Global patterns of species richness: Spatial models for conservation planning using bioindicator and precipitation data

We used birds, butterflies, tiger beetles, mean annual precipitation, and spatial statistical models to investigate the applicability of using indicators of species richness for conservation planning on a continental scale. The models were applied to data collected on three grids of squares (each square 275 or 350 km on a side) covering North America, the Indian subcontinent, and Australia. We applied spatial statistical modeling techniques to determine the viability of using a single or multiple indicators to predict spatial patterns of species diversity of ecologically and phylogenetically unrelated taxa. Spatial models are optimal for these analyses because species data typically are not spatially independent, primarily owing to dispersion effects. Furthermore, spatial models can be used to predict species numbers in areas where no observed data are available. We found that the number of tiger beetle species is a useful indicator of the number of butterfly species in North America and of the number of bird species on the Indian subcontinent, but it is not so useful as an indicator of either the number of bird or butterfly species in Australia or of the number or bird species in North America. We also found that the number of butterfly species is a useful indicator of the number of bird species in North American and Australia and that mean annual precipitation is useful for predicting the number of butterfly species in Australia. Although the general model used on all three continental areas is the same, the relative importance of potential indicators in predicting spatial patterns of other taxa changes from continent to continent. We attribute this change largely to differential biogeographical and ecological history, which must be taken into account in the selection and testing of potential indicators.