A mechanistic model of benthos community dynamics in the River Sihl, Switzerland

Abstract

1. Most of the existing river ecosystem models describe only part of the ecosystem, have site-specific features, and lack an adequate treatment of uncertainty. 2. We developed a dynamic model of the functional groups of non-filamentous and filamentous algae, scrapers, collector-gatherers and predators of the benthos community in rivers. This model is based on relatively universal model formulations and is intended to be a first submodel for a more complete model of the river ecosystem and the biogeochemical processes in the river. 3. The model was tested for the River Sihl in Switzerland. To take advantage of prior knowledge of model parameters, to account for their non-identifiability given the measured data, and to consider uncertainty of model parameters and results, Bayesian inference was used to update the prior probability distribution of the model parameters. 4. The calibrated model showed very good agreement of model results with measured data. None of the posterior parameter marginals showed a conflict with the prior. Some posterior marginals (specific growth rates of the functional groups, specific death rates, parameters characterizing self- and mutual shading of algae, and the critical level of discharge for catastrophic loss) were significantly narrower than the priors, indicating considerable gain of information from the measurements. For other parameters (half-saturation concentrations of algae and invertebrate growth, parameters characterizing temperature dependence of invertebrate growth rates, colonization rates and most coefficients of catastrophic loss), no significant information could be gained from the data. This clearly demonstrated the need for a Bayesian analysis. 5. The model seems to be a promising step towards a more comprehensive model of the river ecosystem and the biogeochemical processes in the river. However, more efforts are required: In particular, extension of the model by a description of the dynamics of benthic biofilms and nutrient dynamics and application of the (extended) model to more rivers to learn about the dependence of model parameters on river characteristics. The difficulty of finding adequate data sets for testing such and extended models demonstrates the need for comprehensive measurement campaigns that get information of periphyton, benthic invertebrates, benthic biofilms, suspended and sedimented organic particles, suspended organisms and nutrient concentrations and transformation rates in the water column. © 2008 The Authors.