摘要
Small, shallow, temperate lakes are predominant landscape features in North America, however, little is known about their long-term ecosystem dynamics, and few data exist on the chironomid fauna they harbor. Using multivariate analyses, we defined relationships between sub-fossil chironomid assemblage composition and environmental variables in 26 shallow lakes of northeastern USA and quantified how differences in taxonomic resolution affect transfer function model performance. Using redundancy analysis, we found that chironomid assemblages are best explained by turbidity, dissolved inorganic carbon and drainage basin/lake area ratio. Turbidity explained the greatest proportion of variance found in the chironomid assemblage (10.4%), followed by total nitrogen. Through ordination analyses and an analysis of similarity, we found that macrophyte density was also a significant predictor of chironomid assemblages. We used partial least squares analysis to develop a robust model for quantitative reconstruction of turbidity, with r jack2 = 0.62. When using a more coarsely resolved taxonomic dataset, we found that model performance statistics were weaker, suggesting the need for fine-resolution taxonomy. Overall, our findings highlight the importance of variables related to lake trophic state in structuring chironomid assemblages in shallow, temperate lakes and provide tools for inferring past ecological changes in these ecosystems.