Doubling up: two-year eBird refresh adds 200 million more observations

Combined update from 2017-2018 nudges GBIF.org past 1.3 billion species occurrence records

Gavia-pacifica
Pacific loon (Gavia pacifica), Utqiagvik, North Slope, Alaska, USA. Photo © 2018 Ian Davies used by permission, via eBird, the Macaulay Library and GBIF.org.

The addition of two years of observations to the eBird Observational Dataset (EOD) has boosted the total number of records available in this dataset from the Cornell Lab of Ornithology (CLO) to 561 million occurrences.

The 55 per cent increase comes from eBird sightings gathered between 1 January 2017 and 31 December 2018, and the world’s largest biodiversity-related citizen science project now includes observations from more than 420,000 individual recorders.

The EOD is a FAIR and open public domain dataset derived from records in the eBird platform each year. Other, richer data products that include additional information associated with observations are available directly from eBird.

"The Cornell Lab of Ornithology has admired the spirit and role that GBIF has played in gathering, archiving, and making available biodiversity data since its inception." said Steve Kelling, Senior Director of Information Science. "We are proud of our contribution of eBird data to this incredibly valuable resource."

Changes in how CLO manages the EOD enabled publication of records from 2018 more quickly than in previous years, allowing for this simultaneous release of two years of observations. These adjustments will enable quicker turnaround in future updates, increasing the speed with which each year's eBird records are made freely and openly available through GBIF.org.

This update signals a consistent growth in observations across all regions as well as steadily increasing adoption of eBird outside North America.

Region 2017-2018 percentage increase Total number of records
Asia 54.7% 21,475,580
Europe & Central Asia 51,7% 23,145,273
Latin America & the Caribbean 45.0% 44,201,386
Africa 41.4% 5,477,267
Oceania 39.8% 18,462,408
North America 32.7% 449,359,703
Antarctica 31.1% 53,668

This latest update offers fresh perspective on longer-term trends in data growth for both eBird and GBIF. In August 2014, the addition of 50 million sightings pushed the EOD to 150 million records and boosted GBIF.org past the 500 million mark. Less than five years later, eBird itself surpasses GBIF's 2014 total number of records, while GBIF.org hosts 1.3 billion species occurrences from nearly 1,400 data-publishing institutions around the world.

eBird's 2018 year in review notes that "more than 100,000 people downloaded raw eBird data for analysis, and more than 3,400,000 people visited the eBird website in 2018". Both of these figures represent significant increases over 2017.

To these can be added users of GBIF.org, who requested more than 35,000 downloads containing records from eBird records in 2017 and 2018. The EOD was also the most frequently visited occurrence dataset page on GBIF.org during the same period, racking up more than 70,000 unique pageviews while edging out iNaturalist research-grade observations for the top spot.

Analysis of data citations in literature also emphasizes the added value of making eBird observations available via GBIF as well as through eBird’s own platform. Since the last refresh in late 2016, 105 peer-reviewed publications in 2017 and 2018 directly cited eBird as a source of data. During the same time, researchers cited downloads from GBIF.org incorporating EOD data in 26 peer-reviewed publications. While review is still underway, preliminary analysis suggest that 19 of these papers are additional to those captured in eBird's literature tracking. As more researchers follow GBIF citation guidelines by using Digital Object Identifiers (DOIs) issued with each download, the onward use of eBird data is becoming more evident, with 10 more uses cited so far in 2019.