Doraditos and Tyrannulets: cryptic species challenge citizen science in South America

Errors resulting from misidentified species can be a major issue for citizen science because many cryptic, or similar species, can be confused. Despite having expert-adjusted filters to prevent errors, these filters fail to recognize the certainty of reports of cryptic species that are often frequent and sympatric. To test the uncertainty level of such records, we compiled data from citizen science platforms of 5 cryptic species of the genera Serpophaga and Pseudocolopteryx. We categorized each record as certain or uncertain according to the documentation submitted, the observer’s experience, and the location and date of the sighting. We performed two MaxEnt models for each species, the first including uncertain records, and the second model only with certain records, and then we evaluated the variation observed between both models. The uncertainty was higher in P. flaviventris (98.4%) and S. subcristata (53.4%), and lower in S. munda (23.6%) and P. citreola (1%). We did not find any misidentified documented records for S. griseicapilla or P. citreola. The models that presented the greatest difference in their distribution areas were from P. flaviventris (31% difference), and S. subcristata (14%); the models of P. citreola were almost identical. To reduce the degree of uncertainty in reports of cryptic species, we suggest training regional reviewers to detect uncertain records, increasing the accuracy of automated filters or creating particular protocols after reporting a cryptic species, and encouraging observers to report species with adequate evidence for identification.