Interactions between soil fertility and climate drive variation in functional traits in New Zealand forests.

Abstract

Plant functional traits provide a mechanistic approach to understanding the processes of environmental filtering and community assembly. Variation along environmental gradients results in changes in the adaptive values of traits. Climate and soil fertility are two dominant factors that drive these patterns of trait variation. These two factors simultaneously select for traits during environmental filtering. However, we do not understand how interactions between climate and soil fertility influence the variation in community-level traits of multiple plant organs. The roles of traits in New Zealand forests are also yet to be studied at a national scale.

This thesis aimed to determine the adaptive values of multiple functional traits across broad climate and soil fertility gradients in New Zealand forests. This was achieved by the following methods. Data were collected for leaf, stem, root, seed, flowering, and whole-plant traits from the 64 most common native trees in forests nationwide. Community composition and soil properties were measured at 324 plots across the country. For each plot, long-term average climate data were extracted from a model. A variable representing variation in soil fertility in the plots was derived by principal components analysis (PCA). Community-weighted mean (CWM) functional traits, i.e. average trait values weighted by the abundance of species, were calculated for each plot. Dimensionality of the specie-trait matrix was determined by PCA. Multiple linear regression was used to model the variation of each of the CWM traits as functions of mean annual temperature (MAT), vapour pressure deficit (VPD), soil fertility, soil fertility × MAT interaction, soil fertility × VPD interaction, total basal area and topography.

Five dimensions of trait variation were identified among New Zealand trees. Soil fertility was a more significant predictor of CWM traits than either of the climate variables. However, both of the interaction effects were significant for most traits and overrode the importance of the main effects. For example, in sites with high fertility soil, leaf economics traits varied from ‘slow’ in cool and dry conditions to ‘fast’ in warm and moist conditions, but in sites with low fertility, these traits were ‘slow’ in all climates. Therefore, the adaptive values of multiple functional traits of New Zealand forests varied depending on both soil fertility and climate.

This thesis provides the first recognition of the significant roles of the interaction effects between soil fertility and climate in driving variation in CWM traits from multiple plant organs. Climate and soil fertility interact in a way that influences CWM trait values independently from the influence of each environmental variable. These interactions are suspected to be important globally and should be tested for worldwide to confirm the generality of their effects. In conclusion, this thesis demonstrates that studying the relationships between CWM traits and soil fertility or climate independently is insufficient, when attempting to understand the process of environmental filtering. It is critical that the interaction effects between climate and soil fertility are included in future studies to enhance our understanding and ability to predict community-level responses to processes such as climate change.