NO BIRD DATABASE IS PERFECT: CITIZEN SCIENCE AND PROFESSIONAL DATASETS CONTAIN DIFFERENT AND COMPLEMENTARY BIODIVERSITY INFORMATION

Citizen science has become a powerful tool for collecting big data on biodiversity. However, concerns have been raised about potential biases in these new datasets. We aimed to test whether citizen science bird databases have more biases than professional scientific databases. Our hypotheses were 1) citizen science databases will have more data on "easy to spot" species, that are widely distributed and have large body sizes; whereas 2) professional databases will have more endangered species and species of special interest for research. We analysed six Spanish bird databases: three professional, two citizen science and one mixed database. Our results show that, in general, occurrences in citizen science databases are better explained by the studied variables than professional databases, but no clear differences were found when analysed individually. Both citizen science and professional databases contain invaluable information on biodiversity but every database comes with a particular history and its stored data is the result of years of field sampling with heterogeneous goals, sampling methods and sampling effort. Consequently, raw observations should not be used directly as an ideal survey of the distribution or abundance of birds. We need to uncover these biases and develop new methods to properly incorporate the extensive and heterogeneous biodiversity data that is readily available to research.