Structural refuges and connectivity alter herbivory, prey abundance and impacts of top-predators in mangrove canopies

Background/Question/Methods

The physical structure of habitats can have important consequences for food-web and community dynamics. Refuges can provide enemy-free space can increase prey density and alter community composition. Additionally, features that facilitate dispersal of prey among patches can subsidize patches that are more exposed to predation and alter prey diversity within patches. In Caribbean mangrove systems, birds are often top-predators of the arboreal community, preying on arthropods belonging to a variety of functional groups. This predation can suppress arthropod populations and reduce herbivory on individual mangrove trees. Natural mangrove stands contain trees that vary in the number of branches connecting them to neighboring trees and the abundance of shelters such as insect-made galleries and dead wood. In order to test whether structural refuges and canopy connectivity alter the top-down suppression of arboreal arthropods, we manipulated 48 small (approx. 1m tall) red mangrove trees (Rhizophora mangle) in the interior of an island off the coast of Belize. The trees were not contiguous with other vegetation and were always surrounded by water. To manipulate dispersal we connected experimental trees to nearby stands using rope, to provide shelters we attached 10 empty Lepidopteran cocoons,  and we excluded top predators using PVC frames covered in propylene netting. The presence vs. absence of rope, shelters, and top predators were crossed in a full factorial design.

Results/Conclusions

Overall, refuges increased the abundance of arthropods (excluding ants) and herbivory when top-predators were present, but not when top-predators were excluded. This result suggests that natural shelters alter arthropod abundance and plant-herbivore interactions primarily by mitigating impacts of bird predation. Connecting trees to neighboring stands increased herbivory, but did not change the abundance of arthropods. Experimental treatments varied in their impacts on particular trophic and taxanomic groups. Ants (62% of arthropods) were always more abundant on trees with shelters, and were not affected by the presence or absence of bird exclosures. Predatory arthropods (mostly spiders) were more common when shelters were provided even in the absence of birds, likely because the shelters were utilized for guarding eggs. Tree crabs were more abundant on trees with rope connections, indicating that they prefer non-aquatic dispersal. Herbivores demonstrated the same trends as the aggregate of all groups. Correlations among arthropod groups were generally weak, except for the negative correlations between ants and both predators and detritivores, which may indicate possible competiton for space or direct antagonism between these groups.