MANY CDR STRATEGIES PROVIDE VIABLE AND REASONABLY LOW-RISK APPROACHES to reducing atmospheric concentrations of CO2. Because the rate of CO2 removal is inherently slow, CDR must be sustained at large scales over very long periods of time to have a significant effect on CO2 concentrations and the associated risks of climate change
The landmark US National Academy of Sciences report, Climate Intervention: Carbon Dioxide Removal and Reliable Sequestration, 2015
The landmark US National Academy of Sciences report, Climate Intervention: Carbon Dioxide Removal and Reliable Sequestration, released in 2015, reviewed the potential contributions to carbon dioxide removal (CDR) from a range of CDR technologies and strategies. These are summarised below.
AFFORESTATION AND REFORESTATION
This currently contributes a net annual uptake of 1 GtCO2, albeit that deforestation, primarily of tropical forests, constitutes the single largest source of land-use related GHG emissions and accounts for about 10% of anthropogenic emissions from all sources. A critical component of any climate mitigation strategy is to prevent additional tropical deforestation, which as an outside limit could add as much as 1,800 GtCO2 to the atmosphere in cumulative emissions—roughly as much CO2 as from all fossil fuel use from the preindustrial period until the present.
Estimates of potential carbon sequestration rates vary. The IPCC 5th Assessment report projected rates of up to 1.5, 9.5 and 14 GtCO2/year in 2030, depending on the mitigation scenario. Other, perhaps more realistic estimates, lie in the range 3.7 to 4-6 GtCO2/year.Nitrogen requirements for enhanced reforestation programmes are an issue: 1-5% is converted to nitrous oxide which has a global warming potential 300 times greater than CO2, leading another study to conclude that consequent N2O emissions can offset stored CO2 by 75 to 310%. Based on land availability and past soil carbon losses, afforestation is estimated to have a cumulative potential impact of 110-180 GtCO2 over 50 years, and 380 GtCO2 over 100 years. Climate change could, however, affect sequestration: forest drought, disease and soil loss could result in terrestrial systems becoming a net carbon source, rather than a sink.
Biological sequestration in forests can be inexpensive: for the US, costs are estimated at $7.5/tCO2 to $22/tCO2. The IPCC Fifth Assessment estimates $20/tCO2 to $100/tCO2 for a global programme.
AGRICULTURAL LAND CARBON SEQUESTRATION
Many cultivated soils have lost 50-70% of their original organic carbon, and intensive soil cultivation has the potential to reduce soil carbon by 25-50% over 30-50 years. Agricultural land can be managed to partially reverse these losses by growing cover crops (adding 0.07-0.7tCO2/year per ha), leaving crop residues to decay, applying manure or compost, using low- or no-till systems (1.2tCO2/year per ha) for the first 20 years, converting croplands back to wetlands (-0.6 to 3.3tCO2eq/year per ha and employing other land management techniques such as low-input, high diversity energy crops (up to 4.4tCO2/year per ha) that increase soil structure and stored organic matter. Of the total 13 billion hectares that make up Earth’s ice-free surface, cropland accounts for around 12%, yielding a global technical potential of 5.2 GtCO2/year in 2030 (IPCC 2014).
ACCELERATED WEATHERING METHODS AND MINERAL CARBONATION
The long term fate of most CO2 released to the atmosphere is to become bicarbonate ions dissolved in the oceans, and later carbonate sediments on the sea floor. These transformations occur as a result of ions provided by carbonate or silicate weathering reactions that typically occur in soils or marine sediments. One class of CDR involves accelerating these processes for CO2 storage as carbonates or bicarbonates in the oceans or as carbonates on land — either in-situ, or ex-situ in an industrial setting.
Absorption of CO2 by the oceans increases their acidity, inhibiting further absorption. The dissolution of calcium carbonate minerals neutralises some of this acidity, allowing more CO2 uptake and increasing the alkalinity of the seawater as follows:
CO2 + Ca CO3 + H2O → Ca 2+ + 2HCO3
It typically takes 2,000-8,000 years for this reaction to return the ocean-surface sediment carbonate system to a steady state following a CO2 perturbation, due to the slow rate of ocean CO2 transport and rates of the natural calcium carbonate cycle involving weathering on land and deposition in marine sediments. It is the acceleration of these processes that are now proposed to reduce atmospheric CO2 and ocean acidity. Silicate weathering reactions are similar, but can store twice as much carbon per mole as carbonate minerals.
Deployed at full scale, these methods would involve mining of substantial masses of mineral — in the order of 100 billion tons to offset current CO2 emissions of around 34 GtCO2/year. For comparison, total world production of coal is around 8 billion tonnes/year (for storage as calcium carbonate, CaCO3, the figure would be 80 billion tonnes, based on a multiple of 2.3 times the mass of the CO2). In-situ mineralisation involves promoting carbonate or silicate mineral weathering reactions on land or in the oceans, rather than in centralised facilities.
Carbonate minerals, silicate minerals and sea water are all abundant and so there are no obvious fundamental physical constraints limiting global scaling. However, at the scale required, economic and environmental considerations may constrain its potential ex-situ. In-situ weathering may, however, offer greater scope. A 2008 study calculated a sequestration potential of 1 trillion tones CO2 within three km of the surface of the Sultanate of Oman, and a similar amount in a section 10km wide by 3 km deep in the world’s slow-spreading ridges.
Ocean-based accelerated weathering has a potential of 1 GtCO2/year at a cost in the range of $50-100/tCO2, and land-based in-situ mineral carbonation may have a potential of 4GtCO2/year, with an estimated cost of $23/tCO2 to $66/tCO2 (IPCC 2014).
A natural biological pump exists in the sea: planktonic algae and other microscopic plants take up CO2 at the ocean surface and convert it to particulate organic matter. Some settles in the deep ocean, and some is taken up in the food chain, but the net result is the sequestration of inorganic carbon, driving stabilisation of atmospheric levels of a few tens to perhaps 100 ppm over decades to century timescales. Various CDR proposals attempt to enhance this process by adding limiting nutrients, such as nitrogen, phosphorus or trace metals, to promote biological sequestration. Whilst the amounts of nitrogen and phosphorus required would be large, micro-nutrients such as iron have a much higher carbon ratio (for carbon to iron, in the range 1,000 to 100,000 mole/mole), and are potentially, therefore, much more effective.
An extensive series of small-scale iron-release experiments has shown that its addition to high-nitrate, low chlorophyll regions in the equatorial Pacific and Southern Ocean does cause increased phytoplankton growth rates and the development of plankton blooms. However, evidence of increased sinking of particulate carbon has proven more elusive, and other questions of environmental impacts remain.
Modelling studies indicate a potential upper limit for a sustained ocean iron fertilisation CO2 sink of around 1.0-3.7 GtCO2/year, although cost estimates vary widely from <$10 to 450/tCO2.
BIOENERGY WITH CARBON CAPTURE AND SEQUESTRATION (BECCS)
BECCS is a process in which biomass is converted to heat, electricity, or liquid or gas fuels, followed by CO2 capture and sequestration. It can produce liquid fuels such as ethanol or methanol, gas fuels such as hydrogen, or engineered algal systems designed to directly produce hydrocarbons. To form liquid fuels, the synthesis gas would be catalytically reacted through a Fischer-Tropsch process.
Current estimates show that if BECCS were deployed to its theoretically maximum, it could account for a significant proportion of the world’s energy supply. However, both the availability of land for biomass cultivation, and the need for the ensuing bulk transportation to processing facilities severely limit the feasible use of bioenergy. Generation of 100 EJ/year may require up to 500 million hectares of land (assuming 10 tonnes of dry biomass per ha of land annually). For comparison, about 1,600 million hectares are currently planted with crops, and 3,400 million hectares are used for pasture.
There are also other issues: there is no difference in net carbon emissions whether CCS is tied to bioenergy or fossil fuel use. Primary forests are more diverse than secondary, so large-scale displacement with biomass plantations would have deleterious biodiversity impacts. Old growth forests and undisturbed grasslands have significant amounts of sequestered carbon, and conversion to other uses leads to large greenhouse gas emissions.
Estimates of the potential contribution of BECCS at scale vary from 1-18GtCO2/year, with mid-range calculations of 3 to 1 GtCO2/year to 2050 (IPCC 2014).
The urgency for large-scale DAC installation both to reduce ongoing emissions and atmospheric CO2 is slowly gaining recognition, but its rate of deployment in various markets will depend on their policy contexts and the growth of other technologies which remove CO2 at source.
In the transportation sector, it is reasonable to assume that ultimately all road, rail, shipping and aviation modes will be powered by a combination of electricity, hydrogen and synthetic fuels (A2F). Rates of market penetration in each of these will depend on a multiplicity of factors, the impact of which will vary between sectors. Electric power, for instance, is compelling for land transportation, but – absent radical innovation -- is unlikely to be viable for air transport, at least within a timeframe which would impact on 2050 decabonisation targets. Battery power density and range issues are also likely to confine land transport mostly to hybrid systems for the immediate future.
Whilst generational times for technological substitution are shortening, it remains the case that for innovation in larger-scale infrastructure, replacement times are more elongated and typically in the order of four decades or more (1). For the fossil fuel sector as a whole, complete system replacement could take up to a century (2). With the advent of DAC, however, that does not necessarily translate to a 100-year timeframe for zero-carbon delivery, due to differential rates of low- and zero-carbon technology take-up between developed and emerging markets, and the potential for decarbonisation through DAC/EOR and carbon offsets. The objective for overall carbon-neutrality, at least for the most advanced markets, should be 2050.
Any market assessment stretching decades into the future is unavoidably conjectural but, in the simplified form of a ‘constraints and opportunities’ exercise, it can nonetheless provide a useful indication of need and potential.
It should be emphasised that the DAC deployment scenario below, although illustrative, is realistic to the extent that it indicates the approximate scale of intervention necessary if the world is to avoid dangerous climate change. It therefore necessarily assumes an idealised policy context, where climate change is addressed not on the basis of political expediency designed to convey an impression of adequacy, but on the measures actually required for a reasonable prospect of solving the problem. Framing policy on the basis of real-world science is the sine qua non of any rational policy response, and it can only be reasonable to assume that, however protracted the current policy process, circumstances must soon prevail where political commitment properly reflects the urgency and seriousness of the situation.
Reflecting the distinction between addressing ongoing emissions and atmospheric reduction, the evaluation below has two limbs: a proposed 40-year build up in DAC plants for low and zero-carbon fuel substitution, and a parallel 40-year deployment of DAC plants for atmospheric CO2 sequestration. In the case of the former, this timeframe extends a decade beyond the date at which carbon-neutrality is targeted (by 2050) in order to provide an additional margin for further market growth in each of the projected sectors, and/or possible slippage in progress toward their decarbonisation. Under this scenario, no plants are constructed beyond 2060: decommissioning of plants after forty years as they reach their end-of-life dates therefore results in a progressive decline in the number of operating plants, falling to zero by 2100 (this assumption is necessarily speculative, but the main purpose of this analysis is to determine the minimum DAC requirements under given assumptions: demand beyond 2060 can be accommodated by simply continuing plant construction thereafter, instead of allowing for facilities retirement as they reach replacement age. Alternatively, plants could be switched between sectors post-2050 as the market dictates – see following).
As will be appreciated from Table 1 below, the wide variation in possible policy contexts and emission-reduction scenarios results in an equally wide range of outcomes for DAC plant deployment. Given their high capital investment and relatively long life cycle, flexibility in responding to market requirements over multi-decadal timeframes will be necessary to enhance investment security and financing ability. For this reason, it is proposed that — at least toward the upper end of global capacity and in particular post-2050 — the ability to switch between strategic markets (EOR/A2S and offsets/atmospheric reduction) may be critical to matching market demand and supply in particular sectors. Thus, for instance, acceleration in the electrification of road transportation could be accommodated through re-allocation of DAC plants from A2F to atmospheric reduction — assuming of course that other A2F markets had reached saturation. Providing for such a possibility narrows locational options to those that can meet the requirements of both.
Figure 3 below indicates possible trajectories for various DAC markets, with the ultimate objective of deploying a sufficient number of installations to achieve zero-carbon status in transportation by 2050 and reduce atmospheric CO2 levels to 350ppm by 2100. The average number of new one-million tonne plants required to be built under this scenario is 940 a year (see Table 2 and panel opposite for breakdown between market sectors).
Whilst such a scale of undertaking undoubtedly constitutes a formidable challenge, it should be recognized that this is perfectly within current global capability and, in terms of comparables, not without precedent. Arguments for a massive escalation in resources for dealing with climate change on the equivalent of a wartime footing have referred to US industrial performance in WW2, where output rocketed to produce 229,000 planes and 5,000 ships in just three years from 1942 to 1944 (3). More recently, Caterpillar’s plant in Corinth, Mississippi, has demonstrated a capability to recycle some 1,700 truckloads of engines a day (4). The Ulsan shipyard in South Korea completes a giant ship every four or five days (5). Global car production in 2015 will exceed ninety million vehicles (6).
This output dwarfs that required for full scale DAC. As can be seen in the illustration of the FOAK plant in figure 1, a 100,000 tonne CO2/year plant comprises 160 contactor (capture) units, each about the size of a shipping container. A one-million tonne plant would need 1,600. Thus, even the peak global installation rate for all sectors of 1,890 plants/year envisaged in figure 3 below would require production of 3 million contactor units a year – one thirtieth of current global car output. This is by no means inconceivable in circumstances where the requisite policy framework is in place.
What is immediately apparent from these projections, however, is the scale of plant construction required for atmospheric CO2 reduction, compared to that for other applications, even adopting the most conservative assumptions for the former of 500 ppm peak levels of CO2, and the most aggressive market ambitions for the latter. Decarbonising the entire aviation sector by 2050, for instance, including allowance for projected market growth and high altitude multiplier impacts, requires construction of around 100 plants a year over a thirty-year timeframe. To reduce atmospheric CO2 levels to 350 ppm by 2100, adopting the most optimistic assumptions for peak concentrations and allowing for 50% of the reduction to be met through other GGR strategies, will require an average build programme of 305 plants a year to 2060.
Table 1 below sets out the DAC plant requirements for each sector. Although carbon neutrality is targeted for the transportation sector by 2050, continuing market growth at similar rates of expansion is assumed for each of the transportation modes to 2060, necessitating continuing deployment. From 2060, as explained above, deployment is negative, reflecting progressive retirement after a 40-year operating life, with replacement subject to evolving market demand.
As explained elsewhere (see Atmospheric CO2 Reduction), driving down atmospheric CO2 by 300ppm (the projected overshoot, based on current trajectories), including allowance for seepage from natural carbon sinks, represents a 4,000 Gt CO2 challenge, or 40 Gt CO2/year for 100 years. If this were to be addressed only through DAC, ramping up to the required 40,000 1 m tonne DAC plants over four decades would necessitate construction of 1,000 plants a year. Even then, the required capacity is only achieved for the remaining 60 years: an additional 20,000 plants/year would be needed to compensate for phasing from a zero start over the first forty years, and a further 20,000 plants would be required in total for end-of-life replacement of those constructed during that build-up period. This is no trivial undertaking.
However, whilst DAC is the most effective of the present GGR (Greenhouse Gas Removal) options in terms of efficiency in land-use, irreversibility and accountability, other technologies and strategies could conceivably be scaled collectively to a similar level of impact within a comparable timeframe and, in some cases, potentially more cost-effectively. These include reforestation, soil carbon sequestration (including organic agriculture), wetland restoration, BECCS (Bio-Energy with Carbon Capture and Storage) and biochar, as well as other possibilities such as accelerated or geo-chemical weathering and ocean liming still at a theoretical or experimental stage (see panel opposite for detail).
For the purposes of this exercise, it is assumed that these approaches can be scaled to account in total for around half of the required atmospheric emissions reduction to bring CO2 levels to 350ppm. It is also assumed that the growing body of evidence indicating that existing and proposed climate policies are seriously behind the curve will result in accelerated action to curb emissions such that global atmospheric CO2 levels peak at 500 ppm, rather than 650ppm as presently contemplated. This halves the reduction requirement to 150 ppm which, allowing as previously for re-emission from natural carbon sinks, increases the figure to 250 ppm, or 3.125 ppm/year over 80 years. Of this, DAC would meet 1.56 ppm or approximately 12.2 GtCO2/year (at 7.8 GtCO2 per ppm). In order to provide for the necessary capacity — allowing for ramping over four decades and downscaling over the following 40 years as plants are retired — peak capacity would need to be doubled to 24.4 GtCO2/year. Thus, a forty-year linear ramp up to 24,400 one-million tonne plants would capture 0.488 trillion tonnes CO2, followed by capture of the same amount in the downcurve, yielding a total reduction of nearly a trillion (0.976) tonnes CO2 by 2100. Although such a trajectory is very much a simplification of potential real-world deployment pathways (which would more likely follow that of an ‘S’ curve), it nonetheless provides a reasonable indication of the ultimate capacity requirement over the given timeframe.
For the 650 ppm scenario, these figures would be doubled.
Road transportation currently represents around 18% of global CO2 emissions, or 6.3 Gt of CO2/year. As seen elsewhere on this site, a projected three-fold increase in car ownership and a two-fold increase in trucking activity from 2014 to 2050 results in an increase – allowing for efficiency improvements -- of between 34% and 106% on 2014 emissions of 5.5 GtCO2 (range 7.37-11.33, average 9.13 GtCO2/year). Eliminating this through DAC would require the construction of 304 plants annually from 2020 to 2050. However, allowing for a further potential emissions increase of 33% from 2050 to 2060 increases this requirement to 405 plants a year.
Current fuel consumption by the aviation industry is 5.2m bbl/day. Reducing CO2 emissions 50% by 2050 equates to a reduction of 31.6m bbl/year over 30 years, requiring the construction of 16 1Mt DAC plants/year. Growth of 4% per annum net of technical efficiency improvements will result in additional consumption of 73m bbl/year, requiring a further 18 plants/year for a 50% carbon intensity (total 34 plants/year). 100% decarbonisation would double this figure to 68 plants/year, and applying a multiplier of 1.63 for impact at high altitude would increase this again to 111 plants/year.
For shipping to become zero-carbon by 2050, its emissions of around 1 Mt/year must be reduced by 33m tonnes annually. Adopting IMO projections of a quadrupling of the shipping market over this timeframe, but a halving in emissions intensity, increases the reduction requirement to 66 Gt per annum. If accomplished through sequestration offsets, it would require plant construction of sixty-six 1m tonne plants a year. Allowing for further growth post-2050, however, increases this figure to 80 plants a year.
As outlined in the section dealing with carbon offsets, the wealthiest 500m globally -- representing 8% of the world’s population -- account for around 50% of anthropogenic emissions. Assuming a market uptake of between 1% and 10%, this would generate a demand of between 250 and 2,500 DAC plants respectively, or a build up of 6.25-62.5 plants/year over the 40 years to 2060.
DAC Locational Criteria
For major installations, these are as follows:
ACCESS: plants ideally need to be serviced with rail and port access for large-scale shipments involved in plant construction; maintenance and replacement; and product distribution (such as synthetic fuels). EOR-related plants require either pipeline distribution or proximity to well locations.
PROXIMITY TO SITES FOR SEQUESTRATION: such as saline aquifers, deep oceans or mafic rock formations. This applies both to dedicated sequestration plants and A2F or EOR plants which may ultimately be switched to sequestration as other renewable energy sources offer competitive advantage in certain markets
SUBSTANTIAL LAND AREAS: As explained below, DAC is land-efficient; however, at the scale of operations envisaged here, it nevertheless requires large land areas that are relatively inexpensive with few competing uses
WATER: a 1Mt DAC plant uses 8m tonnes of water annually in a high evaporation scenario. Water costs vary according to availability, but areas of consistently high insolation tend to be water-scarce, so desalination may be necessary in some of these locations.
SOLAR GENERATION POTENTIAL: current generation DAC plant designs are powered by natural gas. By 2020, cost-reductions in solar generation will enable substitution by solar pv, favouring low latitude, high insolation locations.
ENVIRONMENTAL: Whilst the capture technology performs well under normal atmospheric conditions, the need to minimize downtime and maintenance costs would disqualify areas with a high incidence of dust or sand storms
SECURITY: Given the substantial capital cost of DAC installations and their multi-decadal lifecycle, politically stability will be a critical locational requirement.
'WHILST THE PROPOSED SCALE OF PLANT DEVELOPMENT IS LARGE BY ANY INDUSTRIAL STANDARDS, THE LAND AREAS INVOLVED ARE RELATIVELY INSIGNIFICANT COMPARED TO THOSE AVAILABLE IN SOME OF THE AREAS THAT MAY BE SUITABLE AS LOCATIONS FOR MAJOR PLANT INSTALLATIONS.
FOR ILLUSTRATIVE PURPOSES, EVEN DOUBLING THE EXPECTED LAND REQUIREMENT PER MILLION TONNE CO2 PER ANNUM FACILITY TO TWO SQUARE KILOMETRES, GENERATES A LAND REQUIREMENT OF 65,000 SQ KM FOR SUFFICIENT PLANTS TO ENABLE CAPTURE OF OUR ENTIRE ANNUAL CO2 EMISSIONS FROM FOSSIL FUELS. FOR COMPARISON, THE LAND AREA OF THE NORTHERN TERRITORY IN AUSTRALIA (WITH EXTENSIVE MAFIC GEOLOGICAL FORMATIONS SUITABLE FOR MINERAL SEQUESTRATION) IS 1.4 MILLION SQ. KMS, QUEENSLAND IS 1.8 MILLION SQ. KMS., AND WESTERN AUSTRALIA 2.6 MILLION SQ. KMS.
IN THE US, TEXAS (A POTENTIAL LOCATION FOR SEQUESTRATION IN SALINE AQUIFERS) HAS A LAND AREA OF 696.000 SQ. KMS.'
POTENTIAL FOR DEBT FUNDING
Economic analyses for 1mt plants for the EOR market indicate an IRR of 20% per annum assuming a 75:25 debt to equity funding ratio with debt funding at 6% per annum, net of licence fee income to Carbon Engineering of $20m/year per one million tonnes CO2 captured. Carbon Order is exploring the possibility of 100% debt funding against long-term out-take contracts for low and zero-carbon fuels, using collateralized forward purchase orders as security for plant capital funding. Such a project financing structure would enable major users of these fuels to leverage their substantial purchasing power to develop and own DAC infrastructure - thereby internalizing fuel-supply within their value chain without depleting their own capital in the process. Markets for which this approach could be especially interesting are aviation, shipping, road transportation and carbon offsets. Ultimately, atmospheric reduction is also a possibility, secured through contracts against a global price on carbon.