One use of groundwater flow models is to simulate contributing recharge areas to wells or springs. Particle tracking can be used to simulate these recharge areas, but in many cases the modeler is not sure how accurate these recharge areas are because parameters such as hydraulic conductivity and recharge have errors associated with them. The scripts described in this article (GEN_LHS and MCDRIVER_LHS) use the Python scripting language to run a Monte Carlo simulation with Latin hypercube sampling where model parameters such as hydraulic conductivity and recharge are randomly varied for a large number of model simulations, and the probability of a particle being in the contributing area of a well is calculated based on the results of multiple simulations. Monte Carlo simulation provides one useful measure of the variability in modeled particles. The Monte Carlo method described here is unique in that it uses parameter sets derived from the optimal parameters, their standard deviations, and their correlation matrix, all of which are calculated during nonlinear regression model calibration. In addition, this method uses a set of acceptance criteria to eliminate unrealistic parameter sets.