Statistical analyses
To evaluate the number of taxa/OTUs present in the study area, a
multimodel approach was implemented to fit asymptotes on the species and
OTU accumulation curves. This approach considered 5 different
accumulation models (Lomolino, Michaelis-Menten, Gompertz, asymptotic
regression and logistic curve) and weighted them using the Akaike
Information Criterion (AIC, (34) Aho et al 2014). For each curve, the
accumulation model with the lowest AIC was selected. Accumulation curves
and associated asymptotes were generated using the vegan R package. To
estimate the sampling effort required to achieve a given proportion of
asymptotes, we considered the model selected for accumulation curves.
Then, we extracted the predicted number of samples producing a number of
taxa/OTUs that outreached 90% and 95% of the asymptotes.