The results presented in this paper were implemented using the IMAGE framework. Below, we briefly describe the IMAGE model and the way the planetary boundary control variables are calculated.
Description of the IMAGE model framework
IMAGE is an integrated assessment modelling framework that simulates the global and regional environmental consequences of changes in human activities. Detailed model documentation is available online (www.pbl.nl/IMAGE). The model has been designed to analyse large-scale and long-term interactions between human development and the natural environment and identify response strategies to global environmental change based on assessing options for mitigation and adaptation. The IMAGE framework is structured around the causal chain of key global sustainability issues and comprises two main systems: (1) the human or socioeconomic system that describes the long-term development of human activities relevant for sustainable development; and (2) the Earth system that describes changes in natural systems, such as the carbon and hydrological cycle and climate. The two systems are linked through emissions, land use, climate feedbacks and potential human policy responses. Most of the socioeconomic parameters are simulated for 26 regions, and most environmental parameters are calculated at a geographical grid of 30 × 30 or 5 × 5 min.
Important inputs to the model are descriptions of the future development of the direct and indirect drivers of global environmental change, with exogenous assumptions on population, economic development, lifestyle, policies and technology change forming a key input into the energy system model (Targets Image Energy Regional model, TIMER) and the food and agriculture system model (Modular Agricultural General Equilibrium Tool, MAGNET). TIMER is a system-dynamics energy system simulation model describing key trends in energy use and supply, with changes in model variables calculated based on information from the previous time step. MAGNET is a computable general equilibrium model with high detail for the agricultural sector. It uses information from IMAGE on land availability, suitability and changes in crop yields due to climate change and agricultural expansion into heterogeneous land areas. Together with the drivers described above, the regional consumption of, production of and trade in agricultural commodities are computed. The results from MAGNET on production and endogenous yield changes are used in IMAGE to calculate spatially explicit land-use change and the environmental impacts on the carbon, nutrient and water cycles, biodiversity and climate. Here a rule-based system allocates crop production to the grid based on yields and distance from other agriculture areas. The IMAGE global nutrient model (IMAGE-GNM) is a process-based simulation model that calculates the fate of N (ref. 51) and P from land to sea52,53. It is coupled with the hydrological PCRaster Global Water Balance model, which provides all water fluxes. The IMAGE-GNM describes the flow and retention/removal of the N and P delivery from agricultural and natural soils to surface waters, and the in-stream loss processes in all surface waters (lakes, reservoirs and rivers).
In IMAGE, the main interaction with the Earth system is through changes in energy, food and biofuel production that induce land-use changes and emissions of CO2 and other GHGs. A key component of the Earth system is the Lund–Potsdam–Jena managed land (LPJmL) model, which is hard-coupled to IMAGE. The LPJmL model covers the terrestrial C cycle and vegetation dynamics54. On this basis of the regional production levels and the output of LPJmL, a set of allocation rules in IMAGE determine the actual land cover. Climatic change is calculated as the global mean temperature change using a slightly adapted version of the Model for the Assessment of GHG-Induced Climate Change version 6.0 (MAGICC6) climate model55. The changes in temperature and precipitation in each grid cell are derived from the global mean temperature using a pattern-scaling approach. The model accounts for several feedback mechanisms between climate change and dynamics in the energy, land and vegetation systems.
IMAGE is often used in conjunction with the Global Biodiversity model for policy support (GLOBIO), designed to evaluate the impacts of five human pressures on terrestrial biodiversity intactness, including climate change, land use, atmospheric N deposition, infrastructure and hunting56. Biodiversity intactness is quantified based on the MSA indicator, which represents the mean abundance of original species in an impacted situation compared to their abundance in an undisturbed reference situation, hence being indicative of biosphere integrity. For the present study, we focused on the integrity of terrestrial plant communities. We implemented the relationships between the terrestrial plant MSA and the three human pressures affecting it (climate change, land use and atmospheric N deposition) in the IMAGE modelling framework, allowing us to quantify MSA as a function of these three pressures directly in IMAGE (see also Supplementary Information Section 4 for the exact relationships used).
Description of calculation of planetary boundary control variables in IMAGE
Below, we describe how we used IMAGE to quantify the planetary boundary control variables. This includes model parameters and reference values. The data for the main IMAGE output for all scenarios, as well as the data calculated for the planetary boundaries, can be found via Zenodo at https://doi.org/10.5281/zenodo.10203631.
Climate change
The IMAGE model dynamically calculates GHG emissions from the energy, industry and land-use sectors based on a detailed process-level representation of these systems. The model also calculates full CO2 fluxes on land based on the LPJml and the Bern ocean C model included in IMAGE54. Greenhouse gas emissions are input to the simple climate model MAGICC6, which calculates the radiative forcing and global temperature change55. The 2015 planetary boundary framework proposes a boundary of 350 ppm CO2, with 350–450 ppm CO2 as the zone of increasing risk. These values broadly correspond with scenarios that aim for 1.5 and 2.6 W m−2. Therefore, these values have been used here (instead of the 1.0 and 1.5 W m−2 values provided in the 2015 planetary boundary paper1). The forcing value for 2015 from IMAGE is slightly different than that used in the planetary boundary 2015 paper, partly as a result of the different definition. The difference is, however, within the uncertainty range. The method is also described in refs. 55,57.
Stratospheric O3 depletion
The exogenous scenario output from Coupled Model Intercomparison Project Phase 6 (CMIP6) is based on similar scenarios. The 2015 planetary boundary framework proposes as a boundary a less than 5% reduction from the pre-industrial level of 290 Dobson units (DU) (5–10%), assessed by latitude. Because stratospheric O3 depletion is mostly a problem over Antarctica, we used the 90–60° S band, as assessed based on CMIP6 data33.
Atmospheric aerosol loading
The IMAGE model calculates emissions scenarios for air pollutants based on activities in the energy, industry and land-use sectors in combination with emissions factors57. The outcomes are in line with the SSP projections of other models. The air pollutant emissions are used as input into the Fast Scenario Screening Tool (FASST) model to calculate particulate matter less than 2.5 μm (PM2.5) concentrations. The values used are the population-weighted average annual concentration of PM2.5. We used the WHO interim values of 10 and 25 mg m−3 as the planetary boundary and the high end of the risk zone, respectively58. The planetary boundary value is equal to the most ambitious interim value proposed by the WHO58 and between the value of 15 mg used by the Earth Commission17 and the WHO guideline of 5 mg (ref. 58). For the upper end of the uncertainty zone, we used the second interim value proposed in the WHO guidelines and, as a reference, we used the pre-industrial level indicated by CMIP6 calculations59.
Ocean acidification
Ocean acidification was estimated using a relationship between cumulative CO2 emissions and ocean pH using a correlation of SSP data and the pH numbers reported from CMIP6 (see below in Additional information on methods)60. The 2015 planetary boundary control variable values were translated into pH equivalents using the information available on the correlation between these variables.
Biogeochemical flows (imbalance of N and P cycles)
The IMAGE-GNM model represents the global nutrient cycle of N and P in detail. Key inputs from the IMAGE model are spatial-explicit patterns of cropland and grazing land, livestock numbers and N deposition. The model calculates the balance between inputs and outputs of N and P based on, among other items, water flows and retention and removal processes32. For N, the surplus input on cropland and pastures is used. For P, the total surplus is used. We used the planetary boundary and upper-end values from the 2015 planetary boundary framework, given that the values for the imbalance on agricultural soils are similar to the flow variables defined by the planetary boundary paper for 2015 (ref. 1), and the indicators are also conceptually linked.
Freshwater use
Freshwater withdrawals comprise water used for irrigated agriculture and extraction for municipal, industrial and energy use. Irrigation water availability and use is calculated in LPJmL fully coupled to IMAGE, which dynamically represents the hydrological cycle as well as the growth of crops, grass and natural vegetation using the concept of plant functional types54. The demand for non-agricultural water use is calculated in IMAGE using a detailed end-use-oriented model47. Water demand is met in the order (1) municipal, (2) industrial and energy and (3) irrigation. If insufficient water is available for irrigation, the crop model uses rain-fed yields instead. We used the global indicator proposed in the 2015 planetary boundary assessment—that is, the consumptive use of blue water (from rivers, lakes, reservoirs and renewable groundwater stores) as the global-level control variable, with 4,000 km3 yr−1 as the boundary value and 6,000 km2 yr−1 as the upper-end value.
Land system change
The demand for food, feed, timber and bio-energy is dynamically calculated in IMAGE61. In combination with changes in management and global trade, spatially explicit land-use change is determined as driving, for example, the conversion of natural ecosystems or the abandonment of agricultural land. For the planetary boundary, we used the total forest area from IMAGE in combination with the control values of the 2015 planetary boundary framework.
Biosphere integrity
For biosphere integrity, we used the MSA for terrestrial plants, based on the GLOBIO (version 4) model62. This indicator takes into account the impact of changes in land use and management, climate change and N deposition on terrestrial plant community intactness. The MSA is conceptually similar to the BII used in the 2015 planetary boundary assessment, with it focusing on the ‘naturalness’ of ecological communities compared to a reference state without significant human disturbance. In the 2015 planetary boundary assessment, BII values were only presented for southern Africa. For the MSA, we used a value of 90% for the planetary boundary, as also proposed in the 2015 planetary boundary paper for the BII, representing a highly natural and stable state. For the upper end of the uncertainty zone, we used a value based on the Earth Commission’s suggested range of 50–60% for intact natural systems based on nature’s contribution to people17, but we scaled this number to account for the difference in 2015 between the MSA indicator and the intactness indicator used by the Earth Commission (45–50% versus 54%).
In the 2015 Earth Commission assessment, the values for the Earth Commission and the upper end of the uncertainty zone were used to scale all the values (Extended Data Fig. 1). Because the results here vary over a wider range, we introduced a change on the low end (that only influences values below those of the Earth Commission). We used the pre-industrial value as an additional benchmark, set at zero (in the middle of the circle).
Scenario implementation
The main implementation of the IMAGE scenarios is discussed in the Supplementary Information (IMAGE 3.4 implementation of the scenarios used in: is world development within planetary boundaries possible?), while a more in-depth description of a very similar baseline scenario has been previously published (https://www.pbl.nl/sites/default/files/downloads/pbl-2021-the-2021-ssp-scenarios-of-the-image-3-2-model_4740.pdf). Data on the IMAGE scenarios are available at Zenodo (https://doi.org/10.5281/zenodo.10203631). In addition, Extended Data Table 1 provides a brief overview of the scenarios and Extended Data Table 2 provides an introduction to the main assumptions. A detailed description and a reference to the data files can be found in Supplementary Information Sections 1 and 2 (including the relevant references).
Additional information on methods
For the control variables for ocean acidification and N and P imbalance, some additional information is needed.
Ocean acidification
Plotting cumulative CO2 emissions (from the SSP scenario database) versus ocean pH, as calculated using the CMIP6 model60, provides a good correlation, as shown in Extended Data Fig. 2. Clearly, additional processes might play a role in ocean acidification (for example, the relationship with other biogeochemical cycles), but these are not captured in the current Earth system model results. This relationship is used to calculate the values for ocean acidification, expressed in pH, for the IMAGE scenarios. On the basis of earlier projections of future pH and the surface saturation state with respect to aragonite in the Southern Ocean under various Special Report on Emissions Scenarios, the proposed planetary boundaries for aragonite saturation could be translated in pH values63. It would be useful to better capture this Earth system process in future planetary boundary assessments.
N and P balance
The nitrogen use efficiency (NUE) in 2050 depends on the NUE in 2015 and the change between 1980 and 2015, which is corrected for the future N yield change relative to the historical yield change. The NUE values do not exceed those for SSP1. For regions where the NUE had a negative trend (East Africa, China and Korea), the future NUE was assumed to decline by less than 5% for the period 2015–2050. For China, a constant NUE of 0.38 is assumed after 2015, because current policy in China is actively reducing the N fertilizer load.
