Historic data provide context for the future

By Team eBird August 15, 2016

Henslow's Sparrow by Luke Seitz/Macaulay Library

Every day, thousands of researchers all around the world go out to collect data—helping inform studies that range from short-term graduate research to long-term ecological monitoring. Many of these data go back for decades, or provide snapshots into the status of bird distribution and abundance of a past age. These historic data are incredible valuable, since they provide the context for our understanding of current bird populations—allowing us to understand how a changing world might be affecting the natural ecosystems of our planet. Brett Sandercock, Professor of Wildlife Biology at Kansas State University, recently uploaded more than a thousand historical checklists from Konza Prairie in Kansas, and has kindly written a short piece on the process, and how you easily do the same! Thanks Brett for your contributions, and for this great article. If you have a similar dataset, or even a few notebooks in the attic from times gone by, you can help paint a more complete picture of this changing world—read on to learn how.

Why are historic data of value?  Until recently, the best data on bird populations in North America were from the Breeding Bird Survey program and Christmas Bird Counts – more than a half century of bird records, but only from early summer and mid-winter.  Birders contributing checklists to eBird now provide bird sightings year-round and at a global scale, which has allowed researchers at the Cornell Lab of Ornithology to build exciting new maps for the movements of migratory birds [Team eBird note: stay tuned for more maps in the near future—featuring models of bird abundance at a continental scale!].  eBird was launched in 2002 and most checklists have been collected during the past decade.  Historic data are valuable because they could be used to explore long-term changes in bird populations.  Any historic data that you have can be easily added to this global dataset – learn how here.

Konza Prairie Biological Station is a tallgrass prairie site that is part of a network of Long-term Ecological Research sites funded by the National Science Foundation.  Systematic bird surveys were conducted at Konza Prairie by a series of different observers for a 36-year period from 1981 to 2016.  It seems unlikely that the original investigators could have anticipated how long-term data might be used in the future!  When our bird surveys started in 1981, the Internet was still a university research project and IBM had just released the first personal computer with DOS software from Microsoft.  Investigators at Konza Prairie had the foresight to design surveys based on line transect sampling that collected data on bird species with counts of individuals, along with information on locality, date, and time of surveys.  Our systematic line transects met all of the required criteria for bird sightings collected under the eBird protocol for a Traveling checklist.

Konza Prairie; a truly breathtaking landscape. Photo by Judd Patterson.

Konza Prairie; a truly breathtaking landscape. Photo by Judd Patterson/Birds In Focus.

Importing a large historic dataset into eBird takes a little preparation and here are the main lessons learned:

  • Data can be uploaded in two formats: in “eBird checklist format” (daily accounts) or as “eBird record format” (standard database rows). Learn more.
  • If bird data are uploaded in the eBird record format, the information on survey effort are repeated on each line for every species that was detected. A database trick is to have a variable with a unique name for every checklist, use the unique name to merge the bird records with the survey effort, and then delete the variable before uploading the data to eBird.
  • Bird names must match the standardized common names used in the eBird Taxonomy. Names that fail to match are usually cases of incomplete identification (e.g., “sparrow sp.”) or where bird taxonomy has changed (e.g., “Spotted/Eastern Towhee (Rufous-sided Towhee)”). Helpfully, once you’ve matched a name once in eBird, all future ‘incorrect’ entries with that name will remember that matching.
  • The Date and Start Time need special attention and must be formatted as Date/Time fields, even though the import file is a comma-delimited file (CSV).
  • The observer for the checklist is taken from the name of the eBird account, but the name of the original observer and other details can be included in the Comments field. [Team eBird note: you can also share checklists with the original observers if they have an eBird account, allowing you to maintain the master data set in one account, while ensuring that eBirders get the sightings properly attributed where appropriate]
  • eBird can accept zero-taxon lists where a survey was conducted but no birds were detected. Checklists with no detections must be uploaded in the eBird checklist format.
  • Maximum file size is 1 MB and must be saved as a comma-delimited file (CSV). The CSV files can be larger than the same data in an Access or Excel file.

What does 36-years of systematic bird surveys look like in eBird?  The long-term sampling of the Konza Prairie LTER program resulted in 1,428 checklists collected by five different observers with a grand total of 44,314 individuals of 134 species of birds.  Our historic dataset currently provides First Seen records for about 60-75 bird species in Riley and Geary Counties of Kansas.  Perhaps because the records were based on systematic line transects, the dataset includes relatively few birds that would be flagged as Rare in eBird – a Scarlet Tanager in June 1983, a Pine Warbler in April 1986, and a Golden Eagle in January 1998.  Instead, the most common species in the dataset are birds of the tallgrass prairie: Greater Prairie-Chickens, Upland Sandpipers, Dickcissels, Grasshopper and Henslow’s Sparrows, and Brown-headed Cowbirds, sometimes even on a single line transect.

Greater Prairie-Chickens photographed by Brett Sandercock at Konza Prairie

Greater Prairie-Chickens at Konza Prairie. Photo by Brett Sandercock.

Uploading historic data into eBird serves multiple goals: the information is archived in a stable, long-term repository, all records are publicly available to the birding community, and historic checklists can now be used to model long-term changes in bird abundance or distributions.  Many individuals and programs have valuable historic records that could be suitable for eBird – consider uploading your long-term data too!

Brett K. Sandercock, Kansas State University