Bryophytes of Jau National Park (Amazonas, Brazil): Estimating species detectability and richness in a lowland Amazonian megareserve

摘要

In the past decade, floristic studies have rebounded as checklists are fundamental for executing meta-analyses which address ecological, biogeographic and evolutionary questions of broad geographic scope. Despite the importance of checklists as baseline records of local diversity and distributions, few attempts have been made to quantify sampling effort and species detectability within and among study sites. Quantitative floristics, which combines the use of checklists with statistical methods for estimating local richness, is a promising method for characterizing the completeness of checklists especially for cryptic components of biodiversity. For bryophytes, quantifying levels of detectability among substrate types is of central importance, especially in tropical forests where much of their diversity is harbored in difficult to access habitats such as the tree canopy. In light of the need to establish quantifiable protocols of detectability in poorly studied tropical regions, we present a bryophyte checklist for the Jai National Park (JNP), located in the heart of the Amazon, and estimate local species richness and detectability as it relates to five substrate types (epiphytes, epiphylls, epixylic, epipetric and soil). Identifications from 712 collections made during four excursions over the past decade to JNP revealed 150 species consisting of two new country records and five new state records, along with 20 rarely collected Amazonian endemics. Despite our intensive sampling, which included systematic canopy collections during one of the excursions, Chao richness index estimated that ca. 46 species (nearly one-third of those presently observed) remain undetected from JNP. Furthermore, levels of detectability among substrates varied widely, where observed epiphyte richness, in contrast to the other substrates types, most closely approximated the estimates. Our results illustrate the need for quantitative richness estimates as a means to increase the accuracy of checklist data, particularly when used in meta-analyses addressing global-scale questions.