Large wild herbivores slow down the rapid decline of plant diversity in a tropical forest biodiversity hotspot

Herbivory limits success of vegetation restoration globally

Restoring vegetation in degraded ecosystems is an increasingly common practice for promoting biodiversity and ecological function, but successful implementation is hampered by an incomplete understanding of the processes that limit restoration success. By synthesizing terrestrial and aquatic studies globally (2594 experimental tests from 610 articles), we reveal substantial herbivore control of vegetation under restoration. Herbivores at restoration sites reduced vegetation abundance more strongly (by 89%, on average) than those at relatively undegraded sites and suppressed, rather than fostered, plant diversity. These effects were particularly pronounced in regions with higher temperatures and lower precipitation. Excluding targeted herbivores temporarily or introducing their predators improved restoration by magnitudes similar to or greater than those achieved by managing plant competition or facilitation. Thus, managing herbivory is a promising strategy for enhancing vegetation restoration efforts.

Frugivore Population Biomass, but Not Density, Affect Seed Dispersal Interactions in a Hyper-Diverse Frugivory Network

Mutualistic interactions are regulated by plant and animal traits, including animal body size and population density. In seed dispersal networks, frugivore body size determines the interaction outcome, and species population density determines interaction probability through encounter rates. To date, most studies examining the relative role of body size and population density in seed dispersal networks have examined animal guilds encompassing a narrow range of body sizes (e.g., birds only). Given non-random, body-size dependent defaunation, understanding the relative role of these traits is important to predict and, ideally, mitigate the effects of defaunation. We analyzed a hyper-diverse seed dispersal network composed of birds and mammals that cover a wide range of body sizes and population densities in the Brazilian Pantanal. Animal density per se did not significantly explain interaction patterns. Instead, population biomass, which represents the combination of body size and population density, was the most important predictor for most interaction network metrics. Population biomass was strongly correlated with body size, but not with density. Thus, larger frugivore species dispersed more plant species and were involved in more unique pairwise interactions than smaller species. Moreover, species with larger population biomass had the strongest influence (i.e., as indicated by measures of centrality) on other species in the network and were more generalist, interacting with a broader set of species, compared to species with lower population biomass. We posit that the increased abundance of small-sized frugivores resulting from the pervasive defaunation of large vertebrates would not compensate for the loss-of-function of the latter and the inherent disruption of seed dispersal networks.