Developing diatom-based inference models to assess lake ecosystem change along a gradient of metal smelting impacts:...

Mining and smelting activities have strongly influenced the Sudbury region (Ontario, Canada) since the late 19^th century, leading to acidification and metal contamination in many local ecosystems. Regulations on restricting acidic emissions were enacted in the 1970s, after which a considerable volume of paleolimnological work was completed to study the impacts of acidification on Sudbury-region lakes and their subsequent biological recovery. Twenty years after the last regional diatom-based assessment, many lakes have undergone large changes in limnological variables, including increases in pH and dissolved organic carbon concentrations, as well as decreases in metal concentrations. Additionally, these lakes are under the potential impacts of newly emerging environmental stressors such as climate warming and road salt contamination. Here, we revisited a suite of Sudbury-region lakes (n = 80) to examine the relationships between their current water chemistry and diatom assemblages preserved in surface sediments using a canonical correspondence analysis. Although the pH gradient in our study lakes is shorter (pH ~1.4) than in earlier calibration studies conducted in this region, lake water pH was still identified as the strongest environmental variable shaping diatom distributions and was used to construct a robust inference model (R²_boot  = 0.73; RMSEP = 0.32). By assessing ecological changes experienced by a subset of these Sudbury-region lakes (n = 33) over the past few decades, we identified two major trends: an overall increase in diatom-inferred pH and a rise in the relative abundance of planktonic taxa. Our study provides useful insights into the autecology of major diatom taxa in acidified waters and highlights the importance of considering other anthropogenic stressors when assessing the recovery response of acid-impacted systems.