COS 289-3 - Host–pathogen interactions in hot ponds: Linking local temperature variation to landscape synchrony of chytrid infection in amphibians

Abstract: By regulating a suite of biological processes, local temperature is a key determinant of patch-level host–pathogen interactions. In systems with such temperature dependence, variability in temperature characteristics among patches across a landscape may also alter broader patterns of transmission and disease. For instance, thermally heterogeneous landscapes may be characterized by lower patch-level transmission synchrony and thus lower landscape-level peaks of infection. Yet, few studies have dissected the interplay between local temperature, patch heterogeneity, and landscape-level host–pathogen synchrony, hindering our understanding of how local conditions scale to alter broader disease risk. One system well-suited to studying such processes is infection of pond-breeding amphibians by Batrachochytrium dendrobatidis (Bd), a fungal pathogen that infects amphibian skin. Despite extensive work linking Bd dynamics and temperature, many questions remain regarding how pond-level attributes affect landscape-level processes. Thus, we tested support for the nested hypotheses that (1) pond-level temperature characteristics drive local patterns of Bd and (2) temperature variability among ponds affects the synchrony in infection across a landscape. To do so, we generated a dataset of in-situ temperature, amphibian community characteristics, and Bd dynamics across a landscape of 23 ponds in southeastern Wisconsin, USA. Specifically, from late-winter to mid-summer, we deployed loggers to record hourly temperature, conducted repeated (3×) net and trap surveys of amphibian communities, and swabbed 815 individual amphibians for Bd. Following field-based data collection, we analyzed swabs for Bd presence and load using standard qPCR methods. We analyzed these data using a series of generalized linear mixed models, first treating pond-level Bd prevalence and load as outcomes before considering among-pond synchrony as the focal response. At both stages, we included amphibian density and diversity as covariates and specified site-level random effects. In support of previous work, we found that pond-scale temperature leading up to a sampling event was associated with Bd prevalence among amphibians and that both the variability and predictability of temperature fluctuations were key metrics. We also found that variability in temperature characteristics among ponds was associated with divergent Bd dynamics across the landscape, providing novel evidence that spatial heterogeneity of site-level conditions can alter broader patterns of Bd infection in amphibians. Critically, in this system and others, asynchrony in transmission among patches within a landscape may enable a rescue effect by creating low-infection ‘refuge’ sites that act as population sources from which uninfected hosts can emigrate following die-offs.