eBird is a global monitoring project that gathers standardized bird occurrence records from a network of 200,000 participants who have volunteered almost 15 million hours. Presently, the eBird database holds more than 200 million observations of more than 10,000 species recorded at 2.1 million locations from every country in the world. eBird has been growing exponentially and we anticipate exceeding 1 billion observations by the end of the decade.

Since its inception in 2002 eBird has provided open access to all data, and presently contributes almost one-third of GBIF’s data holdings. Demand for eBird data has required the creation of multiple (3) dataset products each fully versioned and updated quarterly to annually. Tools to access eBird data by either species or region are available through the eBird website, with over 2,000 data requests over the past 12 months. A recent survey of eBird data users has shown that data requests come primarily from academics, governments, non-governmental agencies, and the birding community. Since inception more than 100 peer reviewed publications have used eBird data.

Data quality is a major challenge in eBird and is addressed both during data collection as well as during the entire data analysis workflow. eBird is based on a proven data architecture and infrastructure to support data quality and analysis. More than 8,000 data filters provide lists of likely birds to occur in a region for any day of the year, and more than 800 regional bird occurrence experts edit data submissions. Each year almost 5% (in 2013 2.4 million records) are reviewed. Additionally, analysis is underway on individual eBird contributors to identify the quality of their observations based on their historic data submissions.

eBird data are linked to multiple land-cover, climate, and human population data sources. While data interoperability is an enormous challenge, linking bird occurrences with the covariates obtained from earth observation data and other covariates strengthen the predictive performance of any analysis of eBird data.

A significant component of eBird is data analysis. Focusing on a macro-level perspective researchers have developed novel semi-parametric species distribution models that estimate patterns of occurrence and abundance of bird populations across their entire life history. Dynamic occurrence maps provide the opportunity for data-driven exploratory analysis of bird populations at the hemisphere scale at weekly intervals. These analyses have led to hypothesis-driven tests of eBird data with studies that explored migration drivers and patterns, the impact of climate change, and specie’s evolution.

eBird data and models have been used in a variety of on-the-ground conservation efforts. While eBird analysis has focused on broad spatial and temporal scales, detailed spatial (1 KM) and temporal (daily) drive the model results. These high resolution models have been used for the United States Department of Interior’s State of the Birds Report that highlighted the dependence of bird populations on both Public 2011 and Private 2013 lands. Additionally eBird data are spearheading new opportunities for dynamic land conservation—ensuring that lands are available for migrating or overwintering bird populations when needed.