摘要：
Aims Knowledge of species richness is of prime importance to both basic and applied aspects of ecological studies. However, quantifying plant species richness in the tropics is potentially time-consuming because of high species diversity. Plant species richness estimates are also frequently biased, because many rare species are not detected. To address these problems, the use of a variable area method has been proposed as an alternative to fixed area methods, but its applicability to plants with aggregated distributions has been questioned based on simulation studies. We use empirical data to compare the efficiency and accuracy of a variable area method and a fixed area method for estimating species richness, density and basal area for plants with aggregated distributions, using palms as a model taxon.
Methods Adult palms were sampled in twenty 10×30 m transects in an Atlantic Forest in Rio de Janeiro state, Brazil. All individuals were considered in the fixed area method, while in the variable area method only the first six adults in each transect were sampled; in this case, transect length was defined according to the distance of the sixth adult from the beginning of the transect. When fewer than six individuals were observed in a given transect, transect length was extended up to 50 m to search for additional individuals. The efficiency of both methods was compared based on species rarefaction curves, using the Chao 1 statistic (for abundance data). For each species we calculated mean density per transect and basal area, according to each sampling method. Sampling effort in terms of the number of individual plants and the area necessary to characterize maximum species richness in each sampling method, as well as mean time taken to sample a single transect, were compared as measures of efficiency.
Important findings An accurate estimate of species richness was achieved using both methods, but in the variable area method, a quarter of the number of individuals and half the area was sufficient to characterize maximum species richness. Density and basal area did not differ between methods for any of the species studied. In the fixed area method sampling effort was 90min per transect, whereas in the variable area method it was 30min per transect. The variable area method, with its faster assessment of palm species richness, should facilitate greater spatial representativeness by making it easier to sample a larger number of plots at different spatial scales. We thus find sufficient evidence to recommend the variable area method for rapid and robust evaluations of species richness for palms with aggregate distributions, as well as for other plants with similar spatial patterns, in tropical forests.
