We present a particle-based method to simulate carbonate dissolution at the pore scale directly on the voxels of three-dimensional micro-CT images. The flow field is computed on the images by solving the incompressible Navier-Stokes equations. Rock-fluid interaction is modeled using a three-step approach: solute advection, diffusion, and reaction. Advection is simulated with a semianalytical pore-scale streamline tracing algorithm, diffusion by random walk is superimposed, while the reaction rate is defined by the flux of particles through the pore-solid interface. We derive a relationship between the local particle flux and the independently measured batch calcite dissolution rate. We validate our method against a dynamic imaging experiment where a Ketton oolite is imaged during CO2-saturated brine injection at reservoir conditions. The image-calculated increases in porosity and permeability are predicted accurately, and the spatial distribution of the dissolution front is correctly replicated. The experiments and simulations are performed at a high flow rate, in the uniform dissolution regime – Pe ≫ 1 and PeDa ≪ 1—thus extending the reaction throughout the sample. Transport is advection dominated, and dissolution is limited to regions with significant inflow of solute. We show that the sample-averaged reaction rate is 1 order of magnitude lower than that measured in batch reactors. This decrease is the result of restrictions imposed on the flux of solute to the solid surface by the heterogeneous flow field, at the millimeter scale.