In large lakes, temporal variability is compounded by strong spatial variability associated with mesoscale physical processes such as upwelling and basin-scale circulation. Here we explore the ability of a three dimensional model (ELCOM–CAEDYM) to capture temporal and spatial variability of phytoplankton and nutrients in Lake Erie. We emphasized the east basin of the lake, where an invasion by dreissenid mussels has given special importance to the question of spatial (particularly nearshore–offshore) variability and many comparative observations were available. We found that the model, which did not include any simulation of the mussels or of smaller diffuse nutrient sources, could capture the major features of the temperature, nutrient and phytoplankton variations. Within basin variability was large compared to among-basin variability, especially but not exclusively in the western regions. Consistent with observations in years prior to, but not after, the mussel invasion the model predicted generally higher phytoplankton concentrations in the nearshore than the offshore zones. The results suggest that the elevated phytoplankton abundance commonly observed in the nearshore of large lakes in the absence of dreissenid mussels does not have to depend on localized nutrient inputs but can be explained by the favourable light, temperature and nutrient environment in the shallower and energetic nearshore zone. The model is currently being extended to allow simulation of the effects of dreissenid mussels.