Consistent, small effects of treefall disturbances on the composition and diversity of four Amazonian forests

Abstract

Understanding the resilience of moist tropical forests to treefall disturbance events is important for understanding the mechanisms that underlie species coexistence and for predicting the future composition of these ecosystems. Here, we test whether variation in the functional composition of Amazonian forests determines their resilience to disturbance.We studied the legacy of natural treefall disturbance events in four forests across Amazonia that differ substantially in functional composition. We compared the composition and diversity of all free-standing woody stems 2-10 cm diameter in previously disturbed and undisturbed 20 × 20 m subplots within 55, one-hectare, long-term forest inventory plots.Overall, stem number increased following disturbance, and species and functional composition shifted to favour light-wooded, small-seeded taxa. Alpha-diversity increased, but beta-diversity was unaffected by disturbance, in all four forests.Changes in response to disturbance in both functional composition and alpha-diversity were, however, small (2 - 4% depending on the parameter) and similar among forests. Synthesis. This study demonstrates that variation in the functional composition of Amazonian forests does not lead to large differences in the response of these forests to treefall disturbances, and overall, these events have a minor role in maintaining the diversity of these ecosystems.

Keywords: alpha‐diversity; beta‐diversity; determinants of plant community diversity and structure; functional composition; maximum height; seed mass; tropical forest; wood density.

Figure 1

Box plots of variation in (a) stem density and mean values of (b) wood density, (c) leaf mass per unit area, (d) seed mass, (e) maximum height and (f) mean nitrogen concentration for stems 2–10 cm diameter in forests in four regions of Amazonia (north and south Peru, Central Amazonia (Manaus; CA) and the Guiana Shield (Nouragues; Gui). Significant differences among sites shown as P < 0.05 (*), P < 0.01 (**), P < 0.001 (***).

Figure 2

Principal coordinates analysis using Bray Curtis dissimilarity measure of the variation in (a) genus‐level and (b) family‐level composition of stems 2–10 cm diameter in undisturbed, control 20 × 20 m subplots of four Amazonian forests (north and south Peru, Central Amazonia (Manaus; CA) and the Guiana Shield (Nouragues; Gui). In each graph, the points represent sites, and the lines represent the ordination of genera and families. Key genera and families are labelled on each graph.

Figure 3

Mean and 95% confidence limits on percentage change in stem density, mean community‐level traits (WD – wood density, SM – Seed mass, LMA – leaf mass per unit area, Max Ht – maximum height and N – nitrogen concentration) and diversity (based on rarefied species richness) between disturbed and control subplots in four Amazonian forests.

Figure 4

Relationship between the effect of disturbance on the relative abundance of 121 genera across four Amazonian forests and the mean wood density of those genera. Wood density data from Zanne et al. (2009). Linear regression line (Change in relative abundance = 0.7078–1.1312*Genus wood density) also shown.

Figure 5

Difference in (a) species richness, (b) Shannon Index, (c) Fisher's alpha and (d) rarefied estimates of species richness between disturbed and control subplots in four Amazonian forests (north and south Peru, Central Amazonia (Manaus; CA) and the Guiana Shield (Nouragues; Gui). Positive values indicate higher values in the disturbed subplots. Significant differences for individual sites shown as P < 0.01 (**) and P < 0.001 (***).