The role of forest composition on pool-breeding amphibians: colonization, larval communities, and connectivity
Michael P. Graziano, PhD
Advisor: Stephen N. Matthews
Few studies investigate the intricate effects that the plant community has on amphibian populations. Plants shape ecosystems, affecting both physical and chemical attributes of the landscape. Conspicuous artifacts of the plant community include canopy cover, physical structure, and modified temperature and moisture profiles. However, less conspicuous artifacts of the tree community, namely the physiochemical characteristics of their resulting leaf litter, have the ability to shape their ecosystem just as greatly as their more conspicuous traits. Leaf litter represents the primary energy source in vernal pools and other aquatic systems that are critical amphibian breeding habitat. As plant communities shift across the landscape due to ecosystem degradation, invasion by nonnative species, climate change, and shifting disturbance regimes, there is a critical need to investigate how these potential changes can influence the amphibian community and the mechanisms by which they occur. My research investigates how the vernal pool breeding amphibian community responds to differing plant communities across a heterogeneous forested landscape and throughout their life cycle. As such, my overall objectives are multi-faceted: (1) to determine if the tree community impacts colonization and use of vernal pool-breeding amphibians (2) to scale up mesocosm studies that document the strong regulatory response tree litter inputs can have on growth and development of amphibian larvae to a field setting (3) to determine if small, constructed ridge-top pools are a viable option for enhancing amphibian populations in the landscape, particularly with regards to increasing functional connectivity and maintaining diverse amphibian communities, and (4) to establish a landscape-level study design for conducting future, field-based experiments that can serve as a baseline to document changes in forest ecosystems. These objectives are addressed within each of my primary research pursuits below. The first component of my research investigated the underlying drivers of colonization of novel breeding sites by amphibians, with a focus on the surrounding forest community. Fourteen ponds were created of similar size and depth along a gradient of different tree communities ranging from oak-dominance to maple-dominance Vinton Furnace State Experimental Forest in southeast Ohio in June 2014. In 2015 and 2016 we documented colonizing amphibians, and within 10 months of construction, pool-breeding amphibians had colonized all pools. 1,114 captures were recorded comprising 12 species and 817 unique captures. Generalized linear model-based analyses performed at the community level consistently found that the tree community is significant predictor of amphibians that colonize isolated woodland pools. Additional predictors of the community included year (age), initial pH, percent coverage of leaf litter, and canopy openness of pools. As plant communities shift within the landscape due to invasion by non-native species, climate change, and an alteration of disturbance regimes, amphibian communities may also shift in response. The second component of my research focused on the larval community occupying these constructed pools. The vast majority of studies investigating litter effects utilize mesocosms, and I sought to determine if these effects carry over into the field. We modeled the growth rate, community structure, and size/mass at metamorphosis of four amphibian species to naturally colonize the pools: American toads (Anaxyrus americanus), chorus frogs (Pseudacris crucifer + P. brachyphona), wood frog (Lithobates sylvaticus), and Jefferson salamanders (Ambystoma jeffersonianum). The percentage of allochthonous input derived from oak and hickory trees was a significant predictor in all models for all species, though the effect varied by species. There were also significant effects from year, canopy openness, density, and dissolved oxygen. Our study supports findings from mesocosm experiments suggesting an effect of leaf litter input on amphibian larvae and lends further evidence that shifting plant communities may carry over into the amphibian community. The third component of my research addressed aspects of connectivity and dispersal of pool-breeding amphibians, by examining the degree of breeding site connectivity in a managed forest and the degree to which that effected subsequent colonization by amphibians in constructed ridge-top pools. Most studies on dispersal and connectivity of pool-breeding amphibians focus on fragmented landscapes, and while this research remains important, the ability of species to dispersal through relatively unfragmented landscapes should not be overlooked, as it is these landscapes (i.e. national forests, parks) that harbor some of the most robust amphibian populations. We utilized abundance data of species colonizing constructed ridge-top pools to model landscape connectivity based on three metrics: distance to natural wetlands, slope to natural wetlands, and area of natural wetlands. We found that Jefferson and spotted salamander (Ambystoma jeffersonianum and A. maculatum, respectively) and eastern newt (Notophthalmus viridescens) connectivity was largely dictated by the distance and mean slope to wetlands. Anurans were difficult to accurately assess and were more diverse in their responses than salamanders. My results show connectivity is species specific, and is largely an artifact of distance and slope to existing wetlands in relatively unfragmented landscapes. Further, this shows that connectivity of pool breeding amphibian populations may be enhanced by strategically placing breeding sites on the landscape that capitalize on dispersing individuals. Collectively my results show that the tree community impacts pool-breeding amphibians from the larval stage through maturity, and that wide scale shifts in the plant community may alter the amphibian community as well. Perhaps more importantly, we used a field setting to confirm that results from mesocosms can carry over into the field, highlighting important underlying mechanisms, but also demonstrating additional abiotic determinates that modulate the magnitude of these effects. Further, this research provides insight to patterns and factors that structure initial colonization of novel breeding habitats by pool-breeding amphibians. From an application and management perspective, these results support the construction of small, ridge-top pools as a means of enhancing or buffering local amphibian populations. Additionally, these pools can be quickly and easily constructed in managed forests, potentially enhancing breeding and non-breeding habitat heterogeneity for the amphibian community.