Species distribution models of rare tree species as an evaluation tool for synergistic human impacts in the Amazon rainforest

Tamilis R. Silva, Marcelo B. Medeiros, Sérgio E. Noronha, José Roberto R. Pinto
2017 Revista Brasileira de Botânica  
In the present work, we have considered the vulnerability of some rare tree species to human disturbances in a high biodiversity tropical region. In this context, we aimed to evaluate the combined effect of deforestation and artificial flooding of the large Jirau hydroelectric reservoir on potential distribution areas of 13 locally rare tree species in the southwestern Brazilian Amazon. We performed species distribution modeling (SDM) by using the environmental distance algorithm. Based on
more » ... ithm. Based on these models, we found new sites and subsequently applied rapid ecological assessment to collect further species occurrence data. Additional SDMs were carried out using MaxEnt to determine the potential distribution areas of these rare species. We found that artificial flooding and deforestation caused combined losses of potential distribution areas of rare tree species between 8 and 39% of the total area. The most vulnerable species were Semaphyllanthe megistocaula (K. Krause) L. Andersson (Rubiaceae) (39%), Chrysophyllum colombianum (Aubrév.) T.D. Penn. (Sapotaceae) (34%), Lacunaria jenmanii (Oliv.) Ducke (Quiinaceae) (32%), Brosimum parinarioides Ducke (Moraceae) (32%) and Xylopia benthamii R.E. Fr. (Annonaceae) (30%). These results indicate an additive effect of human disturbances such that artificial flooding, when combined with deforestation, has an overall effect by orders of magnitude. SDMs can be effectively used as a predictive tool in the assessment of human impacts on rare tree species in tropical forests. The results also showed different vulnerability among the rare species, and these results may indicate that some species are more seriously threatened by the extreme loss of potential distribution areas.
doi:10.1007/s40415-017-0413-0 fatcat:cyyl7c52mfhdbdnbdhqs3q3yvm