DIATOM-BASED MODEL TO INFER PAST LITTORAL HABITAT CHARACTERISTICS IN THE ST-LAWRENCE RIVER

Environmental quality measurements for the St. Lawrence River are rela tively scarce for periods prior to the 1970s, yet long-term data are o ften required for effective ecosystem management. For example, in the St. Lawrence system, important questions remain, such as how have shif ts in macrophyte growth responded to past changes in nutrient loading. In this study, we develop a paleolimnological transfer function to pr ovide information on past aquatic habitat structure. Using diatom alga e collected from the three dominant habitats in the river (rocks, macr ophytes, filamentous algae), habitat preferences of the diatom taxa we re identified. Inference models were then developed to reconstruct pas t habitat conditions from fossil diatom assemblages using logistic sta tistical techniques. Logistic equations were developed by performing c orrespondence analysis (CA) on diatom assemblage data, and fitting log istic regressions to the sample ordinations. The three models appeared to reliably predict habitat types from the field data (observed versu s inferred r(2) values: rock = 0.85, Cladophora = 0.94, macrophyte = 0 .97). These diatom-based transfer functions can now be used on fossil diatoms preserved in dated sediment cores to infer past environmental changes, which cart assist the development of rehabilitation efforts i n the St. Lawrence River.