NET ZERO EMISSIONS OF GREENHOUSE GASES BY 2050: ACHIEVABLE AND AT WHAT COST?

1. Introduction

As of late September 2022, about 140 countries have announced or are considering net zero emissions targets by 2050. The nature of these pledges differs and some countries such as China, India, and Saudi Arabia have set the target date after 2050 (Climate Action Tracker, 2022). In a smaller subset of these countries, the target has legal underpinnings, while for many countries the net zero goal has been announced but there is not a legal requirement to achieve it. There are also differences in whether all greenhouse gases (GHGs) and international bunkers within the country are covered, and whether international offsets are allowed. In many cases, the commitment remains somewhat vague regarding such details. The variety of pledges, target dates, lack of legal status, and vagueness of commitments leave open a variety of possibilities in terms of what these pledges will actually mean for global emissions.

As such, net zero emission scenarios have been the focus of a growing number of modeling exercises and research papers. For example, studies have investigated the potential impacts and contribution of the power sector to achieve net zero emission targets (Bistline and Blanford, 2021; Hafner et al., 2021), alternative pathways to achieve a net zero energy system (IEA, 2021), the potential contribution of carbon capture and storage and carbon capture and utilization to net zero targets (Hong, 2022), the need and timing for carbon removal technologies and its dependance on near term mitigation (Fyson et al., 2020; Strefl-er et al., 2021), the time dependency of net-zero CO₂ targets and non-CO₂ mitigation contributions (Ou et al., 2021), the timing of sectoral emissions contributions (Luderer et al., 2018), uncertainties on costs, scalability and public perception about technologies required to achieve net zero targets (Azevedo et al., 2021; Leppert, 2022; Sano et al., 2020; IAEA, 2011), costs of alternative stabilization pathways under alternative assumptions about temperature overshooting (Johansson et al., 2020), and innovation needs and drivers to achieve net-zero emissions (Stern and Valero, 2021). We seek to contribute to this literature by using an integrated earth system–economic model to investigate illustrative net zero emissions scenarios and their implications for the economy, energy and land use systems, and global temperature.

Strengths of the approach we bring to analysis of net zero scenarios are our comprehensive treatment of economic sectors, the inclusion of all major greenhouse gases and their sources and sinks including land use, and integration of human activity with an earth system model of intermediate complexity. Section 2 of the paper discusses some broader considerations regarding net zero targets. Section 3 reviews technological options available to achieve zero and negative emissions. Section 4 briefly describes the integrated earth system–economic model which we use to examine three illustrative “net zero” scenarios: (1) global net zero GHG emissions by 2050, (2) Europe and the USA with specific net zero GHG emissions by 2050 targets and the rest of the world with emissions targets consistent with achieving 1.5^∘C, with global emissions trading, and (3) the same as (2) but with Europe and the USA achieving their net zero targets within their own borders (i.e., not participating in emissions trading with the rest of the world). Within these scenarios we examine issues of emissions trading, negative emissions, energy mix and policy costs. Section 5 discusses the analysis results, and Section 6 concludes.

2. Net Zero — What Does It Mean and Why 2050?

The pledge of “net” zero recognizes that not all emissions can be reduced to zero and so those remaining will be offset by net negative emissions technologies and strategies. Stabilization of GHGs in the atmosphere to avoid dangerous interference with the Earth’s climate is the ultimate goal of the United Nations Framework Convention on Climate Change (UNFCCC). International negotiations have focused on keeping warming from pre-industrial levels below 2^∘C or 1.5^∘C as temperature targets that are hoped to prevent such dangerous interference. In the long term, stabilization of atmospheric levels of GHGs requires that any emissions to the atmosphere be balanced by removal.

The focus on net zero and 2050 for emissions targets appears to be largely a political choice as it is the cumulative emissions over time that determines concentrations. As in our 1.5^∘C scenarios, and as demonstrated in other work, Paris temperature targets can be achieved without net zero global emissions in this century. The Intergovernmental Panel on Climate Change (IPCC) Special Report on 1.5^∘C has 90 individual scenarios with a 50% chance of 1.5^∘C in 2100 (IPCC, 2018). Only 18 of those (20%) have net-zero CO₂ emissions (energy sector and industrial process CO${}_2)$ in 2050. In the latest IPCC Assessment Report (AR6), while all of the scenarios that limit warming to 1.5^∘C with at least a 50% chance and no or limited overshoot reach net zero CO₂ emissions, the timing ranges from 2035 to 2070, and only half of the pathways reach net zero GHG emissions at any point during the second half of the 21st century (IPCC, 2022). Deep cuts in the near term can allow for some emissions later in the century. Perhaps more likely, failure to begin reducing emissions well before 2050 could mean that the world would overshoot GHG concentrations consistent with temperature targets and then would need net negative emissions in later years to compensate (IPCC, 2018; Strefler et al., 2018). Commitments of net zero by 2050 may also be a recognition of the need for climate leadership, especially among wealthier countries, acknowledging the challenges faced by poorer countries to achieve such a goal.

Net zero as a goal means that released emissions need to equal negative emissions, and for atmospheric stabilization that would need to include all greenhouse gases and all sources. However, with regard to policy pledges there remain important questions about what exactly “net zero” means. First, to which emissions does the net zero target apply? Sometimes the target applies to CO₂ only, typically meaning fossil energy and industrial process CO₂ emissions. Sometimes it applies to all greenhouse gases and all sources. CO₂ emissions from land use and land cover changes (LULCC) may or may not be included, or a net zero goal may allow credits from verified reduction in land use emissions or increases in sinks.

Second, what about the pathway to net zero? A net zero goal by a pre-specified date, such as 2050, does not identify the path by which it will be achieved. As illustrated in Fig. 1, different paths can lead to the same end goal, but with quite different cumulative emissions, which ultimately are what is important in determining the climate outcome. Which gases are reduced and when can also affect the climate outcome — since methane has a relatively short lifetime, reductions have a bigger near-term impact, but if that allows more of the longer-lived gases such as CO₂ to be emitted those will then remain in the atmosphere longer making it more difficult to maintain targets. Different pathways to the end target also have different implications for the technologies employed and the cost of achieving the goal.

Third, for any given pathway to net zero, there is also a question of how much positive and negative emissions should be involved. The net zero target does not determine the gross emissions that remain (e.g., from fossil fuel combustion). A net zero target can, in principle, is achieved with deep reductions in emissions and small amounts of negative emissions, or quite large continued emissions and reliance on fossil fuel combustion (without carbon capture and storage) with large offsetting amounts of negative emissions (Fig. 2). There are some difficult-to-eliminate emissions (see next section) and activities for which there are currently no good alternatives, which means some amount of emissions are likely to persist, requiring negative emissions to offset them. One strategy is to reduce emissions as much as possible and offset the remaining difficult-to-eliminate emissions with negative emissions [e.g., Fig. 2(a)]. However, another strategy is to continue to release, and then offset, any emissions with marginal abatement costs greater than the cost per ton of negative emissions [e.g., Fig. 2(b)]. These strategies, and those in between, can result in the same net emissions path, and thus the same or similar climate implications, but quite different implications for technologies, cost of achieving the target, and other areas such as health (e.g., air pollutant emissions associated with fossil fuel combustion) and land use change (e.g., for afforestation or biomass for negative emissions).

3. Hard to Reduce Emissions and Available Technologies

“Hard-to-abate sectors” usually refer to those in the heavy industry, especially the production of iron and steel, cement, and chemicals, and those belonging to the heavy-duty transport, including trucking and aviation (Paltsev et al., 2022; Strategic Sustainability Consulting, 2021; Gross, 2021; Rumayor et al., 2020). Emissions from hard-to-abate sectors or technologies have a relatively lower mitigation potential, as options for decarbonization are more expensive, limited or not yet available.

At the global level, currently the heavy industry sector with the largest carbon footprint is iron and steel making (2.6 Gt-CO₂/year; 7% of total CO${}_2)$, followed by cement production (2.3 Gt-CO₂/year; 6.2% of total) and chemical industry (1.4 Gt-CO₂/year; 3.8% of total), i.e., the three sectors together are sources for about 17% of global CO₂ emissions (IEA, 2020).

It is worth noting that in our study, sectoral emissions, including emissions from hard-to-abate sectors, are direct emissions, which can be in the form of combusted emissions coming from burning fossil fuels, or process (non-combusted) emissions that are embedded in relevant production processes. For instance, the process emissions of iron and steel making constitute about 12% of its overall emissions, and for cement production and chemical industry, the numbers are 65% and 14%, respectively (IEA, 2020). Indirect emissions, emitted by other sectors providing inputs to the sector of interest, are separately accounted for in sectors of origin. As an example, the electricity used by iron and steel making may be provided by coal power plants, and the power generation emissions are allocated to the coal power sector.

Regarding heavy-duty transport, the aviation sector currently accounts for roughly 2% of global combusted CO₂. In recent years, the emissions topped in 2019 at about 1 Gt-CO₂, dropped to 0.6 Gt-CO₂ in 2020 due to the pandemic, went up to over 0.7 Gt-CO₂ in 2021, and are expected to grow and exceed the 2019 level soon (IEA, 2022a). Under this trend, reducing aviation emissions can be arduous, as adopting sustainable aviation fuels, hydrogen-powered or battery-electric airplanes are costly or are still under development (Axel Esqué et al., 2022). Likewise, knocking out emissions from heavy freight trucks, resulting in 1.2 Gt-CO₂ in 2021, is onerous and would require the technology breakthrough that lowers the costs of other low-carbon options (IEA, 2022b).

In addition, cutting emissions from certain activities of the agriculture sector appears to be a tough task as well, unless adopting alternative production processes or using inputs with a lower carbon footprint are feasible, as agriculture emissions can be mobile, dispersed, and therefore harder to capture. For instance, emissions from livestock are around 7.1 Gt-CO₂e/year (Gerber et al., 2013), growing rice produces around 12% of global methane emissions, or 1.5% of total GHGs emissions (Kurnik and Devine, 2022), and the use of nitrogen fertilizer contributes to roughly 2.4% of global GHGs emissions (Greenpeace International, 2021).

Pursuing a net-zero target by 2050 would require a combination of zero-to-low-carbon technologies, which help lower released emissions, and negative emission technologies/approaches, which can be used to offset persisting released emissions. There are many options within each category (Table 1), each with different costs, technology readiness, emissions implications and public perceptions. Zero-carbon and low-carbon options include renewables (wind and solar), hydro, bioenergy, nuclear, abated fossil fuels (e.g., fossil-generation with carbon capture and storage (CCS)). Negative emissions technologies include bioenergy with carbon capture and storage (BECCS) and direct air capture with carbon storage (DACCS), as well as natural-based solutions such as reforestation, afforestation, agricultural practices that sequester carbon in soils, and things like enhanced weathering and ocean fertilization. Griscom et al. (2017) estimated that the nature-climate solutions have the potential of contributing to about 20% of the necessary emissions cut between now and 2050 to keep the less than 2^∘C warming target in check. There is a tradeoff between these two categories in that the more deployment of zero- and low-carbon technologies, the fewer negative emissions are required, and vice versa. To achieve the net-zero target, it is unlikely that any single mitigation option would become the silver bullet, as each technology has its own pros and cons, and distinct regions have various endowments that affect the costs and viability of different technologies.

The potential of each mitigation option will depend on technological feasibility, climate policies, economics, resource constraints, etc. One challenge in many instances is that often the technologies are experimental at best, they have not been applied at scale, their cost and technical performance is highly uncertain and they require significant time to scale up. As a result, any representation of them in models such as ours is highly speculative — a relatively low-cost emissions-free DACCS that can be widely applied, scaled up quickly, and assuming largely unlimited ability to permanently store the capture carbon would result in low costs of meeting the target of net zero by 2050, and might well mean that the world economy can continue to utilize fossil energy resources with little switch to zero carbon alternatives. On the other hand, in “hard to reduce” sectors, failure to represent potential new technologies that can avoid emissions may result in a heavy reliance on net negative emissions technologies, or if those are not easily scaled up, suggest that net zero goals are very expensive or unachievable. How all of these issues are modeled, or not, will affect the mix of technologies employed in a given net zero scenario in our or any modeling framework.

4. Modeling Method

4.1. MIT IGSM

We employ the MIT Integrated Global Systems Model (IGSM), which is a coupled human-Earth system model. It combines the MIT Earth System Model (MESM, Sokolov et al., 2018) with the Economic Projection and Policy Analysis (EPPA) model (Chen et al., 2016, 2022; Paltsev et al., 2005), allowing for the development of emissions pathways consistent with different 21st century temperature outcomes. While some studies using the IGSM system have considered how aspects of the changing climate may affect the economy (e.g., Gurgel et al., 2021; Reilly et al., 2012, 2007), in the version applied here there are no climate feedbacks on economic activity. As a result, measures of economic cost reflect the mitigation cost which will be offset to some degree by avoided climate damages. Inertia in the climate system generally means that much of the climate change we might expect over the next few decades is already “built in” to the system, and so the lack of climate feedbacks in the model will have more effect over the longer term — post ^∼2075.

MESM simulates physical, chemical, and biological Earth’s systems, and is considered an Earth system model of intermediate complexity. All climate-relevant conditions across the Earth system are modeled by MESM, including atmospheric concentrations of greenhouse gases and aerosols, temperature, precipitation, ice and snow extent, sea level, ocean acidity and temperature as well as natural sources and sinks for greenhouse gases that may change due to changing climate and atmospheric composition. EPPA is a multi-sector, multi-region, computable general equilibrium (CGE) model of the world economy, with a recursive-dynamic structure. It represents the evolution of the global economy, projecting economic variables (GDP, energy use, sectoral output, consumption, prices, etc.), energy production and use and land use and land use change (cropland, pastureland, forest, and natural areas). EPPA is linked to MESM through the emissions of greenhouse gases and other pollutants. EPPA projects emissions of both long-lived greenhouse gases (CO₂, CH₄, N₂O, HFCs, PFCs, and SF${}_6)$ and short-lived air pollutants (CO, volatile organic compounds (VOCs), NOx, SO₂, NH₃, and black carbon and organic carbon aerosols) from combustion of carbon-based fuels, industrial processes, waste handling, agricultural activities, and land use change.

With regard to technology options relevant to achieving net zero, the EPPA model includes several zero carbon electricity options including solar, wind, and nuclear. It also includes carbon capture and storage (CCS) technology for the power sector, although the considered CCS technology only captures 90% of the emitted carbon. The assumed costs of generation technologies are based on Morris et al. (2021), following the approach described in Morris et al. (2019a). These costs, and assumptions about the penetration of new technologies (Morris et al., 2019b), are summarized in the Supplementary Material. In addition, bioenergy can be used in the power sector or as a substitute liquid fuel with the net greenhouse gas implications of bioenergy dependent on the endogenously determined land use effects. With regard to negative emissions technologies, the model includes biomass electricity with CCS. Land use change can also lead to increased carbon storage — with land use carbon emissions reductions/storage incentivized through carbon pricing. In private transportation, electric vehicles are a potentially zero carbon option when associated carbon policies eliminate carbon emissions from power production.

Elsewhere in the model, substitution among fuels and between fuels and electricity is represented with Constant Elasticity of Substitution (CES) production functions. Reductions in non-CO₂ greenhouse gases are also represented as a CES substitution of other inputs for emissions, capturing an increasing cost at greater reduction levels based on bottom-up assessment of technology options. In some cases (e.g., methane from fossil fuel production), bottom up studies show that a large fraction of emissions can be eliminated at relatively low cost. In other cases (nitrous oxide from nitrogen fertilizer use/soils), the options are much more limited. A general property of the CES production function is that all inputs are “necessary” and can never be completely eliminated, and so in cases where emissions control options are represented through the CES function they can never be zero. To the extent there may be completely different technologies where such emissions can be completely eliminated, the model may overestimate the need for negative emissions technology. Currently, the model also does not represent carbon free gas (hydrogen, sustainable biogas) or direct air carbon capture technologies.

4.2. Scenarios

We apply the MIT IGSM to examine three illustrative net zero scenarios (Fig. 3). In all scenarios, net zero is defined as covering all greenhouse gases, including those from LUCLCC, which are assumed to be covered by the policy and subject to the same emissions price as energy sector and other emissions. Note that reported emissions are those attributable to human activity as represented in the EPPA component of the IGSM. Changes in the ocean carbon sink, naturally driven changes in land carbon uptake (e.g., CO₂ fertilization in forests) and in CH₄ (from natural wetlands), and N₂O emissions (from soils not associated with application of fertilizers) are simulated in the MESM.

In the first scenario (Global NZE), we consider the case where the world achieves net zero GHG emissions by 2050, and maintains the net zero level through 2100, the horizon of the model, and global emissions trading is allowed. In this scenario, regional emissions pathways decline linearly from 2025 to zero in 2050. With emissions trading, net emissions in some countries remain above net zero through the purchase of allowances from other countries which have net negative emissions, such that emissions for the world as whole are at net zero.

In the second scenario (1.5^∘C), Europe (EUR), and the USA have targets of net zero GHG emissions by 2050 while the rest of the world has emissions targets consistent with their NDCs to 2030 and then with the long-term goal of the Paris agreement to stay below a temperature rise of 1.5^∘C above pre-industrial levels with a 50% likelihood (given an assessment of uncertainty in climate response to greenhouse gas forcing in MESM), with global emissions trading starting in 2035. In this scenario, we assume different years for peak emissions and different rates of decarbonization for G20 countries (except for India) compared to India and other developing countries. The assumed regional GHG emissions constraints are provided in the Supplementary Material.

The third scenario (1.5^∘C_notrade) has the same settings (same regional emissions constraints and therefore same global emissions) as the 1.5^∘C scenario, but with EUR and USA achieving their net zero targets within their own borders (i.e., not participating in emissions trading with the rest of the world and not utilizing international credits).

Within these scenarios, we explore emissions, energy, land use and costs. While our net zero reporting includes only sources and sinks directly related to human activity, the 1.5^∘C pathway was determined by including consideration of natural system feedbacks on emissions and sinks. We also examine uncertainty in the climate outcome given these deterministic emissions targets, by running ensembles of MESM simulations using different values of climate parameters affecting Earth System response to GHG emissions, (e.g., climate sensitivity and the rate of heat and carbon uptake by the deep ocean), sampled from distribution described in Libardoni et al. (2019) and Sokolov et al. (2018).

As we will show, both scenarios have a slight temperature overshoot at median climate settings, but the Global NZE pathway ends the century with global temperature well below 1.5^∘C. Importantly, while the 1.5^∘C scenarios reach near-zero emissions levels by the end of the century, global net zero GHG emissions are not achieved. And, again, while our emissions reporting focuses on those directly related to human activities, the climate results include any natural system emissions and sink changes. Uncertainty in those responses is one reason for uncertainty in the climate outcome despite a fixed human activity related emissions path.

5. Results

5.1. Emissions

In both the Global NZE and 1.5^∘C scenarios, the policies are met by utilizing large amounts of negative emissions (Fig. 4). The Global NZE immediately begins eliminating land use emissions, turning land into a source of negative emissions (via afforestation) by 2030. Afforestation is among the most cost-effective carbon mitigation options, but competition with other land uses and availability of areas with soil-climate-environmental conditions to grow forests, represented in EPPA in accordance with Griscom et al. (2017) and Roe et al. (2021), lead to limits to afforestation which are quickly reached by 2050 when afforestation negative emissions peak. After that negative emissions from afforestation decline as the previously planted forests saturate, and negative emissions from BECCS grow thereafter. Negative emissions are about 25 GT/yr for the second half of the century. These negative emissions offset non-CO₂ GHGs as well as energy fossil CO₂ and small amounts of industrial CO₂ which continue to be released. The 1.5^∘C scenario shows a very similar pattern with the main differences being higher amount of released emissions at mid-century and the more gradual penetration of negative emissions, both of which are driven by the more gradual changes in emissions constraints. Late in the century, this scenario has about 27 GT/yr of negative emissions.

Due to global emissions trading in both of these scenarios, some countries have net emissions above their targets, purchasing emissions allowances from countries that have emissions below their targets (Fig. 5). For Global NZE, regional emissions (blue bars) should be compared to a cap of zero emissions in all regions, and regions below zero are permit sellers while regions above zero are permit buyers. For 1.5^∘C, regional emissions (orange bars) must be compared to the regional caps (gray bars), and regions with emissions below the caps are permit sellers while regions with emissions above the caps are permit buyers. For both scenarios, Africa (AFR) and Latin America (LAM) are the largest sellers of permits (producing negative emissions via both BECCS and afforestation), with Brazil (BRA) being an important seller in 2050 as the largest producer of negative emissions via afforestation. China and India are the largest permit buyers, with the USA and EUR also playing a role as buyers. International emissions trading allows for flexibility in meeting the targets in each region while ensuring the global target is met in the most cost-effective way.

5.2. Energy

In terms of primary energy and electricity generation mixes, the Global NZE and 1.5^∘C scenarios are very similar, with the 1.5^∘C using slightly more bioenergy in the form of BECCS (Fig. 6). Both scenarios have a drop in primary energy use from 2020 to 2050, though it is more rapid in the Global NZE case due to the greater near-term stringency of the policy. Of the coal and gas primary energy used in the second half of the century in both scenarios, much is used in conjunction with CCS, such that unabated coal energy is less than 15% of all coal used by 2100 and unabated gas energy is less than 60% of all gas used. The unabated coal and gas are mainly used by energy-intensive industries. In terms of electricity, coal generation without CCS is phased out globally by 2040 in Global NZE and 2045 in 1.5^∘C. Although nearly all gas generation in the second half of the century is accompanied by CCS, small amounts of unabated gas generation remain in the mix throughout the century for system balancing.

Oil is persistent in the primary energy mix in both scenarios. While personal transportation turns to electric vehicles, other modes of transportation — aviation, shipping, long-haul trucking, rail — continue to rely on oil as the fuel source. Oil use persists in energy-intensive industries as well. Large amounts of negative emissions in each of the scenarios are thus being used to offset continued oil use in transportation as well as oil, gas and coal use in energy-intensive industry.

Of course, the specifics of these energy and electricity mixes depend heavily on the technology assumptions in the model. Different cost assumptions could more heavily favor renewable electricity over CCS in the power sector, while representation of industrial CCS options would likely see sizable deployment of CCS in energy-intensive industry. However, given the high cost of many commercial transportation fuel alternatives and other abatement options for certain industries, combined with the promising economics of BECCS in the power sector, economic deployment of BECCS is relatively robust to changing assumptions about the costs of other technologies. Its deployment however could be limited by political factors, sustainability concerns, public acceptance, and other factors.

5.3. Land use

There are considerable land use changes that accompany the transitions underlying each of the scenarios (Fig. 7). The demand for negative emissions drives afforestation and BECCS, the latter of which needs biomass crop feedstocks. As a result, in both scenarios, there is a significant expansion of natural forests and land for bioenergy from current levels. The Global NZE scenario increases natural forest and bioenergy land earlier than the 1.5^∘C scenario due its earlier ramp up of the policy. By the end of the century there are about 420 additional million hectares (MHa) of natural forest land compared to 2020 under the Global NZE scenario and about 370MHa more under the 1.5^∘C scenario. These scenarios both end up in 2100 with about 800 moreMHa of land used for bioenergy than in 2020. In both scenarios, all of the additional natural forest and bioenergy land displaces a combination of natural grassland, managed forest, pasture and, to a lesser extent, cropland.

5.4. Cost

Looking first at global emissions prices (Fig. 8), they rise much more rapidly and to a higher peak in the Global NZE scenario given that scenario’s earlier ramp up of constraints on emissions. The emissions price peaks at about $1200 in 2050 in the Global NZE scenario. The high peak is driven by limits to the ability to ramp up new technologies quickly, which is represented in the model (Morris et al., 2019b). Given the low cost of afforestation, BECCS does not become economically competitive until 2040, and as a new technology it takes time to build up the capacity to expand BECCS deployment. As a result, a high emissions price is needed to meet the net zero constraint. After the peak price, there is an initial decline in the carbon price as BECCS gets deployed at large scales, and then a gradually increasing price through the rest of the century. While we place no specific physical limit on the availability of BECCS in the model, its supply can be limited by competing uses for land (e.g., for food) as well as the price on land use change emissions which could deter further land conversion for biomass crops. The 1.5^∘C scenario has a similar, but less extreme, carbon price pattern — a mid-century peak at about $500, followed by a slight decline and then a gradual increase, ending the century at about $600. An interesting feature of BECCS is that it jointly produces electricity and negative carbon emissions. With more demand for BECCS over time, there is more electricity production, and as result the electricity price can be lower than it would be otherwise, meaning more of the cost of BECCs must be covered by the carbon price. This can also contribute to a rising carbon price even though the basic technology cost of BECCs is constant. Also, we report no global carbon price in the 1.5^∘C scenario until 2035 because earlier reductions under the Paris pledges through 2030 are achieved with a mix of national policies and measures.

Stringent policies such as the Global NZE and 1.5^∘C scenarios have a sizable impact on economic consumption, shown in Fig. 9 in terms of percentage change in consumption relative to a business-as-usual (BAU) scenario (which assumes increases in renewable penetration consistent with IEA estimates, but no additional climate policy). As noted earlier, these estimates do not include benefits of avoided climate damage. Despite the fact that global emissions fall more quickly in the Global NZE scenario, we see higher consumption losses in early years in the 1.5^∘C scenario. This is driven by the region-specific emissions constraints in 2020–2030 based on countries’ NDC pledges to the Paris Agreement and the assumption of no global trading prior to 2035. From 2035, the Global NZE scenario has a more rapid decline due to greater stringency, reaching peak losses in 2055 of almost 16% globally before improving gradually and ending up with about 12% losses in 2100. The high peak at mid-century is driven by insufficient availability of abatement technologies at that time, including BECCS. As investments in BECCS and other mitigation options are made, the economy becomes better adjusted to a zero emissions world, reducing the consumption impact of the policy. In contrast, the 1.5^∘C scenario has more gradual consumption losses that continue to fall throughout the century, somewhat flattening out by 2070 and ending the century with about 10% losses in 2100. This reflects emissions constraints that are more gradual and continue to decline throughout the century. While these consumption losses, relative to the baseline, might be considered substantial, the world economy continues to expand. Average annual consumption growth from 2020 to 2100 in the baseline is 2.2%, falling to 2.05% and 2.08% in the Global NZE and 1.5^∘C scenarios, respectively.

In both scenarios, the regional distribution of these consumption losses varies greatly (Fig. 10). Brazil (BRA) and the rest of Latin America (LAM) have consumption gains compared to BAU as they sell significant amounts of emissions permits by generating large amounts of negative emissions. These gains are higher in the Global NZE scenario as there is greater demand for negative emissions since all countries must achieve net zero emissions, not just USA and EUR. In particular, Brazil and LAM in 2050 have consumption improvements of nearly 80% compared to BAU in the Global NZE scenario as they can cheaply generate negative emissions from afforestation which are extremely valuable given the globally stringent policy. Russia also sells emissions permits from afforestation in the scenario, seeing consumption gains in 2050 as a result. Africa is also a seller of emissions permits, which helps to minimize consumption losses in that region. The most obvious loser in these scenarios is the Middle East (MES), which has consumption losses nearing 70%. As an oil-dependent region, these policies are very costly to the region, which relies heavily on buying offsets.

An important caveat regarding the regional consumption changes is that they are highly dependent on the initial assigned emissions targets. While the 1.5^∘C scenario assigns different reduction efforts by region, the Global NZE scenario assigns proportional reductions across all regions without consideration of differential abatement potential. The regional consumption losses we simulate can be seen as an indicator of how one might want to reassign regional targets, if a better balance of consumption changes across regions is desirable. For example, wealthier countries whose emissions are growing slower in the baseline tend to have smaller consumption losses, and might be assigned emissions targets with deeper, even negative emissions goals. Those regions with much potential for reforestation and BECCs also might take on deeper cuts, to ease the burden on other regions without those resources. Regions with considerable fossil energy resources, whose value is diminished by the climate policy, might get a larger allocation. And, despite some sizable consumption losses relative to BAU, all regions continue to experience consumption growth over time. Figure 11 shows how the average annual consumption growth rate from 2020 to 2100 differs in the BAU, Global NZE and 1.5^∘C scenarios. With the exception of the Middle East, the stringent policies shave only a small fraction off the regional average annual growth rates.

5.5. Climate

The MESM climate results (Fig. 12) show that median warming in both the Global NZE and 1.5^∘C scenarios slightly exceed 1.5^∘ above pre-industrial (1861–1880 average) by 2050. As noted earlier, there is inertia in the climate system and so not surprisingly there is very little difference in the two scenarios through 2050. The chance of peak temperature over the century remaining below 1.5^∘C in the Global NZE scenario is 43%, dropping to 37% in the 1.5^∘C scenario (Table 2). After 2050, the temperature in the Global NZE scenario drifts down to a median of 1.2C. The 1.5^∘C scenario was constructed to generate a 50% chance of being below 1.5^∘C in 2100, and the results confirm that this is the case. Also, of note is that with the Global NZE scenario there is almost no chance of exceeding 1.5^∘C in 2100, while in the 1.5^∘C scenario the upper (95%) limit on temperature is as much as 1.85^∘C. These limits reflect our current assessment of uncertainty in the earth system based on optimal fingerprint results (Libardoni et al., 2019). With more observation and research some of this uncertainty may be resolved, allowing possible course correction to achieve specific temperature outcomes.

5.6. Achieving net zero without international offsets

For the 1.5^∘C scenario, we also consider an alternative in which Europe and the USA do not participate in global emissions trading with other countries and therefore must achieve their net zero targets within their own borders, without the use of international credits (1.5^∘C_notrade). Whether or not they participate in international emissions trading has significant implications for the emissions, technologies and costs in the USA and Europe. It also affects regions that would be sellers of permits to the USA and Europe, such as Brazil, other Latin America, and Africa.

With emissions trading, net emissions are positive throughout the century in the USA. Relative to the USA, in Europe, net emissions are quite low in the trading case and even slightly negative in some years (i.e., in 2060). Note that in the 1.5^∘C scenario both regions are assumed to follow their Paris NDC’s through 2030 and then move to the global trading regime and that access to international offsets results in a short-term increase in net emissions. Without emissions trading, the net zero requirement must be met in 2050 and thereafter in both countries. Gross emissions are generally lower in both regions with no trade (Fig. 13). In the USA, with trade, BECCS is never employed, as international offsets are used instead. However, without trade BECCS makes a substantial contribution. Carbon uptake by forests makes a somewhat larger contribution in the middle of the century without trade in the US.

In contrast, in Europe, BECCS plays more of a role with emissions trading than without. Upon investigating this surprising result, we found that this is driven by afforestation. Without trade, Europe chooses to pursue more afforestation in early years of the policy than in the scenario with trading since in the absence of international offsets afforestation provides a relatively low-cost source of negative emissions for Europe. That early afforestation then locks land into forest for the remainder of the century, as the carbon price, which covers land use change (LUC) emissions, would strongly discourage deforestation. As a result, in later years of the policy, when there is a strong incentive to pursue BECCS, there is less land available for biocrops as it has already been used for forests. There is therefore less BECCS deployed in the 1.5^∘C_notrade scenario than under the 1.5^∘C scenario with trading. If Europe does not participate in emissions trading but LUC emissions are not covered, it would pursue more BECCS than in the 1.5^∘C with trading scenario. The USA has sufficient land to pursue both afforestation and BECCS, so does not face limited BECCS under the 1.5^∘C_notrade scenario like Europe does.

Total energy use is lower in both regions in the 1.5^∘C_notrade scenario, especially for gas and oil (Fig. 14). In the US with trade, coal reappears later in the century, with CCS (see also, Fig. 15), but it is almost nonexistent in the no trade scenario–the remaining 10% of emissions not captured keep it out of the mix. As seen with net emissions, in the 1.5^∘C scenario when Europe and the USA switch to a global trading regime in 2035 there is a bump up in energy use while in the 1.5^∘C_notrade scenario domestic mitigation must continue more consistently, leading to a smoother early transition. Aligned with the emissions results, under 1.5^∘C_notrade the USA uses more bioenergy than in 1.5^∘C while Europe uses less.

As already noted, in terms of electricity generation technologies, a major difference between 1.5^∘C and 1.5^∘C_notrade is the use of BECCS (Fig. 15). In the USA, in the no trade scenario, BECCS is deployed at large scale, replacing coal with CCS and partly replacing gas with CCS. Gas generation without CCS is also phased out of the mix sooner. In Europe, less BECCS is deployed in the no trading scenario and overall electricity generation is reduced. As described above, this is because without trading, Europe employs more afforestation in early years of the policy, reducing the amount of land that is available for BECCS in later years.

Emissions prices in both the USA and Europe are significantly higher in the 1.5^∘C_notrade scenario [Fig. 16(a)]. With trading, the global emissions price ends up around $650/tCO2e. Without trading, the 2100 price is more than doubled, ending up at about $1725/tonCO2e in the USA and $1800/CO2e in Europe. In terms of policy costs, measured in terms of percentage change in consumption relative to business-as-usual, the lack of trading increases costs, with a greater impact on Europe than the USA [Fig. 16(b)]. This is driven by insufficient land availability in Europe — afforestation is pursued early on, limiting the land available for biocrops for BECCS in later years. Without sufficient domestic negative emissions in the second half of the century, policy costs grow in Europe, ending up with consumption losses of almost 13% in 2100, compared to less than 5% under the 1.5^∘C scenario with trading. Consumption is still growing under 1.5^∘C_notrade (Fig. 17), just slightly less than with trading. Figure 17 also shows that regions that would be sellers of permits to the USA and Europe, such as Brazil, other Latin America, and Africa have slightly lower consumption growth in the 1.5^∘C_notrade scenario.

Consumption in other regions (e.g., Russia, Asia and the Middle East) are also impacted by terms of trade effects under the 1.5^∘C_notrade scenario.

6. Conclusions

We analyzed three illustrative scenarios to assess the economic cost and energy and environmental implications of achieving net zero emissions by 2050. Our Global NZE scenario assumed all countries were assigned a linear emissions target beginning in 2025 going to net zero by 2050, and maintaining it through 2100, the horizon of our modeling experiment. The scenario allowed all regions to trade emissions allowances over the entire period. Our 1.5^∘C scenario assumed only the USA and Europe had net zero by 2050 targets within the context of a global emissions path consistent with being below 1.5^∘C above pre-industrial in 2100 with a 50% likelihood given an assessment of uncertainty in climate response to greenhouse gas forcing. In a nod to greater realism, we modeled countries as staying with their Paris NDC’s through 2030, and only shifting to a global policy regime with trading beginning in 2035, and specific emissions constraints varying by region. With emissions trading individual regions within our modeling framework did not necessarily achieve net zero–some overachieved and sold allowances to underachievers. In a third scenario, we enforced the 1.5^∘C global path, but required the US and Europe to achieve net zero within their borders — they were not allowed to participate in the emissions trading scheme that other countries used.

Our main conclusions include:

Overall, our results indicate a tradeoff between policy costs and likelihood of achieving 1.5^∘C. Targeting global net zero emissions by 2050 is not required to achieve 1.5^∘C climate stabilization, and adds considerable policy costs, especially at mid-century. However, achieving global net zero by 2050 essentially guarantees that the 1.5^∘C target will be achieved, compared to having a 50–50 chance in the scenario without net zero. If net zero by 2050 is chosen as a goal, allowing for international offsets will significantly reduce the cost of achievement.

Acknowledgment

The EPPA model employed in this study is supported by an international consortium of government, industry and foundation sponsors of the MIT Joint Program on the Science and Policy of Global Change (see the list at: https://globalchange.mit.edu/sponsors/current).

Supplementary Information

The Supplementary Materials are available at: https://www.worldscientific.com/doi/suppl/10.1142/S201000782340002X