Elevational zonation of afrotropical forest bird communities along a homogeneous forest gradient

Abstract

AbstractAim This study analyses the distribution and abundance of birds from a forested tropical gradient in order to determine whether elevationally distinct communities are detectable in this habitat.Location An avifaunal census was carried out on a single transect within the tropical forest of the Udzungwa Mountains in the Eastern Arc, Tanzania, covering a range in elevation from 300 to 1850 m.Methods Two complementary data sets on forest birds were analysed, encompassing (1) data derived from standardized 20‐ha spot‐mapping censuses performed at nine elevations over 175‐m intervals from 400 to 1800 m a.s.l., and (2) all observations of birds binned into 32 data points at 50‐m intervals, from 300 to 1850 m a.s.l. The degree of zonation in the avian community along the elevational gradient was examined using the chronological clustering method, an agglomerative hierarchical clustering method that can be carried out with a range of similarity indices.Results The chronological clustering analysis of the data set based on standardized spot‐mapping revealed a clearly defined boundary at c. 1200 m a.s.l., separating lowland from montane communities. Most bird species could be categorized as belonging to one of these two communities. The data set based on all observations revealed a number of potential secondary boundaries, although these boundaries delimited the entire elevational ranges of individual species in only relatively few cases.Main conclusions In contrast to previously published studies, we find evidence of an elevational zonation of distinct communities within a seemingly homogeneous habitat. Although similar boundaries have been assumed to arise as a result of vegetational ecotones, or because of interspecific competition, these mechanisms are poorly corroborated. We suggest that the causes of patterns of zonation are not well understood, and that the interplay among species distributions, species richness, and environmental factors needs more consideration. The chronological clustering method is proposed as an appropriate tool for studying these specific patterns.