Patterns and underlying mechanisms of amphibian response to tropical deforestation

A central goal of ecology is to understand how environmental factors influence species richness and distributions.  Deforestation causes increased isolation of intact forest within a matrix habitat that is often considered unsuitable for many plants and animals.  This results in both the extirpation of populations due to habitat loss and redistribution of populations due to modification of remaining habitat.  Many studies have examined the patterns of amphibian response to habitat modification, but little progress has been made in understanding the mechanisms underlying species-specific responses.  Species that occupy modified habitats should share characteristics that promote survival in the novel habitat, whereas species that are extirpated from novel habitats should share characteristics that render them vulnerable.  My project examines the spatial patterns and underlying mechanisms of how anurans respond to habitat modification in the Osa Peninsula, Costa Rica.  I accomplish these goals in the following projects:  

I.  Spatial patterns of frog occupancy within a mosaic of forest and pasture

A diurnal survey along a forested stream in the Osa Peninsula

I am using visual encounter and vocalization surveys to determine frog species richness, composition, and abundance in aquatic and terrestrial sites within intact forest and pasture habitats.  I will use a geographic information system to examine relationships between these measures of species diversity and local and landscape variables.  My approach will determine species occupancy in pastures and forests and then examine ecological and life-history traits that differ among species occupying either habitat.

A diurnal survey along a pasture stream

II. Frog use of breeding pools across forest-pasture edges in Costa Rica

Little is understood concerning how frogs use edges between intact forest and matrix habitats.  Traditional visual encounter survey methods of adult frogs may be misleading because of the difficulty of distinguishing the observation of an adult from whether a viable population is present in a habitat.  In this study, I examine how species use artificial pools for breeding activity within 50 m of a pasture-forest edge.  Specifically I ask whether species demonstrate differential use of artificial breeding pools at different distances from an edge into pasture and forest and whether time to pool colonization differs depending on distance from the edge. A pasture-forest edge in the Osa Peninsula, Costa Rica

III. Mechanisms underlying adult occupancy patterns: tadpole performance in pasture, edge, and forest sites

            Although many studies have documented differences in amphibian diversity within intact and modified habitats, the mechanisms underlying occupancy patterns are little understood.  Here, I examine whether the performance of tadpoles across habitat types matches patterns of adult occupancy in the same habitats.  I quantified tadpole performance of two species with contrasting life histories and habitat preferences (Physalaemus pustulosus and Dendrobates auratus), in three habitats (forest, edge, and pasture) in the Osa Peninsula, Costa Rica.  Tadpoles were reared in mesocosms in a randomized block design.  To describe each habitat, I measured water temperature and the global site factor, a quantification of the light environment, from hemispherical photographs.  Habitat treatments affected the survival, growth, and development of each species differently.  D. auratus survival was low in pasture (< 5%) but higher in edge and forest (> 18%); in contrast, P. pustulosus survival was equally high in all three habitat treatments (>82%).  D. auratus metamorphs were largest in edges, while P. pustulosus metamorphs were largest in pastures.  Larval period did not differ among treatments for D. auratus but was shortest in pasture for P. pustulosus.  Maximum daily water temperature was at least 8°C higher, and daily variation in temperature was greater in pasture than in edge and forest treatments.  The results suggest that P. pustulosus and D. auratus tadpoles respond differently to abiotic variables within habitats.  Patterns of tadpole performance among habitats reflect those of adult occupancy, with two exceptions.  D. auratus adults are not found in pastures, likely because of physiological limitations on diurnal activity in pastures.  P. pustulosus adults are rarely found in forest, perhaps because edge and pasture habitats offer benefits of faster growth and development.  These findings offer a more detailed understanding of the mechanisms underlying anuran occupancy patterns within neotropical forests and pastures.    

Basins inside a forest fence

Looking for metamorphs

Basins inside a pasture fence

IV.  Heterospecific carnivory and cannibalism among tadpoles in ephemeral pools in tropical pastures

The role of tadpoles in aquatic systems is gradually being redefined, recognizing their potential importance as carnivores and scavengers rather than strict herbivores and detritivores.  In the neotropics, it is not well known whether anuran assemblages in certain habitats are particularly prone to cannibalism and carnivory and the factors that influence the frequency of these feeding modes are poorly understood.  I conducted three experiments to address the following questions: 1) How prevalent is carnivory among tadpoles found in ephemeral pasture pools?, 2) Is the frequency of cannibalism influenced by the size and developmental stage of predators and prey?, and 3) Do density, food availability, and relatedness influence the frequency of cannibalism and performance of tadpoles?  Field surveys and experimental manipulations were conducted in the Osa Peninsula, Costa Rica.  Facultative heterospecific carnivory and cannibalism are common foraging modes among species that reside in ephemeral pasture pools.  Cannibalism in Smilisca phaeota was dependent on the size and stage of predators and prey, with medium and large tadpoles causing almost complete mortality of eggs, but less mortality of mobile hatchlings.  In contrast, small tadpoles caused less mortality of prey at all developmental stages.  S. phaeota cannibalism occurred most often among tadpoles in treatments with low availability of vegetative food.  S. phaeota tadpole performance (survival, metamorph mass, and time to metamorphosis) was influenced by tadpole density and the amount of vegetative food; however, diet supplementation with conspecific eggs and tadpole relatedness had few effects on performance.  These results are consistent with previous work demonstrating the large role density and food limitation have on the frequency of tadpole cannibalism.  Further, the high number of species that exhibit tadpole carnivory in pastures suggests that this feeding mode may be advantageous to tadpoles in these ephemeral pools.
Smilisca phaeota	Cup experiment to examine how predator size and prey stage influence the incidence of cannibalism	Physalaemus pustulosus