Print PageSpecies area relationship (SAR) is one of the most studied pattern in ecology. However, observed SAR is rarely contrasted against proper null-model. One possible reason is that SAR null-model may require detailed abundance data, which is unavailable for many taxa and in most spatial scales. We present a SAR null-model, which is based on random placement of occupancies. The parameters of the null-model describe the link function between the distribution of species from occupancy level 1 (species occurred on exactly 1 isolate) and isolate area. The link function allows for calculating the probability that a species from any given occupancy level is found in an isolate of a given area. We contrasted predictions of two link functions - a random placement link function and the observed link function -- with observed data in an information theory framework. We explored separately the prediction of the null-model for all species, for all species excluding those from occupancy level 1, for rare species and common species. We explored the null-model for 4 datasets.
Results/Conclusions
The four selected datasets represent four different general results. In one dataset, both rare species and all other species are best described by random placement of occupancies. In the second dataset, rare species follow a random placement of occupancies but other species differ from the null-model. In the third dataset, rare species differ from random placement of occupancies, but all other species are best described by the null-model. Finally, in the fourth dataset, both rare and all other species differ from random placement of occupancies. For each of the datasets, we provide possible mechanistic explanations for the emergent pattern that is relevant for that system. As far as we know, the random placement of occupancies SAR null-model, which does not require abundance data, is the only SAR null-model applicable to most datasets, for all taxa, at all spatial scales.
phone 202-833-8773