A comparison of strip transects and fixed-radius point counts for a bird community in a semiarid chaco forest

Abstract

We assessed the performance of strip transects and fixed-radius point counts with respect to their accuracy to estimate bird abundances and cornmunity attributes (diversity, richness, evenness) in the semiarid forests of Santiago del Estero province, Argentina. The study was carried out in a secondary forest and in a sector of the same forest where shrubs were removed. We installed 4 transects and 4 points in each habitat. Our results showed that: 1) estimations of species richness by the point method were lower than by the transect method (P<0.05), even though there were not significant differences in either diversity or evenness (P=0.25 and P=0.56, respectively); 2) there was a positive correlation between bird abundances estimated by both methods in each habitat (P<0.01), and 3) there were not significant differences between methods in their accuracy to estimate bird abundances, recorded by means of the coefficient of variation (CV%). We conclude that, even though both methods showed similar performances, the point method has in our case some advantages over the transects, since they are less difficult to allocate in the study area and do not demand to keep a constant pace in a rugged and scrubby terrain, as the transect method does. The lower species richness detected by the point method could be solved by increasing replicates, as well as by counting birds which are out of the fixed radius.