Afforestation is considered as a major carbon dioxide removal (CDR) technology, but if implemented inappropriately in large scale, can negatively affect food and land systems. Here we quantitatively showed how a forest-type selection in afforestation would enlarge the global carbon sequestration and affect global food and land systems and sustainability. We found that i) afforestation, if its forest type is carefully selected, would increase the carbon sequestration by 25% at maximum compared to the indigenous type of forest while reducing food insecurity. At the same time, ii) compared to bioenergy with carbon capture and storage, afforestation would affect similarly or even more negatively the economy, energy, food, and land systems due to less land efficiency of carbon removal, leading higher price of energy and food, more land expansion for carbon removal and higher food insecurity. This suggests a necessity of a careful selection of forest types in afforestation and at the same time an importance of determining a best mix of land-based CDR technologies to achieve the long-term stringent climate goal without compromising the sustainability.
Article
Forest Types for Afforestation: Benefits for Carbon Sequestration and Food Systems under Stringent Climate Mitigation
https://doi.org/10.21203/rs.3.rs-3095494/v1
This work is licensed under a CC BY 4.0 License
Journal Publication
published 01 Apr, 2024
Read the published version in Communications Earth & Environment →
Version 1
posted
You are reading this latest preprint version
The ambitious climate change mitigation scenario of limiting global warming to 2°C will require extremely rapid reduction in greenhouse gas emissions and large negative emissions in the end of this century1. In most scenarios in IPCC AR6, large-scale CDR (Carbon Dioxide Removable) options or nature-based solutions such as CDR associated with afforestation or reforestation and bioenergy with carbon capture and storage (BECCS) play a vital role to remove CO2 from the atmosphere particularly in the latter half of this century to achieve such a stringent climate goal2, 3, 4. In the scenario that limit warming to 2°C, annual volumes of carbon mitigation is estimated to be 2.75 GtCO2/year for BECCS and 2.98 GtCO2 GtCO2/year for afforestation/reforestation at median level in 20505. The cost of mitigation is estimated to be 50–200 USD/tCO2 for BECCS while that is 0-200 USD/tCO2 for afforestation/reforestation for 20305. A key issue is the feasibility of implementing land-based CDR that may negatively affect food security6, 7, 8, 9, 10 and the biodiversity11, 12, 13, 14. The potential consequences of large-scale CDR in mitigation scenarios with stringent climate goals could be considered infeasible or socially undesirable due to sustainability and intergenerational equity concerns15, 16, 17, 18, 19, 20. Feasibility of the land-based CDR would depend on the stringency of the climate goals, associated emissions pathways and the impacts on other sustainability. Moreover, inappropriate forest expansion, such as afforestation of naturally open habitats, could reduce habitats for non-forest organisms and negatively affect biodiversity11, 12, 13. However, no study has investigated how afforestation should be designed to reduce the negative impacts on other sustainability goals and which forest types of afforestation should be applied.
Here we quantified how a forest-type selection in afforestation and food related measures would affect the carbon sequestration potential and food and land related sustainability. For quantification we used a framework (Fig. S1) that combines an economic model (AIM/Hub)21, a land-use allocation model (AIM/PLUM)22, and a terrestrial vegetation model (Vegetation Integrative Simulator for Trace gases model ; VISIT)23. In the framework, we developed a mechanism to consider a forest-type selection in afforestation and calculate the carbon removal potential of afforestation by assuming different forest types. Regional land demand (17 regions) estimated using AIM/Hub was input into AIM/PLUM to calculate the proportional area of different land categories and amounts of carbon sequestered by afforestation and BECCS at half degree grids. Land of afforestation and other land categories was allocated to maximize profit for landowners based on the biophysically determined land productivity (yield per unit area)22. Ten sets of scenarios were analyzed, differentiated by the two main land-based CDR technologies available for deployment (BECCS, afforestation), three forest type selection schemes, and implementation/non-implementation of food related measures (crop yield improvement, dietary change, trade globalization, improvement of food distribution inequality) along with a conventional scenario in which no climate mitigation measures (Aff, Bio) or food measures are deployed (Table 1). The scenarios with different climate mitigation strategies allow us to explore the benefits and trade-offs/adverse effects of afforestation compared to those of BECCS. The scenarios with differentiation of forest type selection schemes (Aff-Cur, Aff-Div, Aff-Cmax) allows us to explore the impact of forest type selection on carbon sequestration and other sustainability. The comparison of the scenarios with and without the implementation of food related measures (with and without FodPol) allows us to explore the effect of food related measures on carbon sequestration through afforestation and other sustainability. For the deployment of climate mitigation measures in AIM/Hub, either only afforestation or only BECCS was deployed to separate the pure effect of a single technology. Forest type selection was accounted for by assuming growth rate of different forest types. The food relevant measures assumed agricultural intensification, trade globalization, and improved equity in food distribution as policy relevant efforts as well as dietary change as voluntary action. See the methods section for greater detail of assumptions on forest type selection and food related measures. The carbon budget for this century was set to 600 GtCO2, corresponding to the long-term climate target of 2.0°C; no negative emissions were allowed after global achievement of net-zero CO2 emissions (See Figure S5 for emission pathways). Although the absent of negative emissions limits the deployment of large-scale land-based carbon removal technologies, this restriction is adopted because the model could not be solved for scenarios for achieving the 2°C target in which negative emissions are allowed but only afforestation is deployed without BECCS without changing the other socioeconomic conditions such as lowering energy demand24, 25. See the supplementary material for greater detailed description of models.
Increased Carbon sequestration by forest-type selection
Our analysis shows that afforestation, if its forest type is carefully selected from the ones in the same ecological zone (Aff-Div), would increase the carbon sequestration by 2% globally (7.7GtCO2/year) in 2100 compared to the case with the indigenous forest type (7.6GtCO2/year in the Aff-Cur) (Fig. 1-a). Moreover, a fast-growing forest-type (Aff-Cmax) would increase the global carbon sequestration by 25% (9.5GtCO2/year in 2100) compared to the case of indigenous forest. Regional analysis shows that the carbon sequestration substantially increases in all the region excluding OECD and EU (Fig. 1-b, 1-c). For example, the carbon sequestration in LAM increases by 5.2% and 37% in the Aff-Div and Aff-Cmax in 2100 while the amount in REF increases by 4.3% and 18% respectively in the same scenarios and year compared to the case of indigenous forest. As regionally selected forest types, for example in LAM, tropical and subtropical evergreen forest (#1) is currently widely distributed while tropical montane forest (#02) and tropical subtropical dry forest (#03) increases carbon sequestration in the region (Figure S2). In the same way, in REF e.g. South Russia, northern taiga (#11, #12) is currently distributed while the southern taiga (#08) is beneficial for carbon sequestration in the region. In East Asia e.g. South China, currently mid-latitude mixed forest (#04) is distributed while the semiarid wood or low forest (#06) is beneficial in the region.
Land intensity of Carbon sequestration and its impacts on land use
The land intensity of carbon sequestration potential (LIC), which describes potential amount of carbon sequestration per unit area, can be an important factor that determines the difference in the carbon sequestration and area implemented across forest types of afforestation and mitigation options. The LIC of afforestation was calculated by dividing the amount of carbon sequestration potential of afforestation by area of land allocated to afforestation. The land intensity of BECCS was calculated by dividing a multiplication of bioenergy crop yield potential, a regional share of electrification of BECCS, a regional share of CCS usage for bioenergy and carbon capture ratio by area of land allocated to bioenergy crop production. Our results shows that if forest type with highest growth rate in the potential types of forest distributed in the same agroecological zones as the indigenous type or all forest types is selected, the LIC of afforestation becomes higher than the that of the indigenous type (Fig. 2-a). In the Aff-Div and Aff-Cmax, the global mean LIC of afforestation is higher (4.1 tonneCO2 equivalent /ha/year and 5.0tonneCO2 equivalent /ha/year respectively in 2100) than that of the indigenous type of forest (4.0 tonneCO2 equivalent /ha/year for the same years) (Fig. 2-a). This difference is because in some regions or grids, the other types of forest with higher growth rate and more efficient carbon removal than the indigenous type are available in the same agro-ecological zones. See Figure S3 for LIC of different forest type and Figure S4 for the regional geographical distribution of carbon sequestration of different forest type selection.
Given the same long-term climate target, on the other hand, afforestation less contributes to carbon sequestration compared to BECCS. For example, global carbon sequestration of afforestation (8.5 GtCO2/year and 9.5 GtCO2/year in 2050 and 2100 respectively) even in the scenario for maximizing Carbon seq of afforestation (Aff-Cmax) scenario are smaller compared to that of BECCS (7.3 GtCO2/year and 14 GtCO2/year in 2050 and 2100 respectively) in the BECCS only (Bio) scenario in the same years (Fig. 1-a). This difference is because bioenergy crop yield is greater than afforestation yield. Mean annual carbon sequestration of BECCS is more than 10 times that of afforestation (Fig. 2-c, 2-e). As results, the LIC of afforestation is lower than that of BECCS (Fig. 2-a). In the Aff-Cmax, the global mean LIC of afforestation is 7.7 tonneCO2 equivalent /ha/year and 5.0 tonneCO2 equivalent /ha/year in 2050 and 2100 while that of BECCS of 15 tonneCO2 equivalent /ha/year and 19 tonneCO2 equivalent /ha/year for the same years.
Afforestation and BECCS differently affect land-use change. Whereas cropland and pastureland decrease instead of forest area expansion in afforestation-only scenarios, forestland decreases to expand cropland area for bioenergy production in BECCS-only scenarios (Bio) (Fig. 3). What differs here is the change in cropland. In BECCS-only scenarios (Bio), the high sequestration intensity of BECCS means less competition with crop production for land, resulting in a global increase in cropland in response to increased demand of food. In contrast, in afforestation-only scenarios, afforestation increases competition for land with crop production, resulting in decreased cropland.
Regional analysis
From regional perspectives, whereas Carbon sequestration of afforestation is high in OECD countries and reforming economies, that of BECCS is high in Latin America, Asia, and OECD countries (Fig. 1-b, 1-c). These reginal differences in Carbon sequestration potential are the results of the geographical heterogeneity in land availability (e.g., demand for cropland and pastureland), the LIC as well as the potential to deploy renewable energy and to reduce emissions from energy systems such as energy consumption and costs of renewable energy. For example, in high latitude regions such as OECD and EU, and Reforming economics, the LIC of afforestation is high generally because cold-tolerant types of forest (e.g. coniferous forests or taiga) grow faster than tropical and template forests (Figure S2-d for forest growth curves across forest types, Fig. 2-c and Figure S3-a,b,c for geographical distribution of growth ratio of afforestation), and the land available for afforestation is widely distributed in potential area for high-growth forest (Figure S4-a,b). Thus, in high latitude, the current forest types have a high carbon removal potential of afforestation (Figure S3-a), and forest type selection increases the potential (Figure S3-b,c). In contrast, this is not the case in low-latitude regions. In regions with high LIC such as the Amazon in Latin America and Central Africa are already forested and cannot be used for afforestation (Figure S4-a,b). In these regions, instead, the LIC of BECCS is high (Figure S3-d) and thus a smaller area is needed for BECCS than afforestation to achieve a give climate target (Fig. 2-b, 2-d). It is more effective to implement BECCS than afforestation in these regions.
Impacts of food relevant measures
Implementation of food relevant measures considerably increases the Carbon sequestration for all the scenarios of both afforestation and BECCS (Fig. 1-a). For example, the Aff-CmaxFodPol scenario in which forest type selection to maximize Carbon sequestration is combined with implementation of food measures yields the maximum sequestration of afforestation (11.3 GtCO2/year in 2100; 19% more from the level without food measures; 49% more from the level of the indigenous forest without food measures). Among the food relevant measures analyzed, dietary change substantially reduces meat consumption and increases food crop consumption. This, in turn, considerably reduces pastureland area and increases the area of land available for sequestration measures and cropland (Fig. 3-a). Consequently, the area of cropland and pastureland decreases while the afforested area increases in all regions of the world, resulting in increased carbon sequestration (Fig. 2-b, Fig. 3-b).
Impacts on the environment and sustainability.
It is meaningful to explore the impacts of land-based options on the economy (GDP), energy and food systems, because many countries have established long-term climate mitigation goals considering a wide range impacts of climate mitigation measures. In terms of economic efficiency, the afforestation-only set of scenarios (Aff-only) raised carbon prices and caused more GDP loss. This is because economic and industrial structures and energy systems are less economically efficient under these scenarios compared to the BECCS-only scenario with lower land competition and higher economic efficiency. Food measures implemented along with BECCS (2C-Bio-FodPol) lowered carbon price and reduced loss of GDP. In terms of energy systems, the price of energy increases in afforestation-only scenarios where emissions reduction is achieved through deployment of relatively expensive renewable energy (solar, wind) rather than bioenergy; renewal energy accounts for a greater share of energy supply, raising the cost of energy. The higher cost of energy also contributes to losses in GDP (Fig. 3).
In terms of food systems, afforestation and BECCS push up the price of land and food because both of them require land. The risk resulting from afforestation is greater than that resulting from BECCS. The impact of afforestation is greater since afforestation requires more land than BECCS to achieve the carbon sequestration for a given long-term climate target. The impact changes over time depending on the emission pathways (Figure S5) and increases in the latter half of the century. Food relevant measures reduce almost all the food-related negative impacts associated with afforestation and BECCS, except risk of hunger, specifically because the lower food demand due to dietary change and the promotion of production in suitable locations through trade liberalization and yield improvements decreases the land demand for agricultural production and decreases food prices.
We investigated how a forest-type selection in afforestation would enlarge the global carbon sequestration and affect global food and land systems and sustainability. Our results indicate that afforestation, if its forest type is carefully selected, would increase the carbon sequestration by 25% at maximum from the case with the indigenous forest type and that if the implementation combined with food measures the carbon sequestration potential of afforestation increases by 49% at maximum from the case with the indigenous forest type with lower land demand, less populations both at risk of hunger and over-consuming, and less nitrogen and water usage. These results suggested a careful selection of forest type for afforestation combined with appropriate food measure can bring benefits that include increased amount of carbon sequestration, reducing impacts on food and land systems and other sustainability. Although no integrated assessment model accounts for forest types, the consideration about multiple forest types would be beneficial for future evaluations of land-based climate mitigation.
Combination of land-based CDR required in climate mitigation.
At the same time, our results also indicate that BECCS is more effective than afforestation in terms of carbon sequestration and that afforestation has the same or greater negative impact on the economy, energy, food, and land use than BECCS2, 3, 4. This is because lower LIC of afforestation requires more land for a given carbon removal toward the long-term climate target, resulting in higher price of carbon and energy, greater land demand and negative impacts on food security than BECCS although afforestation could bring some benefits from less nitrogen and water usage and the less over-consuming population. Thus, the both land-based CDR technologies will likely negatively impact food and land systems at certain extent. Moreover, the social acceptability and desirability of using BECCS is uncertain. This suggests that all the benefits and risks of implementing land-based CDR technologies should be considered when discussing climate actions and future pathways to be taken. For instance, early climate action that avoids temperature overshoot would reduce reliance on land-based CDR technologies and maximizing emissions reduction while maintaining food and land systems, reducing the risks of implementing land-based CDR technologies26. Given the necessity of CDR technologies to achieve the long-term climate target, it would be important to determine a combination of land-based options with food measures to realize a large-scale CDR while keeping the sustainability. The ambitious target of limiting climate change to 2°C can be achieved through a best mix of land-based mitigation options with a better selection of forest types in proportions that minimize the trade-offs/adverse impacts on economy, energy, food, land systems, and biodiversity.
The importance of improving food distribution inequality when diet changes
Our analysis indicates that raising food distribution equality could is key to mitigating negative impacts of afforestation on food insecurity. In the case where the level of healthy and sustainable food diets is achieved only at the country-mean level without improvement of food distribution equality, hunger likely increases substantially while over-consumption decreases. Although the total population at risk of hunger and the over-consuming population will decrease globally because the reduction in over-consumption is greater than the increase in hunger, regional disparities will occur. Hunger will intensify in developing countries while over-consumption will decrease in advanced countries. Accordingly, improvements in food distribution equality must be accompanied by a shift to a healthy and sustainable diet such as that proposed by EAT-Lancet27. In the additional scenario assuming the food distribution inequality based on past trends as in Hasegawa et al. (2015)28, the population at risk of hunger was greater than for the scenario where there is no dietary change. This suggests that the past trends of food distribution equality improvement is insufficient and that greater reduction of food distribution inequality is needed. To avoid the increased hunger risk accompanying achievement of the EAT-Lancet target, the degree of inequality reduction (reduction in CV for food distribution) must be approximately double that previously assumed28.
Some limitations of this study point at further research avenues. First, there are available land-based CDR technologies that were not considered in this analysis. Currently, IAMs have only been used to model the deployment of BECCS and A/R. Other land-based CDR technologies (e.g. agroforestry, biochar, soil carbon management) have not been considered in IAMs primarily because they are connected to sectors that are not yet included in many models, and because parameterizing these technologies is speculative given that CDR technologies are not currently commercially deployed. It is therefore unlikely that these technologies can be implemented in the current models or considered in further analysis at this time. Second, further analysis using a biodiversity model and its model development on the impacts of land-based CDR or nature-based solutions on the biodiversity help generate more detailed aspects of the biodiversity. Third, cost of food related measures the knowledge of which is still limited and is not considered in this analysis would cause some adverse effects and barriers to realization. Finally, the present study is based on a single set of models; Multiple model analyses would help generate more robust conclusions and clarify sources of uncertainty, particularly with regard to the magnitude of the effects of forest type selection on the carbon removal.
Model framework
We modeled the impact of afforestation with optimal forest types on greenhouse gas emissions, food and land-use systems, and the environment as well as the mitigation potential of afforestation using an integrated framework (Fig. S1) that combines an economic model (AIM/Hub)21 and land-use distribution model (AIM/PLUM)22. AIM/Hub calculates future land demand for 17 regions based on assumed socio-economic conditions including population, GDP, and food measures (yield improvement, dietary change, trade globalization, reduction of food distribution inequality). The output from this model is input into AIM/PLUM to calculate the proportional area of different land uses and amounts of carbon sequestered by afforestation and BECCS in individual grid cells (0.5° × 0.5°) so that total land demand area is satisfied for each region. AIM/PLUM allocates cropland and afforested land to maximize profit to the landowner based on biophysically determined land productivity (yield per unit area)22. The forest type selection scheme for each scenario was accounted for in AIM/PLUM by assigning afforestation yields based on different type selections. Yields from hydrological model H0829 were used for bioenergy cropland productivity (yield) used to calculate revenue; NPPs for afforestation calculated using the terrestrial vegetation model VISIT23 were used for afforestation productivity. Carbon sequestration by forests of different types and ages was calculated based on land-use change over time.
Assumptions on forest types and food measures.
Forest type selection was accounted for by assigning afforestation yields based on different type selections. Annual net primary production (NPP) of the vegetation and soil was calculated at the grid cells for the existing forest type and twelve afforestation types using the VISIT. The Aff-Cur scenario for existing forest types assumed the NPP of currently growing forest types. The Aff-Cmax scenario for forest types that maximizes carbon sequestration assumed the forest type with the highest NPP among the 12 forest types for each grid cell. The Aff-Div scenario for forest types that maximizes carbon sequestration while also considering ecosystem protection assumed the forest types with the highest NPP among the forest types that located in the same climatic and ecosystem zone as the grid cell. Carbon sequestration by different forest types and ages was calculated using tree growth curves30, the parameters of which were estimated from NPP estimated by the VISIT, as well as land-use change over time. The forest types selected for each scenario are shown in Figure S2.
Assumptions on food relevant measures.
The food relevant measures implemented assumed agricultural intensification, trade globalization, dietary change, and improved equity in food distribution. Agricultural intensification and trade globalization were assumed as the same as those for scenario SSP128. To represent dietary change, calorie intake was set so that the EAT-Lancet target will be achieved in 2050 and keep the level through 2100 in all regions. To represent more equitable food distribution, the coefficient of variation (CV) for per-capita dietary energy consumption in each country was set to achieve 0.1 when per-capita GDP is USD 50,000. See Hasegawa et al.28 for details regarding the method for estimating the population at risk of hunger and assumptions regarding CV.
Method for estimating carbon sequestration through afforestation for different forest types
To account for different forest types, annual NPP of vegetation and soil was calculated at the grid cell level for the existing forest types and 12 different forest types using the terrestrial vegetation model VISIT23. Carbon sequestration was calculated for forests of different types, ages, and land-use change trajectories using tree growth curves30. To calculate the growth rates of different forest types in the same environment, the following virtual experiment was performed using VISIT: all vegetation was removed from land in 2010 and afforested simultaneously with a single forest type. The same procedure was performed for all 12 forest types. The change over time in per-unit-area NPP of vegetation and soil in all land grid cells was calculated from 2010 onward. Cumulative NPP of the vegetation and soil was considered the amount of carbon sequestered by trees, and tree growth curves defined by the equation below were fit to the year-over-year change in cumulative NPP for different forest types. Following the method of Sohngen et al. (2008)30, parameters B and δ were set to 30 and 1, respectively; parameter A was estimated using the value for Va calculated by VISIT when tree age is 20. The maximum carbon amount was set to 300 MgC/ha. Tree age was calculated as the number of years from afforestation based on land-use change over time.
Va(MgC/ha)=δ*exp[A-B/age]
Va(MgC/ha): carbon stored in trees, age: tree age, δ, A, B: coefficients
- Riahi K, R. Schaeffer, J. Arango, K. Calvin, C. Guivarch, T. Hasegawa, K. Jiang, E. Kriegler, R. Matthews, G.P. Peters, A. Rao, S. Robertson, A.M. Sebbit, J. Steinberger, M. Tavoni, D.P. van Vuuren. Mitigation pathways compatible with long-term goals. Cambridge, UK and New York, NY, USA: IPCC; 2022.
- Bauer N, Rose SK, Fujimori S, van Vuuren DP, Weyant J, Wise M, et al. Global energy sector emission reductions and bioenergy use: overview of the bioenergy demand phase of the EMF-33 model comparison. Climatic Change 2018.
- Roe S, Streck C, Obersteiner M, Frank S, Griscom B, Drouet L, et al. Contribution of the land sector to a 1.5 °C world. Nature Climate Change 2019, 9(11): 817-828.
- Hanssen SV, Daioglou V, Steinmann ZJN, Doelman JC, Van Vuuren DP, Huijbregts MAJ. The climate change mitigation potential of bioenergy with carbon capture and storage. Nature Climate Change 2020.
- Babiker M, G. Berndes, K. Blok, B. Cohen, A. Cowie, O. Geden, V. Ginzburg, A. Leip, P. Smith, M. Sugiyama, F. Yamba. Cross-sectoral perspectives. Cambridge, UK and New York, NY, USA.: IPCC; 2022.
- Hasegawa T, Sands RD, Brunelle T, Cui Y, Frank S, Fujimori S, et al. Food security under high bioenergy demand toward long-term climate goals. Climatic Change 2020.
- Fujimori S, Wu W, Doelman J, Frank S, Hristov J, Kyle P, et al. Land-based climate change mitigation measures can affect agricultural markets and food security. Nature Food 2022, 3(2): 110-121.
- Doelman JC, Stehfest E, van Vuuren DP, Tabeau A, Hof AF, Braakhekke MC, et al. Afforestation for climate change mitigation: Potentials, risks and trade-offs. Global Change Biology 2020, 26(3): 1576-1591.
- Kreidenweis U, Humpenöder F, Stevanović M, Bodirsky B, Kriegler E, Lotze-Campen H, et al. Afforestation to mitigate climate change: Impacts on food prices under consideration of albedo effects. Environmental Research Letters 2016, 11: 085001.
- Frank S, Havlík P, Soussana J-F, Levesque A, Valin H, Wollenberg E, et al. Reducing greenhouse gas emissions in agriculture without compromising food security? Environmental Research Letters 2017, 12(10): 105004.
- Seddon N, Smith A, Smith P, Key I, Chausson A, Girardin C, et al. Getting the message right on nature-based solutions to climate change. Global Change Biology 2021, 27(8): 1518-1546.
- Portner H-O, Scholes RJ, Agard J, Archer E, Arneth A, Bai X, et al. Scientific outcome of the IPBES-IPCC co-sponsored workshop on biodiversity and climate change (Version 5). Zenodo.; 2021.
- Veldman JW, Overbeck GE, Negreiros D, Mahy G, Le Stradic S, Fernandes GW, et al. Where Tree Planting and Forest Expansion are Bad for Biodiversity and Ecosystem Services. BioScience 2015, 65(10): 1011-1018.
- Krause A, Pugh TAM, Bayer AD, Doelman JC, Humpenöder F, Anthoni P, et al. Global consequences of afforestation and bioenergy cultivation on ecosystem service indicators. Biogeosciences 2017, 14(21): 4829-4850.
- Shue H. Climate dreaming: negative emissions, risk transfer, and irreversibility. Journal of Human Rights and The Environment 2017, 8: 203-216.
- Field CB, Mach KJ. Rightsizing carbon dioxide removal. Science 2017, 356(6339): 706.
- Boysen LR, Lucht W, Gerten D, Heck V, Lenton TM, Schellnhuber HJ. The limits to global-warming mitigation by terrestrial carbon removal. Earth's Future 2017, 5(5): 463-474.
- Morrow D, Svoboda T. GEOENGINEERING AND NON-IDEAL THEORY. Public Affairs Quarterly 2016, 30(1): 83-102.
- Fujimori S, Rogelj J, Krey V, Riahi K. A new generation of emissions scenarios should cover blind spots in the carbon budget space. Nature Climate Change 2019, 9(11): 798-800.
- Rogelj J, Huppmann D, Krey V, Riahi K, Clarke L, Gidden M, et al. A new scenario logic for the Paris Agreement long-term temperature goal. Nature 2019, 573(7774): 357-363.
- Fujimori S, Masui T, Matsuoka Y. AIM/CGE [basic] manual. Tsukuba, Japan: Center for Social and Environmental Systems Research, NIES; 2012.
- Hasegawa T, Fujimori S, Ito A, Takahashi K, Masui T. Global land-use allocation model linked to an integrated assessment model. Science of The Total Environment 2017, 580: 787-796.
- Ito A, Inatomi M. Water-Use Efficiency of the Terrestrial Biosphere: A Model Analysis Focusing on Interactions between the Global Carbon and Water Cycles. Journal of Hydrometeorology 2012, 13: 681-694.
- Grubler A, Wilson C, Bento N, Boza-Kiss B, Krey V, McCollum DL, et al. A low energy demand scenario for meeting the 1.5 °C target and sustainable development goals without negative emission technologies. Nature Energy 2018, 3(6): 515-527.
- van Vuuren DP, Stehfest E, Gernaat DEHJ, van den Berg M, Bijl DL, de Boer HS, et al. Alternative pathways to the 1.5 °C target reduce the need for negative emission technologies. Nature Climate Change 2018, 8(5): 391-397.
- Hasegawa T, Fujimori S, Frank S, Humpenöder F, Bertram C, Després J, et al. Land-based implications of early climate actions without global net-negative emissions. Nature Sustainability 2021.
- Willett W, Rockström J, Loken B, Springmann M, Lang T, Vermeulen S, et al. Food in the Anthropocene: the EAT–<em>Lancet</em> Commission on healthy diets from sustainable food systems. The Lancet 2019, 393(10170): 447-492.
- Hasegawa T, Fujimori S, Takahashi K, Masui T. Scenarios for the risk of hunger in the twenty-first century using Shared Socioeconomic Pathways. Environmental Research Letters 2015, 10(1): 014010.
- Yamagata Y, Hanasaki N, Ito A, Kinoshita T, Murakami D, Zhou Q. Estimating water–food–ecosystem trade-offs for the global negative emission scenario (IPCC-RCP2.6). Sustainability Science 2018, 13(2): 301-313.
- Sohngen B, Golub A, Hertel TW. The Role of Forestry in Carbon Sequestration in General Equilibrium Models; 2008 2008.
Table 1 Scenario settings. Food measures include agricultural intensification, dietary changes, trade globalization, improved food distribution.
There is NO Competing Interest.
- DraftV0230622cleanSI.docx
Supplementary Information