Global Carbon Budget Projections for IPCC Climate Scenarios, 2024–2100(Generated by CTransformer Model)

Accurate quantification of the future global carbon budget (GCB) is vital for climate risk assessment and mitigation strategies. Traditional Earth System Models (ESMs) face high computational costs and uncertainties, limiting large probabilistic ensembles for uncertainty analysis. We present GCBS-Gen, a probabilistic dataset for GCB from 2024–2100, using generative deep learning with CTransformer, trained on 1959–2023 data.It offers 1,000 trajectories per five SSPs, covering 22 variables: fossil emissions (coal, oil, gas), land-use changes (deforestation, regrowth, peat fires), carbon sinks, CO₂ levels, and temperature anomalies. GCBS-Gen provides efficient, consistent open-access scenarios for IPCC frameworks, aiding policy simulation and carbon feedback studies.