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

We compare methods routinely used to count birds in field surveys such as point counts and transects to see differences between them regarding bird grassland richness and abundance. Additionally we test the relationship between an environmental variable (vegetation height) and the parameters recorded by both census methods. From September 2014 to January 2015, we surveyed birds in 132 point counts and 129 transects spread across the southern Brazilian grasslands. We conducted one method in sequence of another, in the same place with the same observer and at the same weather conditions. We standardized data to eliminate effort bias caused by area covered and time in each method. We recorded 4,757 individuals of 188 species in point counts, and 4,440 of 177 species in transects. The two methods did not differ in abundance of birds recorded (p = 0.38), but they had a statistically significant difference in richness (p = 0.02). Both point counts and transects did show a weakly negative relationship of bird abundance with vegetation height (r2 = 0.03; p = 0.05 and r2 = 0.07; p = 0.004, respectively). For richness values both methods did not show relation with vegetation height (r2= 0.002; p = 0.63 to point counts and r2 = 0.02; p = 0.15 to transects). We found that for some parameters analyzed both point counts and transects are similar; however, to detect richness, point counts are more effective.