Significant volumes of world hydrocarbon reserves occur in dolostones, and the majority of these reservoirs are interpreted to be of reflux origin. We used a two-dimensional numerical reactive transport model to investigate systematically the temporal and spatial distribution of replacement dolomitization, dolomite cementation, anhydrite cementation, and porosity evolution in a reflux system. We tested the sensitivity of reflux dolomitization to brine concentration (mesohaline to near-halite-saturated brines), near-surface temperature, flow rate, porosity-permeability feedback relationships, reactive surface area, and the effect of initial porosity-permeability heterogeneity. Simulations support the contention that hypersaline reflux is capable of extensive pervasive dolomitization, and that mesohaline reflux dolomitization is viable. Reflux generated a tabular dolomite body that is thickest close to the brine source and thins basinward. Incorporation of initial porosity and permeability heterogeneity resulted in a dolomite body with pronounced lateral fingers. Replacement dolomitization increased porosity (up to 8%), but postreplacement reflux resulted in the precipitation of minor dolomite cements (overclolomitization). Anhydrite cements that were spatially and temporally associated with replacement dolomitization caused a significant porosity reduction of up to 25%. Simulated rates of replacement dolornitization by reflux are fast, up to three orders of magnitude faster than seawater dolomitization. The rate of dolornitization and anhydrite mineralization proved to be critically sensitive to the flow rate and the brine chemistry. Temperature and reactive surface area were important controls on the rate of dolomitization, whereas the feedback relationship between porosity and permeability was a relatively moderate control. Our reactive transport models predict the general spatial and temporal trends in dolomite, porosity, and anhydrite observed in some major dolomite reservoirs.