Are the effects of an ecosystem engineer and environmental gradient on plant biodiversity independent, additive, or synergistic? Tests with the leaf-cutter ant Atta laevigata in a Neotropical savanna

Background. Species that transform habitats in ways that influence other species are known as Ecosystem Engineers. While the impacts of many engineers have been well described, how the impacts of engineers vary along or even alter environmental gradients remains limited. While complicated – the gradients themselves can be altered by engineers – doing so is necessary to advance models of ecosystem engineering. We used leaf-cutter ants (Atta spp.) as a model system with which to investigate the relative influence of gradients and environmental engineers on the abundance and species richness of woody plant seedlings. Methods. We conducted our research in South American savanna woodlands (Cerrado). We surveyed seedlings along a canopy cover gradient and collected data on environmental conditions at different distances from ant colonies and fit statistical models that addressed the following questions (1) Do the environmental conditions previously shown to influence Cerrado seedling establishment vary with canopy cover? (2) Does A. laevigata alter the canopy cover gradient and the related environmental conditions influencing seedling establishment? (3) If so, what is the spatial extent of Atta impact on the gradient and environmental conditions, and how does this vary along the canopy cover-cover gradient? (4) Do A. laevigata and canopy cover act independently or in synergy to influence seedling abundance and species richness? Results. Environmental conditions that influence seedling establishment varied with canopy cover, but ants are not modifying the canopy cover gradient or canopy cover around nests. However, ants are modifying other environmental factors, and the magnitude and spatial extent of these changes is consistent across the gradient. In contrast to prior studies, we found both seedling abundance and diversity were independent of canopy cover; ant-related factors (e.g., proximity to nests) had the strongest effect on both abundance and richness. Discussion. Atta laevigata in the Cerrado modify environmental conditions in ways that have strong but spatially restricted consequences for seedling communities. We hypothesize that by clearing litter and reducing soil moisture ants indirectly reduce seedling establishment. The alteration of soil nutrients could also reduce seedling growth and survivorship; if so these indirect negative effects of engineering could exacerbate the direct effects of their herbivory. The effects of Atta on environmental conditions and seedling communities appear restricted to the nest mound, but they could be long-lasting because Atta mounds persist long after a colony has died or migrated. These results underscore the hypothesis that leaf-cutter ants play a dominant role in Cerrado plant demography, and that the ecological and economic footprint of these engineers may increase dramatically in coming decades due to ongoing transformations of the Cerrado by human activities.