Amphibians are among the most threatened vertebrate groups worldwide, with population declines driven by climate change, habitat degradation and the spread of emerging infectious diseases. Bd, the causative agent of chytridiomycosis, was first described in 1999 as the agent behind massive amphibian die-offs, but its presence in Mexico dates back to the 1970s. The pathogen is now widespread, particularly in the central and southern regions of the country, and has been reported to infect at least 50 amphibian species in Mexico. Hylidae and Plethodontidae are among the most affected families, with the genera Hyla, Ambystoma and Pseudoeurycea ranking as the most impacted, their species concentrated across central and southern Mexico.

Mexico is a global hotspot of amphibian diversity and endemism, hosting approximately 3.69 percent of the world’s anuran species. The country’s most diverse amphibian family, Plethodontidae, comprises 140 species that primarily inhabit high-elevation forests in the Trans-Mexican Volcanic Belt, the Sierra Madre Oriental and the montane regions of the Sierra Madre del Sur, the Balsas Depression and Oaxaca. Hylidae is the second most diverse family with approximately 101 species, while Ambystomatidae, though lower in species richness with 12 to 18 species depending on the taxonomic framework, includes 16 endemic species, making Mexico the centre of diversity for the group.

The research team selected 13 amphibian species based on endemism, conservation status and ecological relevance within the montane regions of central Mexico. The selection included three Hyla species (H. euphorbiacea, H. eximia and H. plicata), three Ambystoma species (A. altamirani, A. lermaense and A. velasci) and seven Pseudoeurycea species (P. altamontana, P. gadovii, P. juarezi, P. lineola, P. longicauda, P. smithi and P. werleri). All are endemic to Mexico, and most inhabit high-elevation cloud forests or temperate montane ecosystems that are particularly susceptible to climate change, habitat loss and emerging diseases. The species span a broad spectrum of IUCN Red List categories: four are listed as Least Concern, one as Vulnerable, seven as Endangered and one, P. smithi, as Critically Endangered.

Notably, most Pseudoeurycea species included in the study are extremely data deficient. For several of them, only presence records are available, with almost no information on their population sizes, ecology, reproductive behaviour or physiological tolerances. This scarcity of data results in their exclusion from many broad-scale conservation initiatives, despite the likelihood that they are among the most vulnerable to environmental disturbance and chronic infections. By including this taxonomically and ecologically diverse group, ranging from broadly distributed species to microendemics with only a handful of records, the study aimed to provide a more integrative evaluation of the potential impacts of Bd across a gradient of conservation statuses and research gaps.

Methodologically, the researchers compiled occurrence records from online databases including GBIF, iNaturalist and CONABIO, while Bd records were obtained from a previous study and its associated Dryad data repository, covering the period from 2000 to 2022. Records were filtered to remove coordinate errors, duplicates and points outside known geographic ranges, then spatially rarefied to retain only one occurrence per square kilometre, an approach shown to reduce model overfitting. Nineteen bioclimatic variables at roughly one kilometre resolution were obtained from WorldClim, along with elevation and land-cover variables derived from satellite imagery. A Pearson correlation analysis eliminated highly correlated variables, retaining seven final land-cover layers: percentage cover of agriculture, arid vegetation, grassland, fir forest, pine forest, oak forest and cloud forest.

Two distinct modelling methodologies were employed to facilitate comparative analysis. The first used the maximum entropy algorithm implemented through the ENMeval package in R, testing regularisation multipliers and feature class combinations, with 100,000 background points randomly generated to characterise the environmental background. A block method partitioned occurrence data geographically based on latitude and longitude, ensuring that training and evaluation datasets were separated across space. The second approach constructed an ensemble model with the biomod2 package, combining ten algorithms including generalised linear models, random forest, Maxent and XGBOOST, each run ten times with different selections of training and testing datasets. Model performance was assessed using the area under the curve of receiver operating characteristic plots, partial ROC analyses and the true skill statistic.

The results showed that the maximum temperature in the warmest month, the minimum temperature in the coldest month and precipitation in the wettest and driest months were the most important environmental variables for the amphibian species. For most Ambystoma species, the optimal habitat suitability fell between roughly minus two and ten degrees Celsius for the coldest month, while A. lermaense presented an optimum temperature range of 15 to 25 degrees. For Bd, the maximum temperature in the warmest month was positive within a range of approximately 15 to 35 degrees, and the minimum temperature of the coldest month was a negative variable. The optimal temperature range for Bd was found to lie between 10 and 25 degrees, aligning with observations from European studies. Elevation was a positive variable for all species, and pine, oak and cloud forest percentages were positive for nearly all species.

According to the tenth percentile training presence threshold, Bd had the greatest potential distribution among all species studied, covering 267,557 square kilometres, while the most restricted species, P. altamontana, covered just 9,982 square kilometres. Niche overlap analysis revealed that Bd and A. altamirani showed relatively higher overlap values (Schoener’s D = 0.423; Hellinger’s I = 0.572), suggesting a greater degree of environmental similarity compared to other species pairs. In contrast, A. lermaense exhibited the lowest overlap with Bd (D = 0.038; I = 0.178). Barplot analysis indicated that P. juarezi, P. longicauda and P. altamontana exhibited the highest percentage of spatial overlap with Bd, all exceeding 70 percent of their distribution ranges, while A. velasci and H. eximia had the lowest overlap, below 40 percent.

By integrating spatial exposure information with IUCN Red List categories, the researchers developed a composite vulnerability index that ranked P. altamontana, P. longicauda and P. juarezi as the most vulnerable species. A principal component analysis of vulnerability patterns revealed distinct species clusters, with the first two components accounting for 92.3 and 7.7 percent of the variance respectively. Species richness maps showed that amphibian diversity peaks in the central highlands of Mexico, particularly in regions overlapping with the Trans-Mexican Volcanic Belt, where richness values reach up to nine species per pixel, and the overlay analysis demonstrated that Bd suitability was notably high in these same regions. The authors emphasise the urgency of implementing targeted conservation actions, including long-term monitoring, early detection of infections, identification of resistant or tolerant populations and regulation of ecotourism in vulnerable regions. They also highlight the potential importance of range break zones, such as Oaxaca and the Altiplano Sur, as natural barriers that could limit Bd spread, and note that community-based conservation initiatives and voluntary reserves managed by indigenous people offer an opportunity to expand protected areas and implement habitat restoration programmes.

Subject of Research: Ecological niche overlap between the chytrid fungus Batrachochytrium dendrobatidis and endemic amphibians in Mexico's Central-Southern Highlands

Article Title: Mapping Risk Zones: niche overlap between Batrachochytrium dendrobatidis and endemic amphibians in Mexico’s Central-Southern Highlands

Article References: Heredia-Bobadilla, R. L., Manjarrez, J., Gómez-Benítez, A., Reyes-Velázquez, E. A., Carmen-Cristobal, J. M., Sánchez-Lara, A. J., Martínez-Alva, G., Bolom-Huet, R., & Sunny, A. (2026). Mapping Risk Zones: niche overlap between Batrachochytrium dendrobatidis and endemic amphibians in Mexico’s Central-Southern Highlands. Nature Conservation, 63, 299-335. https://doi.org/10.3897/natureconservation.63.163964

Image Credits: AI Generated

DOI: 10.3897/natureconservation.63.163964

Keywords: Batrachochytrium dendrobatidis, chytridiomycosis, amphibian conservation, ecological niche modelling, Mexico, Trans-Mexican Volcanic Belt, Sierra Madre del Sur, Pseudoeurycea, Ambystoma, Hyla, niche overlap, biodiversity hotspots