by social acceptability in many countries due to concerns over risks of accidents and radioactive waste management (Bruckner et al., 2014)¹²⁶. Though comparative risk assessment shows health risks are low per unit of electricity production (Hirschberg et al., 2016)¹²⁷, and land requirement is lower than that of other power sources (Cheng and Hammond, 2017)¹²⁸, the political processes triggered by societal concerns depend on the country-specific means of managing the political debates around technological choices and their environmental impacts (Gregory et al., 1993)¹²⁹. Such differences in perception explain why the 2011 Fukushima incident resulted in a confirmation or acceleration of phasing out nuclear energy in five countries (Roh, 2017)¹³⁰ while 30 other countries have continued using nuclear energy, amongst which 13 are building new nuclear capacity, including China, India and the United Kingdom (IAEA, 2017; Yuan et al., 2017)¹³¹.

Costs of nuclear power have increased over time in some developed nations, principally due to market conditions where increased investment risks of high-capital expenditure technologies have become significant. ‘Learning by doing’ processes often failed to compensate for this trend because they were slowed down by the absence of standardization and series effects (Grubler, 2010)¹³². What the costs of nuclear power are and have been is debated in the literature (Lovering et al., 2016; Koomey et al., 2017)¹³³. Countries with liberalized markets that continue to develop nuclear employ de-risking instruments through long-term contracts with guaranteed sale prices (Finon and Roques, 2013)¹³⁴. For instance, the United Kingdom works with public guarantees covering part of the upfront investment costs of newly planned nuclear capacity. This dynamic differs in countries such as China and South Korea, where monopolistic conditions in the electric system allow for reducing investment risks, deploying series effects and enhancing the engineering capacities of users due to stable relations between the security authorities and builders (Schneider et al., 2017)¹³⁵.

The safety of nuclear plants depends upon the public authorities of each country. However, because accidents affect worldwide public acceptance of this industry, questions have been raised about the risk of economic and political pressures weakening the safety of the plants (Finon, 2013; Budnitz, 2016)¹³⁶. This raises the issue of international governance of civil nuclear risks and reinforced international cooperation involving governments, companies and engineering (Walker and Lönnroth, 1983; Thomas, 1988; Finon, 2013)¹³⁷, based on the experience of the International Atomic Energy Agency.

The growth in electricity storage for renewables has been around grid flexibility resources (GFR) that would enable several places to source more than half their power from non-hydro renewables (Komarnicki, 2016)¹³⁸. Ten types of GFRs within smart grids have been developed (largely since AR5)(Blaabjerg et al., 2004; IRENA, 2013; IEA, 2017d; Majzoobi and Khodaei, 2017)¹³⁹, though how variable renewables can be balanced  without hydro or natural gas-based power back-up at a larger scale would still need demonstration. Pumped hydro comprised 150 GW of storage capacity in 2016, and grid-connected battery storage just 1.7 GW, but the latter grew between 2015 to 2016 by 50% (REN21, 2017)¹⁴⁰. Battery storage has been the main growth feature in energy storage since AR5 (Breyer et al., 2017)¹⁴¹. This appears to the result of significant cost reductions due to mass production for electric vehicles (EVs) (Nykvist and Nilsson, 2015; Dhar et al., 2017)¹⁴². Although costs and technical maturity look increasingly positive, the feasibility of battery storage is challenged by concerns over the availability of resources and the environmental impacts of its production (Peters et al., 2017)¹⁴³. Lithium, a common element in the earth’s crust, does not appear to be restricted and large increases in production have happened in recent years with eight new mines in Western Australia where most lithium is produced (GWA, 2016)¹⁴⁴. Emerging battery technologies may provide greater efficiency and recharge rates (Belmonte et al., 2016)¹⁴⁵ but remain significantly more expensive due to speed and scale issues compared to lithium ion batteries (Dhar et al., 2017; IRENA, 2017a)¹⁴⁶.

Research and demonstration of energy storage in the form of thermal and chemical systems continues, but large-scale commercial systems are rare (Pardo et al., 2014)¹⁴⁷. Renewably derived synthetic liquid (like methanol and ammonia) and gas (like methane and hydrogen) are increasingly being seen as a feasible storage options for renewable energy (producing fuel for use in industry during times when solar and wind are abundant) (Bruce et al., 2010; Jiang et al., 2010; Ezeji, 2017)¹⁴⁸ but, in the case of carbonaceous storage media, would need a renewable source of carbon to make a positive contribution to GHG reduction (von der Assen et al., 2013; Abanades et al., 2017)¹⁴⁹ (see also Section 4.3.4.5). The use of electric vehicles as a form of storage has been modelled and evaluated as an opportunity, and demonstrations are emerging (Dhar et al., 2017; Green and Newman, 2017a)¹⁵⁰, but challenges to upscaling remain.

Climate change has started to disrupt electricity generation and, if climate change adaptation options are not considered, it is predicted that these disruptions will be lengthier and more frequent (Jahandideh-Tehrani et al., 2014; Bartos and Chester, 2015; Kraucunas et al., 2015; van Vliet et al., 2016)¹⁵¹. Adaptation would both secure vulnerable infrastructure and ensure the necessary generation capacity (Minville et al., 2009; Eisenack and Stecker, 2012; Schaeffer et al., 2012; Cortekar and Groth, 2015; Murrant et al., 2015; Panteli and Mancarella, 2015; Goytia et al., 2016)¹⁵². The literature shows high agreement that climate change impacts need to be planned for in the design of any kind of infrastructure, especially in the energy sector (Nierop, 2014)¹⁵³, including interdependencies with other sectors that require electricity to function, including water, data, telecommunications and transport (Fryer, 2017)¹⁵⁴.

Recent research has developed new frameworks and models that aim to assess and identify vulnerabilities in energy infrastructure and create more proactive responses (Francis and Bekera, 2014; Ouyang and Dueñas-Osorio, 2014; Arab et al., 2015; Bekera and Francis, 2015; Knight et al., 2015; Jeong and An, 2016; Panteli et al., 2016; Perrier, 2016; Erker et al., 2017; Fu et al., 2017)¹⁵⁵. Assessments of energy infrastructure adaptation, while limited, emphasize the need for redundancy (Liu et al., 2017)¹⁵⁶. The implementation of  controllable and islandable microgrids, including the use of residential batteries, can increase resiliency, especially after extreme weather events (Qazi and Young Jr., 2014; Liu et al., 2017)¹⁵⁷. Hybrid renewables-based power systems with non-hydro capacity, such as with high-penetration wind generation, could provide the required system flexibility (Canales et al., 2015)¹⁵⁸. Overall, there is high agreement that hybrid systems, taking advantage of an array of sources and time of use strategies, can help make electricity generation more resilient (Parkinson and Djilali, 2015)¹⁵⁹, given that energy security standards are in place (Almeida Prado et al., 2016)¹⁶⁰.

Interactions between water and energy are complex (IEA, 2017g)¹⁶¹. Water scarcity patterns and electricity disruptions will differ across regions. There is high agreement that mitigation and adaptation options for thermal electricity generation (if that remains fitted with CCS) need to consider increasing water shortages, taking into account other factors such as ambient water resources and demand changes in irrigation water (Hayashi et al., 2018)¹⁶². Increasing the efficiency of power plants can reduce emissions and water needs (Eisenack and Stecker, 2012; van Vliet et al., 2016)¹⁶³, but applying CCS would increase water consumption (Koornneef et al., 2012)¹⁶⁴. The technological, economic, social and institutional feasibility of efficiency improvements is high, but insufficient to limit temperature rise to 1.5°C (van Vliet et al., 2016)¹⁶⁵.

In addition, a number of options for water cooling management systems have been proposed, such as hydraulic measures (Eisenack and Stecker, 2012)¹⁶⁶ and alternative cooling technologies (Chandel et al., 2011; Eisenack and Stecker, 2012; Bartos and Chester, 2015; Murrant et al., 2015; Bustamante et al., 2016; van Vliet et al., 2016; Huang et al., 2017b)¹⁶⁷. There is high agreement on the technological and economic feasibility of these technologies, as their absence can severely impact the functioning of the power plant as well as safety and security standards.

The AR5 (IPCC, 2014b)¹⁶⁸ as well as Chapter 2, Section 2.4.2, assign significant emission reductions over the course of this century to CO₂ capture and storage (CCS) in the power sector. This section focuses on CCS in the fossil-fuelled power sector; Section 4.3.4 discusses CCS in non-power industry, and Section 4.3.7 discusses bioenergy with CCS (BECCS). Section 2.4.2 puts the cumulative CO₂ stored from fossil-fuelled power at 410 (199–470 interquartile range) GtCO₂ over this century. Such modelling suggests that CCS in the power sector can contribute to cost-effective achievement of emission reduction requirements for limiting warming to 1.5°C. CCS may also offer employment and political advantages for fossil fuel-dependent economies (Kern et al., 2016)¹⁶⁹, but may entail more limited co-benefits than other mitigation options (that, e.g., generate power) and therefore relies on climate policy incentives for its business case and economic feasibility. Since 2017, two CCS projects in the power sector capture 2.4 MtCO₂ annually, while 30 MtCO₂ is captured annually in all CCS projects (Global CCS Institute, 2017)¹⁷⁰.

The technological maturity of CO₂ capture options in the power sectors has improved considerably (Abanades et al., 2015; Bui et al., 2018)¹⁷¹, but costs have not come down between 2005 and 2015 due to limited learning in commercial settings and increased energy and resources costs (Rubin et al., 2015)¹⁷². Storage capacity estimates vary greatly, but Section 2.4.2 as well as literature (V. Scott et al., 2015)¹⁷³ indicate that perhaps 10,000 GtCO₂ could be stored in underground reservoirs. Regional availability of this may not be sufficient, and it requires efforts to have this storage and the corresponding infrastructure available at the necessary rates and times (de Coninck and Benson, 2014)¹⁷⁴. CO₂ retention in the storage reservoir was recently assessed as 98% over 10,000 years for well-managed reservoirs, and 78% for poorly regulated ones (Alcalde et al., 2018)¹⁷⁵.  A paper reviewing 42 studies on public perception of CCS (Seigo et al., 2014)¹⁷⁶ found that social acceptance of CCS is predicted by trust, perceived risks and benefits. The technology itself mattered less than the social context of the project. Though insights on communication of CCS projects to the general public and inhabitants of the area around the CO₂ storage sites have been documented over the years, project stakeholders are not consistently implementing these lessons, although some projects have observed good practices (Ashworth et al., 2015)¹⁷⁷.

CCS in the power sector is hardly being realized at scale, mainly because the incremental costs of capture, and the development of transport and storage infrastructures are not sufficiently compensated by market or government incentives (IEA, 2017c)¹⁷⁸. In the two full-scale projects in the power sector mentioned above, part of the capture costs are compensated for by revenues from enhanced oil recovery (EOR) (Global CCS Institute, 2017)¹⁷⁹, demonstrating that EOR helps developing CCS further. EOR is a technique that uses CO₂ to mobilize more oil out of depleting oil fields, leading to additional CO₂ emissions by combusting the additionally recovered oil (Cooney et al., 2015)¹⁸⁰.

In a 1.5°C world, local yields are projected to decrease in tropical regions that are major food producing areas of the world (West Africa, Southeast Asia, South Asia, and Central and northern South America) (Schleussner et al., 2016)¹⁸¹. Some high-latitude regions may benefit from the combined effects of elevated CO₂ and temperature because their average temperatures are below optimal temperature for crops. In both cases there are consequences for food production and quality (Cross-Chapter Box 6 in Chapter 3 on Food Security), conservation agriculture, irrigation, food wastage, bioenergy and the use of novel technologies.

Food production and quality. Increased temperatures, including 1.5°C warming, would affect the production of cereals such as wheat and rice, impacting food security (Schleussner et al., 2016)¹⁸². There is medium agreement that elevated CO₂ concentrations can change food composition, with implications for nutritional security (Taub et al., 2008; Högy et al., 2009; DaMatta et al., 2010; Loladze, 2014; De Souza et al., 2015)¹⁸³, with the effects being different depending on the region (Medek et al., 2017)¹⁸⁴.

Meta-analyses of the effects of drought, elevated CO₂, and temperature conclude that at 2°C local warming and above, aggregate production of wheat, maize, and rice are expected to decrease in both temperate and tropical areas (Challinor et al., 2014)¹⁸⁵. These production losses could be lowered if adaptation measures are taken (Challinor et al., 2014)¹⁸⁶, such as developing varieties better adapted to changing climate conditions.

Adaptation options can help ensure access to sufficient, quality food. Such options include conservation agriculture, improved livestock management, increasing irrigation efficiency, agroforestry and management of food loss and waste. Complementary adaptation and mitigation options, for example, the use of climate services (Section 4.3.5), bioenergy (Section 4.3.1) and biotechnology (Section 4.4.4) can also serve to reduce emissions intensity and the carbon footprint of food production.

Conservation agriculture (CA) is a soil management approach that reduces the disruption of soil structure and biotic processes by minimising tillage. A recent meta-analysis showed that no-till practices work well in water-limited agroecosystems when implemented jointly with residue retention and crop rotation, but when used independently, may decrease yields in other situations (Pittelkow et al., 2014)¹⁸⁷. Additional climate adaptations include adjusting planting times and crop varietal selection and improving irrigation efficiency. Adaptations such as these may increase wheat and maize yields by 7–12% under climate change (Challinor et al., 2014)¹⁸⁸. CA can also help build adaptive capacity (medium evidence, medium agreement) (H. Smith et al., 2017; Pradhan et al., 2018)¹⁸⁹ and have mitigation co-benefits through improved fertiliser use or efficient use of machinery and fossil fuels (Harvey et al., 2014; Cui et al., 2018; Pradhan et al., 2018)¹⁹⁰. CA practices can also raise soil carbon and therefore remove CO₂ from the atmosphere (Aguilera et al., 2013; Poeplau and Don, 2015; Vicente-Vicente et al., 2016)¹⁹¹. However, CA adoption can be constrained by inadequate institutional arrangements and funding mechanisms (Harvey et al., 2014; Baudron et al., 2015; Li et al., 2016; Dougill et al., 2017; H. Smith et al., 2017)¹⁹².

Sustainable intensification of agriculture consists of agricultural systems with increased production per unit area but with management of the range of potentially adverse impacts on the environment (Pretty and Bharucha, 2014)¹⁹³. Sustainable intensification can increase the efficiency of inputs and enhance health and food security (Ramankutty et al., 2018)¹⁹⁴.

Livestock management. Livestock are responsible for more GHG emissions than all other food sources. Emissions are caused by feed production, enteric fermentation, animal waste, land-use change and livestock transport and processing. Some estimates indicate that livestock supply chains could account for 7.1 GtCO₂ per year, equivalent to 14.5% of global anthropogenic greenhouse gas emissions (Gerber et al., 2013)¹⁹⁵. Cattle (beef, milk) are responsible for about two-thirds of that total, largely due to methane emissions resulting from rumen fermentation (Gerber et al., 2013; Opio et al., 2013)¹⁹⁶.

Despite ongoing gains in livestock productivity and volumes, the increase of animal products in global diets is restricting overall agricultural efficiency gains because of inefficiencies in the conversion of agricultural primary production (e.g., crops) in the feed-animal products pathway (Alexander et al., 2017)¹⁹⁷, offsetting the benefits of improvements in livestock production systems (Clark and Tilman, 2017)¹⁹⁸.

There is increasing agreement that overall emissions from food systems could be reduced by targeting the demand for meat and other livestock products, particularly where consumption is higher than suggested by human health guidelines. Adjusting diets to meet nutritional targets could bring large co-benefits, through GHG mitigation and improvements in the overall efficiency of food systems (Erb et al., 2009; Tukker et al., 2011; Tilman and Clark, 2014; van Dooren et al., 2014; Ranganathan et al., 2016)¹⁹⁹. Dietary shifts could contribute one-fifth of the mitigation needed to hold warming below 2°C, with one-quarter of low-cost options (Griscom et al., 2017)²⁰⁰. There, however, remains limited evidence of effective policy interventions to achieve such large-scale shifts in dietary choices, and prevailing trends are for increasing rather than decreasing demand for livestock products at the global scale (Alexandratos and Bruinsma, 2012; OECD/FAO, 2017)²⁰¹. How the role of dietary shift could change in 1.5°C-consistent pathways is also not clear (see Chapter 2).

Adaptation of livestock systems can include a suite of strategies such as using different breeds and their wild relatives to develop a genetic pool resilient to climatic shocks and longer-term temperature shifts (Thornton and Herrero, 2014)²⁰², improving fodder and feed management (Bell et al., 2014; Havet et al., 2014)²⁰³ and disease prevention and control (Skuce et al., 2013; Nguyen et al., 2016)²⁰⁴. Most interventions that improve the productivity of livestock systems and enhance adaptation to climate changes would also reduce the emissions intensity of food production, with significant co-benefits for rural livelihoods and the security of food supplies (Gerber et al., 2013; FAO and NZAGRC, 2017a, b, c)²⁰⁵. Whether such reductions in emission intensity result in lower or higher absolute GHG emissions depends on overall demand for livestock products, indicating the relevance of integrating supply-side with demand-side measures within food security objectives (Gerber et al., 2013; Bajželj et al., 2014)²⁰⁶. Transitions in livestock production systems (e.g., from extensive to intensive) can also result in significant emission reductions as part of broader land-based mitigation strategies (Havlik et al., 2014)²⁰⁷.

Overall, there is high agreement that farm strategies that integrate mixed crop–livestock systems can improve farm productivity and have positive sustainability outcomes (Havet et al., 2014; Thornton and Herrero, 2014; Herrero et al., 2015; Weindl et al., 2015)²⁰⁸. Shifting towards mixed crop–livestock systems is estimated to reduce agricultural adaptation costs to 0.3% of total production costs while abating deforestation by 76 Mha globally, making it a highly cost-effective adaptation option with mitigation co-benefits (Weindl et al., 2015)²⁰⁹. Evidence from various regions supports this (Thornton and Herrero, 2015)²¹⁰, although the feasible scale varies between regions and systems, as well as being moderated by overall demand in specific food products. In Australia, some farmers have successfully shifted to crop–livestock systems where, each year, they allocate land and forage resources in response to climate and price trends (Bell et al., 2014)²¹¹. However, there can be some unintended negative impacts of such integration, including increased burdens on women, higher requirements of capital, competing uses of crop residues (e.g., feed vs. mulching vs. carbon sequestration) and higher requirements of management skills, which can be a challenge across several low income countries (Thornton and Herrero, 2015; Thornton et al., 2018)²¹². Finally, the feasibility of improving livestock efficiency is dependent on socio-cultural context and acceptability: there remain significant issues around widespread adoption of crossbred animals, especially by smallholders (Thornton et al., 2018)²¹³.

Irrigation efficiency. Irrigation efficiency is especially critical since water endowments are expected to change, with  20–60 Mha of global cropland being projected to revert from irrigated to rain-fed land, while other areas will receive higher precipitation in shorter time spans, thus affecting irrigation demand (Elliott et al., 2014)²¹⁴. While increasing irrigation system efficiency is necessary, there is mixed evidence on how to enact efficiency improvements (Fader et al., 2016; Herwehe and Scott, 2018)²¹⁵. Physical and technical strategies include building large-scale reservoirs or dams, renovating or deepening irrigation channels, building on-farm rainwater harvesting structures, lining ponds, channels and tanks to reduce losses through percolation and evaporation, and investing in small infrastructure such as sprinkler or drip irrigation sets (Varela-Ortega et al., 2016; Sikka et al., 2018)²¹⁶. Each strategy has differing costs and benefits relating to unique biophysical, social, and economic contexts. Also, increasing irrigation efficiency may foster higher dependency on irrigation, resulting in a heightened sensitivity to climate that may be maladaptive in the long term (Lindoso et al., 2014)²¹⁷.

Improvements in irrigation efficiency would need to be supplemented with ancillary activities, such as shifting to crops that require less water and improving soil and moisture conservation (Fader et al., 2016; Hong and Yabe, 2017; Sikka et al., 2018)²¹⁸. Currently, the feasibility of improving irrigation efficiency is constrained by issues of replicability across scale and sustainability over time (Burney and Naylor, 2012)²¹⁹, institutional barriers and inadequate market linkages (Pittock et al., 2017)²²⁰.

Growing evidence suggests that investing in behavioural shifts towards using irrigation technology such as micro-sprinklers or drip irrigation, is an effective and quick adaptation strategy (Varela-Ortega et al., 2016; Herwehe and Scott, 2018; Sikka et al., 2018)²²¹ as opposed to large dams which have high financial, ecological and social costs (Varela-Ortega et al., 2016)²²². While improving irrigation efficiency is technically feasible (R. Fishman et al., 2015)²²³ and has clear benefits for environmental values (Pfeiffer and Lin, 2014; R. Fishman et al., 2015)²²⁴, feasibility is regionally differentiated as shown by examples as diverse as Kansas (Jägermeyr et al., 2015)²²⁵, India (R. Fishman et al., 2015)²²⁶ and Africa (Pittock et al., 2017)²²⁷.

Agroforestry. The integration of trees and shrubs into crop and livestock systems, when properly managed, can potentially restrict soil erosion, facilitate water infiltration, improve soil physical properties and buffer against extreme events (Lasco et al., 2014; Mbow et al., 2014; Quandt et al., 2017; Sida et al., 2018)²²⁸. There is medium evidence and high agreement on the feasibility of agroforestry practices that enhance productivity, livelihoods and carbon storage (Lusiana et al., 2012; Murthy, 2013; Coulibaly et al., 2017; Sida et al., 2018)²²⁹, including from indigenous production systems (Coq-Huelva et al., 2017)²³⁰, with variation by region, agroforestry type, and climatic conditions (Place et al., 2012; Coe et al., 2014; Mbow et al., 2014; Iiyama et al., 2017; Abdulai et al., 2018)²³¹. Long-term studies examining the success of agroforestry, however, are rare (Coe et al., 2014; Meijer et al., 2015; Brockington et al., 2016; Zomer et al., 2016)²³².

The extent to which agroforestry practices employed at the farm level could be scaled up globally while satisfying growing food demand is relatively unknown. Agroforestry adoption has been relatively low and uneven (Jacobi et al., 2017; Hernández-Morcillo et al., 2018)²³³, with constraints including the expense of establishment and lack of reliable financial support, insecure land tenure, landowner’s lack of experience with trees, complexity of management practices, fluctuating market demand and prices for different food and fibre products, the time and knowledge required for management, low intermediate benefits to offset revenue lags, and inadequate market access (Pattanayak et al., 2003; Mercer, 2004; Sendzimir et al., 2011; Valdivia et al., 2012; Coe et al., 2014; Meijer et al., 2015; Coulibaly et al., 2017; Jacobi et al., 2017)²³⁴.

Managing food loss and waste. The way food is produced, processed and transported strongly influences GHG emissions. Around one-third of the food produced on the planet is not consumed (FAO, 2013)²³⁵, affecting food security and livelihoods (See Cross-Chapter Box 6 on Food Security in Chapter 3). Food wastage is a combination of food loss – the decrease in mass and nutritional value of food due to poor infrastructure, logistics, and lack of storage technologies and management – and food waste that derives from inappropriate human consumption that leads to food spoilage associated with inferior quality or overproduction. Food wastage could lead to an increase in emissions estimated to 1.9–2.5 GtCO₂-eq yr⁻¹ (Hiç et al., 2016)²³⁶.

Decreasing food wastage has high mitigation and adaptation potential and could play an important role in land transitions towards 1.5°C, provided that reduced food waste results in lower production-side emissions rather than increased consumption (Foley et al., 2011)²³⁷. There is medium agreement that a combination of individual–institutional behaviour (Refsgaard and Magnussen, 2009; Thornton and Herrero, 2014)²³⁸, and improved technologies and management (Lin et al., 2013; Papargyropoulou et al., 2014)²³⁹ can transform food waste into products with marketable value. Institutional behaviour depends on investment and policies, which if adequately addressed could enable mitigation and adaptation co-benefits in a relatively short time.

Novel technologies. New molecular biology tools have been developed that can lead to fast and precise genome modification (De Souza et al., 2016; Scheben et al., 2016)²⁴⁰ (e.g., CRISPR Cas9; Ran et al., 2013; Schaeffer and Nakata, 2015)²⁴¹. Such genome editing tools may moderately assist in mitigation and adaptation of agriculture in relation to climate changes, elevated CO₂, drought and flooding (DaMatta et al., 2010; De Souza et al., 2015, 2016)²⁴². These tools could contribute to developing new plant varieties that can adapt to warming of 1.5°C and overshoot, potentially avoiding some of the costs of crop shifting (Schlenker and Roberts, 2009; De Souza et al., 2016)²⁴³. However, biosafety concerns and government regulatory systems can be a major barrier to the use of these tools as this increases the time and cost of turning scientific discoveries into ready applicable technologies (Andow and Zwahlen, 2006; Maghari and Ardekani, 2011)²⁴⁴.

The strategy of reducing enteric methane emissions by ruminants through the development of inhibitors or vaccines has already been attempted with some successes, although the potential for application at scale and in different situations remains uncertain. A methane inhibitor has been demonstrated to reduce methane from feedlot systems by 30% over a 12-week period (Hristov et al., 2015)²⁴⁵ with some productivity benefits, but the ability to apply it in grazing systems will depend on further technological developments as well as costs and incentives. A vaccine could potentially modify the microbiota of the rumen and be applicable even in extensive grazing systems by reducing the presence of methanogenic micro-organisms (Wedlock et al., 2013)²⁴⁶ but has not yet been successfully demonstrated to reduce emissions in live animals. Selective breeding for lower-emitting ruminants is becoming rapidly feasible, offering small but cumulative emissions reductions without requiring substantial changes in farm systems (Pickering et al., 2015)²⁴⁷.

Technological innovation in culturing marine and freshwater micro and macro flora has significant potential to expand food, fuel and fibre resources, and could reduce impacts on land and conventional agriculture (Greene et al., 2017)²⁴⁸.

Technological innovation could assist in increased agricultural efficiency (e.g., via precision agriculture), decrease food wastage and genetics that enhance plant adaptation traits (Section 4.4.4). Technological and associated management improvements may be ways to increase the efficiency of contemporary agriculture to help produce enough food to cope with population increases in a 1.5°C warmer world, and help reduce the pressure on natural ecosystems and biodiversity.

Ecosystem restoration. Biomass stocks in tropical, subtropical, temperate and boreal biomes currently hold 1085, 194, 176, 190 Gt CO₂, respectively. Conservation and restoration can enhance these natural carbon sinks (Erb et al., 2017)²⁴⁹.

Recent studies explore options for conservation, restoration and improved land management estimating up to 23 GtCO₂ (Griscom et al., 2017)²⁵⁰. Mitigation potentials are dominated by reduced rates of deforestation, reforestation and forest management, and concentrated in tropical regions (Houghton, 2013; Canadell and Schulze, 2014; Grace et al., 2014; Houghton et al., 2015; Griscom et al., 2017)²⁵¹. Much of the literature focuses on REDD+ (reducing emissions from deforestation and forest degradation) as an institutional mechanism. However, restoration and management activities need not be limited to REDD+, and locally adapted implementation may keep costs low, capitalize on co-benefits and ensure consideration of competing for socio-economic goals (Jantke et al., 2016; Ellison et al., 2017; Perugini et al., 2017; Spencer et al., 2017)²⁵².

Half of the estimated potential can be achieved at <100 USD/tCO₂; and a third of the cost-effective potential at <10 USD/tCO₂ (Griscom et al., 2017)²⁵³. Variation of costs in projects aiming to reduce emissions from deforestation is high when considering opportunity and transaction costs (Dang Phan et al., 2014; Overmars et al., 2014; Ickowitz et al., 2017; Rakatama et al., 2017)²⁵⁴.

However, the focus on forests raises concerns of cross-biome leakage (medium evidence, low agreement) (Popp et al., 2014a; Strassburg et al., 2014; Jayachandran et al., 2017)²⁵⁵ and encroachment on other ecosystems (Veldman et al., 2015)²⁵⁶. Reducing rates of deforestation constrains the land available for agriculture and grazing, with trade-offs between diets, higher yields and food prices (Erb et al., 2016a; Kreidenweis et al., 2016)²⁵⁷. Forest restoration and conservation are compatible with biodiversity (Rey Benayas et al., 2009; Jantke et al., 2016)²⁵⁸ and available water resources; in the tropics, reducing rates of deforestation maintains cooler surface temperatures (Perugini et al., 2017)²⁵⁹ and rainfall (Ellison et al., 2017)²⁶⁰.

Its multiple potential co-benefits have made REDD+ important for local communities, biodiversity and sustainable landscapes (Ngendakumana et al., 2017; Turnhout et al., 2017)²⁶¹. There is low agreement on whether climate impacts will reverse mitigation benefits of restoration (Le Page et al., 2013)²⁶² by increasing the likelihood of disturbance (Anderegg et al., 2015)²⁶³, or reinforce them through carbon fertilization (P. Smith et al., 2014)²⁶⁴.

Emerging regional assessments offer new perspectives for upscaling. Strengthening coordination, additional funding sources, and access and disbursement points increase the potential of REDD+ in working towards 2°C and 1.5°C limits (Well and Carrapatoso, 2017)²⁶⁵. While there are indications that land tenure has a positive impact (Sunderlin et al., 2014)²⁶⁶, a meta-analysis by Wehkamp et al. (2018a)²⁶⁷ shows that there is medium evidence and low agreement on which aspects of governance improvements are supportive of conservation. Local benefits, especially for indigenous communities, will only be accrued if land tenure is respected and legally protected, which is not often the case (Sunderlin et al., 2014; Brugnach et al., 2017)²⁶⁸. Although payments for reduced rates of deforestation may benefit the poor, the most vulnerable populations could have limited, uneven access (Atela et al., 2014)²⁶⁹ and face lower opportunity costs from deforestation (Ickowitz et al., 2017)²⁷⁰.

Community-based adaptation (CbA). There is medium evidence and high agreement for the use of CbA. The specific actions to take will depend upon the location, context, and vulnerability of the specific community. CbA is defined as ‘a community-led process, based on communities’ priorities, needs, knowledge, and capacities, which aim to empower people to plan for and cope with the impacts of climate change’ (Reid et al., 2009)²⁷¹. The integration of CbA with ecosystems-based adaptation (EbA) has been increasingly promoted, especially in efforts to alleviate poverty (Mannke, 2011; Reid, 2016)²⁷².

Despite the potential and advantages of both CbA and EbA, including knowledge exchange, information access and increased social capital and equity; institutional and governance barriers still constitute a challenge for local adaptation efforts (Wright et al., 2014; Fernández-Giménez et al., 2015)²⁷³.

Wetland management. In wetland ecosystems, temperature rise has direct and irreversible impacts on species functioning and distribution, ecosystem equilibrium and services, and second-order impacts on local livelihoods (see Chapter 3, Section 3.4.3). The structure and function of wetland systems are changing due to climate change. Wetland management strategies, including adjustments in infrastructural, behavioural, and institutional practices have clear implications for adaptation (Colloff et al., 2016b; Finlayson et al., 2017; Wigand et al., 2017)²⁷⁴

Despite international initiatives on wetland restoration and management through the Ramsar Convention on Wetlands, policies have not been effective (Finlayson, 2012; Finlayson et al., 2017)²⁷⁵. Institutional reform, such as flexible, locally relevant governance, drawing on principles of adaptive co-management, and multi-stakeholder participation becomes increasingly necessary for effective wetland management (Capon et al., 2013; Finlayson et al., 2017)²⁷⁶.

Managing coastal stress. Particularly to allow for the landward relocation of coastal ecosystems under a transition to a 1.5°C warmer world, planning for climate change would need to be integrated with the use of coastlines by humans (Saunders et al., 2014; Kelleway et al., 2017)²⁷⁷. Adaptation options for managing coastal stress include coastal hardening through the building of seawalls and the re-establishment of coastal ecosystems such as mangroves (André et al., 2016; Cooper et al., 2016)²⁷⁸. While the feasibility of the solutions is high, they are expensive to scale (robust evidence, medium agreement).

There is low evidence and high agreement that reducing the impact of local stresses (Halpern et al., 2015)²⁷⁹ will improve the resilience of marine ecosystems as they transition to a 1.5°C world (O’Leary et al., 2017)²⁸⁰.  Approaches to reducing local stresses are considered feasible, cost-effective and highly scalable. Ecosystem resilience may be increased through alternative livelihoods (e.g., sustainable aquaculture), which are among a suite of options for building resilience in coastal ecosystems. These options enjoy high levels of feasibility yet are expensive, which stands in the way of scalability (robust evidence, medium agreement) (Hiwasaki et al., 2015; Brugnach et al., 2017)²⁸¹.

Working with coastal communities has the potential for improving the resilience of coastal ecosystems. Combined with the advantages of using indigenous knowledge to guide transitions, solutions can be more effective when undertaken in partnership with local communities, cultures, and knowledge (See Box 4.3).

Restoration of coastal ecosystems and fisheries. Marine restoration is expensive compared to terrestrial restoration, and the survival of projects is currently low, with success depending on the ecosystem and site, rather than the size of the financial investment (Bayraktarov et al., 2016)²⁸². Mangrove replanting shows evidence of success globally, with numerous examples of projects that have established forests (Kimball et al., 2015; Bayraktarov et al., 2016)²⁸³.

Efforts with reef-building corals have been attempted with a low level of success (Bayraktarov et al., 2016)²⁸⁴. Technologies to help re-establish coral communities are limited (Rinkevich, 2014)²⁸⁵, as are largely untested disruptive technologies (e.g., genetic manipulation, assisted evolution) (van Oppen et al., 2015)²⁸⁶. Current technologies also have trouble scaling given the substantial costs and investment required (Bayraktarov et al., 2016)²⁸⁷.

Johannessen and Macdonald (2016)²⁸⁸ report the ‘blue carbon’ sink to be 0.4–0.8% of global anthropogenic emissions. However, this does not adequately account for post-depositional processes and could overestimate removal potentials, subject to a risk of reversal. Seagrass beds will thus not contribute significantly to enabling 1.5°C-consistent pathways.

Urban economies tend to be more energy intensive than national economies due to higher levels of per capita income, mobility and consumption (Kennedy et al., 2015; Broto, 2017; Gota et al., 2018)²⁹⁸. However, some urban systems have begun decoupling development from the consumption of fossil fuel-powered energy through energy efficiency, renewable energy and locally managed smart grids (Dodman, 2009; Freire et al., 2014; Eyre et al., 2018; Glazebrook and Newman, 2018)²⁹⁹.

The rapidly expanding cities of Africa and Asia, where energy poverty currently undermines adaptive capacity (Westphal et al., 2017; Satterthwaite et al., 2018)³⁰⁰, have the opportunity to benefit from recent price changes in renewable energy technologies to enable clean energy access to citizens (SDG 7) (Cartwright, 2015; Watkins, 2015; Lwasa, 2017; Kennedy et al., 2018; Teferi and Newman, 2018)³⁰¹. This will require strengthened energy governance in these countries (Eberhard et al., 2017)³⁰². Where renewable energy displaces paraffin, wood fuel or charcoal feedstocks in informal urban settlements, it provides the co-benefits of improved indoor air quality, reduced fire risk and reduced deforestation, all of which can enhance adaptive capacity and strengthen demand for this energy (Newham and Conradie, 2013; Winkler, 2017; Kennedy et al., 2018; Teferi and Newman, 2018)³⁰³.

Buildings are responsible for 32% of global energy consumption (IEA, 2016c)³⁰⁴ and have a large energy saving potential with available and demonstrated technologies such as energy efficiency improvements in technical installations and in thermal insulation (Toleikyte et al., 2018)³⁰⁵ and energy sufficiency (Thomas et al., 2017)³⁰⁶. Kuramochi et al. (2018)³⁰⁷ show that 1.5°C-consistent pathways require building emissions to be reduced by 80–90% by 2050, new construction to be fossil-free and near-zero energy by 2020, and an increased rate of energy refurbishment of existing buildings to 5% per annum in OECD (Organisation for Economic Co-operation and Development) countries (see also Section 4.2.1).

Based on the IEA-ETP (IEA, 2017g)³⁰⁸, Chapter 2 identifies large saving potential in heating and cooling through improved building design, efficient equipment, lighting and appliances. Several examples of net zero energy in buildings are now available (Wells et al., 2018)³⁰⁹. In existing buildings, refurbishment enables energy saving (Semprini et al., 2017; Brambilla et al., 2018; D’Agostino and Parker, 2018; Sun et al., 2018)³¹⁰ and cost savings (Toleikyte et al., 2018; Zangheri et al., 2018)³¹¹.

Reducing the energy embodied in building materials provides further energy and GHG savings (Cabeza et al., 2013; Oliver and Morecroft, 2014; Koezjakov et al., 2018)³¹², in particular through increased use of bio-based materials (Lupíšek et al., 2015)³¹³ and wood construction (Ramage et al., 2017)³¹⁴. The United Nations Environment Programme (UNEP3)³ estimates that improving embodied energy, thermal performance, and direct energy use of buildings can reduce emissions by 1.9 GtCO₂e yr ⁻¹ (UNEP, 2017b)³¹⁵, with an additional reduction of 3 GtCO₂e yr⁻¹ through energy efficient appliances and lighting (UNEP, 2017b)³¹⁶. Further increasing the energy efficiency of appliances and lighting, heating and cooling offers the potential for further savings (Parikh and Parikh, 2016; Garg et al., 2017)³¹⁷.

Smart technology, drawing on the internet of things (IoT) and building information modelling, offers opportunities to accelerate energy efficiency in buildings and cities (Moreno-Cruz and Keith, 2013; Hoy, 2016)³¹⁸ (see also Section 4.4.4). Some cities in developing countries are drawing on these technologies to adopt ‘leapfrog’ infrastructure, buildings and appliances to pursue low-carbon development (Newman et al., 2017; Teferi and Newman, 2017)³¹⁹ (Cross-Chapter Box 13 in Chapter 5).

Urban form impacts demand for energy (Sims et al., 2014)³²⁰ and other welfare related factors: a meta-analysis of 300 papers reported energy savings of 26 USD per person per year attributable to a 10% increase in urban population density (Ahlfeldt and Pietrostefani, 2017)³²¹. Significant reductions in car use are associated with dense, pedestrianized cities and towns and medium-density transit corridors (Newman and Kenworthy, 2015; Newman et al., 2017)³²² relative to low-density cities in which car dependency is high (Schiller and Kenworthy, 2018)³²³. Combined dense urban forms and new mass transit systems in Shanghai and Beijing have yielded less car use (Gao and Newman, 2018)³²⁴ (see Box 4.9). Compact cities also create the passenger density required to make public transport more financially viable (Rode et al., 2014; Ahlfeldt and Pietrostefani, 2017)³²⁵ and enable combinations of cleaner fuel feedstocks and urban smart grids, in which vehicles form part of the storage capacity (Oldenbroek et al., 2017)³²⁶. Similarly, the spatial organization of urban energy influenced the trajectories of urban development in cities as diverse as Hong Kong, Bengaluru and Maputo (Broto, 2017)³²⁷.

The informal settlements of middle- and low-income cities, where urban density is more typically associated with a range of water- and vector-borne health risks, may provide a notable exception to the adaptive advantages of urban density (Mitlin and Satterthwaite, 2013; Lilford et al., 2017)³²⁸ unless new approaches and technologies are harnessed to accelerate slum upgrading (Teferi and Newman, 2017)³²⁹.

Scenarios consistent with 1.5°C depend on a roughly 15% reduction in final energy use by the transport sector by 2050 relative to 2015 (Chapter 2, Figure 2.12). In one analysis the phasing out of fossil fuel passenger vehicle sales by 2035–2050 was identified as a benchmark for aligning with 1.5°C-consistent pathways (Kuramochi et al., 2018)³³⁰. Reducing emissions from transport has lagged the power sector (Sims et al., 2014; Creutzig et al., 2015a)³³¹, but evidence since AR5 suggests that cities are urbanizing and re-urbanizing in ways that coordinate transport sector adaptation and mitigation (Colenbrander et al., 2017; Newman et al., 2017; Salvo et al., 2017; Gota et al., 2018)³³². The global transport sector could reduce 4.7 GtCO2e yr⁻¹ (4.1–5.3) by 2030. This is significantly more than is predicted by integrated assessment models (UNEP, 2017b)³³³. Such a transition depends on cities that enable modal shifts and avoided journeys and that provide incentives for uptake of improved fuel efficiency and changes in urban design that encourage walkable cities, non-motorized transport and shorter commuter distances (IEA, 2016a; Mittal et al., 2016; Zhang et al., 2016; Li and Loo, 2017)³³⁴. In at least 4 African cities, 43 Asian cities and 54 Latin American cities, transit-oriented development (TOD), has emerged as an organizing principle for urban growth and spatial planning (Colenbrander et al., 2017; Lwasa, 2017; BRTData, 2018)³³⁵. This trend is important to counter the rising demand for private cars in developing-country cities (AfDB/OECD/UNDP, 2016)³³⁶. In India, TOD has been combined with localized solar PV installations and new ways of financing rail expansion (Sharma, 2018)³³⁷.

Cities pursuing sustainable transport benefit from reduced air pollution, congestion and road fatalities and are able to harness the relationship between transport systems, urban form, urban energy intensity and social cohesion (Goodwin and Van Dender, 2013; Newman and Kenworthy, 2015; Wee, 2015)³³⁸.

Technology and electrification trends since AR5 make carbon-efficient urban transport easier (Newman et al., 2016)³³⁹, but realizing urban transport’s contribution to a 1.5°C-consistent pathways will require the type of governance that can overcome the financial, institutional, behavioural and legal barriers to change (Geels, 2014; Bakker et al., 2017)³⁴⁰.

Adaptation to a 1.5°C world is enabled by urban design and spatial planning policies that consider extreme weather conditions and reduce displacement by climate related disasters (UNISDR, 2009; UN-Habitat, 2011; Mitlin and Satterthwaite, 2013)³⁴¹.

Building codes and technology standards for public lighting, including traffic lights (Beccali et al., 2015)³⁴², play a critical role in reducing carbon emissions, enhancing urban climate resilience and managing climate risk (Steenhof and Sparling, 2011; Parnell, 2015; Shapiro, 2016; Evans et al., 2017)³⁴³. Building codes can support the convergence to zero emissions from buildings (Wells et al., 2018)³⁴⁴ and can be used retrofit the existing building stock for energy efficiency (Ruparathna et al., 2016)³⁴⁵.

The application of building codes and standards for 1.5°C-consistent pathways will require improved enforcement, which can be a challenge in developing countries where inspection resources are often limited and codes are poorly tailored to local conditions (Ford et al., 2015c; Chandel et al., 2016; Eisenberg, 2016; Shapiro, 2016; Hess and Kelman, 2017; Mavhura et al., 2017)³⁴⁶. In all countries, building codes can be undermined by industry interests and can be maladaptive if they prevent buildings or land use from evolving to reduce climate impacts (Eisenberg, 2016; Shapiro, 2016)³⁴⁷.

The deficit in building codes and standards in middle-income and developing-country cities need not be a constraint to more energy-efficient and resilient buildings (Tait and Euston-Brown, 2017)³⁴⁸. For example, the relatively high price that poor households pay for unreliable and at times dangerous household energy in African cities has driven the uptake of renewable energy and energy efficiency technologies in the absence of regulations or fiscal incentives (Eberhard et al., 2011, 2016; Cartwright, 2015; Watkins, 2015)³⁴⁹. The Kuyasa Housing Project in Khayelitsha, one of Cape Town’s poorest suburbs, created significant mitigation and adaptation benefits by installing ceilings, solar water heaters and energy-efficient lightbulbs in houses independent of the formal housing or electrification programme (Winkler, 2017)³⁵⁰.

The electrification of urban systems, including transport, has shown global progress since AR5 (IEA, 2016a; Kennedy et al., 2018; Schiller and Kenworthy, 2018)³⁵¹. High growth rates are now appearing in electric vehicles (Figure 4.1), electric bikes and electric transit (IEA, 2018)³⁵², which would need to displace fossil fuel-powered passenger vehicles by 2035–2050 to remain in line with 1.5°C-consistent pathways. China’s 2017 Road Map calls for 20% of new vehicle sales to be electric. India is aiming for exclusively electric vehicles (EVs) by 2032 (NITI Aayog and RMI, 2017)³⁵³. Globally, EV sales were up 42% in 2016 relative to 2015, and in the United States EV sales were up 36% over the same period (Johnson and Walker, 2016)³⁵⁴.

The extent of electric railways in and between cities has expanded since AR5 (IEA, 2016a; Mittal et al., 2016; Zhang et al., 2016; Li and Loo, 2017)³⁵⁶. In high-income cities there is medium evidence for the decoupling of car use and wealth since AR5 (Newman, 2017)³⁵⁷. In cities where private vehicle ownership is expected to increase, less carbon-intensive fuel sources and reduced car journeys will be necessary as well as electrification of all modes of transport (Mittal et al., 2016; van Vuuren et al., 2017)³⁵⁸. Some recent urban data show a decoupling of urban growth and GHG emissions (Newman and Kenworthy, 2015)³⁵⁹ and that ‘peak car’ has been reached in Shanghai and Beijing (Gao and Kenworthy, 2017)³⁶⁰ and beyond (Manville et al., 2017)³⁶¹ (also see Box 4.9).

An estimated 800 cities globally have operational bike-share schemes (E. Fishman et al., 2015)³⁶², and China had 250 million electric bicycles in 2017 (Newman et al., 2017)³⁶³. Advances in information and communication technologies (ICT) offer cities the chance to reduce urban transport congestion and fuel consumption by making better use of the urban vehicle fleet through car sharing, driverless cars and coordinated public transport, especially when electrified (Wee, 2015; Glazebrook and Newman, 2018)³⁶⁴. Advances in ‘big-data’ can assist in creating a better understanding of the connections between cities, green infrastructure, environmental services and health (Jennings et al., 2016)³⁶⁵ and improve decision-making in urban development (Lin et al., 2017)³⁶⁶.

International transport hubs, including airports and ports, and the associated mobility of people are major economic contributors to most large cities even while under the governance of national authorities and international legislation. Shipping, freight and aviation systems have grown rapidly, and little progress has been made since AR5 on replacing fossil fuels, though some trials are continuing (Zhang, 2016; Bouman et al., 2017; EEA, 2017)³⁶⁷. Aviation emissions do not yet feature in IAMs (Bows-Larkin, 2015)³⁶⁸, but could be reduced by between a third and two-thirds through energy efficiency measures and operational changes (Dahlmann et al., 2016)³⁶⁹. On shorter intercity trips, aviation could be replaced by high-speed electric trains drawing on renewable energy (Åkerman, 2011)³⁷⁰. Some progress has been made on the use of electricity in planes and shipping (Grewe et al., 2017)³⁷¹ though no commercial applications have arisen. Studies indicate that biofuels are the most viable means of decarbonizing intercontinental travel, given their technical characteristics, energy content and affordability (Wise et al., 2017)³⁷². The lifecycle emissions of bio-based jet fuels and marine fuels can be considerable (Cox et al., 2014; IEA, 2017g)³⁷³ depending on their location (Elshout et al., 2014)³⁷⁴, but can be reduced by feedstock and conversion technology choices (de Jong et al., 2017)³⁷⁵.

In recent years the potential for transport to use synfuels, such as ethanol, methanol, methane, ammonia and hydrogen, created from renewable electricity and CO₂, has gained momentum but has not yet demonstrated benefits on a scale consistent with 1.5°C pathways (Ezeji, 2017; Fasihi et al., 2017)³⁷⁶. Decarbonizing the fuel used by the world’s 60,000 large ocean vessels faces governance barriers and the need for a global policy (Bows and Smith, 2012; IRENA, 2015; Rehmatulla and Smith, 2015)³⁷⁷. Low-emission marine fuels could simultaneously address sulphur and black carbon issues in ports and around waterways and accelerate the electrification of all large ports (Bouman et al., 2017; IEA, 2017g)³⁷⁸.

Urban land use influences energy intensity, risk exposure and adaptive capacity (Carter et al., 2015; Araos et al., 2016a; Ewing et al., 2016; Newman et al., 2016; Broto, 2017)³⁷⁹. Accordingly, urban land-use planning can contribute to climate mitigation and adaptation (Parnell, 2015; Francesch-Huidobro et al., 2017)³⁸⁰ and the growing number of urban climate adaptation plans provide instruments to do this (Carter et al., 2015; Dhar and Khirfan, 2017; Siders, 2017; Stults and Woodruff, 2017)³⁸¹. Adaptation plans can reduce exposure to urban flood risk (which, in a 1.5°C world, could double relative to 1976–2005; Alfieri et al., 2017)³⁸², heat stress (Chapter 3, Section 3.5.5.8), fire risk (Chapter 3, Section 3.4.3.4) and sea level rise (Chapter 3,

Section 3.4.5.1) (Schleussner et al., 2016)³⁸³.

Cities can reduce their risk exposure by considering investment in infrastructure and buildings that are more resilient to warming of 1.5°C or beyond. Where adaptation planning and urban planning generate the type of local participation that enhances capacity to cope with risks, they can be mutually supportive processes  (Archer et al., 2014; Kettle et al., 2014; Campos et al., 2016; Chu et al., 2017; Siders, 2017; Underwood et al., 2017)³⁸⁴. Not all adaptation plans are reported as effective (Measham et al., 2011; Hetz, 2016; Woodruff and Stults, 2016; Mahlkow and Donner, 2017)³⁸⁵, especially in developing-country cities (Kiunsi, 2013)³⁸⁶. In cases where adaptation planning may further marginalize poor citizens, either through limited local control over adaptation priorities or by displacing impacts onto poorer communities, successful urban risk management would need to consider factors such as justice, equity, and inclusive participation, as well as recognize the political economy of adaptation (Archer, 2016; Shi et al., 2016; Ziervogel et al., 2016a, 2017; Chu et al., 2017)³⁸⁷.

Integrating and promoting green urban infrastructure (including street trees, parks, green roofs and facades, and water features) into city planning can be difficult (Leck et al., 2015)³⁸⁸ but increases urban resilience to impacts of 1.5°C warming (Table 4.2) in ways that can be more cost-effective than conventional infrastructure (Cartwright et al., 2013; Culwick and Bobbins, 2016)³⁸⁹.

Realizing climate benefits from urban green infrastructure sometimes requires a city-region perspective (Wachsmuth et al., 2016)³⁹⁵. Where the urban impact on ecological systems in and beyond the city is appreciated, the potential for transformative change exists (Soderlund and Newman, 2015; Ziervogel et al., 2016a)³⁹⁶, and a locally appropriate combination of green space, ecosystem goods and services and the built environment can increase the set of urban adaptation options (Puppim de Oliveira et al., 2013)³⁹⁷.

Milan, Italy, a city with deliberate urban greening policies, planted 10,000 hectares of new forest and green areas over the last two decades (Sanesi et al., 2017)³⁹⁸. The accelerated growth of urban trees, relative to rural trees, in several regions of the world is expected to decrease tree longevity (Pretzsch et al., 2017)³⁹⁹, requiring monitoring and additional management of urban trees if their contribution to urban ecosystem-based adaptation and mitigation is to be maintained in a 1.5°C world (Buckeridge, 2015; Pretzsch et al., 2017)⁴⁰⁰.

Urban water supply and wastewater treatment is energy intensive and currently accounts for significant GHG emissions (Nair et al., 2014)⁴⁰¹. Cities can integrate sustainable water resource management and the supply of water services in ways that support mitigation, adaptation and development through waste water recycling and storm water diversion (Xue et al., 2015; Poff et al., 2016)⁴⁰². Governance and finance challenges complicate balancing sustainable water supply and rising urban demand, particularly in low-income cities (Bettini et al., 2015; Deng and Zhao, 2015; Hill Clarvis and Engle, 2015; Lemos, 2015; Margerum and Robinson, 2015)⁴⁰³.

Urban surface-sealing with impervious materials affects the volume and velocity of runoff and flooding during intense rainfall (Skougaard Kaspersen et al., 2015)⁴⁰⁴, but urban design in many cities now seeks to mediate runoff, encourage groundwater recharge and enhance water quality (Liu et al., 2014; Lamond et al., 2015; Voskamp and Van de Ven, 2015; Costa et al., 2016; Mguni et al., 2016; Xie et al., 2017)⁴⁰⁵. Challenges remain for managing intense rainfall events that are reported to be increasing in frequency and intensity in some locations (Ziervogel et al., 2016b)⁴⁰⁶, and urban flooding is expected to increase at 1.5°C of warming (Alfieri et al., 2017)⁴⁰⁷. This risk falls disproportionately on women and poor people in cities (Mitlin, 2005; Chu et al., 2016; Ziervogel et al., 2016b; Chant et al., 2017; Dodman et al., 2017a, b)⁴⁰⁸.

Nexus approaches that highlight urban areas as socio-ecological systems can support policy coherence (Rasul and Sharma, 2016)⁴⁰⁹ and sustainable urban livelihoods (Biggs et al., 2015)⁴¹⁰. The water–energy–food (WEF) nexus is especially important to growing urban populations (Tacoli et al., 2013; Lwasa et al., 2014; Villarroel Walker et al., 2014)⁴¹¹.

Isolated efficiency implementation in energy-intensive industries is a necessary but insufficient condition for deep emission reductions (Napp et al., 2014; Aden, 2018)⁴²². Various options specific to different industries are available. In general, their feasibility depends on lowering capital costs and raising awareness and expertise (Wesseling et al., 2017)⁴²³. General-purpose technologies, such as ICT, and energy management tools can improve the prospects of energy efficiency in industry (see Section 4.4.4).

Cross-sector technologies and practices, which play a role in all industrial sectors including small- and medium-sized enterprises (SMEs) and non-energy intensive industry, also offer potential for considerable energy efficiency improvements. They include: (i) motor systems (for example electric motors, variable speed drives, pumps, compressors and fans), responsible for about 10% of worldwide industrial energy consumption, with a global energy efficiency improvement potential of around 20–25% (Napp et al., 2014)⁴²⁴; and (ii) steam systems, responsible for about 30% of industrial energy consumption and energy saving potentials of about 10% (Hasanbeigi et al., 2014; Napp et al., 2014)⁴²⁵. Waste heat recovery from industry has substantial potential for energy efficiency and emission reduction (Forman et al., 2016)⁴²⁶. Low awareness and competition from other investments limit the feasibility of such options (Napp et al., 2014)⁴²⁷.

Recycling materials and developing a circular economy can be institutionally challenging, as it requires advanced capabilities (Henry et al., 2006)⁴²⁸ and organizational changes (Cooper-Searle et al., 2018)⁴²⁹, but has advantages in terms of cost, health, governance and environment (Ali et al., 2017)⁴³⁰. An assessment of the impacts on energy use and environmental issues is not available, but substitution could play a large role in reducing emissions (Åhman et al., 2016)⁴³¹ although its potential depends on the demand for material and the turnover rate of, for example, buildings (Haas et al., 2015)⁴³². Material substitution and CO₂ storage options are under development, for example, the use of algae and renewable energy for carbon fibre production, which could become a net sink of CO₂ (Arnold et al., 2018)⁴³³.

Bio-based feedstock processes could be seen as part of the circular materials economy (see section above). In several sectors, bio-based feedstocks would leave the production process of materials relatively untouched, and a switch would not affect the product quality, making the option more attractive. However, energy requirements for processing bio-based feedstocks are often high, costs are also still higher, and the emissions over the full life cycle, both upstream and downstream, could be significant (Wesseling et al., 2017)⁴³⁴. Bio-based feedstocks may put pressure on natural resources by increasing land demand by biodiversity impacts beyond bioenergy demand for electricity, transport and buildings (Slade et al., 2014)⁴³⁵, and, partly as a result, face barriers in public acceptance (Sleenhoff et al., 2015)⁴³⁶.

Electrification of manufacturing processes would constitute a significant technological challenge and would entail a more disruptive innovation in industry than bio-based or CCS options to get to very low or zero emissions, except potentially in steel-making (Philibert, 2017)⁴³⁷. The disruptive characteristics could potentially lead to stranded assets, and could reduce political feasibility and industry support (Åhman et al., 2016)⁴³⁸. Electrification of manufacturing would require further technological development in industry, as well as an ample supply of cost-effective low-emission electricity (Philibert, 2017)⁴³⁹.

Low-emission hydrogen can be produced by natural gas with CCS, by electrolysis of water powered by zero-emission electricity, or potentially in the future by generation IV nuclear reactors. Feasibility of electrification and use of hydrogen in production processes or fuel cells is affected by technical development (in terms of efficient hydrogen production and electrification of processes), by geophysical factors related to the availability of low-emission electricity (MacKay, 2013)⁴⁴⁰, by associated public perception and by economic feasibility, except in areas with ample solar and/or wind resources (Philibert, 2017; Wesseling et al., 2017)⁴⁴¹.

CO₂ capture in industry is generally considered more feasible than CCS in the power sector (Section 4.3.1) or from bioenergy sources (Section 4.3.7), although CCS in industry faces similar barriers. Almost all of the current full-scale (>1MtCO₂ yr⁻¹) CCS projects capture CO₂ from industrial sources, including the Sleipner project in Norway, which has been injecting CO₂ from a gas facility in an offshore saline formation since 1996 (Global CCS Institute, 2017)⁴⁴². Compared to the power sector, retrofitting CCS on existing industrial plants would leave the production process of materials relatively untouched (Åhman et al., 2016)⁴⁴³, though significant investments and modifications still have to be made. Some industries, in particular cement, emit CO₂ as inherent process emissions and can therefore not reduce emissions to zero without CC(U)S. CO₂ stacks in some industries have a high economic and technical feasibility for CO₂ capture as the CO₂ concentration in the exhaust gases is relatively high (IPCC, 2005b; Leeson et al., 2017)⁴⁴⁴, but others require strong modifications in the production process, limiting technical and economic feasibility, though costs remain lower than other deep GHG reduction options (Rubin et al., 2015)⁴⁴⁵. There are indications that the energy use in CO₂ capture through amine solvents (for solvent regeneration) can decrease by around 60%, from 5 GJ tCO₂⁻¹ in 2005 to 2 GJ tCO₂⁻¹ in the best-performing current pilot plants (Idem et al., 2015)⁴⁴⁶, increasing both technical and economic potential for this option. The heterogeneity of industrial production processes might point to the need for specific institutional arrangements to incentivize industrial CCS (Mikunda et al., 2014)⁴⁴⁷, and may decrease institutional feasibility.

Whether carbon dioxide utilization (CCU) can contribute to limiting warming to 1.5°C depends on the origin of the CO₂ (fossil, biogenic or atmospheric), the source of electricity for converting the CO₂ or regenerating catalysts, and the lifetime of the product. Review studies indicate that CO₂ utilization in industry has a small role to play in limiting warming to 1.5°C because of the limited potential of reusing CO₂ with currently available technologies and the re-emission of CO₂ when used as a fuel (IPCC, 2005b; Mac Dowell et al., 2017)⁴⁴⁸. However, new developments could make CCU more feasible, in particular in CO₂ use as a feedstock for carbon-based materials that would isolate CO₂ from the atmosphere for a long time, and in low-cost, low-emission electricity that would make the energy use of CO₂ capture more sustainable. The conversion of CO₂ to fuels using zero-emission electricity has a lower technical, economic and environmental feasibility than direct CO₂ capture and storage from industry (Abanades et al., 2017)⁴⁴⁹, although the economic prospects have improved recently (Philibert, 2017)⁴⁵⁰.

DRM is a process for designing, implementing and evaluating strategies, policies and measures to improve the understanding of disaster risk, and promoting improvement in disaster preparedness, response and recovery (IPCC, 2012)⁴⁵¹. There is increased demand to integrate DRM and adaptation (Howes et al., 2015; Kelman et al., 2015; Serrao-Neumann et al., 2015; Archer, 2016; Rose, 2016; van der Keur et al., 2016; Kelman, 2017; Wallace, 2017)⁴⁵² to reduce vulnerability, but institutional, technical and financial capacity challenges in frontline agencies constitute constraints (medium evidence, high agreement) (Eakin et al., 2015; Kita, 2017; Wallace, 2017)⁴⁵³.

Risks associated with 1.5ºC warming (Chapter 3, Section 3.4) may increase the demand for options that share and spread financial burdens. Formal, market-based (re)insurance spreads risk and provides a financial buffer against the impacts of climate hazards (Linnerooth-Bayer and Hochrainer-Stigler, 2015; Wolfrom and Yokoi-Arai, 2015; O’Hare et al., 2016; Glaas et al., 2017; Patel et al., 2017)⁴⁵⁴. As an alternative to traditional indemnity-based insurance, index-based micro-crop and livestock insurance programmes have been rolled out in regions with less developed insurance markets (Akter et al., 2016, 2017; Jensen and Barrett, 2017)⁴⁵⁵. There is medium evidence and medium agreement on the feasibility of insurance for adaptation, with financial, social, and institutional barriers to implementation and uptake, especially in low-income nations (García Romero and Molina, 2015; Joyette et al., 2015; Lashley and Warner, 2015; Jin et al., 2016)⁴⁵⁶. Social protection programmes include cash and in-kind transfers to protect poor and vulnerable households from the impact of economic shocks, natural disasters and other crises (World Bank, 2017b)⁴⁵⁷, and can build generic adaptive capacity and reduce vulnerability when combined with a comprehensive climate risk management approach (medium evidence, medium agreement) (Devereux, 2016; Lemos et al., 2016)⁴⁵⁸.

Educational adaptation options motivate adaptation through building awareness (Butler et al., 2016; Myers et al., 2017)⁴⁵⁹, leveraging multiple knowledge systems (Pearce et al., 2015; Janif et al., 2016)⁴⁶⁰, developing participatory action research and social learning processes (Butler and Adamowski, 2015; Ensor and Harvey, 2015; Butler et al., 2016; Thi Hong Phuong et al., 2017; Ford et al., 2018)⁴⁶¹, strengthening extension services, and building mechanisms for learning and knowledge sharing through community-based platforms, international conferences and knowledge networks (Vinke-de Kruijf and Pahl-Wostl, 2016)⁴⁶² (medium evidence, high agreement).

Climate change will exacerbate existing health challenges (Chapter 3, Section 3.4.7). Options for enhancing current health services include providing access to safe water and improved sanitation, enhancing access to essential services such as vaccination, and developing or strengthening integrated surveillance systems (WHO, 2015)⁴⁶³. Combining these with iterative management can facilitate effective adaptation (medium evidence, high agreement).

There is medium evidence and high agreement that indigenous knowledge is critical for adaptation, underpinning adaptive capacity through the diversity of indigenous agro-ecological and forest management systems, collective social memory, repository of accumulated experience and social networks (Hiwasaki et al., 2015; Pearce et al., 2015; Mapfumo et al., 2016; Sherman et al., 2016; Ingty, 2017)⁴⁶⁴ (Box 4.3). Indigenous knowledge is threatened by acculturation, dispossession of land rights and land grabbing, rapid environmental changes, colonization and social change, resulting in increasing vulnerability to climate change – which climate policy can exacerbate if based on limited understanding of indigenous worldviews (Thornton and Manasfi, 2010; Ford, 2012; Nakashima et al., 2012; McNamara and Prasad, 2014)⁴⁶⁵. Many scholars argue that recognition of indigenous rights, governance systems and laws is central to adaptation, mitigation and sustainable development (Magni, 2017; Thornton and Comberti, 2017; Pearce, 2018)⁴⁶⁶.

Human migration, whether planned, forced or voluntary, is increasingly gaining attention as a response, particularly where climatic risks are becoming severe (Chapter 3, Section 3.4.10.2). There is medium evidence and low agreement as to whether migration is adaptive, in relation to cost effectiveness concerns (Grecequet et al., 2017)⁴⁶⁷ and scalability (Brzoska and Fröhlich, 2016; Gemenne and Blocher, 2017; Grecequet et al., 2017)⁴⁶⁸. Migrating can have mixed outcomes on reducing socio-economic vulnerability (Birk and Rasmussen, 2014; Kothari, 2014; Adger et al., 2015; Betzold, 2015; Kelman, 2015; Grecequet et al., 2017; Melde et al., 2017; World Bank, 2017a; Kumari Rigaud et al., 2018)⁴⁶⁹ and its feasibility is constrained by low political and legal acceptability and inadequate institutional capacity (Betzold, 2015; Methmann and Oels, 2015; Brzoska and Fröhlich, 2016; Gemenne and Blocher, 2017; Grecequet et al., 2017; Yamamoto et al., 2017)⁴⁷⁰.

There is medium evidence and high agreement that climate services can play a critical role in aiding adaptation decision-making (Vaughan and Dessai, 2014; Wood et al., 2014; Lourenço et al., 2016; Trenberth et al., 2016; Singh et al., 2017; Vaughan et al., 2018)⁴⁷¹. The higher uptake of short-term climate information such as weather advisories and daily forecasts contrast with lesser use of longer-term information such as seasonal forecasts and multi-decadal projections (Singh et al., 2017; Vaughan et al., 2018)⁴⁷². Climate service interventions have met challenges with scaling up due to low capacity, inadequate institutions, and difficulties in maintaining systems beyond pilot project stage (Sivakumar et al., 2014; Tall et al., 2014; Gebru et al., 2015; Singh et al., 2016b)⁴⁷³, and technical, institutional, design, financial and capacity barriers to the application of climate information for better decision-making remain (Briley et al., 2015; WMO, 2015; L. Jones et al., 2016; Lourenço et al., 2016; Snow et al., 2016; Harjanne, 2017; Singh et al., 2017; C.J. White et al., 2017)⁴⁷⁴.

BECCS has been assessed in previous IPCC reports (IPCC, 2005b, 2014b; P. Smith et al., 2014; Minx et al., 2017)⁶⁰⁰ and has been incorporated into integrated assessment models (Clarke et al., 2014)⁶⁰¹ but also, 1.5°C-consistent pathways without BECCS have emerged (Bauer et al., 2018; Grubler et al., 2018; Mousavi and Blesl, 2018; van Vuuren et al., 2018)⁶⁰². Still, the overall set of  pathways limiting global warming to 1.5°C with limited or no overshoot indicates that 0–1, 0–8, and 0–16 GtCO₂ yr⁻¹ would be removed by BECCS by 2030, 2050 and 2100, respectively (Chapter 2, Section 2.3.4). BECCS is constrained by sustainable bioenergy potentials (Section 4.3.1.2, Chapter 5, Section 5.4.1.3 and Cross-Chapter Box 6 in Chapter 3), and availability of safe storage for CO₂ (Section 4.3.1.6). Literature estimates for BECCS mitigation potentials in 2050 range from 1–85 GtCO⁴. Fuss et al. (2018)⁶⁰³ narrow this range to 0.5–5 GtCO₂ yr⁻¹ (medium agreement, high evidence) (Figure 4.3), meaning that BECCS mitigation potentials are not necessarily sufficient for 1.5°C-consistent pathways. This is, among other things, related to sustainability concerns (Boysen et al., 2017; Heck et al., 2018; Henry et al., 2018)⁶⁰⁴.

Assessing BECCS deployment in 2°C pathways (of about 12 GtCO₂-eq yr⁻¹ by 2100, considered as a conservative deployment estimate for BECCS-accepting pathways consistent with 1.5°C), Smith et al. (2016b)⁶⁰⁵ estimate a land-use intensity of 0.3–0.5 ha tCO₂-eq⁻¹ yr⁻¹ using forest residues, 0.16 ha CO₂-eq⁻¹ yr⁻¹ for agricultural residues, and 0.03–0.1 ha tCO₂-eq⁻¹ yr⁻¹ for purpose-grown energy crops. The average amount of BECCS in these pathways requires 25–46% of arable and permanent crop area in 2100. Land area estimates differ in scale and are not necessarily a good indicator of competition with, for example, food production, because requiring a smaller land area for the same potential could indicate that high-productivity agricultural land is used. In general, the literature shows low agreement on the availability of land (Fritz et al., 2011⁶⁰⁶; see Erb et al., 2016b⁶⁰⁷ for recent advances). Productivity, food production and competition with other ecosystem services and land use by local communities are important factors for designing regulation. These potentials and trade-offs are not homogenously distributed across regions. However, Robledo-Abad et al. (2017)⁶⁰⁸ find that regions with higher potentials are understudied, given their potential contribution. Researchers have expressed the need to complement global assessments with regional, geographically explicit bottom-up studies of biomass potentials and socio-economic impacts (e.g., de Wit and Faaij, 2010; Kraxner et al., 2014; Baik et al., 2018)⁶⁰⁹.

Energy production and land and water footprints show wide ranges in bottom-up assessments due to differences in technology, feedstock and other parameters (−1–150 EJ yr⁻¹ of energy, 109–990 Mha, 6–79 MtN, 218–4758 km³ yr⁻¹ of water per GtCO₂ yr⁻¹; Smith and Torn, 2013; Smith et al., 2016b; Fajardy and Mac Dowell, 2017)⁶¹⁰ and are not comparable to IAM pathways which consider system effects (Bauer et al., 2018)⁶¹¹. Global impacts on nutrients and albedo are difficult to quantify (Smith et al., 2016b)⁶¹². BECCS competes with other land-based CDR and mitigation measures for resources (Chapter 2).

There is uncertainty about the feasibility of timely upscaling (Nemet et al., 2018)⁶¹³. CCS (see Section 4.3.1) is largely absent from the Nationally Determined Contributions (Spencer et al., 2015)⁶¹⁴ and lowly ranked in investment priorities (Fridahl, 2017)⁶¹⁵. Although there are dozens of small-scale BECCS demonstrations (Kemper, 2015)⁶¹⁶ and a full-scale project capturing 1 MtCO₂ exists (Finley, 2014)⁶¹⁷, this is well below the numbers associated with 1.5°C or 2°C-compatible pathways (IEA, 2016a; Peters et al., 2017)⁶¹⁸. Although the majority of BECCS cost estimates are below 200 USD tCO ⁻¹ (Figure 4.2), estimates vary widely. Economic incentives for ramping up large CCS or BECCS infrastructure are weak (Bhave et al., 2017)⁶¹⁹. The 2050 average investment costs for such a BECCS infrastructure for bio-electricity and biofuels are estimated at 138 and 123 billion USD yr⁻¹, respectively (Smith et al., 2016b)⁶²⁰.

BECCS deployment is further constrained by bioenergy’s carbon accounting, land, water and nutrient requirements (Section 4.3.1), its compatibility with other policy goals and limited public acceptance of both bioenergy and CCS (Section 4.3.1). Current pathways are believed to have inadequate assumptions on the development of societal support and governance structures (Vaughan and Gough, 2016)⁶²¹. However, removing BECCS and CCS from the portfolio of available options significantly raises modelled mitigation costs (Kriegler et al., 2013; Bauer et al., 2018)⁶²².

Afforestation implies planting trees on land not forested for a long time (e.g., over the last 50 years in the context of the Kyoto Protocol), while reforestation implies re-establishment of forest formations after a temporary condition with less than 10% canopy cover due to human-induced or natural perturbations. Houghton et al. (2015)⁶²⁶ estimate about 500 Mha could be available for the re-establishment of forests on lands previously forested, but not currently used productively. This could sequester at least 3.7 GtCO₂ yr⁻¹ for decades. The full literature range gives 2050 potentials of 1–7 GtCO₂ yr⁻¹ (low evidence, medium agreement), narrowed down to 0.5–3.6 GtCO₂ yr⁻¹ based on a number of constraints (Fuss et al., 2018)⁶²⁷. Abatement costs are estimated to be low compared to other CDR options, 5–50 USD tCO₂-eq⁻¹ (robust evidence, high agreement). Yet, realizing such large potentials comes at higher land and water footprints than BECCS, although there would be a positive impact on nutrients and the energy requirement would be negligible (Smith et al., 2016b⁶²⁸; Cross-Chapter Box 7 in Chapter 3). The 2030 estimate by Griscom et al. (2017)⁶²⁹ is up to 17.9 GtCO₂ yr⁻¹ for reforestation with significant co-benefits (Cross-Chapter Box 7 in Chapter 3).

Biogenic storage is not as permanent as emission reductions by geological storage. In addition, forest sinks saturate, a process which typically occurs in decades to centuries compared to the thousands of years of residence time of CO₂ stored geologically (Smith et al., 2016a)⁶³⁰ and is subject to disturbances that can be exacerbated by climate change (e.g., drought, forest fires and pests) (Seidl et al., 2017)⁶³¹. Handling these challenges requires careful forest management. There is much practical experience with AR, facilitating upscaling but with two caveats: AR potentials are heterogeneously distributed (Bala et al., 2007)⁶³², partly because the planting of less reflective forests results in higher net absorbed radiation and localised surface warming in higher latitudes (Bright et al., 2015; Jones et al., 2015)⁶³³, and forest governance structures and monitoring capacities can be bottlenecks and are usually not considered in models (Wang et al., 2016; Wehkamp et al., 2018b)⁶³⁴. There is medium agreement on the positive impacts of AR on ecosystems and biodiversity due to different forms of afforestation discussed in the literature: afforestation of grassland ecosystems or diversified agricultural landscapes with monocultures or invasive alien species can have significant negative impacts on biodiversity, water resources, etc. (P. Smith et al., 2014)⁶³⁵, while forest ecosystem restoration (forestry and agroforestry) with native species can have positive social and environmental impacts (Cunningham et al., 2015; Locatelli et al., 2015; Paul et al., 2016⁶³⁶; See Section 4.3.2).

Synergies with other policy goals are possible (see also Section 4.5.4); for example, land spared by diet shifts could be afforested (Röös et al., 2017)⁶³⁷ or used for energy crops (Grubler et al., 2018)⁶³⁸. Such land-sparing strategies could also benefit other land-based CDR options.

At local scales there is robust evidence that soil carbon sequestration (SCS, e.g., agroforestry, De Stefano and Jacobson, 2018)⁶³⁹, restoration of degraded land (Griscom et al., 2017)⁶⁴⁰, or conservation agriculture management practices (Aguilera et al., 2013; Poeplau and Don, 2015; Vicente-Vicente et al., 2016)⁶⁴¹ have co-benefits in agriculture and that many measures are cost-effective even without supportive climate policy. Evidence at global scale for potentials and especially costs is much lower. The literature spans cost ranges of −45–100 USD tCO₂⁻¹ (negative costs relating to the multiple co-benefits of SCS, such as increased productivity and resilience of soils; P. Smith et al., 2014)⁶⁴², and 2050 potentials are estimated at between 0.5 and 11 GtCO₂ yr⁻¹, narrowed down to 2.3–5.3 GtCO₂ yr⁻¹ considering that studies above 5 GtCO₂ yr⁻¹ often do not apply constraints, while estimates lower than 2 GtCO₂ yr⁻¹ mostly focus on single practices (Fuss et al., 2018)⁶⁴³.

SCS has negligible water and energy requirements (Smith, 2016)⁶⁴⁴, affects nutrients and food security favourably (high agreement, robust evidence) and can be applied without changing current land use, thus making it socially more acceptable than CDR options with a high land footprint. However, soil sinks saturate after 10–100 years, depending on the SCS option, soil type and climate zone (Smith, 2016)⁶⁴⁵.

Biochar is formed by recalcitrant (i.e., very stable) organic carbon obtained from pyrolysis, which, applied to soil, can increase soil carbon sequestration leading to improved soil fertility properties.⁵ Looking at the full literature range, the global potential in 2050 lies between 1 and 35 Gt CO₂ yr⁻¹ (low agreement, low evidence), but considering limitations in biomass availability and uncertainties due to a lack of large-scale trials of biochar application to agricultural soils under field conditions, Fuss et al. (2018)⁶⁴⁶ lower the 2050 range to 0.3–2 GtCO₂ yr⁻¹. This potential is below previous estimates (e.g., Woolf et al., 2010)⁶⁴⁷, which additionally consider the displacement of fossil fuels through biochar. Permanence depends on soil type and biochar production temperatures, varying between a few decades and several centuries (Fang et al., 2014)⁶⁴⁸. Costs are 30– 120 USD tCO₂⁻¹ (medium agreement, medium evidence) (McCarl et al., 2009; McGlashan et al., 2012; McLaren, 2012; Smith, 2016)⁶⁴⁹.

Water requirements are low and at full theoretical deployment, up to 65 EJ yr⁻¹ of energy could be generated as a side product (Smith, 2016)⁶⁵⁰. Positive side effects include a favourable effect on nutrients and reduced N₂O emissions (Cayuela et al., 2014; Kammann et al., 2017)⁶⁵¹. However, 40–260 Mha are needed to grow the biomass for biochar for implementation at 0.3 GtCO₂-eq yr⁻¹ (Smith, 2016)⁶⁵², even though it is also possible to use residues (e.g., Windeatt et al., 2014)⁶⁵³. Biochar is further constrained by the maximum safe holding capacity of soils (Lenton, 2010)⁶⁵⁴ and the labile nature of carbon sequestrated in plants and soil at higher temperatures (Wang et al., 2013)⁶⁵⁵.

Weathering is the natural process of rock decomposition via chemical and physical processes in which CO₂ is spontaneously consumed and converted to solid or dissolved alkaline bicarbonates and/or carbonates (IPCC, 2005a)⁶⁵⁶. The process is controlled by temperature, reactive surface area, interactions with biota and, in particular, water solution composition. CDR can be achieved by accelerating mineral weathering through the distribution of ground-up rock material over land (Hartmann and Kempe, 2008; Wilson et al., 2009; Köhler et al., 2010; Renforth, 2012; ten Berge et al., 2012; Manning and Renforth, 2013; Taylor et al., 2016)⁶⁵⁷, shorelines (Hangx and Spiers, 2009; Montserrat et al., 2017)⁶⁵⁸ or the open ocean (House et al., 2007; Harvey, 2008; Köhler et al., 2013; Hauck et al., 2016)⁶⁵⁹. Ocean alkalinization adds alkalinity to marine areas to locally increase the CO₂ buffering capacity of the ocean (González and Ilyina, 2016; Renforth and Henderson, 2017)⁶⁶⁰.

In the case of land application of ground minerals, the estimated CDR potential range is 0.72–95 GtCO₂ yr⁻¹ (low evidence, low agreement) (Hartmann and Kempe, 2008; Köhler et al., 2010; Hartmann et al., 2013; Taylor et al., 2016; Strefler et al., 2018a)⁶⁶¹. Marine application of ground minerals is limited by feasible rates of mineral extraction, grinding and delivery, with estimates of  1–6 GtCO₂ yr⁻¹ (low evidence, low agreement) (Köhler et al., 2013; Hauck et al., 2016; Renforth and Henderson, 2017)⁶⁶². Agreement is low due to a variety of assumptions and unknown parameter ranges in the applied modelling procedures that would need to be verified by field experiments (Fuss et al., 2018)⁶⁶³. As with other CDR options, scaling and maturity are challenges, with deployment at scale potentially requiring decades (NRC, 2015a)⁶⁶⁴, considerable costs in transport and disposal (Hangx and Spiers, 2009; Strefler et al., 2018a)⁶⁶⁵ and mining (NRC, 2015a; Strefler et al., 2018a)⁶⁶⁶⁶.

Site-specific cost estimates vary depending on the chosen technology for rock grinding (an energy-intensive process; Köhler et al., 2013; Hauck et al., 2016)⁶⁶⁷, material transport, and rock source (Renforth, 2012; Hartmann et al., 2013)⁶⁶⁸, and range from 15–40 USD tCO₂⁻¹ to 3,460 USD tCO₂⁻¹ (limited evidence, low agreement; Figure 4.2) (Schuiling and Krijgsman, 2006; Köhler et al., 2010; Taylor et al., 2016)⁶⁶⁹. The evidence base for costs of ocean alkalinization and marine enhanced weathering is sparser than the land applications. The ocean alkalinization potential is assessed to be 0.1–10 GtCO₂ yr⁻¹ with costs of 14– >500 USD tCO₂⁻¹ (Renforth and Henderson, 2017)⁶⁷⁰.

The main side effects of terrestrial EW are an increase in water pH (Taylor et al., 2016)⁶⁷¹, the release of heavy metals like Ni and Cr and plant nutrients like K, Ca, Mg, P and Si (Hartmann et al., 2013)⁶⁷², and changes in hydrological soil properties. Respirable particle sizes, though resulting in higher potentials, can have impacts on health (Schuiling and Krijgsman, 2006; Taylor et al., 2016)⁶⁷³; utilization of wave-assisted decomposition through deployment on coasts could avert the need for fine grinding (Hangx and Spiers, 2009; Schuiling and de Boer, 2010)⁶⁷⁴. Side effects of marine EW and ocean alkalinization are the potential release of heavy metals like Ni and Cr (Montserrat et al., 2017)⁶⁷⁵. Increasing ocean alkalinity helps counter ocean acidification (Albright et al., 2016; Feng et al., 2016)⁶⁷⁶. Ocean alkalinization could affect ocean biogeochemical functioning (González and Ilyina, 2016)⁶⁷⁷. A further caveat of relates to saturation state and the potential to trigger spontaneous carbonate precipitation.⁷ While the geochemical potential to remove and store CO₂ is quite large, limited evidence on the preceding topics makes it difficult to assess the true capacity, net benefits and desirability of EW and ocean alkalinity addition in the context of CDR.

Capturing CO₂ from ambient air through chemical processes with subsequent storage of the CO₂ in geological formations is independent of source and timing of emissions and can avoid competition for land. Yet, this is also the main challenge: while the theoretical potential for DACCS is mainly limited by the availability of safe and accessible geological storage, the CO₂ concentration in ambient air is 100–300 times lower than at gas- or coal-fired power plants (Sanz-Pérez et al., 2016)⁶⁷⁸ thus requiring more energy than flue gas CO₂ capture (Pritchard et al., 2015)⁶⁷⁹. This appears to be the main challenge to DACCS (Sanz-Pérez et al., 2016; Barkakaty et al., 2017)⁶⁸⁰.

Studies explore alternative techniques to reduce the energy penalty of DACCS (van der Giesen et al., 2017)⁶⁸¹. Energy consumption could be up to 12.9 GJ tCO₂-eq⁻¹; translating into an average of 156 EJ yr⁻¹ by 2100 (current annual global primary energy supply is 600 EJ); water requirements are estimated to average 0.8–24.8 km³ GtCO₂-eq⁻¹ yr⁻¹  (Smith et al., 2016b, based on Socolow et al., 2011)⁶⁸².

However, the literature shows low agreement and is fragmented (Broehm et al., 2015)⁶⁸³. This fragmentation is reflected in a large range of cost estimates: from 20–1,000 USD tCO₂⁻¹ (Keith et al., 2006; Pielke, 2009; House et al., 2011; Ranjan and Herzog, 2011; Simon et al., 2011; Goeppert et al., 2012; Holmes and Keith, 2012; Zeman, 2014; Sanz-Pérez et al., 2016; Sinha et al., 2017)⁶⁸⁴. There is lower agreement and a smaller evidence base at the lower end of the cost range. Fuss et al. (2018)⁶⁸⁵ narrow this range to 100–300 USD tCO₂^-1.

Research and efforts by small-scale commercialization projects focus on utilization of captured CO₂ (Wilcox et al., 2017)⁶⁸⁶. Given that only a few IAM scenarios incorporate DACCS (e.g., Chen and Tavoni, 2013; ⁶⁸⁷Strefler et al., 2018b)⁶⁸⁸ its possible role in cost-optimized 1.5°C scenarios is not yet fully explored. Given the technology’s early stage of development (McLaren, 2012; NRC, 2015a; Nemet et al., 2018)⁶⁸⁹ and few demonstrations (Holmes et al., 2013; Rau et al., 2013; Agee et al., 2016)⁶⁹⁰, deploying the technology at scale is still a considerable challenge, though both optimistic (Lackner et al., 2012)⁶⁹¹ and pessimistic outlooks exist (Pritchard et al., 2015)⁶⁹².

Nutrients can be added to the ocean resulting in increased biologic production, leading to carbon fixation in the sunlit ocean and subsequent sequestration in the deep ocean or sea floor sediments. The added nutrients can be either micronutrients (such as iron) or macronutrients (such as nitrogen and/or phosphorous) (Harrison, 2017)⁶⁹³. There is limited evidence and low agreement on the readiness of this technology to contribute to rapid decarbonization (Williamson et al., 2012)⁶⁹⁴. Only small-scale field experiments and theoretical modelling have been conducted (e.g., McLaren, 2012)⁶⁹⁵. The full range of CDR potential estimates is from 15.2 ktCO₂ yr⁻¹ (Bakker et al., 2001)⁶⁹⁶ for a spatially constrained field experiment up to 44 GtCO₂ yr⁻¹ (Sarmiento and Orr, 1991)⁶⁹⁷ following a modelling approach, but Fuss et al. (2018)⁶⁹⁸ consider the potential to be extremely limited given the evidence and existing barriers. Due to scavenging of iron, the iron addition only leads to inefficient use of the nitrogen in exporting carbon (Zeebe, 2005; Aumont and Bopp, 2006; Zahariev et al., 2008)⁶⁹⁹.

Cost estimates range from 2 USD tCO₂⁻¹ (for iron fertilization) (Boyd and Denman, 2008)⁷⁰⁰ to 457 USD tCO₂⁻¹ (Harrison, 2013)⁷⁰¹. Jones (2014)⁷⁰² proposed values greater than 20 USD tCO₂⁻¹ for nitrogen fertilization. Fertilization is expected to impact food webs by stimulating its base organisms (Matear, 2004)⁷⁰³, and extensive algal blooms may cause anoxia (Sarmiento and Orr, 1991; Matear, 2004; Russell et al., 2012)⁷⁰⁴ and deep water oxygen decline (Matear, 2004)⁷⁰⁵, with negative impacts on biodiversity. Nutrient inputs can shift ecosystem production from an iron-limited system to a P, N-, or Si-limited system depending on the location (Matear, 2004; Bertram, 2010)⁷⁰⁶ and non-CO₂ GHGs may increase (Sarmiento and Orr, 1991; Matear, 2004; Bertram, 2010)⁷⁰⁷. The greatest theoretical potential for this practice is the Southern Ocean, posing challenges for monitoring and governance (Robinson et al., 2014)⁷⁰⁸. The London Protocol of the International Maritime Organization has asserted authority for regulation of ocean fertilization (Strong et al., 2009)⁷⁰⁹, which is widely viewed as a de facto moratorium on commercial ocean fertilization activities.

There is low agreement in the technical literature on the permanence of CO₂ in the ocean, with estimated residence times of 1,600 years to millennia, especially if injected or buried in or below the sea floor (Williams and Druffel, 1987; Jones, 2014)⁷¹⁰. Storage at the surface would mean that the carbon would be rapidly released after cessation (Zeebe, 2005; Aumont and Bopp, 2006)⁷¹¹.

There is robust evidence but medium agreement for unilateral action potentially becoming a serious SRM governance issue (Weitzman, 2015; Rabitz, 2016)⁷⁴⁸, as some argue that enhanced collaboration might emerge around SRM (Horton, 2011)⁷⁴⁹. An equitable institutional or governance arrangement around SRM would have to reflect views of different countries (Heyen et al., 2015)⁷⁵⁰ and be multilateral because of the risk of termination, and risks that implementation or unilateral action by one country or organization will produce negative precipitation or extreme weather effects across borders (Lempert and Prosnitz, 2011; Dilling and Hauser, 2013; NRC, 2015b)⁷⁵¹. Some have suggested that the governance of research and field experimentation can help clarify uncertainties surrounding deployment of SRM (Long and Shepherd, 2014; Parker, 2014; NRC, 2015c; Caldeira and Bala, 2017; Lawrence and Crutzen, 2017)⁷⁵², and that SRM is compatible with democratic processes (Horton et al., 2018)⁷⁵³ or not (Szerszynski et al., 2013; Owen, 2014)⁷⁵⁴.

Several possible institutional arrangements have been considered for SRM governance: under the UNFCCC (in particular under the Subsidiary Body on Scientific and Technological Advice (SBSTA)) or the United Nations Convention on Biological Diversity (UNCBD) (Honegger et al., 2013; Nicholson et al., 2018)⁷⁵⁵, or through a consortium of states (Bodansky, 2013; Sandler, 2017)⁷⁵⁶. Reasons for states to join an international governance framework for SRM include having a voice in SRM diplomacy, prevention of unilateral action by others and benefits from research collaboration (Lloyd and Oppenheimer, 2014)⁷⁵⁷.

Alongside SBSTA, the WMO, UNESCO and UN Environment could play a role in governance of SRM (Nicholson et al., 2018)⁷⁵⁸. Each of these organizations has relevance with respect to the regulatory framework (Bodle et al., 2012; Williamson and Bodle, 2016)⁷⁵⁹. The UNCBD gives guidance that ‘that no climate-related geo-engineering activities that may affect biodiversity take place’ (CBD, 2010)⁷⁶⁰.

The literature on the engineering costs of SRM is limited and may be unreliable in the absence of testing or deployment. There is high agreement that costs of SAI (not taking into account indirect and social costs, research and development costs and monitoring expenses) may be in the range of 1–10 billion USD yr⁻¹ for injection of 1–5 MtS to achieve cooling of 1–2 W m⁻² (Robock et al., 2009; McClellan et al., 2012; Ryaboshapko and Revokatova, 2015; Moriyama et al., 2016)⁷⁶¹, suggesting that cost-effectiveness may be high if side-effects are low or neglected (McClellan et al., 2012)⁷⁶². The overall economic feasibility of SRM also depends on externalities and social costs (Moreno-Cruz and Keith, 2013; Mackerron, 2014)⁷⁶³, climate sensitivity (Kosugi, 2013)⁷⁶⁴, option value (Arino et al., 2016)⁷⁶⁵, presence of climate tipping points (Eric Bickel, 2013)⁷⁶⁶  and damage costs as a function of the level of SRM (Bahn et al., 2015; Heutel et al., 2018)⁷⁶⁷. Modelling of game-theoretic, strategic interactions of states under heterogeneous climatic impacts shows low agreement on the outcome and viability of a cost-benefit analysis for SRM (Ricke et al., 2015; Weitzman, 2015)⁷⁶⁸.

For SAI, there is high agreement that aircrafts could, after some modifications, inject millions of tons of SO₂ in the lower stratosphere (at approximately 20 km; (Davidson et al., 2012; McClellan et al., 2012; Irvine et al., 2016)⁷⁶⁹.

Ethical questions around SRM include those of international responsibilities for implementation, financing, compensation for negative effects, the procedural justice questions of who is involved in decisions, privatization and patenting, welfare, informed consent by affected publics, intergenerational ethics (because SRM requires sustained action in order to avoid termination hazards), and the so-called ‘moral hazard’ (Burns, 2011; Whyte, 2012; Gardiner, 2013; Lin, 2013; Buck et al., 2014; Klepper and Rickels, 2014; Morrow, 2014; Wong, 2014; Reynolds, 2015; Lockley and Coffman, 2016; McLaren, 2016; Suarez and van Aalst, 2017; Reynolds et al., 2018)⁷⁷⁰. The literature shows low agreement on whether SRM research and deployment may lead policy-makers to reduce mitigation efforts and thus imply a moral hazard (Linnér and Wibeck, 2015)⁷⁷¹. SRM might motivate individuals (as opposed to policymakers) to reduce their GHG emissions, but even a subtle difference in the articulation of information about SRM can influence subsequent judgements of favourability (Merk et al., 2016)⁷⁷². The argument that SRM research increases the likelihood of deployment (the ‘slippery slope’ argument), is also made (Quaas et al., 2017)⁷⁷³, but some also found an opposite effect (Bellamy and Healey, 2018)⁷⁷⁴.

Unequal representation and deliberate exclusion are plausible in decision-making on SRM, given diverging regional interests and the anticipated low resource requirements to deploy SRM (Ricke et al., 2013)⁷⁷⁵. Whyte (2012)⁷⁷⁶ argues that the concerns, sovereignties, and experiences of indigenous peoples may particularly be at risk.

The general public can be characterized as oblivious to and worried about SRM (Carr et al., 2013; Parkhill et al., 2013; Wibeck et al., 2017)⁷⁷⁷. An emerging literature discusses public perception of SRM, showing a lack of knowledge and unstable  opinions (Scheer and Renn, 2014)⁷⁷⁸. The perception of controllability affects legitimacy and public acceptability of SRM experiments (Bellamy et al., 2017)⁷⁷⁹. In Germany, laboratory work on SRM is generally approved of, field research much less so, and immediate deployment is largely rejected (Merk et al., 2015; Braun et al., 2017)⁷⁸⁰. Various factors could explain variations in the degree of rejection of SRM between Canada, China, Germany, Switzerland, the United Kingdom, and the United States (Visschers et al., 2017)⁷⁸¹.

Institutions – the rules and norms that guide human interactions (Section 4.4.2) – enable or impede the structures, mechanisms and measures that guide mitigation and adaptation. Institutions, understood as the ‘rules of the game’ (North, 1990)⁸²², exert direct and indirect influence over the viability of 1.5°C-consistent pathways (Munck et al., 2014; Willis, 2017)⁸²³. Governance would be needed to support wide-scale and effective adoption of mitigation and adaptation options. Institutions and governance structures are strengthened when the principle of the ‘commons’ is explored as a way of sharing management and responsibilities (Ostrom et al., 1999; Chaffin et al., 2014; Young, 2016)⁸²⁴. Institutions would need to be strengthened to interact amongst themselves, and to share responsibilities for the development and implementation of rules, regulations and policies (Ostrom et al., 1999; Wejs et al., 2014; Craig et al., 2017)⁸²⁵, with the goal of ensuring that these embrace equity, justice, poverty alleviation and sustainable development, enabling a 1.5°C world (Reckien et al., 2017; Wood et al., 2017)⁸²⁶.

Several authors have identified different modes of cross-stakeholder interaction in climate policy, including the role played by large multinational corporations, small enterprises, civil society and non-state actors. Ciplet et al. (2015)⁸²⁷ argue that civil society is to a great extent the only reliable motor for driving institutions to change at the pace required. Kern and Alber (2009)⁸²⁸ recognize different forms of collaboration relevant to successful climate policies beyond the local level. Horizontal collaboration (e.g., transnational city networks) and vertical collaboration within nation-states can play an enabling role (Ringel, 2017)⁸²⁹. Vertical and horizontal collaboration requires synergistic relationships between stakeholders (Ingold and Fischer, 2014; Hsu et al., 2017)⁸³⁰. The importance of community participation is emphasized in literature, and in particular the need to take into account equity and gender considerations (Chapter 5) (Graham et al., 2015; Bryan et al., 2017; Wangui and Smucker, 2017)⁸³¹. Participation often faces implementation challenges and may not always result in better policy outcomes. Stakeholders, for example, may not view climate change as a priority and may not share the same preferences, potentially creating a policy deadlock (Preston et al., 2013, 2015; Ford et al., 2016)⁸³².

International treaties help strengthen policy implementation, providing a medium- and long-term vision (Obergassel et al., 2016)⁸³³. International climate governance is organized via many mechanisms, including international organizations, treaties and conventions, for example, UNFCCC, the Paris Agreement and the Montreal Protocol. Other multilateral and bilateral agreements, such as trade agreements, also have a bearing on climate change.

There are significant differences between global mitigation and adaptation governance frames. Mitigation tends to be global by its nature and based on the principle of the climate system as a global commons (Ostrom et al., 1999)⁸³⁴. Adaptation has traditionally been viewed as a local process, involving local authorities, communities, and stakeholders (Khan, 2013; Preston et al., 2015)⁸³⁵, although it is now recognized to be a multi-scaled, multi-actor process that transcends scales from local and sub-national to national and international (Mimura et al., 2014; UNEP, 2017a)⁸³⁶. National governments provide a central pivot for coordination, planning, determining policy priorities and distributing resources. National governments are accountable to the international community through international agreements. Yet, many of the impacts of climate change are transboundary, so that bilateral and multilateral cooperation are needed (Nalau et al., 2015; Donner et al., 2016; Magnan and Ribera, 2016; Tilleard and Ford, 2016; Lesnikowski et al., 2017)⁸³⁷. The Kigali Amendment to the Montreal Protocol demonstrates that a global environmental agreement facilitating common but differentiated responsibilities is possible (Sharadin, 2018)⁸³⁸. This was operationalized by developed countries acting first, with developing countries following and benefiting from leap-frogging the trial-and-error stages of innovative technology development.

Work on international climate governance has focused on the nature of ‘climate regimes’ and coordinating the action of nation-states (Aykut, 2016)⁸³⁹ organized around a diverse set of instruments: (i) binding limits allocated by principles of historical responsibility and equity, (ii) carbon prices, emissions quotas, (iii) pledges and review of policies and measures or (iv) a combination of these options (Stavins, 1988; Grubb, 1990; Pizer, 2002; Newell and Pizer, 2003)⁸⁴⁰.

Literature on the Kyoto Protocol provides two important insights for the 1.5°C transition: the challenge of agreeing on rules to allocate emissions quotas (Shukla, 2005; Caney, 2012; Winkler et al., 2013; Gupta, 2014; Méjean et al., 2015)⁸⁴¹ and a climate-centric vision (Shukla, 2005; BASIC experts, 2011)⁸⁴², separated from development issues which drove resistance from many developing nations (Roberts and Parks, 2006)⁸⁴³. For the former, a burden-sharing approach led to an adversarial process among nations to decide who should be allocated ‘how much’ of the remainder of the emissions budget (Caney, 2014; Ohndorf et al., 2015; Roser et al., 2015; Giménez-Gómez et al., 2016)⁸⁴⁴. Industry group lobbying further contributed to reducing space for manoeuvre of some major emitting nations (Newell and Paterson, 1998; Levy and Egan, 2003; Dunlap and McCright, 2011; Michaelowa, 2013; Geels, 2014)⁸⁴⁵.

Given the political unwillingness to continue with the Kyoto Protocol approach a new approach was introduced in the Copenhagen Accord, the Cancun Agreements, and finally in the Paris Agreement. The transition to 1.5°C requires carbon neutrality and thus going beyond the traditional framing of climate as a ‘tragedy of the commons’ to be addressed via cost-optimal allocation rules, which demonstrated a low probability of enabling a transition to 1.5°C-consistent pathways (Patt, 2017)⁸⁴⁶. The Paris Agreement, built on a ‘pledge and review’ system, is thought be more effective in securing trust (Dagnet et al., 2016)⁸⁴⁷ and enables effective monitoring and timely reporting on national actions (including adaptation), allowing for international scrutiny and persistent efforts of civil society and non-state actors to encourage action in both national and international contexts (Allan and Hadden, 2017; Bäckstrand and Kuyper, 2017; Höhne et al., 2017; Lesnikowski et al., 2017; Maor et al., 2017; UNEP, 2017a)⁸⁴⁸, with some limitations (Nieto et al., 2018)⁸⁴⁹.

The paradigm shift enabled at Cancun succeeded by focusing on the objective of ‘equitable access to sustainable development’ (Hourcade et al., 2015)⁸⁵⁰. The use of ‘pledge and review’ now underpins the Paris Agreement. This consolidates multiple attempts to define a governance approach that relies on Nationally Determined Contributions (NDCs) and on means for a ‘facilitative model’ (Bodansky and Diringer, 2014)⁸⁵¹ to reinforce them. This enables a regular, iterative, review of NDCs allowing countries to set their own ambitions  after a global stocktake and more flexible, experimental forms of climate governance, which may provide room for higher ambition and be consistent with the needs of governing for a rapid transition to close the emission gap (Clémençon, 2016; Falkner, 2016)⁸⁵² (Cross-Chapter Box 11 in this chapter). Beyond a general consensus on the necessity of measurement, reporting and verification (MRV) mechanisms as a key element of a climate regime (Ford et al., 2015b; van Asselt et al., 2015)⁸⁵³, some authors emphasize different governance approaches to implement the Paris Agreement. Through the new proposed sustainable development mechanism in Article 6, the Paris Agreement allows the space to harness the lowest cost mitigation options worldwide. This may incentivize policymakers to enhance mitigation ambition by speeding up climate action as part of a ‘climate regime complex’ (Keohane and Victor, 2011)⁸⁵⁴ of loosely interrelated global governance institutions. In the Paris Agreement, the ‘common but differentiated responsibilities and respective capabilities’ (CBDR-RC) principle could be expanded and revisited under a ‘sharing the pie’ paradigm (Ji and Sha, 2015)⁸⁵⁵ as a tool to open innovation processes towards alternative development pathways (Chapter 5).

COP 16 in Cancun was also the first time in the UNFCCC that adaptation was recognized to have similar priority as mitigation. The Paris Agreement recognizes the importance of adaptation action and cooperation to enhance such action. Chung Tiam Fook (2017)⁸⁵⁶ and Lesnikowski et al. (2017)⁸⁵⁷ suggest that the Paris Agreement is explicit about multilevel adaptation governance, outlines stronger transparency mechanisms, links adaptation to development and climate justice, and is therefore suggestive of greater inclusiveness of non-state voices and the broader contexts of social change.

1.5°C-consistent pathways require further exploration of conditions of trust and reciprocity amongst nation states (Schelling, 1991; Ostrom and Walker, 2005)⁸⁵⁸. Some authors (Colman et al., 2011; Courtois et al., 2015)⁸⁵⁹

suggest a departure from the vision of actors acting individually in the pursuit of self-interest to that of iterated games with actors interacting over time showing that reciprocity, with occasional forgiveness and initial good faith, can lead to win-win outcomes and to cooperation as a stable strategy (Axelrod and Hamilton, 1981)⁸⁶⁰.

Regional cooperation plays an important role in the context of global governance. Literature on climate regimes has only started exploring innovative governance arrangements, including coalitions of transnational actors including state, market and non-state actors (Bulkeley et al., 2012; Hovi et al., 2016; Hagen et al., 2017; Hermwille et al., 2017; Roelfsema et al., 2018)⁸⁶¹ and groupings of countries, as a complement to the UNFCCC (Abbott and Snidal, 2009; Biermann, 2010; Zelli, 2011; Nordhaus, 2015)⁸⁶². Climate action requires multilevel governance from the local and community level to national, regional and international levels. Box 4.1 shows the role of sub-national authorities (e.g., regions and provinces) in facilitating urban climate action, while Box 4.2 shows that climate governance can be organized across hydrological as well as political units.

Local governments can play a key role (Melica et al., 2018; Romero-Lankao et al., 2018)⁸⁶³ in influencing mitigation and adaptation strategies. It is important to understand how rural and urban areas, small islands, informal settlements and communities might intervene to reduce climate impacts (Bulkeley et al., 2011)⁸⁶⁴, either by implementing climate objectives defined at higher government levels or by taking initiative autonomously or collectively (Aall et al., 2007; Reckien et al., 2014; Araos et al., 2016a; Heidrich et al., 2016)⁸⁶⁵. Local governance faces the challenge of reconciling local concerns with global objectives. Local governments could coordinate and develop effective local responses, and could pursue procedural justice in ensuring community engagement and more effective policies around energy and vulnerability reduction (Moss et al., 2013; Fudge et al., 2016)⁸⁶⁶. They can enable more participative decision-making (Barrett, 2015; Hesse, 2016)⁸⁶⁷. Fudge et al. (2016)⁸⁶⁸ argue that local authorities are well-positioned to involve the wider community in: designing and implementing climate policies, engaging with sustainable energy generation (e.g., by supporting energy communities) (Slee, 2015)⁸⁶⁹, and the delivery of demand-side measures and adaptation implementation.

By 2050, it is estimated three billion people will be living in slums and informal settlements: neighbourhoods without formal governance, on un-zoned land developments and in places that are exposed to climate-related hazards (Bai et al., 2018)⁸⁷⁰. Emerging research is examining how citizens can contribute informally to governance with rapid urbanization and weaker government regulation (Sarmiento and Tilly, 2018)⁸⁷¹. It remains to be seen how the possibilities and consequences of alternative urban governance models will be managed for large, complex problems and for addressing inequality and urban adaptation (Amin and Cirolia, 2018; Bai et al., 2018; Sarmiento and Tilly, 2018)⁸⁷².

Expanding networks of cities are sharing experiences on coping with climate change and drawing economic and development benefits from climate change responses – a recent institutional innovation. This could be complemented by efforts of national governments to enhance local climate action through national urban policies (Broekhoff et al., 2018)⁸⁷³. Over the years, non-state actors have set up several transnational climate governance initiatives to accelerate the climate response, for example, ICLEI (1990), C–40 (2005), the Global Island Partnership (2006) and the Covenant of Mayors (2008) (Gordon and Johnson, 2017; Hsu et al., 2017; Ringel, 2017; Kona et al., 2018; Melica et al., 2018)⁸⁷⁴ and to exert influence on national governments and the UNFCCC (Bulkeley, 2005)⁸⁷⁵. However, Michaelowa and Michaelowa (2017)⁸⁷⁶ find low effectiveness for over 100 of such mitigation initiatives.

Literature has proposed multilevel governance in climate change as an enabler for systemic transformation and effective governance, as the concept is thought to allow for combining decisions across levels and sectors and across institutional types at the same level (Romero-Lankao et al., 2018)⁸⁷⁷, with multilevel reinforcement and the mobilization of economic interests at different levels of governance (Jänicke and Quitzow, 2017)⁸⁷⁸. These governance mechanisms are based on accountability and transparency rules and participation and coordination across and within these levels.

A study of 29 European countries showed that the rapid adoption and diffusion of adaptation policymaking is largely driven by internal factors, at the national and sub-national levels (Massey et al., 2014)⁸⁷⁹. An assessment of national-level adaptation in 117 countries (Berrang-Ford et al., 2014)⁸⁸⁰ found good governance to be the one of the strongest predictors of national adaptation policy. An analysis of the climate responses of 200 large and medium-sized cities across eleven European countries found that factors such as membership of climate networks, population size, gross domestic product (GDP) per capita and adaptive capacity act as drivers of mitigation and adaptation plans (Reckien et al., 2015)⁸⁸¹.

Adaptation policy has seen growth in some areas (Massey et al., 2014; Lesnikowski et al., 2016)⁸⁸², although efforts to track adaptation progress are constrained by an absence of data sources on adaptation (Berrang-Ford et al., 2011; Ford and Berrang-Ford, 2016; Magnan, 2016; Magnan and Ribera, 2016)⁸⁸³. Many developing countries have made progress in formulating national policies, plans and strategies on responding to climate change. The NDCs have been identified as one such institutional mechanism (Cross-Chapter Box 11 in this Chapter) (Magnan et al., 2015; Kato and Ellis, 2016; Peters et al., 2017)⁸⁸⁴.

To overcome barriers to policy implementation, local conflicts of interest or vested interests, strong leadership and agency is needed by political leaders. As shown by the Covenant of Mayors initiative (Box 4.1), political leaders with a vision for the future of the local community can succeed in reducing GHG emissions, when they are supported by civil society (Rivas et al., 2015; Croci et al., 2017; Kona et al., 2018)⁸⁸⁵. Any political vision would need to be translated into an action plan, which could include elements describing policies and measures needed to achieve transition, the human and financial resources needed, milestones, and appropriate measurement and verification processes (Azevedo and Leal, 2017)⁸⁸⁶. Discussing the plan with stakeholders and civil society, including citizens and allowing for participation for minorities, and having them provide input and endorse it, has been found to increase the likelihood of success (Rivas et al., 2015; Wamsler, 2017)⁸⁸⁷. However, as described by Nightingale (2017)⁸⁸⁸ and Green (2016)⁸⁸⁹, struggles over natural resources and adaptation governance both at the national and community levels would also need to be addressed ‘in politically unstable contexts, where power and politics shape adaptation outcomes’.

Multilevel governance includes adaptation across local, regional, and national scales (Adger et al., 2005)⁹⁰². The whole-of-government approach to understanding and influencing climate change policy design and implementation puts analytical emphasis on how different levels of government and different types of actors (e.g., public and private) can constrain or support local adaptive capacity (Corfee-Morlot et al., 2011)⁹⁰³, including the role of the civil society. National governments, for example, have been associated with enhancing adaptive capacity through building awareness of climate impacts, encouraging economic growth, providing incentives, establishing legislative frameworks conducive to adaptation, and communicating climate change information (Berrang-Ford et al., 2014; Massey et al., 2014; Austin et al., 2015; Henstra, 2016; Massey and Huitema, 2016)⁹⁰⁴. Local governments, on the other hand, are responsible for delivering basic services and utilities to the urban population, and protecting their integrity from the impacts of extreme weather (Austin et al., 2015; Cloutier et al., 2015; Nalau et al., 2015; Araos et al., 2016b)⁹⁰⁵. National policies and transnational governance could be seen as complementary, rather than competitors, and strong national policies favour transnational engagement of sub- and non-state actors (Andonova et al., 2017)⁹⁰⁶. Local initiatives are complementary with higher level policies and can be integrated in the multilevel governance system (Fuhr et al., 2018)⁹⁰⁷.

A multilevel approach considers that adaptation planning is affected by scale mismatches between the local manifestation of climate impacts and the diverse scales at which the problem is driven (Shi et al., 2016)⁹⁰⁸. Multilevel approaches may be relevant in low-income countries where limited financial resources and human capabilities within local governments often lead to greater dependency on national governments and other (donor) organizations, to strengthen adaptation responses (Donner et al., 2016; Adenle et al., 2017)⁹⁰⁹. National governments or international organizations may motivate urban adaptation externally through broad policy directives or projects by international donors. Municipal governments on the other hand work within the city to spur progress on adaptation. Individual political leadership in municipal government, for example, has been cited as a factor driving the adaptation policies of early adapters in Quito, Ecuador, and Durban, and South Africa (Anguelovski et al., 2014)⁹¹⁰, and for adaptation more generally (Smith et al., 2009)⁹¹¹. Adaptation pathways can help identify maladaptive actions (Juhola et al., 2016; Magnan et al., 2016; Gajjar et al., 2018)⁹¹² and encourage social learning approaches across multiple levels of stakeholders in sectors such as marine biodiversity and water supply (Bosomworth et al., 2015; Butler et al., 2015; van der Brugge and Roosjen, 2015)⁹¹³.

Box 4.2 exemplifies how multilevel governance has been used for watershed management in different basins, given the impacts on water sources (Chapter 3, Section 3.4.2).

The enhancement of institutional capacity for integrated policy design and implementation has long been among the top items on the UN agenda of addressing global environmental problems and sustainable development (see Chapter 5, Section 5.5) (UNEP, 2005)⁹⁷².

Political stability, an effective regulatory and enforcement framework (e.g., institutions to impose sanctions, collect taxes and to verify building codes), access to a knowledge base and the availability of resources, would be needed at various governance levels to address a wide range of stakeholders and their concerns. The strengthening of the global response would need to support these with different interventions, in the context of sustainable development (Chapter 5, Section 5.5.1) (Pasquini et al., 2015)⁹⁷³.

Given the scale of change needed to limit warming to 1.5°C, strengthening the response capacity of relevant institutions is best addressed in ways that take advantage of existing decision-making processes in local and regional governments and within cities and communities (Romero-Lankao et al., 2013)⁹⁷⁴, and draws upon diverse knowledge sources including indigenous and local knowledge (Nakashima et al., 2012; Smith and Sharp, 2012; Mistry and Berardi, 2016; Tschakert et al., 2017)⁹⁷⁵. Examples of successful local institutional processes and the integration of local knowledge in climate-related decision-making are provided in Box 4.3 and Box 4.4.

Implementing 1.5°C-consistent strategies would require well-functioning legal frameworks to be in place, in conjunction with clearly defined mandates, rights and responsibilities to enable the institutional capacity to deliver (Romero-Lankao et al., 2013)⁹⁷⁶. As an example, current rates of urbanization occurring in cities with a lack of institutional capacity for effective land-use planning, zoning and infrastructure development result in unplanned, informal urban settlements which are vulnerable to climate impacts. It is common for 30–50% of urban populations in low-income nations to live in informal settlements with no regulatory infrastructure (Revi et al., 2014b)⁹⁷⁷. For example, in Huambo (Angola), a classified ‘urban’ area extends 20 km west of the city and is predominantly made up of ‘unplanned’ urban settlements (Smith and Jenkins, 2015)⁹⁷⁸.

Internationally, the Paris Agreement process has aimed at enhancing the capacity of decision-making institutions in developing countries to support effective implementation. These efforts are particularly reflected in Article 11 of the Paris Agreement on capacity building (the creation of the Paris Committee on Capacity Building), Article 13 (the creation of the Capacity Building Initiative on Transparency), and Article 15 on compliance (UNFCCC, 2016)⁹⁷⁹.

One of the novel features of the new climate governance architecture emerging from the 2015 Paris Agreement is the transparency framework in Article 13 committing countries, based on capacity, to provide regular progress reports on national pledges to address climate change (UNFCCC, 2016)¹⁰⁰². Many countries will rely on public policies and existing national reporting channels to deliver on their NDCs under the Paris Agreement. Scaling up the mitigation and adaptation efforts in these countries to be consistent with 1.5°C would put significant pressure on the need to develop, enhance and streamline local, national and international climate change reporting and monitoring methodologies and institutional capacity in relation to mitigation, adaptation, finance, and GHG inventories (Ford et al., 2015b; Lesnikowski et al., 2015; Schoenefeld et al., 2016)¹⁰⁰³. Consistent with this direction, the provision of the information to the stocktake under Article 14 of the Paris Agreement would contribute to enhancing reporting and transparency (UNFCCC, 2016)¹⁰⁰⁴. Nonetheless, approaches, reporting procedures, reference points, and data sources to assess progress on implementation across and within nations are still largely underdeveloped (Ford et al., 2015b; Araos et al., 2016b; Magnan and Ribera, 2016; Lesnikowski et al., 2017)¹⁰⁰⁵. The availability of independent private and public reporting and statistical institutions are integral to oversight, effective monitoring, reporting and review. The creation and enhancement of these institutions would be an important contribution to an effective transition to a low-emission world.

IPCC AR5 assessed that in order to enable a transition to a 2°C pathway, the volume of climate investments would need to be transformed along with changes in the pattern of general investment behaviour towards low emissions. The report argued that, compared to 2012, annually up to a trillion dollars in additional investment in low-emission energy and energy efficiency measures may be required until 2050 (Blanco et al., 2014; IEA, 2014a)¹⁰⁰⁶. Financing of 1.5°C would present an even greater challenge, addressing financing of both existing and new assets, which would require significant transitions to the type and structure of financial institutions as well as to the method of financing (Cochrani et al., 2014; Ma, 2014)¹⁰⁰⁷. Both public and private financial institutions would be needed to contribute to the large resource mobilization needed for 1.5°C, yet, in the ordinary course of business, these transitions may not be expected. On the one hand, private financial institutions could face scale-up risk, for example, the risks associated with commercialization and scaling up of renewable technologies to accelerate mitigation (Wilson, 2012; Hartley and Medlock, 2013)¹⁰⁰⁸ and/or price risk, such as carbon price volatility that carbon markets could face. In contrast, traditional public financial institutions are limited by both structure and instruments, while concessional financing would require taxpayer support for subsidization. Special efforts and innovative approaches would be needed to address these challenges, for example the creation of special institutions that underwrite the value of emission reductions using auctioned price floors (Bodnar et al., 2018)¹⁰⁰⁹ to deal with price volatility.

Financial institutions are equally important for adaptation. Linnerooth-Bayer and Hochrainer-Stigler (2015)¹⁰¹⁰ discussed the benefits of financial instruments in adaptation, including the provision of post-disaster finances for recovery and pre-disaster security necessary for climate adaptation and poverty reduction. Pre-disaster financial instruments and options include insurance, such as index-based weather insurance schemes, catastrophe bonds, and laws to encourage insurance purchasing. The development and enhancement of microfinance institutions to ensure social resilience and smooth transitions in the adaptation to climate change impacts could be an important local institutional innovation (Hammill et al., 2008)¹⁰¹¹.

Effective cooperative institutions and social safety nets may help address energy access and adaptation, as well as distributional impacts during the transition to 1.5°C-consistent pathways and enabling sustainable development. Not all countries have the institutional capabilities to design and manage these. Social capital for adaptation in the form of bonding, bridging, and linking social institutions has proved to be effective in dealing with climate crises at the local, regional and national levels (Aldrich et al., 2016)¹⁰¹².

The shift towards sustainable energy systems in transitioning economies could impact the livelihoods of large populations in traditional and legacy employment sectors. The transition of selected EU Member States to biofuels, for example, caused anxiety among farmers, who lacked confidence in the biofuel crop market. Enabling contracts between farmers and energy companies, involving local governments, helped create an atmosphere of confidence during the transition (McCormick and Kåberger, 2007)¹⁰¹³.

How do broader socio-economic processes influence urban vulnerabilities and thereby underpin climate change adaptation? This is a systemic challenge originating from a lack of collective societal ownership of the responsibility for climate risk management. Explanations for this situation include competing time-horizons due to self-interest of stakeholders to a more ‘rational’ conception of risk assessment, measured across a risk-tolerance spectrum (Moffatt, 2014)¹⁰¹⁴.

Self-governing and self-organ­ised institutional settings, where equipment and resource systems are commonly owned and managed, can poten­tially generate a much higher diversity of administration solutions, than other institutional arrangements, where energy technology and resource systems are either owned and administered individually in market settings or via a central authority (e.g., the state). They can also increase the adaptability of technological systems while reducing their burden on the environment (Labanca, 2017)¹⁰¹⁵. Educational, learning and awareness-building institutions can help strengthen the societal response to climate change (Butler et al., 2016; Thi Hong Phuong et al., 2017)¹⁰¹⁶.

Mitigation and adaptation behaviour is affected by many factors that shape which options are feasible and considered by individuals. Besides contextual factors (see other sub-sections in Section 4.4), these include abilities and different types of motivation to engage in behaviour.

Ability to engage in climate action. Individuals more often engage in adaptation (Gebrehiwot and van der Veen, 2015; Koerth et al., 2017)¹⁰³⁹ and mitigation behaviour (Pisano and Lubell, 2017)¹⁰⁴⁰ when they are or feel more capable to do so. Hence, it is important to enhance ability to act on climate change, which depends on income and knowledge, among other things. A higher income is related to higher CO₂ emissions; higher income groups can afford more carbon-intensive lifestyles (Lamb et al., 2014; Dietz et al., 2015; Wang et al., 2015)¹⁰⁴¹. Yet low-income groups may lack resources to invest in energy-efficient technology and refurbishments (Andrews-Speed and Ma, 2016)¹⁰⁴² and adaptation options (Wamsler, 2007; Fleming et al., 2015b; Takahashi et al., 2016)¹⁰⁴³. Adaptive capacity further depends on gender roles (Jabeen, 2014; Bunce and Ford, 2015)¹⁰⁴⁴, technical capacities and knowledge (Feola et al., 2015; Eakin et al., 2016; Singh et al., 2016b)¹⁰⁴⁵.

Knowledge of the causes and consequences of climate change and of ways to reduce GHG emissions is not always accurate (Bord et al., 2000; Whitmarsh et al., 2011; Tobler et al., 2012)¹⁰⁴⁶, which can inhibit climate actions, even when people would be motivated to act. For example, people overestimate savings from low-energy activities, and underestimate savings from high-energy activities (Attari et al., 2010)¹⁰⁴⁷. They know little about ‘embodied’ energy (i.e., energy needed to produce products; Tobler et al., 2011)¹⁰⁴⁸, including meat (de Boer et al., 2016b)¹⁰⁴⁹. Some people mistake weather for climate (Reynolds et al., 2010)¹⁰⁵⁰, or conflate climate risks with other hazards, which can inhibit adequate adaptation (Taylor et al., 2014)¹⁰⁵¹.

More knowledge on adaptation is related to higher engagement in adaptation actions in some circumstances (Bates et al., 2009; van Kasteren, 2014; Hagen et al., 2016)¹⁰⁵². How adaptation is framed in the media can influence the types of options viewed as important in different contexts (Boykoff et al., 2013; Moser, 2014; Ford and King, 2015)¹⁰⁵³.

Knowledge is important, but is often not sufficient to motivate action (Trenberth et al., 2016)¹⁰⁵⁴. Climate change knowledge and perceptions are not strongly related to mitigation actions (Hornsey et al., 2016)¹⁰⁵⁵. Direct experience of events related to climate change influences climate concerns and actions (Blennow et al., 2012; Taylor et al., 2014)¹⁰⁵⁶, more so than second-hand information (Spence et al., 2011; Myers et al., 2012; Demski et al., 2017)¹⁰⁵⁷; high impact events with low frequency are remembered more than low impact regular events (Meze-Hausken, 2004; Singh et al., 2016b; Sullivan-Wiley and Short Gianotti, 2017)¹⁰⁵⁸. Personal experience with climate hazards strengthens motivation to protect oneself (Jabeen, 2014)¹⁰⁵⁹ and enhances adaptation actions (Bryan et al., 2009; Berrang-Ford et al., 2011; Demski et al., 2017)¹⁰⁶⁰, although this does not always translate into proactive adaptation (Taylor et al., 2014)¹⁰⁶¹. Collectively constructed notions of risk and expectations of future climate variability shape risk perception and adaptation behaviour (Singh et al., 2016b)¹⁰⁶². People with particular political views and those who emphasize individual autonomy may reject climate science knowledge and believe that there is widespread scientific disagreement about climate change (Kahan, 2010; O’Neill et al., 2013)¹⁰⁶³, inhibiting support for climate policy (Ding et al., 2011; McCright et al., 2013)¹⁰⁶⁴. This may explain why extreme weather experiences enhances preparedness to reduce energy use among left- but not right-leaning voters (Ogunbode et al., 2017)¹⁰⁶⁵.

Motivation to engage in climate action. Climate actions are more strongly related to motivational factors than to knowledge, reflecting individuals’ reasons for actions, such as values, ideology and worldviews (Hornsey et al., 2016)¹⁰⁶⁶. People consider various types of costs and benefits of actions (Gölz and Hahnel, 2016)¹⁰⁶⁷ and focus on consequences that have implications for the values they find most important (Dietz et al., 2013; Hahnel et al., 2015; Steg, 2016)¹⁰⁶⁸. This implies that different individuals consider different consequences when making choices. People who strongly value protecting the environment and other people generally more strongly consider climate impacts and act more on climate change than those who strongly endorse hedonic and egoistic values (Taylor et al., 2014; Steg, 2016)¹⁰⁶⁹. People are more prone to adopt sustainable innovations when they are more open to new ideas (Jansson, 2011; Wolske et al., 2017)¹⁰⁷⁰. Further, a free-market ideology is associated with weaker climate change beliefs (McCright and Dunlap, 2011; Hornsey et al., 2016)¹⁰⁷¹, and a capital-oriented culture tends to promote activity associated with GHG emissions (Kasser et al., 2007)¹⁰⁷².

Some indigenous populations believe it is arrogant to predict the future, and some cultures have belief systems that interpret natural phenomena as sentient, where thoughts and words are believed to influence the future, with people reluctant to talk about negative future possibilities (Natcher et al., 2007; Flynn et al., 2018)¹⁰⁷³. Integrating these considerations into the design of adaptation and mitigation policy is important (Cochran et al., 2013; Chapin et al., 2016; Brugnach et al., 2017; Flynn et al., 2018)¹⁰⁷⁴.

People are more prone to act on climate change when individual benefits of actions exceed costs (Steg and Vlek, 2009; Kardooni et al., 2016; Wolske et al., 2017)¹⁰⁷⁵. For this reason, people generally prefer adoption of energy-efficient appliances above energy-consumption reductions; the latter is perceived as more costly (Poortinga et al., 2003; Steg et al., 2006)¹⁰⁷⁶, although transaction costs can inhibit the uptake of mitigation technology (Mundaca, 2007)¹⁰⁷⁷. Decentralized renewable energy systems are evaluated most favourably when they guarantee independence, autonomy, control and supply security (Ecker et al., 2017)¹⁰⁷⁸.

Besides, social costs and benefits affect climate action (Farrow et al., 2017)¹⁰⁷⁹. People engage more in climate actions when they think others expect them to do so and when others act as well (Nolan et al., 2008; Le Dang et al., 2014; Truelove et al., 2015; Rai et al., 2016)¹⁰⁸⁰, and when they experience social support (Singh et al., 2016a; Burnham and Ma, 2017; Wolske et al., 2017)¹⁰⁸¹. Discussing effective actions with peers also encourages climate action (Esham and Garforth, 2013)¹⁰⁸², particularly when individuals strongly identify with their peers (Biddau et al., 2012; Fielding and Hornsey, 2016)¹⁰⁸³. Further, individuals may engage in mitigation actions when they think doing so would enhance their reputation (Milinski et al., 2006; Noppers et al., 2014; Kastner and Stern, 2015)¹⁰⁸⁴. Such social costs and benefits can be addressed in climate policy (see Section 4.4.3.2).

Feelings affect climate action (Brosch et al., 2014)¹⁰⁸⁵. Negative feelings related to climate change can encourage adaptation action (Kerstholt et al., 2017; Zhang et al., 2017)¹⁰⁸⁶, while positive feelings associated with climate risks may inhibit protective behaviour (Lefevre et al., 2015)¹⁰⁸⁷. Individuals are more prone to engage in mitigation actions when they worry about climate change (Verplanken and Roy, 2013)¹⁰⁸⁸ and when they expect to derive positive feelings from such actions (Pelletier et al., 1998; Taufik et al., 2016)¹⁰⁸⁹.

Furthermore, collective consequences affect climate actions (Balcombe et al., 2013; Dóci and Vasileiadou, 2015; Kastner and Stern, 2015)¹⁰⁹⁰. People are motivated to see themselves as morally right, which encourages mitigation actions (Steg et al., 2015)¹⁰⁹¹, particularly when long-term goals are salient (Zaval et al., 2015)¹⁰⁹² and behavioural costs are not too high (Diekmann and Preisendörfer, 2003)¹⁰⁹³. Individuals are more prone to engage in climate actions when they believe climate change is occurring, when they are aware of threats caused by climate change and by their inaction, and when they think they can engage in actions that will reduce these threats (Esham and Garforth, 2013; Arunrat et al., 2017; Chatrchyan et al., 2017)¹⁰⁹⁴. The more individuals are concerned about climate change and aware of the negative climate impact of their behaviour, the more they feel responsible for their actions and think that their actions can help reduce such negative impacts, which can strengthen their moral norms to act accordingly (Steg and de Groot, 2010; Jakovcevic and Steg, 2013; Chen, 2015; Ray et al., 2017; Wolske et al., 2017; Woods et al., 2017)¹⁰⁹⁵. Individuals may engage in mitigation actions when they see themselves as supportive of the environment (i.e., strong environmental self-identity) (Fielding et al., 2008; van der Werff et al., 2013b; Kashima et al., 2014; Barbarossa et al., 2017)¹⁰⁹⁶; a strong environmental identity strengthens intrinsic motivation to engage in mitigation actions both at home (van der Werff et al., 2013a)¹⁰⁹⁷ and at work (Ruepert et al., 2016)¹⁰⁹⁸. Environmental self-identity is strengthened when people realize they have engaged in mitigation actions, which can in turn promote further mitigation actions (van der Werff et al., 2014b)¹⁰⁹⁹.

Individuals are less prone to engage in adaptation behaviour themselves when they rely on external measures such as government interventions (Grothmann and Reusswig, 2006; Wamsler and Brink, 2014a; Armah et al., 2015; Burnham and Ma, 2017)¹¹⁰⁰ or perceive themselves as protected by god (Gandure et al., 2013; Dang et al., 2014; Cannon, 2015)¹¹⁰¹.

Habits, heuristics and biases. Decisions are often not based on weighing costs and benefits, but on habit or automaticity, both of individuals (Aarts and Dijksterhuis, 2000; Kloeckner et al., 2003)¹¹⁰² and within organizations (Dooley, 2017)¹¹⁰³ and institutions (Munck et al., 2014)¹¹⁰⁴. When habits are strong, individuals are less perceptive of information (Verplanken et al., 1997; Aarts et al., 1998)¹¹⁰⁵ and may not consider alternatives as long as outcomes are good enough (Maréchal, 2010)¹¹⁰⁶. Habits are mostly only reconsidered when the situation changed significantly (Fujii and Kitamura, 2003; Maréchal, 2010; Verplanken and Roy, 2016)¹¹⁰⁷. Hence, strategies that create the opportunity for reflection and encourage active decisions can break habits (Steg et al., 2018)¹¹⁰⁸.

Individuals can follow heuristics, or ‘rules of thumb’, in making inferences, which demand less cognitive resources, knowledge and time than thinking through all implications of actions (Preston et al., 2013; Frederiks et al., 2015; Gillingham and Palmer, 2017)¹¹⁰⁹. For example, people tend to think that larger and more visible appliances use more energy, which is not always accurate (Cowen and Gatersleben, 2017)¹¹¹⁰. They underestimate energy used for water heating and overestimate energy used for lighting (Stern, 2014)¹¹¹¹. When facing choice overload, people may choose the easiest or first available option, which can inhibit energy-saving behaviour (Stern and Gardner, 1981; Frederiks et al., 2015)¹¹¹². As a result, individuals and firms often strive for satisficing (‘good enough’) outcomes with regard to energy decisions (Wilson and Dowlatabadi, 2007; Klotz, 2011)¹¹¹³, which can inhibit investments in energy efficiency (Decanio, 1993; Frederiks et al., 2015)¹¹¹⁴.

Biases also play a role. In Mozambique, farmers displayed omission biases (unwillingness to take adaptation actions with potentially negative consequences to avoid personal responsibility for losses), while policymakers displayed action biases (wanting to demonstrate positive action despite potential negative consequences; Patt and Schröter, 2008)¹¹¹⁵. People tend to place greater value on relative losses than gains (Kahneman, 2003)¹¹¹⁶. Perceived gains and losses depend on the reference point or status-quo (Kahneman, 2003)¹¹¹⁷. Loss aversion and the status-quo bias prevent consumers from switching electricity suppliers (Ek and Söderholm, 2008)¹¹¹⁸, to time-of-use electricity tariffs (Nicolson et al., 2017)¹¹¹⁹, and to accept new energy systems (Leijten et al., 2014)¹¹²⁰.

Owned inefficient appliances and fossil fuel-based electricity can act as endowments, increasing their value compared to alternatives (Pichert and Katsikopoulos, 2008; Dinner et al., 2011)¹¹²¹. Uncertainty and loss aversion lead consumers to undervalue future energy savings (Greene, 2011)¹¹²² and savings from energy efficient technologies (Kolstad et al., 2014)¹¹²³. Uncertainties about the performance of products and illiquidity of investments can drive consumers to postpone (profitable) energy-efficient investments (Sutherland, 1991; van Soest and Bulte, 2001)¹¹²⁴. People with a higher tendency to delay decisions may engage less in energy saving actions (Lillemo, 2014)¹¹²⁵. Training energy auditors in loss-aversion increased their clients’ investments in energy efficiency improvements (Gonzales et al., 1988)¹¹²⁶. Engagement in energy saving and renewable energy programmes can be enhanced if participation is set as a default option (Pichert and Katsikopoulos, 2008; Ölander and Thøgersen, 2014; Ebeling and Lotz, 2015)¹¹²⁷.

Policy can enable and strengthen motivation to act on climate change via top-down or bottom-up approaches, through informational campaigns, regulatory measures, financial (dis)incentives, and infrastructural and technological changes (Adger et al., 2003; Steg and Vlek, 2009; Henstra, 2016)¹¹²⁸.

Adaptation efforts tend to focus on infrastructural and technological solutions (Ford and King, 2015)¹¹²⁹ with lower emphasis on socio-cognitive and finance aspects of adaptation. For example, flooding policies in cities focus on infrastructure projects and regulation such as building codes, and hardly target individual or household behaviour (Araos et al., 2016b; Georgeson et al., 2016)¹¹³⁰.

Current mitigation policies emphasize infrastructural and technology development, regulation, financial incentives and information provision (Mundaca and Markandya, 2016)¹¹³¹ that can create conditions enabling climate action, but target only some of the many factors influencing climate actions (see Section 4.4.5.1). They fall short of their true potential if their social and psychological implications are overlooked (Stern et al., 2016a)¹¹³². For example, promising energy-saving or low-carbon technology may not be adopted or not be used as intended (Pritoni et al., 2015)¹¹³³ when people lack resources and trustworthy information (Stern, 2011; Balcombe et al., 2013)¹¹³⁴.

Financial incentives or feedback on financial savings can encourage climate action (Santos, 2008; Bolderdijk et al., 2011; Maki et al., 2016)¹¹³⁵ (see Box 4.5), but are not always effective (Delmas et al., 2013)¹¹³⁶ and can be less effective than social rewards (Handgraaf et al., 2013)¹¹³⁷ or emphasising benefits for people and the environment (Bolderdijk et al., 2013b; Asensio and Delmas, 2015; Schwartz et al., 2015)¹¹³⁸. The latter can happen when financial incentives reduce a focus on environmental considerations and weaken intrinsic motivation to engage in climate action (Evans et al., 2012; Agrawal et al., 2015; Schwartz et al., 2015)¹¹³⁹. In addition, pursuing small financial gains is perceived to be less worth the effort than pursuing equivalent CO₂ emission reductions (Bolderdijk et al., 2013b; Dogan et al., 2014)¹¹⁴⁰. Also, people may not respond to financial incentives (e.g., to improve energy efficiency) because they do not trust the organization sponsoring incentive programmes (Mundaca, 2007)¹¹⁴¹ or when it takes too much effort to receive the incentive (Stern et al., 2016a)¹¹⁴².

While providing information on the causes and consequences of climate change or on effective climate actions generally increases knowledge, it often does not encourage engagement in climate actions by individuals (Abrahamse et al., 2005; Ünal et al., 2017)¹¹⁵² or organizations (Anderson and Newell, 2004)¹¹⁵³. Similarly, media coverage on the UN Climate Summit slightly increased knowledge about the conference but did not enhance motivation to engage personally in climate protection (Brüggemann et al., 2017)¹¹⁵⁴. Fear-inducing representations of climate change may inhibit action when they make people feel helpless and overwhelmed (O’Neill and Nicholson-Cole, 2009)¹¹⁵⁵. Energy-related recommendations and feedback (e.g., via performance contracts, energy audits, smart metering) are more effective for promoting energy conservation, load shifting in electricity use and sustainable travel choices when framed in terms of losses rather than gains (Gonzales et al., 1988; Wolak, 2011; Bradley et al., 2016; Bager and Mundaca, 2017)¹¹⁵⁶.

Credible and targeted information at the point of decision can promote climate action (Stern et al., 2016a)¹¹⁵⁷. For example, communicating the impacts of climate change is more effective when provided right before adaptation decisions are taken (e.g., before the agricultural season) and when bundled with information on potential actions to ameliorate impacts, rather than just providing information on climate projections with little meaning to end users (e.g., weather forecasts, seasonal forecasts, decadal climate trends) (Dorward et al., 2015; Singh et al., 2017)¹¹⁵⁸. Similarly, heat action plans that provide early alerts and advisories combined with emergency public health measures can reduce heat-related morbidity and mortality (Benmarhnia et al., 2016)¹¹⁵⁹.

Information provision is more effective when tailored to the personal situation of individuals, demonstrating clear impacts, and resonating with individuals’ core values (Daamen et al., 2001; Abrahamse et al., 2007; Bolderdijk et al., 2013a; Dorward et al., 2015; Singh et al., 2017)¹¹⁶⁰. Tailored information prevents information overload, and people are more motivated to consider and act upon information that aligns with their core values and beliefs (Campbell and Kay, 2014; Hornsey et al., 2016)¹¹⁶¹. Also, tailored information can remove barriers to receive and interpret information faced by vulnerable groups, such as the elderly during heatwaves (Vandentorren et al., 2006; Keim, 2008)¹¹⁶². Further, prompts can be effective when they serve as reminders to perform a planned action (Osbaldiston and Schott, 2012)¹¹⁶³.

Feedback provision is generally effective in promoting mitigation behaviour within households (Abrahamse et al., 2005; Delmas et al., 2013; Karlin et al., 2015)¹¹⁶⁴ and at work (Young et al., 2015)¹¹⁶⁵, particularly when provided in real-time or immediately after the action (Abrahamse et al., 2005)¹¹⁶⁶, which makes the implications of one’s behaviour more salient (Tiefenbeck et al., 2016)¹¹⁶⁷. Simple information is more effective than detailed and technical data (Wilson and Dowlatabadi, 2007; Ek and Söderholm, 2010; Frederiks et al., 2015)¹¹⁶⁸. Energy labels (Banerjee and Solomon, 2003; Stadelmann, 2017)¹¹⁶⁹, visualization techniques (Pahl et al., 2016)¹¹⁷⁰, and ambient persuasive technology (Midden and Ham, 2012)¹¹⁷¹ can encourage mitigation actions by providing information and feedback in a format that immediately makes sense and hardly requires users’ conscious attention.

Social influence approaches that emphasize what other people do or think can encourage climate action (Clayton et al., 2015)¹¹⁷², particularly when they involve face-to-face interaction (Abrahamse and Steg, 2013)¹¹⁷³. For example, community approaches, where change is initiated from the bottom-up, can promote adaptation (see Box 4.6) and mitigation actions (Middlemiss, 2011; Seyfang and Haxeltine, 2012; Abrahamse and Steg, 2013)¹¹⁷⁴, especially when community ties are strong (Weenig and Midden, 1991)¹¹⁷⁵. Furthermore, providing social models of desired actions can encourage mitigation action (Osbaldiston and Schott, 2012; Abrahamse and Steg, 2013)¹¹⁷⁶. Social influence approaches that do not involve social interaction, such as social norm, social comparison and group feedback, are less effective, but can be easily administered on a large scale at low costs (Allcott, 2011; Abrahamse and Steg, 2013)¹¹⁷⁷.

Goal setting can promote mitigation action when goals are not set too low or too high (Loock et al., 2013)¹¹⁹². Commitment strategies where people make a pledge to engage in climate actions can encourage mitigation behaviour (Abrahamse and Steg, 2013; Lokhorst et al., 2013)¹¹⁹³, particularly when individuals also indicate how and when they will perform the relevant action and anticipate how to cope with possible barriers (i.e., implementation intentions) (Bamberg, 2000, 2002)¹¹⁹⁴. Such strategies take advantage of individuals’ desire to be consistent (Steg, 2016)¹¹⁹⁵. Similarly, hypocrisy-related strategies that make people aware of inconsistencies between their attitudes and behaviour can encourage mitigation actions (Osbaldiston and Schott, 2012)¹¹⁹⁶.

Actions that reduce climate risks can be rewarded and facilitated, while actions that increase climate risks can be punished and inhibited, and behaviour change can be voluntary (e.g., information provision) or imposed (e.g., by law); voluntary changes that involve rewards are more acceptable than imposed changes that restrict choices (Eriksson et al., 2006, 2008; Steg et al., 2006; Dietz et al., 2007)¹¹⁹⁷. Policies punishing maladaptive behaviour can increase vulnerability when they reinforce socio-economic inequalities that typically produce the maladaptive behaviour in the first place (Adger et al., 2003)¹¹⁹⁸. Change can be initiated by governments at various levels, but also by individuals, communities, profit-making organizations, trade organizations, and other non-governmental actors (Lindenberg and Steg, 2013; Robertson and Barling, 2015; Stern et al., 2016b)¹¹⁹⁹.

Strategies can target intrinsic versus extrinsic motivation. It may be particularly important to enhance intrinsic motivation so that people voluntarily engage in climate action over and again (Steg, 2016)¹²⁰⁰. Endorsement of mitigation and adaptation actions are positively related (Brügger et al., 2015; Carrico et al., 2015)¹²⁰¹; both are positively related to concern about climate change (Brügger et al., 2015)¹²⁰². Strategies that target general antecedents that affect a wide range of actions, such as values, identities, worldviews, climate change beliefs, awareness of the climate impacts of one’s actions, and feelings of responsibility to act on climate change, can encourage consistent actions on climate change (van Der Werff and Steg, 2015; Hornsey et al., 2016; Steg, 2016)¹²⁰³. Initial climate actions can lead to further commitment to climate action (Juhl et al., 2017)¹²⁰⁴, when people learn that such actions are easy and effective (Lauren et al., 2016)¹²⁰⁵, when they engaged in the initial behaviour for environmental reasons (Peters et al., 2018)¹²⁰⁶, hold strong pro-environmental values and norms (Thøgersen and Ölander, 2003)¹²⁰⁷, and when initial actions make them realise they are an environmentally sensitive person, motivating them to act on climate change in subsequent situations so as to be consistent (van der Werff et al., 2014a; Lacasse, 2015, 2016)¹²⁰⁸. Yet some studies suggest that people may feel licensed not to engage in further mitigation actions when they believe they have already done their part (Truelove et al., 2014)¹²⁰⁹.

Public acceptability can shape, enable or prevent policy and system changes. Acceptability reflects the extent to which policy or system changes are evaluated (un)favourably. Acceptability is higher when people expect more positive and less negative effects of policy and system changes (Perlaviciute and Steg, 2014; Demski et al., 2015; Drews and Van den Bergh, 2016)¹²¹⁰, including climate impacts (Schuitema et al., 2010b)¹²¹¹. Because of this, policy ‘rewarding’ climate actions is more acceptable than policy ‘punishing’ actions that increase climate risks (Steg et al., 2006; Eriksson et al., 2008)¹²¹². Pricing policy is more acceptable when revenues are earmarked for environmental purposes (Steg et al., 2006; Sælen and Kallbekken, 2011)¹²¹³ or redistributed towards those affected (Schuitema and Steg, 2008)¹²¹⁴. Acceptability can increase when people experience positive effects after a policy has been implemented (Schuitema et al., 2010a; Eliasson, 2014; Weber, 2015)¹²¹⁵; effective policy trials can thus build public support for climate policy (see Box 4.8).

Climate policy and renewable energy systems are more acceptable when people strongly value other people and the environment, or support egalitarian worldviews, left-wing or green political ideologies (Drews and Van den Bergh, 2016)¹²¹⁶, and less acceptable when people strongly endorse self-enhancement values, or support individualistic and hierarchical worldviews (Dietz et al., 2007; Perlaviciute and Steg, 2014; Drews and Van den Bergh, 2016)¹²¹⁷. Solar radiation modification is more acceptable when people strongly endorse self-enhancement values, and less acceptable when they strongly value other people and the environment (Visschers et al., 2017)¹²¹⁸. Climate policy is more acceptable when people believe climate change is real, when they are concerned about climate change (Hornsey et al., 2016)¹²¹⁹, when they think their actions may reduce climate risks, and when they feel responsible to act on climate change (Steg et al., 2005; Eriksson et al., 2006; Jakovcevic and Steg, 2013; Drews and Van den Bergh, 2016; Kim and Shin, 2017)¹²²⁰. Stronger environmental awareness is associated with a preference for governmental regulation and behaviour change rather than free-market and technological solutions (Poortinga et al., 2002)¹²²¹.

Climate policy is more acceptable when costs and benefits are distributed equally, when nature and future generations are protected (Sjöberg and Drottz-Sjöberg, 2001; Schuitema et al., 2011; Drews and Van den Bergh, 2016)¹²²², and when fair procedures have been followed, including participation by the public (Dietz, 2013; Bernauer et al., 2016a; Bidwell, 2016)¹²²³ or public society organizations (Bernauer and Gampfer, 2013)¹²²⁴. Providing benefits to compensate affected communities for losses due to policy or systems changes enhanced public acceptability in some cases (Perlaviciute and Steg, 2014)¹²²⁵, although people may disagree on what would be a worthwhile compensation (Aitken, 2010; Cass et al., 2010)¹²²⁶, or feel they are being bribed (Cass et al., 2010; Perlaviciute and Steg, 2014)¹²²⁷.

Public support is higher when individuals trust responsible parties (Perlaviciute and Steg, 2014; Drews and Van den Bergh, 2016)¹²²⁸. Yet, public support for multilateral climate policy is not higher than for unilateral policy (Bernauer and Gampfer, 2015)¹²²⁹; public support for unilateral, non-reciprocal climate policy is rather strong and robust (Bernauer et al., 2016b)¹²³⁰. Public opposition may result from a culturally valued landscape being affected by adaptation or mitigation options, such as renewable energy development (Warren et al., 2005; Devine-wright and Howes, 2010)¹²³¹ or coastal protection measures (Kimura, 2016)¹²³², particularly when people have formed strong emotional bonds with the place (Devine-Wright, 2009, 2013)¹²³³.

Climate actions may reduce human well-being when such actions involve more costs, effort or discomfort. Yet some climate actions enhance well-being, such as technology that improves daily comfort and nature-based solutions for climate adaptation (Wamsler and Brink, 2014b)¹²³⁴. Further, climate action may enhance well-being (Kasser and Sheldon, 2002; Xiao et al., 2011; Schmitt et al., 2018)¹²³⁵ because pursuing meaning by acting on climate change can make people feel good (Venhoeven et al., 2013, 2016; Taufik et al., 2015)¹²³⁶, more so than merely pursuing pleasure.

Technological systems have their own dynamics. New technologies have been described as emerging as part of a ‘socio-technical system’ that is integrated with social structures and that itself evolves over time (Geels and Schot, 2007)¹²³⁸. This progress is cumulative and accelerating (Kauffman, 2002; Arthur, 2009)¹²³⁹. To illustrate such a process of co-evolution: the progress of computer simulation enables us to better understand climate, agriculture, and material sciences, contributing to upgrading food production and quality, microscale manufacturing techniques, and leading to much faster computing technologies, resulting, for instance, in better performing photovoltaic (PV) cells.

A variety of technological developments have and will contribute to 1.5°C-consistent climate action or the lack of it. They can do this, for example, in the form of applications such as smart lighting systems, more efficient drilling techniques that make fossil fuels cheaper, or precision agriculture. As discussed in Section 4.3.1, costs of PV (IEA, 2017f)¹²⁴⁰ and batteries (Nykvist and Nilsson, 2015)¹²⁴¹ have sharply dropped. In addition, costs of fuel cells (Iguma and Kidoshi, 2015; Wei et al., 2017)¹²⁴² and shale gas and oil (Wang et al., 2014; Mills, 2015)¹²⁴³ have come down as a consequence of innovation.

Since AR5, literature has emerged as to how much future GHG emission reductions can be enabled by the rapid progress of general purpose technologies (GPTs), consisting of information and communication technologies (ICT), including artificial intelligence (AI) and the internet of things (IoT), nanotechnologies, biotechnologies, robotics, and so forth (WEF, 2015; OECD, 2017c)¹²⁴⁴. Although these may contribute to limiting warming to 1.5°C, the potential environmental, social and economic impacts of new technologies are uncertain.

Rapid improvement of performance and cost reduction is observed for many GPTs. They include AI, sensors, internet, memory storage and microelectromechanical systems. The latter GPTs are not usually categorized as climate technologies, but they can impact GHG emissions.

Progress of GPT could help reduce GHG emissions more cost-effectively. Examples are shown in Table 4.9. It may however, result in more emissions by increasing the volume of economic activities, with unintended negative consequence on sustainable development. While ICT increases electricity consumption (Aebischer and Hilty, 2015)¹²⁴⁵, the energy consumption of ICT is usually dwarfed by the energy saving by ICT (Koomey et al., 2013; Malmodin et al., 2014)¹²⁴⁶, but rebound effects and other sustainable development impacts may be significant. An appropriate policy framework that accommodates such impacts and their uncertainties could address the potential negative impacts by GPT (Jasanoff, 2007)¹²⁴⁷.

GHG emission reduction potentials in relation to GPTs were estimated for passenger cars using a combination of three emerging technologies: electric vehicles, car sharing, and self-driving. GHG emission reduction potential is reported, assuming generation of electricity with low GHG emissions (Greenblatt and Saxena, 2015; ITF, 2015; Viegas et al., 2016; Fulton et al., 2017)¹²⁴⁸. It is also possible that GHG emissions increase due to an incentive to car use. Appropriate policies such as urban planning and efficiency regulations could contain such rebound effects (Wadud et al., 2016)¹²⁴⁹.

Estimating emission reductions by GPT is difficult due to substantial uncertainties, including projections of future technological performance, costs, penetration rates, and induced human activity. Even if a technology is available, the establishment of business models might not be feasible (Linder and Williander, 2017)¹²⁵⁰. Indeed, studies show a wide range of estimates, ranging from deep emission reductions to possible increases in emissions due to the rebound effect (Larson and Zhao, 2017)¹²⁵¹.

GPT could also enable climate adaptation, in particular through more effective climate disaster risk management and improved weather forecasting.

Government policy usually plays a role in promoting or limiting GPTs, or science and technology in general. It has impacts on climate action, because the performance of further climate technologies will partly depend on the progress of GPTs. Governments have established institutions for achieving many social, and sometimes conflicting goals, including economic growth and addressing climate change (OECD, 2017c)¹²⁷³, which include investment in basic research and development (R&D) that can help develop game-changing technologies (Shayegh et al., 2017)¹²⁷⁴. Governments are also needed to create an enabling environment for the growth of scientific and technological ecosystems necessary for GPT development (Tassey, 2014)¹²⁷⁵.

While literature on 1.5°C-specific innovation policy is absent, a growing body of literature indicates that governments aim to achieve social, economic and environmental goals by promoting science and a broad range of technologies through ‘mission-driven’ innovation policies, based on differentiated national priorities (Edler and Fagerberg, 2017)¹²⁷⁶. Governments can play a role in advancing climate technology via a ‘technology push’ policy on the technology supply side (e.g., R&D subsidies), and by ‘demand pull’ policy on the demand side (e.g., energy-efficiency regulation), and these policies can be complemented by enabling environments (Somanathan et al., 2014)¹²⁷⁷. Governments may also play a role in removing existent support for incumbents (Kivimaa and Kern, 2016)¹²⁷⁸. A growing literature indicates that policy mixes, rather than single policy instruments, are more effective in addressing climate innovation challenges ranging from technologies in the R&D phase to those ready for diffusion (Veugelers, 2012; Quitzow, 2015; Rogge et al., 2017; Rosenow et al., 2017)¹²⁷⁹. Such innovation policies can help address two kinds of externalities: environmental externalities and proprietary problems (GEA, 2012; IPCC, 2014b; Mazzucato and Semieniuk, 2017)¹²⁸⁰. To avoid ‘picking winners’, governments often maintain a broad portfolio of technological options (Kverndokk and Rosendahl, 2007)¹²⁸¹ and work in close collaboration with the industrial sector and society in general. Some governments have achieved relative success in supporting innovation policies (Grubler et al., 2012; Mazzucato, 2013)¹²⁸² that addressed climate-related R&D (see Box 4.7 on bioethanol in Brazil).

Funding for R&D could come from various sources, including the general budget, energy or resource taxation, or emission trading schemes (see Section 4.4.5). Investing in climate-related R&D has as an additional benefit of building capabilities to implement climate mitigation and adaptation technologies (Ockwell et al., 2015)¹²⁹⁷. Countries regard innovation in general and climate technology specifically as a national interests issue and addressing climate change primarily as being in the global interest. Reframing part of climate policy as technology or industrial policy might therefore contribute to resolving the difficulties that continue to plague emission target negotiations  (Faehn and Isaksen, 2016; Fischer et al., 2017; Lachapelle et al., 2017)¹²⁹⁸.

Climate technology transfer to emerging economies has happened regardless of international treaties, as these countries have been keen to acquire them, and companies have an incentive to access emerging markets to remain competitive (Glachant and Dechezleprêtre, 2016)¹²⁹⁹. However, the complexity of these transfer processes is high, and they have to be conducted carefully by governments and institutions (Favretto et al., 2017)¹³⁰⁰. It is noticeable that the impact of the EU emission trading scheme (EU ETS) on innovation is contested; recent work (based on lower carbon prices than anticipated for 1.5°C-consistent pathways) indicates that it is limited (Calel and Dechezleprêtre, 2016)¹³⁰¹, but earlier assessments (Blanco et al., 2014)¹³⁰² indicate otherwise.

Technology development and transfer is recognized as an enabler of both mitigation and adaptation in Article 10 in the Paris Agreement (UNFCCC, 2016)¹³⁰³ as well as in Article 4.5 of the original text of the UNFCCC (UNFCCC, 1992)¹³⁰⁴. As previous sections have focused on technology development and diffusion, this section focuses on technology transfer. Technology transfer can adapt technologies to local circumstances, reduce financing costs, develop indigenous technology, and build capabilities to operate, maintain, adapt and innovate on technology globally (Ockwell et al., 2015; de Coninck and Sagar, 2017)¹³⁰⁵. Technology cooperation could decrease global mitigation cost, and enhance developing countries’ mitigation contributions (Huang et al., 2017a)¹³⁰⁶.

The international institutional landscape around technology development and transfer includes the UNFCCC (via its technology framework and Technology Mechanism including the Climate Technology Centre and Network (CTCN)), the United Nations (a technology facilitation mechanism for the SDGs) and a variety of non-UN multilateral and bilateral cooperation initiatives such as the Consultative Group on International Agricultural Research (CGIAR, founded in the 1970s), and numerous initiatives of companies, foundations, governments and non-governmental and academic organizations. Moreover, in 2015, twenty countries launched an initiative called ‘Mission Innovation’, seeking to double their energy R&D funding. At this point it is difficult to evaluate whether Mission Innovation achieved its objective (Sanchez and Sivaram, 2017)¹³⁰⁷. At the same time, the private sector started an innovation initiative called the ‘Breakthrough Energy Coalition’.

Most technology transfer is driven by through markets by the interests of technology seekers and technology holders, particularly in regions with well-developed institutional and technological capabilities such as developed and emerging nations (Glachant and Dechezleprêtre, 2016)¹³⁰⁸. However, the current international technology transfer landscape has gaps, in particular in reaching out to least-developed countries, where institutional and technology capabilities are limited (de Coninck and Puig, 2015; Ockwell and Byrne, 2016)¹³⁰⁹. On the one hand, literature suggests that the management or even monitoring of all these UN, bilateral, private and public initiatives may fail to lead to better results. On the other hand, it is probably more cost-effective to adopt a strategy of ‘letting a thousand flowers bloom’, by challenging and enticing researchers in the public and the private sector to direct innovation towards low-emission and adaptation options (Haselip et al., 2015)¹³¹⁰. This can be done at the same time as mission-oriented research is adopted in parallel by the scientific community (Mazzucato, 2018)¹³¹¹.

At COP 21, the UNFCCC requested the Subsidiary Body for Scientific and Technological Advice (SBSTA) to initiate the elaboration of the technology framework established under the Paris Agreement (UNFCCC, 2016)¹³¹². Among other things, the technology framework would ‘provide overarching guidance for the work of the Technology Mechanism in promoting and facilitating enhanced action on technology development and transfer in order to support the implementation of this Agreement’ (this Agreement being the Paris Agreement). An enhanced guidance issued by the Technology Executive Committee (TEC) for preparing a technology action plan (TAP) supports the new technology framework as well as the Parties’ long-term vision on technology development and transfer, reflected in the Paris Agreement (TEC, 2016)¹³¹³.

Box 4.8 shows that the average estimate by seven models of annual investment needs in the energy system is around 2.38 trillion USD₂₀₁₀ (1.38 to 3.25) between 2016 and 2035. This represents between 2.53% (1.6–4%) of the world GDP in market exchange rates (MER) and 1.7% of the world GDP in purchasing power parity (PPP). OECD investment assessments for a 2°C-consistent transition suggest that including investments in transportation and in other infrastructure would increase the investment needs by a factor of three. Other studies not included in Box 4.8, in particular by the World Economic Forum (WEF, 2013)¹³¹⁴ and the Global Commission on the Economy and Climate (GCEC, 2014)¹³¹⁵ confirm these orders of magnitude of investment.

The average increase of investment in the energy sector resulting from Box 4.8 represents a mean value of 1.5% of the total world investment compared with the baselines scenario in MER and a little over 1% in PPP. Including infrastructure investments would raise this to 2.5% and 1.7% respectively.⁹

These incremental investments could be funded through a drain on consumption (Bowen et al., 2017)¹³²⁵, which would necessitate between 0.68% and 0.45% lower global consumption than in the baseline. But, consumption at a constant savings/consumption ratio can alternatively be funded by shifting savings towards productive adaptation and mitigation investments, instead of real-estate sector and liquid financial products. This response depends upon whether it is possible to close the global investment funding gap for infrastructure that potentially inhibits growth, through structural changes in the global economy. In this case, investing more in infrastructure would not be an incremental cost in terms of development and welfare (IMF, 2014; Gurara et al., 2017)¹³²⁶

Investments in other (non-energy system) infrastructure to meet development and poverty-reduction goals can strengthen the adaptive capacity to address climate change, and are difficult to separate from overall sustainable development and poverty-alleviation investments (Hallegatte and Rozenberg, 2017)¹³²⁷. The magnitude of potential climate change damages is related to pre-existing fragility of impacted societies (Hallegatte et al., 2007)¹³²⁸. Enhancing infrastructure and service provision would lower this fragility, for example, through the provision of universal (water, sanitation, telecommunication) service access (Arezki et al., 2016)¹³²⁹.

The main challenge is thus not just a lack of mobilization of aggregate resources but of redirection of savings towards infrastructure, and the further redirection of these infrastructure investments towards low-emission options. If emission-free assets emerge fast enough to compensate for the devaluation of high-emission assets, the sum of the required incremental and redirected investments in the energy sector would (up to 2035) be equivalent to between 3.3% and 5.3% of the average annual revenues of the private capital stock (see Box 4.8) and to between 5.6% and 8.3%, including all infrastructure investments.

The interplay between mechanisms of financial intermediation and the private risk-return calculus is a major barrier to realizing these investments (Sirkis et al., 2015)¹³³⁰. This obstacle is not specific to climate mitigation investments but also affects infrastructure and has been characterised as the gap between the ‘propensity to save’ and the ‘propensity to invest’ (Summers, 2016)¹³³¹. The issue is whether new financial instruments could close this gap and inject liquidity into the low-emission transition, thereby unlocking new economic opportunities (GCEC, 2014; NCE, 2016)¹³³². By offsetting the crowding-out of other private and public investments (Pollitt and Mercure, 2017)¹³³³, the ensuing  ripple effect could reinforce growth and the sustainability of development (King, 2011; Teulings and Baldwin, 2014)¹³³⁴ and potentially trigger a new growth cycle (Stern, 2013, 2015)¹³³⁵. In this case, a massive mobilization of low-emission investments would require a significant effort but may be complementary to sustainable development investments.

This uncertain but potentially positive outcome might be constrained by the higher energy costs of low-emission options in the energy and transportation sectors. The envelope of worldwide marginal abatement costs for 1.5°C-consistent pathways reported in Chapter 2 is 135–5500 USD2010 tCO₂⁻¹ in 2030 and 245–13000 USD2010 tCO₂⁻¹ in 2050, which is between three to four times higher than for a 2°C limit.

These figures are consistent with the dramatic reduction in the unit costs of some low-emission technical options (for example solar PV, LED lighting) over the past decade (see Section 4.3.1) (OECD, 2017c)¹³³⁶. Yet there are multiple constraints to a system-wide energy transition. Lower costs of some supply- and demand-side options do not always result in a proportional decrease in energy system costs. The adoption of alternative options can be slowed down by increasing costs of decommissioning existing infrastructure, the inertia of market structures, cultural habits and risk-adverse user behaviour (see Sections 4.4.1 to 4.4.3). Learning-by-doing processes and R&D can accelerate the cost-efficiency of low-emission technology but often imply higher early-phase costs. The German energy transition resulted in high consumer prices for electricity in Germany (Kreuz and Müsgens, 2017)¹³³⁷ and needed strong accompanying measures to succeed.

One key issue is that energy costs can propagate across sectors and amplify overall production costs. During the early stage of a low-emission transition, an increase in the prices of non-energy goods could reduce consumer purchasing power and final demand. A rise in energy prices has a proportionally greater impact in developing countries that are in a catch-up phase, as they have a stronger dependence on energy-intensive sectors (Crassous et al., 2006; Luderer et al., 2012)¹³³⁸ and a higher ratio of energy to labour cost (Waisman et al., 2012)¹³³⁹. This explains why with lower carbon prices, similar emission reductions are reached in South Africa (Altieri et al., 2016)¹³⁴⁰ and Brazil (La Rovere et al., 2017a)¹³⁴¹ compared to developed countries. However, three distributional issues emerge.

First, in the absence of countervailing policies, higher energy costs have an adverse effect on the distribution of welfare (see also Chapter 5). The negative impact is inversely correlated with the level of income (Harberger, 1984; Fleurbaey and Hammond, 2004)¹³⁴² and positively correlated with the share of energy in the households budget, which is high for low- and middle-income households (Proost and Van Regemorter, 1995; Barker and Kohler, 1998; West and Williams, 2004; Chiroleu-Assouline and Fodha, 2011)¹³⁴³. Moreover, climatic conditions and the geographical conditions of human settlements matter for heating and mobility needs (see Chapter 5). Medium-income populations in the suburbs, in remote areas, and in low-density regions can be as vulnerable as residents of low-income urban areas. Poor households with low levels of energy consumption are also impacted by price increases of non-energy goods caused by the propagation of energy costs (Combet et al., 2010; Dubois, 2012)¹³⁴⁴. These impacts are generally not offset by non-market co-benefits of climate policies for the poor (Baumgärtner et al., 2017)¹³⁴⁵.

A second matter of concern is the distortion of international competition and employment implications in the case of uneven carbon constraints, especially for energy-intensive industries (Demailly and Quirion, 2008)¹³⁴⁶. Some of these industries are not highly exposed to international competition because of their very high transportation costs per unit value added (Sartor, 2013; Branger et al., 2016)¹³⁴⁷, but other industries could suffer severe shocks, generate ‘carbon leakage’ through cheaper imports from countries with lower carbon constraints (Branger and Quirion, 2014)¹³⁴⁸, and weaken the surrounding regional industrial fabric with economy-wide and employment implications.

A third challenge is the depreciation of assets whose value is based on the valuation of fossil energy resources, of which future revenues may decline precipitously with higher carbon prices (Waisman et al., 2013; Jakob and Hilaire, 2015; McGlade and Ekins, 2015)¹³⁴⁹, and on emission-intensive capital stocks (Guivarch and Hallegatte, 2011; OECD, 2015a; Pfeiffer et al., 2016)¹³⁵⁰. This raises issues of changes in industrial structure, adaptation of worker skills, and of stability of financial, insurance and social security systems. These systems are in part based on current holdings of carbon-based assets whose value might decrease by about 38% by the mid-2030s (Mercure et al., 2018)¹³⁵¹. This stranded asset challenge may be exacerbated by a decline of export revenues of fossil fuel producing countries and regions (Waisman et al., 2013; Jakob and Hilaire, 2015; McGlade and Ekins, 2015)¹³⁵².

These distributional issues, if addressed carefully and expeditiously, could affect popular sensitivity towards climate policies. Addressing them could mitigate adverse macroeconomic effects on economic growth and employment that could undermine the potential benefits of a redirection of savings and investments towards 1.5°C-consistent pathways.

Strengthening policy instruments for a low-emission transition would thus need to reconcile three objectives: (i) handling the short-term frictions inherent to this transition in an equitable way, (ii) minimizing these frictions by lowering the cost of avoided GHGs emissions, and (iii) coordinating expectations of multiple stakeholders at various decision-making levels to accelerate the decline in costs of emission reduction, efficiency and decoupling options and maximizing their co-benefits (see the practical example of lowering car use in cities in Box 4.9).

Three categories of policy tools would be available to meet the distributional challenges: carbon pricing, regulatory instruments and information and financial tools. Each of them has its own strengths and weaknesses, from a 1.5°C perspective, policy tools would have to be both scaled up and better coordinated in packages in a synergistic manner.

Economic literature has long argued that climate and energy policy grounded only in regulation, standards and public funding of R&D is at risk of being influenced by political and administrative arbitrariness, which could raise the costs of implementation. This literature has argued that it may be more efficient to make these costs explicit through carbon taxes and carbon trading, securing the abatement of emissions in places and sectors where it is cheapest (IPCC, 1995, 2001; Gupta et al., 2007; Somanathan et al., 2014)¹³⁶¹.

In a frictionless world, a uniform world carbon price could minimize the social costs of the low-carbon transition by equating the marginal costs of abatement across all sources of emissions. This implies that investors will be able to make the right choices under perfect foresight and that domestic and international compensatory transfers offset the adverse distributional impacts of higher energy prices and their consequences on economic activity. In the absence of such transfers, carbon prices would have to be differentiated by jurisdiction (Chichilnisky and Heal, 2000; Sheeran, 2006; Böhringer et al., 2009; Böhringer and Alexeeva-Talebi, 2013)¹³⁶². This differentiation could in turn raise concerns of distortions in international competition (Hourcade et al., 2001; Stavins et al., 2014)¹³⁶³.

Obstacles to enforcing a uniform world carbon price in the short run would not necessarily crowd out explicit national carbon pricing, for three reasons. First, a uniform carbon price would limit an emissions rebound resulting from a higher consumption of energy services enabled by efficiency gains, if energy prices do not change (Greening et al., 2000; Fleurbaey and Hammond, 2004; Sorrell et al., 2009; Guivarch and Hallegatte, 2011; Chitnis and Sorrell, 2015; Freire-González, 2017)¹³⁶⁴. Second, it could hedge against the arbitrariness of regulatory policies. Third, ‘revenue neutral’ recycling, at a constant share of taxes on GDP, into lowering some existing taxes would compensate for at least part of the propagation effect of higher energy costs (Stiglitz et al., 2017)¹³⁶⁵. The substitution by carbon taxes of taxes that cause distortions on the economy can counteract the regressive effect of higher energy prices. For example, offsetting increased carbon prices with lower labour taxes can potentially decrease labour costs (without affecting salaries), enhance employment and reduce the attractiveness of informal economic activity (Goulder, 2013)¹³⁶⁶.

The conditions under which an economic gain along with climate benefit (a ‘double dividend’) can be expected are well documented (Goulder, 1995; Bovenberg, 1999; Mooij, 2000)¹³⁶⁷. In the context of OECD countries, the literature examines how carbon taxation could substitute for other taxes to fund the social security system (Combet, 2013)¹³⁶⁸. The same general principles apply for countries that are building their social welfare system, such as China (Li and Wang, 2012)¹³⁶⁹ or Brazil (La Rovere et al., 2017a)¹³⁷⁰, but an optimal recycling scheme could differ based on the structure of the economy (Lefèvre et al., 2018)¹³⁷¹.

In every country the design of carbon pricing policy implies a balance between incentivizing low-carbon behaviour and mitigating the adverse distributional consequences of higher energy prices (Combet et al., 2010)¹³⁷². Carbon taxes can offset these effects if their revenues are redistributed through rebates to poor households. Other options include the reduction of value-added taxes for basic products or direct benefit transfers to enable poverty reduction (see Winkler et al. (2017)¹³⁷³ for South Africa and Grottera et al. (2016)¹³⁷⁴ for Brazil). This is possible because higher-income households pay more in absolute terms, even though their carbon tax burden is a relatively smaller share of their income (Arze del Granado et al., 2012)¹³⁷⁵.

Ultimately, the pace of increase of carbon prices would depend on the pace at which they can be embedded in a consistent set of fiscal and social policies. This is specifically critical in the context of the 1.5°C limit (Michaelowa et al., 2018)¹³⁷⁶. This is why, after a quarter century of academic debate and experimentation (see IPCC WGIII reports since the SAR), a gap persists with respect to ‘switching carbon prices’ needed to trigger rapid changes. In 2016, only 15% of global emissions are covered by carbon pricing, three-quarters of which with prices below 10 USD tCO₂⁻¹ (World Bank, 2016)¹³⁷⁷. This is too low to outweigh the ‘noise’ from the volatility of oil markets (in the range of 100 USD tCO₂⁻¹ over the past decade), of other price dynamics (interest rates, currency exchange rates and real estate prices) and of regulatory policies in energy, transportation and industry. For example, the dynamics of mobility depend upon a trade-off between housing prices and transportation costs in which the price of real estate and the inert endowments in public transport play as important a role as liquid fuel prices (Lampin et al., 2013)¹³⁷⁸.

These considerations apply to attempts to secure a minimum price in carbon trading systems (Wood and Jotzo, 2011; Fell et al., 2012; Fuss et al., 2018)¹³⁷⁹ and to the reduction of fossil fuel subsidies. Estimated at 650 billion USD in 2015 (Coady et al., 2017)¹³⁸⁰, these subsidies represent 25–30% of government expenditures in forty (mostly developing) countries (IEA, 2014b)¹³⁸¹. Reducing these subsidies would contribute to reaching 1.5°C-consistent pathways, but raises similar issues as carbon pricing around long-term benefits and short-term costs (Jakob et al., 2015; Zeng and Chen, 2016)¹³⁸², as well as social impacts.

Explicit carbon prices remain a necessary condition of ambitious climate policies, and some authors highlight the potential benefit brought by coordination among groups of countries (Weischer et al., 2012; Hermwille et al., 2017; Keohane et al., 2017)¹³⁸³. They could take the form of carbon pricing corridors (Bhattacharya et al., 2015)¹³⁸⁴. They are a necessary ‘lubricant’ through fiscal reforms or direct compensating transfers to accommodate the general equilibrium effects of higher energy prices but may not suffice to trigger the low-carbon transition because of a persistent ‘implementation gap’ between the aspirational carbon prices and those that can practically be enforced. When systemic changes, such as those needed for 1.5°C-consistent pathways, are at play on many dimensions of development, price levels ‘depend on the path and the path depends on political decisions’ (Drèze and Stern, 1990)¹³⁸⁵.

Regulatory instruments are a common tool for improving energy efficiency and enhancing renewable energy in OECD countries (e.g., the USA, Japan, Korea, Australia, the EU) and, more recently, in developing countries (M.J. Scott et al., 2015; Brown et al., 2017)¹³⁸⁶. Such instruments include constraints on the import of products banned in other countries (Knoop and Lechtenböhmer, 2017)¹³⁸⁷.

For energy efficiency, these instruments include end-use standards and labelling for domestic appliances, lighting, electric motors, water heaters and air-conditioners. They are often complemented by mandatory efficiency labels to attract consumers’ attention and stimulate the manufacture of more efficient products (Girod et al., 2017)¹³⁸⁸. Experience shows that these policy instruments are effective only if they are regularly reviewed to follow technological developments, as in the ‘Top Runner’ programme for domestic appliances in Japan (Sunikka-Blank and Iwafune, 2011)¹³⁸⁹.

In four countries, efficiency standards (e.g. miles per gallon or level of CO₂ emission per kilometre) have been used in the transport sector, for light- and heavy-duty vehicles, which have spillovers for the global car industry. In the EU (Ajanovic and Haas, 2017)¹³⁹⁰ and the USA (Sen et al., 2017)¹³⁹¹, vehicle manufacturers need to meet an annual CO₂ emission target for their entire new vehicle fleet. This allows them to compensate through the introduction of low-emission vehicles for the high-emission ones in the fleet. This leads to increasingly efficient fleets of vehicles over time but does not necessarily limit the driven distance.

Building codes that prescribe efficiency requirements for new and existing buildings have been adopted in many OECD countries (Evans et al., 2017)¹³⁹² and are regularly revised to increase their efficiency per unit of floor space. Building codes can avoid locking rapidly urbanizing countries into poorly performing buildings that remain in use for the next 50–100 years (Ürge-Vorsatz et al., 2014)¹³⁹³. In OECD countries, however, their main role is to incentivize the retrofit of existing buildings. In addition of the convergence of these codes to net zero energy buildings (D’Agostino, 2015)¹³⁹⁴, a new focus should be placed, in the context of 1.5°C-consistent pathways, on public and private coordination to achieve better integration of building policies with the promotion of low-emission transportation modes (Bertoldi, 2017)¹³⁹⁵.

The efficacy of regulatory instruments can be reinforced by economic incentives, such as feed-in tariffs based on the quantity of renewable energy produced, subsidies or tax exemptions for energy savings (Bertoldi et al., 2013; Ritzenhofen and Spinler, 2016; García-Álvarez et al., 2017; Pablo-Romero et al., 2017)¹³⁹⁶, fee-bates, and ‘bonus-malus’ that foster the penetration of low-emission options (Butler and Neuhoff, 2008)¹³⁹⁷. Economic incentives can also be combined with direct-use market-based instruments, for example combining, in the United States and, in some EU countries, carbon trading schemes with energy savings obligations for energy retailers (Haoqi et al., 2017)¹³⁹⁸, or with green certificates for renewable energy portfolio standards (Upton and Snyder, 2017)¹³⁹⁹. Scholars have investigated caps on utilities’ energy sales (Thomas et al., 2017)¹⁴⁰⁰ and emission caps implemented at a personal level (Fawcett et al., 2010)¹⁴⁰¹.

In combination with the funding of public research institutes, grants or subsidies also support R&D, where risk and the uncertainty about long-term perspectives can reduce the private sector’s willingness to invest in low-emission innovation (see also Section 4.4.4). Subsidies can take the form of rebates on value-added tax (VAT), of direct support to investments (e.g., renewable energy or refurbishment of buildings) or feed-in tariffs (Mir-Artigues and del Río, 2014)¹⁴⁰². They can be provided by the public budget, via consumption levies, or via the revenues of carbon taxes or pricing. Fee-bates, introduced in some countries (e.g., for cars), have had a neutral impact on public budgets by incentivizing low-emission products and penalizing high-emission ones (de Haan et al., 2009)¹⁴⁰³.

All policy instruments can benefit from information campaigns (e.g., TV ads) tailored to specific end-users. A vast majority of public campaigns on energy and climate have been delivered through mass-media channels and advertising-based approaches (Corner and Randall, 2011; Doyle, 2011)¹⁴⁰⁴. Although some authors report large savings obtained by such campaigns, most agree that the effects are short-lived and decrease over time (Bertoldi et al., 2016)¹⁴⁰⁵. Recently, focus has been placed on the use of social norms to motivate behavioural changes (Allcott, 2011; Alló and Loureiro, 2014)¹⁴⁰⁶. More on strategies to change behaviour can be found in Section 4.4.3.

The redirection of savings towards low-emission investments may be constrained by enforceable carbon prices, implementation of technical standards and the short-term bias of financial systems (Miles, 1993; Bushee, 2001; Black and Fraser, 2002)¹⁴⁰⁷. The many causes of this bias are extensively analysed in economic literature (Tehranian and Waegelein, 1985; Shleifer and Vishny, 1990; Bikhchandani and Sharma, 2000)¹⁴⁰⁸, including their link with prevailing patterns of economic globalization (Krugman, 2009; Rajan, 2011)¹⁴⁰⁹ and the chronic underinvestment in long-term infrastructure (IMF, 2014)¹⁴¹⁰. Emerging literature explores how to overcome this through reforms targeted to bridge the gap between short-term cash balances and long-term low-emission assets and to reduce the risk-weighted capital costs of climate-resilient investments. This gap, which was qualified by the Governor of the Bank of England as a ‘tragedy of the horizon’ (Carney, 2016)¹⁴¹¹ that constitutes a threat to the stability of the financial system, is confirmed by the literature (Arezki et al., 2016; Christophers, 2017)¹⁴¹². This potential threat would encompass the impact of climate events on the value of assets (Battiston et al., 2017)¹⁴¹³, liability risks (Heede, 2014)¹⁴¹⁴ and the transition risk due to devaluation of certain classes of assets (Platinga and Scholtens, 2016)¹⁴¹⁵.

The financial community’s attention to climate change grew after COP 15 (ESRB ASC, 2016)¹⁴¹⁶. This led to the introduction of climate-related risk disclosure in financial portfolios (UNEP, 2015)¹⁴¹⁷, placing it on the agenda of G20 Green Finance Study Group and of the Financial Stability Board. This led to the creation of low-carbon financial indices that investors could consider as a ‘free option on carbon’ to hedge against risks of stranded carbon-intensive assets (Andersson et al., 2016)¹⁴¹⁸. This could also accelerate the emergence of climate-friendly financial products such as green or climate bonds. The estimated value of the green bonds market in 2017 is 155 billion USD (BNEF 2018)¹⁴¹⁹. The bulk of these investments are in renewable energy, energy efficiency and low-emission transport (Lazurko and Venema, 2017)¹⁴²⁰, with only 4% for adaptation (OECD, 2017b)¹⁴²¹. One major question is whether individual strategies based on improved climate-related information alone will enable the financial system to allocate capital in an optimal way (Christophers, 2017)¹⁴²² since climate change is a systemic risk (CISL, 2015; Schoenmaker and van Tilburg, 2016)¹⁴²³.

The readiness of financial actors to reduce investments in fossil fuels is a real trend (Platinga and Scholtens, 2016; Ayling and Gunningham, 2017)¹⁴²⁴, but they may not resist the attractiveness of carbon-intensive investments in many regions. Hence, decarbonizing an investment portfolio is not synonymous with investing massively in low-emission infrastructure. Scaling up climate-friendly financial products may depend upon a business context conducive to the reduction of the risk-weighted capital costs of low-emission projects. The typical leverage of public funding mechanisms for low-emission investment is low (2 to 4) compared with other sectors (10 to 15) (Maclean et al., 2008; Ward et al., 2009; MDB, 2016)¹⁴²⁵. This is due to the interplay of the uncertainty of emerging low-emission technologies in the midst of their learning-by-doing cycle with uncertain future revenues due to volatility of fossil fuel prices (Roques et al., 2008; Gross et al., 2010)¹⁴²⁶ as well as uncertainty around regulatory policies. This inhibits low-emission investments by corporations functioning under a ‘shareholder value business regime’ (Berle and Means, 1932; Roe, 1996; Froud et al., 2000)¹⁴²⁷ and actors with restricted access to capital (e.g. cities, local authorities, SMEs and households).

De-risking policy instruments to enable low-emission investment encompasses interest rate subsidies, fee-bates, tax breaks, concessional loans from development banks, and public investment funds, including revolving funds. Given the constraints on public budgets, public guarantees can be used to increase the leverage effect of public financing on private financing. Such de-risking instruments imply indeed a full direct burden on public budgets only in case of default of the project. They could back for example various forms of green infrastructure funds (de Gouvello and Zelenko, 2010; Emin et al., 2014; Studart and Gallagher, 2015)¹⁴²⁸.¹⁰

The risk of defaulting can be mitigated by strong measurement, reporting and verifying (MRV) systems (Bellassen et al., 2015)¹⁴²⁹ and by the use of notional prices recommended in public economics (and currently in use in France and the UK) to calibrate public support to the provision of public goods in case of persisting distortions in pricing (Stiglitz et al., 2017)¹⁴³⁰. Some suggest linking these notional prices to ‘social, economic and environmental value of voluntary mitigation actions’ recognized by the COP 21 Decision accompanying the Paris Agreement (paragraph 108) (Hourcade et al., 2015; La Rovere et al., 2017b; Shukla et al., 2017)¹⁴³¹, in order to incorporate the co-benefits of mitigation.

Such public guarantees ultimately amount to money issuance backed by low-emission projects as collateral. This explains the potentially strong link between global climate finance and the evolution of the financial and monetary system. Amongst suggested mechanisms for this evolution are the use of International Monetary Fund’s (IMF’s) Special Drawing Rights to fund the paid-in capital of the Green Climate Fund (Bredenkamp and Pattillo, 2010)¹⁴³² and the creation of carbon remediation assets at a predetermined face value per avoided tonne of emissions (Aglietta et al., 2015a, b)¹⁴³³. Such a predetermined value could hedge against the fragmentation of climate finance initiatives and support the emergence of financial products backed by a new class of long-term assets.

Combining public guarantees at a predetermined value of avoided emissions, in addition to improving the consistency of non-price measures, could support the emergence of financial products backed by a new class of certified assets to attract savers in search of safe and ethical investments (Aglietta et al., 2015b)¹⁴³⁴. It could hedge against the fragmentation of climate finance initiatives and provide a mechanism to compensate for the ‘stranded’ assets caused by divestment in carbon-based activities and in lowering the systemic risk of stranded assets (Safarzyńska and van den Bergh, 2017)¹⁴³⁵. These new assets could also facilitate a low-carbon transition for fossil fuel producers and help them to overcome the ‘resource curse’ (Ross, 2015; Venables, 2016)¹⁴³⁶.

Blended injection of liquidity has monetary implications. Some argue that this questions the premise that money should remain neutral (Annicchiarico and Di Dio, 2015, 2016; Nikiforos and Zezza, 2017)¹⁴³⁷. Central banks or financial regulators could act as a facilitator of last resort for low-emission financing instruments, which could in turn lower the systemic risk of stranded assets (Safarzyńska and van den Bergh, 2017)¹⁴³⁸. This may, in time, lead to the use of carbon-based monetary instruments to diversify reserve currencies (Jaeger et al., 2013)¹⁴³⁹ and differentiate reserve requirements (Rozenberg et al., 2013)¹⁴⁴⁰ in the context of a climate-friendly Bretton Woods (Sirkis et al., 2015; Stua, 2017)¹⁴⁴¹.

Adaptation finance is difficult to quantify for two reasons. The first is that it is very difficult to isolate specific investment needs to enhance climate resilience from the provision of basic infrastructure that are currently underinvested (IMF, 2014; Gurara et al., 2017)¹⁴⁴². The UNEP (2016)¹⁴⁴³ estimate of investment needs on adaptation in developing countries between 140–300 billion USD yr⁻¹ in 2030, a major part being investment expenditures that are complementary with SDG-related investments focused on universal access to infrastructure and services and meeting basic needs. Many climate-adaptation-centric financial incentives are relevant to non-market services, offering fewer opportunities for market revenues while they contribute to creating resilience to climate impacts.

Hence, adaptation investments and the provision of basic needs would typically have to be supported by national and sub-national government budgets together with support from overseas development assistance and multilateral development banks (Fankhauser and Schmidt-Traub, 2011; Adenle et al., 2017; Robinson and Dornan, 2017)¹⁴⁴⁴, and a slow increase of dedicated NGO and private climate funds (Nakhooda and Watson, 2016)¹⁴⁴⁵. Even though the UNEP estimates of the costs of adaptation might be lower in a 1.5°C world (UNEP/Climate Analytics, 2015)¹⁴⁴⁶ they would be higher than the UNEP estimate of 22.5 billion USD of bilateral and multilateral funding for climate change adaptation in 2014. Currently, 18–25% of climate finance flows to adaptation in developing countries (OECD, 2015b, 2016; Shine and Campillo, 2016)¹⁴⁴⁷. It remains fragmented, with small proportions flowing through UNFCCC channels (AdaptationWatch, 2015; Roberts and Weikmans, 2017)¹⁴⁴⁸.

Means of raising resources for adaptation, achieving the SDGs and meeting basic needs (Durand et al., 2016; Roberts et al., 2017)¹⁴⁴⁹ include the reduction of fossil fuel subsidies (Jakob et al., 2016)¹⁴⁵⁰, increasing revenues from carbon taxes (Jakob et al., 2016)¹⁴⁵¹, levies on international aviation and maritime transport, and sharing of the proceeds of financial arrangements supporting mitigation activities (Keen et al., 2013)¹⁴⁵². Each have different redistribution implications. Challenges, however, include the efficient use of resources, the emergence of long-term assets using infrastructure as collateral and the capacity to implement small-scale adaptation and the mainstreaming of adaptation in overall development policies. There is thus a need for greater policy coordination (Fankhauser and McDermott, 2014; Morita and Matsumoto, 2015; Sovacool et al., 2015, 2017; Lemos et al., 2016; Adenle et al., 2017; Peake and Ekins, 2017)¹⁴⁵³ that includes robust mechanisms for tracking, reporting and ensuring transparency of adaptation finance (Donner et al., 2016; Pauw et al., 2016a; Roberts and Weikmans, 2017; Trabacchi and Buchner, 2017)¹⁴⁵⁴ and its consistency with the provision of basic needs (Hallegatte et al., 2016)¹⁴⁵⁵.

Carbon prices, regulation and standards, improved information and appropriate financial instruments can work synergistically to meet the challenge of ‘making finance flows consistent with a pathway towards low greenhouse gas emissions and climate-resilient development’, as in Article 2 in the Paris Agreement.

There is growing attention to the combination of policy instruments that address three domains of action: behavioural changes, economic optimization and long-term strategies (Grubb et al., 2014)¹⁴⁵⁶. For example, de-risking low-emission investments would result in higher volumes of low-emission investments, and would in turn lead to a lower switching price for the same climate ambition (Hirth and Steckel, 2016)¹⁴⁵⁷. In the reverse direction, higher explicit carbon prices may generate more low-emission projects for a given quantum of de-risking. For example, efficiency standards for housing can increase the efficacy of carbon prices and overcome the barriers coming from the high discount rates used by households (Parry et al., 2014)¹⁴⁵⁸, while explicit and notional carbon prices can lower the risk of arbitrary standards. The calibration of innovative financial instruments to notional carbon prices could encourage large multinational companies to increase their level of internal carbon prices (UNEP, 2016)¹⁴⁵⁹. These notional prices could be higher than explicit carbon prices because they redirect new hardware investments without an immediate impact on existing capital stocks and associated interests.

Literature, however, shows that conflicts between poorly articulated policy instruments can undermine their efficiency (Lecuyer and Quirion, 2013; Bhattacharya et al., 2017; García-Álvarez et al., 2017)¹⁴⁶⁰. As has been illustrated in Europe, commitment uncertainty and lack of credibility of regulation have consistently led to low carbon prices in the case of the EU Emission Trading System (Koch et al., 2014, 2016)¹⁴⁶¹. A comparative study shows how these conflicts can be avoided by policy packages that integrate many dimensions of public policies and are designed to match institutional and social context of each country and region (Bataille et al., 2015)¹⁴⁶².

Even though policy packages depend upon domestic political processes, they might not reinforce the NDCs at a level consistent with the 1.5°C transition without a conducive international setting where international development finance plays a critical role. Section 4.4.1 explores the means of mainstreaming climate finance in the current evolution of the lending practices of national and multilateral banks (Badré, 2018)¹⁴⁶³. This could facilitate the access of developing countries to loans via bond markets at low interest rates, encouragement of the emergence of new business models for infrastructure, and encouragement of  financial markets to support small-scale investments (Déau and Touati, 2017)¹⁴⁶⁴.

These financial innovations may involve non-state public actors like cities and regional public authorities that govern infrastructure investment, enable energy and food systems transitions and manage urban dynamics (Cartwright, 2015)¹⁴⁶⁵. They would help, for example, in raising the 4.5–5.4 trillion USD yr⁻¹ from 2015 to 2030 announced by the Cities Climate Finance Leadership Alliance (CCFLA, 2016)¹⁴⁶⁶ to achieve the commitments by the Covenant of Mayors of many cities to long-term climate targets (Kona et al., 2018)¹⁴⁶⁷.

The evolution of global climate financial cooperation may involve central banks, financial regulatory authorities, and multilateral and commercial banks. There are still knowledge gaps about the form, structure and potential of these arrangements. They could be viewed as a form of a burden-sharing between high-, medium- and low-income countries to enhance the deployment of ambitious Nationally Determined Contributions (NDCs) and new forms of ‘common but differentiated responsibility and respective capabilities’ (Edenhofer et al., 2015; Hourcade et al., 2015; Ji and Sha, 2015)¹⁴⁶⁸.

An assessment of the degree to which examples of 1.5°C-relevant mitigation options face barriers to implementation, and on which contexts this depends, is summarized in Table 4.11. An explanation of the approach is given in Section 4.5.1 and in supplementary material 4.SM.4.1. Selected options were mapped onto system transitions and clustered through an iterative process of literature review, expert feedback, and responses to reviewer comments. The detailed assessment and the literature underpinning the assessment can be found in supplementary material 4.SM.4.2.

The feasibility framework in Cross-Chapter Box 3 in Chapter 1 highlights that the feasibility of mitigation and adaptation options depends on many factors. Many of those are captured in the indicators in Table 4.10, but many depend on the specific context in which an option features. This Special Report did not have the mandate, space or the literature base to undertake a regionally specific assessment. Hence the assessment is caveated as providing a broad indication of the likely global barriers, ignoring significant regional diversity. Regional and context-specific literature is also just emerging as is noted in the knowledge gaps section (Section 4.6). Nevertheless, in Table 4.11, an indicative attempt has been made to capture relevant contextual information. The ‘context’ column indicates which contextual factors may affect the feasibility of an option, including regional differences. For instance, solar irradiation in an area impacts the cost-effectiveness of solar photovoltaic energy, so solar irradiation is mentioned in this column.

The feasibility assessment highlights six dimensions that could help inform an agenda that could be addressed by the areas discussed in Section 4.4: governance, behaviour and lifestyles, innovation, enhancing institutional capacities, policy and finance. For instance, Section 4.4.3 on behaviour offers strategies for addressing public acceptance problems, and how changes can be more effective when communication and actions relate to people’s values. This section synthesizes the findings in Section 4.4 in an attempt to link them to the assessment in Table 4.11. The literature on which the discussion is based is found in Section 4.4.

From Section 4.4, including the case studies presented in the Boxes 4.1 to 4.10, several main messages can be constructed. For instance, governance would have to be multilevel and engaging different actors, while being efficient, and choosing the form of cooperation based on the specific systemic challenge or option at hand. If institutional capacity for financing and governing the various transitions is not urgently built, many countries would lack the ability to change pathways from a high-emission scenario to a low- or zero-emission scenario. In terms of innovation, governments, both national and multilateral, can contribute to applying general-purpose technologies to mitigation purposes. If this is not managed, some reduction in emissions could happen autonomously, but it may not lead to a 1.5ºC-consistent pathway. International cooperation on technology, including technology transfer where this does not happen autonomously, is needed and can help create innovation capabilities in all countries that allow them to operate, maintain, adapt and regulate a portfolio of mitigation technologies. Case studies in the various subsections highlight the opportunities and challenges of doing this in practice. They indicate that it can be done in specific circumstances, which can be changed.

A combination of behaviour-oriented pricing policies and financing options can help change technologies and social behaviour as it would challenge the existing, high-emission socio-technical regime on multiple levels across feasibility characteristics. For instance, for dietary change, combining supply-side measures with value-driven communication and economic instruments may help make a lasting transition, while an economic instrument, such as enhanced prices or taxation, on its own may not be as robust.

Governments could benefit from enhanced carbon prices, as a price and innovation incentive and also a source of additional revenue to correct distributional effects and subsidize the development of new, cost-effective negative-emission technology and infrastructure. However, there is high evidence and medium agreement that pricing alone is insufficient. Even if prices rise significantly, they typically incentivize incremental change, but typically fail to provide the impetus for private actors to take the risk of engaging in the transformational changes that would be needed to limit warming to 1.5ºC. Apart from the incentives to change behaviour and technology, financial systems are an indispensable element of a systemic transition. If financial markets do not acknowledge climate risk and the risk of transitions, regulatory financial institutions, such as central banks, could intervene.

Strengthening implementation revolves around more than addressing barriers to feasibility. A system transition, be it in energy, industry, land or a city, requires changing the core parameters of a system. These relate, as introduced in Section 4.2 and further elaborated in Section 4.4, to how actors cooperate, how technologies are embedded, how resources are linked, how cultures relate and what values people associate with the transition and the current regime.

This section summarizes the feasibility (defined in Cross-Chapter Box 3, Table 1 in Chapter 1 and Table 4.4) of select adaptation options using evidence presented across this chapter and in supplementary material 4.SM.4.3 and the expert-judgement of its authors (Table 4.12). The options assessed respond to risks and impacts identified in Chapter 3. They were selected based on options identified in AR5 (Noble et al., 2014)¹⁴⁷², focusing on those relevant to 1.5°C-compatible pathways, where sufficient literature exists. Selected options were mapped onto system transitions and clustered through an iterative process of literature review, expert feedback, and responses to reviewer comments.

Besides gaps in the literature around crucial adaptation questions on the transition to a 1.5°C world (Section 4.6), there is inadequate current literature to undertake a spatially differentiated assessment (Cross-Chapter Box 3 in Chapter 1). There are also limited baselines for exposure, vulnerability and risk to help policy and implementation prioritization. Hence, the compiled results can at best provide a broad framework to inform policymaking. Given the bottom-up nature of most adaptation implementation evidence, care needs to be taken in generalizing these findings.

Options are considered as part of a systemic approach, recognizing that no single solution exists to limit warming to 1.5°C and adapting to its impacts. To respond to the local and regional context – and to synergies and trade-offs between adaptation, mitigation and sustainable development – packages of options suited to local enabling conditions can be implemented.

Table 4.12 summarizes the feasibility assessment through its six dimensions with levels of evidence and agreement and indicates how the feasibility of an adaptation option may be differentiated by certain contextual factors (last column).

When considered jointly, the description of adaptation options (Section 4.3), the feasibility assessment (summarized in Table 4.12), and discussion of enabling conditions (Section 4.4) show us how options can be implemented and lead towards transformational adaptation if and when needed.

The adaptation options for energy system transitions focus on existing power infrastructure resilience and water management, when required, for any type of generation source. These options are not sufficient for the far-reaching transformations required in the energy sector, which have tended to focus on technologies to shift from a fossil-based to a renewable energy system (Erlinghagen and Markard, 2012; Muench et al., 2014; Brand and von Gleich, 2015; Monstadt and Wolff, 2015; Child and Breyer, 2017; Hermwille et al., 2017)¹⁴⁷³. There is also need for integration of such energy system transitions with social-ecological systems transformations to increase the resilience of the energy sector, for which appropriate enabling conditions, such as for technological innovations, are fundamentally important. Institutional capacities can be enhanced by expanding the role of actors as transformation catalysts (Erlinghagen and Markard, 2012)¹⁴⁷⁴. The integration of ethics and justice within these transformations can help attain SDG 7 on clean energy access (Jenkins et al., 2018)¹⁴⁷⁵, while inclusion of the cultural dimension and cultural legitimacy (Amars et al., 2017)¹⁴⁷⁶ can provide a more substantial base for societal transformation. Strengthening policy instruments and regulatory frameworks and enhancing multilevel governance that focuses on resilience components can help secure these transitions (Exner et al., 2016)¹⁴⁷⁷.

For land and ecosystem transitions, the options of conservation agriculture, efficient irrigation, agroforestry, ecosystem restoration and avoided deforestation, and coastal defence and hardening have between medium and robust evidence with medium to high agreement. The other options assessed have limited or no evidence across one or more of the feasibility dimensions. Community-based adaptation is assessed as having medium evidence and high agreement to face scaling barriers. Scaling community-based adaptation may require  structural changes, implying the need for transformational adaptation in some regions. This would involve enhanced multilevel governance and institutional capacities by enabling anticipatory and flexible decision-making systems that access and develop collaborative networks (Dowd et al., 2014)¹⁴⁷⁸, tackling root causes of vulnerability (Chung Tiam Fook, 2017)¹⁴⁷⁹, and developing synergies between development and climate change (Burch et al., 2017)¹⁴⁸⁰. Case studies show the use of transformational adaptation approaches for fire management (Colloff et al., 2016a)¹⁴⁸¹, floodplain and wetland management (Colloff et al., 2016b)¹⁴⁸², and forest management (Chung Tiam Fook, 2017)¹⁴⁸³, in which the strengthening of policy instruments and climate finance are also required.

There is growing recognition of the need for transformational adaptation within the agricultural sector but limited evidence on how to facilitate processes of deep, systemic change (Dowd et al., 2014)¹⁴⁸⁴. Case studies demonstrate that transformational adaptation in agriculture requires a sequencing and overlap between incremental and transformational adaptation actions (Hadarits et al., 2017; Termeer et al., 2017)¹⁴⁸⁵, e.g., incremental improvements to crop management while new crop varieties are being researched and field-tested (Rippke et al., 2016)¹⁴⁸⁶. Broader considerations include addressing stakeholder values and attitudes (Fleming et al., 2015a)¹⁴⁸⁷, understanding and leveraging the role of social capital, collaborative networks, and information (Dowd et al., 2014)¹⁴⁸⁸, and being inclusive with rural and urban communities, and the social, political, and cultural environment (Rickards and Howden, 2012)¹⁴⁸⁹. Transformational adaptation in agriculture systems could have significant economic and institutional costs (Mushtaq, 2016)¹⁴⁹⁰, along with potential unintended negative consequences (Davidson, 2016; Rippke et al., 2016; Gajjar et al., 2018; Mushtaq, 2018)¹⁴⁹¹,  and a need to focus on the transitional space between incremental and transformational adaptation (Hadarits et al., 2017)¹⁴⁹², as well as the timing of the shift from one to the other (Läderach et al., 2017)¹⁴⁹³.

Within urban and infrastructure transitions, green infrastructure and sustainable water management are assessed as the most feasible options, followed by sustainable land-use and urban planning. The need for transformational adaptation in urban settings arises from the root causes of poverty, failures in sustainable development, and a lack of focus on social justice (Revi et al., 2014a; Parnell, 2015; Simon and Leck, 2015; Shi et al., 2016; Ziervogel et al., 2016a; Burch et al., 2017)¹⁴⁹⁴, and necessitates a focus on governance structures and the inclusion of equity and justice concerns (Bos et al., 2015; Shi et al., 2016; Hölscher et al., 2018)¹⁴⁹⁵.

Current implementation of urban ecosystems-based adaptation (EbA) lacks a systems perspective of transformations and consideration of the normative and ethical aspects of EbA (Brink et al., 2016)¹⁴⁹⁶. Flexibility within urban planning could help deal with the multiple uncertainties of implementing adaptation (Rosenzweig and Solecki, 2014; Radhakrishnan et al., 2018)¹⁴⁹⁷, for example, urban adaptation pathways were implemented in the aftermath of Superstorm Sandy in New York, which is considered as tipping point that led to the implementation of transformational adaptation practices.

Adaptation options for industry focus on infrastructure resilience and water management. Like with energy system transitions, technological innovation would be required, but also the enhancement of institutional capacities. Recent research illustrates transformational adaptation within industrial transitions focusing on the role of different actors and tools driving innovation, and points to the role of nationally appropriate mitigation actions in avoiding lock-ins and promoting system innovation (Boodoo and Olsen, 2017)¹⁴⁹⁸, the role of private sector in sustainability governance in the socio-political context (Burch et al., 2016)¹⁴⁹⁹, and of green entrepreneurs driving transformative change in the green economy (Gibbs and O’Neill, 2014)¹⁵⁰⁰. Lim-Camacho et al. (2015)¹⁵⁰¹ suggest an analysis of the complete lifecycle of supply chains as a means of identifying additional adaptation strategies, as opposed to the current focus on a part of the supply chain. Chain-wide strategies can modify the rest of the chain and present a win-win with commercial objectives.

The assessed adaptation options also have mitigation synergies and trade-offs (assessed in Section 4.5.4) that need to be carefully considered, while planning climate action.

Monitoring and evaluation (M&E) in adaptation implementation can promote accountability and transparency of adaptation financing, facilitate policy learning and sharing good practices, pressure laggards, and guide adaptation planning. The majority of research on M&E focuses on specific policies or programmes, and has typically been driven by the needs of development organizations, donors, and governments to measure the impact and attribution of adaptation initiatives (Ford and Berrang-Ford, 2016)¹⁵⁰². There is growing research examining adaptation progress across nations, sectors, and scales (Reckien et al., 2014; Araos et al., 2016a, b; Austin et al., 2016; Heidrich et al., 2016; Lesnikowski et al., 2016; Robinson, 2017)¹⁵⁰³. In response to a need for global, regional and local adaptation, the development of indicators and standardized approaches to evaluate and compare adaptation over time and across regions, countries, and sectors would enhance comparability and learning. A number of constraints continue to hamper progress on adaptation M&E, including a debate on what actually constitutes adaptation for the purposes of assessing progress (Dupuis and Biesbroek, 2013; Biesbroek et al., 2015)¹⁵⁰⁴, an absence of comprehensive and systematically collected data on adaptation to support longitudinal assessment and comparison (Ford et al., 2015b; Lesnikowski et al., 2016)¹⁵⁰⁵, a lack of agreement on indicators to measure (Brooks et al., 2013; Bours et al., 2015; Lesnikowski et al., 2015)¹⁵⁰⁶, and challenges of attributing altered vulnerability to adaptation actions (Ford et al., 2013; Bours et al., 2015; UNEP, 2017a)¹⁵⁰⁷.