A synthetic approach for analyzing tropical tree spatial patterns through time

Spatial patterns of tropical tree populations may contribute to the identification of underlying ecological mechanisms such as dispersal, competition and pest and pathogen pressure. Classical structure functions for quantifying spatial patterns have limitations, for example, although Ripley’s K function is useful for identifying differences in spatial patterns through space, this makes it difficult to identify differences in spatial patterns through time. To complement Ripley’s K function, we use the Thomas process, a point pattern model, to quantify temporal changes in spatial patterns for five tropical tree species known to have density-dependent mortality from species-specific pests and/or pathogens on Barro Colorado Island, Panama. We fit the data of each species to the model to estimate the cluster size (σ) and density of cluster centres (ρ). We compared spatial patterns based on the pair correlation function of the Thomas process determined from σ and ρ. We found reduced clustering over the past 20 years for Ocotea whitei (Lauraceae) and Quararibea asterolepis (Bombacaceae), the two species with the best-documented cases of pathogen and pest outbreaks on BCI, suggesting support for the Janzen-Connell Hypothesis; however, we reject the hypothesis because neither species showed spatial regularity throughout the census period. We found that the Thomas process point pattern model used in concordance with Ripley’s K function and other classical structure functions in ecology can be a simplified and powerful method for testing hypotheses regarding changing spatial patterns of populations through time