No abstract

Chironomid assemblages from the uppermost sediments of 435 lakes spanning northern North America were compared to environmental parameters using direct gradient analysis. This large calibration set was merged from several previously developed regional datasets, and increases the number of modern analogues that are available for use for paleoenvironmental interpretations in this region. Air temperature explained the largest amount of variation in the chironomid assemblages with several other environmental factors accounting for statistical significant amounts of the remaining variance. A robust inference model for deriving past mean July air temperatures from subfossil chironomid assemblages was developed and applied to previously published paleoclimate reconstructions from the High-Arctic, Middle-Arctic, Boreal treeline, and Alpine regions of northern North America. The patterns of the temperature reconstructions from the combined dataset were generally similar to the original reconstructions, but with colder inferred temperatures reflecting the incorporation of a larger number of modern sites from colder climates in the combined dataset. This analysis demonstrated that the larger temperature gradient available in the new training set, when compared to the temperature gradients in the original training sets, provides a better estimation of chironomid-environment relationships. In particular, the optima and tolerances estimated using the larger, combined dataset should be more accurate, and therefore, improve midge-based paleoclimate reconstructions for northern North America. Despite the much larger spatial scale and greater associated environmental heterogeneity now incorporated in the combined dataset, this study suggests that in most cases the overarching constraint governing Electronic supplementary material The online version of this article (

No abstract

No abstract