Fitting species-accumulation functions and assessing regional land use impacts on avian diversity

Abstract

As one samples species from a particular assemblage, the initial rapid rate with which new species are encountered declines with increasing effort. Nine candidate models to characterize species-accumulation functions were compared in a search for a model that consistently fit geographically extensive avian survey data from a wide range of environmental conditions. Landscape-specific species-accumulation curves generated under a bootstrap resampling plan were best described by a generalized Weibull cumulative distribution function. Traditional species-area models of cumulative species richness as a function of accumulated sample had notable functional bias. The Weibull model fitted species-accumulation data equally well among sixty-six forested landscapes in the eastern U.S. Landscapes with a greater proportion of agricultural and urban land uses accumulated species more slowly than landscapes which retained a greater proportion of natural habitats (r = -0.64, P<0.001). This finding supports predictions of ecosystem behavior under human land use. There was no evidence that intermediate levels of land use intensity maximized accumulation rates. The approach reviewed in this paper makes no assumptions about the form of the species-abundance distribution or how species are distributed in space, thereby offering some advantages over more conventional diversity indices for characterizing how species assemblages respond to anthropogenic disturbance. Investigation of how species accumulation varies over time in a given geographic area is needed to evaluate fully the potential application of this approach to regional land use planning.