Understanding drivers of eucalypt spatial patterns in temperate forests using point process modelling

The spatial patterning of trees in forests reflects past ecological processes. Mechanisms such as seed dispersal, competition, niche segregation, and facilitation operate across a range of scales, depending on the identity and attributes of the tree and its neighbours. Understanding how these processes contribute to the spatial patterns of trees is a major challenge for forest ecology. In particular, the spatial structure of temperate mixed-eucalypt forests, notable for their high congeneric canopy diversity of the genus Eucalyptus, is poorly understood. In this study, we aim to understand the drivers of fine-scale Eucalyptus spatial pattern in temperate forests. We used data from a 16-ha fully-mapped permanent forest dynamics plot in a mixed-eucalypt forest east of Melbourne, Australia. We applied the saturated pairwise interaction Gibbs point process model to quantify interactions between Eucalyptus trees and how environmental factors influence these interactions. Crown class was a significant driver of Eucalyptus spatial patterning. Subordinate trees were clustered, while more dominant trees had a regular arrangement, likely due to density-dependent mortality. Environmental variability in measures of radiative dryness and topographic wetness influenced coarse-scale spatial patterning of species through niche spatial segregation in the plot. These findings, where density-dependent mortality and niche segregation drive tree arrangement, were consistent with previous studies in tropical and temperate forests. Both crown class and species were essential to describing fine-scale eucalypt spatial arrangement, yet measures of relative tree height or size are rarely used in analyses of spatial patterns. Analyses purely based on environment or interactions can be incomplete. Similarly, analyses conducted at either species or crown class level, but not the combination, can result in uninformative, broad results of tree arrangement that lack fine-scale detail. We find that a fuller understanding of forest community assembly requires accounting for each of species, crown class, environment and interactions.