|dc.description.abstract||Previous studies have shown that changes in community composition in mature kauri forests are driven by topographical gradients. The extent to which functional traits influence the patterns detected in these forests has not been well-studied. The overall objective of this thesis was to enhance our understanding of what drives community assembly in these forests by testing the importance of functional traits in mediating the relationship between environmental factors and community composition.
Puketī Forest (Northland, New Zealand) was selected as the study site because rapid turnover of species composition occurs over very strong topographic and edaphic gradients. Community composition was determined using forty permanent 400 m² plots that spanned the full gradient in soil properties and topographic variation. Leaf, height and wood traits were measured in thirty of these plots on three individuals (> 10 cm dbh) for each species on each plot. A total of 30 species were studied; the criteria for a species to be included was that the species had to attain a relative abundance of at least 0.05 within a single plot and occur on at least 10 percent of the forty plots sampled. Fully-expanded, mature, healthy, well-lit leaves were collected from the canopy using a pole pruner, shotgun or slingshot. Cores were collected using an increment borer. The relationships between community-weighted mean traits and environmental gradients were analysed using linear regression.
Results showed that there is a continuous compositional change along this environmental gradient and consequently there are few visually abrupt changes that occur between forest types. The exception to this is the kauri-dominated stands that are found in low fertility environments that are present on ridges. Soil variables and topography were found to be correlated, and the environmental variables that explain the majority of the variation in the community composition and functional trait data were soil carbon: nitrogen ratio, total soil phosphorus and soil pH, rather than topography.
Trait – environment regressions showed that traits are filtered along the soil fertility gradient. Species that have high SLA, high leaf and litter nitrogen, low leaf thickness, large seeds and low wood densities were found in the more fertile environments, which have a low soil carbon: nitrogen ratio, high total soil phosphorus and a high pH, such as a gully habitat. At the other end of this gradient, species that have low SLA, low leaf and litter nitrogen, high leaf thickness, small seeds and high wood densities were present in less fertile environments, which have a high soil carbon: nitrogen ratio, low total soil phosphorus and a low pH, such as a ridge top habitat.
Functional diversity indices indicate that the strength of environmental filtering changes along the soil fertility gradient. These indices show that trait variability does vary along the soil fertility gradient and that there are different community assembly processes operating in different places. The strength of the soil fertility as an environmental filter is strongest in low fertility soils, as environmental filtering results in convergence towards traits that confer a resource-retaining strategy.
Previous studies in kauri forests have shown changes in community composition to be determined by topography. This study has shown that these well-known patterns are driven by the filtering of leaf and wood traits along the topographically-driven soil fertility gradient.||