Periphytic diatom assemblages from three major substrates (rocks, macrophytes, and the filamentous alga Cladophora glomerata) were sampled from fifty sites in the St. Lawrence River, between the cities of Cornwall and Quebec. Over 300 diatom taxa were identified, 67 of which occurred at relative abundances of >1% in three or more samples. Using multivariate statistical techniques (correspondence analysis, canonical correspondence analysis) the environmental affinities of the diatoms to pollution-related variables and habitat substrates were determined, and quantitative inference models were constructed to infer certain environmental variables (conductivity, suspended solids concentration, and shoreline habitat characteristics).

Sediment cores were taken from two fluvial lakes in the St. Lawrence River: one downstream from Cornwall (Lake Saint-Francois, 3 cores), and the other near Montreal (Lake Saint-Louis, 1 core). Diatoms preserved in the 19th century sediments were primarily benthic taxa with an affinity for relatively clean water. Early in the 20th century, the diatom assemblage became dominated by taxa characteristic of extensive macrophyte growth. By the 1950s, many eutrophic taxa (particularly the planktonic Stephanodiscus species) became dominant. In Lake Saint-François, eutrophic planktonic diatoms were then rapidly replaced by benthic taxa since about the 1970s, indicating that recent remedial actions (e.g. wastewater treatment, removal of phosphates from detergents) may be working.

This study provided the first application of the diatom-based inference model for reconstructing shoreline habitat characteristics, such as macrophyte density. The model appears to sufficiently reconstruct environmental conditions in the St. Lawrence River during the last century. The most notable shift has been an increase in diatom taxa commonly associated with macrophyte substrates.

The paleolimnological information obtained in this study provides baseline data for sound management decisions and remediation goals for St. Lawrence River ecosystem recovery programs. Additionally, ecological data gathered for the river diatoms will aid biomonitoring programs. Moreover, this study illustrates that detailed paleoecological investigations in large rivers are possible, if sites are carefully chosen.