Post‐agriculture rain forest succession on a tropical Pacific island

• Published in: Journal of vegetation science 2021-07, Vol.32 (4), p.n/a
• Main Authors: Webb, Edward L., Lalogafu’afu’a, Avele Iofi, Bult, Martin, Lee, Wei Kit, Fa’aumu, Siaifoi, Izuddin, Muhammad, MacDonald, Mark A., Meyer, Roger, Oh, Rachel Rui Ying, Tagarino, Alden P., Webb, Rachel C., Miles, Adam C., Woods, Kerry
• Format: Article
• Language: English
• Subjects: Abandonment; Agricultural ecosystems; Agricultural land; Agriculture; Biodiversity; Community composition; Composition; Convergence; Diagnostic systems; Dispersal; Dispersion; disturbance; Ecological succession; environmental filtering; Forest ecosystems; Forests; indicator species; Land use; legacy; Plant species; polynesia; Rain; Rainforests; Recruitment; Regeneration; secondary succession; Seed dispersal; seedling ecology; Seedlings; Species diversity; Species richness; Survival; Traditional farming; Tropical forests; tropical secondary forest
• ISSN: 1100-9233; 1654-1103
• DOI: 10.1111/jvs.13064

Questions The effects of agriculture on forest biodiversity and ecosystem properties can persist decades or centuries after abandonment. Many Pacific island nations are experiencing declines in traditional agriculture. Seed dispersal and seedling recruitment are critical to forest regeneration, and this is particularly relevant for Polynesia where native seed disperser diversity is low and species have been extirpated to crisis levels. In this study we asked: are secondary lowland tropical forests converging to mature forest, and are there seedling recruitment and survival differences between forest types? Location The Polynesian island of Tutuila, American Samoa, where traditional agriculture has been on the decline since the 1940s and secondary forests comprise 70% of all forest area. Methods We surveyed 34 vegetation plots (containing 2,232 adult trees and 6,579 tree seedlings) across mature forest and >50 years old secondary lowland rain forest. We quantitatively compared forest diversity, structure and the 3‐year seedling regeneration dynamics of the tree community. Results Although species richness and stem densities were similar between mature and secondary forest, most diversity and structure parameters remained distinguishable between the two forest types. Basal area remained significantly higher in mature forest; the diagnostic tree species of mature and secondary forest differed; and community composition (as indicated by NMDS) remained significantly different. However, seedling survivorship did not differ between forest types and seedling recruitment patterns indicated that there remained a suite of effective seed dispersers on Tutuila. Conclusions We conclude that (a) community differences between mature and secondary lowland rainforest were driven by the abundance of diagnostic species rather than exclusion of species from a forest type, and (b) there was no evidence of biophysical or ecological barriers to forest recovery of lowland secondary rain forest, or its convergence towards mature lowland rain forest. Nevertheless, more than 50 years after agricultural abandonment, secondary forest remains distinct from mature forest, highlighting that land use legacies are long‐term drivers of forest composition and successional trajectory in post‐agricultural landscapes of Polynesia. In a multi‐year study of tree and seedling communities, we found no evidence of biophysical or ecological barriers to post‐agricultural forest succession on a tropical