Defaunation changes leaf trait composition of recruit communities in tropical forests in French Guiana

Abstract

Hunting impacts tropical vertebrate populations, causing declines of species that function as seed dispersers and predators, or that browse seedlings and saplings. Whether and how the resulting reductions in seed dispersal, seed predation, and browsing translate to changes in the tree composition is poorly understood. Here, we assess the effect of defaunation on the functional composition of communities of tree recruits in tropical rainforests in French Guiana. We selected eight sites along a gradient of defaunation, caused by differences in hunting pressure, in otherwise intact old-growth forests in French Guiana. We measured shifts in functional composition by comparing leaf and fruit traits and wood density between tree recruits (up to 5 cm diameter at breast height) and adults, and tested whether and how these compositional shifts related to defaunation. We found a positive relationship with defaunation for shifts in specific leaf area, a negative relationship for shifts of leaf toughness and wood density, and a weak relationship for shifts in fruit traits. Our results suggest that the loss of vertebrates affects ecological processes such as seed dispersal and browsing, of which browsing remains understudied. Even though these changes sometimes seem minor, together they result in major shifts in forest composition. These changes have long-term ramifications that may alter forest dynamics for generations.

Keywords: functional composition; functional trait; hunting; leaf toughness; specific leaf area; wood density.

FIGURE 1

(a) Location of the eight study sites in French Guiana. The hatched area is the area with landscape type “forests of plateaus and high hills” (Guitet et al., 2013). (b) Location of plots (rectangles) relative to the site center point. Three of four cardinal directions were chosen for each site based on a set of criteria. (c) Spatial layout of a plot: Seedlings were sampled in the central plot of 1 × 100 m, saplings, juveniles, and treelets were sampled in the plot of 5 × 100 m and adults were sampled in the surrounding plot of 40 × 125 m.

FIGURE 2

Relationship between the recruit:adult ratios of the community weighted means (CWM) or community weighted proportions (CWP) of eight functional traits and defaunation index based on all censused vertebrate species (D _all) or only ground dwelling browsers (D _leaf). Values greater than one indicate that the CWM or CWP is higher in the recruit community, compared to the adult community. Numbers correspond with site numbers. (a) Leaf nitrogen content, (b) leaf thickness, (c) leaf toughness, (d) specific leaf area, (e) mean fruit length, (f) fleshy fruits, (g) dry fruits, (h) wood density. Seedlings, saplings, juveniles, and treelet communities are represented, respectively, by circles, diamonds, squares, and triangles. Lines indicate least squares regression, with solid lines indicating significant relationships and dashed line indicating non‐significant relationships. Color coding represents, from dark to light, seedling, sapling, juvenile, and treelet classes. Pearson correlations are shown in Table 2.