Phytoplankton associations generally are identified by nonrandom co-occurrence of phytoplankton species, and are considered to be a result of similar species response to properties of the environment. Because an association is characterized by co-occurrence of species across samples, or synchrony of the abundance of species over time, we proposed that the perception of phytoplankton association is subject to spatial and temporal sampling scale and analysis method. We investigated this using an intensive phytoplankton dataset from three lakes of the Greater Yellowstone Ecosystem, with particular concentration on Asterionella formosa (AF) and Aulacoseira subarctica (AS). We compared numerical analyses, including Fager analysis of recurrent groups, rank correlation, and other measures of synchrony, for differences in assessment of co-occurrence. The analyses were conducted on the entire dataset, and then the dataset was reduced to assess the effect of different spatial or temporal sampling regimes on apparent association. While the two diatom species significantly co-occurred across samples, assessment of synchrony was affected by analysis method and sampling regime. AF and AS abundances were positively correlated across lakes and within each lake, but peaks in abundance of the two species did not significantly coincide over time. In addition, species abundances were negatively correlated across depths, and we argue that this is significant to the autecology of each species. These results indicate that the perception of phytoplankton association may simply be an artifact of the limits of our understanding of the distribution and physiology of the organisms.

Key words: associations, Asterionella formosa, Aulacoseira subarctica, synchrony