Review and test of reproducibility of subdecadal resolution palaeoenvironmental reconstructions from microfossil assemblages

Abstract

Very high, even annual, resolution quantitative palaeoenvironmental reconstructions have been generated from microfossil assemblage data using transfer functions. Evidence of the utility of some of these reconstructions is given by the high correlations between the reconstructions and instrumental records. Such performance is despite several ecological and taphonomic issues that can become more problematic at subdecadal resolution than at multidecadal resolution. This paper explores these timescale dependent issues and tests the reproducibility of the several of published reconstructions. It demonstrates that the previously reported exceptionally good performance of several published chironomid-based reconstructions is not reproducible.