Selective observation causes differences in citizen science butterfly data

In times of insect decline, the need for biodiversity monitoring data has become increasingly urgent. However, standardised monitoring of biodiversity is time-consuming and cost-intensive. Citizen science (CS) initiatives therefore may provide valuable data and may complement data collected by professionals. Photo-apps equipped with automated taxonomic identification based on artificial intelligence play a central role in CS, at least for well distinguishable organisms such as the majority of butterfly species. In this study, we analysed butterfly (Papilionoidea) observations collected with three different photo-apps (i.e. Blühendes Österreich, iNaturalist, observation.org ). We compared these data with observations from the Global Biodiversity Information Facility (GBIF). For this purpose, we classified each butterfly species according to its detectability and attractiveness, as well as its ecology and behaviour. Our results show that the observations obtained from the three photo-apps mainly cover mobile, conspicuous and easy-to-identify species, while the rare and sedentary specialist species and species that are difficult to distinguish from other taxa are underrepresented. Our findings show that the observations collected differ significantly between the three apps, and Blühendes Österreich particularly lacks inconspicuous butterfly species. However, a detailed regional analysis of user performance revealed that the differences among the three apps largely stem from Austrian-wide difference in app usage and less to user specific biases in recording. Within single habitats, amateurs and semi-professional users performed similarly in recording. In consequence, a combination of the data from the various apps might provide a largely realistic picture. However, most rare and ecologically demanding species seem to be covered inadequately. Thus, the recording of such species must be continued by experts to obtain a comprehensive picture of regional biodiversity.