Abiotic Factors as Predictors of Distribution in Southern African Bulbuls

Abstract

The closely related Black-eyed Bulbul (Pycnonotus barbatus), Red-eyed Bulbul (P nigricans), and Cape Bulbul (P capensis) occupy parapatric to locally sympatric ranges within southern Africa. We used a multivariate discriminant function analysis to relate the South African distribution of these species at the resolution of a quarter-degree square of latitude and longitude to a suite of six environmental variables extracted from digital mod- els. The analysis correctly classified the distribution of these species for 89% of the 1,426 squares analyzed. Separation on the first discriminant function, which successfully charac- terized the distribution of R nigricans, was mainly by the coefficient of variation in mean annual rainfall and mean July minimum temperature. Separation on the second discriminant function, which successfully characterized R capensis, was mainly by the normalized dif- ference green vegetation index (derived from NOAA satellite scans). Using the standard- ized coefficients for the first discriminant function, a model generated to predict the distri- bution of R nigricans at a finer geographic resolution within our Eastern Cape Province study area was 93% successful. The results suggest that the three species occupy distinctly differ- ent habitats, as characterized by the suite of six environmental variables that we analyzed.