In the world there could be about 50,000 million birds , about six birds per human being, which would belong to 9,700 different species, including flightless birds such as emus and penguins.
This is what a new algorithm has collected, trained with big data extracted from citizen science observations, developed by researchers at the University of New South Wales. This is thus the first comprehensive effort to count a set of other species .
There is a need to continue refining the estimates
The big data for training the algorithm comes from nearly a billion bird sightings recorded in eBird, an online database of bird observations made by citizen scientists. This calculation took into account the "detectability" of each species, that is, the probability that a person had seen that bird and submitted the sighting to eBird . Detectability can include factors such as their size, color, whether they fly in flocks, and whether they live near cities.
Only four species of birds belonged to what researchers call the billion club – species with an estimated global population of more than a billion. The house sparrow (1.6 billion) heads this unique group, which also includes the European starling (1.3 billion), the ring-billed gull (1.2 billion) and the barn swallow (1.1 billion). And about 12% of the bird species included in the study have an estimated global population of less than 5,000. These include species such as the Chinese crested tern, the water blackbird and the invisible rail.
As the study’s lead author Corey Callaghan , who conducted the research while he was a postdoctoral researcher at UNSW Science, explains:
Although this study focuses on birds, our large-scale data integration method could serve as a model for estimating the specific abundance of other groups of animals. Quantifying the abundance of a species is a crucial first step for conservation. By properly counting what’s out there, we learn which species may be vulnerable and we can follow how these patterns change over time – in other words, we can better understand our baselines.
Although the research team is confident in their estimates, they recognize that some degree of uncertainty is unavoidable when working with large data sets like this one. For example, people documenting sightings may be more likely to search for rare species . So the research team plans to repeat their analysis as more data becomes available.