Species richness correlations between taxa calculated from species-area relationships

Predicting the species richness of certain taxa from the richness of others could be of great value in biodiversity conservation. If we know how the species richness of groups of organisms correlate to each other under different circumstances, inventories of only a few well-known taxa could be enough to establish the biodiversity values of different areas. In this master thesis I have examined the richness correlations between taxa from species-area relationships extracted from the literature to evaluate the possibilities of this approach. Calculations of the pairwise correlation between the residuals of the species-area relationships of different taxa found in the literature resulted in 503 correlation coefficients. Of the 79 taxa covered, the most frequently examined groups (several taxa included) were plants > butterflies > true bugs > birds > mammals. Only 157 of the 503 correlations were significant. An average effect size of 0.373 (95 % CL = 0.059) indicates positive but usually weak correlations between taxa in the very heterogeneous dataset. The dataset was divided into a number of categories (according to spatial scale, taxonomic distance, trophic position, biotope and climatic region) to investigate some hypotheses, for example (1) that richness of closely related taxa are likely to be more correlated than distantly related taxa, and (2) that correlations between plants and animals are higher than correlations among plants or among animal groups. Between families – within order comparisons had significantly higher correlation than comparisons in all other taxonomical classes except between kingdoms. These results were in agreement with my hypotheses concerning taxonomic distance and trophic position. No universal indicator taxon has been found to predict the species richness of other taxa. With more careful studies in certain biotopes and at appropriate spatial scale perhaps good indicator groups can be discovered. These are unlikely to be universal and it is probable that indicator taxa ...