Résumé
The process of wastewater treatment was studied by modeling the relationships between physical, chemical, and biological (bacteria, phytoplankton, zooplankton) components of the sewage treatment lagoons of an urban wastewater center, based upon a two-year sampling program. The models of interactions between variables were tested by path analysis. The path coefficients were computed from the results of ridge regression, instead of linear multiple regression. The results show that fecal coliforms were effectively controlled by the environmental variables included in the model, which have a cyclic seasonal behavior. This control grew stronger with distance from the input (R2= 0.71) to the output (R2= 0.88) of the treatment plant, resulting in effective elimination of most enteric bacteria. Simultaneously, the ecosystem's community of aerobic heterotrophic bacteria became more independent from the model's predictive variables, with increased distance from the sewage input, thus demonstrating its maturation as an autonomous community in the lagoon ecosystem. Consequences of modeling are discussed, with respect to the understanding of biological wastewater treatment mechanisms and ecosystem dynamics and to plant management.