High-resolution physical and biological measurements were carried out in May-June 2007 during the ‘HABIT-Pontevedra 2007' survey in Ría de Pontevedra (Galician Rías Baixas, NW Spain) to study small-scale physical-biological interactions in the distribution of microphytoplankton, with special emphasis on harmful species. Longitudinal transects from the Ría to the adjacent shelf were sampled to describe the spring-neap tidal and circadian variability. An in situ particle profiler, a moored ADCP, and a towed undulating CTD (Scanfish) were used during the survey, which took place after an upwelling pulse at neap tides during a downwelling–upwelling cycle and coincided with the annual maximum of Pseudo-nitzschia spp. Cell maxima of P. seriata (2×106 cells L−1) and P. delicatissima (6×105 cells L−1) groups were observed during the first half of the cruise during downwelling, and a significant decrease in cell numbers occurred during subsequent relaxation-upwelling conditions. Thin layers were eroded during downwelling and formed again in the subsequent upwelling pulse. Cells of the P. seriata group were always dominant in terms of biomass but the contribution of the P. delicatissima group increased with stratification. Water exchange between the Ría and the adjacent shelf was mainly controlled by the upwelling/downwelling cycle, and tidal (both semidiurnal and spring-neap) variability appeared as a modulation of the response of the Ría circulation to wind variability. The circadian variability was regulated by tidal forcing, and Pseudo-nitzschia spp. maxima were associated with high stratification during low tide. The magnitude of spring-neap tidal and circadian variability has to be considered when designing and implementing water quality and harmful algae monitoring programmes. Blooms of Pseudo-nitzschia ssp. were not associated with the occurrence of domoic acid in shellfish even when Pseudo-nitzschia australis was dominant. These results confirm that just cell densities of the potential toxin producer are not enough for early warning in monitoring of Pseudo-nitzschia events.