Predation on artificial caterpillars following understorey fires in human-modified Amazonian forests

Tropical forests are facing several impacts from anthropogenic disturbances, climate change and extreme climate events, with potentially severe consequences for ecological functions, such as predation on folivorous invertebrates. Folivory has a major influence on tropical forests by affecting plant fitness and overall seedling performance. However, we do not know whether predation of folivorous arthropods by birds, mammals, reptiles and other arthropods is affected by anthropogenic disturbances such as selective logging and forest fires. We investigated the impacts of both pre-El Niño human disturbances and the 2015-2016 El Niño understorey fires on the predation of 4,500 artificial caterpillars across 30 Amazonian forest plots. Plots were distributed in four pre-El Niño forest classes: undisturbed, logged, logged-and-burned and secondary forests, of which 14 burned in 2015-16. We found a higher predation incidence in forests that burned during the El Niño in comparison to unburned ones. Moreover, logged-and-burned forests that burned again in 2015-16 were found to have significantly higher predation incidence by vertebrates than other forest classes. However, overall predation incidence in pre-El Niño forest disturbance classes was similar to undisturbed forests. Arthropods were the dominant predators of artificial caterpillars, accounting for 91.5% of total predation attempts. Our results highlight the resilience of predation incidence in human-modified forests, although the mechanisms underpinning this resilience remain unclear.

Methods
Predation experiments were conducted between April 10 and June 15 2019, corresponding to the middle-end of the wet season. All artificial caterpillars were green, made from odourless non-toxic coloured plasticine (Lewis NewplastTM) mixing an equal portion of dark and light green to create medium green coloured caterpillars. Artificial caterpillars were crafted to mimic cryptic Lepidoptera larvae. We fixed the caterpillars to the vegetation by passing a wire longitudinally through them. When placing caterpillars in the field, we removed all handling marks to avoid misidentifying them with predator’s marks. 

We established five 20 × 5 m sub-plots within each of the 30 forest plots. We placed 30 caterpillars in each sub-plot, totalling 150 caterpillars per forest and 4,500 caterpillars across the experiment. All caterpillars were installed on understorey vegetation branches, between 1.0 – 4.5 m from the ground, separated by at least 1 m from each other. We evaluated predation in the understorey due to the difficulties of placing caterpillars in the tall canopy (i.e., 30-50 m). After 14 days of field exposure, a single observer (LCR) recovered all caterpillars and assessed whether they had been predated, attributing predation marks to different predator groups. The identification of predation marks on recovered caterpillars was based on the beak, mandibular tooth or radula marks. The predation marks were attributed to one of four predator groups: arthropods, birds, reptiles and mammals. Multiple marks made by the same type of predator on a single caterpillar were categorized as a single predation event.