Climate models reveal two tropical bats facing sharply different futures

Deep in the canopy of Neotropical forests, two closely related fruit bats are quietly performing one of the tropics’ most essential ecological services: dispersing the seeds of countless native plants. Yet a new study published in the journal Web Ecology suggests that these two species, despite sharing a genus and much of their evolutionary history, may be headed toward very different destinies as the climate warms. Using ecological niche modeling, a research team led by Sergio Hernández-Rodríguez of the Universidad Autónoma del Estado de Morelos found that Vampyrodes major, a striped-faced bat ranging from southern Mexico to Colombia, is likely to retain nearly all of its climatically suitable habitat through mid-century. Its South American relative, Vampyrodes caraccioli, could lose up to half of the area where current conditions allow it to thrive, with the losses concentrated in the heart of the Amazon basin.

The findings carry weight well beyond the two bat species themselves. Frugivorous phyllostomid bats are keystone agents of forest regeneration, moving seeds across fragmented landscapes, pollinating chiropterophilous plants, redistributing nutrients through guano deposits, and maintaining genetic connectivity among plant populations. If V. caraccioli contracts across the Amazon, the study’s authors warn, the disruption could ripple through seed-dispersal networks and compromise the natural regeneration of one of the planet’s most biodiverse ecosystems. Both species are currently listed as Least Concern by the IUCN, but the researchers argue that this label, based largely on broad geographic ranges, may mask vulnerabilities that only become visible when climate data are brought into the picture.

The genus Vampyrodes has long been a taxonomic puzzle. Originally proposed as a subgenus in 1889 and later treated as a single species with two subspecies, it was only in 2011 that phylogenetic analysis of mitochondrial cytochrome b gene sequences, combined with a comprehensive review of cranial morphology, confirmed that V. major and V. caraccioli are distinct species. V. major occupies Central America from southern Mexico into Colombia, while V. caraccioli ranges across northern and eastern South America, dominating in the Amazon basin, the Andean foothills, and tropical southern Brazil. Because these bats depend on intact forest for both food and roosting sites, they have traditionally been treated as bioindicators of well-preserved forest, although recent records from restoring secondary vegetation hint that their tolerance may be broader than once assumed.

To build their models, the researchers compiled 190 verified occurrence records, 77 for V. major and 113 for V. caraccioli, drawn from the Global Biodiversity Information Facility and the published literature, deliberately excluding citizen-science observations because the cryptic morphology of these bats raises the risk of misidentification in non-vouchered records. The team filtered the data to remove duplicates and non-georeferenced points, then thinned the records spatially to a minimum separation of ten kilometers to reduce autocorrelation. Fifteen bioclimatic variables from the WorldClim 2.1 database were screened, with four excluded due to geographic discontinuities, and a Spearman correlation threshold of 0.75 narrowed the predictor set to six variables capturing annual mean temperature, temperature seasonality, temperature annual range, warmest-quarter temperature, annual precipitation, and wettest-month precipitation.

The modeling itself used the Maxent algorithm, implemented through the Wallace platform in R, with 120 candidate parameter combinations tested per species. The final model for V. major relied on linear features with a regularization multiplier of one, while V. caraccioli required a more complex linear-quadratic-hinge configuration with a multiplier of three. Model selection followed established statistical criteria, requiring omission rates below five percent and minimizing the corrected Akaike information criterion. Variable importance outputs revealed telling differences: temperature annual range dominated the V. major model at 67.9 percent permutation importance, whereas V. caraccioli’s suitability hinged primarily on annual precipitation at 37.2 percent, followed by temperature annual range and annual mean temperature. In other words, the Central American species is chiefly constrained by thermal variability, while its Amazonian relative is governed by moisture availability.

Niche comparisons in multivariate environmental space showed that the two species share broad climatic ground but are not identical. The first two principal components explained 75.7 percent of environmental variance, and the observed overlap was moderate, with Schoener’s D at 0.398 and the Hellinger-based I at 0.575. Randomization tests found no evidence that similarity exceeded what background environments would predict, and equivalency could not be strictly rejected, but the niche centroids were measurably displaced and the non-overlapping components were asymmetric. The authors interpret this pattern as partial niche conservatism: a shared evolutionary climatic background, consistent with the genus’s recent crown age of roughly two million years, superimposed with early differentiation along axes of temperature seasonality and dry-season moisture. A companion study published in 2026 reached a similar conclusion, linking seasonal climatic gradients to morphological divergence between the two species.

The future projections are where the story turns stark. The team ran an ensemble of five CMIP6 general circulation models under two Shared Socioeconomic Pathways for the period 2041 to 2060: SSP1-2.6, a sustainability-oriented scenario limiting warming to below two degrees, and SSP5-8.5, a fossil-fuel-intensive pathway with substantially higher warming. For V. major, roughly 98 percent of currently suitable area remained stable under both scenarios, with gains and losses each staying below two percent. V. caraccioli told a different story: suitable area shrank by about 44 percent under the low-emission scenario and 51 percent under the high-emission scenario, with the losses concentrated across the Amazon basin in Brazil, Colombia, Venezuela, Guyana, and Suriname. Multivariate environmental similarity surfaces confirmed that most projections fell within the range of training conditions, lending credibility to the forecasts, though the authors caution that extrapolation risk is not uniform across Amazonia.

The protected-area analysis added a sobering dimension. Under current conditions, about 25.5 percent of V. caraccioli’s climatically suitable habitat, roughly 91,217 square kilometers, falls within protected areas, mostly in the Amazon. By mid-century, that figure collapses: protected suitable area drops to about 48,517 square kilometers under SSP1-2.6 and 39,795 square kilometers under SSP5-8.5, meaning more than 46 percent of the species’ current protected habitat could lose its climatic suitability. V. major, by contrast, holds steady, with around 20.5 percent of its suitable area within protected zones across all scenarios. The message is clear: existing reserves, however crucial, may be insufficient to safeguard V. caraccioli if climate and land-use changes continue unabated, and static protected-area networks cannot track dynamically shifting climates.

The authors are careful to note the limitations of their approach. Macroclimatic niche models capture regional suitability but miss fine-scale determinants such as roost microclimates, cave and cavity buffering, riparian corridors, and topography-driven microrefugia that can decouple local conditions from broad climate layers. Occurrence records for V. caraccioli also remain sparse across large portions of Amazonia, meaning the projected contractions could be conservative if suitable climates exist in poorly sampled subregions. Bats’ strong attachment to home ranges and roosts may further limit their ability to track shifting suitability through dispersal, since long-distance migration is restricted to a subset of species. The team recommends pairing multi-season fieldwork on roost and foraging ecology with population genomics to identify locally adapted, climate-resilient populations, an approach shown in other systems to reduce predicted range losses.

The conservation prescriptions that emerge from the study are deliberately species-specific. For V. caraccioli, the priorities are strengthening protected-area networks, restoring degraded habitats, and designing ecological corridors that connect humid forest zones across the Amazon, allowing bats to move toward microclimatic refugia while preserving the seed-dispersal services on which forest regeneration depends. For V. major, climatic stability should not be mistaken for invulnerability: ongoing deforestation and agricultural expansion across Central America continue to fragment habitats and erode connectivity, so corridor networks and intact-forest preservation remain urgent there as well. Because suitable areas and potential refugia span national borders, the authors argue that coordinated transboundary planning will be essential. In a warming world where the Amazon has just recorded unprecedented drought and warmth, the fate of these unassuming seed dispersers may serve as an early warning for the mutualistic webs that hold tropical forests together.

Subject of Research: Climate-driven habitat suitability and conservation of Vampyrodes fruit bats in the Neotropics

Article Title: Habitat characterization and climate-driven niche shifts of Vampyrodes bats reveal contrasting futures for V. major and V. caraccioli

Article References: Hernández-Rodríguez, S., Martínez-Borrego, D., Jácome-Flores, M., & Cruz, D. D. (2026). Habitat characterization and climate-driven niche shifts of Vampyrodes bats reveal contrasting futures for V. major and V. caraccioli. Web Ecology, 26(2), 157-173. https://doi.org/10.5194/we-26-157-2026

Image Credits: AI Generated

DOI: 10.5194/we-26-157-2026

Keywords: Vampyrodes, fruit bats, ecological niche modeling, climate change, seed dispersal, Amazon basin, Maxent, protected areas, Neotropics, habitat loss, SSP scenarios, conservation corridors