Phytoremediation is an attractive treatment technology for many contaminated sites due to its cost effectiveness and public acceptance. We present a sensitivity analysis of important parameters from a screening level model for phytoremediation by grass species of weathered petroleum-contaminated sites. The conceptual framework is that root movement through contaminated soil will enhance contaminant biodegradation by providing a local environment more favorable for petroleum degrading microorganisms--the so-called rhizosphere effect. Common questions in phytoremediation are, "What species should be planted?" and "What management practices should be followed?" These choices may affect degradation kinetics, root biomass (and therefore rhizosphere volume), and the root turnover. Important model parameters are the rate constants, rhizosphere volume, and the rate of root turnover. We present a sensitivity analysis with the aim of identifying the most important factors for improving phytoremediation effectiveness. For simulations of the phytoremediation of weathered diesel range organics, our results indicate that annual species, with higher root turnover, are preferred over perennial species with the caveat of equal degradation rate constants, that is, no species-dependent effects. In addition, the results suggest that the management of nonrhizosphere soil could play an important role in the overall effectiveness of phytoremediation. Finally, the effect of increasing root biomass or increasing the rhizosphere thickness is approximately equivalent with respect to the ultimate removal of the contaminants.