Calibration strategy for dynamic succession models including several phytoplankton groups

Abstract

An accurate description of the relationships between physical environment and ecological conditions, and a proper representation of phytoplankton succession structure, are central topics in water quality modeling. In this work a new calibration strategy is proposed for a dynamic ecological model, that differentiates five phytoplankton groups. Sensitivity analysis and global automatic calibration algorithms are successfully applied to the deterministic, highly parameterized model to solve over parameterization and reduce time of calibration. When the level of description of phytoplankton groups was adequate to system complexity, it was possible to calibrate phytoplankton groups separately and then simulate them toghether without reducing model performance. Model results reproduced the seasonal evolution of most of the algal groups considered in simulation, when an adequate information set was included in the calibration process.Through the applied method we prove that a good representation of phytoplankton succession is feasible without simplifying model structure or limiting number of simulated phytoplankton groups. Our findings illustrate that integration of automatic calibration strategies in complex deterministic ecological models is a useful approach to improve model performance.