A meta‐analysis of the effects of fragmentation on the megadiverse herpetofauna of Brazil

Abstract

Habitat loss and fragmentation are the primary causes of global population decline of amphibians and reptiles. In Brazil, that hosts an extraordinary herpetofaunal richness, amphibians and reptiles are clearly undersampled among vertebrate groups in fragmentation research. This bias may underestimate the effects of fragmentation on herpetofauna in this megadiverse country. Here, we conducted an exhaustive literature review to evaluate the effects and patterns of fragmentation on amphibians and reptiles in Brazil. We analyzed 55 papers between 1994 and 2020, comprising 350 cases of the effect (positive, negative, or neutral) of a given fragmentation metric on a particular biological response. Forest biomes (Amazon and Atlantic Forest) were largely overrepresented in relation to non-forest biomes, comprising 82% of studies. We also found a disproportional prevalence of fragmentation articles on amphibians (75%). Among lower taxonomic groups, Anura and Testudines were significantly overrepresented, whereas Caudata and Crocodylia were neglected. Fragment size reduction, habitat degradation, habitat loss, and matrix contrast were the most studied metrics (70% of studies), while single-species abundance was the most considered response (59%). The effects of fragmentation were not statistically different between amphibians and reptiles. In general, the impacts were predominantly negative or neutral, while positive effects were infrequent or even non-existent. Our findings suggest that amphibians and reptiles respond similarly and may not be as vulnerable to fragmentation. We call future research to consider non-forest biomes and less-studied taxonomic groups (e.g., Squamata) to fully understand the effects of fragmentation on the megadiverse Brazilian herpetofauna and to take well-informed conservation actions.

Fire and water: global fire impacts on physicochemical properties of freshwater ecosystems

“…We analysed signi cant departures from expected frequencies with the Fisher exact test, performed with R software (R Development Core Team 2020). Contingency tables allow the inclusion of the simplest and standard response that can be extracted from the articles (altered -increase, decreaseand unaltered -neutral-; see Teixido et al 2021). Thus, we included as many studies as possible in our assessment, and the many different ways in which results were reported, by testing for signi cant differences in the observed frequency of the responses of water quality parameters to re.…”