Performances (survival and growth) and functional traits of tree seedlings along light and water gradients

A foundational assumption of trait-based ecology is that individual
performances should be predicted by its functional traits. However, the
trait-performance relationships reported in literature were typically
weak, probably due to the ignorance of interactions between
traits and environments, intraspecific trait variability and hard traits
(directly linked to physiological processes of interest). We
conducted an experiment of planting 900 seedlings of six tree species
separately (one seedling per pot) along experimentally manipulated light
and water gradients, monitored their survival and growth, and measured
their morphological, photosynthetic and hydraulic traits. Most
trait-performance relationships depended on the environments, either
marginally changing (weak trait × environment interaction) or even
reversing (strong trait × environment interaction) along light or water
gradients in our experiment. Such trait × environment
interactions were more likely to be detected in growth models
using individual-level traits than models using species mean
traits, but seedling growth was not better modelled with
individual-level traits than species mean traits. Additionally, none of
the hard traits (photosynthetic and hydraulic traits) were better
predictors than soft traits (morphological traits) modeling seedling
growth and survival along light and water gradients. Our study highlights
the necessities of considering trait × environment
interactions when predicting response of plants to changing
environments. The benefits of using individual-level traits or hard traits
to predict plant performance might be reduced or even cancelled if their
measurement errors are not well controlled.