PS 34-14 - Threshold responses of Amazonian stream fish diversity to deforestation

Friday, August 12, 2016
ESA Exhibit Hall, Ft Lauderdale Convention Center
Gabriel L. Brejão1, David J. Hoeinghaus2, Lilian Casatti1, María A. Pérez-Mayorga1 and Silvio F. B. Ferraz3, (1)Department of Zoology and Botany, Sao Paulo State University, São José do Rio Preto, Brazil, (2)Department of Biological Sciences, University of North Texas, Denton, TX, (3)Forestry Sciences, University of São Paulo, Piracicaba, Brazil
Background/Question/Methods

Deforestation is a primary driver of habitat loss and fragmentation affecting terrestrial and aquatic ecosystems and contributes to widespread biodiversity change. Biodiversity of streams draining impacted watersheds may not respond to deforestation in a simple linear relationship in regards to the proportion of the watershed affected or when deforestation occurred. The primary effect by which deforestation influences stream diversity, i.e. habitat simplification and homogenization, does not occur immediately after a deforestation event and populations may also not respond immediately to habitat changes. Threshold responses of stream biodiversity may therefore be expected with regard to degree of and time since deforestation. In this work, we test for threshold responses of taxonomic and functional structure of stream fish assemblages to deforestation and time for 75 streams along a deforestation gradient in the western Brazilian Amazon. Deforestation was assessed using remote sensing for multiple periods between 1984 and 2011. Ten ecomorphological traits related to habitat use, resource capture and locomotion were used to classify 84 fish species into 20 functional groups. Threshold responses at the species/functional group level and at the community level were quantified using TITAN for watershed and riparian zone deforestation.

Results/Conclusions

Significant threshold responses were observed for many species and functional groups along gradients of deforestation and time. Negative threshold responses were more common, and 12 species exhibited negative responses to both watershed deforestation and time since deforestation. Most negative species threshold responses occurred at very low levels of deforestation (e.g. <20%) and soon after impact (e.g. <10 years), whereas positive threshold responses generally occurred at >70% deforestation and >10 years after the impact. Five of the 20 functional groups presented significant negative threshold responses both to deforestation and time since change. Similar to species-level analyses, negative threshold responses occurred at low levels of deforestation and relatively soon after the impact. There were no positive threshold responses for functional groups. Thus, negative thresholds were for species with specific suites of functional traits, whereas species that exhibited positive threshold responses did not have consistent traits and were not functionally unique. Our findings demonstrate that degree of deforestation and time since impact can elicit non-linear responses in tropical stream fish assemblages. Because most negative responses occurred at low levels of deforestation, even minimal landcover change is expected to negatively affect biodiversity, and these responses will likely occur relatively soon after impact.