Assessing the importance of a self-generated detachment process in river biofilm models

1. Epilithic biofilm biomass was measured for 14 months in two sites, located upand downstream of the city of Toulouse in the Garonne River (south-west France). Periodical sampling provided a biomass data set to compare with simulations from the model of Uehlinger, Bürher and Reichert (1996: Freshwater Biology, 36, 249–263.), in order to evaluate the impact of hydraulic disturbance. 2. Despite differences in application conditions (e.g. river size, discharge, frequency of disturbance), the base equation satisfactorily predicted biomass between low and high water periods of the year, suggesting that the flood disturbance regime may be considered a universal mechanism controlling periphyton biomass. 3. However modelling gave no agreement with biomass dynamics during the 7-month long low water period that the river experienced. The influence of other biomassregulating factors (temperature, light and soluble reactive phosphorus) on temporal biomass dynamics was weak. 4. Implementing a supplementary mechanism corresponding to a temperature-dependent self-generated loss because of heterotrophic processes allowed us to accurately reproduce the observed pattern: a succession of two peaks. This case study suggests that during typical summer low water periods (flow stability and favourable temperature) river biofilm modelling requires self-generated detachment to be considered.