A bird survey method for Australian tropical savannas

The tropical savanna of northern Australia is extensive and relatively homogenous compared to the open woodlands of temperate Australia. The avifauna of this biome is unevenly dispersed in the landscape. A standard count method for birds using a timed search along a 100-m transect with eight repeated counts per site over 4 days, has been used extensively over the past decade or more in these savannas, but its effectiveness has never been critically examined. We used data collected from across northern Queensland from 2004 to 2010, comprising 502 single-survey sites (each transect counted eight times) and 60 sites re-sampled four times from 2004 to 2008, to examine species records and accumulation of species over time with respect to time of day, increasing number of repeat counts(from 2 to 32 counts), species turnover, vegetation density effects and distance to first observation. Over 70% of the species in tropical savannas were easily observed and for 93% of the species vegetation structure did not alter detectability. Richness and abundance did not vary significantly across the day. We found that repeated sampling over multiple days, and at different times of the day, provides the best estimate of species richness at a site, and improved detectability