Detecting Density-Dependent Habitat Selection

Publisher Website
Google Scholar
Abstract
Simpson''s Index is frequently used in ecology to express the number and evenness of groups. In order to apply it to the problem of habitat selection, a transformation is derived such that if a species randomly uses all grids at a site, the expected value of the transform will be zero. Moreover, if the transform is plotted against the density of a species minus one, all selective strategies will be straight lines intersecting the origin with slope proportional to selectivity. The transform is mNy - N - m + 1, where N is the total census in all grids, m is the number of grids per site, and y is Simpson''s Index, .SIGMA. ni2 N2 (ni is the number in grid i). Data plotted on these coordinates can be analyzed for selectivity and density dependence using polynomial regression. The transformation was used to analyze the habitat selectivity of six species of rodents in Israel. All three of the psammophiles exhibited habitat selection that decayed after their density had exceeded a certain value. The one most abundant species exhibited a characteristic, almost random distribution at very high densities. The three lithophiles displayed three different patterns: one random; another, density-independent selection; and the third, hyperdispersion possibly caused by the species'' social structure and territoriality.