On-line version ISSN 1667-782X
BEAUMONT FANTOZZI, María Del Mar et al. Effect of size and shape of the sampling unit on the estimation of seed-carrying ants' species richness in the Monte desert, Argentina. Ecol. austral [online]. 2011, vol.21, n.1, pp. 111-119. ISSN 1667-782X.
Species richness is one of the fundamental components for the structure and function of a community. The number of species recorded in a richness assessment depends not only on the characteristics of the focal community but on the size of the sampling area and on sampling effort. The goal of the present study was to establish the optimal size and shape of the sampling unit required to obtain a reliable estimation of the species richness of seed-carrying ants in the central Monte desert, using two simple methods. Ants in two different habitats within the Man and the Biosphere Reserve of Ñacuñán, Mendoza, Argentina, were sampled with 5 m-apart seed baits arranged in three 8x10 grids (2000 m2). After an iterative subsampling protocol, ant species richness was calculated for all possible subgrids of different size (1-80 baits) and shape (from square to elongated), keeping the relative spatial position of the baits. The minimum sampling effort required was evaluated in a graph of mean observed richness in each sub-grid type against sub-grid size (sampling effort). Total species richness in the community (asymptote estimation) was assessed graphically and by fitting three asymptotic models of richness increasing with sampling effort. Different minimum sample sizes were necessary to account for 85% of total ant species richness in grazed and ungrazed landscapes in the Monte desert: 25% of the original sample size (the usual size to estimate the richness of granivorous ants in arid-semiarid areas) would be enough in grazed areas, and >75% in the ungrazed "algarrobal". Therefore, habitat structure appears to influence the relationship between richness estimation and sampling effort. There were no detectable differences among sub-grid shapes: for a given sampling area, thin rectangular grids seem to detect a similar number of ant species than square ones. The complementary use of graphics and asymptotic models showed to be convenient to detect the optimal sampling effort to estimate ant species richness in different habitats of the Monte desert.
Keywords : Sample size; Sampling effort; Sub-sampling; Asymptotic models; Biodiversity.