A previous study noted the occurrence of a subsurface brine plume with more than 160 g/L sodium chloride perched above Welsh salt dome in southwestern Louisiana. Plumes of saline groundwaters have been reported above other Gulf Coast salt domes as well. The U.S. Geological Survey code SUTRA was used to model the formation of brine plumes around a generic salt dome by basin dewatering along the flanks of the dome, and also to study the gravitational instability of a brine plume initially perched above a dome, in the absence of an externally impressed driving force for upwelling. The simulations take into account the effect of salt concentration gradients on brine density. The first set of simulations showed that a brine plume initially perched above a salt dome sank rapidly in the absence of an overpressured section. Where observed, perched brine plumes above salt domes must therefore be either (1) young (< 2 Ma), if formed by basin dewatering and vertical fluid expulsion along the flanks of salt domes, or (2) formed by a continuous mechanism. The second set of simulations modeled the dewatering of a thick initially overpressured sedimentary section extending from a depth of 3 km to a depth of 10 km. The deflation rate of the overpressured section profoundly depended on whether or not a permeable conduit along the flank of the dome existed, which could collect water expelled radially inward to the flank of the dome from the overpressured sediments, and focus it upward along the flank of the dome. In most simulations the volume of compaction water was limited to only 10% of the total volume of overpressured sediment around the dome, and there was an ample amount of compaction water for a perched brine plume to form. In fact, when a perched brine plume formed, it formed very quickly (in < 0.01 m.y.), after a negligible degree of deflation of the overpressured section. It is conceivable that if cycles of conduit sealing and seal rupturing occurred, perched brine plumes could be formed several times during deflation. Conditions necessary for the formation of a perched brine plume are as follows: (1) waters must be focused laterally toward the dome, within in the overpressured section, (2) there must be a conduit of sediment with relatively high vertical permeability along the flank of the dome, and (3) sediments above the dome must be more permeable than sediments along the flank of the dome. For some simulations the average linear velocity of upward flow along the flanks of the dome was 3 m/yr, possibly of large enough magnitude to cause a thermal anomaly. If perched brine plumes form by basin dewatering, the thermal buoyancy effects associated with such a thermal anomaly may make the plumes rise to depths shallower than indicated in the simulations here. The inability of the code SUTRA to simultaneously model heat and solute transport did not permit an evaluation of this.