The effectiveness of point counts and mist-netting in surveying Afrotemperate forest bird community structure in South Africa

Point counts and mist-netting are two frequently employed survey techniques used in assessing forest avian communities, although the reliability of these methods varies according to species composition and habitat. This study investigates how effectively these two methods survey forest bird community structures within South African Afrotemperate forests. Seven forests within the Eastern Cape were surveyed from 140 duplicate point count and 63 mist-netting stations. Both methods were compared for assessing species richness, as determined from bird atlas data. Generalised linear mixed-effect modelling was used to determine functional traits which most impacted species detection, and to identify detection biases for both methods. Both methods compared consistently across the seven forests, which shared similar community structure. Point counts detected 79.2% of the total diversity versus 41.0% using mist-netting, and mixed-effects modelling corroborated that species detection is more effective using point counts. All functional traits tested (body size, primary foraging stratum, feeding guild, habitat specialisation, and dispersal behaviour) affected detection outcome. Point counts better represented all aspects forest bird community structure, including mid- and understorey birds which are presumed to be better detected by mist-netting. Use of mist-netting only slightly enhanced diversity assessments, and combined survey efforts under-represented forest-edge foragers, woodland and grassland habitat generalists (~63.6% total diversity), large birds, Palaearctic migrants, and carnivores.