An updated land emulator dataset for energy models and integrated assessment: GLOBIOM-G4M lookup tables for the new SSPs in the context of ScenarioMIP7

Featuring multi-dimensional energy-land nexus and capturing the complex dynamics of the agriculture, forestry, and other land use (AFOLU) sectors without sacrificing computational efficiency remains a critical challenge in integrated assessment modelling (IAM). To support the generation of new Shared Socioeconomic Pathways (SSPs) for the ScenarioMIP-CMIP7 process, we developed a comprehensive updated set of GLOBIOM-G4M lookup tables available for IAMs and energy models. Operating as multidimensional land emulators, this updated dataset was generated utilizing an enhanced modelling pipeline that incorporates the latest SSPv3 macroeconomic drivers, novel agricultural add-on mitigation technologies, dynamic EAT-Lancet2 dietary shifts), alongside integrated updates in water, nutrient, and biodiversity modules. The resulting emulators provide detailed bioenergy cost-supply curves, agricultural market responses, and AFOLU marginal abatement cost curves (MACCs) across the SSP1-SSP5 as well as the Low Energy Demand (LED) pathways. Compared to CMIP6-era dataset, the updated lookup tables reveal deeper structural mitigation potential. For instance, in the SSP2 pathway and under a 100 USD/tCO₂eq carbon price, the rapid deployment of new agricultural technologies drives enhances the achievable net AFOLU CO₂ sink from approximately -1 GtCO₂/yr to -3 GtCO₂/yr under moderate bioenergy demand by the middle of the century. These emulators directly underpinned the land-use modelling for both the MESSAGEix-GLOBIOM-GAINS and WITCH-GLOBIOM frameworks in the ScenarioMIP7 intercomparison, demonstrating strong cross-model interoperability. To promote transparency and broader community usage, this dataset will be published as an open-access resource on Zenodo. This will enable standalone energy and IAM models to seamlessly integrate SSP-specific, cross-sectoral consistent land-system boundaries and dynamics, preventing unrealistic large-scale BECCS or afforestation deployments, and also reducing the computational burden of hard-linking full-scale endogenous spatial land model with energy models.