Seafloor hydrothermal systems have been intensively studied for the past few decades; however, the location of recharge zones and details of fluid circulation patterns are still largely uncertain. To better understand the effects of anhydrite precipitation on hydrothermal flow paths, we conduct 2-D numerical simulations of hydrothermal circulation at a mid-ocean ridge using a NaCl-H2O numerical code. The simulations focus on East Pacific Rise hydrothermal system at 950N due to availability of key observational data to constrain the models. Seismicity data that is available suggests that fluid flow is primarily along axis and that recharge is focused into a small zone near a 4th order discontinuity in the ridge axis. Simulations are carried out in an open-top square box 1500 m on a side maintained at a surface pressure of 25 MPa, and nominal seawater temperature of 10 C. The sides of the box are assumed to be impermeable and insulated. A constant temperature distribution is maintained along the bottom of the box consisting of a 1000 m long central-heated region maintained at 450 C to represent the axial magma chamber and ensure P-T conditions for phase separation; a linearly decreasing temperature profile from 450 to 300 C is maintained along the 250 m long segments adjacent to the heated region to delineate the recharge zone. We constructed a homogeneous model with a uniform cell size of 25 m with a permeability of 10-13 m2 and a similar model with a 200 m thick layer 2A region with a permeability of 10-12 m2. For the homogeneous model the simulations were run for 100 years to approximate steady state conditions and the model with layer 2A was run for 50 years. Assuming that anhydrite precipitation resulted from the decrease in solubility with increasing temperature as downwelling fluid gets heated, the rate of porosity decrease and sealing time was calculated at 50 and 100 years. The results showed that sealing occurred most rapidly at the bottom of the recharge areas near the base of the high-temperature plumes, where complete sealing occurred after ~55-625 years for an initial porosity of 0.1. The simulations also suggested that sealing would occur more slowly at the margins of the ascending plumes, with times ranging between ~ 80 and 5000 years. The sealing times in the deep recharge zone determined in these simulations are considerably greater than estimated from 1D analytical calculations, suggesting that with a 2D model, focused recharge at the EPR 950N site may occur, at least on a decadal time scale. More detailed analyses are needed to determine whether such focused recharge can be maintained for longer times. / Master of Science