Using citizen science to identify Australia's least known birds and inform conservation action

Citizen science is a popular approach to biodiversity surveying, whereby data that are collected by volunteer naturalists may help analysts to understand the distribution and abundance of wild organisms. In Australia, birdwatchers have contributed to two major citizen science programs, eBird (run by the Cornell Lab of Ornithology) and Birdata (run by Birdlife Australia), which collectively hold more than 42 million records of wild birds from across the country. However, these records are not evenly distributed across space, time, or taxonomy, with particularly significant variation in the number of records of each species in these datasets. In this paper, we explore this variation and seek to determine which Australian bird species are least known as determined by rates of citizen science survey detections. We achieve this by comparing the rates of survey effort and species detection across each Australian bird species' range, assigning all 581 species to one of the four groups depending on their rates of survey effort and species observation. We classify 56 species into a group considered the most poorly recorded despite extensive survey effort, with Coxen's Fig Parrot Cyclopsitta coxeni, Letter-winged Kite Elanus scriptus, Night Parrot Pezoporus occidentalis, Buff-breasted Buttonquail Turnix olivii and Red-chested Buttonquail Turnix pyrrhothorax having the very lowest numbers of records. Our analyses provide a framework to identify species that are poorly represented in citizen science datasets. We explore the reasons behind why they may be poorly represented and suggest ways in which targeted approaches may be able to help fill in the gaps.