The Sky 2050 scenario is designed to meet the goal of net-zero (anthropogenic) CO₂ by 2050; it is a “high overshoot” 1.5°C pathway (i.e., C2 category), based on IPCC’s definition, sitting in the middle of the IPCC scenarios’ C2 range (Sokolov et al., 2023). The scenario is unique among 1.5°C scenarios due to its detailed consideration of NBS, using country level modeling for both energy and the development of NBS, in an integrated narrative. NBS achieve atmospheric CO₂ reduction through the conservation, restoration, or altered management of natural ecosystems (Waring et al., 2023). The Sky 2050 scenario seeks to recognize that implementing change, scaling up technologies or rolling out changes in land-use takes time to build when they are in their infancy today. The land-use analysis in the scenario draws heavily on the academic literature reviewed by the latest assessment of scientific findings (Sixth Assessment Report) by the Intergovernmental Panel on Climate Change (IPCC, 2023) for the long-term and cumulative potentials for NBS.

In particular, data developed by Cook-Patton et al. (2020) and the underlying references of Roe et al. (2021) are used for the modeling of the scale-up of NBS and the total ultimate potential land areas and cumulative CO₂ impact for protection, improved management or restoration. While the Sky 2050 scenario is highly ambitious, it does apply cost constraints for all NBS activities that require a change in land use or management (e.g., reforestation, regenerative agriculture, peatland/forest management). In related literature this is commonly referred to as the “cost-effective” available area and implies that implementing the NBS activity would require no more than 100 USD/tCO₂ (Roe et al., 2021; IPCC, 2023). The analysis is built up from a combination of changing land areas under management together with profiles of CO₂ stored and taken up through time in different ecosystems. Temporal profiles are based on an analysis that uses country-, NBS pathway-, and scenario-specific profiles of how hectares under management (i.e., “enrolled”) scale up.

The approach assesses how five categories of barriers (economic, political, technological, socio-cultural and environmental) will affect the deployment rate, separately across 19 different NBS pathways for 200+ countries, for each scenario. For simplicity, all pathways start this enrollment in 2023, but many do not reach material scale for several years or even decades, dependent on the pathway and country. The greater the number of barriers to overcome, the slower the scale-up to the peak rate (ha/yr) of enrollment. The calibration of peak rates is based on a combination of expert elicitation, comparison with historic enrollment rates of ambitious NBS projects (e.g., US national forest expansion in the 1930s, large-scale reforestation campaigns in China in recent years) and validation against current ambitions (e.g., United Kingdom net-zero reforestation ambitions – Committee of Climate Change, 2020). The resulting anthropogenic land-use emissions fall from a net source of 4 GtCO₂/yr today to a net sink of −6.1 GtCO₂/yr in 2050 (Sokolov et al., 2023). The cumulative anthropogenic land-use emissions in 2023–2100 in the Sky 2050 scenario are −312 GtCO₂.

The second part of the calculation uses existing literature (Roe et al., 2019) to distribute rates of CO₂ removal and avoidance on a unit area and time basis (i.e., tCO₂/ha/yr). This uses country-specific rates for each different NBS type and applies a temporal distribution that describes that activity’s temporal impact on mitigation (e.g., 1 hectare of avoided deforestation generates all emission reductions at the time the area is protected, whereas emission reductions generated by reforestation follow a lognormal distribution starting from the year trees are planted). For more detail on this calculation, see Barros et al. (2023). Many of these temporal profiles are generalized across several countries or climatic zones, but do represent the fact that reforestation in tropical zones may reach peak CO₂ sequestration rates within 6 years whereas reforestation in mid or high latitudes may take 10+ years after saplings have been planted.

Additionally, forests in different regions may reach equilibrium and limit additional sequestration over different time horizons (Cook-Patton et al., 2020). Avoidance pathways use the most recent assessment of current carbon stock levels. For some, such as avoided deforestation, if a hectare is protected then the emission reductions are accounted for in the year in which the hectares are enrolled, but the land area is considered enrolled and managed indefinitely, despite no longer generating any emission reductions. For other NBS activities, such as fire management in savannahs, avoided emissions can be ongoing. An important result from the two steps of the calculation is that both the scaling up of capacity to enroll hectares in land management programs, and factoring in the time from enrollment to CO₂ uptake, act as lags in removing CO₂ from the atmosphere. This bottom-up approach produces projections of land use that account for realistic constraints to the deployment of NBS.

To assess land-use competition between different categories of land, we use the MIT Economic Projection and Policy Analysis (EPPA) model. EPPA is a dynamic multi-sector, multi-region computable general equilibrium (CGE) model of the world economy (Paltsev et al., 2005; Chen et al., 2022). It is designed to develop projections of economic growth, energy transitions and anthropogenic emissions of greenhouse gas and air pollutants. Land use changes in EPPA are explicitly modeled by land use transitions among five major land categories (cropland, pasture, managed forest, natural grassland and natural forest) taking into account costs of transitions and associated emissions and maintaining consistent supplemental physical accounts of land (Gurgel et al., 2016; Gurgel et al., 2021).

Conversion of natural areas to agriculture follows a land supply response based on land conversions observed over the past few decades. Land can move to a less intensely managed use (e.g., from cropland to pasture or forest) or be abandoned completely and return to “natural” grass or forest land if investment in managed land is not maintained. Direct and indirect emissions associated with land use changes are represented in the model. Agricultural goods (e.g., crops, livestock, forestry) are produced by combining land, capital, labor, energy and intermediate inputs under multi-nested Constant Elasticity of Substitution (CES) functions. Alternative bioenergy technologies (e.g., liquid biofuels, bioelectricity, BECCS) are considered in EPPA, and require cropland areas to be deployed.

Future projections in the EPPA model are driven by economic decisions related to savings and investments, productivity improvements in labor, capital, land and energy, changing demand for goods and services due to income growth and international trade, and depletion of natural resources. These economic drivers, combined with imposed policies such as GHG emissions constraints, determine the economic trajectories over time and across scenarios and, consequently, land use changes. In terms of policies, EPPA uses a 1.5°C increase by the end-of-century with a 50% likelihood given an assessment of uncertainty in climate response to greenhouse gas forcing in the MIT Earth System Model (MESM). Global GHG emissions are driven by regional emissions caps based on the existing NDC commitments by 2030 and proportional historical contributions to GHG concentrations beyond that, assuming differentiated responsibilities across regions. In order to capture the potential contribution and land use competition from large adoption of NBS, in EPPA emissions from land use changes are included under the cap and afforestation/reforestation activities can generate carbon credits toward the climate stabilization goal.

A strength of the EPPA model is its economy-wide representation, which captures economic dynamics across multiple sectors of the economy, driving market-based demand for land. However, as a top-down model, EPPA is limited to aggregated representations of land types and land management/mitigation options. It lacks the level of detail considered in the Sky 2050 scenario, particularly details related to NBS. Therefore, to assess potential future land competition among food production, energy production, and carbon sequestration via NBS, we combine the detailed bottom-up estimates of land uses from the Sky 2050 scenario with the top-down modeling of land use change using the EPPA model.

We do this by overlaying the demanded areas for alternative land uses from Sky 2050 in EPPA and evaluating the consistency between them. More specifically, assuming the same climate stabilization target, we combine the NBS adoption rates and projections of land used for renewable energy sources from Sky 2050 with land use projections from EPPA which account for changes in income, food and energy consumption and prices due to the climate stabilization target and future development of societies. We then verify if the detailed bottom-up Sky 2050 land use requirements are compatible with the top-down EPPA market mediated projections on land demand to achieve the 1.5°C target while also satisfying future demand for food, energy and other land-intensive materials.

Global land use was quite stable among broad land use categories before the Industrial Revolution (Figure 2). In the middle of the nineteenth century, changes in land use from natural vegetation to pasture and cropland accelerated—1.08 Gha of natural forests and natural grassland that existed in 1800 became agricultural areas by 1900, and had risen to 3.41 Gha converted by 2000. The most recent 50 years have experienced declining rates of land use conversion worldwide. As shown in Figure 2, strong decarbonization efforts in the 21st century require a halt of historical deforestation trends and a movement toward forest regrowth and more efficient management of existing pasture and grassland areas. It also requires increasing amounts of land used for renewables (bioenergy, wind and solar). At the same time, even with productivity increases, the total amount of cropland must grow somewhat for food production to keep up with population and economic growth, while the significant potential for pasture productivity increases allows land for pasture to slightly decline while still meeting growing demand for products from livestock. Our results are consistent with the IPCC (2023) summary of 1.5°C scenarios that require substantial action on both land and energy.

A relevant aspect of the land use change assessment is the fact that different databases use alternative classifications of land use categories, which poses challenges in comparing specific land use types, as can be noticed by the three alternative datasets displayed in Figure 2. Major differences among the datasets are due to land use/land cover categorization and the number of land use types. Kicklighter et al. (2019) land use is based on land use transitions from Hurtt et al. (2011) from the year 1500 to the year 2100. IPCC land use areas rely on the data and approaches described in Lambin and Meyfroidt (2011), Luyssaert et al. (2014), and Erb et al. (2016).

While a consistent alignment between different land datasets remains challenging, in our modeling we combine the data from FAO (2019) and Kicklighter et al. (2019) to disaggregate total forest into natural and managed forest and to split grasslands into pastures and natural grasslands. We map land use categories from alternative databases to the categories used in the EPPA model (see Supplementary Table SA1 in the Supplemental Material) to allow a proper comparison. The total land area in Kicklighter et al. (2019) and EPPA is consistent and equals to 13.46 Gha, while IPCC reports the global ice-free area as 13 Gha. These numbers are close to the global area reported in Figure 1 (13.5 Gha), when glaciers are removed from the overall number. Overall, the general historical trends are similar in direction and relative size across most databases and in our representation.

Under the decarbonization ambitions of a 1.5°C scenario, there is significant potential for NBS to be deployed at large scale. Figure 3 presents the required annual adoption of NBS projections in the Sky 2050 scenario. The two largest adopted options are related to agricultural areas, such as NBS in cropland (including both biochar and a broad suite of regenerative agricultural practices to reduce emissions and increase soil carbon sequestration) and optimal grazing in pasture areas, which grow fast after 2023, achieve adoption peak rates above 50 Mha per year by 2040, and saturate at the beginning of the 2070s. Natural forest protection also increases, experiencing annual maximum deployment rates of 21 Mha by 2040s and declining thereafter, ceasing completely by 2080. Fire management deployment for grasslands and savannahs expands to 10 Mha by 2040 and keeps this rate of adoption until the end of the century. While this pathway is expressed as 10 Mha/yr, it does not necessitate the enrollment of new area each year; emission reductions can be derived from reducing the intensity/severity of fires in the same area each year. Other relevant NBS practices do not achieve such large rates of adoption, but may accumulate to sizeable amounts by the end of the century, which is the case of reforestation of natural forest areas, achieving 202 Mha, and grassland protection, which covers 246 Mha.

As NBS areas increase through the end of the century, they will compete with other land uses, such as food production, bioenergy and other renewable sources. Figure 4 shows how these competing uses can fit in EPPA’s projections of land requirements under 1.5°C stabilization. Adoption of NBS practices in crop production accelerates after 2030 and covers more than 60% of total cropland area by the end of the century. Bioenergy production requires 286 Mha by the end of the century, while land dedicated to solar and onshore wind generation achieves 347 Mha. Natural forest areas grow by 355 Mha from 2015 to 2100, which is larger than the 202 Mha projected reforestation areas in Sky 2050, since in the EPPA model we also represent market incentives, such as carbon offsets, in reforestation projects that leads to a larger increase in natural forest areas. Almost half of the natural forest area is explicitly convening some NBS by 2100, including forest protection and improved management, and only a very small amount is attributed to reforestation. Despite the expansion of NBS, forest protection and restoration, between 2020 and 2100 forestry products per capita are increased by 213%.

Cropland and forest expansion are possible due to a sharp contraction of 13% (−420 Mha) in pasture areas from 2015 to 2100, as well as some loss of natural grassland areas (−282 Mha). Such strong conversion of pasture areas is driven by increasing GHG prices on land use changes and methane-intensive livestock activities, but also due to current low productivity in grazing activities in several developing countries around the world, which makes the intensification of livestock practices relatively cheap when GHG emissions are constrained. The remaining pasture area by 2100 is still large enough to accommodate the NBS projections from Sky 2050, which covers 42% of the total pasture area. Sky 2050’s projection of NBS related to cropland covers 61% of total cropland by the end of the century, and improved grazing practices are adopted by 42% of total grasslands. We allocate several NBS opportunities not directly related to major agricultural and natural land use types represented in EPPA, such as protection and restoration of peatland, mangrove, seagrass and salt marshes, on EPPA’s “other” land use category, since these are quite small areas compared to other NBS options.

Focusing specifically on agriculture, NBS is projected to grow significantly. The 4.7 Gha of land dedicated to agriculture and bioenergy in 2015 is projected to reach 5.0 Gha by 2030, and then stabilize throughout the period to 2100 (Figure 5), according to EPPA projections under a 1.5°C stabilization scenario. The allocation of agricultural land among different uses will be very different by 2100 from today, and will include large areas of NBS. NBS in pasture areas will grow from 127 Mha in 2030 to 1.0 Gha by 2050 and 1.2 Gha by 2100, while NBS in cropland will evolve from 111 to 990 Mha by mid-century and 1.2 Gha by 2100.

Total land managed for NBS by the end of the century is projected to be about 3.5 Gha, of which 0.77 Gha is related to forest, 1.17 Gha to cropland, 1.17 Gha to pasture, 0.26 Gha to grassland and 0.15 Gha to other land types. Globally, there is enough land in each category to accommodate the NBS projections from Sky 2050, while also ensuring growing demand for food and other land-based products is met. In particular, our results show that despite economic and population growth (MIT Joint Program, 2023) and increasing demand for food, NBS does not interfere with the provision of food, and in fact, between 2020 and 2100 the consumption of total nutrients per capita (based on the overall changes in agriculture and food production) are increased by 161%. Still, it is important to note again that this scenario represents an ambitious expansion of NBS, which can present challenges at the regional scale (see Section 4.5). An increase in total nutrients means that supply of some specific nutrients (such as high-calories energy-yielding nutrients) will grow slower than other nutrients (such as vitamins, minerals, dietary fibers and proteins), but that the overall per capita needs are improving over time if a proper food distribution for healthier diets is in place, especially in developing regions of the world.

Table 2 shows the projected land areas used for bioenergy, onshore wind and solar power generation in 2050 for the selected regions in the Sky 2050 scenario. Table 3 provides the corresponding information for 2100. Global land area dedicated to bioenergy is projected to more than double by mid-century. It grows from about 100 Mha in 2020 (see Table 1) to 242 Mha in 2050 and 286 Mha by 2100. Dedicated biomass growing areas enable growing bioenergy consumption. The total commercial bioenergy use (i.e., including first-generation biofuels, second-generation biofuels and commercial solid biomass, but excluding traditional bioenergy used for heating and cooking) grows from about 20 EJ in 2020 to about 50 EJ in 2050, and about 70 EJ in 2100. We consider this as a conservative projection for bioenergy demand growth, because some alternative scenarios project a larger role for bioenergy with carbon capture (IPCC, 2018; Fajardy et al., 2021). In the Sky 2050 scenario, we take into account sustainability criteria preventing unintended consequences on global food security and environmental quality.

Land used for wind and solar generation grows much faster because of accelerated deployment of these sources of energy in a decarbonized world. In 2020, the global areas for wind and solar generation were 8 Mha and 0.6 Mha, respectively. By 2050, they are projected to grow to 180 Mha and 21 Mha. By the end of the century, the areas grow further to about 300 Mha for wind power and to 45 Mha for solar power. As shown in Tables 2, 3, total energy-dedicated land areas are reaching 7%–9% of the total land areas in the United States, China and India. Globally, land for energy is increasing from less than 1% of total land area in 2020, to about 3% in 2050, and to about 5% in 2100 (about 0.6 Gha). To achieve this scale of transition would require regulatory incentives from policy makers that not only bring the substantial investments needed to scale up low-carbon energy sources, but also address the social and environmental justice issues related to the associated land implications.

The amount of renewable energy in the Sky 2050 scenario is consistent with similar scenarios in other energy outlooks. For example, the International Energy Agency in its net-zero scenario (IEA, 2023) projects that in 2050 the share of wind and solar in global primary energy will be 41% and the corresponding share for bioenergy is 18%. The Sky 2050 scenario is slightly more optimistic about wind and solar generation with their 2050 combined share of 43%, and slightly less optimistic about bioenergy with its share at about 10% of total primary energy. The differences in bioenergy are mostly related to different assumptions about traditional biomass and solid commercial biomass.

Figure 6 presents the CO₂ sequestration contributions from NBS deployed in major land use types in the Sky 2050 scenario. The correspondence between land use types and NBS categories for CO₂ sequestration is provided in Supplementary Table SA2 of the Supplemental Material. Forest NBS opportunities provide the largest contribution and can reach more than 3.7 GtCO₂/year by mid-century, but decline thereafter due to several reforested areas approaching or reaching a maturation age at which point the net sequestration of carbon on living biomass is small. NBS in agricultural activities is the second largest CO₂ sink and may sequester around 2 GtCO₂/year for several decades. NBS in Grassland contributes to sequestering at most 1 GtCO₂/year by mid-century.

Among the NBS options in agriculture areas (Figure 7), biochar provides the largest sequestration from a single activity and is the only one where emission reductions are continuing to increase at the end of the century. Other options, such as optimal grazing practices, legumes in pastures, regenerative arable agriculture, and agroforestry may contribute with substantial CO₂ removals at mid-century, but as carbon in soils reach a saturation level and adopted areas stabilize around equilibrium carbon stocks approaching the end of the century, their contribution to sequester additional CO₂ reduces.

For biochar, there is uncertainty in the literature about application rates, the effects of repeated application and long terms stability (e.g., Jeffery et al., 2017; Tisserant and Cherubini, 2019; Lehmann et al., 2021; Woolf et al., 2021). In our study, we assume that biochar is applied to soils or sediments and no additional impact of the biochar on plant growth or “additional” soil carbon sequestration is factored in. The biochar is applied to land at appropriate rates and in appropriate contexts to ensure negligible effects on existing carbon cycling in that site/ecosystem. It should be noted that we aim to be conservative for the overall NBS estimates and we only include pathways that have more robust scientific evidence. Some researchers are working on other pathways (such as improved aquaculture, enhanced rock weathering, permafrost protection, advanced rice management, advanced manure management and change in cattle diets, etc.), and some of these may well come to fruition. This would potentially increase our overall NBS estimates.

Land use at the global level can accommodate the multiple demands expected under a 1.5°C climate stabilization scenario. At the regional level, however, challenges to integrate all land uses may arise. We explore such potential challenges by considering land use allocation in a set of selected major countries and regions in the world (Figure 8).

In the United States, NBS related to livestock activities occupies 88% of total available pasture land by 2100. Almost half of the other land use category is needed to accommodate other types of NBS and solar energy, while we assume onshore-wind is placed on cropland areas. In the case of Europe, the placement of wind, solar and some NBS on Other land is even more challenging, requiring 74% of it, which means a possible competition with urban and infrastructure areas. NBS in cropland and pastures also may raise concerns, since only 16% and 12% of these land categories, respectively, are not adopting some kind of NBS by 2100.

In emerging economies, we observe mixed outcomes. In China, the amount of required NBS in crops to achieve 1.5°C needs to cover the total cropland available plus most of the area dedicated to bioenergy crops. India faces several land use competition issues: wind, solar and some NBS will require 80% of the Other land category by 2100, around 70% of the already limited pasture areas need to adopt some NBS, and 73% of the 150 Mha of cropland will be required to produce crops using NBS approaches.

Among the major food and bioenergy producers in the world, only Brazil seems to face moderate pressures on land use constraints due to NBS adoption in agriculture and major natural land use categories. However, the country has a very small area of the Other land use category, and the deployment of solar, wind and other NBS types will require the equivalent of 65% of this area, which poses questions about feasibility and competition with urban and infrastructure uses, which may be alleviated if wind, solar and other NBS are placed elsewhere.

Figure 9 provides a snapshot for the shares of land in a particular use in 2050 for the selected major regions and the globe. The figure shows the shares for traditional practices, land for energy, NBS, and proactive land management, such as restoration and protection. Our projections for 2050 illustrate a substantial deployment of advanced land practices, especially for NBS in cropland in Europe and India, NBS in pasture in United States, Europe, China, Brazil, and Africa, and NBS in natural forests in Brazil. At the same time, some regions still rely heavily on traditional land uses, such as pasture in China and Africa, and natural grassland and natural forests in most of the regions.