A influência do delineamento amostral nas estimativas de riqueza e composição de espécies de anfíbios nas margens do alto Rio Madeira (Rondônia, Brasil)

Abstract

Estimates of biodiversity are of fundamental importance for conservation strategy planning. However, differences in sampling design may influence species richness and composition to a greater extent than the natural variability of biological diversity among sites. In this study we tested the effect of the variation in the components of sampling design on the estimates of species richness and composition using data derived from a sampling of amphibians over a 240,000 ha area along the margins of the upper Madeira River (Rondônia, Brasil). We simulated sampling designs with reduction in the number or size of sampling units, or in the number of sampling occasions, and evaluated the effect of geographical distance between sampling units on the estimates of species richness. Estimates of species richness and composition of amphibians were sensitive to the reduction in the sampling effort in these components. Reduction in transect number and size resulted in a respective decrease in 4- 66% and 8-45% of the number of species originally recorded, and in a 4-47% loss of information on species composition. Reduction in sampling occasions resulted in a 21-41% reduction in the number of recorded species, and 18-70% information loss on species composition. Dissimilarity indexes were high independently of the geographical distance between transects pairs. The analysis of results on 18 other amphibian species samplings in Rondônia showed no proportionality among estimates of species richness and the size of the inference area or number of sampling days, which was most likely owed to that sampling units were insufficient and or not homogeneously distributed over the inference area. More detailed comparisons among sampling designs of the 18 studies was hampered by the lack of information on most sampling components in almost all cases.