Editing IPCC:AR6/SR15/Chapter-4 (section)

=== 4.3.7 Carbon Dioxide Removal (CDR) ===

<div id="section-4-3-7-block-1"></div>

CDR methods refer to a set of techniques for removing CO <sub>2</sub> from the atmosphere. In the context of 1.5°C-consistent pathways (Chapter 2), they serve to offset residual emissions and, in most cases, achieve net negative emissions to return to 1.5°C from an overshoot. See Cross-Chapter Box 7 in Chapter 3 for a synthesis of land-based CDR options. Cross-cutting issues and uncertainties are summarized in Table 4.6.

<div id="section-4-3-3-7-1"></div>

<span id="bioenergy-with-carbon-capture-and-storage-beccs"></span>
==== 4.3.7.1 Bioenergy with carbon capture and storage (BECCS) ====

<div id="section-4-3-3-7-1-block-1"></div>

BECCS has been assessed in previous IPCC reports (IPCC, 2005b, 2014b; P. Smith et al., 2014; Minx et al., 2017) <sup>[[#fn:r600|600]]</sup> and has been incorporated into integrated assessment models (Clarke et al., 2014) <sup>[[#fn:r601|601]]</sup> 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) <sup>[[#fn:r602|602]]</sup> . 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 GtCO <sub>2</sub> yr <sup>−1</sup> 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 CO <sub>2</sub> (Section 4.3.1.6). Literature estimates for BECCS mitigation potentials in 2050 range from 1–85 GtCO <sup>[[#fn:4|4]]</sup> . Fuss et al. (2018) <sup>[[#fn:r603|603]]</sup> narrow this range to 0.5–5 GtCO <sub>2</sub> yr <sup>−1</sup> ( ''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) <sup>[[#fn:r604|604]]</sup> .

Assessing BECCS deployment in 2°C pathways (of about 12 GtCO <sub>2</sub> -eq yr <sup>−1</sup> by 2100, considered as a conservative deployment estimate for BECCS-accepting pathways consistent with 1.5°C), Smith et al. (2016b) <sup>[[#fn:r605|605]]</sup> estimate a land-use intensity of 0.3–0.5 ha tCO <sub>2</sub> -eq <sup>−1</sup> yr <sup>−1</sup> using forest residues, 0.16 ha CO <sub>2</sub> -eq <sup>−1</sup> yr <sup>−1</sup> for agricultural residues, and 0.03–0.1 ha tCO <sub>2</sub> -eq <sup>−1</sup> yr <sup>−1</sup> 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., 2011 <sup>[[#fn:r606|606]]</sup> ; see Erb et al., 2016b <sup>[[#fn:r607|607]]</sup> 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) <sup>[[#fn:r608|608]]</sup> 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) <sup>[[#fn:r609|609]]</sup> .

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 <sup>−1</sup> of energy, 109–990 Mha, 6–79 MtN, 218–4758 km <sup>3</sup> yr <sup>−1</sup> of water per GtCO <sub>2</sub> yr <sup>−1</sup> ; Smith and Torn, 2013; Smith et al., 2016b; Fajardy and Mac Dowell, 2017) <sup>[[#fn:r610|610]]</sup> and are not comparable to IAM pathways which consider system effects (Bauer et al., 2018) <sup>[[#fn:r611|611]]</sup> . Global impacts on nutrients and albedo are difficult to quantify (Smith et al., 2016b) <sup>[[#fn:r612|612]]</sup> . 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) <sup>[[#fn:r613|613]]</sup> . CCS (see Section 4.3.1) is largely absent from the Nationally Determined Contributions (Spencer et al., 2015) <sup>[[#fn:r614|614]]</sup> and lowly ranked in investment priorities (Fridahl, 2017) <sup>[[#fn:r615|615]]</sup> . Although there are dozens of small-scale BECCS demonstrations (Kemper, 2015) <sup>[[#fn:r616|616]]</sup> and a full-scale project capturing 1 MtCO <sub>2</sub> exists (Finley, 2014) <sup>[[#fn:r617|617]]</sup> ''',''' this is well below the numbers associated with 1.5°C or 2°C-compatible pathways (IEA, 2016a; Peters et al., 2017) <sup>[[#fn:r618|618]]</sup> . Although the majority of BECCS cost estimates are below 200 USD tCO  <sup>−1</sup> (Figure 4.2), estimates vary widely. Economic incentives for ramping up large CCS or BECCS infrastructure are weak (Bhave et al., 2017) <sup>[[#fn:r619|619]]</sup> . The 2050 average investment costs for such a BECCS infrastructure for bio-electricity and biofuels are estimated at 138 and 123 billion USD yr <sup>−1</sup> , respectively (Smith et al., 2016b) <sup>[[#fn:r620|620]]</sup> .

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) <sup>[[#fn:r621|621]]</sup> . However, removing BECCS and CCS from the portfolio of available options significantly raises modelled mitigation costs (Kriegler et al., 2013; Bauer et al., 2018) <sup>[[#fn:r622|622]]</sup> .

<div id="section-4-3-3-7-1-block-2"></div>

<span id="figure-4.2"></span>
<!-- START IMG -->
<!-- IMG TITLE -->
'''Figure 4.2'''

<span id="evidence-on-carbon-dioxide-removal-cdr-abatement-costs-2050-deployment-potentials-and-key-side-effects."></span>
<!-- IMG CAPTION -->
'''Evidence on carbon dioxide removal (CDR) abatement costs, 2050 deployment potentials, and key side effects.'''
<!-- IMG FILE -->
[[File:d868fe4092f4229ebcec7975ddc2c6f4 fig-4.2-768x1024.jpg]]

Panel A presents estimates based on a systematic review of the bottom up literature (Fuss et al., 2018) <sup>[[#fn:r623|623]]</sup> , corresponding to dashed blue boxes in Panel B. Dashed lines represent saturation limits for the corresponding technology. Panel B shows the percentage of papers at a given cost or potential estimate. Reference year for all potential estimates is 2050, while all cost estimates preceding 2050 have been included (as early as 2030, older estimates are excluded if they lack a base year and thus cannot be made comparable). Ranges have been trimmed to show detail (see Fuss et al., 2018 <sup>[[#fn:r624|624]]</sup> for the full range). Costs refer only to abatement costs. Icons for side-effects are allocated only if a critical mass of papers corroborates their occurrence

Notes: For references please see Supplementary Material Table 4.SM.3. Direct air carbon dioxide capture and storage (DACCS) is theoretically only constrained by geological storage capacity, estimates presented are considering upscaling and cost challenges (Nemet et al., 2018) <sup>[[#fn:r625|625]]</sup> . BECCS potential estimates are based on bioenergy estimates in the literature (EJ yr <sup>−1</sup> ), converted to GtCO <sub>2</sub> following footnote 4. Potentials cannot be added up, as CDR options would compete for resources (e.g., land). SCS – soil carbon sequestration; OA – ocean alkalinization; EW- enhanced weathering; DACCS – direct air carbon dioxide capture and storage; BECCS – bioenergy with carbon capture and storage; AR – afforestation

<!-- END IMG -->
<div id="section-4-3-7-2"></div>

<span id="afforestation-and-reforestation-ar"></span>
==== 4.3.7.2 Afforestation and reforestation (AR) ====

<div id="section-4-3-7-2-block-1"></div>

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) <sup>[[#fn:r626|626]]</sup> 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 GtCO <sub>2</sub> yr <sup>−1</sup> for decades. The full literature range gives 2050 potentials of 1–7 GtCO <sub>2</sub> yr <sup>−1</sup> ( ''low evidence, medium agreement'' ), narrowed down to 0.5–3.6 GtCO <sub>2</sub> yr <sup>−1</sup> based on a number of constraints (Fuss et al., 2018) <sup>[[#fn:r627|627]]</sup> . Abatement costs are estimated to be low compared to other CDR options, 5–50 USD tCO <sub>2</sub> -eq <sup>−1</sup> ( ''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., 2016b <sup>[[#fn:r628|628]]</sup> ; Cross-Chapter Box 7 in Chapter 3). The 2030 estimate by Griscom et al. (2017) <sup>[[#fn:r629|629]]</sup> is up to 17.9 GtCO <sub>2</sub> yr <sup>−1</sup> 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 CO <sub>2</sub> stored geologically (Smith et al., 2016a) <sup>[[#fn:r630|630]]</sup> and is subject to disturbances that can be exacerbated by climate change (e.g., drought, forest fires and pests) (Seidl et al., 2017) <sup>[[#fn:r631|631]]</sup> . 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) <sup>[[#fn:r632|632]]</sup> , 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) <sup>[[#fn:r633|633]]</sup> , 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) <sup>[[#fn:r634|634]]</sup> . 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) <sup>[[#fn:r635|635]]</sup> , 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., 2016 <sup>[[#fn:r636|636]]</sup> ; 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) <sup>[[#fn:r637|637]]</sup> or used for energy crops (Grubler et al., 2018) <sup>[[#fn:r638|638]]</sup> . Such land-sparing strategies could also benefit other land-based CDR options.

<div id="section-4-3-7-3"></div>

<span id="soil-carbon-sequestration-and-biochar"></span>
==== 4.3.7.3 Soil carbon sequestration and biochar ====

<div id="section-4-3-7-3-block-1"></div>

At local scales there is ''robust evidence'' that soil carbon sequestration (SCS, e.g., agroforestry, De Stefano and Jacobson, 2018) <sup>[[#fn:r639|639]]</sup> , restoration of degraded land (Griscom et al., 2017) <sup>[[#fn:r640|640]]</sup> , or conservation agriculture management practices (Aguilera et al., 2013; Poeplau and Don, 2015; Vicente-Vicente et al., 2016) <sup>[[#fn:r641|641]]</sup> 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 tCO <sub>2</sub> <sup>−1</sup> (negative costs relating to the multiple co-benefits of SCS, such as increased productivity and resilience of soils; P. Smith et al., 2014) <sup>[[#fn:r642|642]]</sup> , and 2050 potentials are estimated at between 0.5 and 11 GtCO <sub>2</sub> yr <sup>−</sup> <sup>1</sup> , narrowed down to 2.3–5.3 GtCO <sub>2</sub> yr <sup>−</sup> <sup>1</sup> considering that studies above 5 GtCO <sub>2</sub> yr <sup>−</sup> <sup>1</sup> often do not apply constraints, while estimates lower than 2 GtCO <sub>2</sub> yr <sup>−</sup> <sup>1</sup> mostly focus on single practices (Fuss et al., 2018) <sup>[[#fn:r643|643]]</sup> .

SCS has negligible water and energy requirements (Smith, 2016) <sup>[[#fn:r644|644]]</sup> , 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) <sup>[[#fn:r645|645]]</sup> .

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. <sup>[[#fn:5|5]]</sup> Looking at the full literature range, the global potential in 2050 lies between 1 and 35 Gt CO <sub>2</sub> yr <sup>−</sup> <sup>1</sup> ( ''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) <sup>[[#fn:r646|646]]</sup> lower the 2050 range to 0.3–2 GtCO <sub>2</sub> yr <sup>−</sup> <sup>1</sup> . This potential is below previous estimates (e.g., Woolf et al., 2010) <sup>[[#fn:r647|647]]</sup> , 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) <sup>[[#fn:r648|648]]</sup> . Costs are 30– 120 USD tCO <sub>2</sub> <sup>−1</sup> ( ''medium agreement, medium evidence'' ) (McCarl et al., 2009; McGlashan et al., 2012; McLaren, 2012; Smith, 2016) <sup>[[#fn:r649|649]]</sup> .

Water requirements are low and at full theoretical deployment, up to 65 EJ yr <sup>−1</sup> of energy could be generated as a side product (Smith, 2016) <sup>[[#fn:r650|650]]</sup> . Positive side effects include a favourable effect on nutrients and reduced N <sub>2</sub> O emissions (Cayuela et al., 2014; Kammann et al., 2017) <sup>[[#fn:r651|651]]</sup> . However, 40–260 Mha are needed to grow the biomass for biochar for implementation at 0.3 GtCO <sub>2</sub> -eq yr <sup>−1</sup> (Smith, 2016) <sup>[[#fn:r652|652]]</sup> , even though it is also possible to use residues (e.g., Windeatt et al., 2014) <sup>[[#fn:r653|653]]</sup> . Biochar is further constrained by the maximum safe holding capacity of soils (Lenton, 2010) <sup>[[#fn:r654|654]]</sup> and the labile nature of carbon sequestrated in plants and soil at higher temperatures (Wang et al., 2013) <sup>[[#fn:r655|655]]</sup> .

<div id="section-4-3-7-4"></div>

<span id="enhanced-weathering-ew-and-ocean-alkalinization"></span>
==== 4.3.7.4 Enhanced weathering (EW) and ocean alkalinization ====

<div id="section-4-3-7-4-block-1"></div>

Weathering is the natural process of rock decomposition via chemical and physical processes in which CO <sub>2</sub> is spontaneously consumed and converted to solid or dissolved alkaline bicarbonates and/or carbonates (IPCC, 2005a) <sup>[[#fn:r656|656]]</sup> . 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) <sup>[[#fn:r657|657]]</sup> , shorelines (Hangx and Spiers, 2009; Montserrat et al., 2017) <sup>[[#fn:r658|658]]</sup> or the open ocean (House et al., 2007; Harvey, 2008; Köhler et al., 2013; Hauck et al., 2016) <sup>[[#fn:r659|659]]</sup> . Ocean alkalinization adds alkalinity to marine areas to locally increase the CO <sub>2</sub> buffering capacity of the ocean (González and Ilyina, 2016; Renforth and Henderson, 2017) <sup>[[#fn:r660|660]]</sup> .

In the case of land application of ground minerals, the estimated CDR potential range is 0.72–95 GtCO <sub>2</sub> yr <sup>−1</sup> ( ''low evidence, low agreement'' ) (Hartmann and Kempe, 2008; Köhler et al., 2010; Hartmann et al., 2013; Taylor et al., 2016; Strefler et al., 2018a) <sup>[[#fn:r661|661]]</sup> . Marine application of ground minerals is limited by feasible rates of mineral extraction, grinding and delivery, with estimates of  1–6 GtCO <sub>2</sub> yr <sup>−1</sup> ( ''low evidence, low agreement'' ) (Köhler et al., 2013; Hauck et al., 2016; Renforth and Henderson, 2017) <sup>[[#fn:r662|662]]</sup> . 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) <sup>[[#fn:r663|663]]</sup> . As with other CDR options, scaling and maturity are challenges, with deployment at scale potentially requiring decades (NRC, 2015a) <sup>[[#fn:r664|664]]</sup> , considerable costs in transport and disposal (Hangx and Spiers, 2009; Strefler et al., 2018a) <sup>[[#fn:r665|665]]</sup> and mining (NRC, 2015a; Strefler et al., 2018a) <sup>[[#fn:r666|666]]</sup> <sup>[[#fn:6|6]]</sup> .

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) <sup>[[#fn:r667|667]]</sup> , material transport, and rock source (Renforth, 2012; Hartmann et al., 2013) <sup>[[#fn:r668|668]]</sup> , and range from 15–40 USD tCO <sub>2</sub> <sup>−1</sup> to 3,460 USD tCO <sub>2</sub> <sup>−1</sup> ( ''limited evidence, low agreement'' ; Figure 4.2) (Schuiling and Krijgsman, 2006; Köhler et al., 2010; Taylor et al., 2016) <sup>[[#fn:r669|669]]</sup> . 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 GtCO <sub>2</sub> yr <sup>−1</sup> with costs of 14– &gt;500 USD tCO <sub>2</sub> <sup>−1</sup> (Renforth and Henderson, 2017) <sup>[[#fn:r670|670]]</sup> .

The main side effects of terrestrial EW are an increase in water pH (Taylor et al., 2016) <sup>[[#fn:r671|671]]</sup> , the release of heavy metals like Ni and Cr and plant nutrients like K, Ca, Mg, P and Si (Hartmann et al., 2013) <sup>[[#fn:r672|672]]</sup> , 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) <sup>[[#fn:r673|673]]</sup> ; 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) <sup>[[#fn:r674|674]]</sup> . Side effects of marine EW and ocean alkalinization are the potential release of heavy metals like Ni and Cr (Montserrat et al., 2017) <sup>[[#fn:r675|675]]</sup> . Increasing ocean alkalinity helps counter ocean acidification (Albright et al., 2016; Feng et al., 2016) <sup>[[#fn:r676|676]]</sup> . Ocean alkalinization could affect ocean biogeochemical functioning (González and Ilyina, 2016) <sup>[[#fn:r677|677]]</sup> . A further caveat of relates to saturation state and the potential to trigger spontaneous carbonate precipitation. <sup>[[#fn:7|7]]</sup> While the geochemical potential to remove and store CO <sub>2</sub> 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.

<div id="section-4-3-7-5"></div>

<span id="direct-air-carbon-dioxide-capture-and-storage-daccs"></span>
==== 4.3.7.5 Direct air carbon dioxide capture and storage (DACCS) ====

<div id="section-4-3-7-5-block-1"></div>

Capturing CO <sub>2</sub> from ambient air through chemical processes with subsequent storage of the CO <sub>2</sub> 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 CO <sub>2</sub> concentration in ambient air is 100–300 times lower than at gas- or coal-fired power plants (Sanz-Pérez et al., 2016) <sup>[[#fn:r678|678]]</sup> thus requiring more energy than flue gas CO <sub>2</sub> capture (Pritchard et al., 2015) <sup>[[#fn:r679|679]]</sup> . This appears to be the main challenge to DACCS (Sanz-Pérez et al., 2016; Barkakaty et al., 2017) <sup>[[#fn:r680|680]]</sup> .

Studies explore alternative techniques to reduce the energy penalty of DACCS (van der Giesen et al., 2017) <sup>[[#fn:r681|681]]</sup> . Energy consumption could be up to 12.9 GJ tCO <sub>2</sub> -eq <sup>−1</sup> ; translating into an average of 156 EJ yr <sup>−1</sup> by 2100 (current annual global primary energy supply is 600 EJ); water requirements are estimated to average 0.8–24.8 km <sup>3</sup> GtCO <sub>2</sub> -eq <sup>−1</sup> yr <sup>−1</sup>  (Smith et al., 2016b, based on Socolow et al., 2011) <sup>[[#fn:r682|682]]</sup> .

However, the literature shows ''low agreement'' and is fragmented (Broehm et al., 2015) <sup>[[#fn:r683|683]]</sup> . This fragmentation is reflected in a large range of cost estimates: from 20–1,000 USD tCO <sub>2</sub> <sup>−1</sup> (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) <sup>[[#fn:r684|684]]</sup> . There is lower agreement and a smaller evidence base at the lower end of the cost range. Fuss et al. (2018) <sup>[[#fn:r685|685]]</sup> narrow this range to 100–300 USD tCO <sub>2</sub> <sup>-1</sup> .

Research and efforts by small-scale commercialization projects focus on utilization of captured CO <sub>2</sub> (Wilcox et al., 2017) <sup>[[#fn:r686|686]]</sup> . Given that only a few IAM scenarios incorporate DACCS (e.g., Chen and Tavoni, 2013; <sup>[[#fn:r687|687]]</sup> Strefler et al., 2018b) <sup>[[#fn:r688|688]]</sup> 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) <sup>[[#fn:r689|689]]</sup> and few demonstrations (Holmes et al., 2013; Rau et al., 2013; Agee et al., 2016) <sup>[[#fn:r690|690]]</sup> , deploying the technology at scale is still a considerable challenge, though both optimistic (Lackner et al., 2012) <sup>[[#fn:r691|691]]</sup> and pessimistic outlooks exist (Pritchard et al., 2015) <sup>[[#fn:r692|692]]</sup> .

<div id="section-4-3-7-6"></div>

<span id="ocean-fertilization"></span>
==== 4.3.7.6 Ocean fertilization ====

<div id="section-4-3-7-6-block-1"></div>

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) <sup>[[#fn:r693|693]]</sup> . There is ''limited evidence'' and ''low agreement'' on the readiness of this technology to contribute to rapid decarbonization (Williamson et al., 2012) <sup>[[#fn:r694|694]]</sup> . Only small-scale field experiments and theoretical modelling have been conducted (e.g., McLaren, 2012) <sup>[[#fn:r695|695]]</sup> . The full range of CDR potential estimates is from 15.2 ktCO <sub>2</sub> yr <sup>−</sup> <sup>1</sup> (Bakker et al., 2001) <sup>[[#fn:r696|696]]</sup> for a spatially constrained field experiment up to 44 GtCO <sub>2</sub> yr <sup>−</sup> <sup>1</sup> (Sarmiento and Orr, 1991) <sup>[[#fn:r697|697]]</sup> following a modelling approach, but Fuss et al. (2018) <sup>[[#fn:r698|698]]</sup> 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) <sup>[[#fn:r699|699]]</sup> .

Cost estimates range from 2 USD tCO <sub>2</sub> <sup>−</sup> <sup>1</sup> (for iron fertilization) (Boyd and Denman, 2008) <sup>[[#fn:r700|700]]</sup> to 457 USD tCO <sub>2</sub> <sup>−</sup> <sup>1</sup> (Harrison, 2013) <sup>[[#fn:r701|701]]</sup> . Jones (2014) <sup>[[#fn:r702|702]]</sup> proposed values greater than 20 USD tCO <sub>2</sub> <sup>−</sup> <sup>1</sup> for nitrogen fertilization. Fertilization is expected to impact food webs by stimulating its base organisms (Matear, 2004) <sup>[[#fn:r703|703]]</sup> , and extensive algal blooms may cause anoxia (Sarmiento and Orr, 1991; Matear, 2004; Russell et al., 2012) <sup>[[#fn:r704|704]]</sup> and deep water oxygen decline (Matear, 2004) <sup>[[#fn:r705|705]]</sup> , 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) <sup>[[#fn:r706|706]]</sup> and non-CO <sub>2</sub> GHGs may increase (Sarmiento and Orr, 1991; Matear, 2004; Bertram, 2010) <sup>[[#fn:r707|707]]</sup> . The greatest theoretical potential for this practice is the Southern Ocean, posing challenges for monitoring and governance (Robinson et al., 2014) <sup>[[#fn:r708|708]]</sup> . The London Protocol of the International Maritime Organization has asserted authority for regulation of ocean fertilization (Strong et al., 2009) <sup>[[#fn:r709|709]]</sup> , 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 CO <sub>2</sub> 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) <sup>[[#fn:r710|710]]</sup> . Storage at the surface would mean that the carbon would be rapidly released after cessation (Zeebe, 2005; Aumont and Bopp, 2006) <sup>[[#fn:r711|711]]</sup> .

<div id="section-4-3-7-6-block-2"></div>

<span id="table-4.6"></span>
<!-- START TABLE -->
'''Table 4.6'''

<span id="cross-cutting-issues-and-uncertainties-across-carbon-dioxide-removal-cdr-options-aspects-and-uncertainties"></span>
'''Cross-cutting issues and uncertainties across carbon dioxide removal (CDR) options, aspects and uncertainties'''

<!-- TABLE -->
{| class="wikitable"
|-
! Area of Uncertainty
! Cross-Cutting Issues and Uncertainties
|-
| Technology upscaling
| * CDR options are at different stages of technological readiness (McLaren, 2012) <sup>[[#fn:r712|712]]</sup> and differ with respect to scalability.
* Nemet et al. (2018)  <sup>[[#fn:r713|713]]</sup> find &gt;50% of the CDR innovation literature concerned with the earliest stages of the innovation process (R&amp;D), identifying a dissonance between the large CO <sub>2</sub> removals needed in 1.5°C pathways and the long -time periods involved in scaling up novel technologies.
* Lack of post-R&amp;D literature, including incentives for early deployment, niche markets, scale up, demand, and public acceptance.

|-
| Emerging and niche technologies
| * For BECCS, there are niche opportunities with high efficiencies and fewer trade-offs, for example, sugar and paper processing facilities (Möllersten et al., 2003) <sup>[[#fn:r714|714]]</sup> , district heating (Kärki et al., 2013; Ericsson and Werner, 2016) <sup>[[#fn:r715|715]]</sup> ,  and industrial and municipal waste (Sanna et al., 2012) <sup>[[#fn:r716|716]]</sup> . Turner et al. (2018)  <sup>[[#fn:r717|717]]</sup> constrain potential using sustainability considerations and overlap with storage basins to avoid the CO <sub>2</sub> transportation challenge, providing a possible, though limited entry point for BECCS.
* The impacts on land use, water, nutrients and albedo of BECCS could be alleviated using marine sources of biomass that could include aquacultured micro and macro flora (Hughes et al., 2012; Lenton, 2014) <sup>[[#fn:r718|718]]</sup> .
* Regarding captured CO <sub>2</sub> as a resource is discussed as an entry point for CDR. However, this does not necessarily lead to carbon removals, particularly if the CO <sub>2</sub> is sourced from fossil fuels and/or if the products do not store the CO <sub>2</sub> for climate-relevant horizons (von der Assen et al., 2013)  <sup>[[#fn:r719|719]]</sup> (see also Section 4.3.4.5).
* Methane <sup>[[#fn:8|8]]</sup> is a much more potent GHG than CO <sub>2</sub> (Montzka et al., 2011) <sup>[[#fn:r720|720]]</sup> , associated with difficult-to-abate emissions in industry and agriculture and with outgassing from lakes, wetlands, and oceans (Lockley, 2012; Stolaroff et al., 2012) <sup>[[#fn:r721|721]]</sup> . Enhancing processes that naturally remove methane, either by chemical or biological decomposition (Sundqvist et al., 2012) <sup>[[#fn:r722|722]]</sup> , has been proposed to remove CH <sub>4</sub> . There is ''low confidence'' that existing technologies for CH <sub>4</sub>  removal are economically or energetically suitable for large-scale air capture (Boucher and Folberth, 2010) <sup>[[#fn:r723|723]]</sup> . Methane removal potentials are limited due to its low atmospheric concentration and its low chemical reactivity at ambient conditions.

|-
| Ethical aspects
| * Preston (2013)  <sup>[[#fn:r724|724]]</sup> identifies distributive and procedural justice, permissibility, moral hazard (Shue, 2018) <sup>[[#fn:r725|725]]</sup> , and hubris as ethical aspects that could apply to large-scale CDR deployment.
* There is a lack of reflection on the climate futures produced by recent modelling and implying very different ethical costs/risks and benefits (Minx et al., 2018) <sup>[[#fn:r726|726]]</sup> .

|-
| Governance
| * Existing governance mechanisms are scarce and either targeted at particular CDR options (e.g., ocean-based) or aspects (e.g., concerning indirect land-use change (iLUC)) associated with bioenergy upscaling, and often the mechanisms are at national or regional scale (e.g., EU). Regulation accounting for iLUC by formulating sustainability criteria (e.g., the EU Renewable Energy Directive) has been assessed as insufficient in avoiding leakage (e.g., Frank et al., 2013) <sup>[[#fn:r727|727]]</sup> .
* An international governance mechanism is only in place for R&amp;D of ocean fertilization within the Convention on Biological Diversity (IMO, 1972, 1996; CBD, 2008, 2010) <sup>[[#fn:r728|728]]</sup> .
* Burns and Nicholson (2017)  <sup>[[#fn:r729|729]]</sup> propose a human rights-based approach to protect those potentially adversely impacted by CDR options.

|-
| Policy
| * The CDR potentials that can be realized are constrained by the lack of policy portfolios incentivising large-scale CDR (Peters and Geden, 2017) <sup>[[#fn:r730|730]]</sup> .
* Near-term opportunities could be supported through modifying existing policy mechanisms (Lomax et al., 2015) <sup>[[#fn:r731|731]]</sup> .
* Scott and Geden (2018)  <sup>[[#fn:r732|732]]</sup> sketch three possible routes for limited progress, (i) at EU-level, (ii) at EU Member State level, and (iii) at private sector level, noting the implied paradigm shift this would entail.
* EU may struggle to adopt policies for CDR deployment on the scale or time-frame envisioned by IAMs (Geden et al., 2018) <sup>[[#fn:r733|733]]</sup> .
* Social impacts of large-scale CDR deployment (Buck, 2016) <sup>[[#fn:r734|734]]</sup> require policies taking these into account.

|-
| Carbon cycle
| * On long time scales, natural sinks could reverse (C.D. Jones et al., 2016)  <sup>[[#fn:r735|735]]</sup>
* No robust assessments yet of the effectiveness of CDR in reverting climate change (Tokarska and Zickfeld, 2015; Wu et al., 2015; Keller et al., 2018) <sup>[[#fn:r736|736]]</sup> , see also Chapter 2, Section 2.2.2.2.

|}
<!-- END TABLE -->

<span id="solar-radiation-modification-srm"></span>