Numbers, distribution and behaviour of Australian Sarus Cranes Antigone antigone gillae and Brolgas A. rubicunda at wintering roosts on the Atherton Tablelands, far north Queensland, Australia

The Atherton Tablelands in far north Queensland is the only currently known concentrated flocking area for Australian Sarus Cranes Antigone antigone gillae. Brolgas A. rubicunda also flock there in the non-breeding season, offering a unique opportunity to survey numbers, distribution and roost-sites for both species and their interactions. We searched for dry-season roosts every year from 1997 to 2017, and conducted an annual volunteer count simultaneously at multiple sites from late afternoon to dark as cranes flew in to roost. From one to several thousand cranes winter annually on the Tablelands, with Sarus Cranes highly concentrated in the Atherton Tablelands Key Biodiversity Area (KBA) and Brolgas concentrated to the north-west and south-west of the KBA. No population trends were detected in the context of 21 years of highly variable annual counts, a substantial proportion of birds not identified to species in failing light, and a change in monthly timing of the annual count. Notwithstanding these caveats, we provide the first systematic minimum estimate of 826-3255 Australian Sarus Cranes wintering on the Tablelands, up to 19.5% of the global population. The corresponding estimate for Brolgas is up to 3469 individuals or 4.9% of the global population. These are likely to be underestimates, particularly for Sarus Cranes, which arrived at roosts on average much later than Brolgas, so were more likely to be unidentified. The species shared many roosts and, except at one large shared roost, family groups of the two species intermingled rather than occupying separate parts of the site, an uncommon relationship at mixed-species wintering crane roosts. Further study of these behavioural findings could extend understanding of sympatry in these closely related species. Conducting counts in a consistent month may improve trend detection within the limits imposed by strong fluctuations in annual numbers.