Modeling phytoplankton community in reservoirs. A comparison between taxonomic and functional groups-based models

• Published in: Journal of environmental management 2016-01, Vol.165, p.31-52
• Main Authors: Di Maggio, Jimena, Fernández, Carolina, Parodi, Elisa R., Diaz, M. Soledad, Estrada, Vanina
• Format: Article
• Language: English
• Subjects: Argentina; Comparative analysis; Dynamic parameter estimation; Ecosystem; Models, Biological; Models, Theoretical; Parameter estimation; Phylogeny; Phytoplankton - classification; Phytoplankton - physiology; Phytoplankton community; Phytoplankton functional groups; Plankton; Reproducibility of Results; Reservoirs; Taxonomy; Water Quality; Water quality mathematical model
• ISSN: 0301-4797; 1095-8630
• DOI: 10.1016/j.jenvman.2015.08.027

In this paper we address the formulation of two mechanistic water quality models that differ in the way the phytoplankton community is described. We carry out parameter estimation subject to differential-algebraic constraints and validation for each model and comparison between models performance. The first approach aggregates phytoplankton species based on their phylogenetic characteristics (Taxonomic group model) and the second one, on their morpho-functional properties following Reynolds' classification (Functional group model). The latter approach takes into account tolerance and sensitivity to environmental conditions. The constrained parameter estimation problems are formulated within an equation oriented framework, with a maximum likelihood objective function. The study site is Paso de las Piedras Reservoir (Argentina), which supplies water for consumption for 450,000 population. Numerical results show that phytoplankton morpho-functional groups more closely represent each species growth requirements within the group. Each model performance is quantitatively assessed by three diagnostic measures. Parameter estimation results for seasonal dynamics of the phytoplankton community and main biogeochemical variables for a one-year time horizon are presented and compared for both models, showing the functional group model enhanced performance. Finally, we explore increasing nutrient loading scenarios and predict their effect on phytoplankton dynamics throughout a one-year time horizon. •A water quality model for a non-stratified highly eutrophic reservoir was developed.•We formulate a parameter estimation problem subject to a PDE system.•We compare two approaches for modeling phytoplankton: taxonomic and functional groups.•We show that functional group approach more accurately describes phytoplankton dynamics.