The role of plant functional traits in understanding forest recovery in wet tropical secondary forests.

Simultaneous measurement of plant functional traits and the regeneration environment should shed light on the plant-environment interactions and feedbacks as secondary forest regenerates. However, little of such work has been done in the wet tropics, and even fewer studies have examined soil nutrients. We investigated whether plant functional traits and environmental variables explain the varied recovery of secondary forests in Singapore. Our study plots included three primary forest plots and eight approximately 60-year-old secondary forest plots regenerating from intensive agricultural activities. Using 35 seedling quadrats, we asked: Q1) How do environmental variables explain the variation in seedling functional traits observed in primary and secondary forests? Q2) How do seedling traits, adult traits and environmental variables relate and explain variation in species richness and stem density in secondary forests? We found that both light and soil fertility explained the shifts in plants functional traits from poorly recovering secondary forests to primary forests. Poor forest regrowth was correlated with high soil aluminum levels and lower leaf nitrogen concentrations. Low nutrients and high aluminum saturation were also negatively correlated with seedling species richness, but not stem density, in the secondary forests. Forest recovery is probably slowed by positive feedback between slower nutrient returns from slow decaying litter and further recruitment of nutrient conserving species, as indicated by positive correlations among adult leaf CN ratio, litter depth, soil CN ratio and quadrat level CN ratio. Plant functional traits are indicative of the strategies of successful seedlings and do not necessarily relate to overall forest recovery. Hence, while some specialist plant species are able to accrue high nutrients on degraded soils with aluminum toxicity and low nutrients, species richness on these soils was poor. This underscores the need to concurrently measure environmental variables and plant traits when investigating the mechanisms driving changes during forest recovery. (C) 2018 Elsevier B.V. All rights reserved.