We created a model to study the process in which nutria (Myocastor coypus) feeding activities lead to erosion and loss of marsh area. This model ties together data on nutria population dynamics and feeding behavior from the literature with data from field studies on the phenology of Scirpus americanus and Spartina patens conducted in the Barataria Basin, Louisiana, USA in 1992. The complete model consists of three linked models: a model of nutria population dynamics (nutria model), a model of the annual marsh biomass cycle of Scirpus americanus and Spartina patens (biomass model), and a plant-biomass density-dependent marsh area model (area model). When all three models are linked together, they form the 'nutria-biomass-area model.' Analysis of the models indicated the following. (1) The high population densities and low survivorship rates as reported in the literature are incompatible. (2) the nutria model is sensitive to adult and juvenile survivorship and, to a lesser extent, young born per female. It is not particularly sensitive to gestation periods, impregnation rates, or time to maturity. (3) The marsh area model is not sensitive to the marsh loss equation nor to the density at which loss of marsh area begins but is sensitive to the amount of biomass destroyed per nutria. (4) Nutria numbers do not significantly decrease in the nutria-biomass-area model until the total marsh area approaches zero because marsh loss occurs only during winter when marsh biomass is at its annual low.