Abstract
Although community (or citizen) science has revolutionized biodiversity data collection and expanded its potential application, these datasets are commonly affected by bias. For instance, observers’ attention towards biodiversity is often led by the aesthetic and economic values of organisms, resulting in the under- and overrepresentation of species. Mammals in general are more conspicuous and charismatic than most other groups and therefore hold a unique appeal for observers that are likely to contribute to community-science platforms. Nevertheless, not all mammals are equally attractive to the human observer, and depending on their ecological and phenotypical traits, different species are represented in varying degrees in datasets collected by non-professional scientists. Herein, we assess the contribution of community science observations to global mammal occurrence data, examining how species traits influence the number of contributed observations. We compiled and analyzed spatiotemporal patterns in over 2 million observations globally from the iNaturalist platform. We found that large, crepuscular, and widely distributed species were overrepresented compared to smaller, diurnal species with a narrower distribution. Marine mammals represented 3.1% of species and 7.0% of observations. Nevertheless, the average number of observations per species was 1217.2 for marine species compared to 690.5 for terrestrial species. While bats and rodents were underrepresented, less diverse groups such as elephants and monotremes were overrepresented. Around 55% of mammal species are currently represented in the iNaturalist dataset, and our findings reveal biases linked to species traits, offering opportunities to increase the representation of currently underrepresented mammal species in global biodiversity datasets by adaptive sampling.