We run a comparative study of the results of flume experiments and several dynamic models reproducing the effects of streamflow variability on biofilm (i.e. periphyton) temporal dynamics. During the experiment, two contrasting flow regimes, characterised by a constant and a time-varying discharge temporal sequence, and four different light conditions (from 90% to 27% transmission of incident light) were performed to test the effects of availability and temporal variability of light and streamflows on biofilm growth. Several model formulations, describing growth and loss dynamics, have been explored in order to assess the relevant processes that controlled biofilm temporal pattern. Model identification criteria were used to identify the most suitable model, in which the growth rate is found to be dependent on density-limitation dynamics coupled with a saturating light effect, while the loss rate is linearly proportional to the discharge conditions experienced in the flumes. This model formulation proved able to reproduce remarkably well the observed biofilm dynamics. In order to analyse the stationary behaviour of the best-performing model reproducing biofilm biomass dynamics, we also run a long-term simulation, where no significant biomass differences between the constant and stochastic flow regimes were detected. © 2012 John Wiley & Sons, Ltd.