Bird surveys in grasslands: do different count methods present distinct results?

We compared two methods routinely used to conduct bird community surveys: point counts and transects. Our aim was to look for differences between these two methods regarding detection of bird richness and abundances. Additionally, we analyzed if one of the methods provided higher correlation of bird data with vegetation structure as an important habitat descriptor. From September 2014 to January 2015, we surveyed birds in 264 point counts and 258 transectsspread across the southern Brazilian grasslands. We conducted one method in direct sequence of the other, in the same place with the same observers and at the same weather conditions. We standardized data to eliminate the effort bias caused by area covered and time employed in each method. Total abundance of birds recorded by the two methods did not differ (point counts 4753 and transects4436, P = 0.31), but we found a significant difference in species richness (point counts 187 and transects173, P = 0.01). Abundance of birds sampled with the transectmethod showed a slightly higher correlation with vegetation height (r 2 = 0.07; P = 0.004) than the point counts method (r 2 = 0.03; P = 0.05). While results from both methods were similar, richness detection was more effective in point counts, indicating that this method might be more useful than it currently is. We discuss potential factors that may influence effectiveness of both methods and suggest issues that should be addressed in further research in order to develop standardized sampling methods for bird communities.