Journal Article

The amphibian-killing fungus in a biodiversity hotspot: identifying and validating high-risk areas and rfugia,The amphibian‐killing fungus in a biodiversity hotspot: identifying and validating high‐risk areas and refugia

Valenzuela‐Sánchez, Andrés, Bacigalupe, Leonardo D., Estay, Sergio A., Soto‐Azat, Claudio, Vásquez, Inao A., Alvarado‐Rybak, Mario, Cunningham, Andrew A., Peñafiel‐Ricaurte, Alexandra

2019

Overview

Amphibian chytridiomycosis, due to infection with the fungus Batrachochytrium dendrobatidis ( Bd ), has been associated with the alarming decline and extinction crisis of amphibians worldwide. It is essential for conservation management to identify regions with high or low suitability for Bd . We use a species distribution model to estimate the environmental suitability of Bd in the Chilean Winter Rainfall–Valdivian Forest biodiversity hotspot. Fourteen environmental variables were used as predictors in the statistical modeling (Maxent, generalized linear models, random forest) which also included 56 independent Bd + localities. High‐risk areas (i.e., suitability above a defined threshold) were validated through prospective field surveys conducted in 2017. As results from Maxent, which only uses presence data, were the only results retained, refugia (i.e., suitability below a defined threshold) were validated with the Bd absences (N = 12) used in the GLM and RF modeling. Our results showed that (1) suitability for Bd increased with human footprint and with shorter distances to urban centers and water bodies and decreased with elevation; (2) climate was not a major factor shaping the current distribution of Bd ; and (3) the model predicted high‐risk and refugia areas fairly well. Surveys of 24 new localities in high‐risk areas confirmed that 23 were Bd +; hence, these areas warrant consideration for long‐term Bd surveillance, population monitoring, and disease mitigation. In addition, five localities with apparent Bd absence were found in the predicted high‐risk areas. Our models showed that refugia can exist near high‐risk areas and Bd + sites. Four localities with apparent Bd absence were located within the refugia predicted by the model. Preventing Bd transmission to such refugia is of paramount importance for persistence of Bd ‐susceptible amphibian populations. The identification and validation through prospective field surveys of high‐risk areas and refugia are imperative to develop strategies to prevent further arrival and establishment of Bd and also, by identifying amphibian species or populations of conservation concern in such areas, will help to guide specific actions to reduce the biodiversity loss caused by chytridiomycosis.