The large-scale implementation of Carbon Capture, Utilisation, and Storage (CCUS) requires an available and flexible transport and storage network that can handle varying CO₂ supply rates from different sources. Such a network must provide flexibility in transport, storage, and utilisation capacity to accommodate the supply of captured CO₂. Numerical simulations have been used to predict storage performance and capacity under various conditions, however, they require substantial resources and time to run. A proxy model for saline aquifers characterised by a dome structure was developed, with a focus on the Bunter Closure 36 located in the Southern North Sea in the UK. The proxy model was built using the machine learning technique XGBoost (Extreme Gradient Boosting) to provide estimation for the dome pressure, injection well bottomhole pressure (BHP) and injected CO₂ cumulative under variable injection rates and multi-realisations of storage parameters