Comparing line transect and Indian White-eye mobbing call playback for detecting forest birds in subtropical Yunnan

Detection bias in avian monitoring is a critical constraint on the accurate assessment of community structure. This study presents a detailed field-based comparative case study between the playback of the Indian White-eye ( Zosterops palpebrosus ) mobbing calls—which documents only those bird species and individuals that approached and responded to the playback stimulus—and the conventional line transect method, which records all birds detected visually or acoustically along the transect. The study was conducted in subtropical semi-humid evergreen broad-leaved forests on the central Yunnan Plateau. We conducted 23 replicate surveys using both methods and employed multivariate regression trees to systematically evaluate the detection efficiency of the two methods and their dominant influencing factors. The results indicate the following: (1) The two methods exhibit functional complementarity: the line transect method shows significant advantages in total species richness (84 vs. 51 species) and the detection of larger-bodied birds (body mass ≥ 15 g), while the specific mobbing call playback is more efficient for small-bodied species (particularly the 0–10 g group) and insectivorous/nectarivorous functional guilds. The combined application of both methods can increase the species detection rate by up to 46.9%; (2) Detection efficiency varies seasonally, with significant differences between the two methods during the non-breeding season ( p < 0.05) but no significant difference in the breeding season, reflecting the impact of seasonal changes in avian behavior; (3) Regression tree analysis reveals a hierarchical decision pathway governing detection differences: body mass is the primary differentiating factor (with a threshold of 15 g), followed by habitat type, and finally trophic niche. Specifically, detection rate differences for larger-bodied birds (≥15 g) are more pronounced in shrub and coniferous forests, while small-bodied birds (<15 g) respond more actively to playback in forest habitats, with insectivorous small birds showing the most obvious response. The phylogenetic signal is extremely weak (Blomberg's K = 0.057), further indicating that functional traits rather than phylogenetic relationships dominate interspecific differences in responses to this specific mobbing call. This study provides an empirical case for targeted monitoring using specific mobbing calls and a basis for optimizing avian survey protocols combining multiple methods.