The role of fluid flow in dolomitization is evaluated through hydrologic simulations designed to capture and test numerically the essential elements of some common dolomitization models. A two-dimensional, finite-difference code (Bethke 1985) was used to simulate single-phase flow of variable-salinity fluids driven by compaction, convection, and topography in carbonate platforms and adjacent basinal sediments. A simple hydrostratigraphic model consisting of a platform-reef complex surrounded by basin-filling shale was constructed to examine the magnitude and direction of compaction flow during the initial phase of subsidence. Most pore fluids in the shale migrate to the surface at velocities less than 0.1 mm/yr and basin-derived fluids are focused into the reef from only a limited distance of 10 km. Relatively porous and permeable basinal limestones and sandstones may act as conduits for updip fluid flow into the platform-reef complex. Magnesium mass-balance constraints coupled with hydrologic simulations suggest that burial compaction is not an efficient mechanism for regional dolomitization. Two-layer hydrostratigraphic models involving variable-salinity fluids show that evaporation of seawater results in downward and seaward fluid migration; the magnitude of flow is a function of fluid density. Even fluids at salinities marginally elevated above normal seawater values are capable of penetration several hundreds of meters into the underlying sediments, albeit at rates of less than 1 cm/yr. Paleohydrologic simulations coupled with thermodynamic arguments support that reflux of surface-derived waters evaporated past gypsum saturation is a viable dolomitization mechanism. Calculations utilizing rates of Mg mass flux show that dolomitization of large carbonate platforms (>10 km 3 ) by fluids below halite saturation requires millions of years of reflux-driven flow. Thermal convection of seawater through carbonate platforms has been simulated by varying the basal heat flow. The thermal "pump" produces a convection cell in which lateral and vertical flow achieve rates of 75.0 and 13.0 mm/yr. respectively. The potential for dolomitization is high because a large reservoir of Mg (seawater) is tapped, elevated temperatures drive fluids further into the stability field for dolomite, and kinetic inhibition of dolomite nucleation and growth is decreased. Simulations of high-frequency glacio-eustatic sea-level oscillations suggest that a dynamic hydrologic regime may be established. A sea-level fall of 2 m lasting for 10,000 yr can drive lateral flow at rates up to 0.15 m/yr. Dolomitization is possible because marine waters migrate through the carbonate platform in response to gradients of hydraulic head created by eustatic sea-level fluctuations.