Spatial patterns of tree species distribution in New Guinea primary and secondary lowland rain forest

Questions: How do spatial patterns of tree distribution and species co-occurrence differ between primary and secondary tropical rain forests? What signatures of ecological processes might be discerned by comparing the spatial patterns of trees between primary and secondary forest plots? Location: Tropical rain forest vegetation, lowlands of Papua New Guinea. Methods: All trees over 5 cm DBH were surveyed in two non-replicated 1-ha plots situated in primary and secondary forest. Grid location, DBH, height and species identity were recorded for all surveyed trees. Analysis of the spatial pattern and the autocorrelation of tree sizes and identities were used to assess the structure of the forest found within the plots. Functions combining Ripley's K and the individual species-area relationship were applied to study the spatial distribution of trees and species diversity. Results: The spatial distribution of common species, and all stems collectively, was aggregated in the secondary forest plot but not different from random in the primary forest plot. Diameter and height were also strongly spatially auto-correlated in the secondary forest plot but not in the primary forest plot. Conspecific aggregations were more common in the secondary forest plot. Finally, the secondary forest plot was characterized by the presence of diversity-repelling species and lower diversity than the primary forest plot, where diversity-accumulating species were present. Conclusions: We attribute the weaker autocorrelation of tree size in the primary forest to the development of size hierarchies throughout the course of stand aging. The conspecific aggregation and low local diversity within the secondary forest plot are likely caused by dispersal limitation during a brief period of establishment after disturbance. The higher local diversity of the primary forest can be explained by the reduction of species aggregation through increased mortality of conspecifics. This is caused by strong intraspecific competition, supporting the spatial segregation hypothesis (interspecific spatial segregation).