WA-PLS diatom-based pH, TP and DOC inference models from 42 lakes in the Abitibi clay belt area (Quebec, Canada)

The relationship between surficial sediment diatom species and measured environmental variables was explored in lakes from the Abitibi region of western Quebec. Diatom assemblages in 42 lakes were identified and their relationship with measured environmental variables was examined using multivariate statistical methods. Canonical correspondence analysis with forward selection and Monte Carlo permutation tests revealed that the three environmental variables pH, TP and DOC each accounted for statistically significant fractions of the variation in diatom taxa. A training set with 164 modern diatom taxa was used to derive transfer functions for lake-water pH, TP and DOC using weighted-averaging-partial-least-squares (WA-PLS) techniques. The models were developed to infer lake water pH, TP and DOC within the ranges 4-8 for pH, 2.75-30.0 mug l(-1) for TP, and 2.9-18.5 mg l(-1) for DOC. These quantitative inference models may now be used to help identify and estimate the effects of natural disturbances on the biogeochemistry of Abitibi lakes during their historical development.