Reconstructing Holocene water tables in New Zealand using testate amoebae: differential preservation of tests and implications for the use of transfer functions

Abstract

Testate amoebae analysis is used to provide the first quantitative reconstruction of Holocene surface wetness changes from a Southern Hemisphere ombrogenous bog. Water-table depth and surface moisture are the dominant environmental factors influencing testate amoebae species composition in a modern training set of 62 surface samples from 19 bogs in New Zealand. A transfer function based on partial least squares (PLS) performed well in cross-validation of modern samples, but produced unrealistic predictions of water-table depth and moisture for fossil samples. This resulted from poor overlap of species composition between modern and fossil samples. Peat humification data is used to examine decay-related species diversity trends and ordination is used to explore the differences between modern and fossil assemblages. Poor overlap can be explained by differential preservation in the fossil assemblages of those taxa with more robust tests. This enhances their abundance and reduces the overall diversity of the fossil assemblages. Human activities over the last 150 years have also influenced the surface environments and hydrology of the sampled bogs. In this situation, weighted average (WA) and tolerance downweighted weighted average (WA-Tol) models were more robust than PLS, and were applied, along with semi-quantitative scoring, to reconstruct more realistic Holocene surface wetness changes. Selection of the best model for reconstruction of surface wetness changes should be based on the characteristics of the fossil data rather than the predictive power of the model in cross-validation of the modern data alone.