To limit global warming in 2100 to below 1.5C above pre-industrial levels, many scientists assume that the large-scale use of negative emissions in the latter half of the 21st century will be needed. Negative emissions “suck” CO2 out of the atmosphere, allowing a more gradual reduction of emissions in the near-term.
Integrated assessment models (IAMs) that generate energy and emission pathways to limit warming to 1.5C have generally relied on large amounts of bioenergy with carbon capture and storage (BECCS) to provide the required negative emissions. Many deploy BECCS on a massive scale, allocating a land area up to five times the size of India to growing the biomass needed by 2100.
Integrated Assessment Models: IAMs are computer models that analyse a broad range of data – e.g. physical, economic and social – to produce information that can be used to help decision-making. For climate research, specifically, IAMs are typically used to project future greenhouse gas emissions and climate impacts, and the benefits and costs of policy options that could be implemented to tackle them.
Criticism of model reliance on BECCS has led researchers to examine the potential of “natural climate solutions” (NCS) to remove CO2 in the atmosphere through reforestation, land-use change and other ecosystem-based approaches.
Here, Carbon Brief uses the recently published review of NCS to examine how big a role they could play in contributing to negative emissions. This analysis shows that NCS could provide a sizable portion of the required emissions and reduce the need for BECCS in pathways limiting warming to below 1.5C, particularly when coupled with faster emissions reductions over the next few decades.
Natural carbon sequestration
A recent review paper titled “Natural climate solutions” in the Proceedings of the US National Academy of Sciences (PNAS) examined the different ways that carbon could be captured through conservation, ecosystem restoration and improved land management across global forests, wetlands, grasslands and agricultural lands.
Unlike BECCS, which poses potential challenges to both the natural world and food supplies when adopted at such a massive scale, most NCS can also help improve water filtration, flood protection, soil health and biodiversity habitat. This makes these approaches an attractive alternative to BECCS for meeting the negative emissions needed in scenarios that limit warming to below 1.5C.
The figure below shows Carbon Brief’s estimates of the negative emissions potentials between 2018 and 2100 for various climate solutions assessed in the PNAS paper. These estimates are based on the rate of sequestration (in gigatonnes CO2 per year – GtCO2) and the time horizon before the ecosystem “saturates” and, therefore, can no longer store more carbon.Negative emissions potentials from different NCS, in cumulative GtCO2 between 2018 and 2100. Bars show uncertainties in total potential, while black circles show best-estimates of total potential and diamonds show economic potential at a cost of less than $100 per ton CO2. Estimates are based on both the rate of sequestration and the time horizon over which the sequestration can continue from Griscom et al. 2017.
The total and economically deployable potentials are shown for each NCS, with a black dot representing the best-estimate and a range showing the 95% confidence interval for each solution. The total potential is the maximum possible sequestration of the solution when adopted globally with no consideration of costs. The “cost effective” economic potential is the subset estimated to be economically viable at a cost of less than $100 per tonne CO2.
The total potential is constrained by safeguards for meeting increasing human needs for food and fibre. The authors do not allow existing cropland areas to be reduced or replaced by forests or rangeland. They also exclude any activities that would negatively impact biodiversity or offset the benefits from carbon sequestration due to factors like albedo changes.
The NCS with the largest potential to sequester carbon globally is reforestation. This involves the reversion of non-forest lands to forest in ecologically appropriate areas. The authors of the NCS paper exclude afforestation – the conversion of native grassland, savannah or other non-forest ecosystems to forest – to avoid damaging native ecosystems.
They create a map, below, of all the non-cropland areas globally that would be suitable for reforestation by country. They find a total potential reforestable area of somewhere between 345 and 1,779 million hectares (Mha), and a range of potential sequestration between 2018 and 2100 of 68-447GtCO2 with a best estimate of 253GtCO2. About 30% of this potential sequestration, or 76GtCO2, is cost effective at a price below $100 per tonne CO2.
The vast majority of the reforestation potential is in tropical regions, where rainforests have been cleared for pasture land, palm oil production and timber. The authors point out that their results are somewhat sensitive to assumptions about how much pasture area can be reasonably reforested due to increased efficiency of beef production and/or dietary shifts, as about 42% of the reforestation mitigation potential comes from reforrested tropical pasture land. However, the cost-effective portion of reforestation would only displace about 4% of global grazing lands, as displacing grazing activity has a higher economic cost than reforesting marginal lands.
Natural forest management
Additional carbon can be sequestered through natural forest management. This involves delaying harvests in native forests that are currently used for timber production, as well as reduced-impact logging and other sustainable harvesting practices. Decreases in wood production could be made up for by increased yields from existing timber plantations and additional wood production from reforestation.
The total mitigation potential assumes that timber harvests are deferred by at least 50 years across all native forests, while the cost-effective potential is mostly achieved through a transition to more sustainable logging approaches. The total sequestration potential for natural forest management ranges from 46-411GtCO2 with a best estimate of 74GtCO2. The cost effective potential is estimated at 44GtCO2.
Forest plantations managed specifically for timber production currently occupy approximately 7% of global forest areas and are expanding rapidly to meet the growing demand for timber products. Most plantations are currently run to optimise profits by cutting down trees relatively quickly. Less frequent logging could sequester more carbon and increase wood yields.
The total sequestration potential for improved plantations ranges from 11-66GtCO2 with a best estimate of 29GtCO2. The cost-effective potential is estimated at 17GtCO2.
Biochar is charcoal added to agricultural soils to sequester carbon. While potentially promising, it has not been demonstrated at scale outside of research settings and the potential costs and tradeoffs of a large-scale deployment remain poorly understood.
When converted into charcoal through pyrolysis, biomass can remain stable in soils over long periods. The NCS authors estimate that around 80% of carbon added to the soil via biochar would remain for at least 100 years.
Biochar is estimated to have a sequestration potential of 53-121GtCO2, with a best estimate of 91GtCO2 and a cost-effective potential of 27GtCO2.
Conservation agriculture involves planting cover crops, such as rye or clover, during the part of the year when the main crop is not being grown. This generally involves fields without year-round or winter crops where a fallow period is not required between plantings. These cover crops add additional carbon to the soil, a portion of which remains over long periods of time.
The NCS authors did not include no-till farming, which involves growing crops without disturbing the soil from tilling. They point out that recent studies have questioned the long-term carbon storage and N2O emissions associated with no-till practices.
Conservation agriculture has a sequestration potential of 16-26GtCO2, with a best estimate of 21GtCO2 and a cost-effective potential of 19GtCO2.
Trees in croplands
There are many areas in croplands that are not actively farmed, where trees could be planted without reducing crop yields. These include windbreaks, alley cropping (planting crops between rows of trees) and the scattering of trees within cropland for soil quality and erosion control benefits. The total sequestration potential for trees in croplands ranges from 23-93GtCO2, with a best estimate of 52GtCO2 and a cost-effective potential of 22GtCO2.
Optimal intensity grazing
Changing the pattern of animal grazing can increase grass growth and livestock production. This generally involves rotating cattle and similar animals across rangeland in dense herds, leaving areas to regenerate naturally between grazings.
Optimal intensity grazing can help increase soil carbon and is estimated to have a sequestration potential of 12-58GtCO2, with a best estimate of 12GtCO2 and a cost-effective potential of 7GtCO2.
Legumes in pastures
Plants in the legume family, such as alfalfa or clover, can be planted in pasturelands to help sequester additional carbon in the soil. For this estimate, the NCS authors exclude any regions where the additional N2O emissions from planted legumes would counteract the CO2 reduction. They estimate a sequestration potential of between 1-125GtCO2, with a best estimate of 12GtCO2 and a cost-effective potential of 11GtCO2.
Coastal wetland restoration
Restoring degraded coastal mangroves, salt marshes and seagrass can enhance how much carbon is stored in the soils, grasses and trees. Coastal wetland restoration is estimated to have a sequestration potential of around 5GtCO2 and a cost-effective potential of 1GtCO2.
Natural solutions for mitigation
In addition to measures that sequester carbon in soils and biomass, the NCS authors considered a range of natural measures that would mitigate or avoid emissions. These include measures across forests, agriculture, grasslands and wetlands.
Forest mitigation measures include avoiding deforestation and forest conversion, improving fire management with controlled burns and other measures intended to avoid severe fires, and improving cookstoves to reduce the harvest of woodfuel used for cooking and heating.
Grassland mitigation measures include avoiding the conversion of grasslands into cropland and avoiding N2O emissions by using fertiliser more efficiently. They also examine reducing methane emissions by changing the diets fed to cows and improved livestock breeds and management techniques to reduce the total animal numbers needed to supply the same level of meat and milk demand.
Wetland mitigation measures include improved rice cultivation techniques to reduce methane emissions, restoring peatlands and avoiding additional destruction of coastal wetlands or peatlands.
Taken together, these measures could reduce the need for negative emissions by allowing faster mitigation than is currently included in IAM scenarios. However, some of these measures – particularly avoided deforestation – are already included in future emission scenarios. Disentangling which of these solutions are or are not currently covered by IAM land-use models is beyond the scope of this article. Similarly, while the PNAS paper provides a detailed review, the “best estimates” they calculate for each category of NCS may differ from those found by other researchers.
Negative emissions vs. near-term reductions
If cost is not a concern, the total potential from NCS is between 236- and as much as 1,350GtCO2, with a best estimate of 549GtCO2. NCS have the potential to sequester 97-553GtCO2 between 2018 and 2100 at a cost of less than $100 per ton carbon. The PNAS paper’s best estimate of the total cost-effective potential is 225GtCO2.
IAMs range fairly widely in the amount of negative emissions they assume will be provided by BECCS, from as little as 151GtCO2 to as much as 1,191GtCO2. The amount of negative emissions in models is related to their underlying shared socioeconomic pathway (SSP), which limits how quickly global cooperation on climate change and rapid emission reductions can occur.
IAMs also include NCS, particularly reforestation/afforestation. The scale of NCS already included in IAMs varies considerably between models, with a median negative emissions from land use of 161GtCO2. One model, GCAM4, is particularly bullish on land use change, estimating negative emissions from NCS between 449 and 716GtCO2, more than twice as much NCS as is deployed in any other IAM and higher than the best estimate of cost-effective potential from the PNAS paper.
Models only need 401-1,137GtCO2 of negative emissions in the more sustainability-focused pathway (“SSP1”). They need much higher amounts – 1,198-1,596GtCO2 – in the rapid fossil fuelled-growth pathway (“SSP5”).
The figure below compares the total negative emissions between 2018 and 2100 assumed in IAMs that limit warming to below 1.5C (left) with an estimate of the range of negative emissions potential from both BECCS and NCS (right). The bars show the range between the lowest and highest published estimate. For NCS, the green bar shows the cost-effective potential, while the whiskers show the upper bound of the total potential.(Left) Cumulative negative emission deployment in IAMs limiting warming to below 1.5C in 2100 by SSP (note that SSP4 only has one available model), based on data provided by Rogelj et al (2018). Negative emissions are the sum of BECCS and negative land use emissions. (Right) Range of estimated cost-effective BECCS potential in the literature (see below) and range of NCS cost-effective potential (green bar) and total potential (whisker) estimated from Griscom et al. (2017).
The best-estimate cost-effective NCS potential is 225GtCO2 (diamond marker, green bar). This means BECCS could not be completely replaced with cost-effective natural negative emissions in any of the existing IAM model runs in the SSP database. However, the widespread adoption of cost-effective natural negative emissions approaches could reduce the need for BECCS by up to 56% one model, and by an average of 30% across all the models. As Carbon Brief has previously reported, additional even more stringent mitigation measures might make it possible to completely eliminate the need for BECCS.
The cost-effective NCS estimates are likely the most appropriate values to compare to BECCS. The amount of negative emissions in IAMs is largely driven by the assumption that late-century negative emissions will be more cost-effective than steeper short-term emission reductions. If cost were not a constraint, the potential of BECCS and other negative emissions technologies like direct air capture would also be higher. Similarly, the need negative emissions could be mostly eliminated by simply requiring near-term cuts of global emissions to near-zero.
The figure below, produced by Carbon Brief using data from the NCS study and the SSP database, shows emission trajectories for staying below 1.5C in 2100 from three different models. It examines the role that NCS could play if the amount of BECCS is minimised.Simplified example of how NCS could provide a portion of the negative emissions in different IAMs that limit warming to below 1.5C in 2100. Model CO2 emissions trajectories obtained from Rogelj et al (2018). Positive emissions are based on reported fossil fuel and industrial emissions plus BECCS. Negative emissions are based on BECCS and negative land use emissions.
In the REMIND SSP5 scenario, emissions are reduced more gradually than in other scenarios. This requires a large amount of negative emissions late in the century to avoid exceeding 1.5C in 2100. Here, cost-effective NCS can only provide around 19% of the negative emissions used in the model.
In an SSP2 world, more rapid reductions in emissions are possible, with emissions peaking in 2020 and declining quickly thereafter. The MESSAGE SSP2 scenario has more modest total negative emissions of 622GtCO2, around 36% of which could be replaced by cost-effective NCS.
Finally, in the AIM SSP1 scenario, emissions are reduced rapidly and additional cuts to non-CO2 emissions limit the total amount of negative emissions needed to 401GtCO2, 56% of which could potentially be provided by cost-effective NCS.
These suggest that cost-effective NCS has the potential to provide a sizable portion of negative emissions currently used in IAMs, provided that emission reductions occur quickly. Models with more gradual emissions reductions require much more negative emissions later in the century than cost-effective NCS can provide.
Will BECCS fall short?
IAMs employ cumulative negative emissions from BECCS of as much as 1,191GtCO2 between 2018 and 2100. However, it is far from clear that this potential is actually achievable, as BECCS is largely untested at scale. A number of recent studies have also questioned the high-end estimates of BECCS when emissions associated with converting land from forest or grassland for use in growing energy crops are properly taken into account.
The figure below shows the range of BECCS employed across different IAMs, as well as estimates of the total economic potential for negative emissions from BECCS by a number of recent studies.(Left) Cumulative BECCS deployment in IAMs limiting warming to below 1.5C in 2100 by SSP, based on data provided by Rogelj et al. (2018). The AIM SSP1 scenario is shown as a dot outside the IAMs range to emphasise that it is the only model showing values below 400GtCO2. (Right) Range of estimated BECCS potential in the literature from van Vuuren et al. (2013), Krause et al. (2018), Avoid2 (2015), and Turner et al. (2018).
IAMs that limit warming to 1.5C use a widely varying amount of BECCS. Only one model has estimates below 400GtCO2, however, with the median model in the database using 660GtCO2.
IAMs include land models that provide independent estimates of BECCS potentials, though they do not necessarily use the maximum amount possible. However, IAMs may not always account for all the emissions and other climate effects associated with the conversion of forest or grasslands to bioenergy crops.
More recent work has suggested that the carbon benefits of BECCS could be considerably smaller when converting grassland or forest to cropland to grow energy crops. For example a 2017 paper by Dr Mathilde Fajardy and Dr Niall Mac Dowell of Imperial College London estimates that it would take between six and 32 years for there to be any net carbon benefit from converting grazing land in Brazil to BECCS, due to the additional emissions associated with land-use change. It would take between 22 to more than 50 years for forest converted to cropland.
In a one of the earlier studies on BECCS potential, published in 2013, Prof Detlef van Vuuren and colleagues suggested that the potential for negative emissions from BECCS could be up to 10GtCO2 per year by 2050 and 20GtCO2 per year by 2020. This would amount to around 900GtCO2 total negative emissions between 2018 and 2100. However, they also cautioned that these numbers were highly optimistic and that “more realistic estimates…are likely to be half this potential or less”.
A study earlier this year by Andreas Krause and colleagues used four different dynamic global vegetation models to assess the potential of both afforestation/reforestation and BECCS to provide negative emissions. They looked to see how a cumulative negative emission target of 477GtCO2 could be met by the end of the century. While one of the models was able to achieve all 477GtCO2 from BECCS, most were only able to produce much lower amounts of sequestration from BECCS, with one model showing as little at 70GtCO2.
These large differences across models show that “the size of the removal is highly uncertain and may be much less than previously assumed in IAM scenarios”, Krause et al suggest. They found that these uncertainties mostly stem from different assumptions on crop yields and soil carbon response to land-use change.
Another study, by Dr Andrew Wiltshire of the Met Office and Dr T Davies-Barnard of the University of Exeter for the Avoid2 group examined the potential of BECCS worldwide. They found that while the gross potential for BECCS was quite large, up to 2,346GtCO2, when the additional emissions from land-use conversion are included, the net BECCS potential drops by three-quarters to a maximum of 476GtCO2.
An even lower estimate is found in a 2018 study from Dr Peter Turner and colleagues at the Carnegie Institution for Science at Stanford. Turner et al examined the amount of sustainably harvestable BECCS that would not involve conversion of forestlands. They found that around 3.4GtCO2 per year of negative emissions could be achieved from areas that overlap appropriate geological storage. This amounts to cumulative BECCS of around 238GtCO2, assuming that deployment begins in earnest around 2030. If BECCS is restricted to marginal lands over geological storage, the potential drops to 1.1GtCO2 per year, or around 77GtCO2 total.
All of these estimates of BECCS potential are subject to large uncertainties, as predicting crop yields and available land area in the latter part of the 21st century is a difficult task. As van Vuuren tells Carbon Brief:
“It is very hard to exactly give a limit on BECCS as one does not know yield improvement, response to price changes, etc… I would not say that calculations by the SSPs models are necessary unrealistic. On the contrary, most of the IAM teams behind these numbers have quite some knowledge on land-use based mitigation… It is important to realise that we cannot know the limits for BECCS in 2050 for sure. Some of these limits might not even be biophysical, but related to public acceptance (at the moment, both CCS and bioenergy are seeing opposition). Being a bit more on the lower side of the range might be regarded as taking less risks in terms of non-performance, but also of negative impacts on hunger of biodiversity. Our recent article in Nature Climate Change [recently covered by Carbon Brief] provides some options to reduce the need for BECCS – although it seems (nearly) impossible to avoid the need for negative emissions.”
Ultimately, NCS may be able to play an important role in reducing the need for BECCS in scenarios limiting warming to below 1.5C in 2100. This is particularly important given the uncertainty surrounding whether BECCS can meet the large negative emissions used in most 1.5C IAM scenarios.
NCS has many potential benefits over BECCS associated with enhancement of ecosystem health and services. However, scaling it up may prove challenging given the numerous small-scale landholders involved in land management worldwide.