Survey completeness of a global citizen‐science database of bird occurrence

Measuring the completeness of survey inventories created by citizen‐science initiatives can identify the strengths and shortfalls in our knowledge of where species occur geographically. Here, we use occurrence information from eBird to measure the survey completeness of the world's birds in this database at three temporal resolutions and four spatial resolutions across the annual cycle during the period 2002 to 2018. Approximately 84% of the earth's terrestrial surface contained bird occurrence information with the greatest concentrations occurring in North America, Europe, India, Australia and New Zealand. The largest regions with low levels of survey completeness were located in central South America, northern and central Africa, and northern Asia. Across spatial and temporal resolutions, survey completeness in regions with occurrence information was 55–74% on average, with the highest values occurring at coarser temporal and coarser spatial resolutions and during spring migration within temperate and boreal regions. Across spatial and temporal resolutions, survey completeness exceeded 90% within ca 4–14% of the earth's terrestrial surface. Survey completeness increased globally from 2002 to 2018 across all months of the year at a rate of ca 3% yr –1 . The slowest gains occurred in Africa and in montane regions, and the most rapid gains occurred in India and in tropical forests after 2012. Thus, occurrence information from a global citizen‐science program for a charismatic and well‐studied taxon was geographically broad but contained heterogeneous patterns of survey completeness that were strongly influenced by temporal and especially spatial resolution. Our results identify regions where the application of additional effort would address current knowledge shortfalls, and regions where the maintenance of existing effort would benefit long‐term monitoring efforts. Our findings highlight the potential of citizen science initiatives to further our knowledge of where species occur across space and time, information whose applications under global change will likely increase.