摘要The potential of citizen science projects in research has been increasingly acknowledged,but the substantial engagement of these projects is restricted by the quality of citizen science data.Based on the largest emerging citizen science project in the country—Birdreport Online Database(BOD),we examined the biases of bird-watching data from the Greater Bay Area of China.The results show that the sampling effort is disparate among land cover types due to contributors'preference towards urban and suburban areas,indicating the environment suitable for species existence could be underrepresented in the BOD data.We tested the contributors'skill of species identification via a questionnaire targeting the citizen birders in the Greater Bay Area.The questionnaire show that most citizen birdwatchers could correctly identify the common species widely distributed in Southern China and the less common species with conspicuous morphological characteristics,while failed to identify the species from Alaudidae,Caprimulgidae,Emberizidae,Phylloscopidae,Scolopacidae and Scotocercidae.With a study example,we demonstrate that spatially clustered birdwatching visits can cause underestimation of species richness in insufficiently sampled areas;and the result of species richness mapping is sensitive to the contributors'skill of identifying bird species.Our results address how avian research can be influenced by the reliability of citizen science data in a region of generally high accessibility,and highlight the necessity of pre-analysis scrutiny on data reliability regarding to research aims at all spatial and temporal scales.To improve the data quality,we suggest to equip the data collection frame of BOD with a flexible filter for bird abundance,and questionnaires that collect information related to contributors'bird identification skill.Statistic modelling approaches are encouraged to apply for correcting the bias of sampling effort.