Can tech-based carbon removals mitigate climate change?

Direct air carbon capture and storage (DACCS) — using machines to filter CO₂ out of the air and pumping that CO₂ back into the ground — is one example in a long list of climate mitigation technologies. DACCS is a type of carbon dioxide removal (CDR). This means that it can cancel out carbon emissions from any other source, including those done in the past. While DACCS is a proven technology with several plants either already running or under development, the current scale of operation is minuscule in the context of global greenhouse gas emissions. Some, especially those described as “techno-optimists,” believe that with continued support DACCS can scale to a massive global industry, providing gigatons of carbon dioxide removal and helping to stabilize the climate. Others are more skeptical: they argue that the expense and resource intensity of the DACCS industry will render any potential impact too little and too late. Accordingly, those who doubt the industry argue that investing in DACCS is a dangerous distraction from genuine decarbonization.

Recent research sheds light on the merits of these two perspectives. One study, “Deploying direct air capture at scale: How close to reality?” published in the Journal of Energy Economics, explores a wide range of possibilities for the evolution of DACCS. Another study, “Diverse carbon dioxide removal approaches could reduce impacts on the energy-water-land system,” published in Nature Climate Change, considers DACCS alongside a broader portfolio of potential negative emissions technologies. Both studies use Integrated Assessment Models (IAMs) — models of future scenarios that consider technology choices and resource use across the entire economy — to evaluate the role that DACCS could play in a decarbonized future.

Researchers have been using Integrated Assessment Models to explore future climate change scenarios for decades, helping policymakers understand the interplay between technology changes, greenhouse gas emissions, and the climate system. However, until just a few years ago, carbon dioxide removal technologies were either not included in these models or limited to a single technology. Given the increasingly recognized unavoidable need for negative emissions to stabilize the climate and the ever-increasing urgency of their deployment with each passing year of largely unabated global emissions, researchers have been working to fill this gap.

Most modeled scenarios start with the premise that we will meet a certain climate goal. They almost always set at a warming level of 1.5 or 2°C above pre-industrial levels in 2100 and then ask the question, “What combination of technologies would allow us to achieve this outcome in the cheapest way?” It’s important to note that the models don’t address how — or if — policies will be put in place to achieve this target. They assume it will be done.

It has already been robustly shown that to limit warming levels to 1.5°C by 2100 — and, increasingly, 2°C — we will need both dramatic, rapid emissions reductions and a significant volume of carbon dioxide removal. Both studies’ modeled scenarios follow this pattern: global greenhouse emissions immediately drop precipitously from their current level of ~55 GtCO₂e yr^-1 (gigatons of CO₂ equivalent per year) to well under half that level by 2050. Negative emissions scale up more slowly, growing mostly in the second half of the century to reach levels of about 20 GtCO₂ yr^-1 by 2100.

The sheer scale of this transformation means that any meaningful contribution will also invoke a substantial demand for global resources. Many CDR technologies rely on land directly, whether through growing biomass for energy, spreading rock dust for enhanced weathering (EW), or simply increasing forest cover — and there’s only so much land available. In contrast, DACCS doesn’t use a large land area. Instead, it requires substantial amounts of (net zero carbon) energy to run the machines. This energy could be generated with dispersed renewables, an approach that would demand large amounts of land. Other options include fossil fuels with carbon capture and storage and nuclear, which, while less land-intensive, depend on other resources. IAMs allow the measurement and trading-off of these diverse impacts.

Desport et al. trade-off DACCS with only bioenergy with carbon capture and storage (BECCS) and land-use change as other CDR options. This choice matches previous studies, which often only included BECCS as a proxy for all CDR. Without DACCS, they estimate that BECCS alone would use 8.4 million km² of arable land for biocrops — about the same area that we currently use for growing food.

Fuhrman et al. include a broader range of other CDR technologies. As well as BECCS, they model spreading both rock dust (for enhanced weathering) and biochar on existing agricultural land. While this doesn’t ‘use up’ any land as the same land is also used for growing crops, only a finite area is available. They consider deployment at the maximum physical potential for both these technologies, using all available cropland across the globe and requiring the transport of 13 Gt yr^-1 of rock to achieve 5.5 GtCO₂ yr^-1 of removals in 2100. This level of sequestration is far from guaranteed to be safely possible and is far short of the ~20 GtCO₂ yr^-1 of total CDR required — implying that BECCS, DACCS, and/or some other technology must make up the difference.

Therefore, both studies show that not using DACCS means relying on either devoting a huge area of land to growing biocrops or using some other CDR technique that wasn’t included as an option in these models. Given that these are highly uncertain possibilities, this could mean that large-scale DACCS is required — no matter its price or the challenge of deploying at scale. In particular, Desport et al. find that in a scenario with limited land for BECCS, DACCS would still need to be deployed at significant levels (a cumulative total of 173 GtCO₂ by 2100) even if it remains very expensive.

Supposing, optimistically, that one or more of the other non-DACCS CDR options is available at a sufficient scale to provide all required negative emissions, the question of DACCS’ role in climate mitigation then boils down to how it competes. In the framework of IAMs used by these studies — and, arguably, in the real world — policymakers or removal-credit-purchasers will opt for whichever of the available technologies can remove carbon at the lowest price.

By making assumptions about the base price for each technology and modeling and tracking the prices of required inputs such as energy, transport, labor, and capital over time, these studies consider plausible future price scenarios for each CDR technique in the context of the broader socioeconomic and technological shifts we expect over the rest of the century.

Fuhrman et al. consider a narrower range of cost forecasts: they use a central scenario with plausible but optimistic prices for DACCS and all other CDR options. Based on this, they find that DACCS delivers a minority, but still significant, level of around 5.5 GtCO₂ yr^-1 of removals by 2100, or about 16% of the cumulative CDR total. They project that DACCS will be entirely outcompeted by enhanced weathering over the next few decades, but once enhanced weathering is fully ‘maxed out’ DACCS too will rapidly scale up to meet demand.

In contrast, Desport et al. explore a wide range of base prices (between $180 to $1,000/tCO2) and capture technology and energy supply options for DACCS. At their lower cost estimates, they find that DACCS fully out-competes other CDR options and rapidly scales up starting in 2040 to almost 25 GtCO₂ yr^-1 by 2100 — comparable in scale to today’s oil industry. On the other hand, at their highest cost estimate and with no limits on land use for BECCS, virtually no DACCS is deployed.

Desport et al. supplement this analysis with a further constraint. They consider a scenario without a global market for carbon emissions, but instead 18 separate regional markets. This means that each region must achieve net zero individually and cannot rely on purchased carbon credits from other regions to offset remaining emissions. Within this context, resources such as land for growing biomass and energy generation may be significantly limited in some regions (and therefore expensive) but plentiful (and therefore cheap) in other regions. In this scenario, even with the highest plausible base cost, DACCS remains competitive in some regions — especially China and India, where land is in high demand for growing food. This results in significant global DACCS deployment (135 GtCO₂ by 2100).

Overall, both studies show that under many plausible future scenarios, DACCS is expected to provide a significant and essential role in climate mitigation. At low or moderate price estimates, DACCS could be a competitive option for at least a significant fraction of total CDR, meaning we’d see a massive global industry scaling up especially over the later part of the century — potentially rivaling the scale of the current oil industry. It is also possible that DACCS is fully outcompeted by other CDR options and this industry doesn’t arise, if DACCS costs remain high and/or if other options turn out to be plentiful and cheap. However, if other CDR options are limited by pressures on land use or other physical constraints, or if there isn’t a fully global emissions trading market, DACCS could still be an essential mitigation tool, even if it is expensive.

While the future is uncertain in many ways — including the possibility that another CDR option not included in these studies will prove safe, scalable, and cheap — this research provides evidence that DACCS could be a critical tool in a future that involves successfully mitigating climate change. This means that effective policy support for DACCS over the next few decades, which is expected to be a major influence in delivering lower costs and the ability to scale, could be instrumental to achieving this success. However, this situation is only true under the assumption that we can achieve dramatic emissions reductions and are willing to pay for emissions removals, which is still an open question.