<![CDATA[Carbonate reservoirs are extremely heterogeneous owing to fractures, vugs, micropores, and karst features. This heterogeneity complicates fluid-flow prediction, characterisation, and simulation in both naturally fractured reservoirs and carbon capture and storage (CCS). Dual-porosity modeling remains the dominant approach for capturing matrix–fracture interactions in these systems. This review critically assesses geological, petrophysical, geophysical, and dynamic characterization techniques alongside conventional dual-porosity, dual-permeability, MINC, DFN, and hybrid continuum–discrete methods for carbonate reservoirs, with emphasis on CCS applications. Case studies from Central Luconia and Abu Dhabi demonstrate that dual-porosity workflows outperform single-porosity approaches in multiscale pore-system characterization, permeability prediction, and dynamic behavior forecasting. Nevertheless, significant uncertainties persist regarding fracture connectivity, multimodal pore networks, interporosity transfer functions, and history-matching non-uniqueness, directly impacting CO₂ injectivity, plume migration, and long-term containment security. The review further examines advances in digital rock physics, machine-learning-assisted workflows, uncertainty-aware simulation, and coupled thermal–hydraulic–mechanical–chemical modeling. Though no single framework is universally applicable, future progress depends on improved multiscale characterization, rigorous uncertainty quantification, coupled multiphysics simulation, and computationally efficient workflows that balance geological realism with practical field-scale implementation.]]>
